Datasets:

alon-albalak

/

licensed_pubmed

like 0

PMC3052533

The family *Bunyaviridae* of segmented negative-strand (SNS) RNA viruses contains more than 350 members categorized into five genera: *Orthobunyavirus, Hantavirus, Nairovirus, Phlebovirus* and *Tospovirus*. The prototype of the genus *Orthobunyavirus* is Bunyamwera virus (BUNV), which serves as a model for studying important pathogens within this family. As with all bunyaviruses, the BUNV genome comprises three segments, named small (S), medium (M) and large (L) ([@r9]). These strands are the templates for two fundamentally distinct processes, namely RNA replication and mRNA transcription. In common with other negative-stranded RNA viruses, the active template for these processes is not naked RNA, but is instead the RNA genome bound with virally encoded nucleocapsid (N) protein to form a ribonucleoprotein (RNP) complex. N protein remains associated with replication products throughout the infectious cycle, and thus the N--RNA interaction is critical for virus viability. Several studies have investigated the specificity of this interaction for bunyaviruses. They indicate that, while there is no obligatory encapsidation sequence, there is evidence of preferred binding for viral sequences ([@r14]; [@r15]; [@r17]; [@r18]). It has been proposed that this property may allow selective encapsidation of viral rather than host cell RNAs, and thus selective packaging of only viral RNPs into progeny virions. In contrast, molecular details of the role of the N protein in N--RNA interactions are lacking currently. Here, we have analysed for the first time the roles of individual BUNV N-protein amino acids in binding the RNA template. In addition, we have examined how these roles relate to both RNP assembly and to formation of the active template for both mRNA transcription and RNA replication. Previous structural studies with human respiratory syncytial virus, rabies virus and vesicular stomatitis virus (VSV) show that the mechanism of RNA binding within the non-segmented negative-strand (NNS) RNA-virus RNP relies on multiple interactions, mediated by a variety of residues within specialized RNA-binding grooves ([@r1]; [@r11]; [@r22]). In all three RNPs, positively charged amino acid side-chains interact with the RNA--phosphate backbone. Similar involvement of charged residues in RNA binding has also been described for the influenza virus nucleoprotein, and the crystal structure reveals a potential RNA-binding groove lined with conserved positively charged residues ([@r10]; [@r24]). Because of the important role of positively charged amino acids in RNA binding in the context of both viral RNPs and also in general ([@r23]), in this current study we focused exclusively on these residues. The BUNV N protein contains 31 positively charged amino acids. To identify those likely to possess critical functions, we aligned N proteins from eight orthobunyaviruses (Supplementary Fig. S1, available in JGV Online). We selected 15 of these that were best conserved in terms of identity or positive charge for further analysis. We analysed the role of these conserved residues in RNA binding by exploiting our previously described ability to express and purify soluble N protein from bacterial cells as an N--RNA complex ([@r21]). When purified, the BUNV N protein formed homogeneous and stable soluble assemblies in which N-protein tetramers bound 48 nt sections of bacterial-cell RNA, forming model RNPs ([@r15]). The establishment of this N--RNA assembly system allowed us to introduce mutations into model RNPs and assess how these changes affected RNA-binding ability. Previously described plasmid pET-BUN-N ([@r21]) was modified by using PCR-mediated site-directed mutagenesis to encode 15 N-protein variants incorporating single alanine substitutions at each of the 15 conserved positively charged residues (Supplementary Fig. S1). Briefly, these altered N proteins were expressed in bacteria as N-terminally histidine-tagged fusion proteins and then purified by using affinity chromatography. To test whether the N-protein alterations affected expression, stability and folding, purified N proteins were analysed by using circular dichroism (CD) and SDS-PAGE at various time points post-expression (data not shown). This analysis showed that all altered proteins were expressed at an abundance indistinguishable from wild type. In addition, with the exception of K166A, all proteins were correctly folded and highly stable. In contrast, K166A displayed a somewhat-disordered CD spectrum, and, following several days of incubation at 4 °C, was found to dissociate into two polypeptide fragments at the approximate point of alteration. To allow the functional comparison of all altered proteins, analysis of RNA binding and oligomerization was performed immediately following purification. RNA was harvested from RNPs by phenol--chloroform and chloroform extractions from an equal mass of protein (Fig. [1a](#f1){ref-type="fig"}), which was then visualized by native agarose-gel electrophoresis (Fig. [1b](#f1){ref-type="fig"}) and ethidium-bromide staining, as previously described ([@r15]). We identified three N-protein mutants with RNA-binding abilities significantly different from wild type. The most striking was R94A, for which no RNA binding was detected. In addition, RNA binding of mutants R40A and K50A was reduced significantly. All other mutants displayed RNA binding that was comparable to wild type. As RNA binding of the 48-mer RNA species relies on N-protein tetramer formation, it was also possible that changes in RNA binding reflected changes in multimerization. Therefore, we used size-exclusion chromatography to assess the oligomeric state of wild-type and mutant N proteins, as previously described ([@r15]). Purified N proteins were analysed by using size-exclusion chromatography with a Superdex S200 column (GE Healthcare Life Sciences) immediately following purification and their elution profile determined by spectroscopy and compared with molecular mass standards (Fig. [2](#f2){ref-type="fig"}). As expected from previous results ([@r15]), wild-type N protein was eluted predominantly as a tetramer with an observed molecular mass of 113.7 kDa (115.1 kDa predicted) with a minor peak at 27.2 kDa corresponding to monomer (28.8 kDa predicted) (Fig. [2a](#f2){ref-type="fig"}). Analysis of R94A (Fig. [2a](#f2){ref-type="fig"}) indicated the quantity of monomer was increased, such that monomers and tetramers were equally abundant. However, a major fraction of the expressed material was in the tetramer form, an observation supported by native PAGE analysis of purified R94A protein (not shown). Therefore, while the presence of monomers may contribute to the loss of R94A RNA-binding ability, the lack of detectable bound RNA in R94A tetramers suggested that this residue played an important role in RNA binding. Consistent with this functional assignment, recent experiments performed in the Elliott laboratory show that infectious viruses bearing the R94A mutation could not be rescued ([@r8]), revealing that R94 plays a critical role in the BUNV life cycle. The results of our analysis indicate that this critical role may be RNA binding. The elution profiles of R40A and K50A revealed no significant change in monomer/tetramer ratios, indicating that their reduced RNA binding was not because of reduced multimerization (Fig. [2b](#f2){ref-type="fig"}). Interestingly, both mutants showed an increased proportion of faster-eluting species compared with wild type, with masses \>1000 kDa. Similar large complexes were also detected for K55A, R101A and K179A, all of which displayed wild-type levels of RNA binding (Fig. [2c](#f2){ref-type="fig"}), indicating that formation of such higher-order structures did not correlate with reduced RNA binding. Therefore, these data suggested that R40 and K50 contribute to RNA binding, although they are not essential for it. Interestingly, both these mutants have previously been rescued into infectious viruses ([@r8]), although the resulting viruses exhibited considerably attenuated growth characteristics. As described above for residue R94, our findings, reported here, that R40A and K50A are deficient in RNA binding offer a plausible explanation for why the resulting rescued viruses are either non-viable or growth deficient, consistent with the important role of RNA binding in the virus life cycle. All of the other N-protein mutants exhibited monomer/tetramer ratios that were indistinguishable from wild-type N protein (data not shown), and so we concluded that none of these residues play a critical role in N protein oligomerization. We analysed the ability of all 15 N mutants to support the assembly of BUNV RNP templates, which were active for both transcription and replication. This was achieved by using a previously described assay ([@r4]; [@r7]) that directly detects RNAs generated by the BUNV RNA-dependent RNA polymerase (RdRp) from assembled RNP templates. Briefly, plasmids expressing model segment BUN-M(ren) and the BUNV-S and -L ORFs were transfected into BHK-21 cells that had previously been infected with vaccinia recombinant vTF7-3, which expresses T7 RNA polymerase. BUNV-specific RNAs generated from the resulting RNPs were harvested and detected by primer extension analysis by using ^33^P-end-labelled oligonucleotide REN-PAGE, which was designed to anneal to positive-sense products of both transcription and replication. To assess the role of the 15 charged residues in the formation and activity of RNPs, the plasmid expressing the BUNV N-protein ORF was modified to incorporate each amino acid change individually. Primer-extension analysis detected both mRNAs and anti-genomic RNAs generated from assembled RNPs. By using densitometry to measure the abundance of their corresponding products, we quantitatively assessed the role of each altered residue on transcription and replication activities (Fig. [3](#f3){ref-type="fig"}). Most mutations had a minimal effect on RNP activity, such that levels of replication and transcription were each comparable to those of the wild type. However, mutation R94A exhibited a striking phenotype, which was that it showed dramatically reduced transcriptional activity despite exhibiting robust replication. As described above, the R94A mutation cannot be rescued into infectious virus ([@r8]), and our results presented here suggest that the dramatically reduced transcriptional ability may be responsible. We are curious as to how the N protein is able to impart differential effects to the RdRp functions of transcription and replication. Presumably, any mechanism responsible is unlikely to involve activities that are common to both these RNA synthesis modes. One possibility is that the R94A mutation somehow affects access of the RdRp to *cis*-acting signals for replication versus transcription. BUNV-replication and -transcription signals comprise interacting nucleotides from both ends of each active RNP template, although the critical nucleotides that comprise each signal are distinct ([@r4], [@r5]; [@r2], [@r3]). Perhaps the N protein differentially affects the accessibility of these nucleotides to the BUNV RdRp by altering the exposure of the RNA bases to the polymerase. It is also possible that alterations in base exposure may affect how the ends of the template interact, which also affects RdRp function. Another interesting possibility is that BUNV-replication and -transcription activities may be performed by distinct complexes, as has been proposed for VSV ([@r12]; [@r19]), and differences in RNP composition may differentially affect recruitment of these complexes to the template. The concept that the N protein of a negative-stranded RNA virus can differentially modulate the template activity of the RNP is not new, having been established with the polR1 mutant of the NNS RNA virus VSV ([@r6]). More recently, this concept was reinforced by the identification of additional N-protein residues that differentially affected VSV RNP activity, which was guided by the high-resolution N--RNA structure ([@r13]; [@r16]; [@r20]). However, the RNPs of SNS RNA viruses such as BUNV exhibit important differences from those of the NNS RNPs, including increased sensitivity to RNase digestion, which implies increased exposure of the RNA backbone on the external surface of the RNP. Our results indicate that despite this less intimate N--RNA association, BUNV N-protein residues are likely to exert a major influence on how the RdRp recognizes the RNA template. Taken together, these results show that N proteins of both NNS and SNS RNA viruses possess critical functions outside their roles of RNA encapsidation. Supplementary Material ====================== ::: {.caption} ###### \[Supplementary Figure\] ::: We thank Richard Elliott (University of St Andrews, UK) for continued use of plasmids expressing S and L ORFs as well as BUNV-specific model RNP templates, and in addition for helpful comments during preparation of this manuscript. J. N. B. is a Research Council UK academic fellow, and is further supported by funding from The Wellcome Trust. A supplementary figure is available with the online version of this paper which shows the alignment of the N proteins of selected members of the genus *Orthobunyavirus*. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Analysis of the RNA-binding ability of BUNV N-protein mutants. (a) The high purity of BUNV N-protein samples used to extract RNA was confirmed by subjecting purified proteins to SDS-PAGE analysis and visualization by using Coomassie staining immediately prior to RNA extraction. (b) RNA was extracted from an equal mass of mutant N protein expressed in bacterial cells, electrophoresed on a non-denaturing agarose gel and visualized by ethidium-bromide staining. Each N protein mutant is identified with the number of the corresponding altered residue. RNA size markers are shown by arrowheads. Stds, Standards; WT, wild type. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Analysis of the ability of N-protein mutants to multimerize. Purified BUNV N-protein mutants were analysed by size-exclusion chromatography and elution profiles compared with both wild-type N protein and molecular mass standards. Elution positions of N-protein monomers (1n) and tetramers (4n) are marked with arrows. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### RNA replication and mRNA transcriptional abilities of N-protein mutants were examined by reconstituting BUNV RNPs within BHK-21 cells. Lanes are marked with the number of the altered residue of the mutant used to assemble the corresponding RNPs. RNAs were harvested and analysed by primer extension, which detected T7 primary transcripts, positive-sense replication products and mRNAs. The cDNA encoding RNA template BUN-M(ren) was sequenced with the same primer to act as a size marker. \*, Nucleotide corresponding to the anti-genomic 5′ end. Autoradiographs from two experiments were scanned and analysed by densitometry, and the RNA synthesis characteristics of each RNP were expressed as a ratio of transcriptional activity/replication activity, with each ratio normalized to wild type and which is shown below each corresponding lane. Error bars ([sem]{.smallcaps}) are shown. :::  :::

PubMed Central

2024-06-05T04:04:19.362693

2011-1-01

PMC3052535

Skin lesions in atopic dermatitis (AD) patients display histological alterations such as epidermal thickening and infiltration by eosinophils and mast cells ([@b1]). The pathogenesis of AD is believed to involve repeated abnormal innate and adaptive immune responses to environmental causative agents when the skin barrier is disrupted ([@b2]). Aeroallergens, including house dust mites and pollens, are well known to induce AD through IgE-mediated mechanisms ([@b3]). In addition, microbes represent another important group of extrinsic causative factors in the pathogenesis of AD ([@b4]). *Staphylococcus aureus* (*S. aureus*) appears to be particularly important because it colonizes almost all lesional skin in AD patients, and a reduction in its colonization has been shown to decrease disease severity ([@b5]--[@b7]). Some AD patients produce IgE antibodies specific to staphylococcal enterotoxin A or B (SEA or SEB) ([@b8], [@b9]). Lipoteichoic acid, cell wall component in *S. aureus*, has been detected in the lesions of AD patients and is correlated with AD severity ([@b10]). This finding together with the fact that about 20--40% of AD patients display intrinsic AD with no sensitization to any protein allergen suggests the importance of microbes ([@b11]). Gram-negative bacteria secrete outer membrane vesicles (OMV) ([@b12]), which have pathogenic effects ([@b13], [@b14]). Recently, we demonstrated for the first time that the Gram-positive bacterium *S. aureus* produces OMV-like vesicles called extracellular vesicles (EV) ([@b15]). The EV produced by *S. aureus* are spherical with a diameter of 20--100 nm and are shed from the bacterium\'s membranes. Proteomic analysis revealed that the protein expression pattern in these EV differs from that in whole bacteria, and that they contain various pathogenic molecules. Among these, α-hemolysin and cysteine protease have been linked with AD ([@b16], [@b17]). These findings suggest that EV is a potent initiator of host immune responses. In AD patients, *S. aureus*, rather than invading and infecting the skin, colonizes it. It is thus reasonable to assume that the skin is affected by secretory products from *S. aureus*. Given that staphylococcal secretory products are relevant to the pathogenesis of AD, we hypothesized that *S. aureus* EV, which are complexes of various pathogenic molecules secreted by the bacterium, are involved in the pathogenesis of AD. In this study, we found through *in vitro* and *in vivo* studies that *S. aureus* EV are causative agents in AD, and the clinical observation also supports this hypothesis. Methods ======= Mice ---- SKH-HR1 (hairless) mice were purchased from Charles River Laboratories Japan, Inc. (Yokohama, Japan) and were bred in a pathogen-free facility at Pohang University of Science and Technology (POSTECH; Pohang, Korea). All animal experiments were approved by the POSTECH Ethics Committee. Patients -------- Skin lavage fluids were obtained from two AD patients visiting Pediatric Clinic of Seoul Suncheonhyang Hospital (Seoul, Korea). Serum samples were obtained from 60 AD patients (30 patients aged 6--9 years and 30 patients aged over 10 years) and 20 healthy subjects aged 6--16 years, who were recruited from Seoul Samsung Hospital (Seoul, Korea). Skin lavage fluids and serum samples were isolated after written informed consent had been obtained. The study protocol was approved by the Ethics Committee of Seoul Suncheonhyang Hospital and Seoul Samsung Hospital, respectively. Isolation of *Staphylococcus aureus* extracellular vesicles ----------------------------------------------------------- *Staphylococcus aureus* EV were obtained as described previously ([@b15]). Briefly, *S. aureus* (ATCC14458) was cultured in nutrient broth (Merck, Darmstadt, Germany) and grown at 37°C to 1.0 of optical density (at 600 nm). Bacteria were removed by centrifugation, and the resulting supernatant was filtered through a 0.45-μm vacuum filter. The filtrate was concentrated by ultrafiltration using the QuixStand Benchtop System (Amersham Biosciences, Piscataway, NJ, USA) in conjunction with a 100-kD hollow-fiber membrane (Amersham Biosciences). The resulting concentrated filtrate was passed through a 0.22-μm vacuum filter. Extracellular vesicles were isolated from the resulting filtrate by ultracentrifugation at 150 000 ***g***. The concentration of protein in the EV was measured by Bradford assays (Bio-Rad Laboratories, Hercules, CA, USA). Hereafter, the reported doses of EV refer to the amount of EV protein. Isolated EV were stored at −80°C before use. Bacteria and EV removed, and then \< 100 and \>100-kD soluble fractions of bacterial culture media were concentrated. Protein concentration was measured by Bradford assays, and culture media were stored at −80°C before use. Detection of staphylococcal enterotoxin A and staphylococcal enterotoxin B -------------------------------------------------------------------------- The presence of SEA and SEB in EV and concentrated culture media was analyzed using SDS-PAGE and western blot. SEA and SEB were detected by monoclonal anti-SEA and anti-SEB antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The generation of an atopic dermatitis mouse model using *Staphylococcus aureus* extracellular vesicles ------------------------------------------------------------------------------------------------------- To create a mouse model of AD, *S. aureus* EV were applied to mouse skin according to the following protocol. To disrupt the cutaneous barrier, the dorsal skin of 6-week-old mice was stripped five to six times using Durapore surgical tape (3M Co., St. Paul, MN, USA). Gauze (1.5 × 1.5 cm) soaked with *S. aureus* EV in 100 μl of phosphate buffered saline (PBS) was then placed on the stripped skin and secured using Tegaderm bio-occlusive tape (3M Co.). For the evaluation of inflammation and immune dysfunction, the mice were euthanized 48 h after the final challenge. Histological analysis --------------------- Four-micrometer-thick sections of fixed skin tissues were stained with hematoxylin and eosin (H&E). Mast cells were stained with toluidine blue (TB). Cells were counted in 15--25 high-power fields at a magnification of ×400. Characterization of T-cell subsets ---------------------------------- Single cells of skin-draining lymph nodes (LN) were collected and stimulated with or without 0.1 μg/ml of *S. aureus* EV. Supernatants were harvested after 72 h, and the levels of cytokines measured by ELISA. *In vitro* production of pro-inflammatory mediators from dermal fibroblasts --------------------------------------------------------------------------- Primary mouse dermal fibroblasts were isolated as described previously with some modification ([@b18]). Fibroblasts from passages 1--3 were used. Then, 2 × 10^4^ cells of isolated cells were cultured in 24-well plates then treated with 1 or 10 μg/ml *S. aureus* EV or soluble fractions of bacterial culture media, or SEB (Toxin Technology, Sarasota, FL, USA). Supernatants were collected 24 h after stimulation, and mediator levels measured. Measurement of cytokine and chemokine secretion ----------------------------------------------- The cytokine and chemokine levels were measured by ELISA (R&D Systems, Mineapolis, MN, USA) according to the manufacturer\'s instructions. Isolation of extracellular vesicles from the skin lavage fluids of patients --------------------------------------------------------------------------- Skin lavage fluids were obtained by rinsing patients\' skin lesions three to four times with 50 ml of sterile PBS and were stored at −80°C. To remove bacteria and other debris, 40 ml of skin lavage fluids was centrifuged at 5000 and 10 000 ***g***. After centrifugation, supernatants were filtered through 0.45 and 0.22 μm serially. Then, lavage fluids were concentrated to 1 ml by using Centriprep YM-50 (Millipore, Carringtwohill, Ireland). Same volume of PBS was added on concentrated lavage fluids, and they were ultracentrifuged at 150 000 ***g*** for 3 h at 4°C. The pellet was used as EV fraction. ELISA assay using anti-*Staphylococcus aureus* extracellular vesicles-specific polyclonal antibodies ---------------------------------------------------------------------------------------------------- Anti-*S. aureus* EV-specific polyclonal antibodies were coated on 96-well ELISA plate. Each well was blocked by 1% bovine serum albumin in PBS. After blocking, concentrated lavage fluids and EV fraction were added to each well. Then, biotinylated anti-*S. aureus* EV-specific polyclonal antibodies were added on each well, and then streptavidin-horseradish peroxidase (HRP) was added. After a final wash, chemiluminescence substrates (POD) were added to react with HRP. Luminescence measured using a Wallac 1420 Victor luminometer (American Instrument Exchange, Inc., Haverville, MA, USA). Serum antibody ELISA -------------------- ### Mouse Serum samples were prepared from mouse blood for analysis of the total serum IgG1 and IgE levels by ELISA (Bethyl Laboratories, Montgomery, TX, USA) according to manufacturer\'s instructions. ### Human *Staphylococcus aureus* EV- and SEB-specific IgG1 and IgE levels were measured by ELISA. The wells of 96-well ELISA plates were each coated with 0.1 μg of *S. aureus* EV or 1 μg of SEB. Then, wells were blocked with 3% bovine serum albumin in PBS. After blocking, diluted human serum was added to the wells. Then, HRP-conjugated anti-human IgG1 and IgE antibodies (Southern Biotech, Birmingham, IL, USA) were applied to each well. Chemiluminescence substrates were added, and luminescence was measured by the luminometer. Specific antibody levels were defined as elevated if a value is higher than mean +1 standard deviation of healthy subjects\' values. Statistical analysis -------------------- Statistically significant differences between treatments were identified using Student\'s *t*-test, [anova]{.smallcaps}, or Wilcoxon\'s rank sum test. Multiple comparisons were initially made by [anova]{.smallcaps}. Where significant differences were found, individual *t*-tests or Wilcoxon\'s rank sum tests were used to identify statistically significant differences between treatment group pairs. A *P*-value \< 0.05 was considered to be statistically significant. Results ======= *In vitro* production of pro-inflammatory mediators from mouse dermal fibroblasts treated with *Staphylococcus aureus* extracellular vesicles --------------------------------------------------------------------------------------------------------------------------------------------- Recently, we found that *S. aureus* EV contain 90 proteins, including proteins with pathological function, by proteomic analysis ([@b15]). Scanning electron microscopic images showed that *S. aureus* secreted EV ([Fig. 1A](#fig01){ref-type="fig"}). We evaluated whether EV or soluble fractions of *S. aureus* culture media induce the production of pro-inflammatory mediators from skin fibroblasts. We found that the production of IL-6, thymic stromal lymphopoietin (TSLP), macrophage inflammatory protein (MIP)-1α, and eotaxin by dermal fibroblasts was higher by stimulation with EV than with soluble fraction ([Fig. 1B](#fig01){ref-type="fig"}). It was reported that the presence of IgE antibodies to SEA and SEB was correlated with the severity of skin lesions in children with AD ([@b8], [@b9]). Western blotting using anti-SEA and anti-SEB antibodies showed that SEB was present in soluble fraction, but not in EV, whereas SEA absent in both fractions ([Fig. 1C](#fig01){ref-type="fig"}). We compared *in vitro* activity between EV and SEB on the production of pro-inflammatory mediator. As shown in [Fig. 1D](#fig01){ref-type="fig"}, 1 or 10 μg/ml of SEB did not enhance the production of IL-6, TSLP, MIP-1α, and eotaxin, whereas 1 μg/ml of EV upregulated the production of these mediators. Collectively, these findings suggest that *S. aureus*-derived EV are more potent compared to soluble components, in terms of production of pro-inflammatory mediators. ::: {#fig01 .fig} Figure 1 ::: {.caption} ###### *Staphylococcus aureus* extracellular vesicles (EV) enhance the *in vitro* secretion of immune and pro-inflammatory mediators by mouse dermal fibroblasts. (A) Scanning electron microscopic images that *S. aureus* secrete EV. Arrowheads identify *S. aureus* EV. (B) Levels of pro-inflammatory mediators IL-6, thymic stromal lymphopoietin, macrophage inflammatory protein-1α and eotaxin in supernatants of dermal fibroblasts after stimulation with EV and \>100- and \< 100-kD soluble fractions of bacterial culture media. (C) Western blotting to detect staphylococcal enterotoxin B (SEB) in EV and soluble (Sup) fractions of *S. aureus* culture media. (D) Levels of pro-inflammatory mediators from supernatants of dermal fibroblasts after stimulation with EV or SEB. Assays were performed in duplicate. \**P* \< 0.05; \*\**P* \< 0.01; \*\*\**P* \< 0.001. :::  ::: Local inflammation induced by the application of different doses of *Staphylococcus aureus* extracellular vesicles to tape-stripped mouse skin ---------------------------------------------------------------------------------------------------------------------------------------------- We evaluated the *in vivo* effects of *S. aureus* EV on the induction of skin inflammation. Different doses of EV were applied to tape-stripped mouse skin, and changes in skin inflammation were assessed 4 weeks after initial application ([Fig. 2A](#fig02){ref-type="fig"}). Histological analysis showed that the application of *S. aureus* EV to tape-stripped skin induced AD-like inflammation, including epidermal thickening and infiltration of the dermis by inflammatory cells ([Fig. 2B](#fig02){ref-type="fig"}). Moreover, *S. aureus* EV dose dependently induced epidermal thickening ([Fig. 2C](#fig02){ref-type="fig"}). As with dermal infiltration by inflammatory cells, significantly higher numbers of mast cells were found in the dermis of mice treated with *S. aureus* EV (at 5 or 10 μg, but not 0.1 μg) compared to PBS-treated controls. In addition, the numbers of eosinophils were significantly higher in EV-treated mice than in PBS-treated controls at all doses ([Fig. 2C](#fig02){ref-type="fig"}). Together, these data suggest that the application of *S. aureus* EV to tape-stripped skin induces AD-like inflammation. ::: {#fig02 .fig} Figure 2 ::: {.caption} ###### Application of *Staphylococcus aureus* extracellular vesicles (EV) to tape-stripped mouse skin induces atopic dermatitis-like inflammation. (A) Study protocol: application of different doses of *S. aureus* EV to tape-stripped mouse skin for 4 weeks (*n* = 5 per treatment group). (B) Skin histology \[H&E staining; magnification, ×200 (upper panel) and ×400 (lower panel)\]. Arrowheads identify eosinophils. (C) Histological analysis of epidermal thickness and the numbers of eosinophils and mast cells infiltrating the dermis. \**P* \< 0.05; \*\**P* \< 0.01; \*\*\**P* \< 0.001. :::  ::: *In vivo* immune dysfunction induced by the application of *Staphylococcus aureus* extracellular vesicles to tape-stripped mouse skin for 3 weeks ------------------------------------------------------------------------------------------------------------------------------------------------- To test whether the application of *S. aureus* EV to the skin induces immunological dysfunction, 5 μg of *S. aureus* EV was applied to tape-stripped mouse skin three times per week for 3 weeks. Histological analysis showed that the application of *S. aureus* EV induced epidermal thickening and infiltration of the dermis by inflammatory cells ([Fig. 3A](#fig03){ref-type="fig"}). Tape stripping itself induced epidermal thickening ([Fig. 3B](#fig03){ref-type="fig"}). To characterize the immune dysfunction induced by *S. aureus* EV, we measured the production of Th1, Th17, and Th2 cytokines by T cells from skin-draining LNs following *in vitro* stimulation with *S. aureus* EV. The production of IFN-γ and IL-17 was significantly higher in cells from the EV-treated mice than in those from the PBS-treated animals ([Fig. 3C](#fig03){ref-type="fig"}). IL-4 and IL-5 were not detected in the supernatants from cells isolated from either treatment group (data not shown). In terms of the skin production of Th1, Th17, and Th2 cytokines, the levels of IFN-γ, IL-17, IL-4, and IL-5 in skin homogenates from EV-treated mice were significantly higher than those from PBS-treated mice. Tape stripping did not itself enhance the production of these cytokines ([Fig. 3D](#fig03){ref-type="fig"}). With regard to the production of antibodies, we found the serum levels of total IgG1 and IgE to be similar in the EV- and PBS-treated mice (data not shown). Collectively, these data suggest that exposure to *S. aureus* EV for 3 weeks induces a mixed Th1-/Th17-/Th2-type inflammatory response in the skin, whereas a mixed Th1/Th17 cell response in skin-draining LNs. ::: {#fig03 .fig} Figure 3 ::: {.caption} ###### Three-week exposure of tape-stripped mouse skin to *Staphylococcus aureus* extracellular vesicles (EV) induces a mixed Th1-/Th17-/Th2-type inflammatory response in the skin, and the generation of Th1 and Th17 cells in skin-draining lymph nodes (LNs). Evaluation (*n* = 5 per treatment group) was performed 48 h after the final application of *S. aureus* EV (5 μg) (performed three times a week for 3 weeks) to tape-stripped skin. \**P* \< 0.05; \*\**P* \< 0.01; \*\*\**P* \< 0.001. (A) Skin histology \[H&E staining; magnification, ×200\]. (B) Histological analysis of epidermal thickness. (C) Levels of IFN-γ and IL-17 in supernatants from *S. aureus* EV-treated cells from skin-draining LNs. (D) Levels of IFN-γ, IL-17, IL-4, and IL-5 in skin tissue homogenates. :::  ::: Local inflammation and systemic immune dysfunction induced by long-term (8-week) application of *Staphylococcus aureus* extracellular vesicles to tape-stripped mouse skin -------------------------------------------------------------------------------------------------------------------------------------------------------------------------- To assess the effects of long-term exposure to *S. aureus* EV, *S. aureus* EV (5 μg) were applied to tape-stripped skin three times per week for 8 weeks. Histological analysis showed that prolonged exposure to *S. aureus* EV induced epidermal thickening and increased infiltration of the dermis by inflammatory cells ([Fig. 4A,B](#fig04){ref-type="fig"}). Infiltration of the dermis by mast cells and eosinophils was significantly increased in EV-treated mice relative to PBS-treated controls ([Fig. 4A,B](#fig04){ref-type="fig"}). Furthermore, long-term *in vivo* exposure to *S. aureus* EV enhanced the production of IL-17 by T cells in skin-draining LNs stimulated with *S. aureus* EV *in vitro* ([Fig. 4C](#fig04){ref-type="fig"}). However, it did not enhance the EV-induced production of IFN-γ and IL-4 (data not shown). In terms of antibody production following long-term exposure to *S. aureus* EV, the serum total IgE levels were significantly higher in the EV-treated mice than in the PBS-treated controls, although the serum total IgG1 levels were similar between the two groups ([Fig. 4D](#fig04){ref-type="fig"}). Taken together, these findings suggest that long-term exposure to *S. aureus* EV induces IgE production, whereas Th17-cell response in skin-draining LNs. ::: {#fig04 .fig} Figure 4 ::: {.caption} ###### Long-term exposure of tape-stripped mouse skin to *Staphylococcus aureus* extracellular vesicles (EV) enhances the production of IgE. Evaluation (*n* = 5 per treatment group) was performed 48 h after the final application of *S. aureus* EV (5 μg) (performed three times a week for 8 weeks) to tape-stripped skin. \**P* \< 0.05; \*\**P* \< 0.01; \*\*\**P* \< 0.001. (A) Skin histology \[H&E staining; magnification, ×200 (upper panel) and ×400 (lower panel)\]. Arrowheads identify eosinophils. (B) Histological analysis of epidermal thickness and numbers of eosinophils and mast cells infiltrating the dermis. (C) Levels of IL-17 in supernatants from *S. aureus* EV-treated cells from skin-draining lymph nodes. (D) Serum levels of total IgG1 and total IgE. :::  ::: The presence of *Staphylococcus aureus* extracellular vesicles in the skin lesion of atopic dermatitis patients --------------------------------------------------------------------------------------------------------------- We evaluated the presence of *S. aureus*-derived EV in the skin lesion of AD patients. To test this objective, EV were isolated from the skin lesion of two AD patients and then evaluated whether EV isolated from the skin lesion of AD patients incorporate *S. aureus* EV-specific proteins using anti-*S. aureus* EV-specific polyclonal antibodies. As shown in [Fig. 5](#fig05){ref-type="fig"}, EV isolated from skin lavage fluids of AD patients contained *S. aureus* EV-specific proteins. ::: {#fig05 .fig} Figure 5 ::: {.caption} ###### *Staphylococcus aureus*-derived extracellular vesicles (EV) are present on the skin of atopic dermatitis (AD) patients. This figure showed ELISA assay to detect *S. aureus* EV-specific proteins using anti-*S. aureus* EV polyclonal antibodies; lavage fluids and EV fraction of lavage fluids obtained from two AD patients have *S. aureus* EV-specific proteins (SA\_EV, *S. aureus*-derived EV). :::  ::: Serum *Staphylococcus aureus* extracellular vesicles-specific antibody levels in atopic dermatitis patients ----------------------------------------------------------------------------------------------------------- Finally, we measured serum levels of *S. aureus* EV- and SEB-specific antibodies in AD patients and healthy subjects. The serum levels of *S. aureus* EV- and SEB-specific IgG1 were comparable in AD patients and healthy subjects ([Fig. 6A](#fig06){ref-type="fig"}). By contrast, the serum *S. aureus* EV-specific IgE levels were significantly higher in both 6--9 year aged and \>9 year aged AD patients than in age-matched healthy subjects; this IgE levels were elevated in 33.3% of 6--9 years aged and 40% of \>9 years aged AD patients ([Fig. 6B](#fig06){ref-type="fig"}). In addition, the serum SEB-specific IgE levels were significantly higher in only \>9 years aged AD patients than in age-matched healthy subjects; this IgE levels were elevated in 33.3% of 6--9 years aged and 33.3% of \>9 years aged AD patients ([Fig. 6C](#fig06){ref-type="fig"}). However, neither total IgE nor SEB-specific IgE were correlated with *S. aureus* EV-specific IgE in AD patients (data not shown). These findings indicate that *S. aureus* EV-specific IgE may be a useful biomarker for identifying the cause of AD in individual cases. ::: {#fig06 .fig} Figure 6 ::: {.caption} ###### *Staphylococcus aureus* extracellular vesicles (EV)-specific IgE is elevated in atopic dermatitis (AD) patients than in age-matched healthy subjects. (A) Levels of *S. aureus* EV-specific (left panel) and SEB-specific (right panel) IgG1 in serum from AD patients and healthy subjects. (B) Levels of *S. aureus* EV-specific IgE in serum from AD patients and healthy subjects (left panel), and positive rate of elevated *S. aureus* EV-specific IgE in AD patients (right panel). (C) Levels of SEB-specific IgE in serum from AD patients and healthy subjects (left panel), and positive rate of elevated SEB-specific IgE in AD patients (right panel). Serum samples from AD patients (*n* = 30 in patients aged 6--9 years, and *n* = 30 in patients aged 9--16 years) and healthy subjects aged 6--16 years (*n* = 20); EV, *S. aureus*-derived EV; SEB, staphylococcal enterotoxin B. \**P* \< 0.05; \*\**P* \< 0.01. :::  ::: Discussion ========== Elucidating the pathogenesis of AD has been difficult because of complex interactions between the causative factors and host immune response. Among the known causative factors, microbes are thought to be particularly important in the pathogenesis of AD by elaborating secretary products, including soluble toxins. Recently, we found that *S. aureus* secretes EV, in which pathogenic proteins are incorporated ([@b15]). In the present study, experimental and clinical data support the hypothesis that *S. aureus* EV are involved in the pathogenesis of AD. In terms of host immune responses of AD pathogenesis, recent studies showed that IL-17 producing cells are found in the blood of AD patients and that AD pathogenesis is related with IL-17-mediated immune responses ([@b19], [@b20]). In the present study, we found that the cutaneous application of *S. aureus* EV induced Th17-cell response in skin-draining LNs, suggesting that Th17-cell response is important in the AD pathogenesis. The present data showed that the cutaneous application of *S. aureus* EV induced skin inflammation characterized by infiltration of mast cells and eosinophils. This inflammatory response was associated with enhanced production of not only Th1/Th17-type cytokines, but also Th2-type cytokines in the skin. In addition, the present study showed that *in vitro* stimulation of fibroblasts with *S. aureus* EV increased the secretion of the Th2-type cytokines, such as TSLP and eotaxin ([@b21]). These findings suggest that Th2-type inflammation induced by *S. aureus* EV is mediated by local production of Th2-type cytokines from dermal fibroblasts. *Staphylococcus aureus* can colonize in skin or nasal passage in human ([@b22]). In the present study, we firstly demonstrated that *S. aureus*-derived EV are present on the skin of AD patients. We also found that *S. aureus* EV-specific IgG1 were detected in serums from not only AD patients but also healthy subjects. These findings suggest that *S. aureus* secretes EV in skin, which induce systemic immune responses. Previous studies showed that serum levels of IgE specific to staphylococcal enterotoxins were not only elevated in AD patients, but also correlated with disease severity ([@b8], [@b9]). Our present data indicate that serum levels of both *S. aureus* EV- and SEB-specific IgE were elevated in the AD patients compared to healthy subjects. In addition, the present study showed that *S. aureus* EV did not contain enterotoxins and *S. aureus* EV-specific IgE levels did not correlate with SEB-specific IgE. These findings suggest that the production of EV-specific IgE is not influenced by staphylococcal enterotoxins. Recently, it was reported that IL-17 can induce IgE production in B cells ([@b23]). The present study showed that the cutaneous application of *S. aureus* EV for 8 weeks induce Th17-cell response, but not Th2-cell response, along with the elevation of serum total IgE in mice. These findings suggest that *S. aureus* EV can induce systemic IgE production by Th17-cell response. Recent evidence indicates that EV from Gram-negative bacteria induced systemic inflammatory response ([@b13]). We found that Gram-positive bacteria produced EV ([@b15]). This is the first report to show that EV obtained from Gram-positive bacteria can cause inflammatory disease. Given the abundance of Gram-positive bacteria in our environment, further research will be needed to elucidate the relationships between EV from Gram-positive bacteria and the pathogenesis of immune-based inflammatory diseases. In summary, our present data indicate that *S. aureus*-derived EV can induce AD-like inflammation in the skin, and that the physiological animal model we employed represents a useful tool for performing translational research into AD. Furthermore, our clinical findings provide strong evidence of the importance of *S. aureus* EV in the pathogenesis of AD. We thank Jee-In Lim and Chae-Min Kim for their providing secretarial assistance and members of the POSTECH animal facility for their experimental expertise. This study was supported by grants from the Korea Ministry of Health & Welfare, Republic of Korea (A080711). Conflict of interest ==================== The authors declare that there are no conflicts of interest. [^1]: Re-use of this article is permitted in accordance with the Terms and Conditions set out at <http://wileyonlinelibrary.com/onlineopen#OnlineOpen_Terms>.

PubMed Central

2024-06-05T04:04:19.364369

2011-3-01

PMC3052537

*Bunyamwera virus* (BUNV) is the type species of both the genus *Orthobunyavirus* and the family *Bunyaviridae*, members of which are important human pathogens causing diseases such as febrile illness and haemorrhagic fever. Bunyaviruses possess a tri-segmented, negative-sense RNA genome that encodes four structural proteins: the viral RNA-dependent RNA polymerase (L protein) on the large (L) segment, two glycoproteins (Gn and Gc) on the medium (M) segment, and the nucleoprotein (N) on the small (S) segment. In addition, orthobunyaviruses encode one or two non-structural proteins, NSm on the M segment and NSs (for most but not all orthobunyaviruses; [@r12]) on the S segment ([@r13]; [@r14]). The NSs protein is the major viral interferon (IFN) antagonist involved in evading host innate immune responses ([@r2]; [@r18]). NSs was shown to interact with Med8 ([@r10]), a component of the Mediator complex that is involved in regulating the activity of cellular RNA polymerase II (RNAPII; [@r11]). During infection with wild-type (wt) BUNV cellular RNAPII is degraded, presumably as a result of the interaction between NSs and Med8 ([@r10]). Thus, NSs is thought to antagonize the IFN response by a general block of transcription of all host genes including IFN. This general transcriptional block caused by NSs is at least partially responsible for the observed shut off of host cell protein synthesis ([@r2]; [@r16]; [@r6]). The N and NSs proteins are translated from overlapping open reading frames (ORFs) in a single mRNA transcribed from the S segment. To examine the role of NSs in infection, a recombinant virus lacking NSs (rBUNdelNSs2) was created in which NSs expression was abrogated by changing the tandem AUG initiation codons to ACG codons and converting codon 3 (serine) to a stop codon. In addition, codons 4 and 5 were changed from leucine to proline ([@r2]). Subsequently, a potential downstream AUG-initiation codon at position 30 was changed to ACG (Fig. [1a](#f1){ref-type="fig"}; [@r6]). All the mutations introduced to create rBUNdelNSs2 are silent with respect to the N ORF. During routine analysis of newly prepared viral stocks derived from isolated plaques, lysates of infected BHK cells were checked for expression of NSs by Western blotting as described previously ([@r1]). As expected, wtBUNV expressed the 11 kDa NSs protein (Fig. [1b](#f1){ref-type="fig"}), but surprisingly one particular stock of the NSs-deleted virus, subsequently named mBUNNSs22, expressed a ∼9 kDa protein that cross-reacted with BUNV NSs antiserum (Fig. [1b](#f1){ref-type="fig"}). Analysis of the nucleotide sequence encompassing the N and NSs start codons in the S segment revealed that codon 22 in the NSs ORF was changed from GUG to AUG \[nt 168--170 in the (+) sense S RNA\], creating a new translation initiation codon in the NSs gene (Fig. [1a](#f1){ref-type="fig"}). The NSs protein resulting from initiation at this new AUG (NSs22 protein) is predicted to have a molecular mass of 8.76 kDa, which is consistent with the observed size difference compared with full-length NSs (Fig. [1b](#f1){ref-type="fig"}). The mutation also led to a conservative amino acid change in the N protein from arginine to histidine at residue 28 (Fig. [1a](#f1){ref-type="fig"}). Arginine 28 is not conserved between different orthobunyavirus N proteins and mutations at this position have only a minor impact on virus replication ([@r3]). Treatment of infected cells with the proteasome inhibitor MG132 (10 μM) leads to accumulation and therefore more sensitive detection of NSs ([@r6]). We compared NSs expression in cells infected with another rBUNdelNSs2 stock to those infected with wt- and mBUNNSs22 viruses in the presence of MG132 (Fig. [1c](#f1){ref-type="fig"}). Whereas the latter two lysates showed bands corresponding to full-length and truncated NSs, respectively, no NSs protein could be detected in extracts of rBUNdelNSs2-infected cells. In addition, nucleotide sequence analysis of the S segment of rBUNdelNSs2 confirmed that only the mutations that were originally introduced to create the NSs-deletion mutant were present (codons 1--5 and 30; [@r2]; [@r6]), and codon 22 remained as GUG. Multicycle virus growth in BHK cells was analysed as described previously ([@r15]) at both 33 °C, the temperature at which viral stocks are prepared, and 37 °C, the temperature used for most experimental work with BUNV. At both temperatures both mutant viruses grew to titres about 10-fold lower than wt virus (Fig. [1d, e](#f1){ref-type="fig"}). Interestingly, mBUNNSs22 grew to slightly lower titres than delNSs at 33 °C (4.4×10^5^ p.f.u. ml^−1^ vs 1.5×10^6^ p.f.u. ml^−1^; Fig. [1d](#f1){ref-type="fig"}), but slightly higher titres at 37 °C (4×10^6^ p.f.u. ml^−1^ vs 7.6×10^5^ p.f.u. ml^−1^; Fig. [1e](#f1){ref-type="fig"}), an observation that was consistent in repeated experiments (data not shown). Although mBUNNSs22 was attenuated in the partially IFN-deficient BHK cell line ([@r4]), we speculated that this virus may be less sensitive to the host IFN response than rBUNdelNSs2, since it expressed a large fragment of the viral IFN antagonist. In IFN-competent A549 cells, mBUNNSs22, like rBUNdelNSs2, was severely attenuated, and grew to titres approximately 1000-fold lower than wtBUNV (Fig. [2a](#f2){ref-type="fig"}). The levels of IFN induction in A549 cells were analysed using the biological assay described by [@r12]. Briefly, the medium from infected A549 cells was collected at 24 h post-infection (p.i.), UV-inactivated and then used to induce protection of indicator cells from encephalomyocarditis virus (EMCV) infection. Infection by rBUNdelNSs2 or mBUNNSs22 resulted in secretion of significantly higher amounts of biologically active IFN than infection with wtBUNV (Fig. [2b](#f2){ref-type="fig"}), indicating that mBUNNSs22, like rBUNdelNSs2, is a strong IFN inducer. Finally, we compared the plaque phenotypes of wtBUNV, mBUNNSs22 and rBUNdelNSs2 in A549 cells and in A549-NPro cells that express the bovine viral diarrhea virus NPro protein ([@r5]). NPro induces proteasome-mediated degradation of IRF-3, a cellular transcription factor essential for the production of IFN-*β* in response to virus infection ([@r7]). The cells were infected with approximately 50 p.f.u. of virus and stained after 5 days incubation at 37 °C. Only wt virus produced plaques on naïve A549 cells, but all three viruses formed plaques in A549-NPro cells (Fig. [2c](#f2){ref-type="fig"}). Thus, the attenuation of mBUNNSs22 in naïve A549 cells can be relieved by degradation of IRF-3, suggesting that mBUNNSs22, like rBUNdelNSs2, had lost its IFN-antagonist function. The mechanism by which wtBUNV blocks the IFN response has been proposed to involve NSs-mediated blocking of phosphorylation of serine-2 in the heptad repeat in the RNAPII C-terminal domain (CTD; [@r16]; [@r10]). To test whether mBUNNSs22 was impaired in its ability to inhibit serine-2 phosphorylation, BHK cells were infected with wtBUNV, rBUNdelNSs or mBUNNSs22 and cell lysates analysed by Western blotting using antibodies specific for the serine-2 phosphorylated CTD of RNAPII (Ser2-P RNAPII; H5, Covance Research Products) or for RNAPII irrespective of its phosphorylation state (8WG16; Covance). As observed consistently in repeated experiments, during wtBUNV infection an increase in the signal for NSs correlated with a decrease in the signal for Ser2-P RNAPII and later also RNAPII in any phosphorylation state. Although it cannot be concluded per se that NSs is directly responsible for the degradation of RNAPII, it seems plausible that NSs disturbs serine-2 phosphorylation of the CTD and this leads to a stalled RNAPII complex, which is then targeted for degradation. Generally, no decrease in RNAPII levels was observed in rBUNdelNSs2-infected cell extracts where no NSs was expressed (Fig. [3a](#f3){ref-type="fig"}), confirming that NSs is responsible for RNAPII degradation. In extracts of cells infected with mBUNNSs22 a clear signal for the truncated NSs protein was detected, but no decrease in RNAPII levels could be observed (Fig. [3a](#f3){ref-type="fig"}). These results confirmed that mBUNNSs22 had lost the ability to block phosphorlyation or induce degradation of RNAPII and thus to counteract the host IFN response. Previous analyses had mapped the Med8-interacting domain in NSs to residues 83--91, and showed that NSs proteins with N-terminal truncations of 10, 40 or 49 aa could still interact with Med8 ([@r10]). This implies that the truncated NSs22 protein expressed by the mutant virus would be capable of interacting with Med8, and yet no degradation of RNAPII could be detected. Therefore, the interaction between the NSs C terminus and Med8, though essential, seems not to be sufficient to block phosphorylation of CTD-Ser2 or to promote degradation of RNAPII during infection. The ability of mBUNNSs22 to shut off host protein synthesis was compared to that of wt and rBUNdelNSs2 viruses in A549 and Vero cells (Fig. [3b](#f3){ref-type="fig"}) by metabolic labelling with \[^35^S\]methionine as described previously ([@r10]). Whereas wtBUNV caused host protein synthesis shut off in both cell types, hardly any shut off was observed for rBUNdelNSs2 or mBUNNSs22 (Fig. [3b](#f3){ref-type="fig"}). The shut off observed in wtBUNV-infected cells is the result of the effect of NSs on both cellular translation and transcription ([@r2]; [@r6]; [@r16]) and the blocking of RNAPII activity is a major contributing factor. Thus, the results in Fig. [3(b)](#f3){ref-type="fig"} correlate well with those in Figs [2](#f2){ref-type="fig"} and [3(a)](#f3){ref-type="fig"}, and are in agreement with the hypothesis that inability of NSs to block RNAPII activity prevents virus-mediated shut off of host cell gene expression and consequently leads to the induction of IFN. Compared to the amount of N protein, the level of NSs22 protein produced by mBUNNSs22 appeared to be lower than that of full-length NSs made by wtBUNV, at least for the first 12 h of infection (Figs [1b](#f1){ref-type="fig"} and [3a](#f3){ref-type="fig"}). This difference was consistently observed in repeated Western blot experiments (not shown) and may reflect a lower or slower rate of synthesis of the truncated protein. Alternatively, it could be the result of increased instability of the mutant protein. Two observations seem to point towards the latter explanation: (i) the NSs22 protein was not consistently detected in plasmid-transfected cells (data not shown), and (ii) the C-terminally truncated NSs protein expressed by BUNNSs-T83 ([@r10]) is also difficult to detect in infected cells (unpublished observations). The first NSs-deletion virus generated, rBUNdelNSs9a, was found to express a truncated NSs protein (NSs30) from the AUG at codon 30, but this virus had the same phenotype in BHK cells as the subsequently made rBUNdelNSs2 virus ([@r2]; [@r6]) and was a potent inducer of IFN ([@r18]). The data presented here are in agreement with, and extend, those obtained for rBUNdelNSs9a. It seems unlikely that NSs30 is expressed during wtBUNV infection since most ribosomes will initiate either at the N start codon \[that is in a weak 'Kozak' sequence ([@r9])\] or at the NSs start codon (that is in a slightly stronger Kozak context). On the other hand, the start codon to initiate NSs22 translation is in a relatively strong Kozak context even though it was not originally an initiation codon. This may explain why the mutation that rescued truncated-NSs expression occurred at this site rather than reversion of codon 30. The fact that mBUNNSs22 arose spontaneously suggests a selective pressure for some function of NSs to be regained by the virus. Our results indicate that although mBUNNSs22 expresses a large part of the viral IFN antagonist protein, this mutant virus behaves indistinguishably from rBUNdelNSs2 in its inability to inhibit IFN induction. The selective pressure must thus be for another function of NSs, which could be to inhibit protein translation ([@r6]), to counteract induction of apoptosis ([@r8]) or to regulate the viral RNA polymerase ([@r17]). Attempts to examine the effect of NSs22 on viral polymerase activity using the minigenome assay ([@r17]) were thwarted by the instability of the truncated protein, though other means to measure this effect are currently being explored. The results presented in Figs [2](#f2){ref-type="fig"} and [3](#f3){ref-type="fig"} demonstrate the requirement of the first 21 aa of the NSs protein for its IFN-antagonist function. Although the interaction between NSs and Med8 is thought to be involved in blocking host transcription ([@r10]), the presence of the interacting domain in NSs (located between aa 83 and 91; [@r10]) is apparently in itself not enough to lead to the blocking of RNAPII activity. This suggests that the interaction with Med8 alone is not sufficient to inhibit host transcription and that other factors requiring the NSs N terminus are involved in this mechanism. The identity of these factors and the mechanism of RNAPII inhibition are currently under investigation. We thank Angela McLees for technical support and Benjamin Hale for critical reading of the manuscript. Work in R. M. E.\'s laboratory is supported by grants from the Wellcome Trust, MRC and BBSRC. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Characterization of a mutant Bunyamwera virus that expresses an N-terminally truncated NSs protein. (a) Schematic of S segment RNA sequences surrounding the N and NSs ORF start codons for wtBUNV, rBUNdelNSs2 (delNSs2) and mBUNNSs22 (NSs22). Shown are nt 83--119 and 162--197 of the S (+) RNA. The numbering of the relevant codons is indicated, AUG start codons are underlined and bold, and an asterisk (\*) marks a stop codon. Amino acid sequences for N and NSs are shown above and below the RNA sequences, respectively. (b) Western blot analysis of BHK cells infected with wtBUNV (wt) or mBUNNSs22 (NSs22) or mock-infected (M). Strips of the blot were probed with the antibodies indicated on the right; size markers are indicated on the left. (c) Western blot analysis of BHK cells infected with wtBUNV, mBUNNSs22 (22), rBUNdelNSs2 (del) or mock-infected. MG132 (10 μM) was added to the medium at 5 h p.i. and cell extracts were prepared at 12 h p.i. (d, e) Multi-step growth curves of wtBUNV, rBUNdelNSs2 and mBUNNSs22 in BHK cells at 33 °C (d) and 37 °C (e). Shown are mean values of triplicate infections. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### mBUNNSs22 is attenuated in IFN-competent cells and is a potent IFN inducer. (a) Multi-step growth curves of wtBUNV, rBUNdelNSs2 and mBUNNSs22 virus in A549 cells. Shown are mean values of triplicate infections. (b) Levels of IFN induced in A549 cells after 24 h infection with wtBUNV, rBUNdelNSs2 or mBUNNSs22. The relative IFN content of medium from infected cells was measured by comparing the dilution that could protect indicator cells from EMCV infection. (c) Plaque formation in IFN-competent A549 cells (left panels) and IFN-deficient A549-NPro cells (right panels). Cells were infected with wtBUNV, rBUNdelNSs2 or mBUNNSs22 as indicated and were stained for plaque formation after 5 days incubation at 37 °C. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### mBUNNSs22 does not degrade RNAPII or cause shut off of host protein synthesis. (a) Western blot analysis of BHK cells infected with wtBUNV, rBUNdelNSs2, mBUNNS22 or mock-infected, and harvested at the indicated times p.i. Size markers are indicated on the left, and antibodies used on the right. *α*-RNAPII, antibody against RNAPII-CTD, regardless of its phosphorylation state; *α*-Ser2-P, antibody specific for serine-2-phosphorylated CTD of RNAPII. (b) Metabolic labelling of infected cells. A549 cells (top panel) and Vero cells (bottom panel) were infected with wtBUNV, rBUNdelNSs2, mBUNNSs22 or were mock-infected. Cells were labelled with \[^35^S\]methionine for 1 h prior to the indicated time p.i., and cell lysates were analysed by SDS-PAGE. Viral proteins and their sizes are indicated on the right. :::  :::

PubMed Central

2024-06-05T04:04:19.366484

2010-8-01

PMC3052539

Human cytomegalovirus (HCMV), the prototype species of the subfamily *Betaherpesvirinae*, has a high prevalence in populations worldwide. Although HCMV is recognized to be an important human pathogen, particularly in immunocompromised individuals or following congenital infection, the vast majority of primary infections are subclinical and accompanied by asymptomatic lifelong carriage. HCMV encodes highly effective systems to provide for latency, persistent reactivation and transmission; as part of this process the virus acquired an impressive array of genes that act both to evade and redirect the host immune response ([@r46]). The fact that individuals with genetic defects in their natural killer (NK) cell response are particularly susceptible to severe HCMV disease ([@r8]; [@r22]) provided a rationale to focus attention on this arm of the immune response. NK cells are composed of heterogeneous populations expressing a 'mosaic' of different activating and inhibitory receptors, the function of each cell being regulated by integration of signals received from ligands presented on potential target cells ([@r26]). Inhibitory signals received mainly from autologous MHC class-I molecules normally dominate, to maintain NK cells in a resting state. However, HCMV not only efficiently downregulates MHC-I ([@r1]; [@r21]; [@r24]; [@r41]; [@r44], [@r45]), but also stimulates the expression of recognized NK cell activating ligands, e.g. MHC-I-related chains (MIC) A and B, UL16-binding proteins (ULBP) 1--3, retinoic acid early transcripts (RAET)1E/ULBP4, RAET1G/ULBP5, RAET1L/ULBP6 and CD155 ([@r4]; [@r5]; [@r6]; [@r13]; [@r18]; [@r19]; [@r23]; [@r40]). Despite this, HCMV-infected cells actually prove to be highly resistant to NK cells in functional assays ([@r12]; [@r40]). This resilience can be attributed to a substantial proportion of HCMV genome being directed towards evading the NK cell response. Although HCMV downregulates endogenous MHC-I, the virus also encodes its own MHC-I homologue (gpUL18) that binds the inhibitory receptor LIR-1 (ILT-2) with high affinity ([@r7]; [@r16]; [@r28]) and a peptide in the UL40 leader sequence that acts to promote cell surface expression of the non-classical MHC-I molecule HLA-E, the ligand for the inhibitory receptor CD94/NKG2A ([@r39]; [@r42]; [@r43]). The activating receptor NKG2D is remarkable in recognizing eight ligands. To combat their activation UL16 retains MICB, ULBP1 and ULBP2 in the endoplasmic reticulum (ER); miR-UL112 targets the MICB transcript, while UL142 downregulates MICA ([@r14]; [@r18]; [@r36]; [@r47]). The NK cell activating receptor DNAM-1 (CD226) recognizes both CD155 and CD112 ([@r10]; [@r20]). We previously demonstrated that UL141 elicits efficient protection against NK cell-mediated cytolysis by sequestering CD155 in the ER yet, in isolation, had no effect on CD112 ([@r40]). CD155 is the poliovirus receptor (PVR) or nectin-like molecule-5 (necl-5), while CD112 is also referred to as nectin-2, herpesvirus entry mediator B (HVEB) or poliovirus receptor-related protein 2 (PRR2). CD112 and CD155 are both structurally and functionally related. Nectins and necls are immunoglobulin-like molecules involved in cell adhesion, movement, proliferation, differentiation, polarization, virus entry and immune recognition ([@r38]). In view of its important role as an activating ligand for DNAM-1, we sought to analyse CD112 expression in the context of HCMV infection. Initial flow cytometry studies revealed that CD112 was downregulated by the low passage HCMV strain Merlin, but not high passage strain AD169 (not shown). Strain AD169 has a 15 kb deletion encompassing UL132--UL150 that includes the NK cell evasion genes UL141 and UL142. Merlin was derived from a bacterial artificial chromosome (BAC) containing the entire strain Merlin genome (R. J. Stanton, unpublished data). MerlinΔUL141 was generated using technologies developed previously to facilitate manipulation of the adenovirus genome ([@r35]). Briefly, a selectable cassette comprising ampicillin resistance, lacZ and SacB was PCR amplified and recombineered into the Merlin BAC in place of nt 184597--185412 (relative to published Merlin sequence GenBank accession no. NC\_006273) using primers SacBF-UL141 (5′-[caggtagcataggaaacatacggtgaaaatactccaaaatcccaaaaatgccgcgattccccgagtggcccagggaga]{.underline}cctgtgacggaagatcacttcg-3′, homology to pAL1111 underlined) and SacBR-UL141 (5′-[ccgacgtttgagcggccgacacacggagcaggaacaggcgggcagcgtctctgcgaaaaagggaagaaaagaatcatc]{.underline}ctgaggttcttatggctcttg-3′, homology to pAL1111 underlined). In a second recombineering step, the selectable cassette was removed using oligo delUL141 (5′-atactccaaaatcccaaaaatgccgcgattccccgagtggcccagggagagatgattcttttcttccctttttcgcagagacgctgcccgcctgttcctg-3′), leaving behind a seamless deletion of the first 816 bp of the UL141 ORF. In human fetal foreskin fibroblasts (HFFF) infected with Merlin, cell surface levels of CD155, CD112 and MHC-I were progressively downregulated over the course of infection (Fig. [1](#f1){ref-type="fig"}), with the change in CD112 being more pronounced at 48 h post-infection (p.i.) (Fig. [1b](#f1){ref-type="fig"}). In accord with previous observations ([@r40]), cells infected with MerlinΔUL141 had elevated cell surface levels of CD155, while CD112 levels were comparable with the mock-infected HFFF (Fig. [1](#f1){ref-type="fig"}). Deletion of UL141 therefore ablated downregulation of both CD155 and CD112. This restoration of CD112 expression was unexpected, since UL141 had no overt effect on CD112 when expressed in isolation ([@r40]). Interestingly, a small reproducible decrease in CD112 persisted when MerlinΔUL141-infected and mock-infected cells were compared at 96 h p.i. (Fig. [1d](#f1){ref-type="fig"}). Replicate samples from the flow cytometry study were analysed by immunoblot, in order to further assess the fate of the CD112 protein within the cell. Briefly, cells were extracted with Triton X-114 ([@r9]), proteins were separated on NuPAGE gels (Invitrogen) and blots were analysed with two independent polyclonal anti-CD112 antibodies. In Merlin-infected cells, the loss of CD155 from the cell surface (Fig. [1](#f1){ref-type="fig"}) correlated with the emergence of elevated levels of an immature (endoglycosidase H-sensitive) form of CD155 complexed with gpUL141 in the ER ([@r17]; [@r40]) (Fig. [2a](#f2){ref-type="fig"}). In contrast to CD155, the CD112 signal gradually decreased in Merlin-infected cells and was not detected by 72 h p.i. (Fig. [2a](#f2){ref-type="fig"}). Quantitative real time-PCR showed CD112 mRNA levels to be marginally increased throughout the infection (not shown), consistent with CD112 expression being regulated post-transcriptionally. To determine whether CD112 was targeted for proteolytic degradation, Merlin-infected cells were incubated in the presence of proteasome inhibitors. Treatment with either MG132 or Epoxomycin (Calbiochem) was able to restore CD112 expression, indicating that HCMV targeted CD112 for proteasome-mediated degradation (Fig. [2b](#f2){ref-type="fig"}). UL141 was required for efficient downregulation of both CD112 and CD155 from the cell surface in HCMV-infected cells (Figs [1](#f1){ref-type="fig"} and [3a](#f3){ref-type="fig"}), yet had no effect on CD112 in cells infected with recombinant adenovirus vector encoding UL141 \[RAdUL141 ([@r40]); Fig. [3b](#f3){ref-type="fig"}\]. We reasoned that UL141 acted in partnership with an additional HCMV-encoded function(s) to downregulate CD112. Indeed, the residual level of CD112 suppression mediated by the MerlinΔUL141 (Figs [1d](#f1){ref-type="fig"}, [2a](#f2){ref-type="fig"} and [3a](#f3){ref-type="fig"}) could potentially be mediated by this function operating suboptimally. In cells co-infected with MerlinΔUL141 and RAdUL141, the HCMV deletion mutant was complemented; downregulation of both CD112 and CD155 was restored (Fig. [3c](#f3){ref-type="fig"}). Similarly, co-infection of strain AD169 with RAdUL141 also resulted in the downregulation of both CD112 and CD155 (Fig. [3d](#f3){ref-type="fig"}). These data are consistent with UL141 co-operating with additional HCMV-expressed function(s) to efficiently downregulate CD112, and that function also being intact within AD169 strain (thus excluding UL133--150). Through downregulation of CD112, HCMV eliminates from the cell surface an activating ligand for DNAM-1, which presumably contributes to the enhanced killing of HCMV-infected cells observed when UL141 is deleted from the virus (Fig. [3e, f](#f1){ref-type="fig"}), but not to the protection elicited when UL141 is expressed in isolation ([@r40]). HCMV thus targets both ligands for the NK cell activating receptor DNAM-1. GpUL141 alone is sufficient to sequester CD155 in the ER, while this study predicts that gpUL141 acts in concert with an additional viral function to induce proteasome-mediated degradation of CD112. This additional viral function could either directly co-operate with UL141, or act upon a cellular intermediate. DNAM-1 is remarkable in being expressed on all NK cells and plays a major role in regulating their function. HCMV suppression of CD112 and CD155 may have ramifications that extend beyond the regulation of NK cell function. DNAM-1 is also expressed on activated T, NKT, myeloid and mast cells, megakaryocytes, platelets and a subset of B lymphocytes thereby impacting on a wide range of immunological responses and regulating platelet activation ([@r3]; [@r10]; [@r11]; [@r25]; [@r27]; [@r29]; [@r30]; [@r31], [@r32], [@r33]; [@r48]). For example, the interaction between DNAM-1 and CD112/CD155 has been associated with T-cell differentiation, proliferation, cytotoxicity and cytokine secretion ([@r37]). Furthermore, nectins and necls regulate fundamental processes in cell biology including cell adhesion, movement, proliferation, differentiation, survival, polarization and signalling ([@r38]). HCMV infection is recognized to disrupt focal adhesions and intercellular connections, while inducing cell motility and transendothelial migration ([@r15]; [@r34]). It will be important to determine how the modulation of CD112 and CD115 influences these processes. Flow cytometers and real-time PCR facility were provided by the Cardiff University Central Biotechnology Service. We are grateful to Siân Llewellyn-Lacey for kindly providing technical support. This work was supported by funding from the Wellcome Trust and MRC. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### HFFF were infected (m.o.i.=25) for (a) 24 h (b) 48 h (c) 72 h or (d) 96 h with HCMV strain Merlin, MerlinΔUL141 or mock-infected and cell surface expression of CD112 (Santa Cruz, sc-65333) was analysed by flow cytometry. For reference, expression levels of CD155 (Abcam, ab-3142) and MHC class-I (W632; ATCC) were also monitored, alongside control Ig (cIg). :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### HFFF were infected (m.o.i.=25) for 24, 48, 72 or 96 h p.i. with HCMV strain Merlin (Mer), MerlinΔUL141 (Δ141) or mock-infected (Mock) and cell extracts were analysed by immunoblot using antibodies to: CD112 (R&D, AF2229; Santa Cruz, sc-14799), CD155 \[5D1 ([@r2])\], UL141 \[M550 ([@r40])\] and actin (A-2066; Sigma). (b) HFFF were infected (m.o.i.=25) for 48 h with HCMV strain Merlin (Mer) or mock-infected, then treated for 12 h with proteasome inhibitors MG132 or Epoxomycin as indicated and analysed by immunoblot as in (a). :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### HFFF were infected for 72 h (m.o.i.=25) with HCMV strain Merlin or MerlinΔUL141, as indicated, and analysed for cell surface expression of CD155 and CD112 by flow cytometry. (b) HFFF were infected for 72 h (m.o.i.=200) with replication-deficient adenovirus vectors encoding HCMV UL141 (RAd-UL141) or equivalent empty RAd (RAd-CTRL) ([@r40]), as indicated, and analysed for cell surface expression of CD155 and CD112 by flow cytometry. (c) HFFF were co-infected for 72 h with MerlinΔUL141+RAd-CTRL or MerlinΔUL141+RAd-UL141, as indicated, and analysed for cell surface expression of CD155 and CD112 by flow cytometry. (d) HFFF were co-infected for 72 h with HCMV strain AD169+RAd-CTRL (AD169) or AD169+RAd-UL141, as indicated, and analysed for cell surface expression of CD155 and CD112 by flow cytometry. Control Ig histograms (cIg) were not included in panels (a), (c) and (d) to maintain figure clarity. (e) HFFF were infected for 72 h with HCMV strain Merlin, MerlinΔUL141 or mock infected. Sensitivity to NK cells was measured using alpha interferon (IFN-*α*) activated PBMC in allogeneic CD107a mobilization assay ([@r28]) using the following antibodies: anti-CD107a-FITC (553793; BD Biosciences), anti-CD3-PerCP (SK7; BD Biosciences), anti-CD56-APC (N901; Beckman Coulter). PBMC incubated without targets and K562 cells are shown as controls. (f) RS primary skin fibroblasts were infected for 72 h with HCMV strain Merlin, MerlinΔUL141 or mock infected. Sensitivity to NK cells was measured using IFN-*α* activated RS PBMC in autologous CD107a mobilization assay as described in (e). :::  ::: [^1]: †These authors contributed equally to this work.

PubMed Central

2024-06-05T04:04:19.368064

2010-8-01

PMC3052561

Human adenovirus (HAdV) frequently causes subclinical infections and is additionally associated with acute respiratory tract outbreak, epidemic kerato-conjunctivitis and acute gastroenteritis. Life threatening systemic infections occur in immunosuppressed patients after allogeneic stem cell transplantation (SCT) as well as solid organ transplantation. Incidence of HAdV infection is highest in children after allogeneic SCT. Risk factors are any cause of severe T-cell deficiency post SCT, like T-cell depletion of the graft, alternative donor SCT, delayed T-cell reconstitution and graft-versus-host-disease ([@r3]; [@r6]; [@r9]; [@r17]). In contrast to the widespread use of adenovirus (AdV) for gene therapy or vaccination trials ([@r14]), data on infection of immunologically relevant cell types by clinical adenoviruses are sparse at best. It is yet unknown how the interplay between the innate and adaptive immune response is affected by HAdV, or how antigen presenting cells are modulated by the infection. Dendritic cells (DC) play a crucial role at the crossroad between innate and adaptive immune response as they can prime and activate T cells and modulate the type of T-cell response. This makes them an attractive target for viruses as many viruses are known to impair DC maturation and function. Monocyte-derived DC (mdDC) have evolved as an ideal cell type in order to study the effects of viral infections, for example human cytomegalovirus (HCMV) ([@r5]; [@r7]; [@r10]; [@r11], [@r12]; [@r13]). It was thought until very recently that there were no data on the susceptibility of DC for adenoviruses ([@r1]). Langerhans cells and dermal DC express low levels of the coxsackievirus and adenovirus receptor (CAR), while mdDC lack this main receptor for the clinically most relevant subgroup of HAdV, Ad5. HAdV overcomes this limitation by using integrins as alternative entry receptors. Additionally, it has now been demonstrated that lactoferrin in association with DC-SIGN facilitates entry of recombinant HAdV5 into mdDC by mechanisms independent of CAR ([@r1]). Up to now there has not been any data about the interactions of clinical HAdV strains with DC and their capability to stimulate T cells. We investigated the susceptibility of DC and their precursor cells for wild-type adenovirus infection and analysed their T-cell stimulatory capacity. We infected monocytes, immature and mature mdDC with the clinical HAdV strain BB2000-61 (subgenus C/serotype 2). We isolated BB2000-61 from the blood of an SCT recipient (unpublished data). Subsequently, we analysed the expression of distinct surface markers in response to exposure to HAdV and finally tested the T-cell stimulatory capacity of HAdV-infected DC. It could be demonstrated that both monocytes and mdDC are susceptible to an infection with a clinical HAdV strain. The infection of immature mdDC resulted in an increased expression of major histocompatibility complexes (MHC) class I and class II, while expression of costimulatory molecules was heterogeneous. When mature mdDC were infected with BB2000-61, expression of MHC and costimulatory molecules remained stable or decreased. The infection of monocytes resulted in a block of differentiation towards immature DC. Finally, BB2000-61-infected DC stimulated CD8^+^ T cells. Up to now there has not been any publication as to whether wild-type HAdV can infect DC. In order to address this question, we inoculated immature mdDC with BB2000-61. We analysed the replication of BB2000-61 and an HAdV reference strain of subgenus C/serotype 5 (provided by A. Heim, Medizinische Hochschule Hannover, Germany). DC were monocyte derived ([@r5]). In brief, peripheral blood mononuclear cells (PBMC) were separated from buffy coats by density-gradient centrifugation followed by CD14-based magnetic positive-selection of monocytes. Monocytes were used for experiments or incubated in six-well plates in RPMI with 10 % fetal calf serum, 100 μg gentamicin ml^−1^, 1000 U IL-4 ml^−1^ and 100 ng GM-CSF ml^−1^. After 7 days, immature DC were rinsed off the plates and controlled morphologically as well as for expression of CD1a, CD83, MHC I, MHC II, CD80, CD86 and CD40. For infection experiments, cells were incubated with different m.o.i. amounts of strain BB2000-61 and further cultivated. Thereafter, the cells were spun onto glass slides and stained for viral early (E) and late (L) antigens. Monoclonal antibodies (mAbs) were used directed against the E protein E1A (ab33183; Abcam; IgG2a; Cambridge) and the L hexon protein (C5000; Virion or Chemicon). Cytospins were fixed in acetone for 10 min at room temperature and incubated with the HAdV mAbs for 90 min at 37 °C, followed by incubation with Cy3-conjugated F(ab)′2 fragments of goat anti-mouse IgG antibodies (Jackson ImmunoResearch). Finally, cells were counterstained with 4,6-diamidino-2-phenylindole (DAPI). Primary monocytes were susceptible to infection with the clinical HAdV strain as demonstrated by the expression of the E and L genes (Fig. [1](#f1){ref-type="fig"}). Additionally, immature mdDC could be infected with BB2000-61. When different m.o.i. amounts were compared for the resulting efficiency of infection, the percentage of infected cells ranged between 29 and 39 % at an m.o.i. between 1 and 50 with a peak efficiency of 39 % at an m.o.i. of 10. The range was the same for monocytes and mature mdDC. Intriguingly, higher m.o.i. amounts did not result in higher efficiencies of infection, but more cell debris was observed. The reference strain was also able to infect mdDC and monocytes. Further experiments were done with BB2000-61. This demonstrated for the first time that DC are a target for a clinical HAdV strain and both immature DC and their progenitors can be infected. Additionally, the infection did not abort and both the E and L proteins were expressed. In order to assess if mature mdDC are also susceptible to infection with HAdV, we added 1 μg LPS ml^−1^ or 10 μg poly (I : C) (Sigma Aldrich) ml^−1^ after 7 days and cultivated them for a further 24 h. This induced DC maturation was demonstrated by upregulation of MHC class I, class II, CD40, CD80, CD86 and CD83 expression (data no shown). Mature DC were inoculated with BB2000-61 as described for immature DC and cultivated for a further 5 days. Mature mdDC had approximately the same susceptibility as immature mdDC. This shows that the myeloid lineage from monocytes to mature DC can be a target for HAdV, and replication at least reaches the late phase. The reference strain showed similar properties in infectivity, but was not further included in flow cytometry experiments/T-cell stimulation. Many viruses are known to target DC and impair their function in order to subvert adaptive immunity. For example, HCMV infection impairs formation and intracellular transport of MHC molecules, resulting in a decreased surface expression of the respective molecules ([@r2]; [@r5]; [@r8]; [@r11]; [@r16]). In order to assess how a clinical HAdV strain acts in this regard, we examined the surface expression of immunologically relevant cell markers in mdDC after infection. To this end immature mdDC were infected with BB2000-61 and incubated for a further 5 days. Then the expression of various surface markers was analysed. Cytometry was performed using fluorescein isothiocyanate- or phycoerythrin-conjugated mAbs against surface markers: CD40 (Immunotech) and CD80, CD83, CD86, human leukocyte antigen (HLA)-DR, CD14, HLA-A, -B and C (W6/32), CD1a and IgG1 isotype control (Pharmingen). Infection of immature mdDC with HAdV results in a rather heterogeneous immunophenotype as some markers were upregulated, while others remained stable or were downregulated. In detail, expression of MHC class I and class II as CD86 increased, while CD80 was unaffected. CD83 was slightly downregulated, while CD1a was strongly reduced (Fig. [2a](#f2){ref-type="fig"}). Manipulating the expression of functional surface markers in infected DC is a known mechanism of viral immune evasion ([@r5]; [@r15]). While HCMV as a prominent example causes a somewhat uniform downregulation of expression of both MHC and costimulatory molecules, HAdV induces a more heterogeneous picture. Both MHC class I and class II are upregulated as well as CD86 and CD40. All these molecules are important for a successful interplay between DC and T cells. On the other hand, infected mdDC show no upregulation of CD83, a marker for mdDC maturation. As mature mdDC were also susceptible to HAdV, we analysed the immunophenotype. By impairing the surface expression of MHC and costimulatory molecules, HAdV could possibly interfere with T-cell priming and activation even after the initial activation of the DC. Therefore, we induced maturation of immature mdDC by the addition of 1 μg LPS ml^−1^ or 10 μg poly (I : C) (Sigma-Aldrich) ml^−1^. Cells were further cultivated and infected with BB2000-61 as described above. Stimulation with LPS and poly (I : C) resulted in an increased expression of MHC class II and a more pronounced expression of CD80 and CD86. Comparing the stimulatory performance, poly (I : C) was superior to LPS. Nevertheless, when HAdV-infected samples were compared with mock, HAdV infection resulted in a decreased surface expression of CD80 and CD86. Also, poly (I : C)-stimulated DC expression of MHC class II was decreased, while it remained stable in LPS-matured cells. Intriguingly, CD1a was expressed in higher amounts on HAdV-infected LPS-matured DC than in poly (I : C) stimulated ones. LPS and poly (I : C) induce DC maturation by binding to TLR4 and TLR3, respectively. While TLR4 is located on the cell surface and common on many DC subsets, TLR3 resides in endosomes and is a unique feature of DC of the myeloid lineage. The different pattern recognition receptors, their localization and their signalling pathways possibly are reflected in the variable response to LPS and poly (I : C). Overall, maturation induced by poly (I : C) seems to be more thorough as reflected by a stronger upregulation of surface marker expression (Fig. [2c](#f2){ref-type="fig"}) and additionally is permanent. After infection with HAdV, the immunophenotype of mdDC matured with both LPS and poly (I : C) is skewed to the level or below of immature mdDC. The upregulation of MHC II and primarily CD80 and CD86 that accompanies DC maturation is reverted by HAdV. The infection and subsequent manipulation of mature DC could be one module of a multi-component immune evasion strategy by HAdV. As immature DC respond to HAdV infection with an upregulation of certain activating molecules, formerly uninfected mature DC could be impaired in their function. This would be especially relevant for DC cross-presenting antigens without direct contact to an antigen. One important subtype of DC results from the differentiation of monocytes into immature DC under the influence of cytokines. As we demonstrated, monocytes are susceptible to infection with a clinical HAdV strain. Infection of the direct progenitor of the cell type that links innate and adaptive immunity could be beneficial to a virus, since it could modulate its differentiation process and function. To test this assumption, we infected monocytes and added the cytokines necessary for differentiation into immature mdDC. After 7 days in culture (i.c.), we compared the cells to both the freshly isolated monocytes and non-infected cells with the same treatment regarding their morphology and surface marker expression (Fig. [2b](#f2){ref-type="fig"}). Viewed by microscope, monocytes appear as round bright cells, while DC are less bright and feature their characteristic shape. The majority of the HAdV-infected monocytes remained as round and bright as freshly isolated monocytes even after 7 days i.c., while a small fraction showed the typical shape and colour of macrophages. When the freshly isolated mock-infected monocytes were compared with another fraction of the same monocytes that was differentiated into mdDC, the latter showed all signs of differentiation: CD1a, MHC class II, CD80, CD86 and CD40 expression was upregulated, while CD14 expression was almost diminished. In contrast, HAdV-infected cells showed no signs of differentiation into DC. CD1a, MHC class II, CD80, CD86 and CD40 expression were downregulated, with the exception of CD86 even below the level expressed by monocytes without cytokine stimulation. Additionally, expression of CD14 was increased above monocyte level. All these data, together with the morphological observation, demonstrate that HAdV-infected monocytes cannot be differentiated into immature DC despite stimulation by IL-4 and GM-CSF. DC are crucial for the induction of an adaptive immune response to viral infection. One of the DC subtypes, the mdDC directly differentiates from monocytes in the appropriate cytokine milieu. By the infection of this progenitor cell, HAdV is able to block differentiation to immature DC, thereby probably blocking the presentation of HAdV-derived antigens by these cells. Together with the potential to infect mature DC followed by a decreased expression of immunological activating surface markers, this could overcome the inability to impair expression of MHC and costimulatory molecules in HAdV-infected immature DC. The ability to induce T-cell activation and proliferation is a hallmark of DC function. This ability summarizes all singular features like expression of MHC and costimulatory molecules, acquiring, processing and displaying antigens and secreting cytokines. Therefore, the success of viral subversion of DC function has to be measured by the virus\' capability to evade T-cell proliferation. In order to assess the effect of adenoviral infection of mdDC on their T-cell stimulatory capacity, we performed a mixed lymphocyte reaction. DC were irradiated and co-cultured with autologous PBMC. After restimulation with inactivated viral particles, *in vitro* proliferation was detected with carboxyfluorescein succinimidyl ester (CFSE) ([@r4]). Mock-infected DC stimulated 0.57 % of the autologous CD8^+^ T cells to proliferate, while direct stimulation of T cells with staphylococcus enterotoxin B resulted in a proliferation of 33 % of the CD8^+^ T cells. When HAdV-infected mdDC were incubated with the cells, 5.44 % of them proliferated (Fig. [3](#f3){ref-type="fig"}). This shows that the adenoviral infection of DC does not result in a block of CD8^+^ T-cell stimulation as can be seen in other viruses with similar impact on immune response. In comparison, HCMV leads to a remarkable decrease of the T-cell stimulatory capacity of infected DC ([@r5]; S. Schempp, personal communication). In contrast to this betaherpesvirus, HAdV-infected DC are still able to stimulate CD8^+^ T cells and thereby induce a cellular immune response. Supported by a grant from the German Science Foundation (DFG, SFB 685). ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Expression of HAdV E and L antigens in infected cells (strain BB2000-61). Expression of E and L antigens was visualized 3 and 5 days post-infection. Cells were either mock infected or infected with BB2000-61. Mock is supernatant from cells treated exactly as infected ones except for the addition of HAdV. The m.o.i. of 1 resulted in 31 %, m.o.i. of 10 in 39 %, m.o.i. of 20 in 29 % and m.o.i. of 50 in 31 % positive-stained immature DC with E1A-staining. E1A antibody was diluted 1 : 100. Hexon specific antibody was diluted 1 : 200. Secondary antibody, anti-mouse immunoperoxidase was diluted 1 : 500. Counterstaining with DAPI. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Expression of immunologically relevant surface markers after HAdV infection with an m.o.i. of 10. (a) Freshly isolated DC were stained with antibodies specific for the designated markers or differentiated into mdDC by the addition of IL-4 and GM-CSF and cultivation for 7 days, or infected with HAdV and cultivated for a further 7 days in the presence of IL-4 and GM-CSF then stained and analysed by flow cytometry. (b) Freshly isolated monocytes were infected with the strain BB2000-61 and cultivated for a further 5 days. (c, d) Mature DC were infected with the same virus. The experiments were performed twice. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### Stimulation of autologous T cells detected by CFSE. HAdV-infected DC were irradiated and co-cultured with autologous PBMC. Mock-infected DC stimulated 0.57 % of the autologous cytotoxic T cells to proliferate, while HAdV-infected mdDC induced an increase in proliferation to 5.44 %. Analysis of proliferation among cytotoxic T cells was done using a gate on CD3^+^/CD8^+^ cells. :::  :::

PubMed Central

2024-06-05T04:04:19.369487

2010-5-01

PMC3052562

INTRODUCTION ============ Influenza A virus has a negative-sense RNA genome split into eight viral (v) RNA segments. Unusually for an RNA virus without a DNA-encoding stage, vRNA synthesis occurs in the nucleus ([@r1]). Here, the vRNAs are transcribed by the viral RNA-dependent RNA polymerase to produce 10 major species of capped and polyadenylated mRNA. Although these transcripts are structurally similar to mature host mRNAs, their terminal modifications are achieved through mechanistically distinct routes. Viral mRNAs contain 5′-cap structures that are synthesized by the cellular capping machinery but recycled through the process of 'cap-snatching', in which a cap-dependent endonuclease activity in the viral polymerase generates capped oligonucleotide primers from cellular pre-mRNAs ([@r37]). Poly(A) tails are synthesized by the viral polymerase stuttering on a poly(U) tract near the 5′-end of each vRNA segment ([@r37]), rather than by the host 3′-end processing machinery and poly(A) polymerase. The maturation pathway of most of the viral transcripts also differs from that of cellular mRNAs. While the majority of cellular pre-mRNAs contain introns that are completely removed before nuclear export, most influenza mRNAs (those from segments 1--6, encoding PB2, PB1, PA, HA, NP and NA, respectively, in their largest ORFs) are intronless. Furthermore, although segments 7 and 8 contain introns, the majority (∼90 %) of the primary transcripts are exported unspliced to produce the abundant M1 and NS1 polypeptides, respectively, while the processed minority encode the M2 and NS2/NEP proteins ([@r28]). Thus, the 10 protein-coding influenza virus mRNAs can be classified into three structural classes: intronless, intron-containing but unspliced and fully spliced. Viral mRNAs can also be subdivided into 'early' (segments 1--3, 5 and the unspliced segment 8 transcript) and 'late' (segments 4, 6, 7 and spliced segment 8) classes, based on the temporal expression patterns of their protein products ([@r18]; [@r26]; [@r29]; [@r46]). However, these structural or kinetic classifications do not neatly correspond. The mechanism(s) by which influenza virus mRNAs reach the cytoplasm is unclear. Cellular mRNAs are predominantly exported in the form of messenger ribonucleoprotein complexes (mRNPs) via the NXF1/TAP pathway. Formation and recruitment of these mRNPs to the NXF1 export pathway is closely coupled to Pol II transcription and pre-mRNA maturation, including capping, splicing and polyadenylation steps ([@r12]; [@r13]; [@r21]). During synthesis and processing, a cellular mRNA dynamically acquires a complex set of proteins, including various heterogeneous nuclear ribonucleoproteins (hnRNPs), poly(A)-binding proteins and the transcription-export (TREX) complex. TREX, consisting of the THO complex, UAP56 (BAT1) and Aly (REF/THOC4), is primarily recruited to the pre-mRNA during splicing ([@r31]; [@r48]) but also through interactions with the 5′-cap structure ([@r7]; [@r38]). Aly then recruits NXF1 and its co-factor p15, which displaces Aly and allows NXF1/p15 to bind the RNA directly ([@r19]). NXF1--p15 interactions with the nuclear pore complex then direct export of the mRNP into the cytoplasm. The synthesis of influenza mRNAs by a viral polymerase and the fact that most are either intronless or contain residual introns raises the question of how, or even if, they are fed into the NXF1 pathway. Alternative cellular mRNA export pathways exist, the best characterized of which utilizes the CRM1/exportin 1 factor ([@r5]; [@r25]). Other nuclear replicating viruses such as herpesviruses manage to export intronless mRNAs by encoding adaptor proteins that functionally replace Aly and link the viral transcripts to the NXF1 pathway ([@r3]; [@r47]). Retroviruses export intron-containing mRNAs through *cis*-acting RNA sequences that (directly or via a viral adaptor protein) recruit either NXF1- or CRM1-dependent pathways ([@r9]; [@r14]). Indirect evidence argues both for and against influenza virus possessing analogous mechanisms. Viral gene expression is insensitive to leptomycin B (LMB) treatment ([@r10]), as is the export of mRNAs from segments 6 and 7 ([@r2]; [@r54]), arguing against the use of the CRM1 pathway. However, drugs that inhibit Pol II lead to reversible nuclear retention of segment 4 and 7 but not segment 5 mRNAs ([@r2]; [@r53]). Transcription-dependent export of cellular mRNA has also been demonstrated ([@r22]; [@r52]), suggesting that at least some influenza mRNAs use cellular export machinery that is ordinarily coupled to Pol II transcription. Consistent with this, an siRNA screen for proteins involved in the early stages of influenza virus replication in *Drosophila* cells identified NXF1 as important for expression of a chimaeric reporter gene based on segment 6, as well as for overall virus replication in mammalian cells ([@r16]). It has also been shown that influenza mRNAs from segments 2, 6 and 7 interact with NXF1 ([@r54]). However, a direct test of whether NXF1 functions in the export of any or all the influenza virus mRNAs has not been reported. Furthermore, it has also been proposed that influenza virus NS1 antagonizes the export function of NXF1 as a means of contributing to host-cell shut off ([@r41]). Accordingly, to clarify the mechanism(s) by which influenza A virus mRNAs are exported from the nucleus, we set out to directly test the involvement of the cellular NXF1 and CRM1 pathways. We found no evidence for the involvement of the latter system in influenza mRNA export, whereas export of all viral transcripts tested showed some degree of sensitivity to NXF1 levels. However, a gradient of dependence was evident, with late gene mRNAs being the most sensitive and the intronless early gene transcripts being the least sensitive. RESULTS ======= Previously, we showed that inhibition of RNA Pol II transcription blocked the nuclear export of segment 7 mRNAs (encoding M1/M2), but not segment 5 (NP) mRNA ([@r2]). Analysis of viral protein expression in the presence of the drug 5,6-dichloro-1-*β*-[d]{.smallcaps}-ribofuranosyl-benzimidazole (DRB), an inhibitor of processive Pol II elongation ([@r8]; [@r15]; [@r49]), showed reductions in the synthesis of HA and NS1 which suggested that ongoing Pol II transcription was necessary for the export of other viral mRNAs than just that encoding M1 ([@r2]). However, the processing and export of cellular pre-mRNAs is also associated with the nuclear loading of protein factors that subsequently upregulate translation of the mature transcripts in the cytoplasm ([@r13]; [@r21]). We therefore tested directly whether DRB treatment inhibited export of other influenza virus mRNAs than those from segment 7. Cells were infected (or mock infected) with influenza A/PR/8/34 (PR8) virus and 90 min post-infection (p.i.) were either treated with DRB or left untreated. At 6.5 h p.i., the cells were fixed and analysed by fluorescence *in situ* hybridization (FISH) for the localization of positive-sense RNA from segments 1, 4, 5, 7 and 8. Only background levels of signal were seen from uninfected cells (Fig. [1a](#f1){ref-type="fig"}). In untreated cells all of the viral mRNAs tested were found to be predominantly cytoplasmic (Fig. [1b](#f1){ref-type="fig"}). When infected cells were treated with DRB, segment 5 mRNA still remained mostly cytoplasmic, while as expected ([@r2]) segment 7 transcripts showed pronounced nuclear retention (Fig. [1c](#f1){ref-type="fig"}). Segment 4 mRNA (encoding HA) also showed almost total nuclear retention after DRB treatment, while transcripts from segments 1 (PB2) and 8 (NS1/NS2) showed an intermediate phenotype with obvious nuclear retention, but with residual cytoplasmic staining still clearly visible (Fig. [1c](#f1){ref-type="fig"}). Thus, nuclear export of the majority of influenza mRNAs is sensitive to DRB treatment and the loss of NS1 and HA protein expression under these conditions ([@r2]; [@r50]) directly reflects reduced nuclear export of their mRNAs. However, a gradient of responsiveness to the inhibitor was apparent, with the transcripts of early genes generally showing reduced sensitivity than those of late genes. Given the requirement for ongoing Pol II transcription for both influenza and cellular mRNA export ([@r2]; [@r22]; [@r52]), it was reasonable to hypothesize that the cellular mRNA export machinery was involved. Accordingly, we targeted key components of the cellular bulk mRNA export machinery: NXF1, Aly and UAP56, for siRNA-mediated knockdown. We also targeted hnRNPA1, a protein involved in cellular mRNP biogenesis, as well as glyceraldehyde-3-phosphate dehydrogenase (GAPD) as a control for non-specific siRNA effects. Both UAP56 and hnRNPA1 have been identified as interaction partners of influenza A virus NP/RNPs ([@r32]; [@r34]). Lysates from siRNA-treated human embryonic kidney (293T) cells were first analysed by Western blotting to confirm successful knockdown of the target proteins. In all cases, a reduction in the levels of the targeted protein in comparison with control or untreated cells was apparent (Fig. [2a](#f2){ref-type="fig"}). When blots from replicate experiments were quantified, normalized to actin levels and compared with untreated cells, NXF1 levels were reduced by 75, Aly by 60, UAP56 by 80 and hnRNPA1 by 90 % at 48 h post-transfection (Fig. [2b](#f2){ref-type="fig"}), confirming successful reductions in expression. To establish whether depletion of these cellular proteins affected influenza virus replication, siRNA-treated cells were infected with PR8 virus and the output titres determined by plaque assay. The effect of DRB on virus replication was also assessed, as a known inhibitor of influenza replication that functions at least in part by blocking viral mRNA export and/or gene expression ([@r2]; [@r50]); this treatment reduced virus titres by around 50-fold compared with untreated cells (Fig. [2c](#f2){ref-type="fig"}). Cells in which GAPD had been targeted for knockdown showed only a slight reduction in their ability to replicate virus, as did cells depleted of hnRNPA1 or Aly (Fig. [2c](#f2){ref-type="fig"}). However, targeting of NXF1 reduced the output virus titre approximately 100-fold, while knockdown of UAP56 reduced replication by nearly 10-fold (Fig. [2c](#f2){ref-type="fig"}). These data, consistent with a previous study that examined knockdown of NXF1 ([@r16]), strongly suggested that NXF1 was involved in influenza replication, while UAP56 may also be important. To better define the role of these cellular proteins in the influenza virus life cycle, we next examined viral protein expression in siRNA-treated cells. Lysates from cells treated with siRNAs (or later treated with DRB as a positive control) and then infected with virus were analysed by Western blotting for a panel of viral proteins, including the early class gene products PB2 (as a representative polymerase protein), NP and NS1, the late proteins HA and M1 and the two products of spliced mRNAs, M2 and NS2. RT-PCR analysis of the relative abundance of the M1 and M2 mRNAs indicated that none of the siRNA treatments had a major effect on splicing of segment 7 transcripts (data not shown). The levels of cellular actin were also examined to ensure equal quantities of cell lysate had been tested (Fig. [3a](#f3){ref-type="fig"}, bottom row). Viral proteins were readily detectable in samples from untreated infected cells, but not in mock infected cell lysates (Fig. [3a](#f3){ref-type="fig"}, lanes 7 and 8). As observed previously ([@r2]; [@r50]), DRB treatment profoundly decreased levels of the late viral proteins M1 and HA and reduced NS1 and M2 synthesis, while leaving NP and NS2 accumulation unaffected (Fig. [3a](#f3){ref-type="fig"}, lane 6). The partial retention of segment 1 mRNA seen after DRB treatment (Fig. [1](#f1){ref-type="fig"}) correlated with a threefold reduction in PB2 accumulation (Fig. [3a](#f3){ref-type="fig"}, lane 6; see Fig. [3b](#f3){ref-type="fig"} for quantification). When viral protein levels were examined in siRNA-treated cells, depletion of GAPD or hnRNPA1 had no effect (Fig. [3a](#f3){ref-type="fig"}, lanes 4 and 5; quantification in Fig. [3b](#f3){ref-type="fig"}). Depletion of Aly also had little consequence for viral protein levels, causing at the most a twofold reduction in M2 accumulation (Fig. [3a](#f3){ref-type="fig"}, lane 2). Knockdown of UAP56 had more effect; while PB2, NP and NS2 levels were unaffected, NS1, HA, M1 and M2 accumulation were all reduced by threefold or more, with M2 being the most severely affected (Fig. [3a](#f3){ref-type="fig"}, lane 3 and Fig. [3b](#f3){ref-type="fig"}). However, consistent with its effect on virus titre, depletion of NXF1 had the greatest effect on viral gene expression, severely reducing M2 and HA protein levels and also reducing M1, NS1 and NS2 accumulation by around threefold. Similarly to DRB treatment, NXF1 depletion had little effect on NP accumulation, while PB2 levels were also unaffected. Thus, reduced quantities of certain components of the cellular mRNA export machinery leads to much reduced synthesis of particular influenza virus proteins. The most obvious cause for this would be inhibition of viral mRNA export. To test this we infected siRNA-treated cells and analysed the localization of viral mRNA by FISH. Positive-sense transcripts from segments 1, 4, 5, 7 and 8 were all largely cytoplasmic in cells treated with siRNAs targeting GAPD (Fig. [4a](#f4){ref-type="fig"}), similar to the pattern seen in untreated cells (Fig. [1b](#f1){ref-type="fig"}). This pattern was not altered by siRNA depletion of either hnRNPA1 or Aly (Fig. [4e, c](#f4){ref-type="fig"}, respectively). Quantification of the nuclear : cytoplasmic ratio of staining intensity seen for the FISH probes from multiple cells confirmed that there was no substantial change in localization of any of the viral transcripts examined (Fig. [5](#f5){ref-type="fig"}). Depletion of UAP56 also had little effect on the localization of mRNAs from segments 1, 4 and 5, either by visual inspection (Fig. [4d](#f4){ref-type="fig"}) or by quantification of fluorescence intensity (Fig. [5](#f5){ref-type="fig"}). However, substantial nuclear retention was obvious for mRNAs from segments 7 and 8 (Fig. [4d](#f4){ref-type="fig"}), resulting in between 8 and nearly 30-fold increases in the relative amount of nuclear staining (Fig. [5](#f5){ref-type="fig"}). Depletion of NXF1 caused a similarly large decrease in the amount of cytoplasmic mRNA from segments 7 and 8, but pronounced nuclear retention of segment 4 was also apparent (Figs [4b](#f4){ref-type="fig"} and [5](#f5){ref-type="fig"}). A slight (around twofold) increase in nuclear : cytoplasmic staining ratios was also seen for segment 1 and 5 mRNAs. Thus, UAP56 is required for the export of segment 7 and 8 mRNAs, while NXF1 is required for the efficient nuclear export of most influenza A virus mRNAs, but to a degree that depends on the specific transcript examined. Also, the similar effects DRB treatment and NXF1 depletion have on viral mRNA localization coupled with previous work showing that DRB treatment inhibits bulk cellular mRNA export ([@r22]; [@r52]) suggests that the drug inhibits NXF1-dependent mRNA export. The low sensitivity of intronless early gene transcripts to the effects of NXF1 depletion or DRB treatment could result from their use of a non-NXF1-dependent export pathway. Alternatively, these mRNAs might still use NXF1 but recruit it via a different, more efficient mechanism than that used by the late gene/spliced transcripts, which largely bypasses the blocks introduced by DRB or gene silencing. As a first test of these possibilities, we examined whether influenza virus mRNA export was sensitive to LMB, as CRM1 is the best delineated alternative pathway for mRNA nuclear export ([@r5]; [@r25]). Madin--Darby canine kidney (MDCK) cells were infected with virus, treated (or mock treated) with LMB from 1.5 h p.i. and viral mRNA localization examined by FISH. Parallel analysis of the intracellular localization of NP showed its LMB-dependent nuclear retention as expected ([@r10]), confirming activity of the drug (data not shown). Again, positive-sense transcripts from segments 1, 4, 5 and 7 were largely detected in the cytoplasm in untreated cells (Fig. [6](#f6){ref-type="fig"}, middle panels). Also consistent with previous results ([@r2]; [@r54]), localization of mRNA from segment 7 was unaltered by LMB treatment (Fig. [6](#f6){ref-type="fig"}, lower right). Moreover, trafficking of segment 1, 4 and 5 transcripts were also unaffected by the drug. Similar results were obtained in 293T and BHK cells (data not shown). Thus, while not eliminating the hypothesis that the early intronless viral mRNAs utilize a non-NXF1-dependent export pathway, these data indicate that they do not use a CRM1-dependent mechanism. DISCUSSION ========== Like any virus that transcribes its genome in the nucleus of infected cells, influenza A must ensure that its mRNAs reach the cytoplasm. Retroviruses and nuclear-replicating DNA viruses face a similar problem but for influenza, it is perhaps rendered more difficult by an RNA genome and consequent use of a virally encoded RNA-dependent RNA polymerase. We show here that despite not directly using Pol II to transcribe its genome, influenza virus nevertheless co-opts the cellular NXF1 export machinery for at least a subset of its mRNAs. We also show that not all viral mRNAs interact equally with this nuclear export pathway, with individual viral transcripts displaying a gradient of susceptibility to treatments that disrupt bulk mRNA export. Differing requirements for the helicase UAP56 were also seen. These data allow us to start building a model for how the individual viral mRNAs reach the cytoplasm (Fig. [7](#f7){ref-type="fig"}). Our conclusion that the NXF1 pathway is used for influenza virus mRNA export is based on several strands of evidence. Firstly, it is supported by prior work showing that NXF1 is required for viral replication ([@r16]) and that mRNAs from segments 2, 6 and 7 can be co-immunoprecipitated with NXF1 ([@r54]). More directly, however, we show for the first time that loss of cellular NXF1 strongly inhibited nuclear export of segment 4, 7 and 8 mRNAs (Figs [4b](#f4){ref-type="fig"} and [5](#f5){ref-type="fig"}). The FISH probes we employed do not discriminate between spliced and unspliced transcripts from the latter two vRNAs, but synthesis of all four polypeptides encoded by the segments were reduced (Fig. [3](#f3){ref-type="fig"}), suggesting that export of the minority spliced species (especially that of M2) was also inhibited. Transport of segment 4, 7 and 8 mRNAs, as well as expression of the proteins they encode (excepting NS2), were also strongly retarded by the Pol II inhibitor, DRB (Figs [1](#f1){ref-type="fig"} and [3](#f3){ref-type="fig"}). In contrast, the nucleo-cytoplasmic distribution of the intronless mRNAs from segments 1 and 5 were relatively unaffected by NXF1 depletion, displaying only a twofold change (Fig. [5](#f5){ref-type="fig"}) without a matching reduction in accumulation of the translation products (Fig. [3b](#f3){ref-type="fig"}). DRB also had little effect on NP mRNA localization or translation, while partial nuclear retention of segment 1 mRNA was seen, only a threefold drop in protein accumulation resulted (Figs [1](#f1){ref-type="fig"} and [3b](#f3){ref-type="fig"}). The selective nature of the defects in viral gene expression arising from NXF1 or UAP56 depletion also argue that the requirement for the cellular export machinery is direct (and not simply the consequence of lowered cell viability), since the siRNA-treated cells were still readily infectable with virus and able to support normal synthesis of certain viral polypeptides. This is also consistent with the findings of [@r16] who found that NXF1 depletion of HEK cells did not dramatically affect cell viability over the time-spans used here. The differing susceptibilities individual viral mRNAs showed to siRNA depletion of cellular export factors or DRB correlated better with the kinetic class of the viral gene product than with mRNA structure. Intronless transcripts for early gene products (in particular segment 5/NP mRNA) but also segment 1 (PB2) showed the least dependence on the NXF1 pathway (Fig. [7a](#f7){ref-type="fig"}), while late genes, including the intronless mRNA encoding HA, the spliced mRNA for M2 and the intron-containing but unspliced M1 message showing the clearest dependence (Fig. [7b--d](#f7){ref-type="fig"}). We have not examined the susceptibility of segment 6 (NA) mRNA to NXF1 depletion but [@r54] showed an association between the two molecules, while [@r16] reported that NXF1 depletion blocked expression of an artificial reporter mRNA based on segment 6. It therefore seems plausible that the NA mRNA has a similar export mechanism to the HA mRNA (Fig. [7b](#f7){ref-type="fig"}). The correlation between the degree of dependence on NXF1 and the kinetic class of the viral gene product is not perfect however, as expression of the late protein NS2 (from the spliced segment 8 mRNA) was less sensitive to DRB than expression of the early protein NS1 from the unspliced transcript (Fig. [3b](#f3){ref-type="fig"}) and the export of the majority population of positive-sense mRNA from segment 8 was inhibited by both DRB and NXF1 depletion (Figs [1](#f1){ref-type="fig"}, [4](#f4){ref-type="fig"} and [5](#f5){ref-type="fig"}). The question therefore arises of how the viral mRNAs are recruited to the NXF1/p15 pathway for export. Depletion of Aly, the most thoroughly characterized adaptor protein for cellular mRNA, had little effect on transport of viral messages (Figs [4](#f4){ref-type="fig"} and [5](#f5){ref-type="fig"}) or protein expression (Fig. [3](#f3){ref-type="fig"}). This is perhaps surprising given the dependence cellular mRNAs show on Aly for export ([@r5]; [@r7]; [@r38]), but we saw no more than a twofold drop in M2 levels (Fig. [3](#f3){ref-type="fig"}). We were only able to knockdown Aly expression by ∼60 % however (Fig. [2](#f2){ref-type="fig"}) and although control experiments examining the localization of polyadenylated mRNAs in uninfected siRNA-treated cells showed a degree of nuclear retention (data not shown), it may be that more efficient depletion would reveal a more marked phenotype. Alternatively, our data may indicate a more important role for other cellular adaptor proteins, such as the recently described UIF protein ([@r20]). A strong requirement for UAP56, another component of the TREX complex and responsible for recruiting Aly to cellular pre-mRNAs, was however seen for export and/or expression of certain viral messages (Fig. [7c, d](#f7){ref-type="fig"}). UAP56 depletion caused marked retention of segment 7 and 8 transcripts and concomitant reductions in M1, M2 and NS1 expression (Figs [3](#f3){ref-type="fig"}, [4](#f4){ref-type="fig"} and [5](#f5){ref-type="fig"}). Curiously, however, NS2 expression was resistant to UAP56 depletion, further highlighting that individual viral transcripts show significant differences in their interactions with the cellular export machinery, even when considering the two spliced mRNAs. Also of note was the finding that while UAP56 depletion had little effect on segment 4 localization, (producing less than a twofold increase in the nuclear : cytoplasmic ratio of segment 4 mRNA), a more than threefold reduction in protein expression was seen (Figs [3](#f3){ref-type="fig"} and [5](#f5){ref-type="fig"}). Loss of functional UAP56 has been shown to affect the transport of mRNA within the cytoplasm of *Drosophila* oocytes ([@r33]), so we speculate that the reduction in HA expression seen here results from an effect downstream of mRNA nuclear export. Although we have shown that NXF1 and/or UAP56 are required for export of certain viral transcripts, the mechanism(s) by which these factors are recruited to the mRNAs remains to be determined. Maturation of M2 mRNA resembles that of a normal cellular pre-mRNA: intron removal presumably leads to deposition of the exon junction complex, including UAP56, which will then recruit Aly and NXF1 (Fig. [7d](#f7){ref-type="fig"}). Alternatively or in addition, NXF1 might be directly recruited to the serine/arginine-rich protein splicing factor 2/alternative splicing factor (SF2/ASF) ([@r23]) as the latter polypeptide has been shown to regulate the extent of segment 7 splicing ([@r43]). Recruitment of UAP56 and NXF1 to the unspliced transcripts as well as NXF1 to the intronless late genes is less obvious. UAP56 can be directed to unstructured regions of intronless mRNAs, although if this mechanism applied to the influenza virus transcripts more dependence on Aly might have been expected ([@r51]). Alternatively, the 5′-cap structure of mRNA plays an important role in the export of both spliced and intronless cellular mRNAs, via the cap-binding complex (CBC) recruiting Aly and thus NXF1 ([@r7]; [@r38]). However, the CBC must be displaced from the 5′-cap of cellular pre-mRNAs by the viral polymerase during the process of cap-snatching and although PB2 releases the mRNA cap early after transcription initiation ([@r4]), it is not known whether the CBC rebinds to mature viral mRNAs. Arguing against quick re-recruitment of the CBC, splicing of segment 7 mRNA is controlled by the viral polymerase binding to the 5′-end of the mRNA (including the cap structure as well as an internal polymerase-binding sequence) thereby blocking the 5′-most splice donor site (Fig. [7d](#f7){ref-type="fig"}; [@r45]). It has also been proposed that the 5′-ends of all viral mRNAs are similarly occupied by the viral polymerase as a mechanism to prevent a futile cycle of utilizing viral mRNAs as cap-donors ([@r44]), but this proposal has not been tested *in vivo* yet. Based on numerous precedents from other nuclear-transcribing viruses ([@r42]) it is also possible that viral polypeptide(s) act as an adaptor between the viral mRNA and the cellular nuclear export pathway. For instance, it has been suggested that the viral polymerase complex might functionally replace the cellular CBC for the purposes of nuclear export ([@r44]). It is well established that the viral polymerase interacts with Pol II ([@r11]; [@r30]; [@r32]; [@r40]), potentially placing it in the correct local environment to interact with the export machinery that would normally be recruited co-transcriptionally to a cellular pre-mRNA. Such a mechanism is compatible with the observation that drugs that inhibit Pol II transcription inhibit export of most of the viral mRNAs ([@r2]; [@r53]; [@r54]; this study). NP is also a plausible adaptor candidate: non-RNP-associated NP shuttles between nucleus and cytoplasm ([@r10]; [@r36]; [@r55]) as well as interacting with several cellular proteins involved in mRNA biogenesis and trafficking ([@r27]; [@r32]; [@r34]). While our data here do not support a functionally important role for the NP--hnRNPA1 interaction, they are consistent with (although not proof of) a role for the NP--UAP56 interaction in viral mRNA trafficking. Similarly, circumstantial evidence suggests NS1 might also function as an export adaptor ([@r42]). It interacts with NXF1 and other components of the mRNA export pathway ([@r41]) as well as with at least a subset of the viral mRNAs ([@r54]). However, the near normal export of the NP and PB2 mRNAs seen after NXF1 depletion or DRB treatment when NS1 expression is much reduced, as well as a consideration of the kinetics of protein expression in normal cells suggests that if this mechanism does operate at all, it does not do so for the intronless early gene transcripts. METHODS ======= Materials. ---------- 293T and MDCK cells were cultured as described previously ([@r24]). PR8 virus was grown in cells in serum-free medium in the presence of 1 μg trypsin ml^−1^ and 0.14 % BSA and titrated by plaque assay in MDCK cells as described previously ([@r24]). Virus infections were performed at an m.o.i. of 10. DRB (Calbiochem) was used at 150 μM and LMB (LC Laboratories) at 10 nM. Rabbit anti-NP, anti-M1, anti-PR8 (HA) and anti-PB2 sera have been described previously ([@r2]; [@r39]). Anti-NS2 serum was the kind gift of Dr Agustin Portela (Agencia Española del Medicamento, 28220 Madrid, Spain). Anti-M2 (clone 14C2), anti-Aly (11G5), anti-UAP56 (ab47955), anti-GAPD (6C5), anti-hnRNPA1 (9H10), anti-NXF1 (53H8) and anti-actin (mAbcam8226) were purchased from Abcam. Secondary antibodies were purchased from Molecular Probes, LiCor Biosciences or DAKO. Plasmids pCDNA-PB2, pCDNA-HA, pCDNA-NP, pCDNA-M1 and pCB8− contain cDNA copies of PR8 segments 1, 4, 5, 7 and 8 ([@r2]; [@r6]; [@r17]; [@r35]). siPool siRNAs were purchased from Dharmacon. 293T cells were plated at 0.5×10^5^ cell per well 1 day prior to transfection then transfected with 100 nM siRNA (final concentration) by using Dharmafect-1 following the manufacturer\'s instructions. Microscopy. ----------- FISH was performed essentially as described previously ([@r2]). To synthesize mRNA-binding probes, plasmid templates were linearized with *Hin*dIII for pCDNA-M1, *Xba*I for pCB8− and *Kpn*I for pCDNA-PB2, -NP or -HA, and transcribed using SP6 RNA polymerase in the presence of 0.15 mM cyanine 3-UTP (Perkin Elmer) or 25 μM Dig-UTP (Roche). Fluorescence images were captured using a Leica TCS-NT confocal microscope or an Olympus IX70 fluorescence microscope fitted with a Retiga 2000R camera. Post-capture processing was carried out using Adobe Photoshop to apply linear brightness/contrast adjustments evenly across figure panels. To quantify mRNA nuclear : cytoplasmic ratios, confocal images were analysed using the program ImageJ (NIH). DAPI counterstaining was used to identify the nucleus, which was manually defined using the region of interest (ROI) manager. The cell periphery was also manually delineated using the ROI manager. Nuclear and total cellular intensities for the FISH probe were quantified by using the measure function and the cytoplasmic reading taken as (total-nuclear) intensities. The average intensity of the two compartments was then calculated and used to generate a nuclear : cytoplasmic staining ratio. Protein analyses. ----------------- Cell lysates were analysed by SDS-PAGE and Western blotting according to standard procedures. Blots were imaged by chemiluminescence by using horseradish-peroxidase-conjugated secondary antibodies and X-ray film, or (for quantification) by fluorescence using IRDye 700/800-conjugated secondary antibodies on a Licor Biosciences Odyssey near-infrared imaging platform. Protein levels were quantified using Licor Odyssey version 3 software using background correction followed by normalization to an actin loading control. We thank Professor Bob Lamb and Drs Adrian Whitehouse and Jim Boyne for discussion, Drs Helena Browne and Agustin Portela for gifts of reagents and Dr Helen Wise for assistance with RT-PCR. This work was supported by a grant from the Medical Research Council (no. G0700815) to P. D. E. K. C. R. was supported by a studentship from the Wellcome Trust. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Effect of DRB on influenza virus mRNA localization. 293T cells were infected (or mock infected) with PR8 virus, treated where indicated with DRB at 90 min and 6.5 h p.i. localization of the indicated segments determined by FISH (red) and wide field fluorescence microscopy. Nuclei were counterstained with DAPI and false-coloured blue. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### siRNA depletion of cellular mRNA export factors. (a, b) Lysates from infected siRNA-treated or mock-treated 293T cells at 6.5 h p.i. were analysed by SDS-PAGE and Western blotting for the indicated antigens. (b) Levels of the indicated polypeptides from replicate experiments were quantified and plotted as the mean±[sem]{.smallcaps} relative to untreated cells. (c) siRNA-treated cells were infected at an m.o.i. of 10 and output virus titres at 11 h p.i. from cells treated (or mock treated) with siRNAs targeting the indicated proteins determined by plaque assay. The mean±[sd]{.smallcaps} of three independent experiments is shown. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### Effect of siRNA depletion of cellular mRNA export factors on viral gene expression (a, b) 293T cells were transfected with siRNAs against the indicated proteins \[or mock transfected (--)\] then infected or mock infected with virus. Cell lysates taken at 6.5 h p.i. were analysed by Western blotting for the indicated antigens. (b) Accumulation of the indicated viral polypeptides from replicate experiments were quantified and plotted as the mean±[sem]{.smallcaps} of the level seen in non-siRNA-treated infected cells after normalization for actin levels. :::  ::: ::: {#f4 .fig} Fig. 4. ::: {.caption} ###### Effect of siRNA depletion of cellular mRNA export factors on viral mRNA localization. 293T cells were transfected with siRNAs against the labelled proteins, infected with virus and positive-sense transcripts from the indicated segments detected by FISH followed by confocal microscopy. :::  ::: ::: {#f5 .fig} Fig. 5. ::: {.caption} ###### Quantitative analysis of the effect of siRNA depletion of cellular mRNA export factors on viral mRNA localization. The nuclear : cytoplasmic staining ratios of the FISH probes from replicate confocal FISH images of infected cells treated with siRNAs as described in Fig. [4](#f4){ref-type="fig"} were determined and plotted as the fold change compared with the values seen in GAPD siRNA-treated cells. Error bars represent the [sem]{.smallcaps}. :::  ::: ::: {#f6 .fig} Fig. 6. ::: {.caption} ###### Effect of LMB treatment on viral mRNA localization. MDCK cells were infected or mock infected with virus, treated where indicated with LMB from 90 min p.i. and positive-sense RNA from specific segments (as labelled) detected by FISH and confocal microscopy at 6.5 h p.i. :::  ::: ::: {#f7 .fig} Fig. 7. ::: {.caption} ###### Cartoon model for influenza virus mRNA export. Viral mRNAs split into the four indicated conceptual classes are synthesized in the nucleus by vRNPs. (a) Intronless mRNAs encoding early gene products (PB2, NP, putatively PB1 and PA) show only weak dependence on NXF1 levels for nuclear export. If NXF1 is directly involved in their export, its means of recruitment are unknown. (b) Intronless mRNAs encoding late gene products (HA, probably NA) show strong dependence on NXF1 for export, but its means of recruitment is also uncertain. (c) Unspliced, intron containing mRNAs (M1 and NS1) show strong dependence on NXF1 and UAP56 levels for export. (d) Spliced segment 7 mRNA (M2) shows strong dependence on NXF1 and UAP56 levels and a weak dependence on Aly for its export, which is proposed to follow the normal cellular pattern of splicing dependent deposition of the exon junction complex followed by recruitment of UAP56, Aly and NXF1. The viral polymerase complex (3P) is also known to bind the 5′-end of this mRNA and may also interact with other viral mRNAs. NS2 mRNA is omitted from this model because of insufficient data. :::  :::

PubMed Central

2024-06-05T04:04:19.371277

2010-5-01

PMC3052575

Dear Friends, At the outset I extend my candid wishes to you and your family members for an eventful and prosperous New Year. Beginning January 1, 2011, I have taken over the responsibility of becoming the 'Managing Editor' for 'Toxicology International,' the official journal of the Society of Toxicology, India. I take this opportunity to express my gratitude to my predecessors for their vision and untiring efforts, which have given a strong base for setting the high standards that the journal continues to uphold, with significant visibility among scientists, academia, practitioners, and research students. I understand the challenges to compete with established journals in the area, in the present scenario, however; collective efforts of membership will help us reach the heights. Being a multidisciplinary forum for toxicology research, we welcome for publication, original research articles, with sufficient importance, novelty, and breadth of interest. In addition to that, articles presenting hypotheses and commentaries addressing current issues, which are of immediate interest to other investigators, are also welcome. Updates pertaining to areas of current importance and interests are also published in the form of mini-reviews. Besides clinical reports, occupational and safety evaluation, legal, risk, and hazard assessment, and impact of toxicants on humans, animals, and environment of sufficient novelty to warrant rapid publication, are also considered. My goal is to continue with the policies developed by the earlier Editors, while striving to raise the standards, reputation, and visibility even further. In order to achieve this, the Advisory Board has been expanded, by including seasoned and active scholars working in diversified areas of toxicology. This will facilitate improvement in review processing for each submitted manuscript, by including one member, and eventually it would help in enhancing the consistency of better quality in the articles. I wish to develop a tradition of publishing special issues with a specific focal theme of high priority and relevance falling within the scope of the journal, which will further increase the journal's strength, visibility, and circulation. A new feature, to provide a forum for the discussion and interpretation of data published in earlier issues of the journal, is also under consideration. We have achieved success in a difficult, but important task, of reducing the time between submission and publication through online electronic manuscript submission and a peer review system (<http://www.journalonweb.com/ti/>), which allows submission of articles with tracking of its progress till proof. It will be our endeavor to increase the number of issues in future. As you are aware that the journal is indexed at various sites, including 'Pubmed', it is high time to brain storm to have to procure an impact factor to improve the ranking of our journal among the journals published in the area of Toxicology. I publically express my sincere thanks to all the members of the Editorial and Advisory Board, both past and present, the reviewers, authors, and all the dedicated personnel at Medknow Publications, in advance, for their contribution, cooperation, and services. Perhaps you may find some of my goals to be ambitious and far reaching, but I am confident that with your help and cooperation I will be able to achieve these targets. I look forward to continuing and building on the outstanding work of my predecessor, and the many other visionary and skilful people involved in creating and enriching this journal and guiding its development since the day of its inception. It is an appeal to forward suggestions to improve the quality of the journal, to deliver the best to the scientists and the society. Best wishes,

PubMed Central

2024-06-05T04:04:19.374189

2011-01-01

PMC3052576

INTRODUCTION {#sec1-1} ============ Pollution from industrial and petroleum- related activities are common in many parts of the world. Industrialization and urbanization are the main causes for such pollution. Furthermore, ever increasing number of vehicles and stationary engines has lead to fast growth of automobile workshops in city areas. The washing and servicing of engines and vehicles generate a large volume of oil-based wastes. Diesel used as fuel for automobiles and higher levels of aromatic hydrocarbons than that of crude oil are found in diesel and so the toxicity is higher for the diesel oil.\[[@CIT1]\] Aromatic hydrocarbons are more water soluble and disappear more slowly from solution compared to alkanes. They are accumulated by organisms in greater extent and retained longer than alkanes. These are the main factors contributing to high toxicity of diesel oil. The WSF of Diesel fuel present in water is known to affect the health of fishes and ultimately of human beings. Thus, it becomes necessary to assess the effect of WSF of diesel on fishes, because fish flesh is rich in protein and minerals like calcium, phosphorus and iron. Fishes are a very useful barometer of the real state of purity of water. No aquatic body should be considered in a satisfactory condition unless fish will live and thrive in it. MATERIAL AND METHODS {#sec1-2} ==================== WSF was prepared as per the method given by.\[[@CIT2]\] Different concentration of Diesel fuel was denoted as DF1-5%, DF2-10%, DF3-15%, DF4 -20% and DF5-25%. *Channapunctatus* were collected from local rivers of Doon Valley and brought to the laboratory. After acclimatization the fishes were divided into 6 groups of about same size (15-18cm) and weight (60-80gm). A group of ten fishes was put in six different troughs, 1 serving as control and other 5 as experimental to study the effect of WSF on fishes. The fishes were exposed to various concentrations of WSF of Diesel for 21 days. The fish alive after 21 days was carefully removed from the test solution and sacrified immediately. For histopathological examination the liver were fixed in Bouins fluid for 24 hours followed by dehydration, embedding, sectioning and staining adopting the standard methods.\[[@CIT3]\] Detailed histopathological studies of control and treatment sections (5 microns) were made under microscope and photographs were taken. RESULTS AND DISCUSSION {#sec1-3} ====================== Liver of control fish showed normal hepatic acini arrangement in regular manner. Hepatocytes have polygonal shape with clear cellular border lines and homogenous cytoplasm. The quite concentric nucleus has clear nucleoli \[[Figure 1](#F0001){ref-type="fig"}\]. Comparing the sections of the control and experimental fishes, dilation and thrombosis formation in hepatoportal blood vessel at 2%-DF1 \[[Figure 2](#F0002){ref-type="fig"}\]. The histopathological changes of the liver were more pronounced at 4%-DF2.There was melanomacrophage centers & congestion in blood sinusoids and intervascular haemolysis in hepatic blood vessels and hepatoportal blood vessels \[[Figure 3](#F0003){ref-type="fig"}\]. The normal architecture of liver tissue was markedly disrupted. Sinusoids in most cases were distended, dilation, hemorrhage and lymphocytic infiltration was observed at 6%DF3 \[[Figure 4](#F0004){ref-type="fig"}\]. With increasing the concentration hepatopancreas damage became more conspicuous and aggregation of inflammatory cells between the hepatocytes and focal area of necrosis was observed at 8%-DF4 \[[Figure 5](#F0005){ref-type="fig"}\]. The condition become more critical at higher concentration and Congestion and fibrosis between the hepatocytes were observed at 10%-DF5 \[[Figure 6](#F0006){ref-type="fig"}\]. ::: {#F0001 .fig} Figure 1 ::: {.caption} ###### Liver showing hepatocytes with uniform nuclei and sinusuids (control group) :::  ::: ::: {#F0002 .fig} Figure 2 ::: {.caption} ###### Liver showing degeneration of endothelial lining cells (2%- PF1) 100× :::  ::: ::: {#F0003 .fig} Figure 3 ::: {.caption} ###### Liver showing marked swelling of hepatocytes and diffuse necrosis (4%-PF2) 100× :::  ::: ::: {#F0004 .fig} Figure 4 ::: {.caption} ###### Liver showing hemolysis within the blood vessels (6%-PF3) 100× :::  ::: ::: {#F0005 .fig} Figure 5 ::: {.caption} ###### Liver showing dilation and congestion in blood sinusoids (8%-PF4) 100× :::  ::: ::: {#F0006 .fig} Figure 6 ::: {.caption} ###### Liver showing coagulative necrosis (10%-PF5) 100× :::  ::: The organ most associated with the detoxification and biotransformation process is the liver, and due to its function, position and blood supply\[[@CIT4]\] it is also one of the organ most affected by WSF contamination in the water.\[[@CIT5]\] Melanomacrophage Centres (MMC) recorded in the liver of exposed fish may be suggestive of metabolic disorders and it is commonly associated with dietary deficiency in response to WSF. Similar result were observed by\[[@CIT6]\] who observed the MMC in the liver of Molly fish after exposure to sodium perchlorate. The function of the MMC in the liver of fishes remains uncertain, but some studies have suggested that it is related to destruction, detoxification or recycling of endogenous and exogenous compounds.\[[@CIT7]\] Necrosis of some portions of the liver tissue that were observed probably resulted from the excessive work required by the fish to get rid of the WSF from its body during the process of detoxification by the liver. The inability of fishes to regenerate new liver cells may also have led to necrosis. The present results are in agreement with\[[@CIT8]\] who reported similar changes in the liver of *Astyanax sp*. exposed to WSF of crude oil. Further\[[@CIT9]\] reported, petroleum carcinogenic compounds caused necrosis in hepatopancreatic cell of *Palaemon serratus*. Haemolysis & Hemorrhage is a result of blood channel disruption and is indication of severe physical damage. The haemolysis, dilation, congestion, & fibrosis may be attributed to direct toxic effects of pollutants on hepatocytes, since the liver is the site of detoxification of all types of toxins and chemicals. These four alterations were not found earlier regarding with exposure of WSF of diesel fuel. The present result is in agreement with\[[@CIT4]\] who observed these alterations in *Tilapia zillii* and *Solea vulgaris* under the influence of different pollutants from Lake Qarun, Egypt. According to Saxena\[[@CIT10]\] lymphocytic infiltration in liver with round lymphocytes and dark basophilic nuclei were observed in fish after exposure to polluted water with heavy metals. Since the liver is usually the site of toxicant accumulation and detoxification, it is likely to show very extensive histopathology. CONCLUSION {#sec1-4} ========== In conclusion, the present study proved that WSF of diesel affected the histopathological changes of the liver of *Channa punctatus* and this effect was time dependent. This may not result in fish kill immediately but definitely represents a health hazard to human consumers. Therefore it is important that the waste water should be treated before pouring in to the water bodies. Hene, Government should take remedial measures and pass appropriate legistation. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.374856

2011-01-01

PMC3052577

INTRODUCTION {#sec1-1} ============ Silks are generally defined as protein polymers that are spun into fibers by *Lepidoptera* larvae such as silkworms, spiders, scorpions, mites and flies. Due to its special chemical structure and chemical composition, silkworm silk is highly compatible and absorbed easily with the human skin. Silk is a natural protein that is made up of 25--30% sericin and 70--75% fibroin proteins.\[[@CIT1]\] Silk is a natural protein polymer that has been approved as a biomaterial by the US Food and Drug Administration (FDA).\[[@CIT2]\] The silk sericin and fibroin proteins are prospective wound healing agents and are antioxidants and considered as bio-adhesive mediators of human body\[[@CIT3]\] Furthermore, silk fibroin is considered to be proper for the generation of biomedical products such as blended materials because of its minimal adverse effects on the immune system.\[[@CIT4]\] Silk protein is claimed to have wound healing property and efforts are being made to develop the wound dressing from silk proteins. The present study was designed to evaluate the acute dermal toxicity, dermal irritation and sensitization potential of the silk protein film. In the assessment and evaluation of the toxic characteristics of a substance, determination of acute dermal toxicity is useful where exposure by the dermal route is likely. It provides information on the health hazards likely to arise from a short-term exposure by the dermal route. This test also provides information on dermal absorption and the mode of toxic action of a substance by dermal route. Dermal irritancy evaluation of topical agents was widely examined using Draize test, described by John Draize. Skin sensitization (allergic contact dermatitis) is an immunologically mediated cutaneous reaction to a substance. The sensitivity and ability of tests to detect potential human skin sensitizers are considered important in a classification system for toxicity relevant to public health. MATERIALS AND METHODS {#sec1-2} ===================== The test material for the present study was obtained from a commercial firm (M/s. Sericare, Division of Healthline Pvt. Ltd., Bangalore, India). The test material was composed basically of silk proteins. Acute dermal toxicity study {#sec2-1} --------------------------- This test was conducted as per the OECD Guideline 402.\[[@CIT5]\] Healthy adult Wistar albino rats, females weighing around 150--160 g and males weighing between 200 and 250 g, were used in the study. The females used were nulliparous and non-pregnant. Animals were acclimatized to the experimental conditions for 1 week. They were maintained under standard laboratory hygienic conditions, providing laboratory animal feed and water *ad libitum*. Animals were divided into four groups (*n* = 6) as follows: group I (control group male rats), group II (silk protein film applied male rats), group III (control group female rats) and group IV (silk protein film applied female rats). Approximately 24 hours before the test, the hair coat was removed by closely clipping the dorsal area of trunk of the animals. Care was taken to avoid abrading the skin and only animals with intact skin were used. Films under study were moistened with physiological saline and applied on the shaved area. They were held in contact with the skin using porous gauze dressing and non-irritating tape for a period of 24 hours, whereas for the control groups, gauze moistened with physiological saline was applied and held in contact in a similar way as the treatment groups. Animals were observed within 30 min, 4 and 24 hours after the removal of films and kept under observation for 14 days. Individual weight of animals was determined on the day of application of film and weekly thereafter. They were observed for the changes in skin, eyes and mucous membranes, behavioral patterns, diarrhea, salivation and tremors. Mortality was recorded during the course of study. At the end of the study, survived animals were weighed, sacrificed and subjected to gross necropsy. In rats, the blood samples were collected on day 0, 7, 14 by retro-orbital plexus puncture method and serum biochemical parameters were estimated from the serum samples using clinical chemistry analyzer. Acute dermal irritation study {#sec2-2} ----------------------------- This test was conducted as per the OECD Guideline 404 (2002)\[[@CIT6]\] with a preferred sequential testing strategy. Healthy, young, adult, 4--6 months old, New Zealand White male rabbits, weighing around 1.0--2.0 kg were used as test system. Animals were individually housed and acclimatized to the experimental conditions for 1 week. Animals were free to access food and water *ad libitum*. Approximately 24 hours before the test, hair coat was removed by closely clipping the dorsal area of the trunk of the animals. Care was taken to avoid abrasion of the skin and only animals with intact skin were used for the study. Initial test (*in vivo* dermal irritation test using one animal) {#sec2-3} ---------------------------------------------------------------- Three test patches were applied sequentially, after observing that there was no serious skin reaction produced by the previous patch, at three different sites. The first patch was removed after 3 min and was observed for skin reactions, if any. Similarly, the second patch and third patch was removed after 1 and 4 hours, respectively, and observed for skin reactions. Later, the animals were observed for 14 days after the removal of patches, unless corrosion developed at an earlier time point indicating the need for immediate termination of the test. Confirmatory test (*in vivo* dermal irritation test with additional animals) {#sec2-4} ---------------------------------------------------------------------------- If there was no serious skin reaction observed in the initial test, the irritant or negative response was confirmed using up to two additional animals, each with one patch, for an exposure period of 4 hours. The animals were observed for 14 days after the removal of patches. All the animals were examined for signs of erythema, eschar and edema. The responses were scored at 1, 24, 48 and 72 hours after the patch removal. Skin sensitization study {#sec2-5} ------------------------ The present study was carried out to evaluate the skin sensitization potential of silk protein film as per EPA (Environmental Protection Agency) Health Effects Test Guidelines, OPPTs 870, 2600. Healthy, young, adult male guinea pigs, weighing around 300--450 g, were divided into two groups (*n* = 6). The grouped animals were individually housed and acclimatized to the experimental conditions for 1 week. They were maintained under standard laboratory hygienic conditions, providing laboratory animal feed and water *ad libitum*. Approximately 24 hours before the test, hair coat was removed by closely clipping the dorsal area of the trunk of the animals. ### Induction phase {#sec3-1} On day 0, the animals (groups I and II) were clipped on the left flank to an area of 2 × 2 cm. The test film was moistened with physiological saline and held in contact with the skin of the animals in treated group using an occlusive patch for 6 hours. Similarly, a sterile gauze piece moistened with physiological saline was topically applied as control to the group I animals. On day 7 and 14, the experimental animals were exposed to the same application for 6 hours. ### Challenge phase {#sec3-2} On day 28, occlusive patch was retained for 24 hours. The experimental animals were challenged with the test film on untreated side of the flank. On day 29, the patch was removed and the whole flank was clipped. All the skin samples (in terms of erythema and swelling) and systemic reactions resulting from induction and challenge procedures were observed and recorded at 24 and 48 hours after the patch removal according to the Magnusson/Kligman grading scale. Statistical analysis {#sec2-6} -------------------- Mean values and standard error of mean were calculated and expressed as Mean±SE. The data were analyzed by two-way analysis of variance (ANOVA) followed by Bonferroni's post-test. RESULTS {#sec1-3} ======= In the acute dermal toxicity, there were no appreciable clinical signs observed throughout the observation period of 14 days after the patch removal and there were no appreciable changes in skin, eyes, mucous membranes and behavioral pattern. There was no mortality seen throughout the observation period. Also, there was no significant (*P*\>0.05) difference in body weight in the treatment group compared to the respective control group female rats \[[Table 1](#T0001){ref-type="table"}\]. Serum obtained from blood samples collected on day 0, 7 and 14 of the study period was used to estimate aspartate transaminase (AST), alanine transaminase (ALT), creatinine and blood urea nitrogen (BUN) for all the groups \[Tables [2](#T0002){ref-type="table"}--[5](#T0005){ref-type="table"}\] respectively. None of the biochemical parameters differed significantly (*P*\>0.05) in the treatment group from the control group in both male and female rats. Also, there were no prominent gross lesions observed in vital organs like liver, lung, heart, kidney and spleen of the treatment and control groups in both the sexes. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Effect of treatment on body weight (g) in male and female rats during acute dermal toxicity study ::: Males Females --------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- A~1~ 210 216 219 215 220 222 178 180 192 182 184 189 A~2~ 223 229 231 210 214 217 186 190 193 190 194 199 A~3~ 217 220 223 227 231 233 180 183 187 179 184 188 A~4~ 213 217 220 223 225 228 188 191 194 183 188 191 A~5~ 224 227 231 219 222 225 190 194 197 186 188 192 A~6~\` 208 213 217 222 227 230 179 184 189 191 194 200 Mean±SE 215.84±2.73 220.34±2.61 223.50±2.50 219.34±2.49 223.17±2.42 225.84±2.36 183.50±2.09 187.00±2.23 192.00±1.78 185.17±1.93 188.67±1.84 193.17±2.09 Values are expressed as Mean±SE; *n* = 6, *P*\>0.05 ::: ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Effect of treatment on serum aspartate transaminase concentration (U/l) in male and female rats during acute dermal toxicity study ::: Males Females --------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- ------------- A~1~ 155.50 159.32 156.78 159.43 149.58 155.56 154.60 161.20 151.23 145.50 151.93 158.00 A~2~ 157.60 160.85 152.65 159.76 154.45 159.04 161.14 154.54 164.41 154.70 161.32 154.51 A~3~ 154.35 155.45 160.11 154.70 160.70 154.96 154.87 156.65 151.89 158.59 158.61 153.95 A~4~ 158.89 154.40 149.43 161.42 157.97 156.76 144.59 151.61 158.59 149.59 160.74 160.69 A~5~ 155.93 152.21 155.76 158.96 159.09 160.54 149.89 160.00 152.43 153.98 157.91 159.00 A~6~ 158.87 159.49 154.00 153.47 158.66 161.60 152.49 150.64 158.68 145.50 151.93 158.00 Mean±SE 156.86±0.77 156.96±1.40 154.79±1.50 157.96±1.29 156.75±1.67 158.08±1.12 152.93±2.26 155.74±1.77 156.20±2.14 152.48±2.26 158.11±1.67 157.24±1.31 Values are expressed as Mean±SE; *n* = 6, *P*\>0.05 ::: ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Effect of treatment on serum alanine transaminase concentration (U/l) in male and female rats during acute dermal toxicity study ::: Males Females --------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ A~1~ 45.59 47.11 44.18 45.60 44.72 44.83 48.95 44.57 49.69 44.52 50.96 45.70 A~2~ 44.72 43.34 45.80 48.86 47.68 47.84 47.58 49.60 50.22 53.44 46.09 49.33 A~3~ 47.94 48.51 48.62 49.85 49.54 44.12 43.50 43.57 51.02 43.42 48.01 48.50 A~4~ 45.21 45.62 45.84 43.45 43.45 45.78 48.93 52.44 43.47 47.90 44.54 43.62 A~5~ 50.64 46.51 49.65 44.96 44.81 43.42 49.63 50.68 45.70 45.12 49.96 49.11 A~6~ 48.76 45.89 47.98 45.69 47.58 48.62 51.64 44.21 44.11 45.97 44.62 46.65 Mean±SE 47.15±0.96 46.17±0.71 47.02±0.85 46.41±0.99 46.30±0.95 45.77±0.85 48.38±1.12 47.52±1.57 47.37±1.36 46.73±1.48 47.37±1.12 47.16±0.92 Values are expressed as Mean±SE; *n* = 6, *P*\>0.05 ::: ::: {#T0004 .table-wrap} Table 4 ::: {.caption} ###### Effect of treatment on serum creatinine concentration (mg/dl) in rats during acute dermal toxicity study ::: Males Females --------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- A~1~ 0.52 0.55 0.60 0.59 0.54 0.51 0.56 0.62 0.54 0.59 0.55 0.54 A~2~ 0.63 0.60 0.56 0.58 0.59 0.53 0.59 0.54 0.60 0.64 0.65 0.64 A~3~ 0.60 0.59 0.57 0.55 0.61 0.55 0.62 0.55 0.57 0.53 0.58 0.60 A~4~ 0.55 0.53 0.62 0.59 0.54 0.51 0.54 0.53 0.54 0.59 0.53 0.50 A~5~ 0.57 0.51 0.51 0.58 0.59 0.53 0.63 0.65 0.56 0.52 0.64 0.53 A~6~ 0.56 0.59 0.59 0.55 0.61 0.55 0.58 0.62 0.59 0.55 0.52 0.59 Mean±SE 0.58±0.02 0.57±0.01 0.58±0.02 0.55±0.01 0.58±0.01 0.56±0.02 0.59±0.02 0.59±0.01 0.57±0.02 0.57±0.02 0.58±0.03 0.57±0.03 Values are expressed as Mean±SE; *n* = 6, *P*\>0.05 ::: ::: {#T0005 .table-wrap} Table 5 ::: {.caption} ###### Effect of treatment on blood urea nitrogen concentration (mg/dl) in male and female rats during acute dermal toxicity study ::: Males Females --------- ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ ------------ A~1~ 39.90 41.62 40.50 43.42 43.80 41.34 39.55 38.43 40.64 45.17 43.04 44.07 A~2~ 43.72 38.40 37.45 38.52 44.53 44.57 44.32 42.21 41.18 43.49 41.22 43.40 A~3~ 45.59 44.40 45.09 47.21 39.67 41.25 42.13 39.49 40.08 44.73 39.69 41.15 A~4~ 37.54 47.39 42.47 41.37 37.82 42.34 41.17 40.15 42.19 42.49 43.59 45.15 A~5~ 44.12 42.23 38.04 49.59 38.02 38.16 44.37 41.09 43.55 48.69 47.17 49.10 A~6~ 43.49 45.69 40.08 39.45 45.38 39.59 45.21 47.15 42.49 45.17 43.90 45.59 Mean±SE 42.40±1.24 43.29±1.32 40.60±1.17 43.26±1.80 41.54±1.40 41.21±0.90 42.80±0.89 41.42±1.27 41.69±0.53 44.96±0.87 43.10±1.05 44.75±1.09 Values are expressed as Mean±SE; *n* = 6, *P*\>0.05 ::: In the dermal irritation study, dermal reactions were graded according to Draize\[[@CIT7]\] and recorded at 1, 24, 48 and 72 hours after the patch removal. In the initial test, there was no skin reaction observed in rabbits after any of the three sequential (3 min, 1 and 4 hour) exposures. After a period of 14 days of observation, there was no appreciable skin reaction seen. In the confirmatory test, there was neither erythema nor edema seen in the treated animals and reaction was graded as "0" at all time points of observations. In the skin sensitization study, skin reactions were read and graded 24 and 48 hours after the patch removal according to the Magnusson/Kligman grading scale. The dermal reactions were recorded and graded individually for each animal. No skin reactions were observed on the skin of guinea pigs in the treatment group after the challenge period, i.e., at 24 and 48 hours of patch removal, which was scored 0 based on the Magnusson/Kligman grading scale. Also, there were no systemic reactions observed after the induction and challenge exposure. DISCUSSION {#sec1-4} ========== In the assessment and evaluation of the dermal safety of a substance, acute dermal toxicity and the determination of irritant and sensitization effects on skin are important initial tests. Serum biochemical parameters, viz., AST, ALT, BUN and creatinine did not show any statistically significant alterations in silk protein based film applied on male and female rats when compared to the respective control group animals, indicating that the test compound did not have any effect on serum biochemical parameters estimated. At autopsy, none of the treated and control group rats showed any gross pathological lesions. This lends support to the above findings, and thus, it could be concluded that silk protein based film did not exert any damage to vital organs when applied dermally. Acute dermal toxicity with sericin cream revealed no difference between the body weights of rats in sericin cream and cream base treated groups and there was neither mortality nor any gross lesions observed after sacrificing the animals, and thus, it could be concluded that sericin did not cause an appreciable dermal toxicity in rats.\[[@CIT8]\] In the dermal irritation study, there were no signs of erythema, eschar and edema observed in the film treated group. The scoring was given as "0" for all time points of observation in the treatment group animals according to the Draize scoring system. These findings are in accordance with the reports of Pavankumar,\[[@CIT9]\] wherein the scoring was 0 at all time points of observation based on Draize scoring system when silk protein based films were applied on New Zealand white rabbits. Similarly, Thai silk soap does not cause irritation to the skin of laboratory rabbits when tested at a 10% concentration.\[[@CIT10]\] The silk protein is a biodegradable and highly biocompatible material. Thus, it has not caused any undue effect after application.\[[@CIT11]\] This suggests that silk protein film was not having dermal irritation potential in New Zealand white rabbits. The guinea pigs in treatment group in skin sensitization study did not show any changes on the surface of the skin at 24 and 48 hours of patch removal and the reaction was graded as "0". Moreover, silk fibers used as sutures (FDA approved) are biocompatible and less immunogenic and inflammatory than collagens or polyesters such as polylactic acid.\[[@CIT12]\] The results of present investigation correlate with the study conducted by Pavankumar,\[[@CIT9]\] wherein slight erythema was observed on the skin of the guinea pigs in silk protein based film treated group including control, which was scored "0" based on the scoring system given by Buehler.\[[@CIT13]\] This suggests that silk protein film is not having appreciable skin sensitizing property in guinea pigs. It could be concluded from the present study that the test materials used in the study are safe under acute dermal toxicity, acute dermal irritation and skin sensitization. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.375775

2011-01-01

PMC3052578

INTRODUCTION {#sec1-1} ============ Imidacloprid, a relatively new, systemic insecticide related to nicotine, was introduced to the market in 1991 as the first chloronicotinyl insecticide, and has since become the most successful, highly effective and largest selling insecticide worldwide.\[[@CIT1]\] It has higher selectivity factors for insect versus mammals than organophosphates, methylcarbamates and organochlorines. This is attributable to both target site specificity\[[@CIT2]\] and detoxication. Due to the favorable mammalian safety characteristics, imidacloprid has been developed for veterinary use and is used as flea control agent on cats and dogs. In agriculture, it is most commonly used on rice, cereal, maize, potatoes, vegetables, sugar beets, fruits, cotton, hops and turf for control of sucking insects, coleoptera (beetles) and others.\[[@CIT3]\] Alterations in hematobiochemical parameters following intraperitoneal administration in rats\[[@CIT4]\] and oral administration in cockerels\[[@CIT5]\] have been reported. The present study was conducted to investigate the effect of repeated (28 day) oral administration of imidacloprid on hematobiochemical profiles in male White Leghorn (WLH) chicks. MATERIALS AND METHODS {#sec1-2} ===================== Experimental animals {#sec2-1} -------------------- The present study was conducted on day-old healthy male WLH chicks. The chicks were procured and housed in pens of battery brooder house at Central Poultry Research Station, Anand Agricultural University, Anand. The chicks were provided with standard feed and water *ad libitum*. Experimental design {#sec2-2} ------------------- Birds were acclimatized for a period of 1 week before the start of oral dosing with imidacloprid. All the birds were randomly divided into five groups (C1, C2, I1, I2 and I3), with each group containing 25 birds. Apparent LD~50~ (ALD~50~) of imidacloprid (50 mg/kg) was taken into consideration for calculation of different doses in the groups.\[[@CIT5]\] Birds were treated with imidacloprid at a dose rate of 1/20^th^ of LD~50~, 1/30^th^ of LD~50~ and 1/40^th^ of LD~50~ in three treatment groups for a period of 28 days starting from 7 days of age. Group C1 was given no treatment and served as control. The group C2 was administered groundnut oil and served as vehicle control. Group I1 was given 1/40^th^ of LD~50~ (1.25 mg/kg), group I2 was put on 1/30^th^ of LD~50~ (1.67 mg/kg) and group I3 received 1/20^th^ of LD~50~ (2.5 mg/kg) of imidacloprid suspended in groundnut oil. The daily oral administration was continued for 28 days and the live weight was recorded at weekly intervals and birds were observed for any toxicity symptoms during the entire period of experiment. On 7^th^ day of experiment, the birds were vaccinated with New Castle disease Lasota strain vaccine (Indovax, Gurgaon, India) through intraocular route. The birds were also vaccinated against Infectious Bursal disease (IBD) using (Georgia strain, Indovax) at day 14. Birds were vaccinated with Marek's Disease Bivalent vaccine (Merial, Singapore) at day-old age. After 14 and 28 days of oral administration, the birds were weighed and blood samples were collected before final culling of birds for estimating hematological parameters namely hemoglobin (Hb), packed cell volume (PCV), total erythrocyte count (TEC) and total leukocyte count (TLC), and serum biochemical parameters, namely, serum glutamate pyruvate transaminase (SGPT), serum glutamate oxaloacetate transaminase (SGOT), creatinine (CRT), total protein (TP), serum albumin, serum glucose and serum globulin using standard kits (Coral Clinical Systems, Goa, India). RESULTS AND DISCUSSION {#sec1-3} ====================== The present study was conducted to see the effect of various doses of imidacloprid on the hematobiochemical profile of WLH cockerels. An approximate LD ~50~ of imidacloprid, that is, 50 mg/kg b. wt. in birds, was taken into consideration for the calculation of doses of imidacloprid to be administered to cockerels.\[[@CIT6]\] The effect of administration of imidacloprid at the rate of 1/20^th^ of LD~50~ (1.25 mg/kg b. wt.), 1/30^th^ of LD~50~ (1.67 mg/kg b. wt.) and 1/40^th^ of LD~50~ (2.5 mg/kg b. wt.) once daily for 28 days on body weight, hematological parameters, carbohydrate metabolism, protein metabolism, liver and kidney functions have been investigated in the present study. Transaminases were measured to find out the effect on hepato-biliary system while CRT was measured to evaluate kidney function. TP, total albumin and total globulin were measured to see the effect of imidacloprid on protein metabolism, and serum glucose was measured for monitoring carbohydrate metabolism. There were no apparent clinical signs of toxicity at all the dose levels tested. Besides, there was no significant effect on body weight of the imidacloprid treated birds, giving no indication of stress on the birds due to given doses of imidacloprid. The administration of imidacloprid did not create any significant change in the levels of Hb, PCV and TEC in WLH cockerels after 14 and 28 days of imidacloprid exposure; however, after 28 days of exposure, there was a significant reduction in TLC \[[Table 1](#T0001){ref-type="table"}\] in the birds of group I3 (2.5 mg/kg). It has been suggested that compounds having benzene ring or other ring structure act as a hapten that combines with the protein constituent of leukocytes to form an antigen to which the animal develops antibodies that are toxic to leukocytes, causing either lysis or agglutination.\[[@CIT7]\] Imidacloprid is also a ring-structured compound and thus may have caused leukocytopenia. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Effect of daily oral administration of imidacloprid on hematological parameters in White Leghorn cockerel chicks ::: Group Hb (mean±SE) (g%)) PCV (mean±SE) (%) TEC (mean±SE) (10^6^/mm^3^) TLC (mean±SE) (/mm^3^) ---------------------------- -------------------- ------------------- ----------------------------- ------------------------ ----------- ----------- ------------ -------------------------------------------- Group C1 (control) 9.97±62 8.47±24 29.33±1.02 30.000±73 2.57±08 2.67±08 10,167±530 10,600±520 Group C2 (vehicle control) 9.50±0.71 8.23±0.25 30.00±1.48 29.67±0.95 2.47±0.07 2.57±0.07 10,167±458 10,158±505 Group I1 (LD*50*/40) 9.27±0.46 8.23±0.25 29.33±0.33 28.50±0.62 2.52±0.11 2.67±0.11 10,083±542 10,333±487 Group I2 (LD*50*/30) 9.13±0.55 8.17±0.16 28.00±0.68 28.33±1.08 2.47±0.07 2.57±0.07 10,000±507 9883±399 Group I3 (LD*50*/20) 8.47±0.12 7.97±0.18 28.00±0.73 27.83±1.60 2.47±0.12 2.57±0.12 9967±577 8667±154[\*](#T000F1){ref-type="table-fn"} *n* = 25; \* *P* ≤ 0.05 ::: After 14 and 28 days, imidacloprid treated birds at different doses showed a significant dose-related increase in SGOT level as compared to controls \[[Table 2](#T0002){ref-type="table"}\]. The level being highest in birds of group I3, the elevation of SGOT may be due to degenerative pathological changes in hepatocytes, which cause increase in permeability of cell membrane, resulting in release of transaminases in the blood stream.\[[@CIT8]\] Similar increase in SGOT values by imidacloprid has been reported in cockerels.\[[@CIT5]\] While imidacloprid treated birds at different doses did not show any significant change in SGPT level, SGPT activity is of little diagnostic value in diseases of birds.\[[@CIT9]\] There was no significant effect on serum TP, albumin, globulin and CRT levels, indicating no effect on protein metabolism and kidney function due to administered doses of imidacloprid. These are in agreement with the findings of Premlata *et al*.\[[@CIT4]\] in rats. ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Effect of daily oral administration of imidacloprid on biochemical parameters in White Leghorn cockerel chicks ::: Group SGOT (U/l) (mean±SE) SGPT (U/l) (mean±SE) Serum total protein (g/dl) (mean±SE) Serum albumin (g/dl) (mean±SE) Serum globulin (g/dl) (mean±SE) Serum glucose (mg/dl) (mean±SE) Serum creatinine (mg/dl) (mean±SE) --------------------------- ---------------------- ------------------------------------------------- -------------------------------------- -------------------------------- --------------------------------- --------------------------------- ------------------------------------ ----------- ----------- ----------- ------------------------------------------------- ------------------------------------------------- ------------- ------------- Group C1(control) 160.77±2.57 160.82±5.52 23.26±0.67 24.94±1.54 2.62±0.20 3.51±0.34 0.97±0.07 0.91±0.15 1.65±0.25 2.60±0.22 194.97±5.60 161.20±4.36 0.290±0.007 0.415±0.003 Group C2(vehicle control) 159.88±3.6 162.40±3.96 23.71±0.84 24.05±1.81 2.73±0.23 2.84±0.34 0.87±0.09 0.76±0.11 1.83±0.17 1.91±0.29 193.21±5.17 150.48±3.70 0.300±0.008 0.409±0.005 Group I1(LD~50~/40) 183.07±2.82 202.17±8.33[\*](#T000F2){ref-type="table-fn"} 21.62±1.64 22.88±1.10 2.28±0.06 3.46±0.15 0.85±0.13 1.02±0.05 1.42±0.13 2.43±0.16 191.216±4.64 110.23±6.46[\*\*](#T000F3){ref-type="table-fn"} 0.298±0.008 0.409±0.005 Group I2(LD~50~/30)) 193.70±5.12 227.02±8.93[\*\*](#T000F3){ref-type="table-fn"} 21.50±1.06 22.10±0.95 2.33±0.04 3.53±0.23 0.87±0.15 1.06±0.21 1.46±0.13 2.47±0.35 201.27±5.05 102.47±4.93[\*](#T000F2){ref-type="table-fn"} 0.307±0.003 0.401±0.008 Group I3(LD~50~/20) 214.03±8.58 237.47±9.15[\*\*](#T000F3){ref-type="table-fn"} 23.50±0.93 25.22±1.73 2.29±0.23 2.99±0.18 0.64±0.08 1.13±0.09 1.65±0.27 1.95±0.17 169.40±6.03[\*\*](#T000F2){ref-type="table-fn"} 98.34±5.53[\*\*](#T000F2){ref-type="table-fn"} 0.302±0.006 0.387±0.011 n = 25; \* *P* ≤ 0.05 \*\* *P* ≤ 0.01 ::: At 14 days, imidacloprid treated birds at highest dose (I3) showed a marked decrease in glucose levels, while there was no significant effect on birds of groups I1 and I2. However, after 28 days of exposure, there was a significant decrease in glucose level in all the insecticide treated groups. The birds in group I3 exhibited the highest decrease in glucose levels. Hypoglycemia induced by the given doses of imidacloprid may be due to the fact that thyroid is especially sensitive to imidacloprid. In 1998, an US EPA report indicated that imidacloprid can affect thyroid function in animals, which may cause decrease in blood glucose levels. In conclusion, imidacloprid seems to be less toxic for multiple systems in WLH chicks at the doses administered. However, dose-dependent decrease in blood glucose and increased SGOT indicative of degenerative changes in liver were caused by imidacloprid. Besides, imidacloprid caused leukocytopenia at the dose of 2.5 mg/kg/day. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.378118

2011-01-01

PMC3052579

INTRODUCTION {#sec1-1} ============ Dairy cows fed on high-quality forage produce more milk with less supplemental concentrate than the cows fed lower-quality forages. Forages with high concentrations of crude protein (CP) are considered high quality because feeding high-protein forage cuts down the need of supplemental protein. Secondly, CP content is positively correlated to energy content of forages. High-protein forages generally are more digestible and provide more energy than low-protein forages. Application of nitrogen (N) fertilizer improves both quality and yield by increasing CP content of forage markedly and the available energy. Due to this reason, the use of N-fertilizers on agricultural land has been immensely increased around the world.\[[@CIT1]\] Forages take up and assimilate nitrogen as NH~4~^+^, NO~3~^-^ and soluble organic compounds such as urea (CO(NH~2~)~2~) and amino-acids.\[[@CIT2]\] Nitrate is the primary nutrient form of the nitrogen in soils and is a normal constituent of plants. Occasionally, excessive amounts of nitrate accumulate in plants and result in livestock mortalities. Outbreaks of nitrate toxicity due to consumption of fodder containing high amounts of nitrate have occurred in farm animals throughout the world.\[[@CIT3]\] Most commonly, nitrate poisoning occurs in cattle and sheep. In ruminants, nitrate is reduced by microbial reductases to nitrite. The rumen microbes utilize this nitrite by converting it into ammonia as a nitrogenous source. However, excessive nitrite gets accumulated in rumen, from where it is readily absorbed into blood stream and combines with ferrous ion of hemoglobin (Hb) to form met-hemoglobin (met-Hb). The met-Hb is a poor transporter of oxygen in the body and the animal suffers from oxygen deficiency. In livestock, poisoning due to nitrate ions is influenced by several factors that include plant, environmental, management factors and health status of the animal. The plant factors are the most important amongst these because nitrate toxicity in livestock is chiefly caused by consumption of plants rich in nitrates. The factors that influence the accumulation of nitrate in fodder crops are species, stage of growth, part of plant, pH of soil, use of fertilizers and climatic conditions. These factors had not been studied recently and the guidelines found today in literature are based on limited research data obtained in the 60s and 70s, and have not been updated to more recent research and field experiences. Keeping this in view, factors contributing toward accumulation of nitrate in forages were studied and correlated with recent research and the field problems encountered in dairy animals due to excessive exposure to nitrate rich plants. MATERIALS AND METHODS {#sec1-2} ===================== Six forage crops, viz., berseem (*Trifolium alexandrinum*), oats (*Avena sativa*), sorghum (*Sorghum vulgare*), bajra (*Pennisetum typhoides*), toriya (*Brassica napus*), and maize (*Zea mays*), commonly used as fodder for farm animals, were evaluated for the nitrate content and the factors associated with toxicity and its accumulation were studied. The nitrate content was estimated in samples of bajra, berseem, maize, oats, sorghum and toriya that were collected randomly from different villages of five districts (Ludhiana, Moga, Jalandhar, Patiala and Sangrur) of Punjab state. Samples were also collected from fields of the Punjab Agricultural University, where fodder was grown without the use of chemical fertilizers. Nitrate estimations were done separately as whole plant, stem and leaves of plants. Samples of guinea grass were collected from different cuttings (first to fourth cut) from fields and the nitrate content was estimated. The fodder samples received in the toxicology laboratory from farmers of Punjab state, suspected to be contaminated, were also tested for nitrate levels and their background history was collected. The fodder samples were initially subjected to qualitative test with diphenylamine blue (DPB-1% in concentrated sulfuric acid) to know the presence/absence of nitrate.\[[@CIT4]\] The same fodder samples were finally subjected to quantitative estimation of nitrates as per the method described by Cataldo *et al*,.\[[@CIT5]\] RESULTS AND DISCUSSION {#sec1-3} ====================== The amount of nitrate accumulated within the plant depends upon the rate of nitrogen uptake by the plant from the soil and the rate of its reduction by the plant. There is no accumulation when the rate of reduction equals the rate of uptake and when uptake exceeds the rate of reduction, nitrate starts getting accumulated. Some species, viz., sorghum, oats, sudan grass, etc., are known nitrate accumulators and may cause sudden death in animals.\[[@CIT6]\] Nitrate can be detected in traceable amounts in all plants but it becomes dangerous when it exceeds the safe limit of 2500 ppm NO~3~-N and forages having more than 4500 ppm NO~3~-N are considered highly toxic. Seven outbreaks due to consumption of fodder having excessive amount of nitrate have been recorded in bovines in Punjab during 2002--2007 (Annual Progress Reports, Animal Disease Research Centre, GADVASU, Ludhiana; 2002-2007; unpublished data). Two dreadful outbreaks associated with consumption of fodder containing toxic amount of nitrate occurred recently in Tajpur and Haibowal Dairy Complexes, Ludhiana, Punjab, and 105 dairy cattle died in these unfortunate incidents. Even the low level of nitrate in fodder may lead to slight respiratory distress and help respiratory infections to flourish. Reduced weight gain and impaired fertility has been reported in cattle grazing pasture having NO~3~-N in the range of 801--3400 ppm.\[[@CIT7]\] Therefore, six forage crops commonly used as fodder for farm animals in India were evaluated for nitrate content and the factors influencing its accumulation were studied. The mean values of nitrate content determined in common fodder crops collected from villages of Punjab are presented in [Table 1](#T0001){ref-type="table"}. The nitrate content on dry matter basis was highest in sorghum (1695 ppm), followed by oats (1525 ppm) and toriya (1510 ppm). The nitrate levels in maize, berseem and bajra were 825, 990 and 776 ppm, respectively. The study revealed that oats, sorghum and toriya are the natural accumulators of nitrate. In previous literature, it was found that sorghum is one of the most notorious accumulators of nitrate but no reference suggesting excessive amounts of nitrogen in toriya could be traced.\[[@CIT3]\] There was no environmental stress when these samples were collected. The reason could be the presence of excess soil nitrogen than that required for maximum growth of oats, sorghum and toriya. In Punjab, farmers have been using nitrate containing fertilizers (urea, ammonium nitrate) in their fields for 30 years without following any fertility program, to obtain higher yields of crops. Therefore, it is assumed that in general, the farming soil might be rich in nitrogen in the state. This needs to be investigated further. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Mean nitrate content in fodder samples (whole plant) collected from villages of Punjab state and farms of Punjab Agricultural University (PAU), Ludhiana ::: Crop Number of samples (*n*) Village sample[\*](#T000F1){ref-type="table-fn"} NO~3~-N (ppm) PAU farm samples[\*](#T000F1){ref-type="table-fn"} NO~3~-N (ppm) PAU farm samples[\*\*](#T000F2){ref-type="table-fn"} NO~3~-N (ppm) --------- ------------------------- ---------------------------------------------------------------- ------------------------------------------------------------------ -------------------------------------------------------------------- Oats 65 1525 (985--2035) 935 (850--1075) 1480 (1260--1855) Sorghum 30 1695 (1025--2240) 1080 (805--1250) NA Berseem 82 990 (650--1970) 675 (510--945) 1056 (805--1265) Bajra 30 776 (450--775) 565 (524--762) 760 (574--795)[\*\*\*](#T000F3){ref-type="table-fn"} Maize 50 825 (610--1234) 515 (505--670) 745 (610--940) Toriya 50 1510 (820--1950) 1220 (805--1455) 1670 (1150--1852) \* Samples collected during normal days; \*\* Samples collected during harsh winter days and cloudy environment; \*\*\* Samples collected during dry, very hot and humid weather; NA -- Data not available ::: It was interesting to note that oats and toriya grown together with barseem in the same field were found to contain two times of nitrate compared to barseem. It indicates the intrinsic ability of oat and toriya to accumulate nitrate \[[Table 2](#T0002){ref-type="table"}\]. Two outbreaks of nitrate poisoning that occurred recently in Ludhiana were the result of huge consumption of toriya alone by hungry cattle and buffaloes, which were not previously fed this fodder (data not shown). The nitrate level of toriya on dry matter basis was 5200--5500 ppm at Tajpur Dairy Complex outbreak and 6820--7040 ppm at Haibowal Dairy Complex outbreak. At Haibowal Dairy Complex, the fodder comprised barseem + toriya. The fodder was grown together in the same field and barseem contained nitrate of 800 ppm. This suggested the inherent ability of toriya to accumulate nitrate. Previous literature supports the differences in potential for nitrate accumulation among the species and varieties of forages.\[[@CIT4]\] Another consistent finding was higher nitrate content in radish than in berseem, oats and toriya samples despite being grown together in the same field. The nitrate concentration in radish was 5--6 times higher than barseem and 2-3 times more when compared to oats and toriya. It is confirmed that nitrate accumulation in fodder may vary depending upon the genetic variability in nitrate uptake by a plant. The farmers should select particular forage for feeding of livestock keeping in mind the nitrate content of forages and crops available at their disposal. ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Nitrate content in fodder samples grown as mixed crops in the same field (mean value, *n* = 30) ::: Crop combination NO~3~-N (ppm) ---------------------------------- --------------- ------ ------ --------------------------- Barseem + oats + radish 704 1675 NA Root = 3500 Leaves = 1145 Barseem + toriya + radish 876 NA 1425 Root = 5000 Leaves = 1075 Barseem + oats + toriya + radish 825 1594 1500 Root = 4700 Leaves = 1267 NA -- Data not available ::: Nitrate accumulation varies with the stage of plant growth. Rate of uptake diminishes with the maturity of the plant. Therefore, immature (young) crop contain more nitrate than the mature crop.\[[@CIT3]\] The increased frequency of occurrence of outbreaks in farm animals due to nitrate poisoning in the months of December and January every year may be correlated with this factor (data not shown). In these extreme winter months, young crops rich in nitrate are fed due to scarcity of green fodder. This was reflected in the recent outbreaks of Tajpur and Haibowal Dairy Complexes, where feeding young toriya took the lives of 105 dairy cattle. To determine the effect of age of plant on nitrate levels, berseem, bajra, maize, oats and toriya were obtained at two levels of cuttings (mature and immature) from the same fields to determine the nitrate content. The nitrate level reduced in all forages with the age of plant. The difference in nitrate level of mature and immature forages was statistically significant at 5% level of significance \[[Table 3](#T0003){ref-type="table"}\]. The effect of plant maturity yielded interesting results in maize and bajra where nitrate concentrated - double in stem than in leaves. It showed that distribution of nitrate varies with the age of plant. It might be due to the distribution of some nitrate in the grains of maize and bajra which appear at the mature stage. The nitrate level was quite high in samples of young crops of oats, toriya and sorghum received from farmers in the laboratory for testing as compared to mature samples of these crops (young crops NO~3~-N = 1568-2054 ppm; mature crops NO~3~-N = 972-1232 ppm). The results confirmed the previous reports of reduced nitrate levels with the maturity of plant.\[[@CIT8]\] Lower nitrate levels in mature plant may be due to decreased uptake or increased enzyme activity to convert the nitrate into intermediate compounds ready for evaporation or used by the plant. It is suggested to delay the harvesting of known nitrate accumulators (toriya, oats and sorghum) for reducing the dangerous nitrate levels. ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Distribution of nitrate content in fodder samples collected from Punjab Agricultural University farms (mean value; *n* = 30) ::: NO~3~-N (ppm) -------------- --------------- ------ ------ Barseem Leaves 615 430 Stem 835 576 Whole plant 745 505 Oats Leaves 1195 965 Stem 1465 1170 Whole plant 1275 1005 Toriya Leaves 1024 832 Stem 1560 1046 Whole plant 1245 995 Maize Leaves 390 195 Stem 598 420 Whole plant 564 405 Napier bajra Leaves 472 306 Stem 915 635 Whole plant 748 525 ::: Guinea grass samples taken from four cuttings revealed significant differences in the nitrate content amongst cuttings \[[Table 4](#T0004){ref-type="table"}\]. The nitrate content was lowest in fourth cutting and highest in second cutting. The reason for highest nitrate level in second cutting could be plentiful growth of crop and use of fertilizer (urea) in the fields after first cutting by the farmers. Four farmers applied urea after third cutting which was reflected as exceptionally high nitrate content (1705, 1525, 1678 and 2005 ppm) in fourth cutting plants collected from those fields. The data suggested that nitrate content decreases with the maturity of plant and use of nitrate containing fertilizers directly affects the nitrate accumulation in plants. ::: {#T0004 .table-wrap} Table 4 ::: {.caption} ###### Nitrate content in different cuttings of guinea grass (*n* = 20) ::: Number of cuttings Nitrate content (ppm) -------------------- ----------------------- 1^st^ 580--695 2^nd^ 1560--2090 3^rd^ 1200--1260 4^th^ 870--1005 ::: The concentration of nitrate differs with the parts of plants when accumulation occurs. To determine the distribution of nitrate in the plants, the nitrate content was determined in leaves, stem and whole plant of berseem, bajra, maize, oats and toriya \[[Table 3](#T0003){ref-type="table"}\]. Results showed the variability in nitrate level in different parts of plants. All forages exhibited similar results of having higher levels of nitrate in stem than in leaves. The difference of nitrate levels in stem and leaves was highly significant in maize and bajra, followed by berseem, oats and toriya. This might be due to differences in the ability of roots to take up nitrogen from the soil. These findings have confirmed that plant parts vary in nitrate content; parts close to the ground contain more nitrates and as we go higher along the length of plant, nitrate content goes on decreasing. Roots and stems have more nitrate content, followed by leaves,\[[@CIT3]\] whereas flowers and grains usually contain little or no nitrate.\[[@CIT9]\] To study the effect of use of nitrate containing fertilizers on the nitrate accumulation in plants, the samples of berseem, bajra, maize, oats, sorghum and toriya were harvested from fields of Punjab Agricultural University, Ludhiana, where no chemical fertilizer was used. The mean values of nitrate content obtained in those fodder samples are shown in [Table 1](#T0001){ref-type="table"}. Environmental factors contribute to variation in nitrate level in plants; however, in this study the effect of environmental factors was neutralized, as climatic conditions were the same for the samples collected from the villages and university fields. The samples collected on the same day from both the places were included in this study to minimize the effect of environmental factors. The nitrate levels of university field forages were lower than that of the field samples and were quite below the maximum safe limit (2500 ppm). The trend was same in all seasons (summer, winter, spring and rainy season). This indicated excessive use of urea in the village fields by the farmers to have a better yield. There is an urgent need to determine the soil nitrate content throughout India and to educate the ignorant farmers regarding the consequences of this excessive use of fertilizers. Besides the above factors, weather conditions influence nitrate accumulation in plants significantly. Unfavorable weather conditions for plant growth, viz., drought, frost, extreme cold and cloudy weather, may increase nitrate accumulation in plants.\[[@CIT4]\] In the present study, nitrate level was significantly higher in forages when determined in adverse growing conditions, viz., cloudy, cold with frost, wet in winters and very hot drought conditions in summer compared to nitrate concentrations found in same fields under normal weather conditions. This is one of the reasons why cases of nitrate poisoning in Punjab are frequently encountered during the months of December and January when crops are affected by extreme cold and frost.\[[@CIT10]\] In very hot summer, the nitrification of bacteria in soil increases many folds which leads to nitrate accumulation in fodder. The cold and cloudy weather decreases the nitrate reductase activity and hence increases nitrate accumulation by plants. The conditions detrimental for plant growth including frost enhance accumulation of nitrate by reducing the surface area of the plant available for evaporation and photosynthesis.\[[@CIT9]\] Two outbreaks of nitrate poisoning in cattle due to Jumbo grass (sorghum hybrid) were reported in New Zealand. In both the incidents, the climatic factors were responsible for increasing nitrate content in forage. Accidental rain after a long dry summer, causing very rapid growth of grass, was a favorable condition for nitrate accumulation.\[[@CIT11]\] Similarly, sudden deaths due to nitrate poisoning occurred in cattle grazing ryegrass pasture for 6 hours in Australia. The quantitative analysis of ryegrass showed levels of 12.5% nitrate/nitrite on dry matter basis. The peculiar weather conditions, viz., summer, newly sown pasture, regenerated lush grass and cloudy environment, contributed to high nitrate levels in the grazed pasture.\[[@CIT12]\] A very cold, wet and windy weather contributed to another nitrate poisoning in the cattle fed sudax hay (*Sorghum* sp.) in South Wales, Australia. The toxicological laboratory determined the nitrate levels in hay to be up to 3.1%.\[[@CIT13]\] Soil type also plays an important role in nitrate accumulation. The plants grown in acidic and phosphorus deficient soils are known to have greater nitrate content. The uptake of NO~3~^-^ is largely increased at slightly acid pH levels because of the higher H^+^ gradient across the plasma membrane at low pH, and possibly because increased H^+^ influx reduces the membrane potential and facilitates NO~3~^-^ uptake.\[[@CIT14]\] Crops take up and assimilate nitrogen as NH~4~^+^, NO~3~^-^ and various soluble organic compounds such as urea (CO(NH~2~)~2~) and amino acids.\[[@CIT2]\] In aerated soils with a pH above 4, NO~3~^-^ is the prevailing N compound and NH~4~^+^ is found only in low concentrations; but in waterlogged soils, the ratio of NO~3~^-^ and NH~4~^+^ is reversed mainly as a consequence of depressed bacterial nitrification activity and denitrification of NO~3~^-^.\[[@CIT14]\] It is concluded that all plants contain nitrates but levels vary with the species, stage of growth, parts of plants and intrinsic ability to accumulate nitrate. Grasses are more sensitive to nitrate toxicity than cereals and pulses.\[[@CIT8]\] Addition of nitrate containing fertilizer should be avoided in the crops that tend to accumulate nitrate, viz., sorghum, oats, toriya, and if need arises, the quantity should be strictly adhered to the recommendations of agriculturists. The soil should be checked for its type, pH and nitrate levels before sowing the forages. In fields having higher nitrate levels, lower part of the crop near the ground should not be used as fodder because roots and stem store more nitrate than leaves. However, fodder can be used in the form of silage as ensiled forages contain less nitrate than fresh plants, but hay usually contains the same amount of nitrate as present in the fodder.\[[@CIT3]\] The forage containing high nitrate content (sorghum, oats and toriya) may be used by mixing with other crops having a low amount of nitrate to reduce the chances of toxicity in dairy animals. The forages should not be grown in the fields receiving sewage water or sludge or industrial effluents which may lead to serious nitrate toxicosis in the farm animals. Besides high levels of nitrate, other commonly used chemicals (pesticides, other fertilizers, etc.) in agricultural practices also occur in the environment. Synergistic or additive effect of these chemicals needs to be determined to minimize the risk of adverse effects on human and animal health. Given it a multi-factorial problem, there is an urgent need to take on further research on the nitrate toxicity due to fodder crops in livestock using integrated approach involving different disciplines. The known information gathered from existing literature put forward many serious questions on man-made global nitrate pollution endangering the health and survival of man, livestock and wildlife population if agricultural and urban development continues without the incorporation of strict regulations to reduce the impact of nitrate on our environment. The current study may serve as background information to conduct further research for the better understanding of factors contributing to serious potential of nitrate accumulation in forages leading to animal toxicity. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.380354

2011-01-01

PMC3052580

INTRODUCTION {#sec1-1} ============ The use of ethanol as a fuel additive has increased the ground water contamination problems.\[[@CIT1]\] The breakdown of ethanol in surface waters through biological and chemical processes potentially results in the consumption of significant quantities of dissolved oxygen in the surface water body which will adversely affect aquatic life, resulting in fish kills.\[[@CIT2]\] Toxicants of various origins disturb the osmoregulatory potential and other physiological processes of fish.\[[@CIT3]\] The fish gill epithelium has been extensively studied as a model for ion-transporting epithelia. Gill, which is the main osmoregulatory surface tissue in aquatic animals, seems to be the primary site of uptake of waterborne pollutants; therefore, it is expected to be the first site where the sub-lethal effects of chemicals are observed.\[[@CIT4]\] ATPases are the membrane-bound enzymes concerned with immediate release of energy and are responsible for a large part of basic metabolic and physiological activities. Na^+^/K^+^ ATPase has long been studied as a target for ethanol. The activity of ATPases in fish gill represents a useful cell-membrane biomarker of pollution as it can be easily quantified.\[[@CIT5]\] Tilapia of the genus *Oreochromis* is selected as a suitable biological model for studying the mechanism of osmoregulation in teleost fish since it is capable of tolerating a wide range of salinity.\[[@CIT6]\] As there is very limited information on fish exposed to ethanol, the present study was carried out to explore the effect of ethanol on branchial ATPase activities in a fresh water fish *Oreochromis mossambicus* (Peters). An attempt has also been made to assess the possibility of using gill ATPase estimation as a suitable biomarker in ethanol related studies. MATERIALS AND METHODS {#sec1-2} ===================== Experimental design {#sec2-1} ------------------- Fresh water fish, *O. mossambicus* (Peters), commonly known as Tilapia, was selected as the animal model for the study, considering its hardy nature, ease of rearing, maintenance, availability, resistance and economic viability. They were collected from local hatcheries and acclimatized to the laboratory conditions for 15 days in large tanks filled with dechlorinated water (500 l). The tanks were previously washed with potassium permanganate to free the walls of the tanks from fungal infections. The tank had a continuous and gentle flow of tap water. The physico-chemical parameters of water were estimated daily\[[@CIT7]\] and were maintained constant throughout the experiment. The mean values for the parameters were as follows: dissolved oxygen of 8.16 ppm, total hardness 13±2 mg/l, total alkalinity 4±2 mg/l, temperature 26±2°C, pH 7.0±0.33 and salinity at 0 ppt. For conducting lethal toxicity studies, fishes were exposed to different concentrations of ethanol ranging from 1.27 to 127 g/l, in which they exhibited erratic movements, loss of equilibrium, grouping, increase in respiratory rhythm, excess secretion of mucus, followed by a gradual onset of inactivity. Each experiment was repeated three times at the selected ethanol concentration, every time noting the number of fish killed at each concentration up to 96 hours. A control without the toxicant was also maintained for both lethal and sub-lethal studies.\[[@CIT8]\] It was observed that at 13.01 g/l dose, 50% of the fishes were dead within 96 hours. The LC~50~ value for 96 hours was found to be 13.01 g/l and it was confirmed following Probit analysis.\[[@CIT9]\] From this method, it was calculated that 1.3 g/l was the sub-lethal ethanol concentration for *O. mossambicus*. To know the effect of higher concentrations as well as lower concentrations of ethanol, three sub-lethal concentrations (0.65, 1.3 and 2.6 g/l) along with control were taken as sub-lethal dose, which correspond to 1/20^th^, 1/10^th^ and 1/5^th^ of LC~50~ value, respectively. For conducting experimental studies, *O. mossambicus* of 10±2 g were taken in three separate tubs (capacity 60 l) which contained desired concentration of ethanol (0.65, 1.3 and 2.6 g/l, respectively) along with tap water. Six replicates were kept for each experiment. A control was also maintained in the water without the addition of ethanol. Whereas in the sub-lethal toxicity study, water was changed daily and the test solutions were renewed every 24 hours to maintain the dissolved oxygen concentration at the optimum level.\[[@CIT10]\] The fishes were fed on the same commercial diet *ad libitum*. The exposure periods such as 7 and 21 days were selected as per Organization for Economic Cooperation and Development (OECD) guideline program meant for aquatic organisms.\[[@CIT11]\] During the experimental period of 21 days, the animals were fed on the same diet so as to avoid the effects of starvation on normal physiological processes and antioxidant stress. Any other factor likely to influence toxicity was nullified by maintaining a suitable control. On completion of fixed exposure period, gill tissues were dissected out. They were then washed in ice-cold 0.33 M sucrose (pH 7.5). 10% of gill homogenate was taken for the present study. Extraction of the enzyme {#sec2-2} ------------------------ Cell fractionation of the gill homogenate was carried out according to the method of Davis (1970) with slight modifications. 10% of gill homogenate was centrifuged at 3000 *g* for 15 min in a cold refrigerated centrifuge. Supernatant was taken. It was again centrifuged at 12,000 *g* for 30 min. Clear supernatant thus obtained was taken. It was then centrifuged at 35,000 *g* for 30 min. Supernatant so obtained was discarded. The pellet obtained corresponds to the heavy microsomal fraction which was then resuspended in cold 0.33 M sucrose which served as the enzyme source. This was used for the experimental studies. They were then stored at −20°C until assayed. Immediately after thawing, gill microsomal preparations were used for the branchial ATPase activity assays. One unit of ATPase activity was expressed as micromoles of inorganic phosphate (Pi) produced by ATP decomposition per milligram protein per hour. Total ATPase activity was estimated from the amount of Pi liberated by the method of Evans.\[[@CIT12]\] Na^+^/K^+^ ATPase activity was estimated by the method of Bonting,\[[@CIT13]\] Ca^2+^ ATPase activity by the method of Hjerten and Pan\[[@CIT14]\] and Mg^2+^ ATPase activity was estimated according to the method of Ohnishi *et al*,\[[@CIT15]\] The inorganic phosphorus liberated was estimated by the method of Fiske and Subbarow.\[[@CIT16]\] The protein content of the samples was estimated by the method of Lowry *et al*.,\[[@CIT17]\] using bovine serum albumin as the standard. All the reagents used were of analytical grade. Statistical analysis {#sec2-3} -------------------- All processing of data was conducted with the software packages Microsoft Excel XP (for data storage) and SPSS version 15.0 (for statistical evaluation). Results are presented as mean±standard deviation (SD). Data distributions were examined to fit a normal distribution and homogeneity of variance was tested using analysis of variance (ANOVA). Data from six fish in each group were statistically analyzed adopting two-way ANOVA supplemented by multiple comparison test using Dunnett's method. Statistical significance was accepted at *P*\<0.001. RESULTS AND DISCUSSION {#sec1-3} ====================== Aquatic pollutants exert a biological effect on the ATPase system by partitioning in the enzyme complex, which may cause an allosteric change that results in decreased ATPase activity.\[[@CIT18]\] Introduction of large amount of ethanol in an aquatic system results in a variety of toxicological consequences of which reduced oxygen supply is a marked respiratory effect. Biophysical studies indicate that ethanol alters membrane function by disintegrating the membrane and changing the mobility of membrane lipids and proteins. Ethanol membrane interaction specifically affects some of the membrane-bound enzymes. Fish gills being sensitive to changes are referred to as important indicators of waterborne toxicants, thereby reducing oxygen consumption and disrupting its osmoregulatory function.\[[@CIT19]\] The reduction in the activity of ATPases was correlated to the altered ionic transport and decreased ATP breakdown which had impaired the metabolic and vital physiological activities.\[[@CIT20]\] In the present investigation, exposure of fish to 0.65, 1.3 and 2.6 g/l of ethanol for 21 days with a periodical sampling at 7 days caused marked significant (*P*\<0.001) inhibition in branchial ATPase activities. Tables [1](#T0001){ref-type="table"} and [2](#T0002){ref-type="table"} indicate a significant decrease (*P*\<0.001) in the total ATPase activity of *O. mossambicus* when exposed for 7 and 21 days, respectively, to different sub-lethal doses of ethanol. Damages in the membrane architecture may be the reason for the enzyme inhibition during the sub-lethal treatment with ethanol. Another possible reason may be the non-availability of substrates like ATP molecules, which resulted in the inhibition of ATPase. Observations made by Suhel *et al*,\[[@CIT21]\] and Jayantha *et al*,\[[@CIT22]\] support the present study. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Effect of exposure to different concentrations of ethanol for 21 days on the activity of gill ATPases (micromoles of Pi liberated/hour/mg protein) in *O. mossambicus* ::: Parameter Concentrations of ethanol ------------------- --------------------------- -------------- -------------- -------------- Total ATPase 27.47±0.2251 23.87±0.3247 20.82±0.8414 16.20±0.1102 Na^+^/K^+^ ATPase 12.08±0.0143 9.888±0.1349 8.392±0.7664 7.096±0.1284 Ca^2^+ ATPase 7.728±0.0997 7.116±0.0333 6.248±0.1573 4.032±0.4294 Mg^2^+ ATPase 7.682±0.1070 6.827±0.1283 6.201±0.0520 5.196±0.1073 Values are Mean±SD of six observations ::: ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Effect of exposure to different concentrations of ethanol for 21 days on the activity of gill ATPases (micromoles of Pi liberated/hour/mg protein) in *O. mossambicus* ::: Parameter Concentrations of ethanol ------------------- --------------------------- -------------- -------------- -------------- Total ATPase 26.64±0.6226 20.71±0.7307 15.51±0.1586 12.20±0.0541 Na^+^/K^+^ ATPase 11.38±0.1606 8.960±0.6726 6.670±0.4731 5.391±0.1692 Ca^2^+ ATPase 7.984±0.8211 4.760±0.3814 3.498±0.3426 2.675±0.1744 Mg^2^+ ATPase 7.110±0.0328 6.932±0.0352 5.340±0.1084 4.180±0.2537 Values are Mean±SD of six observations ::: The findings put forth by Simkiss\[[@CIT23]\] explain that fishes living in polluted rivers slightly impaired the main biochemical systems without causing death by resulting in the inhibition of the gill Na^+^/K^+^ ATPase activity. The present results are in agreement with previous reports stating that ethanol brings about an inhibition in Na^+^/K^+^ ATPase activity. A significant decrease (*P*\<0.001) in the Na^+^/K^+^ ATPase activity \[Tables [1](#T0001){ref-type="table"} and [2](#T0002){ref-type="table"}\] of *O. mossambicus* observed when exposed to 7 and 21 days mainly refers to the changes in the membrane lipid content, which in turn have been shown to influence the inhibition in Na^+^/K^+^ ATPase activity. Also, fishes entering into a state of severe hypoxia exhibited reduction in Na^+^/K^+^ ATPase activity. The findings of Kim *et al*,\[[@CIT24]\] support the present data. Tables [1](#T0001){ref-type="table"} and [2](#T0002){ref-type="table"} also present a significant decrease (*P*\<0.001) in the Ca^2+^ ATPase activity of *O. mossambicus* when exposed to 7 and 21 days to different sub-lethal concentrations of ethanol. Findings put forth by Ross *et al*,\[[@CIT25]\] support the present study. Mg^2+^ ATPase enzyme is always found to be associated with Na^+^/K^+^ ATPase in fish. It is also essential for the integrity of the cellular membrane and for the stabilization of branchial permeability.\[[@CIT26]\] A significant decrease (*P*\<0.001) \[Tables [1](#T0001){ref-type="table"} and [2](#T0002){ref-type="table"}\] observed in the Mg^2+^ ATPase activity of *O. mossambicus* when exposed for 7 and 21 days to various sub-lethal concentrations of ethanol is due to reduced ATP production which therefore results in disruption in cellular and ionic regulation. The findings of Racker *et al*,\[[@CIT27]\] support the present data. Further, ethanol suppresses the ATPase activities through activation of lipid peroxidation. The free radicals generated during the catalytic cycle of ethanol would have induced the peroxidation process in membrane lipids. The inhibition observed in the ATPase activities when *O. mossambicus* was exposed to ethanol for 21 days indicates peroxidation occurring in the damaged tissues, which brings about a change in the structure and inactivates a number of membrane-bound enzymes and protein receptors, which finally disrupts branchial membrane integrity. This statement was supported by Sato and Yonei.\[[@CIT28]\] Two-factor ANOVA in [Table 3](#T0003){ref-type="table"} reveals that total ATPase, Ca^2+^ ATPase and Mg^2+^ ATPase levels varied significantly between days and between concentrations at *P*\<0.001. Also, while taking into consideration the days of exposure as well as concentrations effect together (interaction), a significant difference was observed (*P*\<0.001). In the case of Na^+^/K^+^ ATPase activity, significant difference (*P*\<0.001) was observed between days as well as between concentrations, whereas when the effects between days of exposure as well as between concentrations were taken into account (interaction), a significant difference was observed (*P*\<0.01) \[[Table 3](#T0003){ref-type="table"}\]. Subsequent pair wise comparisons carried out using Dunnet's method \[[Table 4](#T0004){ref-type="table"}\] exhibited significant difference (*P*\<0.001) in all the four ATPase activities when all the three sub-lethal ethanol concentrations were compared with the control. ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Summary of two-factor ANOVA ::: Parameter Between days of exposure Between concentrations Days of exposure × concentration (interaction) ------------------- ---------------------------------------- ---------------------------------------- ------------------------------------------------ Total ATPase 0.000[a](#T000F1){ref-type="table-fn"} 0.000[a](#T000F1){ref-type="table-fn"} 0.000[a](#T000F1){ref-type="table-fn"} Na^+^/K^+^ ATPase 0.000[a](#T000F1){ref-type="table-fn"} 0.000[a](#T000F1){ref-type="table-fn"} 0.005[b](#T000F2){ref-type="table-fn"} Ca^2^+ ATPase 0.000[a](#T000F1){ref-type="table-fn"} 0.000[a](#T000F1){ref-type="table-fn"} 0.000[a](#T000F1){ref-type="table-fn"} Mg^2^+ ATPase 0.000[a](#T000F1){ref-type="table-fn"} 0.000[a](#T000F1){ref-type="table-fn"} 0.000[a](#T000F1){ref-type="table-fn"} The values are significant at a *P*\<0.001 and b *P*\<0.01 x= Interaction effect; 0.000 indicates that the values are significant at *P*\<0.001 ::: ::: {#T0004 .table-wrap} Table 4 ::: {.caption} ###### Multiple comparison test ::: Groups Total ATPase Na^+^/K^+^ ATPase Ca^2+^ ATPase Mg^2+^ ATPase ----------- ---------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- ---------------------------------------- Dunnett's Control vs. 0.65 g/l 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} Control vs. 1.3 g/l 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} Control vs. 2.6 g/l 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} 0.000[a](#T000F3){ref-type="table-fn"} The values are significant at a *P*\<0.001 ::: CONCLUSION {#sec1-4} ========== The present findings warrant future studies to explore ATPases as possible biomarkers of ethanol related incidents in ecotoxicology. The major findings of the present experiment validate that ATPase activity can be taken as a meaningful index of cellular activity and forms a useful toxicological tool. Thus, it may be concluded by stating that fish gill can be used as a model system to study the effect of ethanol on ion transport mechanisms across cellular and epithelial membranes. First author is grateful to C0 ochin U0 niversity of Science and Technology, CUSAT, Kochi, Kerala, India, for providing financial assistance and lab facilities to carry out the work. **Source of Support:** Cochin University of Science and Technology, CUSAT, Kochi, Kerala, India **Conflict of Interest:** None declared

PubMed Central

2024-06-05T04:04:19.382746

2011-01-01

PMC3052581

INTRODUCTION {#sec1-1} ============ Now-a-days, hydrogenated vegetable oils, which are solids at room temperature and have better physical properties, shelf life and flavor stability, are preferred over animal fats for mixing in feeds. However, the process of hydrogenation leads to the formation of *trans* fatty acids which cause many deleterious effects on health. In parts of India, *trans* fats from hydrogenated vegetable oil in the form of vanaspati are consumed in greater quantity than in the United States.\[[@CIT1][@CIT2]\] In India, cardiovascular diseases cause 3 million deaths per year, accounting for 25% of all mortality.\[[@CIT3]\] The World Health Organization predicted that deaths due to circulatory system diseases are projected to double between 1985 and 2015.\[[@CIT4]\] Dietary factors that may contribute to a high ischemic heart disease (IHD) risk in India include low intakes of vitamin B6 and folate\[[@CIT5]\] and high intakes of *trans* fatty acids, which have been associated with risk in studies conducted in the West.\[[@CIT6]\] Further, a wide range of organic and inorganic compounds may occur in feedstuffs, including pesticides, industrial pollutants, radionuclides and heavy metals. Pesticides that may contaminate feeds originate from most of the major groups, including organochlorine, organophosphate and pyrethroid compounds.\[[@CIT7]\] The interaction of the contaminants with added fats was not addressed earlier. In view of the increasing use of hydrogenated vegetable oil in both poultry and human diets, the present study was taken up to assess the effects of supplementation of vanaspati in layer diet alone and in combination with various stressors on sero-biochemical and immunological parameters. MATERIALS AND METHODS {#sec1-2} ===================== A total of 160 White Leghorn pullets, aged 18 wk, were procured from a local farm and were housed in the poultry Layer House of College of Veterinary Science, Rajendranagar, Hyderabad, with 16 hours of light. The birds were given standard layer feed formulated by the Department of Poultry Science, College of Veterinary Science, Rajendranagar and *ad libitum* water. Feed composition {#sec2-1} ---------------- The feed was formulated using the following ingredients (for 100 kg): maize 52 kg; soyabean 20 kg; sunflowercake 12 kg; de-oiled rice bran 6 kg; shell grit 8 kg; di-calcium phosphate 1.33 kg; trace minerals 100 g; vitamins A, B1, D3 -- 15 g; vitamin B12 -- 250 g; salt 333 g; digestible crude protein (DCP) 18.70%; metabolizable energy (ME) 2648 kcal/kg; calcium 3.67%; available phosphorus 0.69%; lysine 0.91% and methionine 0.32%. The feed supplemented with 5% vanaspati had an energy content of 2965 kcal/kg. Chemicals {#sec2-2} --------- Vanaspati (Dalda; Bunge Pvt. Ltd., Mumbai, India), chlorpyrifos EC 20% (Dursban; NOCIL, Pvt. Ltd., Mumbai), ferrous sulfate hepta hydrate and cadmium chloride (Qualigens Fine Chemicals, Mumbai), New castle disease virus (NDV) vaccine, R2B strain from Indovax, Indovet Pvt. Ltd. New Delhi. All the chemicals used were procured from Qualigens Pvt. Ltd. and were of analytical grade. Experimental design {#sec2-3} ------------------- After an acclimatization period of 2 wk, the experiment was started with the birds attaining an age of 20 wk and the birds were divided into eight groups of 20 birds each as follows. Group 1: Basal feed (20--54 wk) Group 2: Vanaspati feed (5%) (20--54 wk) Group 3: Basal feed (20--54 wk) + 1% ferrous sulfate (42--54 wk) Group 4: Basal feed (20--54 wk) + 100 ppm chlorpyrifos (42--54 wk) Group 5: Basal feed (20-54 wk) + 100 ppm cadmium (42-54 wk) Group 6: Vanaspati feed (20--54 wk) + 1% ferrous sulfate(42--54 wk) Group 7: Vanaspati feed (20--54 wk) + 100 ppm chlorpyrifos (42--54 wk) Group 8: Vanaspati feed (20--54 wk) + 100 ppm cadmium (42--54 wk) Birds of all the groups were vaccinated with NDV vaccine R2B at 46 wk of age and a booster subsequently after 21 days. The period from 38 to 42 wk, from 42 to 46 wk, from 46 to 50 wk and from 50 to 54 wk was considered as months 0, 1, 2 and 3, respectively. Blood samples were collected from the jugular vein at the end of every month from 10 birds in each group and serum was separated for the estimation of alkaline phosphatase (ALP), alanine transaminase (ALT), total protein, albumin, globulin, A/G ratio, total cholesterol, high density lipoprotein (HDL), triglycerides and creatinine, using standard kits supplied by Qualigens Pvt. Ltd. Parameters like globulins and A/G ratio were calculated by using the values from relevant parameters. HI titer\[[@CIT8]\] and PHA index\[[@CIT9]\] were estimated at the end of the experimental period. Statistical analysis {#sec2-4} -------------------- The data for sero-biochemical parameters were analyzed by two-way analysis of variance (ANOVA) followed by Tukey's *post hoc* test, and the data for immunological parameters were analyzed by one-way ANOVA using Statistical Package for Social Sciences (SPSS) 15^th^ version. The significance value was set at *P*\<0.05. Logarithmic transformation was used to the correct non-normal data and Welch ANOVA was used if the assumption of homogeneity of variance was found to be violated. RESULTS AND DISCUSSION {#sec1-3} ====================== The results of sero-biochemical profile and immunological parameters are depicted in Tables [1](#T0001){ref-type="table"}--[3](#T0003){ref-type="table"}, respectively. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Marginal means of sero-biochemical parameters in different groups ::: Group Cholesterol (mg/dl) HDL (mg/dl) Triglycerides (mg/dl) Alkaline phosphatase (IU/l) ALT (IU/l) Creatinine (mg/dl) ------------------------------ ----------------------------------------------- ------------------- ----------------------- ----------------------------- ---------------- -------------------- -- -- A  Normal 111.19±1.06^b^ 200.36±46.80^a^ 630.79±1.14^b^ 200.36±46.80^a^ 10.76±0.84^a^ 0.40±0.07^ab^  Vanaspati 103.19±1.06^ab^ 315.17±46.18^abc^ 669.24±1.13^b^ 315.17±46.18^abc^ 14.67±0.84^bc^ 0.36±0.07^a^  Ferrous sulfate 99.23±1.06^ab^ 267.97±47.50^ab^ 455.65±1.13^b^ 267.97±47.50^ab^ 15.08±0.84^bc^ 0.48±0.07^ae^  Chlorpyrifos 108.88±1.07^ab^ 445.41±46.18^bc^ 600.74±1.14^b^ 445.41±46.18^bc^ 11.95±0.84^ab^ 0.40±0.07^ad^  Cadmium 104.01±1.06^ab^ 672.76±46.18^d^ 243.48±1.13^a^ 672.76±46.18^d^ 18.50±0.84^c^ 0.58±0.07^de^  Ferrous sulfate + vanaspati 89.81±1.06^ab^ 524.89±49.30^c^ 406.72±1.14^ab^ 524.89±49.30^c^ 13.51±0.84^bc^ 0.48±0.07^bcde^  Chlorpyrifos + vanaspati 91.38±1.06^ab^ 262.48±47.50^ab^ 660.60±1.13^b^ 262.48±47.50^ab^ 14.99±0.84^bc^ 0.41±0.07^ac^  Cadmium + vanaspati 84.43±1.06^a^ 613.40±46.18^d^ 242.78±1.13^a^ 613.40±46.18^d^ 18.22±0.84^c^ 0.64±0.07^de^ B  Vanaspati 99.22±3.64 10.69±1.26^NS^ 600.05±31.57^NS^ 323.65±25.55^NS^ 14.15±2.74^NS^ 0.47±0.04^NS^  Normal 114.96±3.69[\*](#T000F1){ref-type="table-fn"} 11.57±1.28^NS^ 611.77±32.00^NS^ 296.68±25.27^NS^ 20.62±2.80^NS^ 0.46±0.04^NS^ A: effect of individual treatment; B: overall effect of vanaspati supplementation; Values are Mean±SE. Two-way ANOVA followed by Tukey's Honestly Significant Difference (HSD) *post hoc* test; Means with different superscripts are significantly different (*P*\<0.05); ALT, alanine transaminase; \* Significantly different (*P*\<0.05) ::: ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Marginal means of sero-biochemical parameters in different groups ::: Group Protein (g/dl) Albumin (g/dl) Globulin (g/dl) A/G ratio ----------------------------- ---------------- ---------------- ----------------- --------------- A  Normal 6.05±1.04^NS^ 1.70±0.04^NS^ 4.34±1.05^NS^ 0.40±0.02^NS^  Vanaspai 6.01±1.04^NS^ 1.65±0.04^NS^ 4.33±1.05^NS^ 0.40±0.02^NS^  Ferrous sulfate 5.37±1.04^NS^ 1.63±0.04^NS^ 3.70±1.05^NS^ 0.46±0.02^NS^  Chlorpyrifos 5.89±1.04^NS^ 1.68±0.04^NS^ 4.20±1.06^NS^ 0.40±0.02^NS^  Cadmium 5.95±1.04^NS^ 1.73±0.04^NS^ 4.20±1.05^NS^ 0.42±0.02^NS^  Ferrous sulfate + vanaspai 5.51±1.04^NS^ 1.58±0.04^NS^ 3.91±1.05^NS^ 0.41±0.02^NS^  Chlorpyrifos + vanaspai 5.77±1.04^NS^ 1.61±0.04^NS^ 4.12±1.05^NS^ 0.41±0.02^NS^  Cadmium + vanaspai 5.32±1.04^NS^ 1.56±0.04^NS^ 3.75±1.05^NS^ 0.42±0.02^NS^ B  Vanaspai 5.83±0.013^NS^ 1.60±0.02 4.23±0.13^NS^ 0.41±0.10^NS^  Normal 5.96±0.014^NS^ 1.69±0.02^\*^ 4.27±0.13^NS^ 0.42±0.10^NS^ A: effect of individual treatment; B: overall effect of vanaspati supplementation.; Values are Mean±SE. Two-way ANOVA followed by Tukey's HSD *post hoc* test; Means with different superscripts are significantly different (*P*\<0.05) ::: ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Marginal means of immunological parameters in different groups ::: Group HI (log~2~) PHA index (mm) ----------------------------- ------------------- ---------------- A  Normal 424.00±103.29^b^ 0.43±0.05^ab^  Vanaspai 69.00±30.64^a^ 0.65±0.06^bcd^  Ferrous sulfate 432.00±13.13^b^ 0.47±0.09^ac^  Chlorpyrifos 512.00±125.69^b^ 0.42±0.10^a^  Cadmium 1536.00±193.52^c^ 0.63±0.09^bcd^  Ferrous sulfate + vanaspai 62.00±11.49^a^ 0.75±0.18^bcd^  Chlorpyrifos + vanaspai 114.00±25.20^a^ 0.82±0.06^d^  Cadmium + vanaspai 82.50±18.68^a^ 0.56±0.09^ad^ B  Vanaspai 81.86±11.35 0.69±0.06^\*^  Normal 726.00±107.96^\*^ 0.49±0.04 A: effect of individual treatment; B: overall effect of vanaspati supplementation; Values are Mean±SE. Two-way ANOVA followed by Tukey's HSD *post hoc* test; Means with different superscripts are significantly different (*P*\<0.05); HI, hemagglutination inhibition; PHA, phytohemagglutination ::: In the present study, the activities of ALT and ALP were determined to assess the degree of damage to the liver as the levels of certain enzymes like ALT, AST, gamma glutamyl transferase (GGT), etc., are shown to be elevated following hepatocellular injury.\[[@CIT10]\] In this study, the activities of ALT and ALP were significantly elevated in the cadmium toxic control group, whereas ALT activity was elevated in ferrous sulfate group, suggesting the hepatocellular insult following administration of cadmium and ferrous sulfate. The supplementation of vanaspati with ferrous sulfate resulted in an elevated activity of ALP. The results are in accordance with those of Demerdesh *et al*.,\[[@CIT11]\] Kara *et al*,\[[@CIT12]\] and Yadav *et al*.,\[[@CIT13]\] who reported an increase in the activities of ALT and ALP in plasma following cadmium toxicity. An increase in the total cholesterol, triglycerides and low density lipoprotein (LDL) with the decrease in HDL in serum will serve as a biomarker for hepatopathy, cardiac damage as well as renal failure.\[[@CIT10]\] In the present study, no changes in the serum cholesterol profile and protein profile were detected except that cadmium paradoxically caused a significant elevation of HDL and significant reduction in triglycerides. CPS produced a non-significant decrease in HDL and increase in total lipids of liver. Supplementation of vanaspati along with the stressors showed no effect on the above parameters except that HDL was increased in cadmium group. The overall effect of vanaspati supplementation revealed a significantly decreased serum cholesterol and albumin. The increase in HDL and decrease in triglyceride concentration of cadmium group is attributed to the possible interference of cadmium with the test procedures. HDL was, however, low in CPS group, which is explained by the fact that HDL that is associated with paraoxonase enzyme, which helps in breaking down organophosphorus (OP) pesticides\[[@CIT14]\] and hence is bound to CPS oxon.\[[@CIT15]\] The supplementation of vanaspati resulted in decreased serum cholesterol level. Such a reduction in cholesterol level is attributed to the significant amounts of polyunsaturated fatty acids (PUFAs) present in vanaspati, despite hydrogenation, which gets incorporated into and consequently weakens the cell membranes when insufficient saturated fats are not available. This leads to deposition of cholesterol to strengthen the cell membranes and consequently results in lowering of blood cholesterol.\[[@CIT16]\] It was observed that the albumin level was reduced upon supplementation of vanaspati. This is due to fact that fatty acids are transported in a bound form with albumin in the blood stream.\[[@CIT17]\] In the present study, albumin was estimated by dye binding method and hence the method could not be used to estimate fatty acid bound albumin. Non-protein nitrogenous (NPN) substances, such as serum creatinine, are increased only when renal function is below 30% of its original capacity in birds. In the present study, creatinine was significantly increased in the cadmium group and supplementation of vanaspati had no effect. The increased creatinine levels in serum indicate kidney damage. The elevated levels of creatinine are due to the damage caused by cadmium to the hepatic and renal tissue, leading to the liberation of marker enzymes into serum. The results are in accordance with those of Kara *et al*.,\[[@CIT12]\] and Yadav *et al*.\[[@CIT13]\] In the present study, stressors showed little effect upon immunity. Paradoxically, cadmium showed a significant increase in antibody titer. A similar paradoxical increase in antibody titer was reported by Robohm.\[[@CIT18]\] However, it is assumed that the increase in antibody titer is probably a case of interference of cadmium with hemagglutination of RBC because the corresponding increase in globulin level of serum was not observed. The overall supplementation of vanaspati reduced HI titer. The PHA index, which is a measure of cell mediated immunity and T-cell proliferation, was significantly increased in vanaspati fed groups. This is probably due to the fact that higher intakes of partially hydrogenated vegetable oils are associated with elevated concentrations of inflammatory biomarkers like C-reactive protein, tumor necrosis factor-α, interleukin-6 and soluble intercellular adhesion molecule-1.\[[@CIT19]\] Further, it was reported that the CD4:CD8 T-lymphocytes ratio was increased by dietary *trans* fatty acids.\[[@CIT20]\] The results of the present study enunciated the fact that cadmium was more potent in inducing damage and supplementation of vanaspati was of no ameliorative value in any of the toxic models. Further, the toxicities of CPS and cadmium were aggravated with the supplementation of vanaspati. Hence, it is concluded that consumption of vanaspati is harmful and leads to deleterious effects on health. **Source of Support:** Nil **Conflict of Interest:** None declared

PubMed Central

2024-06-05T04:04:19.384487

2011-01-01

PMC3052582

INTRODUCTION {#sec1-1} ============ Nowadays, use of vitamin D~3~ (cholecalciferol) in commercial pet, livestock and infant feed supplements, multivitamin preparations and as a rodenticide has increased the risk of its toxicity. Various plant species that have a high concentration of vitamin D analogue have been reported as a source of vitamin D~3~ toxicity in livestock. The most common source of vitamin D~3~ toxicity in dogs and cats is accidental ingestion of rodenticide baits containing cholecalciferol.\[[@CIT1]\] So, the present study was designed to examine the clinical signs, hematological, biochemical and histopathological aspects and ameliorative effect of *Aloe vera* on vitamin D~3~ toxicity in rats. MATERIALS AND METHODS {#sec1-2} ===================== Thirty-two healthy adult male *Wistar* albino rats were weighed and housed in laboratory animal cages. The rats were fed on standard laboratory animal feed for rat and mice, supplied by Godrej Agrovet Limited, Mumbai. The rats were kept for 7 days to acclimatize in laboratory condition prior to start of sampling protocols. On day 7, the rats were randomly divided into four groups (A, B, C and D) of eight rats in each group. Group A was given vitamin D~3~ orally at 2 mg/kg single dose daily, dissolved in groundnut oil. Group B was given vitamin D~3~ orally at 2 mg/kg dissolved in groundnut oil single dose daily and 2.5% commercial *Aloe vera* preparation in drinking water daily. Group C served as control group and was administered groundnut oil only. Group D was given 2.5% commercial *Aloe vera* preparation in drinking water daily. The rats were anesthetized by using diethyl-ether inhalation anesthesia method and blood samples were collected by retro-orbital plexus method on day 0, day 10 or just before death in 1% heparin solution for biochemistry and in ethylenediaminetetraacetic acid (EDTA) for hematological analysis. Necropsy was performed on rats died due to toxicity and relevant tissues were collected in 10% buffered formalin saline for histopathological studies. The experiment was approved by Institutional Animal Ethics Committee (IAEC), GADVASU, Ludhiana. Estimation of biochemical parameters {#sec2-1} ------------------------------------ The collected blood samples were centrifuged immediately at 1500 rpm for 15 minutes and plasma was separated and stored at 4°C. These plasma samples were used for estimation of biochemical parameters. The biochemical parameters were estimated on auto analyzer by using diagnostic reagent kits (Autopak) supplied by Bayer Diagnostics India Limited, Baroda, Gujarat, India. The mean values of concentrations estimated in plasma of respective groups were compared by using one-way analysis of variance (ANOVA) test. Collection, processing and staining of tissue samples for histopathology {#sec2-2} ------------------------------------------------------------------------ Necropsy of rats that died due to toxicity was performed as early as possible and the tissue/specimens consisting of pieces of various organs were collected in 10% formalin for histopathological examination. The control group rats were anesthetized by using diethyl-ether inhalation anesthesia method and killed by exsanguinations. The formalin-fixed tissues were washed overnight in running tap water, dehydrated in ascending grades of alcohol and cleared in benzene. The 4--5 *μ*m thick tissue sections were cut from the paraffin embedded tissues and were stained with hematoxylin and eosin stain (HandE) for routine histopathology. Wherever necessary, duplicate sections were stained to confirm calcification in tissues with Von Kossa stain.\[[@CIT2]\] RESULTS AND DISCUSSION {#sec1-3} ====================== Clinical signs observed {#sec2-3} ----------------------- On day 8 of treatment, the rats of group A and group B were dull and showed decrease in feed and water intake. These signs became more intense with course of treatment. On days 10, 11 and onward, some rats of the treatment groups stopped taking feed and water. They were dull, showing difficulty in movement and rigidity of limbs. They showed ataxic movements or could not even open their mouth. The body coat was ruffled and the animals showed severe cachexia. Diarrhea was observed in rats of groups A and B besides difficulty in respiration, shivering and epistaxis. Before death, there were nervous signs like aimless running, rolling and epileptic seizures. On day 10, one rat from group A and two rats from group B were found dead. Mortality of rats in groups A and B was observed between day 10 and day 19 of treatment. A significant decrease in body temperature of groups A and B was recorded on day 10 of treatment as compared to their average body temperature before treatment on day 0, while no significant change was found in the body temperature of rats in groups C and D \[[Table 1](#T0001){ref-type="table"}\]. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Effect on body temperatures of treated and control group ::: Day Group A (°F) (Mean±SD) Group B (°F) (Mean±SD) Group C(°F) (Mean±SD) Group D (°F) (Mean±SD) ------------------------------------------------------ ---------------------------------------------- ---------------------------------------------- ------------------------------------------------ ------------------------------------------------ Day 0 (*n* = 8) 99.76±0.83 99.31±0.42 99.58±0.43 99.13±0.58 Day 10 (for group A *n* = 7 and for group B *n* = 6) 95.30±1.12[\*](#T000F1){ref-type="table-fn"} 95.10±0.96[\*](#T000F1){ref-type="table-fn"} 99.44±0.26[\*\*](#T000F2){ref-type="table-fn"} 99.06±0.57[\*\*](#T000F2){ref-type="table-fn"} \* Represents values differing significantly from day 0 value of that group at 5% level of significance by one-way ANOVA test; \*\* Represents values differing non-significantly from day 0 value of that group at 5% level of significance by one-way ANOVA test ::: A significant decrease in body weight of rats in groups A and B was observed on day 10 of treatment as compared to their average body weights before treatment (day 0). The total decrease in average body weight was 23.17% (76.67 g) and 22.61% (80.84 g) in groups A and B, respectively. No significant change was recorded in body weights of groups C and D \[[Table 2](#T0002){ref-type="table"}\]. ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Effect on body weight of treated and control groups ::: Day Group A (g) (Mean±SD) Group B (g) (Mean±SD) Group C (g) (Mean±SD) Group D (g) (Mean±SD) ------------------------------------------------------- ------------------------------------------------- ------------------------------------------------- --------------------------------------------------- --------------------------------------------------- Day 0 (*n* = 8) 330.83±41.03 357.50±19.42 354.28±33.09 357.85±49.23 Day 10 (for group A *n* = 7, and for group B *n* = 6) 254.16 ±33.82[\*](#T000F3){ref-type="table-fn"} 276.66 ±17.79[\*](#T000F3){ref-type="table-fn"} 357.85 ±32.25[\*\*](#T000F4){ref-type="table-fn"} 359.28 ±48.08[\*\*](#T000F4){ref-type="table-fn"} \* Represents values differing significantly from 0 day value of that group at 5% level of significance by one-way ANOVA test; \*\* Represents values differing non-significantly from 0 day value of that group at 5% level of significance by one-way ANOVA test ::: The clinical signs of vitamin D~3~ toxicity like anorexia, diarrhea, progressive emaciation/weight loss, dehydration, weakness and difficulty in movement observed in the present study are in consonance with earlier findings in different species.\[[@CIT3][@CIT4]\] Morrow\[[@CIT5]\] reported that with unregulated increase in plasma calcium and phosphorus in vitamin D~3~ toxicity, their product (calcium × phosphorus) can rise above 60, which causes mineralization of tissues/organs like kidneys, gastrointestinal tract (GIT), cardiac muscles and blood vessels, leading to structural damage with eventual decreased functional capacity. This loss of function contributes to the development of ongoing and end-stage clinical signs as well as long-term signs in animals that survive. In the present study, hypercalcemia and hyperphosphatemia along with mineralization of various organs was observed. Thus, the clinical signs observed were the sequelae of biochemical and histopathological changes due to vitamin D~3~ toxicity. Biochemical findings {#sec2-4} -------------------- The plasma concentrations of various biochemical parameters before treatment and on day 10 of treatment are shown in Tables [3](#T0003){ref-type="table"} and [4](#T0004){ref-type="table"}, respectively. On day 10 of treatment, the plasma concentration of calcium, phosphorus and blood urea nitrogen (BUN) in vitamin D~3~ treated groups (A and B) was found to be significantly increased, while the plasma concentration of total plasma protein and albumin was found to be significantly decreased before death as compared to their average values in control groups \[Tables [3](#T0003){ref-type="table"} and [4](#T0004){ref-type="table"}\]. ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Effect on biochemical parameters of treated and control groups on day 0 ::: Parameter Groups Mean±SD Range ----------------------------- ------------------------------------------------------- ------------ -------------- Calcium (mg/dl) Group A[\*\*](#T000F5){ref-type="table-fn"} (*n* = 8) 9.68±0.27 9.2--10.0 Group B[\*\*](#T000F5){ref-type="table-fn"} (*n* = 8) 9.62±0.27 9.3--10.0 Group C (*n* = 8) 9.61±0.30 9.1--10.3 Group D[\*\*](#T000F5){ref-type="table-fn"} (*n* = 8) 9.50±0.36 9.1--10.1 Phosphorus (mg/dl) Group A[\*\*](#T000F5){ref-type="table-fn"} 4.55±0.31 4.1--4.9 Group B[\*\*](#T000F5){ref-type="table-fn"} 4.50±0.32 4.1--5.0 Group C 4.42±0.23 4.1--5.1 Group D[\*\*](#T000F5){ref-type="table-fn"} 4.60±0.34 4.1--4.9 Albumin (g/dl) Group A[\*\*](#T000F5){ref-type="table-fn"} 2.87±0.19 2.6--3.2 Group B[\*\*](#T000F5){ref-type="table-fn"} 2.77±0.24 2.5--3.1 Group C 2.63±0.41 2.0--3.2 Group D[\*\*](#T000F5){ref-type="table-fn"} 2.78±0.25 2.5--3.2 Total plasma protein (g/dl) Group A[\*\*](#T000F5){ref-type="table-fn"} 7.12±0.35 7--8 Group B[\*\*](#T000F5){ref-type="table-fn"} 7.0±0.0 7.00 Group C 7.0±0.0 7.00 Group D[\*\*](#T000F5){ref-type="table-fn"} 7.12±0.35 7--8 BUN (mg/dl) Group A[\*\*](#T000F5){ref-type="table-fn"} 22.43±3.62 18.72--27.60 Group B[\*\*](#T000F5){ref-type="table-fn"} 21.74±3.42 18.14--27.86 Group C 22.26±2.86 18.24--26.78 Group D[\*\*](#T000F5){ref-type="table-fn"} 22.24±2.93 18.24--26.14 ^\*^Represents values differing significantly from value of control rats at 5% level by one-way ANOVA test; \*\* Represents values differing non-significantly from value of control rats at 5% level by one-way ANOVA test ::: ::: {#T0004 .table-wrap} Table 4 ::: {.caption} ###### Effect on biochemical parameters of treated and control groups on day 10 ::: Parameter Groups Mean±SD Range ----------------------------- ------------------------------------------------------- ------------- ------------- Calcium (mg/dl) Group A[\*](#T000F6){ref-type="table-fn"} (*n* = 7) 13.76±1.73 10.20-16.30 Group B[\*](#T000F6){ref-type="table-fn"} (*n* = 6) 13.30±0.95 11.70-14.40 Group C (*n* = 8) 9.72±0.32 9.2-10.3 Group D[\*\*](#T000F7){ref-type="table-fn"} (*n* = 8) 9.52±0.31 9.0-10.0 Phosphorus (mg/dl) Group A[\*](#T000F6){ref-type="table-fn"} 5.33±0.89 4.6-6.8 Group B[\*](#T000F6){ref-type="table-fn"} 5.36±1.032 4.4-7.20 Group C 4.37±0.14 4.1-4.6 Group D[\*\*](#T000F7){ref-type="table-fn"} 4.38±0.21 4.1-4.7 Albumin (g/dl) Group A[\*](#T000F6){ref-type="table-fn"} 0.75±0.30 0.40-1.3 Group B[\*](#T000F6){ref-type="table-fn"} 0.87±0.44 0.2-1.50 Group C 2.73±0.22 2.40-3.10 Group D[\*\*](#T000F7){ref-type="table-fn"} 2.82±0.29 2.4-3.30 Total plasma protein (g/dl) Group A[\*](#T000F6){ref-type="table-fn"} 5.93±0.30 5.3-6.3 Group B[\*](#T000F6){ref-type="table-fn"} 5.76±0.27 5.3-6.1 Group C 7.31±0.45 7-8 Group D[\*\*](#T000F7){ref-type="table-fn"} 7.37±0.51 7-8 BUN (mg/dl) Group A[\*](#T000F6){ref-type="table-fn"} 51.54±13.09 37.31-74.86 Group B[\*](#T000F6){ref-type="table-fn"} 51.70±10.07 37.0-68.31 Group C 22.06±2.48 18.40-25.90 Group D[\*\*](#T000F7){ref-type="table-fn"} 22.13±2.95 18.10-26.14 \* Represents values differing significantly from value of control rats at 5% level by one-way ANOVA test; \*\* Represents values differing non-significantly from value of control rats at 5% level by one-way ANOVA test ::: The increase in calcium levels or hypercalcemia in cholecalciferol toxicity in the present study is in consonance with the earlier studies on vitamin D~3~ toxicity by different workers. In vitamin D~3~ toxicity, the active metabolites of cholecalciferol have been reported to increase the blood calcium (hypercalcemia) level by increased resorption/mobilization of calcium from bone, increased absorption of calcium from intestine and decreased calcium excretion by kidney. The net result is high concentration of blood calcium level (hypercalcemia) and death reported due to renal and cardiac failure.\[[@CIT6]\] The significant increase found in the plasma phosphorus levels in the present study is in accordance with findings of earlier studies on vitamin D~3~ toxicity. Morrow\[[@CIT5]\] reported that in case of acute toxicity of cholecalciferol, there is a moderate rise in serum/plasma phosphorus concentration up to 11 mg/dl and it is due to stimulation of transfer of phosphorus along with calcium from bone to plasma. The net result is mineralization of the kidney, GIT, cardiac muscles and blood vessels, causing structural damage and decrease in functional capacity of these tissues and organs. The elevated level of BUN in this study indicates renal damage. Histopathologically, hypercalcemia-induced extensive glomerular and renal tubular damage (nephrocalcinosis) was seen in the present study as reported earlier in vitamin D~3~ toxicity.\[[@CIT7]\] The increased level of BUN in the present study was in agreement with that reported earlier by various workers.\[[@CIT8]\] The significant decrease found in the levels of albumin and total plasma protein was thought to be due to loss of protein due to extensive renal damage. Hematological findings {#sec2-5} ---------------------- On day 10, the concentration of hemoglobin and total leukocyte count of groups A and B increased non-significantly as compared to group C (control) rats. The mean relative neutrophil counts of groups A and B increased significantly from group C (control) rats on day 10, whereas counts of other cells differed non-significantly from control rats. The hemoglobin, total leukocyte count and differential leukocyte count of group D rats differed non-significantly from control rats on day 10. The relative increase in neutrophil count could be due to ongoing tissue necrosis as confirmed by histopathological observations. Gross postmortem changes observed {#sec2-6} --------------------------------- The rats died due to vitamin D~3~ toxicity (groups A and B) or sacrificed showing signs of vitamin D~3~ toxicity had severe emaciation and ruffled body coat. The heart was felt hard with white chalky deposits on epicardial surface. The stomach and intestines showed diffused white chalky deposits on serosal surface. The stomach and intestines revealed bloody ingesta in the lumen with marked hemorrhages on mucosa. Pin point white chalky deposits were observed on capsules of both kidneys \[[Figure 1](#F0001){ref-type="fig"}\]. The liver, lung and other organs grossly appeared normal. Gross pathological changes like bloody ingesta in stomach and intestine were observed in the present study as also reported by Long,\[[@CIT8]\] due to vitamin D~3~ toxicity in pigs. ::: {#F0001 .fig} Figure 1 ::: {.caption} ###### Necropsy of rats that died due to vitamin D~3~ toxicity, showing the presence of white deposits of mineralization on epicardial surface of heart, serosal surface of stomach and intestine and pin point white deposits on capsular surface of kidney. Arrows indicate sites of calcification (mineralization) :::  ::: Histopathological findings {#sec2-7} -------------------------- The following histopathological changes were observed in the tissues collected from rats died due to vitamin D~3~ toxicity (groups A and B) or sacrificed rats showing clinical signs of toxicity. The histopathological changes are presented below according to severity of lesions: ### Tongue {#sec3-1} The lamina propria, connective tissue in-between muscle fibers and tunica intima of blood vessels of tongue showed purple deposits of calcification. Around the calcified areas, there was mild lymphomononuclear cell infiltration. The calcification was also confirmed by Von Kossa staining as black deposits of calcification \[Figures [2](#F0002){ref-type="fig"} and [3](#F0003){ref-type="fig"}\]. ::: {#F0002 .fig} Figure 2 ::: {.caption} ###### Section of tongue showing purple deposits of calcification in between muscle fibers and lymphomononuclear cell infiltration (H and E, ×150). Arrow indicate sites of calcification (mineralization) :::  ::: ::: {#F0003 .fig} Figure 3 ::: {.caption} ###### Section of tongue showing calcification within wall of blood vessel (H and E, ×150). Arrow indicate sites of calcification (mineralization) :::  ::: ### Stomach {#sec3-2} Calcification was observed in mucosa, muscularis mucosa and muscularis externa layers of stomach. Calcification of submucosal and serosal blood vessels was observed. Around the area of calcification, there was lymphomononuclear cell infiltration. Hemorrhagic gastritis was also observed in some rats. The calcification was confirmed by Von Kossa staining \[Figures [4](#F0004){ref-type="fig"}--[6](#F0006){ref-type="fig"}\]. ::: {#F0004 .fig} Figure 4 ::: {.caption} ###### Section of stomach showing calcification of mucosa, muscularis mucosa and muscularis externa (H and E, ×75). Arrows indicate sites of calcification (mineralization) :::  ::: ::: {#F0005 .fig} Figure 5 ::: {.caption} ###### Section of stomach showing calcification within mucosa of stomach (H and E, ×150). Arrow indicate sites of calcification (mineralization) :::  ::: ::: {#F0006 .fig} Figure 6 ::: {.caption} ###### Section of stomach showing black deposits of mineralization within mucosa, muscularis mucosae, muscularis externa and within walls of blood vessels (Von Kossa stain, ×75). Arrows indicate sites of calcification (mineralization) :::  ::: ### Small and large intestine {#sec3-3} The lesions observed in small intestine included calcification of the Brunner's gland epithelium and sloughing of intestinal villi epithelium. The histopathological lesions in large intestine included calcification of muscularis externa layer \[[Figure 7](#F0007){ref-type="fig"}\]. ::: {#F0007 .fig} Figure 7 ::: {.caption} ###### Section of large intestine showing calcification of muscularis externa layer (H and E, ×75). Arrow indicate sites of calcification (mineralization) :::  ::: ### Kidney {#sec3-4} Mineralization (nephrocalcinosis) was observed mainly in cortex and medulla. The tubular epithelium showed calcification and coagulative necrosis along with presence of proteinaceous casts in the lumen. The calcification was also observed in basement membranes of tubules. In medulla, mineralization and degeneration was mainly observed in the collecting tubules. The walls of renal blood vessels showed calcification. Mineralization was observed in glomerular capsules and inside the glomerulus. The deposits of mineralization were also present in interstitium \[Figures [8](#F0008){ref-type="fig"} and [9](#F0009){ref-type="fig"}\]. ::: {#F0008 .fig} Figure 8 ::: {.caption} ###### Section of kidney showing necrosis and calcification of basement membrane of tubular epithelium (H and E, ×750). Arrows indicate sites of calcification (mineralization) :::  ::: ::: {#F0009 .fig} Figure 9 ::: {.caption} ###### Section of kidney showing black deposits of mineralization within cortical tubules (Von Kossa stain, ×75). Arrows indicate sites of calcification (mineralization) :::  ::: ### Larynx and trachea {#sec3-5} Calcification was observed in lamina propria of larynx. The tracheal cartilage and basement membrane of tracheal mucosal epithelium showed mineralization. The tracheal mucosa was infiltrated with inflammatory cells with exfoliated mucosal epithelial cells \[[Figure 10](#F0010){ref-type="fig"}\]. ::: {#F0010 .fig} Figure 10 ::: {.caption} ###### Section of trachea showing black deposits of calcification within cartilage and basement membrane of mucosal epithelium (Von Kossa stain, ×150). Arrows indicate sites of calcification (mineralization) :::  ::: ### Lungs {#sec3-6} Calcification of bronchial mucosal epithelium was observed. In lungs, brown dirty color granular deposits of calcium were observed in alveolar and interalveolar septae. There was infiltration of lymphocytes, macrophages, occasional neutrophils and few giant cells in the thickened alveolar septae. Other histopathological changes in lungs included marked hemorrhage, emphysema, and edema. The calcification in bronchi and lung was confirmed by Von Kossa staining. The Von Kossa staining of lung sections showed extensive mineralization of alveolar septae \[[Figure 11](#F0011){ref-type="fig"}\]. ::: {#F0011 .fig} Figure 11 ::: {.caption} ###### Section of lung showing black deposits of mineralization within alveolar septae and bronchial mucosa (Von Kossa stain, ×150). Arrows indicate sites of calcification (mineralization) :::  ::: ### Heart {#sec3-7} Calcification was observed in epicardium, myocardium and endocardium, cardiac valves and coronary arteries. The histopathological changes included calcification, degeneration and replacement of cardiac muscle cells with fibrous tissue along with lymphomononuclear cell infiltration. The epicardium and myocardium were severely affected with extensive fibrous tissue proliferation and severe lymphomononuclear infiltration \[Figures [12](#F0012){ref-type="fig"} and [13](#F0013){ref-type="fig"}\]. ::: {#F0012 .fig} Figure 12 ::: {.caption} ###### Section of heart showing marked calcification within wall of coronary blood vessel and myocardial muscle fibers (H and E, ×150). Arrows indicate sites of calcification (mineralization) :::  ::: ::: {#F0013 .fig} Figure 13 ::: {.caption} ###### Section of heart showing black deposits of calcification within myocardial fibers and coronary blood vessel (Von Kossa stain, ×75). Arrows indicate sites of calcification (mineralization) :::  ::: ### Aorta {#sec3-8} The histopathological changes observed in aorta included calcification and structural derangement of tunica media. The deposits of calcium were observed in between the elastic fibers of tunica media. The elastic fibers of tunica media were widely separated due to calcium deposits in between them causing structural derangement. The calcification in aortic media was confirmed by Von Kossa staining \[Figures [14](#F0014){ref-type="fig"} and [15](#F0015){ref-type="fig"}\]. ::: {#F0014 .fig} Figure 14 ::: {.caption} ###### Section of aorta showing calcification within tunica media (H and E, ×75). Arrows indicate sites of calcification (mineralization) :::  ::: ::: {#F0015 .fig} Figure 15 ::: {.caption} ###### Section of aorta showing black deposits of mineralization within tunica media (Von Kossa stain, ×30). Arrows indicate sites of calcification (mineralization) :::  ::: ### Spleen {#sec3-9} The histopathological changes found in spleen included calcification of central arteries of lymphatic nodules and calcification of capsular surface of spleen \[[Figure 16](#F0016){ref-type="fig"}\]. ::: {#F0016 .fig} Figure 16 ::: {.caption} ###### Section of spleen showing calcification of wall of central arteries of lymphatic nodules (H and E, ×300). Arrows indicate sites of calcification (mineralization) :::  ::: ### Liver {#sec3-10} The histopathological changes observed in liver included fatty/vacuolative degeneration of hepatocytes. ### Adrenal {#sec3-11} There was vacuolation of adrenal cortical cells and marked congestion of medulla. ### Brain {#sec3-12} Calcification of choroid plexus arteries was observed along with other histopathological changes like neuronophagia and satellitosis in cerebral cortex \[[Figure 17](#F0017){ref-type="fig"}\]. The findings are in consonance with earlier findings reported by Morita *et al* (1995) and Hilbe *et al* (2000) in cholecalciferol toxicity. Morita *et al* (1995) and Hilbe *et al* (2000) reported the calcification of choroid plexus blood vessels in cats and dog respectively. ::: {#F0017 .fig} Figure 17 ::: {.caption} ###### Section of cerebral hemisphere showing calcification of choroid plexus blood vessel (H and E, ×750). Arrows indicate sites of calcification (mineralization) :::  ::: ### Vas deferens {#sec3-13} The histopathological changes observed in vas deferens included the calcification and degeneration of muscle fibers of tunica muscularis layer \[[Figure 18](#F0018){ref-type="fig"}\]. ::: {#F0018 .fig} Figure 18 ::: {.caption} ###### Section of vas deferens showing calcification within muscular layer (H and E, ×150). Arrows indicate sites of calcification (mineralization) :::  ::: The effects of hypercalcemia on cells include altered cell membrane permeability, altered calcium pump activity, decreased cellular energy production and cellular necrosis. As the toxicity increases, clinical signs including hypertension, polyuria and extremely elevated serum calcium level may cause cardiac arrhythmias.\[[@CIT5]\] The mechanism of mineralization or calcification in vitamin D~3~ toxicity is well explained by Morrow,\[[@CIT5]\] who reported that with unregulated increase in plasma calcium and phosphorus in vitamin D~3~ toxicity, their product (calcium × phosphorus) can rise above 60, which causes mineralization of tissues/organs like kidneys, GIT, cardiac muscles, blood vessels and causes structural damage that leads to decreased functional capacity of these tissues and organs. The loss of function contributes to the development of ongoing end-stage clinical signs as well as long-term signs in animals that survive. The cause of death reported in vitamin D~3~ toxicity includes cardiac and renal failure. The histopathological changes observed within various organs in present study are in agreement with earlier reports\[[@CIT6]--[@CIT8]\] on vitamin D~3~ toxicity in different animal species. Effect of *Aloe vera* on vitamin D3 toxicosis {#sec2-8} --------------------------------------------- The biochemical, hematological and histopathological parameters of group B (vitamin D~3~ at 2 mg/kg dissolved in groundnut oil daily single dose plus 2.5% *Aloe vera* juice in drinking water daily) were evaluated to assess the protective role of *Aloe vera* on vitamin D~3~ toxicity. These parameters showed same clinical signs, gross postmortem, biochemical, hematological and histopathological changes as observed in group A (vitamin D~3~ at 2 mg/kg dissolved in groundnut oil single dose daily), in comparison with control rats. The biochemical and hematological changes in group B (on day 10) differed non-significantly from group A rats. Kawashima *et al*.,\[[@CIT10]\] reported the inhibitory effect of aspirin on vitamin D~3~ -induced hypercalcemia. Heggers *et al*.,\[[@CIT11]\] reported that *Aloe vera* gel has anti-inflammatory properties like aspirin. The earlier studies also reported the protective effect of *Aloe vera* in toxicants like lindane and arsenic.\[[@CIT12][@CIT13]\] In the present study, no protective effect of *Aloe vera* was found in vitamin D~3~ toxicity. In conclusion, the study revealed clinical signs of severe progressive emaciation, diarrhea, difficulty in respiration and movement, nervous signs like aimless running, rolling and epileptic seizures and subnormal body temperature before death in vitamin D~3~ toxicity. The biochemical findings included increased levels of blood calcium, phosphorus and BUN, while a significant decrease was observed in the levels of total protein and albumin. The histopathological findings included calcification of various organs throughout the body. *Aloe vera* was found having no ameliorative effect against vitamin D~3~ toxicity. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.385939

2011-01-01

PMC3052583

INTRODUCTION {#sec1-1} ============ Chlorpyrifos (CPS) is used in controlling a variety of insects, flees, termites, lice etc. It is used as an insecticide on grain, fruit nut and vegetable crops. Residual amounts of CPS are detected in water, soil, fabric and on surfaces for months to years. Goel *et al*.,\[[@CIT1]\] reported that CPS induces toxicity by generating free radicals and by altering the levels of enzymatic and non-enzymatic antioxidant defenses. Several herbs, spices and many Indian medicinal plants are considered to be potential sources of antioxidants.\[[@CIT2]\] Therefore, the present study was planned on broilers, using certain herbs and herbal combinations against CPS-induced toxicity with reference to lipid and protein profile. MATERIALS AND METHODS {#sec1-2} ===================== A total of 225 sexed male broiler chicks (*Cobb* strain) of a day-old age were procured for the study. The chicks were randomly divided into 15 groups consisting of 15 chicks in each group. All the birds were provided with feed and water *ad libitum* throughout the experiment. All the groups were maintained as per the following treatment schedule for 6 wk: 1. Group 1: Basal diet control (1-42 days) 2. Group 2: Chlorpyrifos (100 ppm) toxic control (1--42 days) 3. Group 3: Chlorpyrifos (1--28 days) + *Withania somnifera* at 0.1% of feed (29--42 days) 4. Group 4: Chlorpyrifos (1--28 days) + *Ocimum sanctum* at 0.1% of feed (29--42 days) 5. Group 5: Chlorpyrifos (1--28 days) + *Asparagus racemosus* at 0.1% of feed (29--42 days) 6. Group 6: Chlorpyrifos (1--28 days) + *Andrographis paniculata* at 0.1% of feed (29--42 days) 7. Group 7: Chlorpyrifos (1--28 days) + *Murraya koenigii* at 0.1% of feed (29--42 days) 8. Group 8: Chlorpyrifos (1--28 days) + shilajit at 0.1% of feed (29--42 days) 9. Group 9: Chlorpyrifos (1--28 days) + *Gymnema sylvestre*at 0.1% of feed (29--42 days) 10. Group10: Chlorpyrifos (1--28 days) + *Allium sativum* at 0.1% of feed (29--42 days) 11. Group 11: Chlorpyrifos (1--28 days) + spirulina at 0.1% of feed (29--42 days) 12. Group 12: Chlorpyrifos (1--28 days) + ginseng at 0.1% of feed (29--42 days) 13. Group 13: Chlorpyrifos (1--28 days) + *Withania somnifera + Asparagus racemosus + Andrographis paniculata* at 0.05% of each in feed (29-42 days) 14. Group 14: Chlorpyrifos (1-28 days) + *Withania somnifera* + *Murraya koenigii* + *Allium sativum* at 0.05% of each in feed (29+42 days) 15. Group 15: Chlorpyrifos (1-28 days) + spirulina + shilajit + *Gymnema sylvestre* at 0.05% of each in feed (29-42 days) Group 1 was negative control and group 2 was toxic control throughout the study. Groups 3--15 were toxic controls up to the end of 4^th^ week (1--28 days), and subsequently, these were kept on treatment with herbs and herbal combinations for the next two weeks (29--42 days). The blood samples were drawn from wing vein at the end of 4^th^ and 6^th^ weeks for the estimation of lipid and protein profile, by using commercially available diagnostic kits (Qualigens Pvt. Ltd., Mumbai, India). The birds were sacrificed at the end of 6^th^ week and liver tissues were collected for histological examination. The data were subjected to statistical analysis by applying one-way analysis of variance (ANOVA) using statistical package for social sciences (SPSS) 10^th^ version. Differences between means were tested using Duncan's multiple comparison test and significance was set at *P*\<;0.05. RESULTS AND DISCUSSION {#sec1-3} ====================== The concentrations of total cholesterol, low density lipoprotein (LDL) cholesterol and triglycerides (mg/dl) recorded at the end of 4^th^ week in the basal diet control (group 1) were 128.78±0.67, 70.11±1.11 and 31.83±1.29, respectively, which were significantly (*P*\<;0.05) increased in the CPS toxic controls (groups 2--15) with the values ranging from 143.77±0.49 to 147.81±1.21, from 94.73±0.60 to 98.56±1.13 and from 41.01±0.99 to 42.50±0.64, respectively. Following supplementation with herbs in groups 3--15, there was a significant (*P*\<;0.05) decrease in the cholesterol levels at the end of 6^th^ week with the values ranging from 170.25±0.53 to 181.00±1.99, from 102.59±1.30 to 114.06±3.24 and from 47.91±0.76 to 55.75±3.52, respectively, as compared to the CPS toxic control group 2 (195.51±1.48, 133.08±1.31 and 67.52±1.60, respectively). However, these values were significantly (*P*\<;0.05) higher when compared to the basal diet control (162.90±3.44, 92.02±3.23 and 31.83±1.29, respectively). The high density lipoprotein (HDL) cholesterol concentration (mg/dl) recorded at the end of 4^th^ week in the basal diet control (group 1) was 52.31±0.71, which was significantly (*P*\<;0.05) decreased in CPS toxic controls (groups 2--15) with the values ranging from 40.03±0.38 to 40.98±0.13. Following supplementation with herbs in groups 3--15, there was significant (*P*\<;0.05) increase in the HDL cholesterol concentration at the end of 6^th^ week (values ranged from 55.82±2.28 to 60.42±0.83) as compared to their corresponding 4^th^ week values and the CPS control group 2 (31.83±1.29). However, the groups 5--8 and 10--15 showed no significant difference as compared to the basal diet control (62.15±0.86), but the remaining groups showed significantly (*P*\<;0.05) lower levels of HDL cholesterol. An increase in the total cholesterol, triglycerides and LDL with a decrease in HDL in serum serves as an index for hepatopathy, cardiac damage as well as renal failure.\[[@CIT3]\] In the present study, concentrations of the total cholesterol, LDL and triglycerides were significantly increased in toxic control with a significant decrease in HDL concentration as compared to the basal diet control group. All these findings can be further substantiated from the histopathology and biomarker studies of liver and kidney. Treated groups showed significant decrease in total cholesterol, LDL and triglycerides, with a significant increase in the HDL concentration. The beneficial effects of herbs in test could be attributed to their lipid lowering property. This effect of ginseng may be attributed to the ginsenosides and other active principles. *A. sativum* decreases hepatic 3-methylglutaryl-COA reductase, cholesterol-7-hydroxylase and fatty acid synthetase, thus lowering the serum lipid profile.\[[@CIT4]\] Spirulina significantly increases lipoprotein lipase activity and thus reduces the serum cholesterol.\[[@CIT5]\] *M. koenigii* has tertiary and quaternary alkaloids, flavonoids and glycoside components that reduce lipid levels in animals.\[[@CIT6]\] *G. sylvestre* increases lecithin-cholesterol acyl transferase activity and thus decreases serum cholesterol.\[[@CIT7]\] The concentrations of total protein, albumin and globulins (g/dl) recorded at the end of 4^th^ week in the basal diet control (group 1) were 3.50±0.140, 1.92±0.041 and 1.58±0.168, respectively, which were significantly (*P*\<;0.05) decreased in the CPS toxic controls (groups 2-15) with the values ranging from 1.19±0.084 to 1.28±0.045, from 0.49±0.007 to 0.51±0.008 and from 0.67±0.077 to 0.78±0.051, respectively. However, following supplementation with herbs and herbal combinations in test, there was a significant (*P*\<;0.05) increase in the total protein concentration in groups 3-15 at the end of 6^th^ week with the values ranging from 4.16±0.011 to 4.70±0.026, from 2.28±0.022 to 2.79±0.014 and from 1.83±0.048 to 1.96±0.043, respectively, as compared to their corresponding 4^th^ week values and toxic control group 2 (1.94±0.015, 0.96±0.019 and 0.98±0.024, respectively). Amongst all the groups, the protein profile remained significantly (*P*\<;0.05) higher in the basal diet control (4.95±0.030, 3.00±0.018 and 1.94±0.046, respectively). The A/G ratio recorded at the end of the 4^th^ week in the basal diet control (group1) was 1.28±0.145, which was significantly (*P*\<;0.05) decreased in CPS toxic controls (groups 2--15) with the values ranging from 0.64±0.041 to 0.79±0.065. Following supplementation with herbs in groups 3--15, there was a significant (*P*\<;0.05) increase at the end of 6^th^ week with the values ranging from 1.20±0.026 to 1.52±0.055 as compared to their corresponding 4^th^ week values and CPS toxic control group 2 (0.98±0.042). The protein profile was studied by assessing total protein, albumin, globulin and A/G ratio in serum of birds. They serve as indicators for hepatic damage. The study revealed a significant decrease in the concentrations of total protein, albumin and globulin and the A/G ratio in CPS toxic control group, which could be attributed to the reduced capacity of the liver to synthesize them. This could be due to the peroxidative damage of liver, which is the exclusive site of protein synthesis.\[[@CIT3]\] The altered protein profile may be due to the hepatic damage caused by oxidative stress and this is substantiated from the histological sections of liver in CPS toxic control group, which showed marked central vein congestion, hydropic degeneration, mild bile duct hyperplasia and dilated sinusoidal spaces with congestion in the sinusoidal spaces, whereas the groups supplemented with herbs showed regenerative changes in liver and improvement in protein profile, suggesting the therapeutic potential of the herbs in test. In conclusion, the results of the present investigation show that CPS exerted alterations in lipid and protein profile along with histological alterations in liver. Use of herbs and herbal combinations in test could counter the adverse effects of CPS pre-induced toxicity to a major extent, suggesting their beneficial effects. **Source of Support:** Nil **Conflict of Interest:** None declared

PubMed Central

2024-06-05T04:04:19.388308

2011-01-01

PMC3052584

INTRODUCTION {#sec1-1} ============ Cadmium is a hazardous heavy metal that is widely distributed in the environment, and it is present in trace levels in sea water and in a broad range of animal and plant species.\[[@CIT1]\] Many of the poultry industries are located near highways and in the proximity of industries that liberate Cd as pollutant. So, it is a common finding that poultry industry is affected with Cd toxicity. Cd is a potent inducer of apoptosis, which is mediated via induction of oxidative stress.\[[@CIT2]\] Cadmium is also known to cause stress by increasing lipid peroxidation or by changing intracellular glutathione (GSH) levels.\[[@CIT3]\] Keeping these toxic effects in consideration, an experimental study was conducted to evaluate its effects on certain organ biomarkers and to evaluate the benefit of using *Emblica officinalis*, vitamin E and polyherbal formulation. MATERIALS AND METHODS {#sec1-2} ===================== A total of 80 day-old sexed male broiler chicks of *Cobb* strain obtained from Venkateswara Hatcheries (Hyderabad) were randomly divided into eight groups consisting of 10 chicks in each. Group 1 was maintained on basal diet for 6 wk, group 2 on cadmium (100 ppm in feed) for 6 wk, group 3 on a combination of Cd + **E. officinalis** (500 ppm in feed) for 6 wk, group 4 on a combination of Cd + vitamin E (300 ppm in feed) for 6 wk, group 5 on a combination of Cd + polyherbal formulation (1 g/kg feed) (stressroak; Ayurvet Ltd., Himachal Pradesh), consisting of *Withania somnifera, Ocimum sanctum*, Shilajit, Amla, *Mangifera indica*) for 6 wk. Groups 6, 7 and 8 were fed with Cd for the first 4 wk followed by E. *officinalis*, vitamin E and polyherbal formulation, respectively, during the subsequent 2 wk. Sera samples were separated for the estimation of renal and hepatic biomarkers by using commercially available diagnostic kits (Qualigens Pvt. Ltd., Mumbai, India). The data were analyzed by one-way analysis of variance (ANOVA) using statistical package for social sciences (SPSS), version 10. *P*\<0.05 was considered as significant. RESULTS AND DISCUSSION {#sec1-3} ====================== The activity of alanine transaminase (ALT; IU/l) in group 1 was 19.346 ± 0.131, which was significantly (*P*\<0.05) increased in Cd toxic control groups 2, 6, 7 and 8 to 60.279 ± 0.166, 60.115 ± 0.022, 59.728 ± 0.208 and 60.030 ± 0.080, respectively, at the end of 4^th^ wk. In groups, 6, 7 and 8 following treatment, the activity was significantly (*P*\</0.05) reduced to 41.743 ± 0.331, 42.075 ± 0.056 and 40.023 ± 0.088, respectively, as compared to their respective 4^th^ wk values and that of group 2 (68.071 ± 0.012) at the end of 6^th^ wk. The activity of alkaline phosphatase (ALP; units/ml) in group 1 was 70.785 ± 0.620, which was significantly (*P*\<0.05) increased in the Cd toxic control groups 2, 6, 7 and 8 to 80.211 ± 1.149, 78.674 ± 0.515, 78.206 ± 0.337 and 78.161 ± 0.314, respectively, at the end of 4^th^ wk. In groups 6, 7 and 8 following treatment, the ALP activity was significantly (*P*\<0.05) decreased (72.034 ± 0.065, 70.061 ± 0.198 and 70.04 ± 0.012, respectively) when compared to their respective 4^th^ wk values and that of group 2 (82.693 ±0.216) at the end of 6^th^ wk. The activities of ALT and ALP are elevated following hepatocellular injury.\[[@CIT4]\] Demerdesh *et al*.\[[@CIT5]\] reported an increase in the activity of ALT and ALP in plasma, when rats were fed with Cd at 5 mg/kg. Treatment with *E. officinalis*, vitamin E and stressroak following discontinuation of cadmium resulted in a significant reduction in the activity of ALT. The hepatocellular injury due to cadmium could be attributed to the cadmium-induced generation of free radicals and the reversal of the findings following treatment could be attributed to the antioxidant and the hepatoprotective potential of the drugs in test. Alcoholic and aqueous extracts of the fruits of *Emblica* have shown hepatoprotective properties in experiments in rats; it inhibited hepatic lipid peroxidation and the increase of serum levels of ALT, aspartate transaminase (AST) and lactate dehydrogenase (LDH). These results support the use of *Emblica* fruit for hepatoprotection.\[[@CIT6]\] However, simultaneous supplementation of drugs in the test along with cadmium revealed a significant increase in ALT and ALP activities at different time intervals when compared to the plain basal diet control, but the values were significantly lower in comparison to pure cadmium toxic control group 2. This finding suggests the prophylactic potential of these drugs to prevent cadmium-induced toxic manifestations, though there was no complete prevention of changes. The renal indices such as serum urea and creatinine were significantly (*P*\<0.05) increased in toxic controls 2, 6, 7 and 8 at the end of 4^th^ wk when compared to group 1 \[[Table 1](#T0001){ref-type="table"}\]. This could be due to the oxidative damage of cadmium on renal tissue. However, in groups 6, 7 and 8 following treatment with *E. officinalis*, vitamin E and polyherbal formulation, respectively, the renal profile revived to normal, at the end of 6^th^ wk, suggesting the involvement of oxidative damage prior to treatment. Non-protein nitrogenous (NPN) substances such as serum urea and creatinine are increased only when renal function is below 30% of its original capacity in birds. Plasma urea appears to be the single most useful variable for detection of pre-renal causes of renal failure.\[[@CIT4]\] In the present study, the serum urea and creatinine levels were significantly increased in toxic controls at the end of 4^th^ wk as compared to the remaining groups. The nephrotoxic metal cadmium at micromolar concentrations induces apoptosis of rat kidney proximal tubule (PT) cells within 3-6 hours of exposure. This involves a complex and sensitive interplay of signaling cascades involving mitochondrial pro-apoptotic factors, calpains and caspases, whose activation is determined by cadmium concentration and the duration of cadmium exposure.\[[@CIT7]\] Cadmium-induced apoptosis in rat kidney epithelial cells involves decrease in nuclear factor-kB (NF-kB) activity.\[[@CIT8]\] Treatment in groups 6, 7 and 8 with *E. officinalis*, vitamin E and stressroak, respectively, following discontinuation of cadmium, resulted in significant decrease in serum creatinine and blood urea nitrogen (BUN) as compared to Cd toxic control group 2, which suggests their protective role. However, simultaneous supplementation of drugs in test along with cadmium revealed a significant increase in serum creatinine and BUN at different time intervals when compared to basal diet control, though the values were significantly lower in comparison to pure cadmium toxic control group 2. This finding suggests the prophylactic potential of these drugs to prevent cadmium-induced toxic manifestations, though there was no complete prevention of the changes. The beneficial renal protective actions of drugs in test may be attributed to their antioxidant/free radical scavenging actions and protection of protein thiols from deleterious actions of cadmium in kidney. In conclusion, the study revealed that cadmium has the potential to induce hepatotoxicity and nephrotoxicity and supplementation with **E. officinalis**, vitamin E and polyherbal formulation (stressroak) has a beneficial role in preventing the adverse effects. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Profile of renal biomarkers ::: Group Creatinine (mg/dl) Urea (mg/dl) ----------------------------------------------------------- -------------------- ------------------ ------------------ ------------------ Basal diet (1-42 days) 0.343±0.018^aA^ 0.424±0.046^aB^ 16.752±0.035^aA^ 17.208±0.149^aA^ Cadmium (1-42 days) 2.854±0.021^cA^ 3.235±0.012^eB^ 43.589±0.194^dA^ 49.287±0.245^eB^ Cadmium+ **Emblica* officinalis* E (1-42 days) 2.277±0.031^bA^ 2.232±0.073^bA^ 33.332±0.341^cA^ 33.047±0.168^dA^ Cadmium + vitamin E (1-42 days) 2.239±0.312^bA^ 2.216±0.011^bA^ 32.193±0.245^cA^ 30.199±0.168^cB^ Cadmium + stressroak (1-42 days) 2.276±0.211^bA^ 2.214±0.011^bA^ 26.210±0.266^bA^ 25.925±0.381^bA^ Cadmium (1-28 days); **Emblica* officinalis* (29-42 days) 2.856±0.231^cA^ 2.443±0.064^dB^ 43.304±0.168^dA^ 31.338±0.266^cB^ Cadmium (1-28 days); vitamin E (29-42 days) 2.823±0.181^cA^ 2.365±0.031^cB^ 43.019±0.133^dA^ 31.345±0.206^cB^ Cadmium (1-28 days); stressroak (29-42 days) 2.876±0.041^cA^ 2.395±0.051^cdB^ 42.450±0.320^dA^ 30.758±0.207^cB^ Values are mean + SE of 6 observations, Means with different alphabets as superscripts differ significantly (*P*\<0.05) ANOVA. Capital alphabets (Horizontal comparison) Small Alphabets (Vertical comparison) ::: **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.389216

2011-01-01

PMC3052585

INTRODUCTION {#sec1-1} ============ Synthetic pyrethroid insecticides are now being substituted for pest control and increased production.\[[@CIT1]\] They are most widely used group of pesticides due to their low toxicity to mammals, birds and insects.\[[@CIT2]\] Cypermethrin, a composite pyrethroid, is a broad-spectrum insecticide and fast acting neurotoxin with good contact and stomach action.\[[@CIT3]\] Consistent with its lipophilic nature, it has been found to accumulate in body fat, skin, liver, kidneys, adrenal glands, ovaries and brain.\[[@CIT3]\] Cypermethrin and dichlorodiphenyltrichloroethane (DDT) have been detected in human milk from malarial endemic areas in South Africa.\[[@CIT4]\] There have been extensive histomorphological, toxicological and biochemical reports on cypermethrin toxicity in different species of animals.\[[@CIT5][@CIT6]\] However, no information is available about the effect of cypermethrin in female rats. So, the present study was conducted to elucidate the toxicity of cypermethrin at low dose levels in female albino rats. MATERIALS AND METHODS {#sec1-2} ===================== Sexually mature female albino rats of 3 months age, weighing 110-150 g, were obtained from Guru Angad Dev Veterinary and Animal Sciences University (GADVASU), Ludhiana. The animals were housed in groups of two rats per cage. The rats were acclimatized for 10 days before using them for experimentation. Their ovarian cyclicity was checked by daily microscopic examination of vaginal smears. Rats showing a 4-day estrous cycle were selected for the present investigation. Commercial formulation of cypermethrin (Rallis India Limited, Secunderabad, Hyderabad) having 25% Emulsifiable Concentrate (EC) was used for the present study. Adequate dilutions were made with vegetable oil to achieve the test concentration of 50 mg/kg. The test concentration of cypermethrin was calculated from the percentage of active ingredient of commercial formulation of cypermethrin. Rats were intubated orally with cypermethrin at 50 mg/kg b. wt. daily for 2 and 4 weeks. Group of rats receiving similar amount of vegetable oil and distilled water served as control rats. All the animals (control and cypermethrin treated) were observed daily for clinical symptoms like salivation, activity, irritability, fecal pellet conditions, diarrhea, eyeball movement, weakness, coarse tremor, paralysis of limb, wounds and mortality, etc. Feed and water intake was also noted for control and cypermethrin-treated rats. Cyclicity was checked daily for the control and treated rats throughout the experimentation period. The body weight of the rats was taken before the start of the treatment in control as well as in cypermethrin treatment groups. During the duration of experiment, the rats were weighed weekly to determine the change in body weight. On completion of the experiment, the rats were sacrificed and various organs like liver, spleen, heart, lungs, kidney and endocrine glands were excised and weighed. RESULTS AND DISCUSSION {#sec1-3} ====================== General toxicity symptoms {#sec2-1} ------------------------- In the first and second weeks of treatment, no morphological or toxicological effect of pesticide was observed. However, the chronic exposure of cypermethrin for 4 weeks resulted in loose fecal pellets and hyperirritability in treated rats. Various studies have investigated the toxic effects of cypermethrin in mammals, revealing increase in salivation, lack of coordination, muscle tremor and convulsions.\[[@CIT7]\] These signs of toxicity indicate that the target for this compound is the central nervous system in mammals.\[[@CIT7][@CIT8]\] It was observed that there was no effect on feed and water intake in treated rats as compared to control groups \[[Table 1](#T0001){ref-type="table"}\]. Rats exposed to liquid mosquito repellent (LMR) containing allethrin (3.6% w/w) showed no significant effect on food consumption.\[[@CIT2]\] Lambda cyhalothrin also had no significant effect on water intake.\[[@CIT9]\] However, cypermethrin when given to white rabbits showed reduced feed and water intake in treated animals as compared to control groups.\[[@CIT10]\] The estrous cycle was slightly disturbed in all the treated rats. Persistent estrous was observed in some of treated rats at third week of treatment as compared to control groups. Mortality also occurred on 17^th^ and 24^th^ days of treatment in two rats treated with cypermethrin, while all the control rats remained healthy throughout the experimentation. Treatment-related deaths (two with low dose and five with high dose) have also been observed in lambda cyhalothrin treated rats.\[[@CIT9]\] Table 1Effect of cypermethrin on feed and water intake in control and treated female albino rats after 2 and 4 weeks of treatmentComponentsAfter 2 weeksAfter 4 weeksControl IControl IITreatedControl IControl IITreatedFeed intake (g/100 g b. wt.)9.25±0.088.23±0.108.60±0.019.72±0.078.43±0.018.50±0.02Water intake (ml/100 g b. wt.)26.50±0.1625.60±0.1125.10±0.1826.90±0.2025.50±0.2126.00±0.31[^1] Body weight {#sec2-2} ----------- The net body weight gain in all the treated rats was less at both 2 and 4 weeks as compared to the control rats \[[Table 2](#T0002){ref-type="table"}\]. In accordance with the present observations, Hussain *et al*,\[[@CIT6]\] have also observed significantly lower body weight gain in cypermethrin-treated rats as compared to control rats. Lakkawar *et al*.\[[@CIT10]\] observed decrease in body weights in rabbits treated with cypermethrin. Lambda cyhalothrin administered orally to rats has also resulted in reduced body weight gain in both male and female rats.\[[@CIT9]\] Table 2Effect of cypermethrin treatment on body weight of treated female albino rats as compared to control rats after 2 and 4 weeks of treatmentTreatmentAfter 2 weeksAfter 4 weeksInitial weight (g/100 g b. wt.)Final weight (g/100 g b. wt.)Net body wt. gain (g/100 g b. wt.)Initial weight (g/100 g b. wt.)Final weight (g/100 g b. wt.)Net body wt. gain (g/100 g b. wt.)Control I116.60±13.60140.0±14.7020.50±0.80116.60±13.60160.00±6.2337.20±0.89Control II125.00±0150.00±2.3520.00±0.90125.00±0167.50±1.7634.00±0.74Treated125.00±0145.00±2.3516.00±0.20[a](#T000F1){ref-type="table-fn"}150.00±20.40186.66±27.9024.40±0.32[a](#T000F1){ref-type="table-fn"}[b](#T000F2){ref-type="table-fn"}[^2][^3] Vital organ weight {#sec2-3} ------------------ Weights of all vital organs, viz., liver, spleen, heart, lungs and kidneys were determined on the day of sacrifice and expressed as g/100 g body weight. No significant changes in the weights of vital organs like heart and lungs were observed as compared to their relative control rats after 2 and 4 weeks of treatment. Nagarjuna and Doss\[[@CIT11]\] have observed pathological changes in hearts of rats orally exposed to cypermethrin for a long duration. Weight of liver and spleen decreased in treated rats at both 2 and 4 wks as compared to the control groups \[[Table 3](#T0003){ref-type="table"}\], while a significant increase (*P* ≤ 0.05) in the weight of kidneys was observed in cypermethrin-treated rats as compared to control and vehicle treated rats after both 2 and 4 wks of treatment \[[Table 3](#T0003){ref-type="table"}\]. Liver relative weight also varied nonsignificantly in early stages, but in later stages of the experiment, it significantly decreased in cypermethrin fed rats.\[[@CIT6]\] Yavasoglu *et al*.\[[@CIT12]\] have observed no significant change in relative liver weight of cypermethrin-treated rats when compared with control animals at lower dose but at high dose significant decrease in liver weight of treated rats was observed. Some workers have observed dose-dependent increase in liver weight of cypermethrin-treated rats at some acute and subacute dose levels.\[[@CIT2]\] Adverse changes in liver tissues in test animals have also been observed in cypermethrin-treated male rats as compared to control rats.\[[@CIT13]\] ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Effect of cypermethrin treatment at 50 mg/kg b. wt. on weights (g/100 g b. wt.) of different organs in control and treated rats after 2 and 4 weeks of treatment ::: Name of tissue/organ After 2 weeks After 4 weeks ---------------------- --------------- --------------- -------------------------------------------- ----------- ----------- -------------------------------------------- Liver 4.61±0.28 4.62±0.66 4.40±0.01[a](#T000F3){ref-type="table-fn"} 4.58±0.20 4.82±0.12 4.43±0.07[a](#T000F3){ref-type="table-fn"} Spleen 0.24±0.07 0.29±0.04 0.22±0.02[a](#T000F3){ref-type="table-fn"} 0.26±0.02 0.27±0.07 0.23±0.01[a](#T000F3){ref-type="table-fn"} Heart 0.34±0.03 0.40±0.02 0.36±0.03 0.33±0.04 0.36±0.02 0.35±0.06 Lungs 0.34±0.02 0.37±0.05 0.35±0.03 0.32±0.02 0.34±0.01 0.33±0.03 Kidney 0.32±0.01 0.36±0.01 0.39±0.04[a](#T000F3){ref-type="table-fn"} 0.33±0.08 0.35±0.01 0.40±0.01[a](#T000F3){ref-type="table-fn"} a Statistically significantly different (*P* ≤ 0.05) between groups; All the values are mean±SE values of six animals in each group ::: In the present study, weight of spleen was decreased in the cypermethrin-treated rats after 2 and then further after 4 wks of treatment \[[Table 3](#T0003){ref-type="table"}\]. The decrease in spleen weight may be indicative of depression of immune status of animals and might reflect a depression of cellular response. Cypermethrin caused increase in kidney weight of treated rats after both 2 and 4 weeks of treatment. Rats exposed to LMR containing allethrin (3.6% w/w) have also shown a significant increase in the weight of kidneys.\[[@CIT2]\] In male rats, long-term feeding studies with cypermethrin have shown increased kidney weights.\[[@CIT13]\] Cypermethrin resulted in alteration in the distribution pattern of oxidoreductase in the kidneys of rats. The general loss of oxidoreductases in kidneys of rabbits because of cypermethrin toxicity may be suggestive of decreased metabolism of physiological processes due to degenerative changes in various segments of nephron resulting in concomitant increase of enzymes in the extracellular fluid.\[[@CIT14]\] Endocrine gland weight {#sec2-4} ---------------------- Endocrine glands such as adrenal, thyroid and parathyroids were weighed after 2 and 4 weeks of treatment in control and treated rats. Significant increase (*P* ≤ 0.05) in weights of adrenal and thyroid was observed in treated rats at both 2 and 4 weeks as compared to the control rats, and the weight of parathyroid gland decreased significantly in the treated rats at both 2 and 4 weeks as compared to the control rats \[Table 4\]. Limited data are available regarding endocrine effects in animals following oral exposure to pyrethroids. Significant increase in adrenal weight has been observed in cypermethrin and cybil treated male rats.\[[@CIT15]\] Increased adrenal weight might reflect a state of physiological stress in the body of rats. Glucocorticoids released from adrenals might play an important role in mediation of depressed immune response. Rats exposed to LMR containing allethrin (3.6% w/w) also had a significant increase in relative weights of adrenal glands in the male rats.\[[@CIT2]\] In the present study, the weight of the thyroid gland increased in the treated rats at 2 and 4 weeks of treatment. Rats treated with fenvalerate caused a nonsignificant increase in T~3~ and T~4~ concentration.\[[@CIT16]\] Similarly, increase in thyroid weight has also been observed in rats treated with LMR containing allethrin.\[[@CIT2]\] It may be concluded that the changes that were observed in body weight and various organ weights in rats treated with cypermethrin may be indicative of intensity of cellular and tissue damage. **Source of Support:** Nil **Conflict of Interest:** None declared. [^1]: All the values are mean±SE values of six animals in each group [^2]: a Statistically significantly different (*P* ≤ 0.05) between groups; [^3]: b statistically significantly different (*P* ≤ 0.05) between two different time periods; All the values are mean±SE values of six animals in each groups

PubMed Central

2024-06-05T04:04:19.390045

2011-01-01

PMC3052586

INTRODUCTION {#sec1-1} ============ Free radicals are continuously produced by the body and are also generated through environmental pollution, radiation, drugs, chemicals, pesticides, etc.\[[@CIT1]\] Iron is the most common cofactor within the oxygen handling biological machinery\[[@CIT2]\] and interacts with molecular oxygen and generates reactive oxygen species (ROS) through Haber-Weiss and Fenton reactions,\[[@CIT3]\] leading to oxidative stress.\[[@CIT4]\] High tissue iron concentrations have been associated with the development and progression of several pathological conditions including certain cancers, liver and heart diseases, diabetes, hormonal abnormalities and immune system dysfunctions.\[[@CIT5]\] Lipid peroxidation and subsequent oxidative stress could be mitigated by antioxidants. Endogenous and passively acquired exogenous antioxidant defense systems do not accelerate in maturation until late in the third trimester.\[[@CIT6]\] Therefore, provision of proper neonatal nourishment is essential to make the organism to resist the adversities in further life. Keeping the above facts in view, an experimental study was planned to evaluate the potential of feeding herbal Neonatal Chick care (NNCC) immediately after hatch in amelioration of iron-induced oxidative stress and injury to the biological system. MATERIALS AND METHODS {#sec1-2} ===================== A total of 130 day-old sexed male broiler chicks (*Vencobb* strain) were randomly divided into six groups consisting of 25 chicks each in groups 1--4 and 15 each in groups 5 and 6, to evaluate the efficacy of herbal NNCC (Ayurvet Ltd., Himachal Pradesh) in improving immune response, overall growth and development of digestive organs. All the birds were provided with feed and water *ad libitum* throughout the experiment and were maintained as per the following treatment schedule for 6 weeks: Group 1: Basal diet controlGroup 2: Herbal NNCC 6 g paste/chick/day immediately after hatch for 2 days and normal feed from 3^rd^ day onward (there was no left over paste)Group 3: Herbal NNCC 8 g paste/chick/day immediately after hatch for 2 days and normal feed from 3^rd^ day onward (there was no left over paste)Group 4: Basal diet + stressor (ferrous sulfate) at 0.5% in feedGroup 5: Basal diet + stressor (ferrous sulfate) + NNCC as per groups 2 and 4Group 6: Basal diet + stressor (ferrous sulfate) + NNCC as per groups 3 and 4 Herbal NNCC is a polyherbal formulation containing *Terminalia chebula, Terminalia bellerica, Vitis vinifera, Aegle marmelos, Spirulina and Phyllanthus emblica.* Superoxide dismutase (SOD)\[[@CIT7]\] and catalase\[[@CIT8]\] were estimated on 28^th^ and 42^nd^ days in RBC. Thiobarbituric acid reactive substances (TBARS)\[[@CIT9]\] and reduced glutathione (GSH)\[[@CIT10]\] were estimated in liver, kidney and heart, and protein carbonyls in liver,\[[@CIT11]\] and the immune status was assessed by performing HI and phytohemagglutinin (PHA) tests at the end of 6^th^ wk. Serum samples (*n* = 8) were separated for estimation of alanine transaminase (ALT), glucose and calcium (Span Diagnostics Ltd., Surat, India) by using diagnostic kits. The data were analyzed by one-way analysis of variance (ANOVA) using statistical package for social sciences (SPSS), version 15. *P*\<0.05 was considered as significant. RESULTS AND DISCUSSION {#sec1-3} ====================== The TBARS concentrations \[μM malondialdehyde (MDA)/mg protein\] of liver, heart and kidney were significantly (*P*\<0.05) increased in group 4 (0.58±0.03, 16.48±0.27, and 20.11±0.75, respectively) at the end of 6^th^ week. The TBARS concentrations of liver, heart and kidney in groups 5 and 6 showed a significant (*P*\<0.05) decrease as compared to group 4 \[[Table 1](#T0001){ref-type="table"}\]. The GSH concentrations (μM/mg protein) in liver, heart and kidney were significantly (*P*\<0.05) decreased in group 4 (0.21±0.03, 0.17±0.01 and 2.24±0.16, respectively), while groups 5 and 6 showed a significant (*P*\<0.05) increase in GSH concentration as compared to group 4 \[[Table 1](#T0001){ref-type="table"}\]. The activity of erythrocytic SOD (U/mg protein) and catalase (μg H~2~ O~2~ decomposed/mg protein/min) was significantly (*P*\<0.05) decreased in group 4 (10.37±0.71 and 1.29±0.18, respectively), while groups 5 and 6 showed a significant (*P*\<0.05) increase in SOD and catalase activity as compared to group 4 \[[Table 2](#T0002){ref-type="table"}\]. The concentration of protein carbonyls (nM/100 mg protein) in liver was significantly (*P*\<0.05) increased in group 4 (19.28±0.73) at the end of 6^th^ week, while groups 5 and 6 showed a significant (*P*\<0.05) decrease in protein carbonyls concentration as compared to group 4 \[[Table 3](#T0003){ref-type="table"}\]. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Results of oxidative stress and antioxidant defenses ::: Group Liver Heart Kidney ------------------------------------------------------- ---------------- --------------- --------------- TBARS (µM of MDA/mg protein)  1. Basal diet 0.12±0.02^a^ 5.17±0.551^a^ 8.52±0.43^a^  2. Herbal neonatal chick care 0.16±0.01^ab^ 5.76±0.36^a^ 9.12±0.54^a^  3. Herbal neonatal chick care 0.23±0.01^b^ 5.47±0.63^a^ 9.00±0.73^a^  4. FeSO~4~ + basal diet 0.58±0.03^d^ 16.48±0.27^c^ 20.11±0.75^c^  5. Herbal neonatal chick care + FeSO~4~ + basal diet 0.41±0.05^c^ 7.98±0.47^b^ 11.53±0.85^b^  6. Herbal Neonatal Chick care + FeSO~4~ + basal diet 0.45±0.06^c^ 7.92±0.43^b^ 13.06±0.67^b^ GSH (µM/mg protein)  1. Basal diet 0.66±0.08^d^ 0.93±0.08^d^ 9.38±0.6^d^  2. Herbal neonatal chick care 0.51±0.02^c^ 0.51±0.04^c^ 6.12±0.59^c^  3. Herbal neonatal chick care 0.52±0.02^c^ 0.47±0.04^c^ 5.95±0.59^c^  4. FeSO~4~ + basal diet 0.21±0.03^a^ 0.17±0.01^a^ 2.24±0.16^a^  5. Herbal neonatal chick care + FeSO~4~ + basal diet 0.26±0.02^ab^ 0.35±0.02^b^ 3.68±0.41^ab^  6. Herbal neonatal chick care + FeSO~4~ + basal diet 0.34±0.04^b^ 0.30±0.01^b^ 4.63±0.61^bc^ Protein carbonyls (nM/100mg protein)  1. Basal diet 5.68±0.84^a^ -- --  2. Herbal neonatal chick care 12.17±0.87^bc^ -- --  3. Herbal neonatal chick care 10.06±0.72^b^ -- --  4. FeSO~4~ + basal diet 19.28±0.73^d^ -- --  5. Herbal neonatal chick care + FeSO~4~ + basal diet 13.54±1.35^c^ -- --  6. Herbal neonatal chick care + FeSO~4~ + basal diet 12.05±0.91^bc^ -- -- Values are Mean±SE (*n* = 10); one-way ANOVA (SPSS). Means with different alphabets as superscripts differ significantly (*P*\<0.05) ::: ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Activity of SOD, catalase and ALT, and concentration of glucose and calcium ::: Group 4^th^ wk 6^th^ wk ---------------------------------------------------------- ----------------- ----------------- SOD (U/mg protein)  1. Basal diet 30.75±3.18^cA^ 33.71±0.72^dA^  2. Herbal neonatal chick care (6 g/chick/day) 19.03±2.22^bA^ 22.38±0.41^cA^  3. Herbal neonatal chick care (8 g/chick/day) 20.69±1.14^bA^ 21.80±0.42^cA^  4. FeSO~4~ control 6.92±1.74^aA^ 10.37±0.71^aA^  5. Herbal neonatal chick care (6 g/chick/day) + FeSO~4~ 14.18±2.36^bA^ 17.07±1.11^bA^  6. Herbal neonatal chick care (8 g/chick/day) + FeSO~4~ 16.67±3.09^bA^ 18.53±1.51^Ba^ Catalase (µg H~2~O~2~ decomposed/mg protein/min)  1. Basal diet 8.29±0.41^eA^ 9.57±0.3^dB^  2. Herbal neonatal chick care (6 g/chick/day) 6.68±0.13^dA^ 7.08±0.17^cA^  3. Herbal neonatal chick care (8 g/chick/day) 5.91±0.12^cdA^ 6.69±0.36^cA^  4. FeSO~4~ control 0.98±0.31^aA^ 1.29±0.18^aA^  5. Herbal neonatal chick care (6 g/chick/day) + FeSO~4~ 5.11±0.72^cA^ 4.89±0.43^bA^  6. Herbal neonatal chick care (8 g/chick/day) + FeSO~4~ 3.98±0.25^bA^ 5.08±0.27^bB^ ALT (U/l)  1. Basal diet 6.98±3.01^aA^ 8.08±7.12^aA^  2. Herbal neonatal chick care (6 g/chick/day) 9.89±1.47^aA^ 11.35±6.78^aA^  3. Herbal neonatal chick care (8 g/chick/day) 11.20±1.91^aA^ 14.56±5.23^aA^  4. FeSO~4~ control 56.38±5.40^CA^ 69.67±2.34^cA^  5. Herbal neonatal chick care (6 g/chick/day) + FeSO~4~ 36.12±6.72^bA^ 44.34±7.64^bA^  6. Herbal neonatal chick care (8 g/chick/day) + FeSO~4~ 34.14±3.56^bA^ 36.37±9.00^bA^ Glucose (mg/dl)  1. Basal diet 35.34±1.28^aA^ 36.75±1.78^aA^  2. Herbal neonatal chick care (6 g/chick/day) 41.30±1.47^abA^ 41.68±1.82^abA^  3. Herbal neonatal chick care (8 g/chick/day) 45.00±1.56^bA^ 52.80±2.06^cA^  4. FeSO~4~ control 65.88±3.8^dA^ 65.58±2.83^dA^  5. Herbal neonatal chick care (6 g/chick/day) + FeSO~4~ 45.69±2.27^bA^ 53.40±4.06^cA^  6. Herbal neonatal chick care (8 g/chick/day) + FeSO~4~ 52.22±1.89^cA^ 54.85±3.27^bcA^  Calcium (mg/dl)  1. Basal diet 2.55±0.27^aA^ 6.01±0.78^aB^  2. Herbal neonatal chick care (6 g/chick/day) 2.96±0.23^aA^ 7.01±0.44^aB^  3. Herbal neonatal chick care (8 g/chick/day) 3.59±0.17^aA^ 6.93±0.51^aB^  4. FeSO~4~ control 3.93±0.52^bA^ 9.06±0.31^aB^  5. Herbal neonatal chick care (6 g/chick/day) + FeSO~4~ 3.96±0.44^bA^ 9.43±0.16^aB^  6. Herbal neonatal chick care (8 g/chick/day) + FeSO~4~ 4.79±0.41^bA^ 8.93±0.24^aB^ Values are Mean±SE (*n* = 8); one-way ANOVA (SPSS). Means with different alphabets as superscripts differ significantly (*P*\<0.05). Capital alphabets (horizontal comparison); small alphabets (vertical comparison); SOP, superoxide dismutase; ALT, alanine transaminase ::: ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Results of immunological parameters ::: Group 6^th^ wk ---------------------------------------------------------- -------------- HI (Log^2^) titer  1. Basal diet 5.67±0.33^c^  2. Herbal neonatal hick care (6 g/chick/day) 4.67±0.61^c^  3. Herbal neonatal chick care (8 g/chick/day) 4.86±0.34^c^  4. FeSO~4~ control 2.50±0.22^a^  5. Herbal neonatal chick care (6 g/chick/day) + FeSO~4~ 4.17±0.60^b^  6. Herbal neonatal chick care (8 g/chick/day) + FeSO~4~ 4.23±0.75^b^ Phytohemagglutinin (PHA) assay (mm)  1. Basal diet 0.85±0.04^d^  2. Herbal neonatal chick care (6 g/chick/day) 0.68±0.03^c^  3. Herbal neonatal chick care (8 g/chick/day) 0.73±0.06^c^  4. FeSO~4~ control 0.22±0.06^a^  5. Herbal neonatal chick care (6 g/chick/day) + FeSO~4~ 0.52±0.03^b^  6. Herbal neonatal chick care (8 g/chick/day) + FeSO~4~ 0.48±0.03^b^ Values are Mean±SE (*n* = 8); one-way ANOVA (SPSS). Means with different alphabets as superscripts differ significantly (*P*\<0.05) ::: The TBARS and protein carbonyls were elevated and the antioxidant defenses like GSH, SOD and catalase were decreased in the toxic control group 4 that was maintained on iron throughout the study. All the changes in the antioxidant defense profile were significantly reversed in the groups (5 and 6) that were treated with herbal NNCC. These results may be attributed to the antioxidant properties aqueous extract of *T. chebula\>*,\[[@CIT12]\] *T. bellerica*,\[[@CIT13]\] grape seed proanthocyanidines,\[[@CIT14]\] *A. marmelos*,\[[@CIT15]\] *Spirulina plantensis*,\[[@CIT16]\] etc. Therefore, the polyherbal formulation, NNCC, showed potent antioxidant actions in this study. The activity of ALT (U/l) and the concentration of glucose (mg/dl) was significantly (*P*\<0.05) increased on 4^th^ and 6^th^ weeks in group 4 as compared to group 1 \[[Table 2](#T0002){ref-type="table"}\], while groups 5 and 6 revealed significant (*P*\<0.05) decrease as compared to group 4. The increased activity of ALT and concentration of glucose in the iron toxic control group 4 suggests the hapatocellular insult following administration of iron. Treatment with herbal NNCC resulted in significant reduction in the activity of ALT. The hepatocellular injury due to iron could be attributed to the iron-induced generation of ROS or free radicals and the reversal of the findings following treatment could be attributed to the antioxidant and the hepatoprotective potential of the ingredients of herbal NNCC, namely, *A. marmelos*,\[[@CIT17]\] *T. chebula*,\[[@CIT18]\] *P. emblica*,\[[@CIT19]\] etc. The PHA index (mm) and HI titer (Log^2^) were significantly (*P*\<0.05) decreased in the iron control group 4 (0.22±0.06 and 2.50±0.22, respectively) \[[Table 3](#T0003){ref-type="table"}\], while the thickness and HI titer in groups 2, 3, 5 and 6 were significantly (*P*\<0.05) increased as compared to group 4. The beneficial immune competent actions of NNCC may be attributed to its synergistic herbs. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.391719

2011-01-01

PMC3052587

INTRODUCTION {#sec1-1} ============ Iron, by virtue of its ability to participate directly as a donor or acceptor in electron transfer reactions, is an essential trace element for cell function. This property makes iron the most common cofactor within the oxygen handling biological machinery.\[[@CIT1]\] However, the very property that enables iron to participate in oxygen metabolism explains its potential damaging effects: if not handled properly by the cell, iron interacts with molecular oxygen, generating reactive oxygen species (ROS) through Haber-Weiss and Fenton reactions.\[[@CIT2]\] Uncontrolled ROS production leads to oxidative damage of cellular components, a condition termed oxidative stress.\[[@CIT3]\] To prevent peroxidative tissue damage, there are protective mechanisms *in vivo*, such as an enzymatic defense system (antioxidant enzymes) and free-radical scavengers (antioxidants). Iron interferes with the activities of different antioxidants and hence may contribute to progression of liver damage and damage to other organ systems.\[[@CIT4]\] Oxidative damage by free radicals can be prevented by the use of antioxidants such as ascorbic acid, tocopherols, carotenoids and certain herbs.\[[@CIT5]\] Of late, several herbs have been reported to counter the peroxidative stress due to several stressors. Keeping the above facts in view, an experimental study was planned to evaluate the mechanism of iron-induced oxidative stress and injury to the biological system, and to evaluate the therapeutic potential of certain herbs and herbal combinations against experimental iron toxicosis in broilers. MATERIALS AND METHODS {#sec1-2} ===================== A total of 225 sexed male broiler chicks (*Cobb* strain) of a day-old age were randomly divided into 15 groups consisting of 15 chicks in each group. All the birds were provided with feed and water *ad libitum* throughout the experiment. The groups were maintained as per the following treatment schedule for 6 weeks: Group 1: Basal diet control (1--42 days)Group 2: FeSO~4~ (0.5% of feed) toxic control (1--42 days)Group 3: FeSO~4~ (1--28 days) + *Withania somnifera* at 0.1% of feed (29--42 days)Group 4: FeSO~4~ (1--28 days) + *Ocimum sanctum* at 0.1% of feed (29--42 days)Group 5: FeSO~4~ (1--28 days) + *Asparagus racemosus* at 0.1% of feed (29--42 days)Group 6: FeSO~4~ (1--28 days) + *Andrographis paniculata* at 0.1% of feed (29--42 days)Group 7: FeSO~4~ (1--28 days) + *Murraya koenigii* at 0.1% of feed (29--42 days)Group 8: FeSO~4~ (1--28 days) + shilajit at 0.1% of feed (29--42 days)Group 9: FeSO~4~ (1--28 days) + *Gymnema sylvestre* at 0.1% of feed (29--42 days)Group 10: FeSO~4~ (1--28 days) + *Allium sativum* at 0.1% of feed (29--42 days)Group 11: FeSO~4~ (1--28 days) + Spirulina at 0.1% of feed (29--42 days)Group 12: FeSO~4~ (1--28 days) + (*W. somnifera* + *As. racemosus* + *An. paniculata* at 0.05% of each in feed for 29--42 days)Group 13: FeSO~4~ (1--28 days) + (*W. somnifera* + *M. koenigii* +*Al. sativum* at 0.05% of each in feed 29--42 days)Group 14: FeSO~4~ (1--28 days) + (Spirulina + shilajit + *G. sylvestre* at 0.05% of each in feed for 29--42 days)Group 15: FeSO~4~ (1--28 days) + (*An. paniculata* + *O. sanctum* + *As. racemosus* at 0.05% of each in feed 29--42 days) The blood samples were drawn from wing vein at the end of 4^th^ and 6^th^ weeks from the birds in each group for the assay of superoxide dismutase (SOD)\[[@CIT6]\] and catalase.\[[@CIT7]\] Sera samples were separated from the blood for the estimation of alanine transaminase (ALT) and serum creatinine by using diagnostic kits (Qualigens Pvt. Ltd., Mumbai, India). The birds were sacrificed at the end of 6^th^ wk and tissues were collected for the assay of reduced glutathione (GSH)\[[@CIT8]\] and thiobarbituric acid reactive substances (TBARS)\[[@CIT9]\] in liver and kidney homogenates. The data were subjected to statistical analysis by applying one-way analysis of variance (ANOVA) using statistical package for social sciences (SPSS), 10^th^ version. Differences between means were tested using Duncan's multiple comparison test and significance was set at *P*\<0.05. RESULTS AND DISCUSSION {#sec1-3} ====================== The concentrations of TBARS \[nanomoles malondialdehyde (MDA)/g protein\] of liver and kidney in basal diet control (group 1) were 153.42±5.943 and 106.709±3.788, respectively, which was significantly (*P*\<0.05) increased in iron toxic control group 2 (185.256±3.796 and 177.763±3.306, respectively) at the end of 6^th^ wk. Groups 3-15 that were supplemented with herbs and their combinations during the last 2 weeks following discontinuation of iron revealed a significant (*P*\<0.05) decrease in the concentration (from 140.145±3.172 to 154.777±6.680 and from 119.247±5.725 to 142.039±4.62 in liver and kidney, respectively) as compared to group 2. The concentrations of GSH (mg/g protein) in liver and kidney in the basal diet control (group 1) were 70.425±2.471 and 31.573±1.215, respectively, which was significantly (*P*\<0.05) reduced in iron toxic control group 2 (33.237±1.557 and 18.707±0.59, respectively) at the end of 6^th^ wk. Groups 3--15 that were supplemented with herbs and their combinations during the last 2 wk following discontinuation of iron revealed a significant (*P*\<0.05) increase (from 60.683±2.129 to 69.095±2.099 and from 24.740±0.938 to 31.274±0.668 in liver and kidney, respectively). The activities of erythrocytic SOD (IU/g protein) and catalase (μM/min) at the end of 4^th^ wk in the basal diet control (group 1) were 41.158±0.471 and 2.55±0.055, respectively, which were significantly (*P*\<0.05) increased in all the remaining groups with the values ranging from 64.187±0.345 to 66.305±0.69 and from 3.464±0.013 to 3.868±0.084, respectively. However, following supplementation with herbs and their combinations in test, there was a significant (*P*\<0.05) decrease in the activity of SOD in groups 3-15 at the end of 6^th^ wk with the values ranging from 55.288±0.52 to 56.883±0.531 and from 2.852±0.045 to 3.085.±0.092, respectively, as compared to their corresponding 4^th^ wk values and the activity in iron toxic control group 2 (81.852±0.834 and 4.312±0.178, respectively). In this study, the lipid peroxidation biomarkers like TBARS and the antioxidant defenses like SOD and catalase were elevated, while GSH was reduced in the toxic control group that was maintained on iron throughout the study. All the changes in the antioxidant defense profile were significantly reversed when treated with herbs and herbal combinations. These results are attributed to the antioxidant properties of steroidal lactones such as withanolides, which are the main active components of *W. somnifera*.\[[@CIT10]\] *O. sanctum* has been reported to reduce lipid peroxidation and increase GSH concentration in blood.\[[@CIT11]\] *An. paniculata* has been reported to increase the activities of SOD, catalase, glutathione peroxidase and glutathione reductase as well as the concentration of glutathione, with a subsequent decrease in lipid peroxidation. Similar reports were documented for shilajit in rats.\[[@CIT12]\] Spirulina decreased the MDA levels and increased the GSH levels in the goat liver homogenate *in vitro*.\[[@CIT13]\] The activity of ALT (units/l) in basal diet control (group 1) was 18.656±0.583 at the end of 4^th^ wk, which was significantly (*P*\<0.05) increased in all the remaining groups with the values ranging from 43.142±0.865 to 44.891±1.250. However, following supplementation with herbs and their combinations in test, there was a significant (*P*\<0.05) decrease in the activity of ALT in groups 3-15 at the end of 6^th^ wk with the values ranging from 36.729±2.633 to 39.644±1.475 as compared to their corresponding 4^th^ wk values and the activity in iron toxic control group (66.462±1.56). Amongst all the groups, the activity remained significantly (*P*\<0.05) lower in basal diet control (24.778±0.949). The activity of ALT was determined to assess the degree of damage to the liver as the levels of certain enzymes like alanine transaminase (ALT), aspartate transaminase (AST), gamma glutamyl transferase (GG T), etc., are reported to be elevated following hapatocellular injury.\[[@CIT14]\] In this study, the activity of ALT was significantly elevated in the iron toxic control group, suggesting the hapatocellular insult following administration of iron. Treatment with herbs and their combinations following discontinuation of iron resulted in significant reduction in the activity of ALT. The hepatocellular injury due to iron could be attributed to the iron-induced generation of ROS or free radicals and the reversal of the findings following treatment could be attributed to the antioxidant and the hepatoprotective potential of the herbs\[[@CIT15]\] to prevent iron-induced toxic manifestations, though there was no complete prevention of changes. The serum creatinine concentration (mg/dl) in the basal diet control (group 1) was 0.402±0.016 at the end of 4^th^ wk, which was significantly (*P*\<0.05) increased in all the remaining groups with the values ranging from 1.274±0.034 to 1.385±0.106. However, following supplementation with herbs and their combinations in test, there was a significant (*P*\<0.05) decrease in the concentration of serum creatinine in groups 3--15 at the end of 6^th^ wk, with the values ranging from 1.000±0.011 to 1.137±0.009 as compared to their corresponding 4^th^ wk values and that of iron toxic control group 2 (2.154±0.075). Amongst all the groups, the concentration remained significantly (*P*\<0.05) lower in basal diet control (0.573±0.038) at the end of 6^th^ wk. Lowest concentration of serum creatinine was recorded in group 12 (1.000±0.011) at the end of 6^th^ wk. In the present study, serum creatinine was significantly increased in the toxic controls at 4^th^ wk, which could be attributed to the free radical induced oxidative damage by iron on kidney. These findings can be further substantiated from the results of oxidative stress which indicated elevated TBARS and reduced GSH. The groups that were treated with herbs and herbal combinations following discontinuation of iron resulted in significant decrease in serum creatinine as compared to toxic control group, which further confirms the therapeutic potential of the herbs in test. The beneficial renal protective actions of herbs in test may be attributed to their antioxidant/free radical scavenging actions and protection of protein thiols from deleterious action of iron in kidney. In conclusion, the study revealed that iron induces toxicity to liver and kidney by means of oxidative stress by either generating excess free radicals or by disturbing antioxidant defenses and supplementation of herbs with antioxidant potential is useful in the management of accidental iron toxicity in broilers. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.392694

2011-01-01

PMC3052588

INTRODUCTION {#sec1-1} ============ Lead is one of the widely dispersed toxic substances, and sources of lead in environment include lead paints, automobiles, industrial lead emission and lead in food/feed and water. However, lead gasoline combustion in vehicles has accounted for as much as 90% of the total anthropogenic sources of environmental lead.\[[@CIT1]\] Such accumulated lead is toxic in most of its chemical forms, whether it is inhaled or ingested through water or feed. Lead affects the metabolism of other minerals and has affinity for bone, where it acts by replacing calcium. Thus, the highest concentrations of lead are usually found in bone, kidney and liver.\[[@CIT2]\] Studies have reported that lead has a potential for inducing oxidative stress and acts as a catalyst in the oxidative reactions of biological molecules by producing free radicals or reactive oxygen species (ROS; hydroxyl radicals, peroxyl radicals, alkyl radicals, etc.). Lead interferes with different antioxidant defenses like glutathione peroxidase (GSH-Px), glutathione reductase (GSH-R), catalase, superoxide dismutase (SOD) and reduced glutathione (GSH). Reduced amounts of antioxidants may contribute to progression of liver damage and damage to other organ systems including kidney and nervous system.\[[@CIT3]\] Several herbs have been reported to counter the peroxidative stress in biological system due to several stressors including pesticides, mycotoxins and heavy metals. Keeping the above facts in view, an experimental study was planned to evaluate herbal methionine and *Mangifera Indica* against experimental lead toxicosis in broilers. MATERIALS AND METHODS {#sec1-2} ===================== A total of 90 sexed male broiler chicks (*Cobb* strain) of day-old age were randomly divided into six groups consisting of 15 in each group. The chicks were provided with feed and water ad *libitum* throughout the experiment. The groups were maintained as per the following feeding schedule for 5 wk in order to study the antioxidant role of herbal methionine (combination of *Cicer arienticum, Triticum sativum, Phaseolus mungo, Mucuna puriens* and *Allium cepa*) and *M. indica* against lead-induced oxidative damage. Group 1: Basal diet controlGroup 2: Lead toxic control (300 mg/kg feed)Group 3: Herbal methionine control (25% above normal level in feed, i.e., 1.4 g/kg feed)Group 4: *M. indica* control (0.1% in feed)Group 5: Lead + herbal methionineGroup 6: Lead + **M. indica** The birds were sacrificed at the end of 5^th^ week and tissues were collected for the assay of reduced glutathione (GSH),\[[@CIT4]\] superoxide dismutase (SOD),\[[@CIT5]\] catalase,\[[@CIT6]\] protein carbonyls\[[@CIT7]\] and thiobarbituric acid reactive substances (TBARS)\[[@CIT8]\] in liver and kidney homogenates. The activities of Na^+^/K^+^ ATPase, Mg^2+^ATPase\[[@CIT9]\] and cytochrome P~450~ (CYP~450~)\[[@CIT10]\] were estimated in the liver homogenate. The data were analyzed by one-way analysis of variance (ANOVA) using statistical package for social sciences (SPSS) version 10. *P*\<0.05 was considered as significant. RESULTS AND DISCUSSION {#sec1-3} ====================== The concentration of TBARS \[μM of malondialdehyde (MDA)/mg protein\] in liver and kidney of basal diet control (group 1) was 2.37±0.13 and 1.90±0.12, respectively, which was significantly (*P*\<0.05) increased in lead toxic control group 2 at the end of 5^th^ week. Groups 5 and 6 that were supplemented with herbs revealed significantly (*P*\<0.05) decreased concentration of TBARS as compared to group 2. The concentration of protein carbonyls (nM/100 mg protein) in liver and kidney of basal diet control (group 1) was 18.79±1.02 and 10.13±0.75, respectively, which was significantly (*P*\<0.05) increased in lead toxic control group 2 at the end of 5^th^ week. Groups 5 and 6 that were supplemented with herbs revealed significantly (*P*\<0.05) decreased concentration of protein carbonyls as compared to lead control. The concentration of GSH (μM/mg protein) in liver and kidney of basal diet control (group 1) was 9.29±0.62 and 14.88±1.58, respectively, which was significantly (*P*\<0.05) decreased in lead toxic control group 2 at the end of 5^th^ week. Groups 5 and 6 that were supplemented with herbs revealed significantly (*P*\<0.05) increased concentration of GSH as compared to lead control. All these parameters in groups 3 and 4 were comparable to that of basal diet control \[Tables [1](#T0001){ref-type="table"} and [2](#T0002){ref-type="table"}\]. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Results of oxidative stress and enzymes in liver ::: Group TBARS (*μ*M of MDA/mg protein) GSH (μM/mg protein) SOD (U/mg protein) Catalase (*μ*g of H~2~O~2~ decomposed/ mg protein/min) Protein carbonyls (nM/100 mg protein) Na^+^/K^+^ ATPase (μM of Pi released/ mg protein/30 min) Mg^2+^ ATPase (*μ*M of Pi released/mg protein/30 min) CYP~450~ (*μ*M/ mg microsomal protein) ------- -------------------------------- --------------------- -------------------- -------------------------------------------------------- --------------------------------------- ---------------------------------------------------------- ------------------------------------------------------- ---------------------------------------- 1 2.37±0.13^a^ 9.29±0.62^c^ 13.36±0.57^d^ 11.65±0.58^cd^ 18.79±1.02^a^ 6.81±0.50^c^ 13.53±0.59^c^ 0.09±0.01^c^ 2 5.35±0.18^c^ 2.41±0.37^a^ 5.22±1.63^a^ 4.48±1.45^a^ 37.93±1.60^c^ 2.73±0.19^a^ 10.16±0.44^a^ 0.03±0.01^a^ 3 2.39±0.14^a^ 8.04±1.11^c^ 11.66±0.92^cd^ 11.62±0.67^c^ 19.36±0.82^a^ 6.75±0.38^c^ 13.35±0.42^c^ 0.09±0.01^c^ 4 1.89±0.08^a^ 9.89±0.53^c^ 14.15±1.08^d^ 13.23±1.82^d^ 18.07±1.83^a^ 6.99±0.76^c^ 14.73±0.83^c^ 0.10±0.02^c^ 5 3.99±0.38^b^ 5.37±0.79^b^ 8.58±0.72^b^ 8.10±0.42^b^ 30.14±0.97^b^ 4.28±0.50^b^ 11.80±0.28^b^ 0.06±0.01^b^ 6 3.89±0.32^b^ 5.95±0.70^b^ 10.13±0.53^bc^ 9.19±0.80^bc^ 26.57±3.44^b^ 4.79±0.39^b^ 11.81±0.22^b^ 0.06±0.01^b^ Values are Mean±SE (*n* = 8); one-way ANOVA (SPSS). Means with different alphabets as superscripts differ significantly (*P*\<0.05); TBARS, thiobarbituric acid reactive substances; MDA, malondialdehyde; GSH, glutathione; SOD, superoxide dismutase ::: ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Results of oxidative stress in kidney ::: Group TBARS concentration (*μ*M of MDA/mg protein) GSH concentration (*μ*M/mg protein) SOD activity (U/mg protein) Catalase activity (*μ*g of H~2~O~2~ decomposed/mg protein/min) Protein carbonyls concentration (nM/100 mg protein) ------- ---------------------------------------------- ------------------------------------- ----------------------------- ---------------------------------------------------------------- ----------------------------------------------------- 1 1.90±0.12^a^ 14.88±1.58^c^ 9.02±0.47^c^ 10.50±0.34^cd^ 10.13±0.75^a^ 2 5.33±0.91^c^ 3.11±0.57^a^ 1.59±0.13^a^ 4.07±0.26^a^ 34.22±1.20^c^ 3 2.02±0.13^a^ 12.45±0.32^bc^ 11.02±0.94^d^ 9.61±0.20^cd^ 13.14±0.93^a^ 4 1.17±0.14^a^ 15.48±1.58^c^ 9.95±0.31^cd^ 11.97±1.76^d^ 9.76±1.22^a^ 5 3.76±0.74^b^ 9.99±1.18^b^ 6.17±0.51^b^ 7.19±0.24^b^ 20.59±1.54^b^ 6 3.58±0.57^b^ 10.11±1.37^b^ 7.04±0.82^b^ 7.35±0.24^b^ 19.53±4.35^b^ Values are Mean±SE (*n* = 8); one-way ANOVA (SPSS). Means with different alphabets as superscripts differ significantly (*P*\<0.05); TBARS, thiobarbituric acid reactive substances; MDA, malondialdehyde; GSH, glutathione; SOD, superoxide dismutase ::: The activity of SOD (U/mg protein) in liver and kidney in basal diet control (group 1) was 13.36±0.57 and 9.02±0.47, respectively, which was significantly (*P*\<0.05) decreased in lead toxic control group 2 at the end of 5^th^ week. Groups 5 and 6 that were supplemented with herbs revealed significantly (*P*\<0.05) increased activity of SOD as compared to that of lead control. The activity of catalase (μg of H~2~O~2~ decomposed/mg protein/min) in liver and kidney of basal diet control (group 1) was 11.65±0.58 and 10.50±0.34, respectively, which was significantly (*P*\<0.05) decreased in lead toxic control group 2 at the end of 5^th^ week. Groups 5 and 6 that were supplemented with herbs revealed significantly (*P*\<0.05) increased activity of catalase as compared to lead control \[Tables [1](#T0001){ref-type="table"} and [2](#T0002){ref-type="table"}\] The concentration of TBARS and protein carbonyls in liver and kidney was significantly increased in toxic control group, while the activities of SOD and catalase, and the concentration of GSH in liver and kidney were reduced, suggesting the ongoing peroxidative stress and compromised antioxidant defense mechanisms. Enhanced oxidative stress contributes to lead-induced toxicity, where restoration of a cells' antioxidant capacity appears to provide a partial remedy. Several studies are underway to determine the effect of antioxidant supplementation following lead exposure. Data suggest that antioxidants may play an important role in abating some hazards of lead.\[[@CIT2]\] It has been reported earlier that lead impairs the antioxidant defenses of the cells and renders them more susceptible to oxidative attacks.\[[@CIT11][@CIT12]\] All the changes in the antioxidant defense profile were significantly reversed when treated with herbal methionine and *Ma. indica*. The beneficial effects of herbal methionine are attributed to antioxidant and anti-stress principles, namely, *Mu. pruriens*\[[@CIT13]\] and *A. cepa,*\[[@CIT14]\] besides other constituent principles. *Ma. indica* has proven antioxidant potential.\[[@CIT15]\] The aqueous stem bark extract of *Ma. indica* L. has been reported to have antioxidant properties. In a study, Hernandez *et al*,\[[@CIT16]\] reported that *Ma. indica* extract attenuated accumulation of reactive oxygen species (ROS) and intracellular free Ca^2+^, and consequently, down-regulated CD95L mRNA expression and CD95-mediated activation-induced cell death (AICD). *Mu. pruriens* has the ability to scavenge DPPH (α,α-diphenyl-β-picrylhydrazyl) radicals, ABTS radicals and reactive oxygen species. It significantly inhibits the lipid oxidation\[[@CIT17]\] and effectively scavenges hydroxyl radicals (OH') and superoxide anion radicals (O~2~'^−^).\[[@CIT18]\] Treatment with *Mu. pruriens* decreased the levels of lipid peroxidation and increased the levels of glutathione (GSH), superoxide dismutase (SOD) and catalase (CAT).\[[@CIT13]\] The activity of Na^+^/K^+^ ATPase and Mg^2+^ ATPase (μM of Pi released/mg protein/30 min) in liver of basal diet control group 1 was 6.81±0.50 and 13.53±0.59, respectively, which was significantly (*P*\<0.05) decreased in lead toxic control group 2 at the end of 5^th^ week. Groups 5 and 6 that were supplemented with herbs revealed significantly (*P*\<0.05) increased activity of ATPases as compared to lead control. The activity in groups 3 and 4 was comparable to that of basal diet control. The activity of CYP~450~ (μM/mg microsomal protein) in liver of basal diet control group 1 was 0.09±0.01, which was significantly (*P*\<0.05) decreased in lead toxic control group 2 at the end of 5^th^ week. Groups 5 and 6 that were supplemented with herbs revealed significantly (*P*\<0.05) increased activity of CYP~450~ as compared to lead control. The activity in groups 3 and 4 was comparable to that of basal diet control \[[Table 1](#T0001){ref-type="table"}\]. The activities of Na^+^/K^+^ ATPase, Mg^2+^ ATPase and CYP~450~ were significantly reduced in the liver of lead toxic control. Lipid peroxidation is a natural and deleterious process. The decrease in the levels of Na^+^/K^+^ ATPase, Mg^2+^ ATPase and CYP~450~ could be due to enhanced lipid peroxidation by free radicals in lead treated groups. Since these membrane-bound enzymes are sulfhydryl (SH) group containing and are lipid-dependent, any alteration in the membrane lipids owing to lipid peroxidation may significantly alter their activity. The treated groups revealed slight improvement in the membrane bound enzymes though the activities were not exactly similar to those of controls. It has been reported that dietary lead alters fatty acid composition and membrane peroxidation in chick liver microsomes.\[[@CIT19]\] These findings could be further substantiated from the results of the present study, which revealed a significant increase in lipid peroxidation (indicated by increased TBARS) in liver of the lead toxic control as compared to the remaining groups. The results of the present study enunciated that supplementation of herbal methionine and *Ma. indica* could significantly reverse the oxidative stress induced by lead. The beneficial actions of herbal methionine and *Ma. indica* may be attributed to their antioxidant, anti-stress and hepatoprotective principles. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.393425

2011-01-01

PMC3052589

INTRODUCTION {#sec1-1} ============ Lead is one of the widely dispersed toxic substances and lead gasoline combustion in vehicles has accounted for as much as 90% of the total anthropogenic sources of environmental lead.\[[@CIT1]\] Lead produces acute and chronic poisoning and induces a broad range of physiological, biochemical and behavioral dysfunctions resulting in reduced performance and death in livestock. Lead affects the metabolism of other minerals and has affinity for bone, where it acts by replacing calcium. Thus, the highest concentrations of lead are usually found in bone, kidney and liver.\[[@CIT2]\] Lead is toxic to poultry at much lower levels than previously recognized.\[[@CIT3]\] Renal activities of Na^+^/K^+^-ATPase and K^+^-dependent phosphatase were decreased to 50--70% of control values by long-term (12 weeks) ingestion of lead in rats.\[[@CIT4]\] *In vitro* lead at 50 and 100 *μ*M significantly inhibited ATPase activity in all regions of the rat brain.\[[@CIT5]\] However, literature is scanty on the effect of lead on the hepatocytic membrane ATPases and the influence of adaptogens on lead-induced derangement of ATPases. Keeping the above facts in view, the present study was taken up to examine the effect of lead on hepatic enzyme profile and histology, and to assess the ameliorating effect of certain adaptogens and polyherbal formulations (PHFs). MATERIALS AND METHODS {#sec1-2} ===================== A total of 225 sexed male broiler chicks (*Cobb* strain) of day-old age were randomly divided into 15 groups consisting of 15 chicks in each. Feed and water were provided *ad libitum* throughout the experiment. Groups 1, 2, 3, 4 and 5 were maintained on basal diet, PHF (stressroak; Ayurvet Ltd., Himachal Pradesh; composition: *Withania somnifera, Ocimum sanctum, Phyllanthus emblica, Mangifera indica* and shilajith at 100 ppm in feed), shilajith (100 ppm in feed), amla (500 ppm in feed) and vit E + Se (300 ppm + 0.3 ppm in feed), respectively, and groups 6 and 7 were the toxic controls that were kept on lead for 6 and 4 weeks, respectively. Groups 8, 9, 10 and 11 were given lead along with PHF, shilajith, amla and vit E + Se, respectively, for 6 weeks. Groups 12, 13, 14 and 15 were maintained on lead for the first 4 weeks and on PHF, shilajith, amla and vit E + Se, respectively, for the subsequent 2 weeks. Blood samples were drawn from the wing vein at the end of 4^th^ and 6^th^ weeks for the assay of alanine transaminase (ALT), using commercial diagnostic kits (Qualigens Pvt. Ltd., Mumbai, India). For preparation of liver microsomes,\[[@CIT6]\] liver pieces were collected in cold conditions by using pre-cooled petri dishes and were blotted free of blood and tissue fluid. After being weighed, the pieces were cross-chopped with surgical scalpel into fine slices, chilled in ice-cold 0.25 M sucrose and then blotted on a filter paper. The tissues were minced and homogenized in 10 mM Tris-HCl buffer (pH 7.4) at a concentration of 10% w/v with 25 strokes of tight Teflon homogenizer at a speed of 2500 rpm. The prolonged homogenization under hypotonic condition disrupted the ventricular structure of cells so as to release soluble proteins and leave membrane and non-vascular matter in sediment form. The homogenates were centrifuged at 10,000× *g* at 4°C for 20 min.\[[@CIT7]\] The supernatant was re-centrifuged for 1 hour at 100,000× *g* at 4°C using high-speed cooling centrifuge. The sediment was re-suspended in Tris-HCl buffer (pH 7.4) to get the final concentration of 10% and was used for the assay of membrane-bound enzymes. Total ATPase and Mg^2+^ATPase activities were determined by the method of Sadasivudu *et al*,\[[@CIT8]\] Na^+^/K^+^-ATPase activity was estimated as the difference between the total ATPase and Mg^2+^ATPase. Ca^2+^ATPase activity was determined by the method described by Kodama *et al*.\[[@CIT9]\] For CYP assay, livers were removed rapidly, weighed and perfused with ice-cold 1.15% KCl solution containing 0.05 M ethylenediaminetetraacetic acid (EDTA). Each of the perfused livers was homogenized with two volumes of ice-cold 0.25 M sucrose in the homogenizer. The homogenate was centrifuged at 10,000× *g* for 30 min in a refrigerated centrifuge. Microsomes from the supernatant fraction were isolated by the procedure of Cinti *et al*,\[[@CIT10]\] The microsomal pellet was washed once with 1.15% KCl solution containing 0.05 mM EDTA, re-suspended in phosphate buffer (pH 7.4) and the suspension was used for microsomal enzyme assays.\[[@CIT11]\] Total protein was estimated by the method described by Lowry *et al*,\[[@CIT12]\] Microsomal suspension in phosphate buffer (1.0 mg protein/ml) was treated with a few grains of sodium dithionate and taken in the reference and sample cuvettes and the baseline was recorded from 400 to 510 nm. The contents of the sample cuvette were transferred into a tube and bubbled gently with CO for 60 seconds. Then, the contents were transferred back to the sample cuvette and the difference in spectrum was recorded from 400 to 510 nm again. The difference in absorbance between 450 and 490 nm was then used for the calculation of CYP~450~content using the extinction coefficient difference (ΔE~450-490~ nm of 91 cm/mM).\[[@CIT13]\] The data were subjected to statistical analysis by applying two-way analysis of variance (ANOVA). Differences between means were tested using Duncan's multiple comparison test and significance was set at *P*\<0.05. RESULTS AND DISCUSSION {#sec1-3} ====================== The activity of ALT (units/l) in basal diet control (group 1) was 19.217±2.156, which was significantly (*P*\<0.05) elevated in lead toxic control groups 6, 7, 12, 13, 14 and 15 at the end of 4^th^ week. The activities of therapeutic control groups 2--5 were comparable to that of group 1 on 4^th^ and 6^th^ weeks. The groups 8--11 revealed a significant (*P*\<0.05) increase in the activity of ALT at the end of 4^th^ and 6^th^ weeks as compared to group 1 and other therapeutic controls (groups 2--5) at respective time intervals. In groups 12--15, there was a significant (*P*\<0.05) decrease in the activity at the end of 6^th^ week as compared to groups 6 and 7 \[[Table 1](#T0001){ref-type="table"}\]. The histological sections of the liver from the lead toxic control (group 6) showed moderate focal lymphoid aggregates \[[Figure 1](#F0001){ref-type="fig"}\], while group 11 showed mild central vein congestion and paracentral infiltration \[[Figure 2](#F0002){ref-type="fig"}\]. The lesions were very mild or absent in treated groups 12 through 15 and plain controls. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Activity of ALT and CYP~450~ in different groups of broiler chicks ::: Group ALT (IU/l) CYP~450~ (nmol/mg microsomal protein) --------------------------------------------------- ------------------- --------------------------------------- ------------------ Basal diet (1--42 days) 19.217±2.156^cA^ 24.962±1.453^bA^ 0.2520±0.009^d^ PHF (stressroak) (1--42 days) 13.588±1.737^aA^ 20.712±1.015^aA^ 0.2489±0.004^d^ Shilajith (1--42 days) 15.306±0.755^abA^ 21.185±0.940^aA^ 0.2529±0.010^d^ Amla (1--42 days) 17.433±0.928^bcA^ 21.807±0.808^aA^ 0.2510±0.008^d^ Vit E + Se (1--42 days) 16.906±1.15^bcA^ 23.236±0.981^abA^ 0.2499±0.010^d^ Lead (1--42 days) 60.114±3.110^hA^ 69.082±1.671^jA^ 0.1262±0.009^a^ Lead (1--28 days); basal diet (29--42 days) 60.531±1.481^hA^ 62.023±0.536^fgA^ 0.1523±0.011^ab^ Lead + PHF (stressroak) (1--42 days) 50.160±1.076^dA^ 61.345±1.113^fA^ 0.1850±0.007^bc^ Lead + shilajith (1--42 days) 52.941±1.236^eA^ 64.278±1.374^ghA^ 0.1890±0.010^c^ Lead + amla (1--42 days) 55.340±0.796^efA^ 66.061±1.026^hiA^ 0.1869±0.010^c^ Lead +Vit E + Se (1--42 days) 57.079±0.871^fgA^ 67.164±1.922^ijA^ 0.1820±0.006^bc^ Lead (1--28 days); PHF (stressroak) (29--42 days) 59.028±1.539^ghA^ 32.453±1.609^cA^ 0.1922±0.0155^c^ Lead (1--28 days); shilajith (29--42 days) 58.250±1.341^ghA^ 35.226±0.981^dA^ 0.1912±0.009^c^ Lead (1--28 days); amla (29--42 days) 58.929±1.988^ghA^ 37.344±1.182^deA^ 0.1890±0.003^c^ Lead (1--28 days); vit E + Se (29--42 days) 59.115±0.663^ghA^ 37.767±1.062^eA^ 0.1932±0.007^c^ Values are mean±SE of eight observations. Means with different alphabets as superscripts differ significantly (*P*\<0.05); ANOVA. Capital alphabets (horizontal comparison); small alphabets (vertical comparison); ALT, alanine transaminase ::: ::: {#F0001 .fig} Figure 1 ::: {.caption} ###### Liver of group 6 showing moderate focal lymphoid aggregates (H and E ×100) :::  ::: ::: {#F0002 .fig} Figure 2 ::: {.caption} ###### Liver of group 11 showing mild central vein congestion and paracentral infiltration :::  ::: The activity of ALT was determined to assess the degree of damage to liver as the levels of certain enzymes like ALT, aspartate transaminase (AST), gamma glutamyl transferase (GGT), etc., are elevated following hepatocellular injury.\[[@CIT14]\] In this study, the activity of ALT was significantly elevated in the lead toxic control group, suggesting the hepatocellular insult following administration of lead. These results are further substantiated from the histopathological studies of liver, which revealed moderate focal lymphoid aggregates in lead toxic control. Treatment with PHF (stressroak), shilajith, amla and vit E + Se, following discontinuation of lead resulted in significant reduction in the activity of ALT. The hepatocellular injury due to lead could be attributed to the lead-induced generation of reactive oxygen species (ROS) or free radicals and the reversal of the findings following treatment could be attributed to the antioxidant and the hepatoprotective potential of the drugs in test. Alcoholic and aqueous extracts of the fruits of *Emblica* have shown hepatoprotective properties in experiments in rats.\[[@CIT15]\] Further, the treatment controls, which were not exposed to lead did not manifest any significant change in the activity of ALT, and also there were no significant lesions in the liver tissues. However, simultaneous supplementation of drugs in the test along with lead revealed a significant increase in ALT activity at different time intervals when compared to the plain control groups, though the values were significantly lower in comparison to pure lead toxic control groups 6 and 7. This finding suggests the prophylactic potential of these drugs to prevent lead-induced toxic manifestations, though there was no complete prevention of changes. The CYP~450~ activity (nmol/mg microsomal protein) in basal diet control (group 1) was 0.2520±0.009, which was significantly (*P*\<0.05) decreased in lead toxic control groups 6 and 7 at the end of 6^th^ week. The activities of therapeutic control groups 2--5 revealed no significant difference in CYP~450~ activity as compared to group 1, while the groups 8--11 revealed a significant (*P*\<0.05) decrease in the activity as compared to group 1, and groups 12-15 revealed a significant (*P*\<0.05) increase in the activity as compared to group 6 at the end of 6^th^ week \[[Table 1](#T0001){ref-type="table"}\]. The activity of Na^+^/K^+^-ATPase, Mg^2+^ATPase and Ca^2+^ATPase (mM of Pi liberated/mg protein) in basal diet control (group 1) was 2.5919±0.062, 1.8691±0.098, and 1.6609±0.081, respectively, which was significantly (*P*\<0.05) decreased in lead toxic control groups 6 and 7 at the end of 6^th^ week. The activities of therapeutic control groups 2-5 revealed no significant difference in the activity of ATPases as compared to group 1 at the end of 6^th^ week. Groups 8-11 revealed a significant (*P*\<0.05) decrease in the activity as compared to group 1 at the end of 6^th^ week. Groups 12-15 revealed a significant (*P*\<0.05) increase in the activity as compared to group 6 \[[Table 2](#T0002){ref-type="table"}\]. ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Activities of Na^+^/K^+^-ATPase, Mg^2+^ATPase and Ca^2+^ATPase in the liver of different groups of broiler chicks ::: Group Na^+^/K^+^-ATPase (mM of Pi liberated/mg protein) Mg^2^+ATPase (mM of Pi liberated/mg protein) Ca^2+^ATPase (mM of Pi liberated/mg protein) --------------------------------------------------- --------------------------------------------------- ---------------------------------------------- ---------------------------------------------- Basal diet (1--42 days) 2.5919±0.062^e^ 1.8691±0.098^d^ 1.6609±0.081^e^ PHF (stressroak) (1--42 days) 2.5833±0.085^e^ 1.8817±0.074^d^ 1.6553±0.043^e^ Shilajith (1--42 days) 2.4485±0.102^de^ 1.8911±0.026^d^ 1.6845±0.063^e^ Amla (1--42 days) 2.6009±0.091^e^ 1.8699±0.036^d^ 1.6695±0.046^e^ Vit E + Se (1--42 days) 2.6324±0.062^e^ 1.8831±0.044^d^ 1.6806±0.061^e^ Lead (1--42 days) 0.9713±0.045^a^ 1.1284±0.054^a^ 0.7723±0.037^a^ Lead (1--28 days); basal diet (29--42 days) 1.1699±0.097^ab^ 1.3206±0.077^ab^ 0.9886±0.042^ab^ Lead + PHF (stressroak) (1--42 days) 2.2062±0.063^cd^ 1.5609±0.059^bc^ 1.4574±0.091^de^ Lead + shilajith (1--42 days) 2.0524±0.121^c^ 1.3261±0.052^ab^ 1.2499±0.081^cd^ Lead + amla (1--42 days) 1.8905±0.076^c^ 1.3296±0.107^ab^ 1.2450±0.056^cd^ Lead + vit E + Se (1--42 days) 1.9141±0.069^c^ 1.3519±0.078^ab^ 1.2836±0.074^cd^ Lead (1--28 days); PHF (stressroak) (29--42 days) 2.1842±0.139^cd^ 1.6240±0.056^cd^ 1.1567±0.113^bc^ Lead (1--28 days); shilajith (29--42 days) 2.1204±0.109^cd^ 1.4530±0.1196^bc^ 1.0929±0.049^bc^ Lead (1--28 days); amla (29--42 days) 1.3955±0.076^b^ 1.3439±0.094^ab^ 1.0499±0.056^bc^ Lead (1--28 days); vit E + Se (29--42 days) 2.1639±0.159^cd^ 1.5565±0.071^bc^ 1.1474±0.015^bc^ Values are mean±SE of eight observations. Means with different alphabets as superscripts differ significantly (*P*\<0.05); ANOVA ::: In this study, significant reduction in the activities of Na^+^/K^+^-ATPase, Ca^2+^ATPase, Mg^2+^ATPase and CYP~450~ in the liver of lead toxic control groups 6 and 7 may be attributed to the lipid proxidation by free radicals in lead treated groups, since these membrane-bound enzymes are SH group containing enzymes, which are lipid-dependent and their activity is likely to be altered with the change in the composition of membrane lipids.\[[@CIT6]\] The treated groups revealed slight improvement in the membrane-bound enzymes though the activities were not exactly similar to those of controls. It has been reported that dietary lead alters fatty acid composition and membrane peroxidation in chick liver microsomes.\[[@CIT16]\] These findings could be further substantiated from the results of the present study, wherein there was a significant increase in the activity of ALT (suggesting hepatocellular injury) and histopathological changes in liver in the toxic control group 6 as compared to the plain controls and treated groups. In conclusion, the study enunciated that lead induces damage to the hepatocytes leading to elevation of ALT along with reduction in CYP and membrane ATPases. Supplementation of anti-stress agents is of use to reverse the hepatocytic damage caused by lead. Amongst the drugs in test, PHF (stressroak) was found to be superior owing to its synergistic antioxidant herbs, followed by shilajith, amla and vit E + Se in the given order. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.394852

2011-01-01

PMC3052590

INTRODUCTION {#sec1-1} ============ Long-term low-level exposure to lead may result in reduced free T~4~concentration.\[[@CIT1]\] There is a strong correlation between maternal and umbilical cord blood lead levels, indicating the transfer of lead from mother to fetus.\[[@CIT2]\] Several data suggest that for lead, the main toxic event is prenatal exposure.\[[@CIT3]\] In addition, lead levels in breast milk also increase proportionately with the lead level in maternal blood, posing an additional risk to the neonates/nursing infants.\[[@CIT4]\] Zaidi *et al*.,\[[@CIT5]\] reported that upon exposure to the dust and liquid formulation of endosulfan, monocrotophos (MCP), parathion and fenvalerate, levels of TSH were elevated and T~4~ levels were decreased as compared to the values of control group. Therefore, based on these earlier findings, it is hypothesized that organophosphate an d lead compounds could affect maternal thyroid hormone availability for the fetus and therefore affect the development of the neonates. Further, studies on the combination of these compounds (MCP + lead acetate), a natural feature in environmental pollution, have rarely been attempted. Also, there is a need to establish a link, if any, between the exposure of mothers to these pollutants and altered thyroid homeostasis in both the dams and fetuses, and the onset of a syndrome of toxicological manifestations in the neonates. Hence, this study was taken up to evaluate the hepatotoxic effects of these pollutants individually and in combination on pups whose mothers were treated from 3^rd^ day of gestation till the weaning of the pups. MATERIALS AND METHODS {#sec1-2} ===================== Female albino rats of *Wistar kyoto* strain, weighing about 200-250 g, were procured from National Institute of Nutrition, Hyderabad. The animals were housed in solid bottom polypropylene cages. Animals were given commercial standard mash feed and provided water *ad libitum*. Experiment was conducted according to the guidelines of Institutional Animal Ethics Committee. Five groups comprising eight dams in each were mated and the progeny raised were maintained from postnatal day (PND) 1 to 90. The dams were maintained as per the following schedule from day 3 of pregnancy till the weaning of pups (PND 21). 1. Group 1: Sham 2. Group 2: Methimazole 0.02% in drinking water 3. Group 3: MCP at 0.3 mg/kg orally 4. Group 4: Lead acetate at 0.2% in drinking water 5. Group 5: Combination of MCP + lead acetate as detailed in groups 3 and 4 Only thyroid profile was studied in dams, while the progeny were subjected to the liver tissue analysis. Blood samples were collected from the dams on 14^th^ day of gestation and sera samples were separated to estimate thyroid hormone profile by radioimmunoassay. Eight pups from each group were euthanized on PND 21 and 90 for analysis. Thiobarbituric acid reactive substances (TBARS),\[[@CIT6]\] protein carbonyls\[[@CIT7]\] and reduced glutathione (GSH)\[[@CIT8]\] of liver were studied on PND 21 and 90, while the activities of Na^+^ /K^+^ATPase\[[@CIT9]\] and Mg^2+^ ATPase\[[@CIT10]\] in the liver were studied on PND 90. Protein estimation was done by the method of Lowry *et al*.,\[[@CIT11]\] The data were analyzed by one-way analysis of variance (ANOVA) using statistical package for social sciences (SPSS) version 15. Differences between means were tested using Duncan's multiple comparison test and significance was set at *P*\<0.05. RESULTS AND DISCUSSION {#sec1-3} ====================== The mean concentrations of T~3~ (ng/dl) and T~4~ (*μ*g/dl) were significantly (*P*\<0.05) reduced, while that of TSH (*μ*U/ml) was significantly (*P*\<0.05) increased in all the treated groups of dams as compared to that of control group 1 (158.68±21.46, 3.83±0.14 and 0.06±0.02, respectively). Thyroid hormone profile was estimated in order to establish the influence of MCP and lead acetate on the functioning of thyroid. It is well documented that acute exposure to high levels of persistent organic pollutants and heavy metals can affect human health including thyroid function\[[@CIT12]\] by significantly decreasing the concentration of T~3~ and T~4~ and significantly increasing TSH. Methimazole is a known anti-thyroid drug, while MCP and lead acetate are strong inducers of oxidative stress. Hence, it is hypothesized that the free radicals generated in dams following exposure to these chemicals might have induced dysfunctioning of thyroid. The concentration of TBARS \[*μ*M malondialdehyde (MDA)/mg protein\] in liver revealed a significant (*P*\<0.05) rise in progeny rats of all the test groups on 21^st^ and 90^th^ days as compared to group 1 rats (0.12 0.02 and 0.18±0.03, respectively). The concentration of protein carbonyls (nM/100 mg protein) in liver recorded a significant (*P*\<0.05) increase in progeny of all the test groups on 21^st^ and 90^th^ days as compared to group 1 rats (1.42±0.18 and 1.78±0.07, respectively) \[[Table 1](#T0001){ref-type="table"}\]. Lead is known to produce oxidative damage in the liver tissue by enhancing peroxidation of membrane lipids, a deleterious process solely carried out by free radicals.\[[@CIT13]\] Organophosphates have been reported to increase TBARS concentration (an indicator of lipid peroxidation) on sub-chronic exposure.\[[@CIT14]\] Free radicals cause peroxidation of lipids, resulting in the formation of aldehydes such as TBARS, while oxidation of proteins results in the formation of carbonyls. In this study, concentration of TBARS and protein carbonyls was increased in liver in the progeny pups of treatment groups, suggesting ongoing oxidative stress. This finding is in accordance with an earlier report by Elvira *et al*.,\[[@CIT15]\] which shows similar findings on oxidative stress following prenatal exposure to lead. The concentration of GSH (mM/g tissue) in liver revealed a significant (*P*\<0.05) reduction in the progeny rats of all the test groups on 21^st^ and 90^th^ days as compared to group 1 rats (4.23±0.39 and 4.04±0.17, respectively) \[[Table 1](#T0001){ref-type="table"}\]. Lead also has a very high affinity for the sulfhydryl groups of GSH, which has implications for the maintenance of the thiol-disulfide balance in the cell.\[[@CIT16]\] Hepatic and extrahepatic GSH depletion and glutathione-S-transferase (GST) inhibition were observed in different tissues of albino rats that were given a single oral dose of MCP.\[[@CIT17]\] In the present study, concentration of GSH was reduced significantly in the treated groups in liver as compared to pups of control group. The results of oxidative stress on the biomarkers revealed a correlation between the action of methimazole and toxic test compounds (MCP and lead acetate). As the stress parameters of methimazole can be attributed directly to the anti-thyroid effects of the drug, the simulating actions of both MCP and lead acetate can be correlated to anti-thyroid actions in fetuses and neonates, besides direct free radical generating ability of these test chemicals, which enter into the fetus through maternal circulation and to the neonates through milk while nursing. There was a significant (*P*\<0.05) decrease in the activity (*μ*M of Pi released/mg protein/30 min) Na^+^ /K^+^ ATPase and Mg^2+^ ATPase in the liver of all the groups (2-5) of test progeny as compared to that in group 1 (16.56±1.73 and 10.75±0.86, respectively) \[[Table 1](#T0001){ref-type="table"}\]. The action of membrane-bound enzymes depends on the cellular membrane integrity; an altered integrity of membrane following derangement of membrane lipids due to lipid peroxidation may affect the functioning of membrane ATPases, subsequently leading to ionic imbalances in the cells. This is evident from the results of the present investigation, which revealed increased lipid and protein peroxidation. In conclusion, the study revealed that the hepatotoxic effects recorded in the progeny of dams treated with MCP and lead acetate were comparable to those of the progeny of dams administered with methimazole, a proven anti-thyroid drug. Therefore, it may be possible that both MCP and lead acetate might have a possible influence on the thyroid gland of dams as the thyroid hormone levels, which play an important role in early prenatal life in the proper development of fetus, were significantly altered in dams and were comparable among the treatment groups. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Results of oxidative stress and ATPases in liver of progeny rats ::: Group TBARS (uM of MDA/mg protein) Protein carbonyls (nM/100 mg protein) GSH (mM/g tissue) Na/K^+^ ATPase (*μ*M of Pi released/mg protein/30 min) Mg^2+^ (*μ*M of Pi released protein/30 min) ------------------- ------------------------------ --------------------------------------- ------------------- -------------------------------------------------------- --------------------------------------------- --------------- --------------- --------------- Control 0.12±0.02^aA^ 0.18±0.03^aA^ 1.42±0.18^aA^ 1.78±0.07^aA^ 4.23±0.39^aA^ 4.04±0.17^aA^ 16.56±1.73^a^ 10.75±0.86^a^ Methimazole 0.19±0.03^bA^ 0.29±0.02^bA^ 2.76±0.26^bA^ 3.33±0.19^cA^ 1.46±0.37^cA^ 1.37±0.21^cA^ 5.32±0.71^c^ 5.69±0.18^c^ MCP 0.21±0.01^bA^ 0.31±0.04^bA^ 2.31±0.32^bA^ 2.52±0.11^bA^ 3.31±0.21^bA^ 2.87±0.19^bA^ 9.20±1.62^b^ 6.67±0.36^b^ Lead acetate 0.22±0.02^bA^ 0.34±0.01^bA^ 2.53±0.41^bA^ 2.70±0.19^bA^ 2.90±0.29^bA^ 2.41±0.15^bA^ 12.54±1.41^b^ 8.99±0.40^b^ MCP+ lead acetate 0.27±0.02^cA^ 0.44±0.02^cA^ 2.81±0.31^bA^ 3.43±0.15^cA^ 1.53±0.22^cA^ 1.42±0.36^cA^ 5.46±0.44^c^ 3.26±0.62^d^ Values are Mean±SE (*n* = 8); one-way ANOVA (SPSS). Means with different alphabets as superscripts differ significantly (*P*\<0.05); capital alphabets indicate horizontal comparison between time intervals and small alphabets indicate vertical comparison among groups; TBARS, Thiobarbituric acid reactive substances (TBARS); GSH, reduced glutathione; MCP, monocrotophos ::: **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.396138

2011-01-01

PMC3052591

INTRODUCTION {#sec1-1} ============ Cell phone usage is a major public health concern because of potential risk of chronic exposure to low level of radiofrequency and microwave radiation that pulse off the phone antennae in close proximity to the head.\[[@CIT1]\] These concerns have induced a large body of research, both epidemiological and experimental, in humans and animals. Honeybees are reliable indicators of environmental status and possess several important ecological, ethological, and morphological characteristics. They are the best experimental animals to study the effect of electromagnetic waves because they possess in their abdomen magnetite granules which help the bees in their orientation flight. Moreover, the integument of bees has semiconductor functions. It is in the light of these characteristics of honey bees that the present investigation was planned to study the metabolic changes with respect to proteins, carbohydrates, and lipids in hemolymph of worker honeybee of *Apis mellifera* L. exposed to cell phone radiation. MATERIALS AND METHODS {#sec1-2} ===================== Study area {#sec2-1} ---------- The samples of *A. mellifera* adult worker bees were drawn from the colonies maintained by Department of Zoology, Punjab University, Chandigarh. Experimental design {#sec2-2} ------------------- A specially designed wooden box called the observation hive was used for the experiment. Front and back of the box were made up of glass while the two sides had wire gauze to ensure proper ventilation. Two such boxes, one experimental and the other control, were taken for the present study. The phones used were of the same make and model and had the same network. Phones were kept in listen-talk mode for 40 min using a tape recorder. Ten honeybees were collected from the exposed frame at intervals of 10, 20 and 40 min. Ten honeybees were collected from the control at the same time intervals. Sample preparation {#sec2-3} ------------------ Hemolymph of the worker bees was extracted with the help of the micropipette inserted into the inter-segmental region of the bee's abdomen. Equal volume of hemolymph from all bees was dissolved in 1 ml of normal saline. Biochemical estimation {#sec2-4} ---------------------- ### Estimation of total carbohydrates {#sec3-1} The quantitative estimation of total carbohydrates in the test and control samples of *A. mellifera* was carried out by following the method of Sawhney and Singh.\[[@CIT2]\] ### Estimation of glycogen {#sec3-2} Seifter's method\[[@CIT3]\] was followed for the estimation of glycogen in treated and non-treated hemolymph of *A. mellifera*. ### Glucose estimation {#sec3-3} For the estimation of glucose in the hemolymph of *A. mellifera*, the method of Somogyi and Nelson\[[@CIT4]\] was employed. ### Estimation of total lipids {#sec3-4} The quantitative estimation of total lipids in the treated and non-treated hemolymph extract of *A. mellifera* was carried out by following the procedure of Fringes and Dunn.\[[@CIT5]\] ### Cholesterol estimation {#sec3-5} Method of Zlatki's *et al*.\[[@CIT6]\] was employed for the estimation of cholesterol in treated and non-treated sample. ### Estimation of protein {#sec3-6} The total quantity of protein in the test and control sample of *A. mellifera* was determined by following the standard procedure of Lowry.\[[@CIT7]\] RESULTS AND DISCUSSION {#sec1-3} ====================== Total carbohydrates {#sec2-5} ------------------- ### Control {#sec3-7} In the non-treated or control sample, the concentration of total carbohydrates in the hemolymph was found to be 1.29±0.02 mg/ml. ### Test {#sec3-8} In the hemolymph of treated bees, the concentration was 1.5±0.04 mg/ml in 10 min, 1.73±0.01 mg/ml in 20 min and 1.61±0.02 mg/ml in 40 min exposure samples. ### Glycogen {#sec3-9} In the treated sample, the glycogen content (mg/ml) was found to be 0.019 0.001 as compared to 0.047 0.001, 0.076±0.001 and 0.028±0.002 in 10, 20 and 40 min exposure samples, respectively. ### Glucose {#sec3-10} The glucose content (mg/ml) in control sample was 0.218±0.0005, while in the various treated samples the concentration was 0.231±0.002, 0.277±0.001 and 0.246±0.002 in 10, 20 and 40 min exposure samples, respectively. ### Total lipids {#sec3-11} The concentration of total lipids (mg/ml) in the hemolymph of control worker bee was found to be 2.06±0.02. For the treated sample, the concentration of total lipids was 3.03±0.02, 4.50±0.035 and 3.10±0.02 in 10, 20 and 40 min exposure samples, respectively. ### Cholesterol {#sec3-12} The cholesterol concentration (mg/ml) in the non-treated sample was 0.230 0.001. In the treated sample, the concentration was 1.381±0.002, 2.565±0.002 and 1.578±0.002 in 10, 20 and 40 min exposure samples, respectively. ### Total protein content {#sec3-13} In the hemolymph of control sample, the protein concentration (mg/ml) was 0.475±0.002. In the treated sample, the protein concentration was 0.525±0.003, 0.825±0.0001 and 0.650±0.0003 in 10, 20 and 40 min exposed samples, respectively. Very little work has been done on biochemical, metabolic and physiological influences of cell phone radiations pertaining to health risk in man.\[[@CIT8]\] Therefore, the present investigations on the influence of cell phone radiations on some biochemical and physiological aspects of honeybee biology were undertaken. That the behavior of honeybee is altered to some extent by high or low energy fields or electromagnetic radiations has been known for quite some time.\[[@CIT9]\] During the present investigation, it was observed that there was an increase in concentration of total carbohydrates in the bees exposed to cell phone radiation for 10 min as compared to unexposed or control bees. Increasing the exposure time to 20 min resulted in further increase in the concentration, while an exposure of 40 min had a reverse effect and there was a decline in carbohydrate concentration, though it was still higher as compared to control. Hemolymph glycogen and glucose content also showed the same trend, i.e., there was increase in content up to 20 min exposure after which there was a slight decline in the concentration which remained more than the control. Sharma\[[@CIT10]\] had also reported increase in glycogen and glucose levels in the exposed pupa of *A. mellifera*. Lipids are the major energy reserves of insects. Certain lipid classes are structure components of membranes while others are raw materials for a variety of hormones and pheromones. Estimation of total lipids and cholesterol during the present study showed that the trend was similar to that of carbohydrates. After an initial increase in concentration at the 10 and 20 min exposure period, a decline was observed in the concentration of total lipids and cholesterol at 40 min exposure. It was interesting to note that during the present study as the exposure time increased, it appeared that the bees having assessed the source of the disturbance decided to move and a large scale movement of the workers toward the talk-mode (not toward the listening mobile) was observed. Also, the bees became slightly aggressive and started beating their wings in agitation. This mobility of the bees could be responsible for increase utilization of energy sources and consequent decrease in concentration of carbohydrates and lipids in the 40 min exposed sample. Research facilities provided by Department of Zoology, Punjab University, Chandigarh, are gratefully acknowledged. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.397363

2011-01-01

PMC3052592

INTRODUCTION {#sec1-1} ============ Advent of agricultural and industrial revolution in India has added many pollutants in the environment, which are potentially hazardous, out of which some may be toxic, inflammable, explosive or corrosive. Of the various pollutants present in the hydrosphere and lithosphere, heavy metals and pesticides are toxic to livestock as well as human beings. However, not all the pesticides are actually toxic for humans or other non-target species.\[[@CIT1]\] Endosulfan (6, 7, 8, 9, 10, 10 -hexachloro- 1, 5, 5a, 6, 9, 9a -methano -2, 4, 3 -benzodioxathiepin- 3 -oxide) is an organochlorine insecticide and acaricide, and acts as a contact poison in a wide variety of insects and mites.\[[@CIT2]\] It is being easily absorbed by the stomach, lungs and through the skin, which means that all routes of exposure can pose a hazard. It enhances the effect of estrogens and acts as an endocrine disruptor, causing reproductive and developmental damage in animals and humans, as well as cancer. Recent studies indicate that pesticide intoxication produces oxidative stress by the generation of free radicals and by inducing tissue lipid peroxidation in mammals and other organisms.\[[@CIT3]\] Hincal *et al*,.\[[@CIT4]\] reported the oxidative stress inducing effects of endosulfan, with an increase of lipid peroxidation and a significant alteration in glutathione (GSH) redox cycle in cerebral and hepatic tissues of rats. The organophosphorus (OP) insecticide, chlorpyrifos (*O*,*O*-diethyl O-3, 5, 6-trichloro-2-pyridyl-phosphorothioate), is widely used for a variety of agricultural and human health applications. OPs produce a wide range of toxicity in mammals by inhibiting acetylcholinesterase (AChE), and the consequent accumulation of the neurotransmitter acetylcholine (ACh) in synaptic junction leads to excessive stimulation of postsynaptic cells causing cholinergic toxicity.\[[@CIT5]\] In fact, one of the molecular mechanisms of the toxicity of some pesticides seems to be lipid peroxidation. As a consequence, these compounds can disturb the biochemical and physiological functions of the red blood cells (RBCs).\[[@CIT6]\] The susceptibility of RBC to oxidative damage is due to the presence of polyunsaturated fatty acid, heme iron and oxygen, which may produce oxidative changes in RBC.\[[@CIT7]\] Major contributors to non-enzymatic protection against lipid peroxidation are vitamin E and vitamin C, well-known free radical scavengers.\[[@CIT8]\] Vitamin E is a lipid soluble, chain-breaking antioxidant playing a major protective role against oxidative stress and prevents the production of lipid peroxides by scavenging free radicals in biological membranes. Some investigators reported that administering Vitamin E may be useful in controlling the toxic effect of insecticides and chemicals.\[[@CIT9]\] Keeping these points in mind, the present study was planned to establish the antioxidant role of vitamin E on oxidative stress induced in rat RBCs by pesticides. MATERIALS AND METHODS {#sec1-2} ===================== Ten male healthy rats weighing about 100--200 g were used in the study. About 3 ml peripheral blood was obtained from ocular vein/heart puncture of rats, using ethylenediaminetetraacetic acid (EDTA)-sodium salt as the anticoagulant for assays. Blood was centrifuged at 2000 rpm for 10 min. Plasma and buffy coat was removed. Subsequently, the cells were washed three times with phosphate buffered saline (PBS), pH 7.2. The final red cell suspension was taken in test tubes for chlorpyrifos, endosulfan and vitamin E treatment, each in triplicate set. Pesticides and vitamin E (chlorphyrifos, endosulfan, vitamin E, chlorphyrifos and vitamin E, and endosulfan and vitamin E) were dissolved in dimethyl sulfonic acid (DMSO) and the solution was made up 100 ppm stock solution of each group, respectively. The above combinations were mixed and the desired 1 ppm concentration was made in all groups (DMSO 5% of total volume and test was performed in triplicate set). Also, 5% DMSO was dissolved/mixed in control group. The tubes were incubated for 3 hours at 37°C in a shaking water bath. At the end of incubation, the tubes were removed and subjected to biochemical analysis. Biochemical analysis {#sec2-1} -------------------- ### Chemicals for analysis {#sec3-1} Alpha endosulfan (100%) and chlorpyrifos (100%) were purchased from Accustandard Inc., NH, USA. Tris cacodylic acid, diethylenetriamine-penta-acetic acid (DTPA) (99% pure), nitroblue tetrazolium (98%), pyrogallol (\>98%), sodium dodecyl sulfate (\>99%), bovine serum albumin (97%), triton X-100 and thiobarbituric acid (99%) were purchased from Sigma Chemicals (USA). Nitric acid (69%), NaH~2~PO~4~ (98%), KH~2~PO~4~ (99.5%), pyridine (99%) and 1-butanol (0.99%) extra pure grade were purchased from Qualigens Chemicals, Mumbai, India. ### Lipid peroxidation in erythrocytes {#sec3-2} The level of lipid peroxides (LPOs) in erythrocyte hemolysate was determined spectrophotometrically following the method of Placer *et al*.\[[@CIT10]\] Lipid peroxidation was calculated using 1.56 × 10^5^ as extinction coefficient\[[@CIT11]\] to express the value in nanomoles of malondialdehyde (MDA) per milliliter of hemolysate. Hemoglobin (Hb) in hemolysate was estimated spectrophotometrically by cyanomethemoglobin method\[[@CIT12]\] and LPO level in the erythrocytes was expressed as nanomoles of MDA per mg of hemoglobin. ### Antioxidant enzymes {#sec3-3} The erythrocyte homogenate was used for analysis of antioxidant enzymes after suitable dilution. Activity of superoxide dismutase (SOD) was measured in the 10% RBC hemolysate, following the method of Marklund and Marklund\[[@CIT13]\] with certain modifications suggested by Minami and Yoshikawa\[[@CIT14]\]. Each unit of SOD activity is defined as the quantity of enzyme that inhibits auto-oxidation of pyrogallol by 50% under suitable experimental conditions. Catalase (CAT) activity in RBC hemolysate was estimated spectrophotometrically at 240 nm wavelength after appropriate dilution, following the method of Cohen *et al*.\[[@CIT15]\] and the values were expressed in units per milligram of hemoglobin for erythrocytes. Glutathione-*S*-transferase (GST) activity in RBC hemolysate was estimated spectrophotometrically at 340 nm wavelength as per the method of Habig *et al*.\[[@CIT16]\] and the values were expressed in units per milligram of hemoglobin for erythrocytes. Statistical analysis {#sec2-2} -------------------- The data obtained were statistically analyzed using two-way analysis of variance to find out significance of difference within a group at different periods of observation and Student's *t*-test to find out the significance between the groups at a particular period of observation.\[[@CIT17]\] RESULTS AND DISCUSSION {#sec1-3} ====================== Data pertaining to Hb concentration and enzyme activity are presented in [Table 1](#T0001){ref-type="table"}. There was significantly higher (*P*\<0.05) Hb level in the endosulfan exposed group and lowest level was found in the control group (without exposure and treatment) erythrocyte cell lysate. The level of LPO in cell lysate was significantly (*P*\<0.05) greater in the group exposed to endosulfan and chlorpyrifos, and lowest in the control group. The LPO level in cell lysate reduced significantly (*P*\<0.05) in vitamin E treated groups as compared to their respective non-treated pesticide exposed group. The SOD activity was significantly lower in endosulfan and chlorpyrifos exposed groups (18.16±0.44 and 26.14±0.52, respectively). Rats' erythrocytes exposed to endosulfan and chlorpyrifos at 1 ppm concentration and on treatment with vitamin E showed greater SOD activity compared with non-treated pesticide exposed group. It was with duration treatment (3 hrs) E treatment improved SOD activity with duration treatment in both pesticide groups. However, vitamin E treatment in both pesticide exposures increased CAT activities in erythrocytes. The CAT activity in erythrocytes decreased gradually following both the pesticide exposures. The GST activity in erythrocytes increased significantly (*P*\<0.05) in pesticide exposed groups as compared to their respective vitamin E treated group. The GST activity was significantly higher in endosulfan and chlorpyrifos exposed groups (2.16±0.04 and 3.63±0.03, respectively). ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Activities of antioxidant enzymes in erythrocytes of the treatment groups ::: Hb (g/dl) LPO (nmoles MDA/mg of Hb) SOD (units/mg of Hb) CAT (units/mg of Hb) GST (units/mg of Hb) --------------------------- ---------------- --------------------------- ---------------------- ---------------------- ---------------------- Control 7.33±0.01^a^ 8.45±0.00^a^ 32.55±0.02^d^ 115.02±0.47^f^ 1.13±0.00^b^ Chlorphyrifos 7.42±0.02^abc^ 9.21±0.03^c^ 26.14±0.52^b^ 80.00±1.06^c^ 3.63±0.03^c^ Endosulfan 7.52±0.04^c^ 10.43±0.08^d^ 18.16±0.44^a^ 46.00±0.73^a^ 2.16±0.04^d^ Vitamin E 7.30±0.03^ab^ 8.78±0.00^b^ 32.08±0.07^d^ 109.00±0.37^e^ 1.03±0.01^a^ Chlorphyrifos + Vitamin E 7.40±0.06^abc^ 8.81±0.06^b^ 29.06±0.15^c^ 98.02±0.89^d^ 1.58±0.02^b^ Endosulfan + Vitamin E 7.45±0.06^bc^ 9.10±0.05^c^ 25.99±0.10^b^ 71.83±0.83^b^ 1.23±0.04^c^ Values bearing different superscripts in a row differ significantly (*P*\<0.05) ::: Our studies comprise a part of comparative toxicology studies aimed to identify the biochemical and physiological alterations in RBCs exposed to two different pesticides. The approach is known to help in understanding the mechanisms of toxic action due to xenobiotics.\[[@CIT18]\] There are several pathways by which the pesticide is thought to induce oxidative stress. It inhibits the mitochondrial electron-transfer chain reaction, leading to accumulation of semi ubiquitous, which enables it to transfer one electron (e^--^) to molecular oxygen to form superoxide radicals.\[[@CIT19]\] Further, it may also interfere with cellular antioxidant defense system via alteration in activities of antioxidant enzymes, viz., SOD and CAT and status of GSH.\[[@CIT20]\] LPO level in rat erythrocytes treated with vitamin E was comparable to that of control, suggesting that the pesticides (endosulfan and chlorphyrifos) act as catalysts in the oxidative deterioration of biological macromolecules and this effect could be minimized by treatment with antioxidants. These indirectly suggest an increased production of oxygen free radicals in erythrocytes. Highly reactive oxygen metabolites, especially hydroxyl radicals, act on unsaturated fatty acids of phospholipid components of membranes to produce MDA, a lipid peroxidation product. Chlorpyrifos has been reported to induce oxidative stress, as shown by enhanced MDA production.\[[@CIT21]\] The use of vitamin E in conjunction with chlorpyrifos affected such an elevation in the level of MDA, bringing it within the normal limits. The normalization of LPO following vitamin E treatment is very likely due to its antioxidant properties. Our results revealed that endosulfan and chlorpyrifos caused a statistically significant decrease (*P*\<0.05) in SOD activity in rat erythrocytes compared to the control value. Supplementation of vitamin E to endosulfan and chlorpyrifos treated groups of rat erythrocytes normalized the levels of SOD. Treatment with vitamin E alone did not result in significant alteration in SOD activity compared to control treatment. The decrease in the activity of SOD in chlorpyrifos-intoxicated animals may be attributed to the consumption of this enzyme in converting O^2--^ to H~2~O. The dismutation of O^2--^ to H~2~O is catalyzed by SOD which contains both copper and zinc. In comparison to the control group, the activity of GST was significantly (*P*\<0.01) higher in chlorpyrifos treated rat erythrocytes. Considering that GSTs are detoxifying enzymes that catalyze the conjugation of a variety of electrophilic substrates to the thiol group of GSH, producing less toxic forms, the significant increase of GST activity in the rat erythrocytes after exposure to endosulfan and chlorpyrifos may indicate sufficient detoxification of pesticide in rat erythrocytes while the use of vitamin E with pesticide aproaches the control group. CAT is ubiquitously present in a wide range of aerobic cell types, with the highest activities in mammals being found in liver, kidney and RBCs. Endosulfan and chlorpyrifos caused significant decrease in CAT activity in erythrocytes of rats in this study. In comparison, vitamin E with endosulfan and chlorpyrifos treated erythrocytes maintained the levels of CAT at the normal values. In conclusion, treatment with vitamin E potentially reduced the free radicals in erythrocytes and ameliorated the oxidative stress as evidenced from lower concentrations of LPOs and GST and higher activities of SOD and CAT in erythrocytes. The efficacy of vitamin E in ameliorating pesticide-induced oxidative stress was higher for chlorpyrifos than for endosulfan. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.398188

2011-01-01

PMC3052593

INTRODUCTION {#sec1-1} ============ *Mimosa invisa* is a shrubby herbaceous plant, which is widespread in central and southern parts of Kerala. Toxicity due to consumption of this plant is very common in Kerala during rainy season. The earlier reports by Rajan *et al*,\[[@CIT1]\] indicated the toxic symptoms and pathological changes produced by this plant in calves. Detailed studies about the toxicity of this plant are lacking. Hence, an experiment was designed to study the toxicity produced by selected pooled fraction of *M. invisa*. An attempt has been made to protect the body from the toxicity with a decoction containing *Boerhaavia diffusa, Hygrophila auriculata* and *Tribulus terrestris*. MATERIALS AND METHODS {#sec1-2} ===================== Alcoholic extract of *M. invisa* was prepared from the dried pulverized plant material. Alcoholic extract was again fractionated in chromatographic columns so that four fractions, namely, chloroform fraction, *n*-butanol fraction, aqueous fraction and water insoluble residue were separated. The pilot studies conducted revealed that the *n*-butanol fraction and aqueous fraction were toxic. Hence, these two fractions were pooled and the toxicity of this pooled toxic fraction was studied in this experiment. Eighteen adult rabbits procured from Small Animal Breeding Station, Kerala Agricultural University, were used for the study. The animals were divided into three groups of six animals each. Group I animals served as control, group II was administered with pooled toxic fraction (0.4 g/kg) of *M. invisa* and group III received pooled toxic fraction (0.4 g/kg) followed by a decoction containing *H. auriculata, T. terrestris* and *B. diffusa* (5 g/kg). Blood was collected before the administration of extract and at day 1, 3, 5, 10, 15 and 20. The serum was separated and biochemical parameters like alanine transaminase (ALT), aspartate transaminase (AST), gamma glutamyl tranferase (GGT), creatine kinase (CK), alkaline phosphatase (ALP), creatinine and urea were estimated. The analysis was conducted using Ecoline Kits manufactured by E. Merck Limited (India), in a semiautomatic blood analyzer (Microlab 200). The data were analyzed statistically by using "*t*" test.\[[@CIT2]\] All the animals were sacrificed at the end of the experiment (20^th^ day) and observed for gross pathological lesions in liver, kidney and heart. These tissues were preserved in 10% formalin for histopathological examination. RESULTS AND DISCUSSION {#sec1-3} ====================== The results are presented in [Table 1](#T0001){ref-type="table"}. Group II rabbits, which received the pooled toxic fraction of *M. invisa*, showed clinical symptoms like inappetence, dullness, lethargy and loss of body condition. These symptoms were observed from the second day onward. Similar symptoms were reported by Alex *et al*,\[[@CIT3]\] in heifers with *Mimosa* toxicity. Alikutty and Pillai\[[@CIT4]\] also reported reduced appetite in a clinical case of *Mimosa* poisoning in a buffalo. Group III (pooled toxic fraction + decoction) also showed similar symptoms from second day onward, but from the fifth day, they started taking feed and water and became very active gradually. This may be due to the effect of decoction, which reversed the changes produced by the toxin. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Biochemical parameters before and after administration of pooled toxic fraction of *M. invisa* (0.4 g/kg) and decoction (5 g/kg) ::: Parameter Group Day ------------ ------- -------------- ----------------------------------------------- -------------------------------------------------- -------------------------------------------------- -------------------------------------------------- ------------------------------------------------- ------------------------------------------------- ALT I 54.67±4.62 56±4.57 56±3.61 56.67±3.5 55.67±3.34 56.5±3.91 55.83±3.63 II 55.5±5.69 85.5±8.05[\*](#T000F1){ref-type="table-fn"} 126.5±10.55[\*\*](#T000F2){ref-type="table-fn"} 136±13.06[\*\*](#T000F2){ref-type="table-fn"} 150.17±12.81[\*\*](#T000F2){ref-type="table-fn"} 162.67±8.09[\*\*](#T000F2){ref-type="table-fn"} 170.5±5.54[\*\*](#T000F2){ref-type="table-fn"} III 37±2.72 110±3.6 74.83±3.64[\*](#T000F1){ref-type="table-fn"} 65.67±4.51[\*](#T000F1){ref-type="table-fn"} 49.83±0.6 42±0.82 39±2.02 AST I 50.17±3.11 49.5±2.58 49.83±3.0 50.5±3.21 50.83±2.44 51.17±2.65 50.5±2.74 II 51±3.49 110±8.14[\*\*](#T000F2){ref-type="table-fn"} 154.17±15.08[\*\*](#T000F2){ref-type="table-fn"} 167.67±17.45[\*\*](#T000F2){ref-type="table-fn"} 188.5±11.65[\*\*](#T000F2){ref-type="table-fn"} 174.5±5.37[\*\*](#T000F2){ref-type="table-fn"} 186.83±5.78[\*\*](#T000F2){ref-type="table-fn"} III 45.17±1.20 81.67±2.68[\*](#T000F1){ref-type="table-fn"} 82±3.48[\*](#T000F1){ref-type="table-fn"} 66±2.10 57.83±2.37 42.17±1.33 42.83±1.91 GGT I 5.5±0.56 5.67±0.42 5.5±0.43 5.5±0.43 5.17±0.31 5.17±0.48 5.33±0.56 II 5.5±0.56 6.67±0.42 8.17±0.60[\*\*](#T000F2){ref-type="table-fn"} 9.83±0.87[\*\*](#T000F2){ref-type="table-fn"} 10.67±0.33[\*\*](#T000F2){ref-type="table-fn"} 11.5±0.22[\*\*](#T000F2){ref-type="table-fn"} 11.0±0.26[\*\*](#T000F2){ref-type="table-fn"} III 4.83±0.40 5.0±0.52 5.0±0.45 6.5±0.34 6.17±0.48 6.83±0.31 4.5±0.34 CK I 132.67±7.0 134.5±6.39 129.67±6.63 135±6.64 132±7.89 133.5±5.6 131.5±7.0 II 133.83±10.11 170.67±15.6[\*](#T000F1){ref-type="table-fn"} 199.33±13.76[\*](#T000F1){ref-type="table-fn"} 215.33±15.81[\*\*](#T000F2){ref-type="table-fn"} 170.33±6.98[\*\*](#T000F2){ref-type="table-fn"} 161.5±3.28[\*\*](#T000F2){ref-type="table-fn"} 140.5±3.43 III 199.67±6.04 211±3.99 201.5±4.2 204±4.54 197.5±4.9 200.5±4.17 200.83±5.67 ALP I 43.5±9.19 44.17±8.45 43.33±8.48 44±8.02 44±8.71 43.5±8.53 45.33±8.72 II 42.83±10.64 46.5±11.05 62.17±12.25[\*](#T000F1){ref-type="table-fn"} 79.17±14.58[\*](#T000F1){ref-type="table-fn"} 63.67±10.63[\*](#T000F1){ref-type="table-fn"} 64.5±9.59[\*](#T000F1){ref-type="table-fn"} 53.67±8.19 III 50.67±5.18 84±4.49 53.83±2.3 52.5±2.05 53.5±2.05 47.5±2.26 47.17±1.66 Creatinine I 2.5±0.22 2.33±0.21 2.67±0.21 2.50±0.22 2.67±0.21 2.67±0.21 2.5±0.22 II 2.17±0.31 4.67±0.33[\*](#T000F1){ref-type="table-fn"} 7.17±0.40[\*\*](#T000F2){ref-type="table-fn"} 8.33±0.49[\*\*](#T000F2){ref-type="table-fn"} 9.83±0.31[\*\*](#T000F2){ref-type="table-fn"} 10.80±0.40[\*\*](#T000F2){ref-type="table-fn"} 12.17±0.31[\*\*](#T000F2){ref-type="table-fn"} III 2.67±0.21 4.5±0.22[\*](#T000F1){ref-type="table-fn"} 5±0.26[\*](#T000F1){ref-type="table-fn"} 5.5±0.22[\*](#T000F1){ref-type="table-fn"} 4.83±0.17[\*](#T000F1){ref-type="table-fn"} 4.67±0.21[\*](#T000F1){ref-type="table-fn"} 2.5±0.22 Urea I 59.83±2.87 58.83±2.41 59±2.41 58.33±2.73 58.67±2.93 58.33±2.90 60.00±2.92 II 59.17±3.17 64.33±3.9 142.83±26.2[\*\*](#T000F2){ref-type="table-fn"} 157.33±23.92[\*\*](#T000F2){ref-type="table-fn"} 175.67±12.07[\*\*](#T000F2){ref-type="table-fn"} 195.83±8.99[\*\*](#T000F2){ref-type="table-fn"} 217.0±10.96[\*\*](#T000F2){ref-type="table-fn"} III 42.83±1.54 71.83±4.71[\*](#T000F1){ref-type="table-fn"} 69.17±2.12[\*](#T000F1){ref-type="table-fn"} 67.5±2.36[\*](#T000F1){ref-type="table-fn"} 63±1.61 55.35±3.48 47±5.01 \* *P*\<0.05; \*\* *P*\<0.01; *n* = 6; ALT, alanine transaminase; AST, aspartate transaminase; GGT, gamma glutamyl tranferase; CK, creatine kinase; ALP, alkaline phosphatase ::: There were significant changes in biochemical parameters in group II when compared with control animals (group I). A very significant (*P*\<0.01) increase in serum ALT and AST was observed from the very next day after the administration of pooled toxic fraction of *M. invisa* in group II. Similar changes were observed with other plants. Flaoyen *et al*,\[[@CIT5]\] reported increased ALT levels after feeding flower stem of *Narthecium ossifragum* in goats. Flaoyen *et al*,\[[@CIT6]\] observed an increase in AST levels in sheep during experimental toxicity with *N. ossifragum*. According to Burtis and Ashwood,\[[@CIT7]\] in liver diseases associated with hepatic necrosis, the levels of ALT and AST rise even before the start of clinical symptoms. Hence, the increase in ALT and AST in the present study may be attributed to liver damage. In group III, a transient increase in AST and ALT levels followed by a gradual decrease was observed. Hewawasam *et al*,\[[@CIT8]\] observed a decrease in serum ALT and AST levels while studying the protective effect of *Hygrophila spinosa* extract in mouse liver injury induced by carbon tetrachloride and paracetamol. They suggested that this might be due to reduced leakage of enzymes from hepatocytes. The GGT levels showed a significant increase in group II. Craig *et al*,\[[@CIT9]\] reported a continuous increase in GGT levels in toxicity due to pyrrolizzidine alkaloids. The group III animals exhibited considerable improvement in GGT levels when compared with group II animals. The CK levels showed an increase followed by a decrease in Group II animals. Burtis and Ashwood (1996) explained that this may be due to muscular dystrophies. In group III animals, there was no significant increase in CK levels. There was significant increase in ALP values throughout the experiment in group II animals, whereas group III animals showed a transient increase followed by normalcy. It may be due to hepatoprotective effect of *H. auriculata* as reported by Hewawasam *et al*,\[[@CIT8]\] The biochemical parameters, serum creatinine and urea showed a significant increase throughout the experiment in group II animals, and group III animals exhibited a decrease when compared to Group II animals. The rapid increase in urea and creatinine indicates typical impairment of kidney function.\[[@CIT10]\] The decrease in creatinine and urea observed in group III animals may be due to nephroprotective effect of *H. auriculata, T. terrestris* and *B. diffusa*. Gross pathological lesions include necrotic patches and petechial hemorrhages in the kidney and severe congestion and necrosis of the liver. Heart did not show any gross lesions. The histopathological examination of tissue collected from group II animals revealed tubular dilatation, diffuse tubular degeneration, shrunken glomeruli in kidney \[[Figure 1](#F0001){ref-type="fig"}\], fatty change and necrosis in liver \[[Figure 2](#F0002){ref-type="fig"}\] and focal myolysis and diffuse hyalinization in heart \[[Figure 3](#F0003){ref-type="fig"}\]. Postmortem examination of group III animals did not show any gross lesions; however, histopathological examination revealed regenerative changes \[Figures [4](#F0004){ref-type="fig"}--[6](#F0006){ref-type="fig"}\]. ::: {#F0001 .fig} Figure 1 ::: {.caption} ###### Kidney -- pooled fraction -- tubular dilatation, diffuse tubular degeneration, shrunken glomeruli (H and E, ×100) :::  ::: ::: {#F0002 .fig} Figure 2 ::: {.caption} ###### Liver -- pooled fraction -- fatty change, necrosis (H and E, ×400) :::  ::: ::: {#F0003 .fig} Figure 3 ::: {.caption} ###### Heart -- pooled fraction -- focal myolysis, diffuse hyalinization (H and E, ×100) :::  ::: ::: {#F0004 .fig} Figure 4 ::: {.caption} ###### Kidney -- pooled fraction + decoction -- tubules with intact normal lining cells (H and E, ×100) :::  ::: ::: {#F0005 .fig} Figure 5 ::: {.caption} ###### Liver -- pooled fraction + decoction -- sinusoidal dilatation, regenerative changes (H and E, ×400) :::  ::: ::: {#F0006 .fig} Figure 6 ::: {.caption} ###### Heart -- pooled fraction + decoction -- intact heart muscle fibers (H and E, ×100) :::  ::: CONCLUSION {#sec1-4} ========== The results of the study reveal that the pooled fraction of *M. invisa* has the potential to produce nephrotoxic and hepatotoxic effects in rabbits. Administration of the decoction containing *H. auriculata, T. terrestris* and *B. diffusa* could protect the liver and kidney as evinced by the biochemical parameters and histopathological changes. **Source of Support:** Nil **Conflict of Interest:** None declared

PubMed Central

2024-06-05T04:04:19.399688

2011-01-01

PMC3052594

INTRODUCTION {#sec1-1} ============ Chlorpyrifos, a non-systemic broad-spectrum organophosphate insecticide, is used for the control of a large number of insect pests of various crops. It is a cholinesterase inhibitor.\[[@CIT1]\] Since *Ocimum sanctum* leaf extract has time tested healing value in the traditional Indian medicinal system, it was thought that it would be interesting to know if it has a genoprotective effect against aberrations induced by chlorpyrifos in *in vivo* mouse system. There is a continued interest and need to identify and develop non-toxic genoprotective compounds. An efficient genoprotectant could prove useful in occupational and therapeutic settings where genotoxic chemicals are used or where exposure occurs. *O. sanctum*, commonly called *"Tulsi"* (Family Labiateae), is easily available in the whole of tropical and subtropical India. It is held sacred by Hindus, and various parts of the plant have been traditionally used in *Ayurveda* and *Siddha* systems of medicine for the treatment of diverse hepatic disorders, cold, cough and as an antidote for snakebite.\[[@CIT2]\] It has also been reported to have anticarcinogenic activity,\[[@CIT3]\] as well as radioprotective effects.\[[@CIT4]\] Flavonoids isolated from *O. sanctum* scavenged the free radicals *in vitro* and showed antilipoperoxidant activity *in vivo* at a very low concentration.\[[@CIT5]\] It is well known that pesticides are genotoxic to experimental subjects (rat/mice) and have been shown to cause the same effects in human subjects also.\[[@CIT6]\] The extensive application of pesticides in modern agriculture requires an intensive investigation of the impact of these chemicals on the environment and public health. With the dispersal of hundreds of millions of kilograms each year, these agents must be analyzed for their mutagenic properties. Therefore, looking at the extensive application of chlorpyrifos and the possible genoprotective role of *O. sanctum*, it was considered worthwhile to undertake this study. Thus, the present study was undertaken to investigate the genoprotective effect of *O. sanctum* extract (OE) on mitotic index (MI) and chromosomal aberration (CA) percentage in bone marrow cells of rats induced with 1/15 and 1/5 LD~50~ of chlorpyrifos. Our goal was also to evaluate the cytogenetic effects of single exposure to chlorpyrifos although pesticide sprayers receive a chronic exposure to this commonly used pesticide. MATERIALS AND METHODS {#sec1-2} ===================== Comet assay to assess DNA damage {#sec2-1} -------------------------------- Comet assay or single cell gel electrophoresis assay was used for the evaluation of DNA damage in individual cells.\[[@CIT7]\] The experimental sample consisted of blood subjected to chlorpyrifos at 75 *μ*g/ml (the concentration is one-fourth of what is sprayed by farmers in the field) for 2 hours at 37°C. Blood sample treated with 40 mM H~2~O~2~ for 10 min at room temperature formed the positive control. The negative control was untreated blood. A small number of cells are immersed in agarose gel, lysed, subjected to an electrophoretic field and then stained with silver stain.\[[@CIT8]\] The assay was run in triplicate for experimental and control samples. Fifty cells for each sample were scored for DNA damage visually under the light microscope and were classified into six categories\[[@CIT9]\] as shown below: Category A: Undamaged cellsCategories B-E: Cells with progressively greater DNA damageCategory F: Apoptotic cells *in vitro* lymphocyte culture and *in vivo* studies {#sec2-2} --------------------------------------------------- ### Preparation of OE {#sec3-1} Fresh leaves of *O. sanctum*, collected locally, were air dried, powdered and extracted with 50% ethyl alcohol and 50% distilled water in a soxhlet apparatus by refluxing for 68 hours (at 4 hours/day for 17 days) at 60°C. The extract was evaporated to obtain it in a powder form. For oral administration, the extract was constituted in 0.5 ml distilled water and administered at 50 mg of extract/kg b.wt. to rats, since this dose of OE gave protection against radiation injury.\[[@CIT10]\] The dose of the insecticide was calculated as 1/15 and 1/5 of the recommended LD~50~(135 mg/kg b.wt.) for rats. Group 1 -- control, untreated ratsGroup 2 -- treated with 1/5 LD~50~ of chlorpyrifosGroup 3 -- pretreated with OE and i.p. injection of 1/5 LD~50~ of chlorpyrifosGroup 4 -- treated with 1/15 LD~50~ of chlorpyrifosGroup 5 -- pretreated with OE and i.p. injection of 1/15 LD~50~ of chlorpyrifos ### Animals {#sec3-2} Experiments were conducted on albino rats weighing approximately 75 g. The rats were acclimatized to the laboratory conditions for 48 hours. They were maintained on standard rat feed and water *ad libitum*. The experimental group for each treatment consisted of six animals. The experiments were conducted according to the recommendations of the institutional ethical committee. The first treatment (1/5 of LD~50~) consisted of an i.p. injection of chlorpyrifos at 27 mg/kg b.wt. The animal was sacrificed 24 hours later and bone marrow chromosome preparations were made as per the standard hypotonic/air drying/Giemsa technique. For the second treatment, rats were fed OE at 50 mg/kg per day for 21 days and thereafter given an i.p. injection of 1/5 LD~50~ of chlorpyrifos. The rats were sacrificed 24 hours later and bone marrow preparations were made in the usual manner. The same experimental schedule was followed for 1/15 LD~50~ of chlorpyrifos. For the controls, the rats were given 0.5 ml distilled water orally for 21 days. ### Human lymphocyte culture {#sec3-3} The assessment of the genoprotective role of OE was also carried out *in vitro* in cultured human lymphocytes. The chromosome preparations were made from peripheral blood cultures following the method of Moorhead *et al*.\[[@CIT11]\] As a first step, the CA percentage was assessed using 75 *μ*g/ml. In the other samples, *O. sanctum* extract was added at zero hour at 12 *μ*g/ml and chlorpyrifos was added after 48 hours to the culture at 75 *μ*g/ml. Experiment 1: Forty-eight hours after setting up the cultures, the lymphocytes were treated with chlorpyrifos at 75 *μ*g/ml of culture.Experiment 2: Lymphocytes were maintained in culture for 72 hours with OE (added at zero hour).Experiment 3: Forty-eight hours after setting up the cultures, already treated with OE (at zero hour), the lymphocytes were treated with chlorpyrifos at 75 *μ*g/ml of culture, allowing the cells to be in contact with OE for two cell cycles.Untreated (control) 4: Lymphocytes maintained in culture for 72 hours received only distilled water. CAs were scored under a light microscope at a magnification of 100×. Two hundred metaphase plates were examined per treatment (for *in vivo* studies). Different types of aberrations, such as chromatid breaks, chromosome breaks, fragments and numerical aberrations, were scored to give the total CA percentage for each treatment. Statistical analysis {#sec2-3} -------------------- The data were analyzed using student's *t*-test. RESULTS AND DISCUSSION {#sec1-3} ====================== Comet assay to assess DNA damage {#sec2-4} -------------------------------- The assay was run in triplicate for experimental and control samples. Fifty cells for each sample were scored for DNA damage visually under the light microscope and were classified into six categories as shown in [Table 1](#T0001){ref-type="table"}. ::: {#T0001 .table-wrap} Table 1 ::: {.caption} ###### Percentage of cells in different categories of DNA damage ::: Cells (%) in different categories of DNA damage -------------------------------- ------------------------------------------------- ------------ ------------ ----------- ----------- ----------- Negative control (non-treated) 79.00±2.00 7.00±1.00 5.00±0.33 5.00±0.5 2.00±0.30 2.00±0.50 Treated with chlorpyrifos 63.00±3.50 12.00±1.50 12.00±2.50 5.00±1.00 6.00±0.50 2.00±0.30 ::: *in vivo* study {#sec2-5} --------------- ### a) Analysis of mitotic index {#sec3-4} For investigation of mitotic index (MI), 3000 cells for each treatment were scored. In the first treatment (1/5 LD~50~), the mean value of MI% in untreated controls was 5.46±0.88, which fell to 1.28±0.52 in animals treated with chlorpyrifos only. The depression caused in MI% by chlorpyrifos treatment was significant (*P*\<0.001) as compared to the control values. There was a significant (*P*\<0.01) increase in MI% (2.65±0.45) in animals that were given oral OE and 1/5 LD~50~of chlorpyrifos than the MI value of only chlorpyrifos treated rats (1.28±0.52). In the second treatment (1/15 LD~50~), the mean value of MI% in untreated controls was 5.46±0.88, which fell to 3.88±0.35 in animals treated with chlorpyrifos only. The depression caused in MI% by chlorpyrifos treatment was significant (*P*\<0.01) as compared to the control values. There was a significant (*P*\<0.05) increase in MI% value (4.30±0.30) in animals that were given oral OE and 1/15 LD~50~of chlorpyrifos than the MI value of only chlorpyrifos treated rats. A significant (*P*\<0.05) decrease in MI% value was found in rats treated with 1/5 of LD~50~of chlorpyrifos when compared to those treated with 1/15 LD~50~ of the pesticide \[[Table 2](#T0002){ref-type="table"}\]. ::: {#T0002 .table-wrap} Table 2 ::: {.caption} ###### Mitotic index in bone marrow of different groups of rats ::: Treatment MI% -------------------------- --------------------------------------------- Control 5.46±0.88 Chlor (1/5 LD~50~) 1.28±0.52[\*](#T000F1){ref-type="table-fn"} Chlor (1/5 LD~50~) + OE 2.65±0.45[Ψ](#T000F2){ref-type="table-fn"} Chlor (1/15 LD~50~) 3.88±0.35[†](#T000F3){ref-type="table-fn"} Chlor (1/15 LD~50~) + OE 4.30±0.30[ф](#T000F4){ref-type="table-fn"} \* *P*\<0.001 (comparison between control and chlor 1/5 LD~50~); Ψ *P*\<0.01 (comparison between chlor 1/5 LD~50~ and chlor 1/5 LD~50~ + OE); † *P*\<0.01 (comparison between control and chlor 1/15 LD~50~); \* *P*\<0.05 (comparison between chlor 1/15 LD~50~ and chlor 1/15 LD~50~ + OE) ::: ### b) Analysis of chromosomal aberrations {#sec3-5} For the investigation of chromosomal aberrations (CA%), 200 cells for each treatment were scored. In the first treatment (1/5 LD~50~), the mean value of CA for controls was 2.0±0.9 and the value was 27.5±3.0 for animals treated with only chlorpyrifos. Thus, there was a significant increase (*P*\<0.001) in the frequency of aberrant cells in bone marrow of rats treated with 1/5 LD~50~ of chlorpyrifos as compared to the control values. However, in the animals pretreated with OE, there was a significant (*P*\<0.05) decrease in the frequency of aberrant cells (mean CA% 16.0±1.0) as compared to the chlorpyrifos treated rats. In the second treatment (1/15 LD~50~), the mean value of CA for controls was 2.0±0.9 and it was 3.5±0.30 for animals treated with only chlorpyrifos. However, there was an increase in the frequency of aberrant cells in bone marrow of rats treated with 1/15 LD~50~ of chlorpyrifos but the increase was not significant. In the animals pretreated with OE, there was a decrease in the frequency of aberrant cells (mean CA% 2.5±0.8) as compared to the chlorpyrifos treated rats \[[Table 3](#T0003){ref-type="table"}\]. ::: {#T0003 .table-wrap} Table 3 ::: {.caption} ###### Frequency of chromosomal aberration in different groups of rats ::: Group (treatment) Number of plates Ctb Aneu Hap Poly Ct-const Ctg PCD Total CA% -------------------------- ------------------ ----- ------ ----- ------ ---------- ----- ----- ---------------------------------------------- Control 200 -- 01 -- -- -- 02 01 02.0±0.9 Only 1/5 LD~50~ chlor 200 01 25 09 6 11 02 01 27.5±3.0[\*](#T000F5){ref-type="table-fn"} OE and 1/5 LD~50~ chlor 200 01 07 04 01 01 6 12 16.0±1.0[\*\*](#T000F6){ref-type="table-fn"} Only 1/15 LD~50~ chlor 200 -- 1 02 1 2 1 -- 03.5±0.3 OE and 1/15 LD~50~ chlor 200 -- 02 03 -- -- -- -- 02.5±0.8 Ctb, chromatid break; Aneu, aneuploidy; Hap, haploidy; Poly, polyploidy, Ct-const, chromatid constriction; PCD, precocious centromere dissociation; Ctg, chromatid gap; CA, chromosomal aberrations; \* *P*\<0.001; \*\* *P*\<0.05 ::: *in vitro* study {#sec2-6} ---------------- A significant increase (*P*\<0.05) in CAs was observed in lymphocytes treated with chlorpyrifos and a statistically significant (*P*\<0.05) decrease was found in cultures pretreated with OE. MI decreased significantly (*P*\<0.05) in cultures treated with chlorpyrifos and a slight increase in MI was found in OE pretreated lymphocytes but the increase was not statistically significant \[[Table 4](#T0004){ref-type="table"}\]. ::: {#T0004 .table-wrap} Table 4 ::: {.caption} ###### Chromosomal aberrations and mitotic index of cultures pretreated with *Ocimum sanctum* extract ::: Treatment CA% (mean±SE) MI% (mean±SE) ----------------------------------------- ----------------------------------------------- -------------------------------------------- Control 1.78±0.17 3.75±0.38 Chlorpyrifos (75 μg/ml) 5.52±0.95[\*](#T000F7){ref-type="table-fn"} 2.32±0.16[Ψ](#T000F9){ref-type="table-fn"} OE (12 μg/ml) + chlorpyrifos (75 μg/ml) 2.41±0.18[\*\*](#T000F8){ref-type="table-fn"} 2.88±0.35 \* *P*\<0.05 (statistical difference between control and chlorpyrifos); \*\* *P*\<0.05 (statistical difference between chlorpyrifos and OE + chlorpyrifos); Ψ *P*\<0.05 (statistical difference between control and chlorpyrifos); CA, chromosomal aberrations; MI, mitotic index ::: [Table 1](#T0001){ref-type="table"}shows that in non-treated samples, cells of category A (zero damage) formed 79% of the total cells counted, whereas in the blood samples treated with 75 *μ*g/ml of chlorpyrifos, cells of this category were 63%. The chlorpyrifos treated samples showed 35% cells with DNA damage ranging from categories B-E, whereas in non-treated samples only 19% cells fell in this category (*P*\>0.05). Thus, the study showed that 75 *μ*g/ml of chlorpyrifos caused statistically nonsignificant damage to DNA as determined by comet assay. There seems to be no other reference available on such assessment of genotoxicity of chlorpyrifos. However, similar work has been done on some other pesticides and herbicides which are outlined below. Assessment of genotoxic effects of chlorpyrifos and acephate by comet assay in mice leukocytes was done by Rahman *et al*.\[[@CIT12]\] Evaluation of herbicide-induced DNA damage in human lymphocytes by comet assay was done by Ribas *et al*,\[[@CIT13]\] They found that alachlor, atrazine, maleic hydrazide, paraquat and trifluralin gave positive results for genotoxicity by increasing the comet tail length. Chlorpyrifos-induced DNA damage in rat liver and brain cells was assessed through comet assay by Mehta *et al*.,\[[@CIT14]\] who classified the DNA damage in various classes from zero to four. DNA damaging effects of pesticides were measured by comet assay and CAs in Chinese hamster ovary cells by Vigreux *et al*,\[[@CIT15]\] and they found that chlorothalonil was toxic to CHOK1 cells but carbendazim did not induce DNA strand breaks in comet assay. Occupational exposure of workers employed in pesticide production was found to cause an increase in mean tail length of comet in a study by Paramjit *et al*.\[[@CIT16]\] In the classification of pesticides, chlorpyrifos is classified as a "moderately hazardous pesticide" by World Health Organization (WHO).\[[@CIT1]\] Our results indicate a moderate toxicity of chlorpyrifos at a concentration of 75 *μ*g/ml of blood, which agrees with the WHO data. This study was also aimed to evaluate the genoprotective effects of *O. sanctum* on chlorpyrifos-induced genotoxicity. The use of pesticides has become a routine mainly in underdeveloped countries, but the genotoxic potential of these substances is not yet well established.\[[@CIT17]\] Most of the farmers responsible for the application are at risk for cytotoxicity and genotoxicity. Awa *et al*,\[[@CIT18]\] detected a positive correlation between the risk of genetic diseases in populations and the level of cytogenetic damage, whereas Au *et al*,\[[@CIT19]\] hypothesized that CAs were in the background of carcinogenesis and that the determination of their incidence was an important parameter for the effect of various agents on the health status of mammals and man. Thus, the increased frequency of CAs is related to higher risk of development of malignancies. Maximum number of aneuploidy cells was observed in mice treated with 1/5 of LD~50~ of chlorpyrifos and the number significantly decreased in mice pretreated with OE. There was a total absence of metaphase plates showing precocious centromeric separation in rats pretreated with OE; perhaps the flavonoids and other active components help in polymerization of spindle fibers so that all the cell divisions are in phase. Radiation and chemical toxins produce biological damage by forming reactive oxygen species like singlet oxygen and superoxides, hydroxyl and hydroperoxy radicals, hydrogen peroxide and organic peroxides.\[[@CIT20]\] The genoprotective effect of *O. sanctum* is associated with the presence of its flavonoids, such as orientin and vicenin, which take part in scavenging reactive intermediates that are capable of binding to proteins and DNA.\[[@CIT5]\] Chlorpyrifos was found to increase the activities of superoxide dismutase, glutathione peroxidase and catalase. Melatonin causes decrease in the above enzymes and an increase in thiobarbituric acid reactive substances. Production of reactive oxygen species could be a cause of DNA damage. *in vitro* and *in vivo* generation of reactive oxygen species, DNA damage and lactate dehydrogenase (LDH) leakage by selected pesticides was studied by Bagchi *et al*.\[[@CIT21]\] According to them, brain lipid peroxidation and DNA single strand breaks are two indices of oxidative stress and oxidative tissue damage. Thus, the most likely mechanism of DNA damage and chromosome breakage by chlorpyrifos seems to be through the production of reactive oxygen species, and the present investigation indicates that the pre-treatment of rats with OE at 50 mg/kg per day for 21 days has a significant (*P*\<0.001) positive effect in the MI depression caused by chlorpyrifos. *O. sanctum* also had a genoprotective effect on the CA%. It was also found out that OE caused a significant decrease in CA% in *in vitro* lymphocyte cultures. Thus, the genoprotective effect of *O. sanctum* was confirmed both *in vivo* and *in vitro*. **Source of Support:** Nil **Conflict of Interest:** None declared.

PubMed Central

2024-06-05T04:04:19.401080

2011-01-01

PMC3052595

Hepatitis C virus (HCV) is estimated to infect some 123 million individuals ([@r21]), and establishes a chronic infection that can ultimately result in liver fibrosis, cirrhosis or hepatocellular carcinoma. Combination therapy comprising pegylated alpha interferon (IFN-*α*) and ribavirin is only successful in approximately 50 % of patients. HCV, a member of the family *Flaviviridae*, is an enveloped virus with a positive-sense RNA genome 9.6 kb in length and containing a single ORF. An internal ribosome entry site mediates cap-independent translation of the 3000 aa polyprotein, which is cleaved co- and post-translationally by host-cell and viral proteases to release the structural proteins (core, E1, E2 and p7) and non-structural proteins (NS2, NS3, NS4A, NS4B, NS5A and NS5B). NS5A has been shown to have many functions: foremost, as a component of the RNA replication complex, it is absolutely required for viral RNA replication. Structural analysis has revealed that NS5A comprises three domains separated by short low-complexity regions ([@r23]) (Fig. [1](#f1){ref-type="fig"}, top). The structure of domain I has been determined; it coordinates a zinc ion and is postulated to dimerize ([@r10]; [@r24]). Domains II and III are less structured and more flexible; in particular, domain III can accommodate a GFP insert at the C terminus with no adverse effects ([@r2]; [@r15]; [@r16]). By SDS-PAGE and Western blotting, two forms of the protein with different apparent mobilities can be observed: these correspond to alternatively phosphorylated forms of NS5A -- a basally phosphorylated form (apparent molecular mass 56 kDa) and a hyperphosphorylated form (58 kDa). Proline-directed kinases such as casein kinase II (CKII) have been implicated in basal phosphorylation ([@r20]), and CKI*α* has been implicated in hyperphosphorylation ([@r19]); however, there is little consensus as to either the locations or the number of phosphorylation sites on both forms of NS5A. Basal phosphorylation sites have been shown to be present in both domains II and III ([@r22]) and hyperphosphorylation sites have also been mapped to domain II ([@r7]; [@r22]). Interestingly, inhibition of hyperphosphorylation either pharmacologically or by mutation enabled replication of a non-culture-adapted genotype 1b subgenomic replicon ([@r2]; [@r17]). In order to assign phosphorylation sites unambiguously within NS5A, we adopted a mass-spectrometric (MS) approach. We showed previously that a recombinant baculovirus containing a tetracycline-responsive mammalian promoter (a BacMAM vector termed FBrepcon1neo; [@r14]) could be used to drive high levels of expression of a Con1-derived genotype 1b NS3-5B subgenomic replicon in HepG2 cells. Importantly, HepG2 supported the replication of a subgenomic replicon delivered via the baculovirus route, as judged by induction of an IFN-*β* response that was not seen following delivery of a GND-mutant (replication-defective) replicon ([@r13]). Furthermore, NS5A expressed from this vector exhibited both basal and hyperphosphorylated species (p56 and p58) ([@r14]), similar to those observed in Huh7 cells stably harbouring subgenomic replicons. HepG2 cells were therefore transduced with a BacMAM vector expressing the tetracycline transactivator (BactTA) together with FBrepcon1neo; NS5A was purified from cell lysates by using protein G--agarose and polyclonal sheep anti-NS5A serum, separated by SDS-PAGE and stained with colloidal blue stain. The p58 band was excised and submitted for MS analysis. Peptides from a trypsin digest were analysed on two systems independently: firstly, a 4000QTRAP system, where precursor ion scanning was used to identify phosphorylated peptides (Fingerprints Proteomics Facility, University of Dundee, UK) (Fig. [1](#f1){ref-type="fig"}), and secondly, a Waters Synapt HDMS system (Astbury Centre for Structural Molecular Biology, University of Leeds, UK) (data not shown). Potential phosphorylated peptides were then identified by MS/MS. In both cases, only one major phosphorylated peptide was identified. This was HDpSPDADLIEANLLWR \[phospho-serine 249 (serine 2221 in polyprotein numbering)\], which is situated at the junction of the first low-complexity sequence and domain II. This site has not been identified previously as a phosphorylation site, although it is situated close to a cluster of serine residues that have been postulated to be involved in hyperphosphorylation (serines 2197, 2201 and 2204) ([@r22]). There are a number of reasons why other phosphorylated peptides were not identified, including inefficient ionization, low stoichiometry and failure of some peptides to be captured by the ion trap. Therefore, the failure to identify other phosphorylation sites does not mean that the BacMAM-expressed NS5A in HepG2 cells is not phosphorylated on other residues. However, the identification of phospho-serine 249 by using two different systems in independent experiments suggests that this is a major site of phosphorylation within the hyperphosphorylated form of NS5A. In order to assess the potential significance of phosphorylation at this residue for viral genomic replication, we generated two mutant forms of the FK5.1 culture-adapted subgenomic replicon ([@r8]) in which serine 249 was mutated to either an alanine (phosphoablative) or an aspartic acid (phosphomimetic) by PCR (oligonucleotide sequences are available upon request). We chose FK5.1 because the Con1 replicon is not culture-adapted and does not replicate at a sufficient level to allow discrimination between wild type and mutants with impaired replication efficiency. *In vitro* transcripts of the FK5.1 replicons were electroporated into Huh7 cells; cells were then selected for 3 weeks with 1 mg G418 ml^−1^, prior to either staining for colony-forming assays or generation of stable polyclonal replicon-harbouring cell lines. Neither mutation had any effect on either the number of colonies (Fig. [2a](#f2){ref-type="fig"}) or the p56 : p58 ratio (Fig. [2b](#f2){ref-type="fig"}). In order to exclude the possibility that the replicons had reverted to the wild-type sequence, RNA was extracted from cells, reverse-transcribed, amplified using PCR, cloned and sequenced (data not shown). All clones examined contained the original mutation and had thus not reverted to wild type. These data were also confirmed in the context of a transient, luciferase-based FK5.1 subgenomic replicon. *In vitro* transcripts of the FK5.1 luciferase replicons were electroporated into Huh7 cells, and luciferase activity was measured at both 4 h post-transfection (to assess transfection efficiency and translation of input RNA) and 72 h (to assess RNA replication) (Fig. [2c](#f2){ref-type="fig"}). Relative replication levels for both mutants were similar to those of the wild type. Serine 249 is almost completely conserved in all genotype 1 isolates of HCV; however, this residue is not conserved in other genotypes ([@r9]). Notably, the equivalent residue in the only isolate that has the capacity for replication in cell culture, JFH-1, is threonine. In addition to this, a serine is situated two residues N-terminally (serine 247) in JFH-1 NS5A (Fig. [3a](#f3){ref-type="fig"}). The data in Fig. [2](#f2){ref-type="fig"} demonstrated that phosphorylation of serine 249 was dispensable for RNA replication; therefore, to test whether the presence of a phosphorylatable residue at this position was important for the role of NS5A in virus assembly ([@r3]; [@r6]; [@r12]; [@r25]), we mutated either serine 247 or threonine 249 to either alanine or aspartic acid in the context of full-length JFH-1 virus ([@r27]), previously modified to contain synonymous mutations generating unique restriction sites flanking NS5A ([@r6]). Somewhat surprisingly, neither mutation had a substantial effect on either virus assembly or release (Fig. [3a](#f3){ref-type="fig"}). HCV protein synthesis in infected cells (Fig. [3b](#f3){ref-type="fig"}) and the NS5A p56 : p58 ratio were also unaffected. We conclude that phosphorylation of these residues is not important for either virus genome replication or particle assembly; however, we cannot rule out the possibility that, unlike Con1, in the context of FK5.1 or JFH-1 these residues are not phosphorylated. To test this, it would be necessary to express either the FK5.1 subgenomic replicon or full-length JFH-1 in HepG2 cells using a BacMAM vector and analyse the phosphorylation status of NS5A. Such experiments are under way in our laboratory. What, then, might be the role of phosphorylation of serine 249? It is possible that, in the context of the culture-adapted FK5.1 subgenomic replicon or the highly efficiently replicating JFH-1 genome, there is no requirement for phosphorylation of this residue. However, in the context of the non-culture-adapted infectious clones (e.g. H77, J4, Con1), the phosphorylation of serine 249 might be important in some aspect of NS5A function. Indeed, the acquisition of culture-adaptive mutations in NS5A has been shown to correlate with a loss of hyperphosphorylation, even when the mutations were not at serine residues ([@r4]). It will be of interest to investigate the potential role of serine 249 phosphorylation in the context of the infectious genotype 1b virus described recently ([@r18]). However, as acknowledged in that publication, the very low replicative capacity of this isolate in Huh7 cells will render such an investigation technically challenging. One attractive hypothesis that may be more tractable to investigation is the potential structural role of serine 249. As it is followed by a proline residue, phosphorylation of serine 249 might influence the recognition of this proline by peptidyl--prolyl isomerases such as cyclophilin A, a known NS5A-interacting partner that is important for viral RNA replication. The location of serine 249, precisely at the N terminus of domain II of NS5A, might allow phosphorylation to influence the spatial orientation of that domain with respect to domain I by mediating the *cis*--*trans* isomerization of the peptidyl--prolyl bond. Structural studies (e.g. NMR) on purified wild-type and mutant NS5A, phosphorylated *in vitro*, might shed light on this issue, but represent significant technical challenges. This work was supported by the Wellcome Trust (078358). We thank Volker Lohmann and Ralf Bartenschlager (University of Heidelberg, Germany) for pFK5.1, and Takaji Wakita (National Institute for Infectious Diseases, Tokyo, Japan) for pJFH-1. We also thank Dougie Lamont (Fingerprints Proteomics Facility, University of Dundee, UK) and James Ault (Astbury Centre for Structural Molecular Biology, University of Leeds, UK) for help with the MS analysis. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### MS identification of serine 249 phosphorylation. At the top, a schematic of the structure of NS5A is depicted, showing the location of the phosphopeptide. LCS, Low-complexity sequence. For MS analysis, HepG2 cells were seeded onto rat tail collagen-coated plates and transduced with BactTA and FBrepcon1neo at 800 p.f.u. per cell each, as described previously ([@r14]). Cells were lysed 24 h post-transduction and NS5A was purified by using protein G beads and a polyclonal sheep anti-NS5A serum ([@r11]). Following SDS-PAGE, the colloidal blue-stained band corresponding to p58 was excised and digested with trypsin. The tryptic digests were analysed by liquid chromatography--MS with precursor ion scanning on a 4000QTRAP system. Sums of the peptide masses detected by precursor ion scanning in the negative-ion mode across the 45 min HPLC separation are shown. The peptide was identified from the MS/MS fragmentation spectra by database searching and manual inspection of the spectra. amu, Atomic mass units. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Role of serine 249 phosphorylation in genotype 1b RNA replication. (a) Huh7 cells (4×10^6^) were electroporated with the indicated *in vitro*-transcribed FK5.1 RNAs (2 μg), selected in the presence of G418 (1 mg ml^−1^) for 3 weeks and colonies were stained with Coomassie brilliant blue. Number of colonies produced is presented as a percentage of the initial number of electroporated cells. (b) Western blot analysis of lysates from stable cell lines harbouring the indicated FK5.1 replicons with sheep polyclonal antiserum to either NS3 ([@r1]) or NS5A ([@r11]). (c) Huh7 cells were electroporated with replicon RNA and harvested into passive lysis buffer (Promega) at the indicated time points (black bars, 4 h; grey bars, 72 h). Luciferase activity (relative luciferase units; RLU) was measured as described previously ([@r11]). In both graphs, error bars show [sem]{.smallcaps}; data from three independent experiments are shown. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### Role of serine 247/threonine 249 phosphorylation in assembly and release of infectious HCV. (a) Sequence of the region surrounding serine 249 in Con1, J4 and JFH-1 NS5A. Mutated residues are shown in bold. Huh7 cells (4×10^6^) were electroporated with the indicated *in vitro*-transcribed virus RNAs (10 μg) and virus release into the culture supernatant (black bars) was measured by focus-forming assay ([@r6]). Intracellular virus titres (grey bars) were measured by focus-forming assay following cell disruption by repetitive freeze--thaw at 72 h post-transfection (p.t.). (b) Huh7 cells were electroporated with the indicated virus RNAs and harvested at 48 h p.t. by lysis in Glasgow lysis buffer ([@r5]). Protein (10 μg) was analysed by Western blotting with antiserum to NS5A or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as indicated. :::  ::: [^1]: †**Present address:** Leeds Institute of Molecular Medicine, Faculty of Medicine and Health, University of Leeds, Leeds LS9 7TF, UK.

PubMed Central

2024-06-05T04:04:19.403629

2010-10-01

PMC3052599

INTRODUCTION ============ Herpesviruses are widespread pathogens that use immune evasion to establish persistent infectivity in immunocompetent hosts ([@r51]). Most can be neutralized *in vitro*. However, herpesvirus carriers not only remain infectious, but can be superinfected ([@r22]) and show little tendency to select viral antigenic variants. Therefore in the context of natural infection, virus neutralization works poorly. Understanding why is important, as this could potentially reveal new means of infection control. Studies of gammaherpesvirus neutralization have concentrated on Epstein--Barr virus (EBV). Antibodies to gp350 can stop EBV infecting B cells ([@r46]). However, the gp350-specific antibodies of EBV carriers do not render them non-infectious, and vaccination with gp350 failed to reduce the incidence of EBV infection ([@r40]). This could reflect gp350-independent host entry ([@r25]), which gp350-specific antibodies might even enhance ([@r47]). Neutralization studies of both EBV and the Kaposi\'s sarcoma-associated herpesvirus (KSHV) have been hampered by the difficulty of infecting cells at more than a very low level. Thus, an anti-gH/gL mAb can block EBV infection of epithelial cells, an anti-gp42 mAb can block EBV infection of B cells ([@r27]), and KSHV can be neutralized by sera raised against the whole virus ([@r8]), gH/gL ([@r31]), gB ([@r2]) or K8.1 ([@r38]), but much of this neutralization has been weak (\<threefold reduction in infection) in settings where infection may be suboptimal -- for example, EBV infects epithelial cells poorly *in vitro*. Its quantitative significance is therefore hard to assess. In contrast to EBV and KSHV, murid herpesvirus-4 (MuHV-4) efficiently and productively infects a wide range of cell types (although *in vitro* B-cell infection remains problematic). It also infects mice. However, while antibody reduces MuHV-4 lytic replication *in vivo* ([@r42]; [@r26]; [@r9]), such protection depends on IgG Fc receptor engagement rather than neutralization ([@r50]), perhaps because cell association allows neutralization-resistant virus spread ([@r45]). The likely key setting for MuHV-4 neutralization and its evasion -- the transmission of cell-free virions between hosts -- is not yet experimentally accessible. Nevertheless, MuHV-4 allows us to explore how gammaherpesvirus neutralization might be improved through immunization. MuHV-4 infects fibroblasts and epithelial cells by first attaching to heparan sulfate via gp70 ([@r14]) or gH/gL ([@r17]). These interactions can be blocked by mAbs derived from virus carriers, and a block to cell binding is the main mechanism by which immune sera stop MuHV-4 infecting fibroblasts ([@r10]). However, the apparently neutralized virions can still infect myeloid cells via IgG Fc receptors ([@r37]), and the limited capacity of immune sera to block host entry ([@r15]) suggests that an equivalent rescue operates *in vivo*. In contrast to this redundancy in cell binding, infection always requires membrane fusion. Therefore, the virion fusion machinery -- gB and gH -- is potentially a more universal neutralization target, and mAbs specific for gB or gH/gL (the predominant form of gH on extracellular virions) can block the infection of both fibroblasts and macrophages ([@r15]). However, several factors limit the vulnerability of gB and gH to neutralization. First, gB and gH/gL are poorly immunogenic in the context of whole virus ([@r16]). Neither is conformationally stable without the other ([@r18]), so native virion epitopes tend to be lost unless the gH/gL/gB complex remains intact. In contrast, native gp150 epitopes are preserved even in small protein fragments ([@r16]). Thus gp150, which is not a neutralization target, dominates the antibody response of MuHV-4 carrier mice. gp150-specific antibodies both drive IgG Fc receptor-dependent infection and tend to suppress neutralizing responses ([@r16]), presumably by competition for antigen. Boosting carrier mice with recombinant gH/gL can overcome the immunodominance of gp150 and improve neutralization titres; boosting with gB works less well ([@r15]). Second, a complicated entry mechanism allows key fusion complex epitopes to remain hidden on extracellular virions. MuHV-4 infects by endocytosis ([@r10]). Fusion follows gH/gL dissociation in late endosomes ([@r18]), when gH changes to an antigenically very different form (gH-only). Since gH-only is inaccessible pre-endocytosis, gH-specific antibodies must neutralize by blocking the gH/gL to gH-only transition. gB, which forms a complex with gH/gL ([@r11]), may also not be fully revealed until after endocytosis ([@r20]). The gB N terminus (gB-NT, residues 25--65) is accessible and is a neutralization target for several herpesviruses ([@r33]; [@r34]) including MuHV-4 ([@r13]), but is readily masked by *O*-linked glycans, making this neutralization strongly cell type-dependent ([@r12]). Most MuHV-4 gB-specific mAbs recognize it either before or after virion endocytosis, consistent with an extensive fusion-associated conformation switch ([@r19]). The recombinant gB extracellular domain (residues 1--724) displays mainly post-fusion epitopes. All known pre-fusion epitopes, including all known neutralization epitopes, lie within residues 1--423 (gB-N) ([@r19]), of which residues 81--423 form gB domains I+II ([@r23]). Most gB-specific neutralizing mAbs have been IgMs and therefore difficult to reproduce by vaccination ([@r13], [@r19]). However, we have recently identified a potent neutralizing IgG mAb, SC-9E8, that recognizes gB residues 81--423 ([@r20]). Delivering its epitope could therefore improve MuHV-4 neutralization. We tested here whether recombinant forms of gB could elicit MuHV-4-neutralizing antibodies and protect against MuHV-4 challenge. RESULTS ======= Protection against primary MuHV-4 lytic replication by vaccinia virus-expressed gB (VAC-gB) ------------------------------------------------------------------------------------------- The MuHV-4 gB can be expressed at cell surfaces independent of other virion components by replacing its transmembrane and cytoplasmic domains with a glycosyl-phosphatidyl-inositol (GPI) anchor ([@r28]). VAC-gB-GPI displayed both pre-fusion and post-fusion epitopes, but like transfected gB-GPI ([@r18]) showed a much greater representation of post-fusion epitopes than MuHV-4-infected cells (Fig. [1a](#f1){ref-type="fig"}). Immunization intraperitoneally (i.p.) with VAC-gB (Fig. [1b](#f1){ref-type="fig"}) induced a stronger gB-specific serum antibody response than i.p. infection with MuHV-4, as measured by flow cytometric staining of gB-GPI-transfected CHO cells. Immunization also afforded significant protection against intranasally (i.n.) MuHV-4 challenge (Fig. [1c](#f1){ref-type="fig"}). However, sera from the immunized mice showed no MuHV-4 neutralization (Fig. [1d](#f1){ref-type="fig"}). Similar results were obtained whether sera were used fresh or after heating (56 °C, 30 min) to inactivate complement, and whether or not reconstituted mouse complement (Sigma Chemical Co.) was added to the antibody-coated virions. Expression of gB-N from vaccinia virus -------------------------------------- We reasoned that gB-N (domains I+II) might more effectively present pre-fusion epitopes than the full-length extracellular domain, as it cannot switch to a post-fusion form. The neutralization epitope recognized by mAb SC-9E8 was presented at least as well by VAC-gB-N as by VAC-gB, as was the distinct neutralization epitope recognized by IgM mAb BN-6E1 (Fig. [2a](#f2){ref-type="fig"}). Both pooled (Fig. [2b](#f2){ref-type="fig"}) and individual sera (Fig. [2c](#f2){ref-type="fig"}) from mice immunized with VAC-gB or VAC-gB-N showed gB/gB-N cross-reactivity, but substantially stronger recognition of the cognate form. Protection against primary MuHV-4 lytic replication by VAC-gB-N --------------------------------------------------------------- Priming with either VAC-gB or VAC-gB-N reduced MuHV-4 lytic replication after i.n. challenge (Fig. [3](#f3){ref-type="fig"}). Viral luciferase expression (Fig. [3a, b](#f3){ref-type="fig"}) and plaque assays (Fig. [3c](#f3){ref-type="fig"}) yielded similar results. VAC-gB protected marginally better than VAC-gB-N, particularly in noses, but both significantly reduced infection compared with controls. As in previous studies ([@r43]), reducing lytic replication did not affect long-term latency titres: by 2 months post-infection, splenic infectious centre titres were equivalent between vaccinated and unvaccinated mice (Fig. [3d](#f3){ref-type="fig"}). We were more concerned whether the protection reflected virus neutralization. Despite good gB-specific antibody responses, sera from gB-N-primed mice failed to neutralize MuHV-4 either as pools (Fig. [4a, b](#f4){ref-type="fig"}) or as individual sera (Fig. [4c](#f4){ref-type="fig"}). gB-primed sera even showed a modest enhancement of RAW-264 cell infection compared with naive controls (Fig. [4b](#f4){ref-type="fig"}). Post-challenge serum neutralization titres were also similar between vaccinated and unvaccinated mice (Fig. [4d](#f4){ref-type="fig"}). Therefore, although both gB and gB-N presented neutralizing epitopes, protection was independent of detectable neutralization. Protection by vaccination with gB-N depends on IgG Fc receptors --------------------------------------------------------------- Both neutralizing and non-neutralizing passive antibodies reduce MuHV-4 lytic replication by IgG Fc receptor-dependent mechanisms ([@r50]). The protection by VAC-gB-N appeared to be similar, as it was lost in IgG Fc receptor-deficient mice (Fig. [5a](#f5){ref-type="fig"}). Immune sera transferred from VAC-gB-N-immunized to naive mice also gave IgG Fc receptor-dependent protection (Fig. [5b](#f5){ref-type="fig"}). Boosting neutralizing antibody responses with VAC-gB and VAC-gB-N ----------------------------------------------------------------- We next tested whether VAC-gB-N could boost neutralizing antibodies in MuHV-4 carrier mice. Previously ([@r15]), it was found that VAC-gB was effective in only a minority of mice. C57BL/6 mice infected i.n. with MuHV-4 were 3 months later boosted i.p. with VAC-gB, VAC-gB-N or an irrelevant VAC control. The mice were bled before boosting, 10 days after (acute response) and 30 days after (longer-term response) (Fig. [6](#f6){ref-type="fig"}). VAC-gB-N boosted gB-N-specific antibodies significantly better than VAC-gB without improving the recognition of CHO-gB cells, again emphasizing that gB-N and gB are antigenically quite different (Fig. [6a](#f6){ref-type="fig"}). The magnitude of gB- and gB-N-specific antibody responses varied considerably between individual mice (Fig. [6b](#f6){ref-type="fig"}). The gB-N-specific antibody responses of non-boosted carriers (Fig. [6c](#f6){ref-type="fig"}) were also highly variable, and showed no correlation with responses to full-length gB or the MuHV-4 gp70 (Fig. [6d](#f6){ref-type="fig"}). Thus, even homozygous sibling mice varied considerably in antibody response. Pooled C57BL/6 mouse sera (Fig. [7a](#f7){ref-type="fig"}) neutralized MuHV-4 for BHK-21 cell infection better after boosting with either VAC-gB or VAC-gB-N compared with the control. Boosting improved the neutralization of RAW-264 cell infection only at the 10 day time point. The effects were not large, and individual neutralization titres (Fig. [7b, c](#f7){ref-type="fig"}) overlapped between groups, consistent with the varied boosting efficacy observed by flow cytometry (Fig. [6](#f6){ref-type="fig"}). Thus, VAC-gB and VAC-gB-N were both capable of boosting neutralizing antibodies, but worked well in few mice, presumably because few mice made primary gB-specific neutralizing responses. VAC-gB and VAC-gB-N similarly increased significantly gB-N- and gB-specific antibody responses in BALB/c carrier mice (Supplementary Fig. S1, available in JGV Online) with only a small improvement in neutralization (Supplementary Fig. S2, available in JGV Online). As the boosting of neutralization was similar between VAC-gB and VAC-gB-N, the limiting factor appeared to be the composition of the primary antibody response to MuHV-4 rather than the level of subsequent gB-N delivery. DISCUSSION ========== We sought to understand why MuHV-4-infected mice make poor neutralizing antibody responses to gB. Specifically, we tested whether recombinant forms of gB containing known neutralization epitopes could stimulate neutralizing responses in either naive or carrier mice. Immunization with either the full-length gB extracellular domain or its N-terminal half reduced subsequent MuHV-4 lytic replication. However, neither construct elicited detectable primary neutralizing responses, even though both were recognized by a potent neutralizing mAb. Some boosting of neutralizing antibodies in carrier mice was possible, but the effect was not large, there was considerable variation between individual mice, and gB-N was no more effective than gB. We conclude that gB-directed MuHV-4 neutralization is fundamentally limited by the difficulty of targeting the few vulnerable epitopes it presents. How does an essential component of the MuHV-4 entry complex avoid being a good neutralization target? It seems likely that post-endocytic conformation changes allow key gB epitopes to remain hidden on extracellular virions. Thus, while gB-N binds to cells ([@r12]), this must occur downstream in entry as blocking heparan sulfate binding by gp70 and gH/gL largely abrogates virion binding ([@r21]). The increased susceptibility of gL^−^ virions to gB-directed neutralization ([@r20]) implies that gH/gL helps to hide gB; the regulation of virion binding by gp150 ([@r7]) suggests that it could also be involved. While the neutralization epitope defined by mAb SC-9E8 is accessible it is evidently difficult to target, much like broadly neutralizing human immunodeficiency virus epitopes ([@r4]). Such resistance to neutralization may be a common feature of persistent viruses. Herpes simplex virus (HSV) and human cytomegalovirus (HCMV) may be more easily neutralized *in vitro* by gB-specific antibodies ([@r5]; [@r24]; [@r48]) because they fuse at the plasma membrane rather than in late endosomes: if gB conformation switches are reversible as for vesicular stomatitis virus glycoprotein G ([@r36]) and some switching precedes actual fusion, then mAbs specific for the post-fusion HSV or HCMV gB could lock it non-productively in that form. Such a mechanism would be consistent with HSV and HCMV neutralization epitopes mapping to prominent features of post-fusion gB ([@r23]). In contrast, endocytic entry hides the MuHV-4 gB conformation changes from antibody. HCMV may be protected partly by competition between neutralizing and non-neutralizing antibodies ([@r41]), but viral transmission would also segregate plasma membrane fusion from the antibody in immune donors, thereby protecting gB epitopes revealed after cell binding. Reducing viral lytic replication may reduce transmission events associated with primary infection, but optimally protecting naive subjects against infection by virus carriers is likely to require mucosal vaccination, either boosting IgA responses in carriers or generating primary IgA responses in naive subjects. First it is necessary to induce good neutralization. Attention so far has focussed on IgG, since IgG responses are more readily elicited by vaccination. Alpha- and betaherpesvirus gB vaccine studies ([@r49]; [@r32]; [@r35]) have generally not distinguished neutralization from IgG Fc receptor-dependent effector functions. This first description of gB immunization against a gammaherpesvirus gave protection consistent with the alpha- and betaherpesvirus studies, but showed that neutralization was not involved. Therefore, it seems unlikely that vaccination with gB alone will reduce infection prevalence. METHODS ======= Mice. ----- C57BL/6, BALB/c, Sv129 (Harlan UK Ltd) and C57BL/6×Sv129 FcR*γ*^−/−^Fc*γ*RII^−/−^ mice (Taconic Europe) were housed at the Cambridge University Department of Pathology and infected when 6--8 weeks old under the Home Office Project Licence 80/1992. All experiments conformed to local and national ethical regulations. MuHV-4 (10^4^ p.f.u.) was given i.n. under general anaesthesia. Vaccinia viruses (10^6^ p.f.u.) were given i.p. For luciferase imaging, mice were injected i.p. with luciferin (2 mg per mouse), anaesthetized with isoflurane and scanned with an IVIS Lumina (Caliper Life Sciences). Quantitative comparisons were made by using Living Image software. Sera were collected from tail veins, or by cardiac puncture under terminal anaesthesia. Cells. ------ BHK-21 fibroblasts (ATCC CCL-10), TK^−^143 cells (ATCC CRL-8303), NMuMG epithelial cells (ATCC CRL-1636), NIH-3T3-CRE cells ([@r44]), RAW-264 monocytes (ATCC TIB-71), CHO-K1 cells (ATCC CCL-61) and CHO-K1 derivatives expressing either gp70, the main product of ORF4 ([@r14]), a fusion protein of the gH and gL extracellular domains with a GPI anchor ([@r16]), or a GPI-linked gB extracellular domain (CHO-gB) ([@r28]), were grown in Dulbecco\'s modified Eagle\'s medium (DMEM; Invitrogen Corporation) supplemented with 10 % FCS (PAA Laboratories), 2 mM glutamine, 100 U penicillin ml^−1^ and 100 μg streptomycin ml^−1^. To make CHO-gB-N cells, the coding sequence for gB N-terminal to its furin cleavage site (aa 1--423) was cloned upstream of a GPI anchor attachment site as for full-length gB ([@r28]). CHO-K1 cells transfected with the gB-N expression plasmid were selected with hygromycin (250 μg ml^−1^). Viruses. -------- All MuHV-4 derivatives were generated from a BAC-cloned genome ([@r1]). Luciferase expression from an ectopic MuHV-4 lytic promoter has been described previously ([@r30]). The MuHV-4 BAC contains an HCMV IE1 promoter-driven eGFP expression cassette. However, this is poorly transcribed in some cell types ([@r37]). Therefore to assay *in vitro* infection, we inserted into the MuHV-4 genome a separate, intergenic eGFP expression cassette with an EF1*α* promoter. We first mutated the internal *Bgl*II site of the EF1*α* promoter in pBRAD from AGATCT to AGGTCT by overlap PCR, then PCR-amplified the modified promoter, adding *Bgl*II and *Eco*RI sites to its respective 5′ and 3′ ends, and cloned upstream of a poly(A) site in pSP73 ([@r29]). The eGFP coding sequence was PCR-amplified from pEGFP-N3 (Clontech), adding 5′ *Eco*RI and 3′ *Xho*I sites, and cloned into *Eco*RI/*Sal*I sites between the EF1*α* promoter and poly(A) site. The resulting eGFP expression cassette was then subcloned as a blunted *Bgl*II--*Xho*I fragment into the *Mfe*I site of a 3.4 kb *Bgl*II genomic clone, between the 3′ ends of ORFs 57 and 58 ([@r29]). Inserts here have no obvious effect on MuHV-4 replication ([@r3]). The eGFP expression cassette and its genomic flanks were then subcloned into the pST76K-SR shuttle vector and recombined into the MuHV-4 BAC by standard methods ([@r1]). Recombinant clones were identified and checked for genome integrity by restriction enzyme mapping. Infectious virus was recovered by transfecting BAC DNA into BHK-21 cells. The BAC cassette was then removed by passage through NIH-3T3-CRE cells. Virus stocks were grown in BHK-21 cells. Virions were harvested from infected cell supernatants by ultracentrifugation, and cell debris removed by low speed centrifugation ([@r7]). Vaccinia viruses expressing a GPI-linked gB extracellular domain or gH/gL fusion protein have been described previously ([@r15]). To make VAC-gB-N, the coding sequence for gB aa 81--423 was amplified by PCR, digested with *Xba*I at a site internal to gB and cloned as a blunt *Xba*I fragment into the *Sna*BI/*Xba*I sites of gB in pBRAD-gB. The predicted gB signal sequence is aa 1--24, and *Sna*BI cuts after residue 22, so the coding sequence for residues 23--25 was included in the 5′ PCR primer, thereby reconstituting the native signal sequence without residues 26--80. The modified gB-N coding sequence plus C-terminal GPI attachment site was PCR-amplified from pBRAD-gB-N using 5′ *Nhe*I-restricted and 3′ *Hin*dIII-restricted primers and cloned into the *Nhe*I/*Hin*dIII sites of the vaccinia recombination plasmid pMJ601 ([@r6]). pMJ-601-gB-N-GPI was then transfected into vaccinia virus-infected TK^−^143 cells. Thymidine kinase-deficient recombinants were selected by passage in 25 μg 5′-bromo-2′-deoxyuridine (Sigma Chemical Co.) ml^−1^, identified by *β*-galactosidase expression using X-Gal, and plaque purified. Virus stocks were grown and titrated using TK^−^143 cells. As a TK^−^ control, we used vaccinia virus expressing the mouse invariant chain coding sequence with ovalbumin residues 323--339 substituted for the CLIP peptide ([@r39]). Virus titres. ------------- Virus stocks and *ex vivo* samples were titrated for infectivity by plaque assay on BHK-21 cells ([@r7]). Lungs and noses were removed from mice post-mortem, freeze--thawed, then homogenized in DMEM. Nose samples included the turbinates and nasal septum, which contain all the nasal luciferase signal of mice infected with luciferase^+^ MuHV-4 ([@r30]). Serial dilutions of each sample were incubated (2 h, 37 °C) with BHK-21 cell monolayers, then overlaid with DMEM plus 0.3 % carboxymethylcellulose. After 4 days the monolayers were fixed in 4 % formaldehyde, stained with 0.1 % toluidine blue and plaques were counted with a plate microscope. Neutralization assays. ---------------------- EF1*α*-eGFP MuHV-4 was incubated with serum dilutions (2 h, 37 °C), then added to BHK-21 fibroblasts or RAW-264 monocytes. After 2 h, phosphonoacetic acid was added (100 μg ml^−1^) to prevent lytic spread. After 16 h, the cells were harvested and assayed for eGFP expression by flow cytometry. In preliminary experiments, virus titres by EF1*α*-driven eGFP expression in BHK-21 cells equalled or exceeded plaque assay titres, and in RAW-264 cells equalled or exceeded BAC cassette-associated eGFP expression maximized by lipopolysaccharide treatment ([@r37]). Flow cytometry. --------------- Cells exposed to eGFP^+^ viruses were washed twice in PBS and analysed directly for green channel fluorescence. For specific staining, cells were incubated (1 h, 4 °C) with MuHV-4 gB-specific mAbs or with immune sera, followed by fluorescein-conjugated rabbit anti-mouse IgG pAb (Dako Cytomation). mAbs BN-1A7 (IgG2a), BN-6E1 (IgM) and SC-9E8 (IgG2a) all recognize epitopes in gB-N and are specific for pre-fusion gB, whereas MG-1A12 is specific for post-fusion gB ([@r19]). All cells were washed twice in PBS after each antibody incubation and analysed on a FACS Scan that runs the CellQuest software (BD Biosciences). Supplementary Material ====================== ::: {.caption} ###### \[Supplementary Figures\] ::: This work was supported by the Wellcome Trust (GR076956MA and WT089111MA) and by the Medical Research Council (G0701185). Supplementary figures are available with the online version of this paper. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Expression of gB epitopes by vaccinia virus-expressed gB and gB-N. (a) BHK-21 cells were left uninfected (UI) or infected (1 p.f.u. per cell, 14 h) with either MuHV-4 or a vaccinia virus expressing a GPI-linked gB extracellular domain (VAC-gB). Cell surface gB expression was analysed by flow cytometric staining of pre-fusion gB with mAb BN-1A7, and post-fusion gB with MG-1A12. (b) C57BL/6 mice were infected i.p. with either MuHV-4 or VAC-gB. Sera collected 2 weeks later were used to stain CHO cells either untransfected or transfected with gp70 or GPI-linked gB. Each point shows the median fluorescence intensity for a 1/500 serum dilution from one mouse. The bars show mean values. CHO-gB staining was significantly greater after VAC-gB infection than after MuHV-4 infection (*P*\<0.01 by Student\'s two-tailed *t*-test). In this and subsequent experiments, consistent results were obtained with least two further serum dilutions. (c) Mice were immunized i.p. with either VAC-gB or a control, and 1 month later infected i.n. with MuHV-4. Infectious MuHV-4 in lungs was plaque assayed at 5 days post-challenge. Each point shows the titre of one mouse. The bars show geometric means. (d) Sera from mice infected with either VAC-gB or MuHV-4 were used to neutralize eGFP^+^ MuHV-4 for BHK-21 cell and NMuMG cell infections. Infection was assayed by flow cytometry for eGFP expression 18 h after exposure to virus/antibody. Each point shows mean±[sd]{.smallcaps} of triplicate infections. The dashed lines show infection with virus alone. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Primary antibody responses to gB vaccines in C57BL/6 and BALB/c mice. (a) BHK-21 cells were left uninfected (UI) or infected (1 p.f.u. per cell, 14 h) with vaccinia virus expressing either the full-length gB extracellular domain (VAC-gB), just its N-terminal domains (VAC-gB-N), or a control insert (VAC-cont). gB expression was analysed by flow cytometry. mAbs BN-6E1 and SC-9E8 recognize neutralizing epitopes on pre-fusion gB. MG-1A12 recognizes a post-fusion epitope that is not contained within gB-N. (b) Sera from mice infected i.p. 1 month previously with VAC-gB, VAC-gB-N or a control virus were pooled from six mice per group and assayed for reactivity to gB, gB-N and gp70 (as a negative control) by flow cytometry of transfected CHO cells. Bars show mean±[sd]{.smallcaps} values. For both C57BL/6 and BALB/c mice, VAC-gB-N primed CHO-gB-N-specific responses significantly better than did VAC-gB, and VAC-gB primed CHO-gB-N-specific responses significantly better than did VAC-gB-N (*P*\<0.0001 by Student\'s two-tailed *t*-test). (c) Individual sera from the C57BL/6 mice in (b) were assayed for reactivity to gB and gB-N. Each point shows staining of the relevant CHO cells by serum from one mouse. The bars show mean values. VAC-gB-primed mice showed significantly stronger staining of both CHO-gB (*P*\<0.0001) and CHO-gB-N (*P*\<0.01), as did VAC-gB-N-primed mice (*P*\<0.0001 by Student\'s two-tailed *t*-test). Each group also showed significantly stronger staining of the cognate form of gB (*P*\<0.0001). :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### Protection against MuHV-4 lytic replication by vaccination with gB. (a) BALB/c mice were immunized i.p. with vaccinia viruses expressing gB, gB-N or an irrelevant insert (VAC-cont), and 1 month later challenged i.n. (10^4^ p.f.u.) with luciferase^+^ MuHV-4. Viral replication in lungs and noses was monitored by luciferin injection and CCD camera scanning. A representative image is shown of mice 5 days post-challenge. (b) Mice were immunized then infected with luciferase^+^ MuHV-4 as in (a). Each point shows the luciferase signal for one mouse. The bars show mean values. Dashed lines show the lower limits of signal detection. Priming with either VAC-gB or VAC-gB-N significantly reduced luciferase signals in lungs at all time points (*P*\<0.03 by Student\'s two-tailed *t*-test). Luciferase signals in noses were also significantly reduced in VAC-gB-primed mice at days 2 (*P*\<0.01) and 5 (*P*\<0.04), and in VAC-gB-N-primed mice at day 2 (*P*\<0.01). (c) BALB/c or C57BL/6 mice were immunized i.p. with vaccinia virus recombinants then challenged i.n. with MuHV-4 as in (b), but with luciferase^−^ MuHV-4. Lungs and noses were titrated for infectious virus at 6 days post-infection by plaque assay. Each point shows the titre of one mouse. The bars show geometric means. Both BALB/c (*P*\<0.04) and C57BL/6 mice (*P*\<0.01) lung titres were significantly reduced by priming with either VAC-gB or VAC-gB-N. C57BL/6 mice nose titres were also reduced, although this was only statistically significant for VAC-gB (*P*\<0.01). (d) Latent virus loads in spleens were measured by infectious centre assay at 2 months post-challenge of vaccinated C57BL/6 mice. The bars show mean±[sd]{.smallcaps} titres of five mice per group. :::  ::: ::: {#f4 .fig} Fig. 4. ::: {.caption} ###### gB-primed sera fail to neutralize MuHV-4. (a) BALB/c mice were primed with vaccinia virus expressing gB, gB-N or a control insert. After 1 month, sera pooled from six mice per group were tested for their capacity to reduce BHK-21 or RAW-264 cell infections by eGFP^+^ MuHV-4, as assayed by flow cytometry 16 h after exposure to virus. Neutralization by sera pooled from MuHV-4-infected mice (Imm ser) is also shown. Dashed lines show the level of infection by virus alone. (b) In a similar protocol to (a), VAC-gB-primed and VAC-gB-N-primed sera, each pooled from six C57BL/6 mice, showed no significant MuHV-4 neutralization. (c) Sera from the mice in (b) were tested individually for neutralization of BHK-21 cell infection. Again none was observed. The dashed line shows neutralization by pooled, MuHV-4-immune sera. (d) Sera pooled from unprimed or VAC-primed mice (*n*=7) were collected either before or 16 days after MuHV-4 challenge and used to neutralize MuHV-4 virions for BHK-21 cell infection. Each line shows mean±[sem]{.smallcaps} results. No significant difference was observed between VAC-gB-N-primed and control mice. :::  ::: ::: {#f5 .fig} Fig. 5. ::: {.caption} ###### Protection of naive mice and induction of MuHV-4 neutralization by vaccination with a gHL fusion protein. (a) 129Sv×C57BL/6 FcR*γ*^−/−^Fc*γ*RII^−/−^ mice (FcR^−/−^) or 129Sv controls were primed i.p. with vaccinia virus recombinants (10^5^ p.f.u.) and 1 month later challenged i.n. with MuHV-4 (10^4^ p.f.u.). Infectious virus titres in lungs and noses were measured by plaque assay 6 days later. Each point shows the result for one mouse. The bars show geometric means. VAC-gB significantly reduced virus replication in lungs (*P*\<0.00001) and noses (*P*\<0.04) of 129SV but not FcR^−/−^ mice. (b) Sera were pooled from mice (*n*=10) 3 months after i.p. infection with VAC-gB-N or VAC-cont, then given i.p. to 129Sv or FcR^−/−^ mice (500 μl per mouse) at the same time as i.n. MuHV-4 (5000 p.f.u.). Lungs were titrated for infectious virus by plaque assay 5 days later. Each point shows the result for one mouse, and the bars show geometric means. gB-N-immune sera significantly reduced virus titres in 129Sv mice (*P*\<0.0001 by Student\'s two-tailed *t*-test) but not in FcR^−/−^ mice (*P*=0.86). :::  ::: ::: {#f6 .fig} Fig. 6. ::: {.caption} ###### Boosting gB-specific antibody responses by post-exposure vaccination of MuHV-4 carrier mice. (a) C57BL/6 mice were infected i.n. with MuHV-4 and 3 months later boosted i.p. with vaccinia virus expressing either the full-length gB extracellular domain (VAC-gB), just its N-terminal half (VAC-gB-N) or a control insert (VAC-cont). Sera were taken before vaccinia infection (pre-boost), 10 and 30 days later, and tested for gB reactivity by flow cytometric staining of CHO cells expressing gB or gB-N. CHO cells either untransfected or transfected with the MuHV-4 gp70 provided staining controls. Bars show mean±[sd]{.smallcaps} fluorescence intensities for sera pooled from six mice each. VAC-gB-N and VAC-gB both significantly increased CHO-gB-N staining at days 10 and 30 compared with VAC-cont (*P*\<0.0001 by Student\'s two-tailed *t*-test). The increase with VAC-gB-N was significantly greater than with VAC-gB (*P*\<0.0001). VAC-gB but not VAC-gB-N significantly increased CHO-gB staining at days 10 and 30 compared with VAC-cont (*P*\<0.0001). (b) Sera from the individual day 30 boosted mice in (a) were assayed for CHO-gB and CHO-gB-N reactivity. Although boosting increased staining, this did not reach statistical significance (*P*\>0.05) because of wide individual variation. Each point shows an individual mouse and the bars show geometric means. (c) C57BL/6 MuHV-4 carrier mice (3 months post-infection) were analysed by flow cytometry for serum binding to CHO-gB-N cells. Each point shows the result for one mouse. (d) A separate set of MuHV-4-infected C57BL/6 mice were compared for serum reactivity to CHO-gB, CHO-gB-N and CHO-gp70 cells. Each point shows the result for one mouse. There was no obvious correlation for individual mice between the responses to each target. :::  ::: ::: {#f7 .fig} Fig. 7. ::: {.caption} ###### Boosting gB-specific neutralizing antibody responses by post-exposure vaccination of MuHV-4 carrier mice. (a) Sera taken pre-boosting, 10 or 30 days post-boosting with vaccinia viruses expressing the full-length gB extracellular domain (VAC-gB), its N-terminal half (VAC-gB-N) or an irrelevant protein (VAC-cont), were pooled from groups of six C57BL/6 mice each. eGFP^+^ MuHV-4 virions were incubated with dilutions of the pooled sera and then used to infect either BHK-21 cells (0.3 p.f.u. per cell) or RAW-264 cells (3 p.f.u. per cell). Infected cells were enumerated 16 h later by flow cytometry. Dashed lines show infection with untreated virus. Each point gives the result for 10 000 cells. Thus by *χ*^2^ test, the visible reductions in infection with VAC-gB or VAC-gB boosting were all highly significant (*P*\<0.0001). The day 30 RAW-264 cell infection differences were not considered significant as they were not maintained over more than 1 serum dilution. (b) Individual sera taken at day 30 post-boosting were analysed for neutralization of BHK-21 cell infection as in (a). With 4 μl serum, both the VAC-gB (*P*\<0.05) and VAC-gB-N boosted groups showed a significant reduction in infection (*P*\<0.03 by Student\'s two-tailed *t*-test) compared with the control group. With 1.3 μl serum, only the VAC-gB-N boosted group showed a reduction (*P*\<0.03). (c) Sera taken at day 30 post-boosting were analysed for neutralization of RAW-264 cell infection (3 p.f.u. per cell). The boosted groups showed a wider distribution than the unboosted, but no significant reduction. :::  :::

PubMed Central

2024-06-05T04:04:19.404819

2010-10-01

PMC3052601

INTRODUCTION ============ Scrapie is a naturally occurring transmissible spongiform encephalopathy (TSE) that has been endemic in the British national sheep flock for over 250 years ([@r30]; [@r32]), and causes progressive neurological degeneration and death. It is associated with an abnormal form of the prion protein (PrP^Sc^) ([@r3]). Prioritization of scrapie eradication in Great Britain (GB) occurred after the link between bovine spongiform encephalopathy (BSE) in cattle and variant Creutzfeldt--Jakob disease (vCJD) in humans was discovered ([@r7]; [@r2]; [@r18], [@r19]). The potential for sheep to be a vCJD risk was based on four concerns: (i) sheep were demonstrated to be susceptible to BSE under experimental conditions ([@r7]), (ii) they were potentially exposed to the same contaminated feed that had caused the outbreak in cattle ([@r24]), (iii) the distribution of BSE-infected tissue in sheep ([@r22]; [@r35]) has made horizontal transmission a distinct possibility ([@r24]) and (iv) as sheep affected by BSE and scrapie have very similar clinical signs and pathology, scrapie has the potential to mask an incipient BSE epidemic in sheep ([@r21]). As a result, the National Scrapie Plan (NSP) ([@r4]) was implemented in GB in July 2001 (<http://www.defra.gov.uk/animalhealth/managing-disease/NSPAC/>, discussed in [@r23]). The NSP\'s primary objectives were to eradicate scrapie and breed for TSE resistance in the national sheep flock (<http://www.defra.gov.uk/animalhealth/managing-disease/NSPAC/>, discussed in [@r1]), thereby minimizing the likelihood that BSE could be present and not detected in the national flock. The strategy was adopted due to the clear relationship between polymorphisms in the gene that encodes the prion protein and susceptibility to scrapie and BSE ([@r9]). At the time, any strategy based purely on identifying cases of disease was considered prohibitively risky and expensive because the possible low incidence of BSE in sheep may be masked by the presence of scrapie. Thus, a genetically based breeding strategy targeting susceptibility, rather than disease, provided the most reasonable chance for success. In 2004, the programme was augmented by a slaughter and replacement scheme. This was initially a voluntary programme, the voluntary scrapie flock scheme (VSFS) aimed at flocks known to have been scrapie-affected since 1998, but after July 2004, control became mandatory for all flocks with confirmed cases from that date, as required by EC Reg 999/2001. Upon confirmation, the compulsory scrapie flock scheme (CSFS) ([@r5]) requires that all sheep be slaughtered, or all susceptible sheep in the flock slaughtered and only resistant replacement animals bought in. While this has been undoubtedly effective, the relatively small incidence of scrapie means that large numbers of healthy sheep are culled. The recent development of a live test for scrapie ([@r10]) suggests that preclinical testing may prove a more cost-effective disease-based strategy for scrapie eradication. Immunohistochemical staining of lymphoid biopsies from tonsil ([@r33]) and third eyelid ([@r28]) have long been recognized as a method for preclinical diagnosis of TSEs. In this test, rectal biopsies are used, as anaesthesia is not required for collection, and lymphoid follicles are abundant in the rectal mucosa. Further, because the analysis can be adapted to rapid, high-throughput methods ([@r11]), this approach is potentially feasible for large-scale application in the field. Previous mathematical models have considered scrapie control at the flock ([@r36]; [@r16]; [@r8]; [@r37]) and national level ([@r23]; [@r14]; [@r15]). However, none have considered a strategy targeting infected sheep. Here, the trade-off between the number of sheep culled and the potential for onward transmission of disease after the implementation of a single (conducted at a single point in time) or sequential (conducted once each year over 3 years) live preclinical test(s) are contrasted to the CSFS. As there is strong experimental ([@r31]; [@r26]) and epidemiological evidence ([@r27]; [@r1]) that scrapie transmission is related to infected breeding sheep, the number of infected sheep likely to be sold on for breeding is used to compare between these strategies. RESULTS ======= Flocks were characterized according to the risk posed to other farms. The number of infected breeding sheep sold was calculated from the number of sheep sold by each flock type and the prevalence of scrapie in the absence of a control strategy (in each year of a 15 year epidemic). A summary of each of the flock parameters and the risk they posed to other flocks is presented in Fig. [1](#f1){ref-type="fig"}. The mean number of infected breeding sheep sold by different flock types to other farms each year in the first year of an epidemic is described in Fig. [2](#f2){ref-type="fig"}. The mean number of infected breeding sheep sold per 100 sheep (in the flock) by different flock types per year in each year of a 15 year epidemic is described in Fig. [3](#f3){ref-type="fig"}. No intervention --------------- Assuming a mean genotype distribution (39 % susceptible genotypes in the flock), over a 15 year period, scrapie prevalence decreased in small flocks (i.e. commercial flocks with ≤200 sheep and pure-bred flocks with ≤100). In these flocks, the smaller the proportion of home-bred sheep, the lower the prevalence of scrapie in the flock. In large commercial flocks (≥500 sheep), the prevalence of scrapie remained constant over a 15 year period if the proportion of home-bred sheep was small (≤0.10). However, if the proportion of home-bred sheep was large (≥0.89), scrapie prevalence in large commercial (≥500 sheep) and pure-bred flocks (≥700 sheep) increased over time. When the proportion of susceptible genotypes in the flock was small (0.39), small pure-bred flocks with large proportions of home-bred sheep and small commercial flocks with large and small proportions of home-bred sheep sold on average less than one infected breeding sheep per year and thus were considered a low risk for onward transmission (Figs [1](#f1){ref-type="fig"}, [2](#f2){ref-type="fig"} and [3](#f3){ref-type="fig"}). A small proportion of large commercial flocks with small proportions of home-bred sheep sold on ≥1 infected sheep per year (the number of infected sheep sold increased over time) (Figs [1](#f1){ref-type="fig"}, [2](#f2){ref-type="fig"} and [3](#f3){ref-type="fig"}). These flocks were defined as moderate-risk flocks due to the potential for the relatively small numbers of infected breeding sheep sold on each year. The majority of large pure-bred flocks posed a high risk of onward transmission (≥1 infected sheep sold per year). A smaller proportion of large commercial flocks posed similar risks. Both of these flock types were considered to be high risk due to the relatively high potential to sell on infected breeding sheep (Figs [1](#f1){ref-type="fig"}, [2](#f2){ref-type="fig"} and [3](#f3){ref-type="fig"}). A threshold susceptible proportion was calculated for each flock type. This threshold was the proportion of susceptible genotypes in the flock which reduced the maximum risk of onward transmission to ≤1 infected sheep sold per year. In high-risk flocks, the threshold susceptible proportion was less than 5 %; in moderate-risk flocks, less than 39 % (the national average) and in low-risk flocks with large proportions of home-bred sheep, less than 50 %. In contrast, low-risk flocks with small proportions of home-bred sheep posed little risk of onward transmission even when the threshold susceptible proportion was very high (≥0.95). CSFS strategy (strategy 2) -------------------------- While the CSFS strategy was always successful in eradicating disease, this invariably resulted in higher numbers of sheep removed and culled than the two alternative preclinical testing strategies (strategies 3 and 4), as all sheep with susceptible genotypes in the flock would be removed, regardless of disease status. Preclinical testing strategies (strategies 3 and 4) --------------------------------------------------- There was no difference in the results between either of the testing strategies when the tests were applied to sheep at 20 or 12 months of age in any of the flock types, due to the early detection assumed for highly susceptible genotypes. However, as test sensitivity decreased, the mean number of sheep removed from the flock decreased and the risk of onward transmission increased. Even in high-risk flocks, the within-flock prevalence of scrapie was low \[≤0.22 (LPLH; see Table [2](#t2){ref-type="table"})\]. Therefore, both the positive predictive value (PPV) of the test (100 %) and the negative predictive value (NPV) (\>97 %) were high if the test sensitivity was at least 90 %. At a similar prevalence, but with lower test sensitivities (70 and 35 %, respectively), the PPV remained unchanged, but the NPV decreased (greater than 93 and 86 %, respectively). Therefore in high-risk flocks, in the worst case scenario (with a test sensitivity of 35 %) the probability of retaining an undetected diseased sheep in the flock was 14 %. Conversely, in low-risk flocks (prevalence \<1 %), with a test sensitivity of 35 %, the probability of retaining an undetected diseased sheep was less than 1 %. Both sequential and single testing strategies were more effective at reducing prevalence and onward transmission of infection when tests were applied and assumed to be effective in sheep at 6 months of age than when testing was delayed until 12 and 20 months of age. In all flock types, sequential-testing strategies resulted in increased numbers of infected sheep culled compared with a single-test strategy (due to the increased opportunity to detect infected sheep in subsequent years). However, sequential-testing strategies still resulted in a smaller proportion of the flock culled (percentage of flock culled in years 1--15 : 0.2--19 % in large pure-bred flocks, 0.2--0.3 % in large commercial flocks with small proportions of home-bred sheep and 0.3--11 % in large commercial flocks with large proportions of home-bred sheep) compared with the CSFS (36 % of all flocks culled). Both live testing strategies substantially reduced the risk of onward transmission of infection in high-risk flocks (Fig. [3](#f3){ref-type="fig"}). The effectiveness of each strategy was dependent on when disease was detected during the course of a within-flock epidemic. At high test sensitivities (≥90 %), if scrapie was detected early in the epidemic (\<5 years), sequential testing reduced the risk of onward transmission (≤1 infected breeding sheep). However, if scrapie was detected later in the epidemic (≥5 years), the majority of flocks still sold on one or more infected sheep depending on the year of implementation. A single test strategy in these high-risk flocks was not effective as the majority of flocks still sold one or more than one infected breeding sheep in the first year of the epidemic after implementation of the test. In moderate- and low-risk flocks, sequential and single testing strategies were equally effective in reducing the risk of onward transmission to other flocks (number of sheep sold \<1) at any time during the epidemic. DISCUSSION ========== This study illustrates important differences in the demographics, trading patterns and ability of different flock types to sustain or perpetuate scrapie epidemics that are important for understanding the epidemic dynamics at a national level, and targeting control strategies. While this approach is invaluable for identifying broad, strategic approaches to control disease, it is not intended to provide detailed predictions about specific outbreaks. While always successful at eliminating a within-flock epidemic of classical scrapie the CSFS strategy relies on the removal of the entire susceptible population in the flock, which potentially includes many healthy sheep. [@r37] showed that trading restrictions alone have little power to limit transmission and a more efficient implementation could concentrate on culling or breeding restrictions; however, they did not consider the detailed demographic data from the postal surveys, which show clear distinctions amongst different flock types. Small low-risk flocks (≤200 commercial sheep and ≤100 pure-bred sheep) were less likely to sustain an epidemic or sell on infected breeding sheep. This suggests that in these flocks, draconian control measures such as the CSFS may be unnecessary in preventing the onward spread of disease. For smaller flocks, particularly with few home-bred animals, there is both little opportunity for transmission and few susceptibles, especially with replacement of clinical cases with resistant stock. Thus, the number of susceptible animals will probably rapidly decline below a density threshold that is necessary to support endemic infection ([@r41]). Similarly, the CSFS may be unnecessary for those flocks classified in this study as 'moderate-risk flocks'. These flocks would typically be mule or other cross-bred sheep (M. Dawson, personal communication) and virtually all production would be finished lamb. Ewes would normally be culled at the end of their breeding life and as such, sheep sold on from these flocks are unlikely to be breeding sheep. Therefore, the risk for disease transmission from the sale of sheep from these flocks may be smaller than estimated in this study. Conversely, large flocks (≥500 commercial sheep and ≥700 pure-bred sheep), with large proportions of home-bred sheep (which retained all their infected lambs), posed the greatest risk to other farms due to the increased opportunity for an epidemic to take hold. In these flocks, live preclinical testing dramatically reduced the number of sheep culled compared with the CSFS. However, testing strategies are unlikely to eradicate scrapie completely within a scrapie-positive flock due to the imperfect test sensitivity and probability of retaining infected sheep in the flock through misclassification. Even intermittent onward sale may result in epidemic persistence, if, for example, larger flocks were found to preferentially buy and sell from each other, either deliberately or inadvertently. Thus, the reduction in cost to the farmer must be balanced against the risk posed to other flocks at a national level, in particular if the test sensitivity is low in preclinical animals or if the test is too costly or difficult to implement or interpret. The onward risk of transmission by these flocks depends on the combination of having large numbers of animals at risk ([@r17]; [@r20]), large numbers of replacement animals providing a steady supply of new susceptibles, and large numbers of sheep sold on to other farms. These factors therefore identify likely scenarios under which stringent control measures may be necessary and in these high-risk flocks, the existing CSFS strategy may remain the most appropriate method of control and eradication. Use of a single test strategy may be cheaper and easier to implement. However, this strategy may result in the unacceptable exposure of additional flocks via the sale of breeding sheep from high-risk flocks. Sequential-testing strategies resulted in increased numbers of infected sheep identified and culled and are thus likely to be more expensive compared with a single-test strategy. However, as retesting in subsequent years was performed using the same rectal biopsy test, the improvement in test sensitivity from using this approach is not as great as predicted if the tests were conditionally independent ([@r6]). In this study, the predicted prevalence of within-flock scrapie infection in high-risk pure-bred and commercial flocks is higher than the estimated within-flock prevalence reported by [@r29] and [@r14]. This may be due to differences with respect to the stage of detection of the within-flock epidemic, ability to detect disease, or most likely, the effect of stratification of flock types according to size and proportion of home-bred sheep. Nevertheless, a higher estimate of prevalence in this study (compared with others) highlights the fact that only a small proportion of the flocks in GB (high-risk pure-bred and commercial flocks) pose the greatest risk for onward transmission of scrapie infection and these should be identified and targeted in future control strategies (Fig. [1](#f1){ref-type="fig"}). Future work may investigate different live-testing strategies such as testing specific genotypes only or random sampling of sheep within a flock to determine if this has an impact on the efficiency and cost effectiveness of the diagnostic test. Further information to refine flock parameters (such as different genotypic distributions for different flock types) would also assist in the identification of high-risk flocks. The logistics of testing both at a flock and a national scale must also be considered, and additional information is still required to determine the cost and ease of test implementation, the cost of culling, restrictions in breeding and trading to the farmer at an individual and national level, before further economic evaluation and comparison with the CSFS can be made. In practice, the high-risk classification used in this study may not pertain to all large pure-bred or commercial flocks. As an example, the majority of hill breed flocks do not sell pure-bred females, although some flocks will buy in breeding ewes during particularly poor years. Typically, the best ewe lambs are retained and the remainder are finished or sold as stores. In a minority of such flocks, some ewes are 'drafted' to lower ground to produce mule or half-bred ewe lambs, which are then sold on to other flocks. These are the progeny of older hill ewes and may not present the same risk of carrying scrapie as the progeny of younger ewes (M. Dawson, personal communication). This study has shown that large flocks with many home-bred breeding sheep are likely to be a disproportionate risk for onward transmission. While interpretation of the absolute risk associated with particular flock types must be viewed with caution for all the reasons stated above, the marked difference in the relative risk, due to potential for persistent within-flock transmission and onward transmission via trade suggests that scrapie could be controlled by concentrating on these few, 'high-risk' flocks, obviating the need for radical control measures for most flocks. In order to determine the role that these flocks might play in the persistence and spread of scrapie in the British national flock, existing datasets characterizing the detailed flock-to-flock movement of sheep in GB ([@r25]; [@r13]; [@r40]) will have to be refined to consider only those movements that are important for scrapie transmission, most likely those movements involving breeding ewes. This will be the subject of future work. METHODS ======= Within-flock spread of infectivity. ----------------------------------- A difference equation model adapted from previous analyses [@r24]; [@r8]) was used to describe the within-flock spread of scrapie, incorporating age, infection and breeding structure, and the inheritance and impact of the PrP gene. The model takes into account flock size, the number of breeding ewes, proportion of ewes that are home-bred, proportion of ewes sold as replacements, and the genotype distribution of the flock. A full description of the model variables, parameters and equations is presented in [@r8]. Model outputs were compared to determine the impact of the different control and eradication strategies on different flock structures (Supplementary Material, available in JGV Online). Flock parameters. ----------------- In this model, flocks were characterized according to the breeding and commercial structure of the UK industry and the role of different flock types in onward transmission of disease. These characteristics were derived from data collected in the 2002 and 2006 anonymous postal surveys (see [@r34]), which describe the demographics of British sheep flocks with and without scrapie. Flock parameters selected for the model were based on flock type (pure-bred, commercial or mixed), flock size (25th and 75th percentiles), the proportion of home-bred sheep (25th and 75th percentiles), the likelihood of selling sheep on to other flocks and whether scrapie has ever been identified in the flock. Further information on flock parameter selection is presented in Supplementary Material. Ultimately, there were six different flock structures defined by the number of sheep in the flock and the proportion of home-bred sheep (Fig. [1](#f1){ref-type="fig"}, Supplementary Material). In the postal survey data, of flocks that have ever had scrapie, commercial flocks represented the majority (68 %). Pure-bred flocks (14.1 %) formed smaller proportions of the GB flock demographics ([@r34]). Pure-bred flocks were highly skewed with respect to the distribution of flock size (mean 366, median 153, range 9--3270) and proportion of home-bred sheep (mean 0.85, median 0.98, range 0--1.0). Similarly, commercial flocks were skewed with respect to flock size (mean 331, median 204, range 13--5512), but normally distributed with respect to the proportion of home-bred sheep (mean 0.58, median 0.53, range 0--1.0). Consistent with the postal survey data, it was assumed that 98 % of the breeding sheep within each flock are ewes and the remaining 2 %, rams. Although some pure-bred flocks will breed all their own replacements, in this model, all flocks were assumed to sell ewes as replacements. Genotype distribution. ---------------------- An average genotype distribution (equal to the national distribution) was used to determine the impact of different control strategies on the different flock types with respect to the number of sheep culled and infected sheep sold on to other farms. This was based on the prevalence of genotypes described in abattoir active surveillance data ([@r38], [@r39]). There were six genotypes described (in order of increasing susceptibility to infection): ARR/ARR, ARR/AXX, AXX/AXX, ARR/VRQ, AXX/VRQ and VRQ/VRQ. The AXX genotype refers to ARQ, AHQ and ARH genotypes. Genotype properties are described in Table [1](#t1){ref-type="table"} (see also [@r8]). In order to consider the impact of flocks with unusually large proportions of susceptible genotypes (i.e. not ARR/ARR or ARR/AXX), the number of infected sheep sold on to other farms was also compared for different flock types with different genotype distributions. We calculated a threshold susceptible proportion for each flock type. This threshold was the proportion of susceptible genotypes in the flock which reduced the maximum risk of onward transmission to ≤1 infected sheep sold per year. We examined a range of proportions of susceptible genotypes in each flock type (0, 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 0.95). The relative distribution of each of the six individual genotypes within each of these scenarios was derived using the ratios of each genotype in the abattoir screening data (Table [1](#t1){ref-type="table"}). Control and eradication strategies. ----------------------------------- The model was adapted to consider the impact of three proposed control and eradication strategies for scrapie and contrasted with a strategy where no intervention occurs for up to 15 years (the estimated average duration of a within-flock epidemic) ([@r12]). In the model, it was assumed that the presence of an index case has identified an ongoing scrapie epidemic, defined as a holding that has at least one case of home-bred scrapie in the last year or at least two cases of scrapie in the last 5 years, of which at least one was home-bred ([@r8]). The model assumed that most sheep get infected within the first few months of life but that no sheep under the age of 12 months die from scrapie ([@r8]). The model runs for a period of 5 years after the year of strategy implementation to consider the short-term impact of the differing strategies, assuming that the year of detection occurred anywhere between 1 and 15 years after introduction of at least one infected sheep. This was done to take into account preclinical infection that may otherwise go undetected due to the long incubation period of the disease. ### Strategy 1. In this strategy, no intervention was implemented. In these flocks, no scrapie-positive sheep were culled or removed from the flock at any point in time during the epidemic. ### Strategy 2. This strategy was modelled on the current CSFS (<http://www.defra.gov.uk/animalhealth/publications/NSPAC/>). All sheep are tested for genotype. This is a reactive policy enforced in response to detected disease in the flock. Susceptible genotypes were culled and resistant genotypes (ARR/ARR ewes and rams and ARR/AXX ewes) retained ([@r4]). For 2 years after the first year of implementation of this strategy, only resistant sheep were bought in (ARR/ARR ewes and rams and ARR/AXX ewes). After 2 years of restrictions, replacement sheep were bought in, in proportion to the distribution of the genotype of the national flock. ### Strategy 3. A preclinical diagnostic test was conducted each year over three consecutive years. A 3 year testing strategy was chosen initially to reflect similar time and trade restrictions imposed by the CSFS. Like the CSFS, for 2 years after the first year of implementation of this strategy, the model restricted the genotype of bought in sheep to those genotypes resistant to scrapie (ARR/ARR ewes and rams and ARR/AXX ewes). After 2 years of restrictions, replacement sheep were bought in, in proportion to the distribution of the genotype of the national flock. ### Strategy 4. A preclinical diagnostic test was conducted once (at a single time within a single year with subsequent restrictions enforced on sale, purchase and breeding on the holding for a period of 3 years), otherwise as for strategy 3. Live test parameters. --------------------- The live preclinical test implemented in strategies 3 and 4 was based on the experimental support for rectal biopsies of lymphoid tissue and is described in detail by [@r10]. This test is not currently a pen-side test. Under experimental conditions, when the prevalence of disease was 100 %, [@r10] found that the risk of a false-negative result in a preclinical rectal biopsy sample was 9.3 % if the sample contained 10 follicles. The probability of obtaining a sample containing at least 10 follicles was 87 %. The risk of a false-positive result was believed to be negligible (L. González, personal communication). For this model, it was assumed that every sheep in the flock (over a specified age) was tested and that at least 10 follicles were retrieved in every single biopsy sample. An initial individual test sensitivity of 90 % and specificity of 100 % were chosen for the model. Lower individual test sensitivities (\<90 %) were also investigated. Specificity was assumed to be 100 % in each scenario. PPV (the probability that given a positive test result, the sheep has the disease) and NPV (the probability that given a negative test result, the sheep does not have the disease) were calculated for different flock structure prevalences using different test sensitivities (90, 70 and 35 %). In the model, the live preclinical testing strategies were applied to sheep of different ages (greater than 20, 12 and 6 months). These ages were based on data from [@r10] which suggest that the first positive tests in the rectal mucosa appear at statistically similar average proportions of the incubation period of experimentally infected sheep (0.5 in AXX/AXX sheep, 0.49 in AXX/VRQ and 0.43 in VRQ/VRQ sheep). Although the chances of a positive test result increase with age, infected VRQ/VRQ sheep can be detected as early as 4--8 months post-infection; ARQ/VRQ and ARQ/ARQ sheep can be detected as early as 8 months post-infection. Most of these genotypes are detected at 16 months post-infection; ARR/VRQ sheep are detected as early as 24 months post-infection but most are detected at 36--48 months post-infection) (L. Gonzalez, personal communication) Model outputs. -------------- The mean number of breeding sheep culled due to scrapie and the mean prevalence of scrapie per year were calculated for each flock type and year past the initial intervention. In addition, data from the scrapie postal survey (2002) on all flocks were used to estimate the potential number of infected breeding sheep sold to other farms after each intervention for each flock type is investigated. A qualitative comparison of the number of sheep sold by each flock type indicates that there is a negligible difference between scrapie-positive flocks and scrapie-negative flocks, so data from all flocks were used. The number of sheep sold was recorded categorically in the scrapie postal survey (0, 1--5, 6--20, 21--50, 50--100 and greater than 100 sheep sold). In order to obtain the maximum number of breeding sheep sold on from an infected flock per annum, categorical data were interpreted in terms of the upper bounds of each category (i.e. 0, 5, 20, 50, 100 and \>100 sheep). An estimate of the upper bounds of the final category (\>100 sheep) was based on the average size of the breeding ewe flock in each flock type (in the final category), assuming that each breeding ewe produces at least one ewe lamb per year (Table [2](#t2){ref-type="table"}). This is an upper estimate for hill flocks, as most pure-bred ewes produce an average of 0.5--0.75 ewe lambs per year (M. Dawson, personal communication). The average prevalence of scrapie in the flock, after each strategy was implemented, was used to calculate the number of infected sheep sold on subsequently to other flocks each year for each flock type (Table [2](#t2){ref-type="table"}). For each flock type, the mean number of infected breeding sheep sold onwards per flock per year was calculated for years 1--15 of an epidemic. The mean number of infected breeding sheep sold on was divided by the flock size for each flock type to obtain a per capita onward risk so that direct comparisons could be made between different flock types. For each flock type, this output was multiplied by 100 to obtain the number of infected breeding sheep sold per 100 sheep in the flock. Model outputs were compared to determine the impact of the different control and eradication strategies on different flock structures. Supplementary Material ====================== ::: {.caption} ###### \[Supplementary Material\] ::: The authors would like to acknowledge Defra for funding this study (project SE0250). R. R. K. is funded by the Wellcome Trust, F. H. by the Royal Society. Postal survey data were provided by the Veterinary Laboratories Agency. Supplementary material is available with the online version of this paper. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Summary of the flock parameters used in the model and description of the potential number of infected breeding sheep sold to other farms by each flock type if scrapie is present in the flock over a 15 year period. The number of flocks in each flock category reported in the postal survey data are described in each bubble. Pure-bred flocks with small proportions of home-bred sheep were not examined in this model as these flock types are extremely rare. The percentage of the flock subgroup, major flock category (commercial or pure-bred) and total flocks (commercial and pure-bred) are reported, respectively, in parentheses. For example, in the high-risk category, 83 % (91/109) of large pure-bred flocks with a large percentage of home-bred sheep sold more than one infected breeding sheep each year in a 15 year epidemic. This high-risk group comprised 14 % (91/635) of all pure-bred flocks and 2 % \[91/(635+3832)\] of all pure-bred and commercial flocks. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Violin plot showing the distribution of flocks with respect to the average number of infected breeding sheep sold to other farms each year by each flock type in the first year of an epidemic. The Violin plot is similar to a box plot, except that it also shows the probability density of the data at different values. The line=1 represents one infected sheep sold on per year and denotes the cut-off point between low-risk and moderate-risk flocks. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### Mean number of infected breeding sheep sold by different flock types per year (in each year of an epidemic) after implementation of live-test strategies. The CSFS strategy reduced the number of infected sheep sold to zero in all cases (and is not shown on these graphs). The upper bars indicate the potential maximum number of infected breeding sheep sold per year (per 100 sheep in a flock). The minimum number of infected sheep sold per year (per 100 sheep in a flock) is zero in all years. The black line represents one infected sheep sold. :::  ::: ::: {#t1 .table-wrap} Table 1. ::: {.caption} ###### Average sheep genotypes: prevalence, relative susceptibility and mean age at onset of scrapie (reproduced from [@r8]) [na]{.smallcaps}, Not applicable. ::: **Genotype** **Prevalence in abattoir screening -- national genotype (%)** **Estimated relative susceptibility to scrapie** **Mean age of onset of scrapie symptoms (years)** -------------- --------------------------------------------------------------- -------------------------------------------------- --------------------------------------------------- ARR/ARR 19.5 0 [na]{.smallcaps} ARR/AXX 41.9 0.001 5.0 AXX/AXX 26.5 0.026 3.8 ARR/VRQ 5.5 0.119 5.9 AXX/VRQ 6.2 0.359 3.8 VRQ/VRQ 0.4 1.000 3.2 ::: ::: {#t2 .table-wrap} Table 2. ::: {.caption} ###### Prevalence and proportion of flocks within each flock type selling on sheep to other farms L, Large; S, small; P, pure-bred; C, commercial; H, home-bred. ::: **Flock type** **No. flocks in postal survey (% of all study flocks *n*=5022)** **Prevalence (range years 1--15)** **Proportion (and number) of each flock type selling sheep to other farms (by category of sheep sold)** **Estimated maximum number of sheep sold (in \>100 category)** -------------------------- ------------------------------------------------------------------ ------------------------------------ --------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------- ----------- ----------- ----------- ------------ ------ **Low-risk flocks** SPLH 147 (3 %) 1.0--0.2 % 0.32 (47) 0.05 (8) 0 (0) 0.63 (92) 0 (0) 0 (0) 50 SCLH 37 (1 %) 1.0--0.2 % 0.71 (466) 0.02 (10) 0.09 (59) 0.13 (86) 0.04 (28) 0.01 (9) 103 SCSH 516 (12 %) 0.4--\<0.001 % 0.83 (429) 0.004 (2) 0.03 (14) 0.05 (25) 0.04 (23) 0.04 (23) 100 **Moderate-risk flocks** LCSH 26 (0.6 %) 0.3--0.9 % 0.73 (103) 0 (0) 0.01 (2) 0.01 (1) 0.04 (6) 0.19 (27) 993 **High-risk flocks** LPLH 91 (2 %) 1--22 % 0.10 (11) 0 (0) 0.02 (2) 0.05 (5) 0.08 (9) 0.75 (82) 1210 LCLH 154 (3 %) 2--27 % 0.46 (154) 0.01 (2) 0.01 (4) 0.06 (21) 0.14 (48) 0.32 (106) 1373 :::

PubMed Central

2024-06-05T04:04:19.407472

2010-10-01

PMC3052602

INTRODUCTION ============ Chronic wasting disease (CWD) is a prion disease affecting free-ranging and captive cervids, including mule deer, white-tailed deer, Rocky Mountain elk and moose ([@r64], [@r65]; [@r63]; [@r6]). Like all mammalian prion diseases, which include Creutzfeldt--Jakob disease (CJD), kuru and variant CJD (vCJD) in humans and bovine spongiform encephalopathy (BSE) in cattle, the central event in CWD infection is the post-translational conversion of the host-encoded, cellular prion protein (PrP^C^), to an abnormal isoform, designated PrP^Sc^ ([@r44]; [@r15]). Progressive accumulation of PrP^Sc^ in the central nervous system ([@r23]) is associated with clinical signs of CWD which include weight loss, behavioural changes, excessive salivation, difficulty swallowing, polydipsia, polyuria, and ataxia prior to death ([@r64], [@r65]; [@r63]). International concern over CWD is growing as infected cervids have now been reported in 14 states in North America, two Canadian provinces and in South Korea ([@r30]; [@r63]; [@r52]; [@r51]). To date, CWD has not been reported in Europe, although surveillance has been limited. The prevalence of CWD infection can reach levels of up to 30 % in free-ranging herds in North America and up to 90 % in animals housed in CWD research facilities ([@r63]). Infectious prions in the saliva ([@r37]; [@r24]; [@r38]), urine ([@r24]) and faeces of CWD-infected animals ([@r55]) may underlie the highly efficient natural transmission of CWD among cervids through environmental contamination ([@r38]). Protease-resistant cervid prion protein has recently been demonstrated in an environmental water sample from a CWD endemic area ([@r42]). Despite efficient horizontal transmission of CWD prions among cervids, to date there is no clear evidence for natural disease transmission to other species. A recent survey for transmissible spongiform encephalopathy in scavengers of white-tailed deer carcasses in a CWD endemic area of Wisconsin found no evidence for cross-species transmission ([@r28]). Nevertheless, the zoonotic transmission of BSE prions ([@r18]; [@r27]; [@r11]; [@r4]; [@r57]) has dramatically highlighted the potential risk posed to humans from dietary exposure to CWD prions ([@r7]; [@r51]). Infectious prions are present in the blood ([@r37]), skeletal muscle ([@r2]) and fat ([@r46]) of CWD-infected deer and CWD prions have been shown to be experimentally transmissible after oral inoculation of elk and deer and cervid PrP expressing transgenic mice ([@r25]; [@r21]; [@r56]). Consumption of hunted deer and elk is widely practised in North America and a survey conducted by the American Red Cross and other blood banking establishments has reported that ∼40 % of USA blood donors have consumed venison obtained from the wild ([@r8]). To date, however, epidemiological surveillance has not indicated any link between human disease and CWD exposure ([@r7]; [@r39]; [@r1]). However, incubation periods in human prion disease even in the absence of a transmission barrier can exceed 50 years ([@r19], [@r20]). Accordingly, there has been intense research interest in establishing the host range of CWD prions through experimental transmissions to laboratory animals ([@r54]; [@r49]; [@r53]; [@r26]) and through the use of *in vitro* prion amplification systems ([@r48]; [@r33]). Concern that CWD prions might be transmissible to humans was heightened in 2005 by the finding that squirrel monkeys can be infected by intracerebral inoculation with CWD mule deer brain homogenate ([@r36]). However, a more recent study has shown that cynomolgus macaques (that are evolutionarily closer to humans) differ significantly from squirrel monkeys with respect to their susceptibility to infection with CWD prions, with no evidence for clinical disease in macaques at 70 months post-inoculation ([@r47]). Crucially however, because prion transmission barriers and prion strains are intimately related by conformational selection ([@r13]; [@r15]) the ability of CWD prions to propagate in humans cannot be inferred by studying the interaction of CWD prions with distinct (albeit highly conserved) PrP sequences from other species. To date, two studies have reported that transgenic mice expressing human PrP with methionine at polymorphic residue 129 are resistant to intracerebral challenge with CWD prions. The first of these studies used two lines of transgenic mice expressing human PrP at either one or two times the endogenous level of mouse brain. After inoculation with CWD-infected elk brain homogenate, none of these transgenic mice showed clinical signs of prion disease or detectable accumulation of abnormal PrP by either immunohistochemistry or immunoblotting ([@r31]). Although these mice are susceptible to infection with atypical BSE prions, their susceptibility to classic BSE prions or vCJD prions has not been reported ([@r32]). The second study used hemizygous transgenic mice expressing human PrP at two times the endogenous level of murine PrP expression in mouse brain. No evidence of clinical prion disease was observed following intracerebral challenge with CWD-infected elk, mule deer or white-tailed deer brain homogenate; however, importantly subclinical infection was not excluded ([@r54]). Susceptibility of these mice to infection with BSE or vCJD prions has not been reported. Here, to investigate further the potential risks for transmission of cervid prions to humans, we have transmitted mule deer CWD prions to lines of transgenic mice overexpressing human PrP two- to sixfold with either methionine or valine at polymorphic residue 129 in which we have extensive experience of transmission of a wide range of human acquired, sporadic and inherited prion disease isolates, including kuru and multiple vCJD cases ([@r16], [@r17], [@r18]; [@r27]; [@r61]). Extensive comparative data are available on transmission of multiple cattle BSE isolates ([@r27]; [@r4], [@r5]; [@r59]) as well as BSE experimentally passaged or naturally transmitted to multiple mammalian species and these mice are therefore suitable for comparative assessment of the zoonotic potential of CWD prions. RESULTS ======= Immunoblot analysis of CWD-infected brain and spinal cord --------------------------------------------------------- CWD-infected mule deer brain (from animal D10) and spinal cord (from animal D08) originated from captive animals housed at the Colorado Division of Wildlife, Wildlife Research Centre, Colorado, USA. Homogenates (10 % w/v) of these tissues were prepared in PBS and examined for proteinase K (PK)-resistant PrP by immunoblotting. Both samples showed a high level of cervid PrP^Sc^ (Fig. [1](#f1){ref-type="fig"}) with a PrP glycoform ratio that showed a dominant diglycosylated conformer, typical of that associated with CWD prions ([@r45]). In contrast, identical analysis of brain homogenates prepared from uninfected mule deer showed no detectable PK-resistant PrP (Fig. [1](#f1){ref-type="fig"} and data not shown). CWD prions do not transmit prion disease to transgenic mice overexpressing human prion protein ---------------------------------------------------------------------------------------------- PrP^Sc^-positive CWD-infected brain and spinal cord homogenates were used to prepare inocula for transmission studies in transgenic mice overexpressing human PrP with either methionine or valine at polymorphic residue 129. 129MM Tg35, 129MM Tg45 and 129VV Tg152 transgenic mice overexpress human PrP in brain at levels of two, four and six times that of human brain, respectively ([@r17], [@r18]; [@r27]; [@r4]). These lines of mice have been extensively used by us for over 15 years and have proven susceptibility to infection with human or BSE prions ([@r17], [@r18]; [@r27]; [@r4], [@r5]; [@r59], [@r60], [@r61]). Following intracerebral inoculation with CWD brain or spinal cord, groups of 10 transgenic mice were observed throughout their life time for clinical signs of prion disease. As reported in Table [1](#t1){ref-type="table"}, we observed no clinical prion disease in any inoculated mouse, including those with post-inoculation intervals greater than 700 days (Table [1](#t1){ref-type="table"}). Accordingly, brains from mice culled as a result of inter-current illness or senescence were examined for subclinical prion transmission. In all cases examined, pathological PrP accumulation in brain was undetectable by either immunoblotting (Fig. [2](#f2){ref-type="fig"}, Table [1](#t1){ref-type="table"}) or immunohistochemistry (Fig. [3](#f3){ref-type="fig"}, Table [1](#t1){ref-type="table"}). Futhermore, neuropathological examination of CWD-inoculated transgenic mouse brain, showed no evidence of spongiform change or gliosis consistent with prion disease and their appearance was indistinguishable from the brain of age matched control mice inoculated with normal mule deer brain (Fig. [3](#f3){ref-type="fig"} and data not shown). In summary, we conclude that intracerebral challenge of these transgenic mice with CWD prions caused no clinical or subclinical prion infection, indicating that both methionine and valine 129 polymorphs of human PrP are refractory to pathological conversion by CWD prions. DISCUSSION ========== In this study, we have shown that transgenic mice overexpressing human PrP of both residue 129 polymorphic forms, known to be susceptible to a wide range of human and other prions, are highly resistant to infection with mule deer CWD prions. These findings agree with those of others who have previously reported an inability of CWD prions to transmit disease to transgenic mice expressing human PrP 129 methionine ([@r31]; [@r54]) or a poor ability of human PrP to act as a substrate for CWD prions in *in vitro* conversion assays ([@r48]; [@r33]). Importantly, the transgenic mice used in our study have proven susceptibility to infection with BSE prions \[[@r27]; [@r4], [@r5]; [@r59] (Table [1](#t1){ref-type="table"})\]. The negative transmissions that we report here therefore strongly support the conclusion that the transmission barrier associated with the interaction of human PrP and these CWD prions is greater than that associated with interaction of human PrP and the prion strain causing epizootic BSE in cattle. The failure to show propagation of CWD prions using human PrP as a substrate either *in vivo* in transgenic mice or *in vitro* in biochemical conversion assays suggests that potential zoonotic threat from CWD is low. However, an important caveat in this regard is that the number of prion strains propagated in CWD is currently unknown ([@r10]; [@r49]; [@r22]; [@r3]). Because prion strains can adapt and mutate on passage in new species ([@r15]; [@r9]; [@r12]; [@r14]), and also within species as a result of PrP polymorphisms and other genetic factors ([@r4]; [@r34]; [@r59]; [@r40]; [@r35]), the risk that each prion strain poses to public health must be evaluated directly. There is now growing evidence that polymorphisms of cervid PrP may dictate prion strain selection ([@r43]; [@r41]; [@r22]; [@r3]). Thus, while the available experimental data appear reassuring, further transmission studies will be of vital importance to evaluate the properties of distinct cervid prion strains as they are isolated. METHODS ======= Mule deer tissues. ------------------ Importation, storage and use of CWD-infected tissues was performed under licence granted by Defra under the terms of the Importation of Animal Pathogens Order 1980. CWD-infected mule deer brain (from animal D10) and spinal cord (from animal D08) originated from naturally infected captive animals from Colorado, USA that had clinical signs consistent with terminal stages of prion disease. CWD-infection in these animals was confirmed by the presence of histopathological lesions in the brain, including spongiform degeneration of the perikaryon, by immunohistochemical or immunoblot detection of disease-related PrP and by positive transmission of prion disease to transgenic mice expressing cervid PrP ([@r10]; [@r2]; [@r22]). Brain from uninfected mule deer fawns (FPS 6.98 and FPS 3.98) was used as negative controls. Transgenic mice. ---------------- Transgenic mice homozygous for a human PrP 129V transgene array and murine PrP null alleles (*Prnp^o/o^*) designated Tg(HuPrP129V^+/+^ *Prnp^o/o^*)-152 mice (129VV Tg152 mice) or homozygous for a human PrP 129M transgene array and murine PrP null alleles (*Prnp^o/o^*) designated Tg(HuPrP129M^+/+^ *Prnp^o/o^*)-35 mice (129MM Tg35 mice) or Tg(HuPrP129M^+/+^ *Prnp^o/o^*)-45 mice (129MM Tg45 mice) have been described previously ([@r17], [@r18]; [@r27]; [@r4], [@r5]; [@r59], [@r60], [@r61]). Transmission studies. --------------------- All procedures were carried out in a microbiological containment level 3 facility with strict adherence to safety protocols. Care of mice was according to institutional guidelines. Mule deer tissues were prepared as 10 % (w/v) homogenates in sterile PBS lacking Ca^2+^ and Mg^2+^ ions by serial passage through needles of decreasing diameter, and subsequently diluted to 1 % (w/v) in PBS. Following intracerebral inoculation with 30 μl of 1 % (w/v) tissue homogenate as described previously ([@r4], [@r5]; [@r59]), mice were examined daily and were killed if exhibiting signs of distress or once a diagnosis of clinical prion disease was established. Brains from inoculated mice were analysed by PrP immunoblotting or immunohistochemistry and by neuropathological examination. Immunoblotting. --------------- All procedures were carried out in a microbiological containment level 3 facility with strict adherence to safety protocols. Tissue homogenates (10 % w/v) were prepared in PBS lacking Ca^2+^ or Mg^2+^ ions. PK digestion (50 or 100 μg ml^−1^ final protease concentration, 1 h, 37 °C), electrophoresis and immunoblotting was performed as described previously ([@r58], [@r62]). Immunoblot detection was performed using anti-PrP monoclonal antibody ICSM35 (D-Gen) for cervid PrP or 3F4 ([@r29]) for human PrP in transgenic mice. Brain homogenates scored negative for PrP^Sc^ after analysis of 10 μl 10 % (w/v) brain homogenate were re-analysed by sodium phosphotungstic acid precipitation of PrP^Sc^ ([@r50]) from 250 μl of 10 % (w/v) brain homogenate as described previously ([@r58]). Neuropathology and immunohistochemistry. ---------------------------------------- All steps prior to prion decontamination with formic acid were performed within a microbiological containment level 3 facility with strict adherence to safety protocols. Brain was fixed in 10 % buffered formal saline and then immersed in 98 % formic acid for 1 h and paraffin wax embedded. Serial sections of 4 μm thickness were pre-treated by boiling for 10 min in a low ionic strength buffer (2.1 mM Tris, 1.3 mM EDTA, 1.1 mM sodium citrate, pH 7.8) before exposure to 98 % formic acid for 5 min. Abnormal PrP accumulation was examined using anti-PrP monoclonal antibody ICSM35 (D-Gen) on a Ventana automated immunohistochemical staining machine (Ventana Medical Systems) using proprietary secondary detection reagents (Ventana Medical Systems) before development with 3′3-diaminobenzedine tetrachloride as the chromogen ([@r62]). Harris haematoxylin and eosin staining was done by conventional methods. Appropriate positive and negative controls were used throughout. Photographs were taken on an ImageView digital camera and composed with Adobe Photoshop. We thank our biological service team for animal care and R. Young for the preparation of figures. We thank Michael W. Miller and the Colorado Division of Wildlife for access to the CWD-infected deer samples. This research was funded by the Medical Research Council (UK) and the European Union. Conflict of interest statement: J. C. is a Director and J. C. and J. D. F. W. are shareholders and consultants of D-Gen Limited, an academic spin-out company working in the field of prion disease diagnosis, decontamination and therapeutics. D-Gen markets the ICSM35 antibody used in this study. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Detection of PrP^Sc^ in the brain and spinal cord from CWD-infected mule deer. Immunoblots show the analysis of 5 μl aliquots of 10 % (w/v) homogenates of uninfected mule deer brain or CWD-infected mule deer brain or spinal cord, before (−) or after (+) digestion with PK. Immunoblots were analysed by enhanced chemiluminescence with anti-PrP monoclonal antibody ICSM35. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Failure to detect PrP^Sc^ in the brain of CWD prion-inoculated transgenic mice. The high sensitivity immunoblot using anti-PrP monoclonal antibody 3F4 shows PK-digested sodium phosphotungstic acid pellets recovered from 10 % (w/v) transgenic mouse brain homogenates. Lanes 1 and 2, positive controls showing efficient recovery of PrP^Sc^ after spiking 2 μl 10 % (w/v) BSE-inoculated 129MM Tg45 and 129MM Tg35 transgenic mouse brain homogenates ([@r4]) into 100 μl 10 % (w/v) uninfected 129MM Tg45 and 129MM Tg35 mouse brain homogenates, respectively. Lane 3, PK-digested sodium phosphotungstic acid pellet from 250 μl 10 % (w/v) brain homogenate from a 129MM Tg45 mouse inoculated with normal mule deer brain. Lanes 4--9, PK-digested sodium phosphotungstic acid pellets from 250 μl 10 % (w/v) brain homogenates from 129MM Tg35, 129MM Tg45 and 129VV Tg152 mice inoculated with CWD-infected mule deer brain. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### Failure to detect abnormal PrP deposition in the brain of CWD prion-inoculated transgenic mice. Representative PrP immunohistochemistry using anti-PrP monoclonal antibody ICSM35. Panels (a--f) show no abnormal PrP deposition in either the hippocampus or thalamus of 129VV Tg152, 129MM Tg45 or 129MM Tg35 mice inoculated with CWD-infected brain homogenate. These mice were culled 517, 529 and 559 days post-inoculation, respectively. Panels (g) and (h) show hippocampus and thalamus from an age matched control 129MM Tg45 mouse brain inoculated with 10 % (w/v) uninfected mule deer brain homogenate. In contrast, extensive deposition of abnormal PrP is seen in the hippocampus and thalamus of a BSE-infected 129MM Tg45 mouse with subclinical prion disease (panels i and j) ([@r4]). Bar, 500 μm. :::  ::: ::: {#t1 .table-wrap} Table 1. ::: {.caption} ###### Primary transmission of CWD and BSE prions to transgenic mice Data for BSE transmissions have been published previously ([@r27]; [@r4]). ::: **Transgenic line** **CWD brain** **CWD spinal cord** **BSE brain** --------------------- --------------- --------------------- --------------- 129VV Tg152 0/8† 0/6‡ 10/26 129MM Tg45 0/7§ 0/6\|\| 9/12 129MM Tg35 0/9¶ 0/5\# 14/49 \*All mice were inoculated with 30 μl of 1 % (w/v) tissue homogenate. Attack rate is defined as the total number of both clinically affected and subclinically infected mice as a proportion of the number of inoculated mice. Subclinical prion infection was assessed by sodium phosphotungstic acid precipitation of 250 μl 10 % brain homogenate and analysis for PrP^Sc^ by immunoblotting and/or immunohistochemical examination of brain. †Mice culled at 274, 316, 321, 436, 517, 517, 587 and 781 days post-inoculation. ‡Mice culled at 354, 364, 463, 541, 704 and 724 days post-inoculation. §Mice culled at 322, 322, 395, 400, 529, 656 and 736 days post-inoculation. \|\|Mice culled at 275, 345, 396, 462, 462 and 532 days post-inoculation. ¶Mice culled at 341, 559, 662, 662, 680, 707, 707, 747 and 748 days post-inoculation. \#Mice culled at 392, 414, 542, 699 and 732 days post-inoculation. :::

PubMed Central

2024-06-05T04:04:19.411165

2010-10-01

PMC3052717

INTRODUCTION ============ Hepatitis C virus (HCV) is a major cause of liver disease globally ([@r1]). The virus is able to evade host innate and adaptive immune responses in immunocompetent adults and to set up persistent infection in the majority of people. Those persistently infected are at risk of progressive liver fibrosis, cirrhosis and cancer. A minority of those infected, however, clear the virus spontaneously, a feature that is reproduced in animal models ([@r4]). The immunological response to HCV has an important bearing not only on the acute outcome (i.e. persistent infection versus spontaneous resolution), but also potentially on the long-term outcome in chronic carriers, although our understanding of the immune responses in acute outcome is more complete. Successful outcome is associated with the maintenance of broadly directed CD4^+^ and CD8^+^ T-cell responses, with maintained functionality ([@r4]). Specific HLA class I and II molecules are associated with a successful response ([@r38]; [@r40]), and depletion studies in animals strongly support the critical role of both CD4^+^ and CD8^+^ T-cell subsets in acute control ([@r20]; [@r52]). The responses in those with established persistent infection are typically described as highly attenuated, especially in the blood ([@r4]; [@r26]; [@r32]; [@r33]; [@r45]). Chronically infected human and chimpanzee studies based on proliferation responses in blood, or analysis of T-cell lines after stimulation, have typically found either no or only very limited responses ([@r14]; [@r39]). In contrast to these studies, recent *ex vivo* analyses based around gamma interferon (IFN-*γ*) secretion have indicated some CD4^+^ T-cell responsiveness in those with persistent infection: we and others have found responses to pools of core peptides in a range of patients with chronic disease, associated with low proliferative capacity and loss of interleukin (IL)-2 secretion ([@r2]; [@r22]; [@r46]; [@r49], [@r50]); the responses obtained were reproducible and robust and were not detected in healthy normal controls. Responses using an identical assay against antigens from genotype 1 NS3--NS5b typically yield minimal responses (undetectable or weak) and are difficult to study further. Loss of CD4^+^ T-cell responses to these peptides is seen in the context of human immunodeficiency virus (HIV) infection, associated with an increased viral load ([@r22]). The mechanism behind the failure of CD4^+^ T-cell responses to contain virus in chronic infection has not been established, nor has the mechanism underlying their low levels in blood ([@r4], [@r5]; [@r26]). Individual reports have indicated that immune escape may occur in epitopes targeted by HCV-specific CD4^+^ T cells ([@r15], [@r16]; [@r42]; [@r54]; [@r55]). Variation within epitopes has also been described as affecting T-cell cytokine secretion ([@r54]; [@r55]). It is well established that escape from HCV-specific CD8^+^ T cells can occur and may be critical for persistence ([@r10]; [@r13]; [@r17]; [@r44]; [@r56]). However, as CD4^+^ T-cell responses have been considered difficult to detect during chronicity, this issue has not been systematically addressed for T helper populations. To address this, we analysed CD4^+^ T-cell responses in a large, well-defined cohort of individuals with chronic HCV genotype 1 infection. We systematically analysed the *ex vivo* CD4^+^ T-cell response and defined the fine specificity of this response in relation to the sequence of autologous virus. Surprisingly, we found a persistent IFN-*γ* response in the majority of donors, often targeting multiple epitopes, including a newly defined immunodominant peptide in the core. We evaluated the relationship between T-cell responses targeting individual peptides and the sequence variation within those peptides, and defined the functional consequences of viral mutation and the impact of CD4^+^ T-cell escape in this cohort. RESULTS ======= *Ex vivo* CD4^+^ T-cell responses are detectable in the majority of donors -------------------------------------------------------------------------- In this study, we excluded CD8^+^ T-cell responses at the outset by the use of CD8-depleted peripheral blood mononuclear cells (PBMCs). We measured *ex vivo* responses in an IFN-*γ* enzyme-linked immunosorbent spot (ELISPOT) assay using established robust methods and identified a response against the pool of genotype 1 core peptides in 32/61 genotype 1-positive donors. These donors were collected prospectively and independently of previous immunological or clinical data (Table [1](#t1){ref-type="table"}). Alanine transaminase (ALT) levels, viral loads, age and prior treatment status did not differ significantly between the two groups. This frequency of response is very similar to that defined in two previous studies on HCV^+^ chronically infected cohorts using whole undepleted PBMCs and peptide pools ([@r49], [@r50]). Fine specificity of CD4^+^ T-cell responses in persistent infection ------------------------------------------------------------------- In this study, we wished initially to define the peptide targets of the observed CD4^+^ T-cell responses. The results of this are illustrated in Table [2](#t2){ref-type="table"} and Fig. [1(a)](#f1){ref-type="fig"}. Table [2](#t2){ref-type="table"} shows the raw data for individual responses to single peptides derived from the core peptide sequence, as well as responses to a pool of the 18 core peptides (core pool). Only data from those responding to one or more peptides are shown (*n*=32). Individuals responded to diverse peptides across the sequence \[spot forming cells (s.f.c.) per 10^6^ CD8-depleted PBMCs, mean 170; range 50--732\]. If responses to two adjacent peptides were considered as a single target (a conservative estimate), we observed, on average, responses to two epitopes per subject (median 2; 15 subjects with one response, 12 subjects with two responses, four subjects with three responses and one subject with four responses). The group data are shown in Fig. [1(a)](#f1){ref-type="fig"}. This revealed responses to the majority of peptides tested, with the commonest targets being peptide 31--50 ([@r31]; [@r37]) and peptide 61--80 ([@r35]). Both of these peptides were recognized in 9/32 individuals. The magnitude of the response varied from 50 to 732 s.f.c. per 10^6^ CD8-depleted PBMCs (Table [2](#t2){ref-type="table"}) and representative ELISPOT wells are shown in Fig. [1(b)](#f1){ref-type="fig"}. These responses were reproduced in repeat assays, and titrations of the individual peptide, including epitope 61--80, showed that concentrations as low as 1--0.1 μg ml^−1^ were still able to stimulate 50 % maximal responses (Fig. [1c](#f1){ref-type="fig"}). Impact of T-cell responses on viral sequence -------------------------------------------- We next addressed the relationship between the presence of CD4^+^ T-cell responses and viral sequence evolution. This is important for three reasons: (i) to define to what extent the detectable T-cell responses are targeting autologous sequences; (ii) to define the impact of potential T-cell-mediated selection pressure; and (iii) to analyse whether the non-responder group had accumulated significant mutations as a potential cause of T-cell failure. Sequences were therefore analysed as bulk sequence and, in addition, clonally derived sequences were obtained in a subset of patients. Firstly, we analysed the relatedness of sequences to address whether responder and non-responder sequences clustered separately. Phylogenetic analysis of the core region sequences revealed a typical subdivision into genotype 1a and 1b subtypes (Fig. [2](#f2){ref-type="fig"}). No correlation was observed between HCV subgenotype and the CD4^+^ T-cell response. In one case, a CD4 responder and non-responder both shared identical core regions, placing them on the same branch of the tree. Forty-six sequences were obtained: 23 CD4 responder core sequences and 23 CD4 non-responder sequences. Amongst the responders, T-cell responses to previously defined peptides were observed in 12 donors and responses to novel peptides (i.e. not previously published as epitopes) in ten donors. From these, all but two donors (304 and 379) showed conserved sequences within the peptides targeted (i.e. identical to the peptides used in the assay). This indicated that the selection of escape mutants through immune pressure from CD4^+^ T cells targeting the HCV core region appears to be rare, at least in this group of individuals. However, single variants were observed within previously defined epitopes, peptide 61--80 (A68V) in one individual and peptide 141--160 (A147V) in another (Fig. [3a](#f3){ref-type="fig"}). Two additional mutations were observed in both individuals within peptide 101--120 (donor 304: T110S; donor 379: S106G); however, CD4 responses were not detected at this or neighbouring peptides. To determine the frequency at which these variants were observed within these two donors, the core--E1 region was cloned and sequenced (donor 304: *n*=21 sequences; donor 379: *n*=14 sequences). At aa 68 within donor 304, the mutation to valine was observed in the majority of the cloned population (Fig. [3b](#f3){ref-type="fig"}; 71.4 %). Within donor 379, the quasispecies population at amino acid position 147 remained identical to the mutant variant valine. To establish whether T-cell reactivity was linked overall to positive selection within the core, the frequencies of synonymous (dS) and non-synonymous (dN) substitutions and mean pairwise diversity (*π*) were calculated for the CD4 responders and non-responders; these gave an indication of, respectively, the putative levels of selection and nucleotide diversity within each population (Table [3](#t3){ref-type="table"}). Patients exhibiting a CD4^+^ T-cell response to the core peptides did not display a greater level of nucleotide diversity or level of selection (*P*\>0.05). The values of dS/dN were substantially lower, as expected, in E1 compared with the core, but this did not differ between the two groups. Further analysis identified no evidence for significant positive selection at individual codons within the core region for both CD4 responders and non-responders. Cloning of the core--E1 region for patients 304 and 379 also identified no positively selected sites. Functional impact of epitope mutants ------------------------------------ To test whether the donor CD4^+^ T cells were able to recognize autologous viral sequence in the two cases where variants were observed (donors 304 and 379), *ex vivo* ELISPOT assays were performed using wild-type and mutant peptides (Fig. [3c](#f3){ref-type="fig"}). The peptides were titrated to determine the functional avidity of the response. In both cases, there was a clear-cut lack of recognition of the autologous mutant peptide, despite consistent reactivity against the wild-type peptide, suggesting in both cases that CD4^+^ T cells were unable to respond to the endogenous variant. DISCUSSION ========== Virus variation is a major feature of HCV infection, and immune escape from cellular and humoral immune responses is thought to play a significant role in the evolution of chronic infection ([@r4], [@r5]; [@r10]; [@r44]). During chronic infection, it is well recognized that there is a loss of CD8^+^ and CD4^+^ reactivity in the blood towards a range of peptides, compared with patients where infection has been resolved ([@r4]; [@r9]; [@r12]; [@r32]; [@r36]; [@r39]; [@r49], [@r50]). Although a link between immune pressure mediated by CD8^+^ T cells and selection of immune escape variants has been clearly shown ([@r4], [@r5]; [@r9]), few comparable data exist for CD4^+^ T cells. The impact of virus variation on T-cell function can be profound ([@r54]), and CD4^+^ escape has been documented in other viral infections ([@r6]), as well as in specific cases in HCV ([@r42]). However, evaluation of such CD4^+^ T-cell responses in chronic hepatitis C is extremely difficult as they are typically described as 'weak' or 'absent' in persistent infection. However, the use of conventional assays of T-cell proliferation in these analyses may miss populations of CD4^+^ T cells with different functional profiles. We have recently shown using fresh PBMCs in both ELISPOT assays and a carboxyfluorescein succinimidyl ester proliferative evaluation ([@r49], [@r50]) that certain CD4 populations are maintained, but that they are low in both IL-2 production and proliferative capacity, features similar to HIV-specific CD4^+^ T-cell populations ([@r11]). Assays may also be limited by substantial sequence mismatching between the viral antigens used in the assays and those present in the donor\'s circulating virus -- this matters particularly in cases where superinfection has occurred ([@r21]; [@r48]). In this study, we first sought to examine HCV-specific CD4^+^ T-cell responses in a large, carefully selected cohort of chronic HCV patients in order to clearly map the frequency of the peptide-specific response. The HCV genotype 1 core was selected as an appropriate target due to previously identified reproducible responses to this antigen in studies of persistently infected patients ([@r2]; [@r22]; [@r49], [@r50]; [@r56]). In addition, the core is a relatively conserved protein and, as a result, sequence variation between the peptides used in the study and the viral strains present in donors should be limited. Peptides were tested individually in order to analyse the breadth of responses and to avoid competitive inhibition. The use of CD8-depleted PBMCs and a sensitive *ex vivo* ELISPOT assay for IFN-*γ* enabled us to detect relatively low-frequency cell populations and those with limited proliferative capacity, and thus provided a specific and accurate picture of the HCV CD4^+^ population in the cohort. Responses to non-structural antigens were very weak in this study, and similar results were also obtained using independently generated overlapping peptide sets spanning the entire genome (data not shown). Although such proteins would represent ideal targets for future studies, the lack of responses to these antigens in *ex vivo* assays severely limits the practicality and scale of such analyses. HCV core-specific CD4^+^ T cells were readily detectable in just over 50 % of chronically infected patients using the IFN-*γ* ELISPOT assay, similar to results found in other studies ([@r49], [@r50]). The majority of core peptides were recognized, but responses focused particularly on aa 31--50 and 61--80. Within an individual, up to four different regions were recognized, with the majority of donors recognizing two or more distinct peptides even within this small region of the viral genome. Such data suggest that the level of multi-specificity of *ex vivo* responses in blood during chronic infection may previously have been underestimated. Even our analysis may be an underestimate, as the use of overlapping peptides as described might still fail to detect certain responses. The HLA class II types of those responding to individual peptides were not consistent in each case (data not shown), but data from other studies of virus-specific CD4^+^ T cells have shown that peptides typically bind and are presented by diverse HLA class II molecules ([@r19]; [@r31]). However, the finding of highly targeted peptides within this conserved region, even though it is an underestimate of the total antiviral CD4^+^ T-cell response, may still provide an important focus in future for analyses of CD4^+^ T-cell responses, e.g. with MHC class II tetramers. Sequencing of the core region of autologous virus revealed only two mutations within the recognized epitope, indicating that escape from recognition by specific CD4 cells is not a common occurrence. The CD4^+^ T-cell non-responder group acted as important controls in this respect, as it could be argued that the responses are most likely to be sustained in those with intact sequences; however, in the non-responder group, mutation was also rare. Phylogenetic and selection analysis did not resolve the two groups, and no signature of immune selection was detected within the region. Larger studies looking for HLA footprints may be required to define the impact of relatively infrequent selection events ([@r44]). In the two cases where mutations did occur, these were novel mutations not found in sequence databases ([@r7]; [@r23]) and both were associated with a functional impact on T-cell recognition. In the absence of the infecting strain, it is not possible to prove that the mutations occurred within the donors. However, in the case of donor 304, infection with HCV occurred as a result of a blood transfusion with no further apparent exposure. This makes it unlikely that superinfection with a mutant strain had occurred, but rather that this change has emerged within the host through immune selection pressures, with two-thirds of the viral clonal population possessing the A68V mutation. As the core is a conserved viral gene, the failure to generate escape mutants may be related to the fitness cost to the virus of such changes. Interestingly, the A147V mutation lies at the beginning of a structural helix in the lipid droplet-binding domain (D2) of the core. [@r51] showed that this amino acid substitution could enhance virus production through improved binding of the core to lipid droplets. Thus, loss of fitness does not seem to be a major argument for conservation of this residue. The impact of viral mutation may be greater on other sites within the core, although escape mutation sites in conserved regions of the core and NS3, which are targeted by CD8^+^ T cells, have been identified ([@r27]; [@r44]; [@r53]). The impact of virus-specific CD4^+^ T cells on virus sequence and vice versa have been addressed in previous studies, although with more limited patient and hence responder numbers. In a single case report, induction of substantial CD4^+^ T-cell responses against HCV in a chimpanzee was associated temporally with the emergence of a mutation that eliminated recognition ([@r42]). In studies of T-cell lines derived from individuals with chronic infection, it has been observed that mutations can impact on the functional responses of T cells, including a switch away from a Th1 phenotype ([@r54]). Despite these data, the overall role of virus escape in evasion of CD4^+^ T cells has not been fully addressed. Our data suggest that the phenomenon may occur, but that it is not common, at least among readily detectable responses. A recently performed longitudinal analysis of four infected chimpanzees concurs with this conclusion; these mutations occurred only rarely in this chronically infected HCV model ([@r18]). Exactly what the impact is of such persistent CD4^+^ T-cell responses with intact viral sequences on viral load and liver pathology is still not clear and may demand a larger study, incorporating larger numbers of antigenic targets. Studies to define the impact of target-specific CD8^+^ T-cell responses on HIV viral load (a setting where T-cell responses are much more readily detectable) have required cohorts in the order of 300--400 ([@r24]). Alternatively, it may be that similar studies of liver-derived cells may be required to define the specificity and magnitude of the responses and their function at the site of infection ([@r41]; [@r47]). Longitudinal studies may also reveal specific mutations, although these patients have been chronically infected for many years and acute infecting sequences are not available. For studies of CD8^+^ T-cell responses, cross-sectional approaches have revealed very obvious 'footprints' using a bulk sequencing approach, with similar or even smaller patient numbers, so although there are limitations in the approach used, it can be quite sensitive if selection pressure is strong ([@r13]). In conclusion, we observed sustained CD4^+^ T-cell responses in a large group of individuals persistently infected with HCV genotype 1, which targeted multiple peptides within the core protein. The responses were not associated with a clear virological footprint, and virologically responders and non-responders did not differ significantly, but in those cases where this was observed, mutation had an impact on T-cell recognition. This is inconsistent with a major role for immune escape in impairment of CD4^+^ T-cell responses against HCV, although it may occur in specific instances or in distinct antigenic targets. In this study, we focused on responses to the HCV core in genotype 1; it will be of importance to analyse whether the same process occurs in those responses targeting other viral proteins and/or in other genotypes, which may be under different functional constraints. Similarly, although cross-sectional studies do have significant power to detect escape, sequential studies of patients tracked through acute disease may provide an important alternative strategy. Defining the mechanisms underlying the failure of CD4^+^ T cells to contain HCV and the factors that determine their magnitude and function, longitudinally and especially in the liver, remain important goals in future studies. METHODS ======= Study subjects. --------------- Sixty-one subjects were included in the study from the Hepatitis Clinic at the John Radcliffe Hospital, Oxford, UK. All patients had persistent genotype 1 HCV infection and consented according to a locally approved protocol (COREC 04.OXA.010). Patients who had received treatment ending within the last 12 months were excluded. Of the 61 patients, 15 had received prior unsuccessful treatment with interferon and/or ribavirin (mean end 4.1 years previously, range 4--14 years). The clinical details of the patients are shown in Table [1](#t1){ref-type="table"}. T-cell assays ------------- ### IFN-*γ* ELISPOT. Fresh blood was obtained from the 61 individuals and PBMCs were separated on a density gradient. CD8^+^ cells were depleted using magnetic beads (Dynal) and the CD8^−^ PBMCs assayed in an IFN-*γ* ELISPOT assay (MabTech) using 2×10^5^ cells per well, against a pool of 18 core peptides (10 μg ml^−1^ final concentration for each peptide) to observe overall responses and a panel of 18 individual overlapping 20mer peptides covering the HCV genotype 1 core (aa 1--191; see Supplementary Table S1, available in JGV Online; 10 μg ml^−1^ final concentration) to provide fine detail. In addition, recombinant genotype 1 NS3--NS5 (Chiron; 2 μg ml^−1^), cytomegalovirus-infected cell lysate (Virusys; 2 μg ml^−1^) and phytohaemaglutinin were included as positive controls. Each antigen was tested in duplicate wells and the frequency of IFN-*γ*-producing cells was calculated by subtracting the mean number of s.f.c. per 10^6^ CD8-depleted PBMCs in the negative-control wells (cells/medium alone) from the mean number of s.f.c. in the test wells ([@r49]). A positive response was regarded as one in which the difference above the negative-control value of the s.f.c. per well was calculated to be significant (*P*\<0.05; Excel [binomdist]{.smallcaps}). In specific experiments, a pool of genotype 1-derived NS5A peptides, consisting of 40 overlapping 18mers (BEI Resources) was used. In further experiments to define the efficacy of the responses or examine the effect of mutations on T-cell recognition, wild-type and/or mutant peptides were tested in serial dilutions in RPMI 1640 in pairs, using the ELISPOT technique as above. ### HCV sequence evaluation. Viral RNA was extracted from plasma samples using a Viral RNA Extraction kit (Qiagen). Using a combined reverse transcription and first-round PCR step to amplify the core--E1 region, a 5063 bp external fragment was amplified using 10 pmol of primers utr-246 (5′-GACTGCTAGCCGAGTAGTGTTG-3′) and NS-5315 (5′-CGACCTCYARGTCNGCYCACATRC-3′) ([@r34]). A second-round PCR was performed with the inner primer utr-277 (5′-CCTTGTGGTACTGCCTGATAG-3′) and a modification of the outer primer C-E1 (5′-GTDGGNGACCARTTCATCATCAT-3′) ([@r8]). Using a SuperScript III One-Step RT-PCR kit with Platinum *Taq* DNA polymerase (Invitrogen), RT-PCR cycling conditions were as follows: 55 °C for 30 min and 94 °C for 2 min, followed by 39 cycles of 15 s at 94 °C, 30 s at 57 °C and 5.5 min at 68 °C, with a final extension of 68 °C for 10 min. The inner PCR conditions were as follows: 94 °C for 2 min and ten cycles of 15 s at 94 °C, 30 s at 56 °C and 1 min at 72 °C, followed by 20 cycles of 15 s at 94 °C, 30 s at 56 °C and 1 min increasing by 5 s every cycle at 72 °C, with a final extension of 72 °C for 20 min using a high-fidelity *Taq* DNA polymerase (Expand High Fidelity PCR System; Roche). PCR fragments were gel purified (Qiagen) and the population was sequenced bidirectionally using Prism Big Dye (Applied Biosystems) on an ABI 3100 DNA automated sequencer. Where necessary, PCR products were also cloned (TOPO TA; Invitrogen) and the DNA purified as above (Qiagen) and sequenced as above. Sequences were edited using Sequencher v4.8 (Gene Codes) and aligned with the Se-Al v2.0 sequence alignment editor ([@r43]). Kimura\'s two-parameter model was implemented using [mega]{.smallcaps} 4.0 to create neighbour-joining phylogenetic trees ([@r25]). Bootstrap analyses were carried out with 1000 replicates. Levels of dS and dN mutations and *π* were calculated using [mega]{.smallcaps} 4.0. To detect evidence of selection at individual codons, single likelihood ancestor counting was used as implemented in DataMonkey ([@r29], [@r30]). The analysis was conducted with both the HKY85 and general reversible models of nucleotide substitution with a cut-off *P* value of 0.1. ### HCV load. Based on the method outlined by [@r28], HCV viral load quantification was carried out using real-time PCR with SYBR Green I detection (Roche). The 5′ HCV non-coding region was transcribed into cDNA using primer RC21 (5′-CTCCCGGGGCACTCGCAAGC-3′) ([@r3]) and following the manufacturer\'s instructions for SuperScript II Reverse Transcriptase (Invitrogen). Real-time PCR was carried out with 1 μl cDNA with 10 pmol primer RC1 (5′-GTCTAGCCATGGCGTTAGTA-3′) and primer RC21 in a final volume of 25 μl. The reaction was performed in a LightCycler 480 (Roche). The PCR cycling conditions were as follows; an initial denaturation step at 95 °C for 15 min, followed by 45 cycles of 95 °C for 30 s, 60 °C for 30 s and 72 °C for 15 s. For each step, the temperature transition rate varied between 2.2 and 4.4 °C s^−1^, with fluorescence measurements taken after each elongation step. Conversion of copies ml^−1^ to IU ml^−1^ was performed using the HCV RNA genotype panel (National Institute for Biological Standards and Control, UK). ### Statistical analysis. Levels of ALT, viral load and T-cell responses were compared using the Mann--Whitney test. Pearson\'s chi-squared test was used to compare genetic diversity between groups for the core and E1. *P*\<0.05 was considered statistically significant. Statistics were analysed using Prism V (Graphpad Software) and Excel (Microsoft). Supplementary Material ====================== ::: {.caption} ###### \[Supplementary Table\] ::: This work was funded by the Wellcome Trust, the MRC, the James Martin 21st Century School, Oxford, UK, the EU (VIRGIL) programme and the NIHR Biomedical Research Centre Programme (Oxford, UK). The GenBank/EMBL/DDBJ accession numbers for the sequences determined in this study are FN665803 and FN666283--FN666408. A supplementary table showing the HCV core peptide amino acid sequences used in this study is available with the online version of this paper. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### IFN-*γ* ELISPOT response to core peptides. (a) The frequency of CD4^+^ T-cell responses against the panel of core peptides is indicated. Data were derived from Table [2](#t2){ref-type="table"}. The sequences of the individual peptides are available in Supplementary Table S1. All donors were tested against all peptides, and the total numbers of donors positive from the 61 tested is shown. (b) Typical ELISPOT assay from patient 306, showing a positive IFN-*γ* response to HCV core peptide 61--80 (s.f.u. per 10^6^ CD8-depleted PBMCs=140). (c) Titration experiment using fresh *ex vivo* CD8-depleted PBMCs from donor 304. The peptide concentration used is displayed on the *y*-axis, with the background subtracted. Peptides: ▴, 31--50; ▪, 61--80; ▾, 151--170. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Relatedness of sequences amongst responders and non-responders. Phylogenetic tree of the CD4 responder (•) and non-responder (○) HCV core region based on the neighbour-joining method using 1000 bootstrap replicates (scores \<30 are indicated by an asterisk). Bar, nucleotide substitutions per site. :::  ::: ::: {#f3 .fig} Fig. 3. ::: {.caption} ###### Sequence mutants in targeted epitopes. (a) An alignment of the core region is shown. The upper line indicates the group consensus. The lower lines indicate donors 304 and 379 with mutations within targeted epitopes indicated. Dots indicate amino acids identical to the consensus sequence. (b) An alignment of the core region for cloned donor 304 is shown. Each clone was compared with the bulk sequencing product. The frequency of the variant within the epitope 61--80 is indicated by shading: A68V was observed in the majority of the sequenced population. (c) Peptide titrations using PBMCs from donors 304 and 379, using wild-type (▪) and mutant (□) peptide as indicated in Fig. [3(a)](#f3){ref-type="fig"}. The assays were performed as in Fig. [1](#f1){ref-type="fig"}. :::  ::: ::: {#t1 .table-wrap} Table 1. ::: {.caption} ###### Clinical features of the study cohort [na]{.smallcaps}, Not available. ::: **Patient no.** **M/F** **Age (years)** **ALT level (U l^−1^)** **Viral load (IU ml^−1^)** **Core response** **Core region sequenced?** ----------------- --------- ----------------- ------------------------- ---------------------------- ------------------- ---------------------------- 106 F 71 135 110 925 Positive Yes 112 M 53 62 5 017 Positive Yes 117 M 45 39 61 929 806 Positive Yes 141 M 28 31 26 183 Positive Yes 160 M 50 34 [na]{.smallcaps} Positive No 168 M 59 40 17 913 Negative Yes 171 M 45 39 51 739 Negative No 172 M 50 38 478 192 Positive Yes 182 F 52 38 71 729 Negative Yes 183 M 48 199 2 920 108 Negative No 188 F 48 29 7 172 883 Negative Yes 191 M 56 168 [na]{.smallcaps} Negative No 193 M 46 103 685 931 Positive Yes 201 M 52 78 235 960 Positive Yes 205 M 56 111 56 642 Negative Yes 208 M 28 59 52 131 Negative Yes 210 M 43 443 5 056 294 Negative No 215 M 49 77 1 046 535 Negative Yes 252 M 44 17 92 895 Negative Yes 263 F 50 36 999 500 Positive No 269 M 73 26 489 951 Negative Yes 287 M 54 32 1 152 365 Negative Yes 289 F 49 25 960 304 Negative No 303 M 45 34 27 681 Positive Yes 304 F 47 24 111 709 Positive Yes 305 F 41 67 132 091 Negative Yes 306 M 54 56 89 367 Positive No 308 F 62 33 10 504 549 Positive No 310 F 50 186 169 719 Positive Yes 311 F 47 47 1 226 837 Positive Yes 317 F 46 34 15 992 Positive Yes 318 F 42 21 603 620 Negative Yes 319 M 61 327 486 031 Negative Yes 320 F 50 15 9 995 Positive Yes 323 F 44 50 127 425 Positive Yes 324 F 40 76 333 559 Positive Yes 328 M 38 96 36 139 Negative Yes 338 M 47 94 97 990 Positive Yes 342 F 40 33 93 679 Negative Yes 343 M 47 57 86 231 Positive Yes 348 M 27 51 11 092 491 Negative Yes 349 M 52 40 23 674 Negative Yes 350 F 71 81 82 704 Positive Yes 365 F 55 115 175 206 Positive Yes 369 M 55 91 148 553 Negative Yes 370 M 45 101 33 669 Positive No 371 F 40 58 1 046 535 Negative Yes 375 F 59 130 228 121 Positive Yes 376 M 58 134 41 940 Positive No 379 M 50 47 226 553 Positive Yes 382 F 54 115 153 649 Negative Yes 384 M 52 40 212 051 Negative Yes 386 M 45 85 635 Positive No 388 M 36 12 5 997 000 Negative Yes 393 M 54 230 587 941 Negative Yes 396 M 47 24 438 996 Positive Yes 401 M 49 50 3 484 532 Positive Yes 402 M 40 167 [na]{.smallcaps} Positive No 403 F 52 115 [na]{.smallcaps} Negative No 404 M 53 34 [na]{.smallcaps} Positive No 405 M 57 27 [na]{.smallcaps} Negative No ::: ::: {#t2 .table-wrap} Table 2. ::: {.caption} ###### HCV core CD4^+^ IFN-*γ* ELISPOT responses of 32 positive subjects Results are given as s.f.c. per 10^6^ CD8-depleted PBMCs. ::: **Patient no.** **Core pool** **Peptide** **No. of epitopes** ----------------- --------------- ------------- --------------------- ----- ----- ----- ----- ----- ----- ---- -- -- ----- ----- ----- ---- ----- ----- -- --- 106 195 190 1 112 160 135 1 117 283 150 1 141 60 65 1 160 250 120 120 2 172 250 280 175 150 3 193 56 350 560 2 201 100 320 1 263 350 170 1 303 100 65 60 2 304 297 77 110 77 220 173 3 306 870 313 140 2 308 140 124 159 1 310 169 97 150 2 311 114 198 1 317 291 80 80 90 127 4 320 166 155 145 2 323 60 182 1 324 110 80 1 338 335 180 268 2 343 137 168 1 350 120 125 1 365 250 405 175 732 254 3 370 105 70 1 375 185 128 85 2 376 772 675 1 379 294 109 80 2 386 50 50 1 396 100 66 50 2 401 100 150 80 2 402 250 110 90 125 3 404 85 90 95 2 ::: ::: {#t3 .table-wrap} Table 3. ::: {.caption} ###### Genetic diversity of the HCV core and E1 in CD4^+^ T-cell responders and non-responders ::: **Core** **E1** -------------------------------- ---------- -------- ------ ------- All individuals (*n*=46) 26.86 0.048 8.13 0.149 CD4^+^ responders (*n*=23) 26.88 0.055 9.23 0.167 CD4^+^ non-responders (*n*=23) 25.83 0.041 6.93 0.132 ::: [^1]: †These authors contributed equally to this work.

PubMed Central

2024-06-05T04:04:19.412775

2010-6-01

PMC3052722

INTRODUCTION ============ Human cytomegalovirus (HCMV; species *Human herpesvirus 5*) can infect most major organs and a wide range of cell types ([@r3]; [@r27]; [@r34], [@r37]). This broad tropism is lost during adaptation of the virus to fibroblasts in cell culture ([@r19]; [@r20]; [@r22]; [@r28]; [@r35]; [@r40], [@r41]). Adaptation to fibroblasts is associated strongly with mutations in a region encompassing three small genes at nt 176309--178146 in reference strain Merlin (GenBank accession no. AY446894). These genes are UL128, UL130 and UL131A, which together are termed the UL128 locus (UL128L). The mutations are predicted to ablate gene functions, and include nucleotide substitutions that introduce in-frame translational termination codons or compromise splicing, small insertions or deletions (indels) that introduce frameshifts into coding regions, and larger deletions or inversions that affect one or more genes. UL128L was originally described as containing four unspliced coding regions: UL128, UL129, UL130 and UL131 ([@r7]). Redefinition based on comparative genomics led instead to three genes: UL128 (three exons), UL130 (unspliced) and UL131A (two exons) ([@r9]). This arrangement was confirmed by transcript mapping ([@r2]; [@r19]). Each gene was predicted to encode a small protein that transits through the secretory pathway ([@r2]; [@r19]). The UL128 protein has sequence similarity to CC chemokines ([@r2]; [@r19]), and the UL130 protein contains a putative CX chemokine fold ([@r24]; [@r44]). The observation that virus passaged in fibroblasts is usually mutated in only one of the UL128L genes implies that the proteins act in concert ([@r2]; [@r19]), and the implication that this function is detrimental to growth of HCMV in these cells has found experimental support ([@r1]; [@r42]). In contrast, all three genes are known to be essential for growth of HCMV in endothelial cells ([@r19]) and dendritic cells ([@r16]), and at least one (UL131A) for growth in epithelial cells ([@r42]). These observations are consistent with an involvement of UL128L in cell tropism, specifically for infection of non-fibroblast cells ([@r33]; [@r36]). The UL128L proteins are present in virions ([@r1]; [@r25]; [@r43]), where they form a complex with virus glycoproteins H and L (gH--gL), as an alternative to glycoprotein O (gO) in the gH--gL--gO complex, for entry into non-fibroblast cells via endocytosis and low pH-dependent fusion ([@r29], [@r31]) or direct fusion with the plasma membrane ([@r26]). This key role for the gH--gL--UL128L complex is supported by the observation that antibodies against any of the UL128L proteins inhibited infection of endothelial and epithelial cells, but not fibroblasts ([@r1]; [@r21]; [@r43]). Also, endogenous expression of the gH--gL--UL128L complex rendered epithelial cells, but not fibroblasts, refractory to infection by viruses in which UL128L was intact, thus implying that specific receptors may exist on non-fibroblast cells ([@r30]). Mutations in other regions of the genome have been reported for strains passaged in fibroblasts, and most appear to be sporadic. However, several mutations affect members of the RL11 gene family, which has 14 related members (RL5A, RL6, RL11, RL12, RL13, UL1, UL4, UL5, UL6, UL7, UL8, UL9, UL10 and UL11), most of which are predicted to encode class I membrane glycoproteins ([@r7]; [@r10]). Lesions take the form of termination, frameshift or deletion mutations, and have been identified in RL13 (nt 11189--12070 in reference strain Merlin), RL5A, RL6 and UL9 ([@r8]; [@r11]). The observations described above depend largely on targeted regions in the genomes of various HCMV strains. We set out to paint a full picture of what happens dynamically to entire HCMV genomes when adapted from clinical samples to cell culture. Our aim was to identify patterns of mutation that occur during adaptation and to assess whether HCMV is genetically more stable in non-fibroblast than in fibroblast cells. RESULTS ======= Genomes sequenced completely ---------------------------- As multiple HCMV strains are common in clinical samples, 14 isolates were tested during isolation by PCR amplification of hypervariable genes UL146 and UL139, followed by sequencing the inserts in several plasmids ([@r4]). The epidemiologically distinct AF1, U8 and U11 isolates (see Methods) were each found to consist of a single strain and were chosen for further passaging. The analysis was complemented by data from a fourth isolate (VR1814; see Methods) that had been passaged earlier. All sequence data obtained were fully in accordance with the presence of a single, unique strain in each isolate. The strategy for adapting strains AF1, U8 and U11 to growth in cell culture is outlined in Fig. [1](#f1){ref-type="fig"}. The strains were isolated from clinical samples by using human fetal foreskin fibroblasts (HFFF-2) grown in shell vials, followed by expansion of the infected cells by a limited number (four/five) of passages. Infectivity was transferred in a single passage by co-culture of infected HFFF-2 cells with hTERT-immortalized human retinal pigmented epithelial (RPE-1) cells or primary human umbilical vein endothelial cells (HUVECs). Cell-free virus, obtained by sonication of the co-cultures, was then used to establish passage series in epithelial and endothelial cells, and passaging in fibroblasts was also continued. The strains were subjected weekly to collateral passaging (CP) in the three cell types by co-culture until at least passage 50 (p50), except that passaging of strain U11 in endothelial cells failed at an early stage for unknown reasons. Strain VR1814 was passaged by using similar methods, but much more extensively: 60 times in fibroblasts (isolation) followed by CP, attaining p214 in fibroblasts and p199 in endothelial cells (Fig. [1](#f1){ref-type="fig"}). In the fibroblast series, passage numbering started with the first isolation passage after growth in shell vials. In the epithelial and endothelial cell series, numbering started with the initiation of CP. Passage descriptions were abbreviated according to strain name, cell type (F, fibroblasts; R, epithelial cells; E, endothelial cells) and passage number. Thus, passage 35 of strain U8 in epithelial cells was denoted U8Rp35. At the termination of CP, the complete HCMV genomes in the final passages were sequenced, yielding a total of ten sequences: AF1Fp60, AF1Rp50 and AF1Ep50; U8Fp65, U8Rp51 and U8Ep51; U11Fp63 and U11Rp52; and VR1814Fp214 and VR1814Ep199. Genomes sequenced partially --------------------------- Two further passage series (II and III) of strain U11 were carried out in fibroblasts, in addition to the original series (I). Series II involved initiating CP with the virus isolated previously during series I (p4), whereas series III commenced with reisolation of virus from the clinical sample. RL13 and UL128L were sequenced from the final passages of series II and III (p51 and p40, respectively). Strain VR1814 is the parent of the bacterial artificial chromosome (BAC) FIX-BAC, which was generated by cloning a genome present at p46 in fibroblasts ([@r18]) and sequenced ([@r23]). A virus (RVFIX7) reconstituted from FIX-BAC that had been passaged approximately five times in fibroblasts ([@r18]) was treated with anti-mycoplasma agents during five further passages in fibroblasts. This virus was then subjected to CP by 117 passages in fibroblasts or 94 passages in endothelial cells to derive RVFIXFp117 and RVFIXEp94. Sequence data were derived for the region containing RL1--UL20 (nt 1772--28092, which includes RL13) for VR1814Fp22, VR1814Fp48, FIX-BAC, RVFIXFp117 and RVFIXEp94, and for the region containing UL122--UL150 (nt 168563--194150, which includes UL128L) for VR1814Fp48, FIX-BAC, RVFIXFp117 and RVFIXEp94. Identification of mutations --------------------------- Mutations in a particular strain were identified from differences between the genome sequences at the final passages, whether or not they were predicted to affect gene functions. The mutated and wild-type versions were differentiated from each other in two ways: directly on the basis of targeted data from earlier passages and the clinical samples, and indirectly by reference to sequence alignments for other HCMV strains. Most mutations consisted of substitutions or indels, but several sizeable deletions were also detected. Mutations that did not result in differences between the final sequences, but that had been present in the clinical sample or had arisen during isolation, would not have been detected except where they potentially ablated gene functions. Also, extensive deletions or rearrangements present in a proportion of, but not all, genomes might not have been detected. Inspection of sequence traces facilitated the identification not only of predominant mutations, but also of mutations represented in a significant minority (\>10 %) of the HCMV genome population. The dynamics by which mutations arose were monitored by reference to targeted sequence or PCR data from earlier passages or the clinical samples. The approaches used did not permit the complete absence of a particular sequence (mutant or wild-type) to be inferred. The mutations detected in the four strains are summarized in Tables [1](#t1){ref-type="table"}--[4](#t4){ref-type="table"}. Mutations were detected at each stage: a few in the clinical samples (not included in Tables [1](#t1){ref-type="table"}--[4](#t4){ref-type="table"}, but described below), some during isolation (e.g. the substitution in UL72 at nt 106148 of strain AF1; Table [1](#t1){ref-type="table"}), and the majority during CP (e.g. the substitution in UL72 at nt 105981 of strain AF1; Table [1](#t1){ref-type="table"}). It is notable that events seem to have moved at a slower pace in epithelial cells than in fibroblasts or endothelial cells. The HCMV genome has the structure *ab*-U~L~-*b*′*a*′*c*′-U~S~-*ca*, where U~L~ and U~S~ denote long and short unique regions and *ba*/*b*′*a*′ and *ca*/*c*′*a*′ indicate inverted repeats flanking the unique regions. The region most prone to mutations was the *b*/*b*′ inverted repeat, with a lower frequency in the flanking regions of the *a*/*a*′ and *c*/*c*′ inverted repeats. Owing to the locations of the PCR products utilized for sequencing, mutations in separate copies of the inverted repeats could be evaluated for the whole of *b*/*b*′ and *a*/*a*′ and the proximal part of *c*/*c*′, and were found to be approximately equimolar. This observation is consistent with recombination between the repeats during virus DNA replication. Strain AF1 ---------- RL13 mutants were detected in fibroblasts (nt 11460 and 11667; the same genome population), epithelial cells (nt 11488--11489) and endothelial cells (nt 10984) (Table [1](#t1){ref-type="table"}). Mutation of UL128L (UL130; nt 177046) occurred only in fibroblasts, and was first detected in a later passage (p20) than that in RL13 (p15). Seven clustered substitutions occurred in UL123 (nt 172215--173747; the same genome population) in endothelial cells. The strain AF1 genome in the clinical sample was mutated in two genes: RL6 containing an in-frame translational termination codon due to a C to A substitution at nt 5936, and UL9 containing a frameshift due to an additional A residue at nt 17079. Strain U8 --------- RL13 mutants were detected in fibroblasts (nt 11924), epithelial cells (nt 11269--11270) and endothelial cells (nt 11890) (Table [2](#t2){ref-type="table"}). Mutation of UL128L (UL131A; nt 178000) occurred only in fibroblasts, and was first detected in a later passage (p18) than that in RL13 (p16). Four clustered substitutions occurred in fibroblasts in the region containing UL142 and UL141 (nt 182822--184383; the same genome population). Complex, unresolved sequence heterogeneity (not included in Table [2](#t2){ref-type="table"}) was evident as various deletions in the region containing the *b*--U~L~ junction in fibroblasts and possibly also epithelial and endothelial cells. No obvious mutations were identified in the strain U8 clinical sample. Strain U11 ---------- Strain U11 was passaged successfully only in fibroblast and epithelial cells (Table [3](#t3){ref-type="table"}, series I). An RL13 mutation (nt 11378) was detected in both cell types. Mutation of UL128L (UL128; nt 176252 and 176307--176311; different genome populations) was detected in fibroblasts, and was first detected in a later passage (p15) than that in RL13 (p8). The same mutations were also detected at low levels in epithelial cells. Generation of the five concurrent substitutions at nt 176307--176311 probably occurred in a single event involving switches of replication polarity or recombination in a hairpin ([ACGCCGT]{.underline}CAAGA[ACGGCGT]{.underline}; inverted repeat underlined). The strain U11 genome in the clinical sample was mutated in a single gene (UL9) due to a frameshifting 23 bp deletion near the 3′ end (between nt 17102 and 17103). Several examples of mutations present in both final passages were detected, including those in RL13 and UL128, implying that they may have arisen during isolation. The reason for the apparently faster rate of mutant selection in strain U11 than in strains AF1 and U8 is not known. To investigate whether the mutations detected in both final passages had been present in the clinical sample or had originated during isolation, two additional passage series were carried out in fibroblasts. In series II, in which CP was initiated with the virus isolated during series I, three mutations in RL13 were detected (different genome populations) and one in UL128L (again, UL128). One of the RL13 mutations (nt 11378) reiterated a mutation detected in series I. In series III, which commenced with reisolation of virus from the clinical sample, two mutations in RL13 were detected (different genome populations) and one in UL128L (again, UL128), none corresponding to mutations observed in series I and II. In a further investigation, PCR products from the RL13 and UL128 regions were generated from the clinical sample. None of the 48 plasmids generated from each product contained any of the mutations detected in series I, II and III. Strain VR1814 ------------- Compared with strains AF1, U8 and U11, strain VR1814 had been subjected to more extensive passaging during isolation and CP (Fig. [1](#f1){ref-type="fig"}) and more mutations were observed (Table [4](#t4){ref-type="table"}). However, the analysis was limited in that the clinical sample was not available, and it was possible to differentiate the mutated and wild-type versions of a particular sequence only indirectly, by reference to sequence alignments for other HCMV strains. RL13 mutants were present in both fibroblasts and endothelial cells (nt 10999--12018 and 11775; different genome populations), and these were first detected during isolation (p48; Table [4](#t4){ref-type="table"}). A frameshift-inducing mutation in UL128L (UL131A; nt 177905) in fibroblasts was first detected during CP (p95). In addition, non-synonymous substitutions in UL128L (UL130) that left the coding region intact were detected in fibroblasts (p48; nt 177250) and endothelial cells (p199; nt 176856). Complex, unresolved sequence heterogeneity (not included in Table [4](#t4){ref-type="table"}) was detected in the *b*/*b*′ and *a*/*a*′ inverted repeats. No obvious mutations in the clinical sample were inferred. FIX-BAC and derived viruses RVFIXFp117 and RVFIXEp94 contained one of the RL13 mutations (nt 11775) and one of the UL128L (UL130) mutations (nt 177250). Passages of strain VR1814 in fibroblasts were assessed for endotheliotropism (p55, p154 and p214) and leukocyte transfer (p22, p48, p55, p95, p121, p143, p154, p179, p194 and p214). Endotheliotropism was lost between p55 and p154, and remained negative at p214. Leukocyte transfer was lost between p95 (2000 positive cells per 2×10^5^ leukocytes) and p154 (no positive cells), with reductions registered at p121 (540 positive cells) and p143 (65 positive cells) and leukocyte transfer remaining negative at p214. These findings are consistent with detection of the UL128L (UL131A) mutation at p95 and its predominance at p121 (Table [4](#t4){ref-type="table"}). Both RVFIXFp117 and RVFIXEp94 tested positive for leukocyte transfer and endotheliotropism. Growth properties ----------------- The passage series derived for strains AF1, U8 and U11 provided a means for obtaining virus stocks that were largely non-mutated in RL13 and UL128L (RL13^+^UL128L^+^), mutated in RL13 (RL13^−^UL128L^+^) or mutated in both locations (RL13^−^UL128L^−^). The process of generating stocks was protracted, taking approximately 2 weeks for RL13^−^UL128L^−^ viruses, 8 weeks for RL13^−^UL128L^+^ viruses and 10 weeks for RL13^+^UL128L^+^ viruses. To distinguish between the ability of cell-free virus in these stocks to initiate infection and to form a plaque, titrations in fibroblasts were either subjected to immunofluorescence (IF) analysis or stained for plaque counting. Regardless of strain, the ratio of the number of IF-positive cells to that of plaques was of the order of 100 for RL13^+^UL128L^+^ viruses, 10 for RL13^−^UL128L^+^ viruses, and \<10 for RL13^−^UL128L^−^ viruses. Kinetic-yield experiments in fibroblasts (Fig. [2](#f2){ref-type="fig"}) showed that the RL13^−^UL128L^−^ viruses produced \>10-fold more cell-released virus (CRV) than RL13^−^UL128L^+^ viruses, which in turn produced \>100-fold more CRV than RL13^+^UL128L^+^ viruses. Indeed, no CRV was detected from RL13^+^UL128L^+^ viruses. All viruses produced measurable amounts of cell-associated virus (CAV) at 8 days post-infection (p.i.), and the relative order was the same as that for CRV. DISCUSSION ========== Isolation and passaging of strains ---------------------------------- Strains AF1, U8 and U11 were isolated by a few passages in fibroblasts, thus subjecting them to fibroblast-associated selection pressures prior to CP. Indeed, mutations that apparently arose at this stage were identified for each strain, in particular strain U11. Isolation in fibroblasts was necessary because repeated attempts to isolate strains AF1, U8 and U11 directly in epithelial and endothelial cells were unsuccessful. This failure is consistent with the far greater ease with which HCMV may be isolated in fibroblasts than in other cell types ([@r15]). This phenomenon may be connected in part with the neutralizing activity of antibodies against the UL128L proteins ([@r17]), which are potentially bound to virus particles in the inoculum. Historically, strains in clinical samples have been isolated by passage in fibroblasts and then transferred to other cell types either by co-culture of infected fibroblasts and target cells or by infection of target cells with cell-free virus produced by disruption of fibroblast cultures ([@r13], [@r15]; [@r22]; [@r35]; [@r41]). We utilized co-culture for CP, which would have applied less selection in favour of virus release than passage of cell-free virus. Mutations in RL13 and UL128L ---------------------------- In strains AF1 and U8, RL13 mutated in fibroblast, epithelial and endothelial cells (at a different location in each cell type), and UL128L (UL130 and UL131A, respectively) mutated only in fibroblasts. In strain U11, for which endothelial-cell passaging failed, mutations in RL13 and UL128L (UL128) were detected in both fibroblasts and epithelial cells, indicating that they had occurred during isolation. Consistent with the known requirement for UL128L in non-fibroblast cells, the UL128L mutations in strain U11 were predominant in fibroblasts, but present at low levels in epithelial cells, where they were presumably maintained by complementation. In strain VR1814, RL13 mutated during isolation, and the UL128L (UL131A) mutant was detected later in fibroblasts, but not endothelial cells. In all four strains, mutation of RL13 in fibroblasts was detected before that of UL128L, and mutations in both loci quickly became predominant. The loss of UL128L and RL13 fits with previous observations on the presence of mutations in these loci in strains passaged in fibroblasts. However, in contrast to UL128L, RL13 mutated not only in fibroblasts, but also in epithelial and endothelial cells. Moreover, kinetic-yield experiments suggested that RL13 and UL128L suppress growth in fibroblasts at least partially independently, pointing to the involvement of different functional pathways. CP-derived viruses were used to investigate growth properties, but were of limited utility because they contained small amounts of alternative genotypes, and other mutations were also present in various proportions. Nonetheless, the similar growth properties obtained for viruses derived from strains AF1, U8 and U11 support the view that whether RL13 and UL128L were mutant or non-mutant was the main factor affecting differences in growth properties. In accord, a collaborative study has shown that virus derived from a strain Merlin BAC containing the wild-type gene complement (RL13^+^UL128L^+^) invariably mutated in RL13 in both fibroblasts and epithelial cells, with UL128L mutants being detected later only in fibroblasts (R. J. Stanton, unpublished data). Future investigations into the growth properties of RL13^+^UL128L^+^ virus are likely to be made to best effect using the Merlin system. The number of passages required for RL13 and UL128L (UL131A) to mutate in strain VR1814 appeared to be greater than that for the other strains. The observations that mutation of UL128L (UL131A) was detected at a particularly late stage (p95) and that the FIX-BAC-derived viruses (RVFIXFp117 and RVFIXEp94; both ostensibly RL13^−^UL128L^+^) did not mutate during passage in fibroblasts suggest genetic compensations elsewhere. In this connection, it is notable that strain VR1814 grown in fibroblasts had an additional, earlier mutation in UL128L, specifically a non-synonymous substitution in UL130. This mutation was inherited by FIX-BAC and the viruses derived from it. It is possible that this mutation reduced the selection pressure for further mutations in UL128L during growth in fibroblasts, whilst permitting the retention of a degree of endotheliotropism and ability to transfer to leukocytes. This might explain why UL128L mutations engineered into FIX-BAC-derived viruses did not result in significantly enhanced growth in fibroblasts in kinetic-yield experiments ([@r19]; commented on by [@r42]). In a similar vein, strain TB40/E ([@r38]) has a unique non-synonymous substitution in UL130 that might account for its ability to grow in both fibroblasts and endothelial cells ([@r11]). These comments prompt caution about assuming wild-type status for UL128L in fibroblast-passaged HCMV strains (as viruses or BACs) that retain the ability to grow in non-fibroblast cells. It is not fully understood why UL128L is detrimental to growth in fibroblasts, although inhibition of virion release has been proposed ([@r1]; [@r42]). Preliminary studies indicate that the RL13 product is a 100 kDa type I transmembrane glycoprotein (K. Baluchova, unpublished data), but its function and the reason for its inhibitory properties are unknown. It is possible that its suppressive effects in cell culture reflect similar properties operating *in vivo*, where HCMV persistence might be facilitated by mechanisms that temper virus replication. Alternatively, the effect might result from an interference with replication in cell culture that does not occur *in vivo*. Nonetheless, there remains a striking contrast between the enfeebled growth properties of RL13^+^UL128L^+^ virus in cell culture and the ability of HCMV to replicate well in clinical situations ([@r12]). Mutations in UL145--UL140 ------------------------- After passaging in endothelial cells, strain U8 had a 551 bp deletion in UL144 that also had the potential for affecting UL145 transcription (Table [2](#t2){ref-type="table"}). After passage in fibroblasts and epithelial cells, strain U11 had a single nucleotide insertion that caused a frameshift in UL141 (Table [3](#t3){ref-type="table"}). After passage in fibroblasts and endothelial cells, strain VR1814 had a 3173 bp deletion that affected UL142, UL141 and UL140 and possibly also UL144 transcription (Table [4](#t4){ref-type="table"}). Only strain AF1 was unaffected in this region. Lesions usually took the form of deletions, suggesting that more than one gene was selected against, although not as strongly as RL13 and UL128L. Mutations in a region containing five genes (UL145, UL144, UL142, UL141 and UL140) have been identified for strain AD169 varUC in fibroblasts (ablating UL144, UL142, UL141 and perhaps affecting UL140 transcription), strain TB40/E in endothelial cells (UL141, with a derivative additionally compromised in UL145 and UL144) and strain VR3216B in endothelial cells (UL142) ([@r5]; [@r38]; [@r39]; GenBank accession no. GQ222018). In addition, the genomes of the most widely used HCMV strains, AD169 (varUK and varATCC) and Towne (varS), each lack a sizeable region (U~L~/*b*′; 15 or 19 genes, respectively) that includes UL145--UL140 ([@r5]; [@r6]; [@r9]; [@r23]). It is possible that these extensive deletions represent one end of a spectrum of mutations whose key effect is to disable UL145--UL140. Other mutations --------------- All four strains exhibited numerous mutations in the inverted-repeat regions. The dynamics by which these mutations arose was not determined. Their frequent occurrence might be promoted in part by recombination, but other factors are probably important, as the major part of *c*/*c*′ was not affected. Many of the mutations in other regions of the genome were non-synonymous substitutions, which suggests that at least some were selected because they provided growth advantages. A cluster of mutations in the same genome population was evident for strains AF1 and U8, in UL123 for the former and in the region of UL142 and UL141 for the latter. These clusters may have resulted from local, error-prone replication of a genome whose progeny was then selected. Origins of mutations -------------------- The passaging experiments show that the HCMV genome was unstable during adaptation to cell culture, regardless of virus strain and cell type. However, the identification of mutations that apparently arose during isolation raised the question of whether they might have been present at low levels in the clinical sample. This possibility is technically difficult to rule out. If it is the case, the data from reisolation of strain U11 indicate the existence of a swarm of different mutants, each present at a very low level in the clinical sample. However, there was no evidence from sequence traces for the prior existence of any mutation. Moreover, none of the RL13 or UL128 mutants in strain U11 was detected in sets of PCR-derived plasmid clones generated from the clinical sample, thus excluding an approximate frequency per mutation of \>2 %. In contrast to the lack of evidence for the presence of the mutations described above in the clinical samples, a small number of mutations (in RL6 and UL9) were identified as characterizing the entire populations of HCMV genomes in the clinical samples from which strains AF1 and U11 were derived. Their identification strengthens the view that HCMV mutants lacking functions, particularly in the RL11 gene family, exist in clinical settings ([@r8]; [@r32]). METHODS ======= Clinical samples. ----------------- Clinical samples were obtained in accordance with the relevant national ethical guidelines. Strains AF1, U8 and VR1814 originated from Pavia, Italy, and strain U11 from London, UK. Strain AF1 was derived from amniotic fluid from a pregnant woman infected with HCMV, strains U8 and U11 from urine from infants infected congenitally with HCMV, and strain VR1814 from the cervical secretions of a pregnant woman with a primary HCMV infection. Cell lines. ----------- Strains AF1, U8 and U11 were isolated and passaged in Glasgow, UK, in three different cell lines (Fig. [1](#f1){ref-type="fig"}). HFFF-2 cells (ECACC 86031405) were grown in Dulbecco\'s modified Eagle\'s medium (DMEM)/10 % (v/v) fetal calf serum. hTERT-immortalized RPE-1 cells (Clontech C4000-1) were grown in DMEM/Ham\'s F-12 supplement/10 % (v/v) fetal calf serum. Primary HUVECs were prepared from umbilical cords from consenting anonymous volunteers attending the Queen Mother\'s Maternity Hospital, Glasgow, UK. Cords were obtained in accordance with National Health Service ethical guidelines. HUVECs were grown using the EGM-2 BulletKit system (Lonza) without heparin and subjected to IF using mouse antibody MCA 127T (Serotec) to confirm the presence of the von Willebrand factor. Each cell line was checked regularly for mycoplasma contamination by using a Minerva Biolabs VenorGeM mycoplasma detection kit (Cambio). Strain VR1814 was isolated and passaged similarly in Pavia, Italy, using human embryo lung fibroblasts (VO cells) and primary HUVECs. Isolation of HCMV strains. -------------------------- The AF1, U8 and U11 clinical samples were diluted 1 : 3 with DMEM and 500 μl was inoculated onto fibroblasts grown on glass coverslips in shell vials. Following centrifugation at 450 ***g*** for 30 min at 18 °C, the cultures were incubated overnight at 37 °C. The cultures were then trypsinized and the infected cells were expanded by passaging, reaching the stage of 75 cm^2^ cell-culture flasks. Plaques first appeared for strains AF1, U8 and U11 at p2, p4 and p2, respectively, and infectivity was transferred to epithelial and endothelial cells by co-culture with infected fibroblasts at p5, p5 and p4, respectively. Co-cultures were harvested by scraping the cells into the medium and disrupting them by ultrasonic treatment. Cell-free virus was then used to infect fresh monolayers of epithelial and endothelial cells, thus avoiding overgrowth by fibroblasts. Passaging in fibroblasts was also continued. Strain VR1814 was isolated and transferred similarly, except that it was passaged 60 times in fibroblasts (isolation) and then transferred to endothelial cells or continued in fibroblasts for CP. Passaging of HCMV strains. -------------------------- Strains AF1, U8 and U11 were subjected to CP (\>50 times) in fibroblast, epithelial and endothelial cells on a weekly basis by co-culture of infected cells with uninfected cells. Split ratios (infected : uninfected cells) depended on the extent of cytopathic effect (CPE). Typically, they were 1 : 2 to 1 : 4 for epithelial and endothelial cells and 1 : 2 to 1 : 250 for fibroblasts. At most passages, aliquots of the infected cultures were withdrawn for storage in liquid nitrogen, initiating the next passage, or extracting infected-cell DNA by using a FlexiGene kit (Qiagen). Following isolation, strain VR1814 was subjected to CP a further 154 times (i.e. a total of 214 times) in fibroblasts and 199 times in endothelial cells, and material from certain passages was retained. DNA sequencing. --------------- The HCMV genome present in infected-cell DNA at each final passage was sequenced in its entirety by a PCR-based approach involving the use of a large library of conserved, HCMV-specific primers for generating and sequencing a set of overlapping PCR products ([@r8]). Data were obtained in the form of sequence traces from an Applied Biosystems 3730 instrument at the BHF Glasgow Cardiovascular Research Centre (University of Glasgow, UK). The processed data were assembled by using the Staden Pregap4 and Gap4 programs, and the traces were inspected visually throughout. Most PCR products were sequenced directly, but some were cloned into four to six plasmids. In some instances, clones were utilized to determine whether closely located mutations were present in the same genome population. Targeted sequencing or PCR (for deletions) of HCMV DNA in selected lower passages or the clinical specimen was also carried out. Production of HCMV stocks. -------------------------- Infected cells at passages in which RL13^+^UL128L^+^, RL13^−^UL128L^+^ or RL13^−^UL128L^−^ viruses were predominant were recovered from liquid nitrogen and seeded on monolayers of the same cell line. Infectivity was amplified by several rounds of co-culture with uninfected cells and, in the case of RL13^+^UL128L^+^ and RL13^−^UL128L^+^ viruses, by a final 1 : 1 co-culture step with fibroblasts to boost titres. CRV and CAV were harvested as cell-free virus from at least two 175 cm^2^ flasks when complete CPE was achieved. In the case of RL13^+^UL128L^+^ and RL13^−^UL128L^+^ viruses, low-titre CRV was concentrated 100-fold by centrifugation at 26 000 ***g*** for 1 h at 18 °C. Infectivity of HCMV stocks. --------------------------- Titrations of cell-free virus were performed simultaneously in fibroblasts either in 30 mm plates or on glass coverslips. The cultures on 30 mm plates were fixed and stained for plaque counting at 12 days p.i. Those on coverslips were fixed at 3 days p.i. and permeabilized for IF analysis. The coverslips were treated with a mixture of antibodies recognizing the immediate-early UL123/UL122 proteins (IE1/IE2) (kindly provided by Edward Mocarski, Jr, Emory University School of Medicine, Atlanta, GA, USA) labelled using fluorescein isothiocyanate (FITC), and FITC-pre-conjugated antibody CCH2 (DAKO) recognizing the early UL44 protein (DNA polymerase processivity factor). The ability of strain VR1814 at various passages in fibroblasts or endothelial cells to transfer by co-culture to leukocytes (leukocyte transfer) or endothelial cells (endotheliotropism) was monitored by using published methods ([@r13], [@r14]; [@r19]). Leukocyte transfer was determined by IF detection in leukocytes of the UL82 protein (pp65; expressed as the number of positive cells per 2×10^5^ leukocytes) via a cytospin preparation, and endotheliotropism was assessed from cell-free virus-induced CPE, with infection of 80--100 % of cells being scored as positive. Kinetics of virus growth. ------------------------- The passages used were: AF1Rp38 (RL13^+^UL128L^+^), AF1Ep50 (RL13^−^UL128L^+^) and AF1Fp60 (RL13^−^UL128L^−^); U8Rp39 (RL13^+^UL128L^+^), U8Ep51 (RL13^−^UL128L^+^) and U8Fp54 (RL13^−^UL128L^−^); and U11Rp27 (RL13^+^UL128L^+^), U11Rp37 (RL13^−^UL128L^+^) and U11Fp51 (RL13^−^UL128L^−^). For each virus stock, one 35 mm plate containing 2×10^5^ fibroblasts was infected with 1 infectious unit of virus per cell (determined from IF titres). After adsorption for 1 h at 37 °C, unbound virus was removed by washing with medium and the cells were overlaid with 2 ml medium. Aliquots (1 ml) of medium were removed for titration of CRV in fibroblasts. Cells in these aliquots that had detached from the monolayer were pelleted by centrifugation at 500 ***g*** for 10 min at 18 °C and returned to the infected culture in 1 ml fresh medium. Virus yields were stored at −70 °C. At 8 days p.i., CAV was harvested by scraping the cells into 2 ml fresh medium and subjecting them to ultrasonic disruption. This work was supported by the Medical Research Council and the Wellcome Trust. We thank Wai Kwong Lee and Stewart Laing for DNA sequencing services, Gabriele Hahn for FIX-BAC, and James Cordiner (Consultant Obstetrician) for facilitating the provision of umbilical cords. The GenBank/EMBL/DDBJ accession numbers for the genome sequences of HCMV strains AF1, U8, U11 and VR1814 are GU179291, GU179288, GU179290 and GU179289, respectively. ::: {#f1 .fig} Fig. 1. ::: {.caption} ###### Scheme for isolation of HCMV strains from clinical samples in fibroblasts and CP in fibroblast, epithelial and endothelial cells. The passage number (p) at which CP commenced and terminated is shown for each strain. The final passage numbers for fibroblasts include those involved in virus isolation, whereas those for non-fibroblast cells do not. [nd]{.smallcaps}, Not done. :::  ::: ::: {#f2 .fig} Fig. 2. ::: {.caption} ###### Kinetics of virus growth in fibroblasts of RL13^+^UL128L^+^ (⧫), RL13^−^UL128L^+^ (▪) and RL13^−^UL128L^−^ (▴) viruses derived from strains AF1, U8 and U11. The connected points show CRV titres and the unconnected points on the right show CAV titres. Values below the limit of detection (dotted line) are assigned to the *x*-axis. :::  ::: ::: {#t1 .table-wrap} Table 1. ::: {.caption} ###### Mutations that occurred during passage of strain AF1, with the final passages being F (p60), R (p50) and E (p50) ::: **Location\*** **Region†** **Sequence‡** **Coding effect** **Dynamics§** **Intermediate passages\|\|** ------------------- ------------- --------------- ------------------- --------------- ------------------------------- -------------- ---------------- ---------------------- 10984 RL13 G G G **A∧** Termination 14/18 5 10 12 14 16 11460 RL13 C **G∧** C C Substitution 15/23 5 10 15 16 18 20 11488--11489 RL13 2 2 **Δ−** 2 Frameshift 30 10 20 11667 RL13 C **T∧** C C Termination 15/23 5 10 15 16 18 20 57436 UL44 \[T\] **C=** T T Substitution 50 5 12 25 105981 UL72 C C C **T∧** Substitution 30/30 5 12 106148 UL72 G **A+** **A+** G Substitution 60 (F); 50 (R) None (F); none (R) 134792--134836 UL89 in \[45\] 45 **Δ∧** 45 None 30/40 10 20 30 172215 UL123 \[C\] C C **T=** Substitution 50 10 20 30 40 172317 UL123 \[C\] C C **T=** Substitution 50 None 172375 UL123 \[C\] C C **T=** None 50 None 173346 UL123 \[C\] C C **T=** Substitution 46 10 20 30 40 42 173351 UL123 \[C\] C C **T=** Substitution 46 10 20 30 40 42 173582 UL123 \[C\] C C **T=** Substitution 44 10 20 30 40 42 173747 UL123 in \[C\] C C **T=** None 46 10 20 30 40 177046 UL130 C **A∧** C C Termination 20/23 10 12 14 15 16 18 20 194890/944 *b*′ C **A+** C C None 60 None 194979/855 *b*′ C **A=** C C None 60 None 195025/809 *b*′ G **C+** G G None 60 None 195037/797 *b*′ C C C **A=** None 50 None 195049/785 *b*′ T T **G−** **G=** None 60 (R); 50 (E) None (R); none (E) 195053/781 *b*′ T **G−** T T None 60 None 195056/778 *b*′ G **C−** G G None 60 None 195058/776 *b*′ T **G−** T T None 60 None 195081/753 *b*′ C **A−** C C None 60 None 195143/691/235892 *a*′ G **T+** G G None 60 None 196025/235010 *c*′ A **C+** A A None 60 None \*Location (nt) in the genome sequence. Multiple locations relate to duplications (*b*/*b*′ and *c*/*c*′) or triplications (*a*/*a*′/*a*) in the inverted repeats. †Regions affected by mutations are protein-coding or non-coding. The latter are located in the inverted repeats (*a*′/*b*′/*c*′) or introns (in) or are intergenic (IG). ‡N, Not passaged (clinical sample); F, passaged in fibroblast cells; R, passaged in epithelial cells; E, passaged in endothelial cells. The mutated nucleotide at each location is shown in bold, followed by a sign indicating relative abundance in the final passage: ∧, only the mutated nucleotide was detected; +, the mutated nucleotide predominated over the wild-type nucleotide; =, the mutated and wild-type nucleotides were present in approximately equal amounts; −, the mutated nucleotide was less abundant than the wild-type nucleotide. Where a deletion (Δ) occurred, the size (nt) is specified for the wild-type sequence. Where an insertion occurred, the original nucleotide is specified for the wild-type sequence and is followed by additional nucleotides (or replaced by the total number of nucleotides for larger insertions) for mutated sequences. Square brackets indicate that the sequence was not determined directly, but indirectly by reference to sequence alignments for other HCMV strains. [nd]{.smallcaps}, Passaging not done. §A single passage no. is that at which the mutation was first detected; a second passage no. is that at which only the mutation was detected. Square brackets (Table [4](#t4){ref-type="table"} only) indicate that the mutation detected in the E series was first detected during isolation in F cells. \|\|Passages analysed prior to that at which the mutation was first detected or that at which only the mutation was detected are listed. Later passages were also analysed. ::: ::: {#t2 .table-wrap} Table 2. ::: {.caption} ###### Mutations that occurred during passage of strain U8, with the final passages being F (p65), R (p51) and E (p51) ::: **Location\*** **Region†** **Sequence‡** **Coding effect** **Dynamics§** **Intermediate passages\|\|** ------------------- ---------------- --------------- ------------------- --------------- ------------------------------- -------------- ------------------- ---------------------- 11269--11270 RL13 2 2 **Δ−** 2 Frameshift 40 30 36 38 11890 RL13 A A A **AA∧** Frameshift 16/51 10 14 16 20 21 22 30 11924 RL13 G **A∧** G G Termination 16/20 11 12 14 16 18 35813 UL29 \[A\] **G=** A A Substitution 65 10 20 30 39 50 49479 UL36 \[C\] C C **T+** Substitution 10 10 49959 UL36 \[C\] **CC=** C C Frameshift 50 10 20 30 39 58564 UL45 \[C\] C **A−** C Substitution 40 20 30 100119 UL69 \[C\] **T∧** C C Substitution 20/65 10 20 30 39 50 106348 UL72 \[G\] G G **A+** Substitution 30 10 20 146722 UL100 \[G\] **T=** G G Substitution 65 10 20 30 39 50 146904 UL100 \[C\] C C **T∧** Substitution 39/50 10 20 30 39 149267 UL102 \[T\] T T **C∧** Substitution 30/50 10 20 30 40 178000 UL131A C **A∧** C C Termination 18/20 10 12 14 15 16 18 181584 UL145 C **G=** C C Substitution 32 30 181906--182456 UL144 551 551 551 **Δ∧** Deletion 30/51 30 182822 UL144/UL142 IG C **T=** C C None 32 30 183281 UL144/UL142 IG C **T+** C C None 32 30 183294 UL142 C **T+** C C Substitution 32 30 184383 UL141 C **T+** C C Substitution 32 9 20 30 186935 UL139/UL138 IG \[C\] **G=** C C None 39 10 20 30 194454/1113 *b*′ T **G∧** T T None 65/65 None 194666/901 *b*′ C **A∧** C C None 65/65 None 194676/891 *b*′ T **G=** T T None 65 None 194677/890 *b*′ C **A=** C C None 65 None 194690/877 *b*′ C **A+** C C None 65 None 194749/818 *b*′ C **A∧** **A=** C None 65/65 (F); 51 (R) None (F); none (R) 194774/793 *b*′ C **A∧** C C None 65/65 None 194792/775 *b*′ C **A+** C C None 65 None 194874/693/235647 *a*′ \[G\] **C∧** G G None 65/65 None See Table [1](#t1){ref-type="table"} for footnotes. ::: ::: {#t3 .table-wrap} Table 3. ::: {.caption} ###### Mutations that occurred during passage of strain U11, with the final passages in series I being F (p63) and R (p52) ::: **Location\*** **Region†** **Sequence‡** **Coding effect** **Dynamics§** **Intermediate passages\|\|** --------------------------------------------------------------------------------------------------------------------------------------------------------- ------------- --------------- ------------------- ------------------ ------------------------------- -------------------- ------------------------------------------ **Series I** 11378 RL13 A **AA∧** **AA∧** Frameshift 8/9 (F); 30/52 (R) 4 6 7 8 (F); 25 28 30 33 35 37 39 40 (R) 83326 UL55 C **T∧** C Substitution 58/62 20 30 40 45 50 58 60 61 176252 UL128 C **A=** **A--** Termination 15 (F); 20 (R) 5 8 10 (F); 10 (R) 176307--176311 UL128 CAAGA **TCTTG=** **TCTTG−** Substitutions 15 (F); 37 (R) 5 8 10 (F); 10 20 25 35 (R) 184505 UL141 G **GT=** **GT−** Frameshift 63 (F); 52 (R) None (F); none (R) 193985/735 *b*′ C **A−** C None 63 None 194097/623 *b*′ T **G=** T None 63 None 194099/621 *b*′ C **A−** **A=** None 63 (F); 52 (R) None (F); none (R) 194104/616 *b*′ C **A−** C None 63 None 194144/576 *b*′ T **G+** **G+** None 63 (F); 52 (R) None (F); none (R) 194192/528 *b*′ T **G−** T None 63 None 194216/504/234725 *a*′ T **C=** **C−** None 63 (F); 52 (R) None (F); none (R) 194220/500/234721 *a*′ G **T+** G None 63 None 194222/498/234719 *a*′ A **G=** **G+** None 63 (F); 52 (R) None (F); none (R) 194263/457/234678 *a*′ T **A−** **A−** None 63 (F); 52 (R) None (F); none (R) 194268/452/234673 *a*′ G G **A−** None 52 None 194795/234146 *c*′ G **T=** G None 63 None 194829/234112 *c*′ C **G+** **G+** None 63 (F); 52 (R) None (F); none (R) 222612 US26 C C **T+** Substitution 20 None **Series II, which started with virus isolated in series I and shows mutations that occurred in RL13 and UL128L, with the final passage being F (p51)** 10795 RL13 G **T−** [nd]{.smallcaps} Termination 20 None 10976--10977 RL13 2 **Δ−** [nd]{.smallcaps} Frameshift 20 None 11378 RL13 A **AA−** [nd]{.smallcaps} Frameshift 40 13 175795 UL128 C **A∧** [nd]{.smallcaps} Not spliced 13/20 5 13 15 **Series III, which started with virus reisolation and shows mutations that occurred in RL13 and UL128L, with the final passage being F (p40)** 11232--11233 RL13 2 **Δ−** [nd]{.smallcaps} Frameshift 21 10 11484 RL13 1 **Δ+** [nd]{.smallcaps} Frameshift 21 7 10 175739 UL128 G **A∧** [nd]{.smallcaps} Termination 30/40 7 10 15 21 25 30 33 See Table [1](#t1){ref-type="table"} for footnotes. ::: ::: {#t4 .table-wrap} Table 4. ::: {.caption} ###### Mutations that occurred during passage of strain VR1814, with the final passages being F (p214) and E (p199) ::: **Location\*** **Region†** **Sequence‡** **Coding effect** **Dynamics§** **Intermediate passages\|\|** ------------------- ------------------- --------------- ------------------- --------------- ------------------------------- ---------------------------- ---------------------------------------- 10999--12108 RL13 UL1 \[1110\] **Δ+** **Δ+** Deletion 48 (F); \[48\] (E) 22 25 30 41 (F); none (E) 11775 RL13 \[A\] **AA+** **AA+** Frameshift 48 (F); \[48\] (E) 22 25 30 41 (F); none (E) 28163--30607 UL23 UL24 UL25 \[2445\] 2445 **Δ+** Deletion 199 None 41424 UL32 \[G\] **A∧** G Substitution 214 None 51484 UL38 \[G\] G **GG=** Frameshift 199 None 57447 UL44 \[T\] T **C=** Substitution 199 None 73746 UL50 \[C\] C **CGAG+** In-frame insertion 199 None 82573 UL55 \[C\] **A=** C Substitution 214 None 82890 UL55 \[T\] T **C∧** Substitution 199 None 85615 UL56 \[C\] C **T+** Substitution 199 None 89021 UL57 \[T\] T **C∧** Substitution 199 None 94456--94470 OriLyt \[15\] 15 **Δ∧** None 199 None 94474 OriLyt \[C\] **111∧** **111∧** None 214/214 (F); 199/214 (E) None (F); none (E) 139372 UL89 \[C\] **G+** C Substitution 214 None 145063 UL98 \[A\] **C∧** **C∧** Substitution 214/214 (F); 199/214 (E) None (F); none (E) 163725 UL112 \[T\] T **TGGT+** In-frame insertion 199 None 171135 UL122 \[G\] G **A∧** Substitution 199 None 176856 UL130 \[G\] G **T∧** Substitution 199 None 177250 UL130 \[A\] **G∧** A Substitution 48/48 22 177905 UL131A \[T\] **TT∧** T Frameshift 95/121 22 25 30 41 48 55 95 181943 UL145/UL144 IG \[T\] T **TT+** None 199 None 182580--185752 UL142 UL141 UL140 \[3173\] **Δ∧** **Δ∧** Deletion 48/143 (F); \[48\]/199 (E) 22 25 30 41 48 55 95 121 (F); none (E) 194218/1118 *b*′ \[C\] C **A∧** None 199 None 194225/1111 *b*′ \[T\] **G−** T None 214 None 194290/1046 *b*′ \[C\] C **A+** None 199 None 194352/984 *b*′ \[C\] **A−** C None 214 None 194367/969 *b*′ \[C\] C **A∧** None 199 None 194389/947 *b*′ \[C\] **A−** **A+** None 214 (F); 199 (E) None (F); none (E) 194498/838 *b*′ \[C\] C **A∧** None 199 None 194563/773 *b*′ \[C\] **G−** C None 214 None 195182/154/234767 *a*′ \[T\] **G−** **G∧** None 214 (F); 199 (E) None (F); none (E) 195286/50/234663 *a*′ \[A\] A **C∧** None 199 None 195477/234472 *c*′ \[A\] **C=** **C=** None 214 (F); 199 (E) None (F); none (E) 195488/234461 *c*′ \[C\] **A−** C None 214 None 195545/234404 *c*′ \[G\] **T+** **T∧** None 214 (F); 199 (E) None (F); none (E) 195552/234397 *c*′ \[A\] **C−** **C∧** None 214 (F); 199 (E) None (F); none (E) 195703/234246 *c*′ (IRS1) \[A\] **C∧** **C+** Substitution 214 (F); 199 (E) None (F); none (E) 195709--197642 IRS1 \[1934\] 1934 **Δ∧** Deletion 199 None 198925 US1/US2 IG \[T\] T **C=** None 199 None 225998 US28 \[T\] T **TTT+** Frameshift 199 None 230525 US34 \[G\] G **GAG=** Frameshift 199 None See Table [1](#t1){ref-type="table"} for footnotes. ::: [^1]: †**Present address:** Department of Biochemistry and Immunology, Wellcome Building, Trinity College, Dublin 2, Ireland.

PubMed Central

2024-06-05T04:04:19.417150

2010-6-01

PMC3052910

Near-infrared (NIR) is an electromagnetic radiation that simultaneously exhibits both wave and particle properties and is strongly absorbed by water, hemoglobin, or myoglobin. We previously reported that NIR irradiation simulating solar NIR at specific wavelengths with pre- and parallel-irradiational cooling can penetrate the skin and non-thermally affect dermis,[@B1]^-^[@B3] superficial muscles,[@B4],[@B5] and so forth.[@B6] To clarify the possible effect of NIR irradiation on other subcutaneous tissues, we sought to evaluate how NIR non-thermally affects subcutaneous adipocytes and bones in rats. MATERIALS AND METHODS {#S1} ===================== Animals {#S1-1} ------- Thirty-five male Wistar rats (*Rattus norvegicus albinu*s) weighing 360 to 440 g were used. Experiments were performed in a temperature-controlled environment (24°C ± 1.5°C) under a 12-hour light-dark cycle with free access to water and standard rat chow. All animals were treated humanely and in compliance with the recommendations of the animal care local committee. The study was approved by our institutional ethics committee for animal experiments. Animals were anesthetized with an intra-abdominal dose of sodium pentobarbitone (50 mg/kg, IP) and were sacrificed by intracardiac administration of ketamine (150 m/kg) upon completion of the experiment. NIR irradiation {#S1-2} --------------- Near-infrared irradiation was generated using a broadband NIR source (Titan; Cutera, Brisbane, Calif). The device emitted a NIR spectrum ranging from 1100 to 1800 nm, filtering wavelengths between 1400 and 1500 nm, thereby simulating solar radiation. This allows us to deliver NIR without the wavelengths that are strongly absorbed by water and hemoglobin and allow safe delivery of NIR energy deep into the tissue. The horizontal spot size of irradiation was 10 × 30 mm. The system delivered energy with a fluence range from 5 to 56 J/cm^2^ using continuous energy single irradiation pulses ranging from 4 to 10 seconds. The sapphire contact cooling tip was set to a fixed temperature of 20°C to provide contact cooling. Pre- and parallel-irradiational cooling of the superficial layers was accomplished using a temperature-controlled sapphire window, which further prevented excessive superficial heating. Thirty-five rats were either irradiated (*n* = 25) or not (control; *n* = 10). The center of the dorsal portion of the irradiated rats was subjected to 3 rounds of irradiation at 40 J/cm^2^ on days 0, 7, and 14 without application of topical anesthesia. One round of irradiation consisted of 2 passes of NIR irradiation. Histological evaluation {#S1-3} ----------------------- Specimens, which included the spinous process of the sixth lumbar vertebra and the overlying subcutaneous tissues, were isolated from the experimental group (5 rats per time point) at 7, 30, 60, 90, and 180 days after the final dose of irradiation (days 7, 30, 60, 90, and 180, respectively). Control samples were only isolated at days 0 and 180 (5 rats per time point). The specimens were fixed in 20% neutral buffered formalin, processed for paraffin embedding, and serially sectioned along the sagittal plane (3- to 4 µm thickness) Tissue sections were stained with hematoxylin and eosin as well as with an anti-CD34 antibody for immunohistochemical analysis to determine CD34-positive cells in subcutaneous tissue and hematopoietic stem cells in bone marrow. CD34 staining positive cells in subcutaneous tissue are not an indicator for CD34 positive adipose-derived stem cells. Thus, we did not calculate the number of CD34-positive cells on the panniculus carnosus in this study. The transferase-mediated dUTP nick-end labeling (TUNEL) assay was used to stain apoptotic cells. The percentage of the area occupied by subcutaneous adipocytes was calculated for all time points in an area 0.5 mm high × 2 mm wide on the panniculus carnosus, which was located in the middle of spinous process. The percentage of the area occupied by bone marrow adipocytes, the cortical bone mass, as well as hematopoietic bone marrow cell counts and CD34-positive stem cell counts was calculated in a superficial area 1.5 mm deep in the spinous process. Images were scanned and quantified in 5 representative fields per section and subsequently averaged to obtain a final score. The sections were photographed under a BIOREVO BZ-9000 microscope (Keyence, Osaka, Japan). The digital photographs were processed using Adobe Photoshop (Adobe, San Jose, Calif). Statistical analyses {#S1-4} -------------------- The differences between groups at each time point were examined for statistical significance using the Mann-Whitney U test. *P* \<.05 was set as a cutoff for statistical significance. Data are represented as means±standard deviation. RESULTS {#S2} ======= Subcutaneous adipocytes on the panniculus carnosus overlying the spinous process were abruptly induced at day 7 and subsequently showed a gradual decrease (Figs [1](#F1){ref-type="fig"}--[3](#F3){ref-type="fig"}). Statistically significant increases were observed at days 7, 30, 90, and 180, compared with non-irradiated controls at days 0 and 180 (*P* \<0.05). Subcutaneous adipocytes on the panniculus carnosus overlying the spinous process were not observed in the non-irradiated controls at either time point. Bone marrow adipocytes were also induced dramatically at day 7 and after which numbers gradually decreased (Figs [1](#F1){ref-type="fig"}--[3](#F3){ref-type="fig"}). Significant increases were observed at days 7, 30, 90, and 180, compared with the non-irradiated controls at days 0 and 180 (*P* \<0.05). Cortical bone mass increased steadily over the 6-month period (Figs [1](#F1){ref-type="fig"} and [3](#F3){ref-type="fig"}). Significant increases in cortical bone mass were only observed at days 90 and 180 compared with non-irradiated controls (*P* \<0.05). Statistically significant differences were not achieved between control samples at day 0 and experimental samples at either days 7 or 30 (*P* = 0.8340 and *P* = 0.3472, respectively), as well as control samples at days 0 and 180 (*P* = 0.7540). Hematopoietic bone marrow cell counts in the irradiated groups decreased significantly at day 7 and then increased gradually from day 30 to day 180 (Figs [1](#F1){ref-type="fig"}--[3](#F3){ref-type="fig"}). Significant decreases were observed between samples at days 7, 30, 90, and 180 and the non-irradiated controls at days 0 and 180 (*P* \<0.05) (Figs [1](#F1){ref-type="fig"}--[3](#F3){ref-type="fig"}). CD34-positive cells on the panniculus carnosus appeared to increase at day 7 and then slowly decrease (Fig [2](#F2){ref-type="fig"}). Numbers of CD34-positive stem cells in the bone marrow were dramatically increased at day 7 and this increase persisted till day 180. Significant increases in cell number were observed at days 7, 30, 90, and 180 compared with the non-irradiated controls at days 0 and 180 (*P* \<0.05) (Figs [2](#F2){ref-type="fig"} and [3](#F3){ref-type="fig"}---lower panel). A majority of CD34-positive stem cells in bone marrow were observed at the inner surface of the bone cortex (Fig [2](#F2){ref-type="fig"}). TUNEL staining of the bone marrow showed almost no apoptotic cells in control samples at day 0 or irradiated samples at day 60. However, positive staining was seen at days 7 and 30 (Fig [4](#F4){ref-type="fig"}). There were no evidences of necrosis, including inflammation and hyperplasia of fibroblasts or lymphocytes, in irradiated samples obtained at days 7 and 30. DISCUSSION {#S3} ========== Because of its dual wave and particulate properties, NIR irradiation is able to penetrate the thin layer of skin (1000-1300 µm in this study) and minimal subcutaneous tissue covering of the spinous process. Thus, the bone marrow within the spinous process in rats is exposed to the NIR irradiation. Near-infrared irradiation dramatically increased subcutaneous adipocytes on the panniculus carnosus as well as CD34-positive cells surrounding the subcutaneous adipocytes. Histological analysis of the adipose tissue has revealed that CD34-positive cells are widely distributed among adipocytes and play a role in vascular stabilization.[@B7] Adipogenesis is tightly associated with angiogenesis, and the expression of adipocytes is linked to the development of its vascularure.[@B8] These reports suggest that NIR irradiation may stimulate CD34-positive cells to affect adipogenesis and angiogenesis on the panniculus carnosus. Near-infrared irradiation resulted in an early increase in the number of bone marrow adipocytes[@B9]^-^[@B10] and gradually increased the cortical bone mass. This was accompanied by an enrichment in CD34-positive stem cells especially at the inner surface of the bone cortex. CD34 is a marker for activated stem cells[@B11] and is highly expressed on hematopoietic cells at the earliest stages of their development.[@B12] The stem cell microenvironment (niche) is composed of mostly osteoblasts, which line the inner surface of bone and produce hematopoietic growth factors that regulate the formation and differentiation of hematopoietic stem cells.[@B13],[@B14] An increase in the number of osteoblasts results in a concurrent increase in the hematopoietic stem cell population and the cortical bone mass.[@B15],[@B16] Near-infrared irradiation dramatically decreased hematopoietic bone marrow cell counts by apoptosis. TUNEL identifies DNA fragmentation and can thus stain positive in cells undergoing either apoptosis or necrosis. In this study, we did not detect any evidence of necrosis. Therefore, positive TUNEL staining suggested that NIR irradiation induces a kind of apoptosis. NIR irradiation has previously been reported to induce strand breaks and cell death by apoptosis.[@B17] Actively proliferating cells also show increased sensitivity to red and NIR radiation.[@B18],[@B19] To address the physiological relevance of our study, we can make the following calculations: The energy of the NIR irradiation we used was 40 J/cm^2^, each treatment consisted of 2 passes of NIR irradiation, and a total of 3 rounds of NIR treatment were performed on each rat. Thus, each rat was exposed to a total NIR energy of 40 × 2 × 3 J/cm^2^. The total incident solar energy at sea level is reported to be 0.0747 W/cm^2^.[@B20] The energy distribution emitted by the NIR device used at wavelengths 1100 to 1800 nm is mathematically estimated to constitute 10.2% of total incident solar energy at sea level. Thus, incident NIR (1100--1800 nm) in solar energy will be 0.0076194 W/cm^2^. Thus, the energy emitted by these NIR treatments is approximately equivalent to sunbathing for 8 hours and 45 minutes in North America (40 \[J/cm^2^\] × 2 \[pass\] × 3 \[round\]/0.0076194 \[W/cm^2^\] = 31498.8188 \[s\]). Sunburn is believed to occur when skin has been burned by ultraviolet radiation, most often after prolonged exposure to the sun. However, non-thermal damages to the subcutaneous tissues can occur when skin is exposed to NIR radiation while sunbathing. Therefore, exposed skin should be protected with sunscreens that block not only ultraviolet but also NIR radiation. Near-infrared irradiation damages superficial muscle[@B4],[@B5] as well as bone marrows cells. In contrast, it thickens the dermis and may enrich the stem cells that induce the number of adipose cells and increase cortical bone mass. These findings are relevant to facial skin that is constantly exposed to the sun as follows: solar NIR radiation may thicken the facial dermis, thereby increasing the number of superficial granular adipocytes and weakening the superficial facial muscles, which results in eyelid and facial skin ptosis. Although further studies are required to confirm our results, they may have major implications in facial aging. CONCLUSIONS {#S4} =========== Near-infrared irradiation that simulated solar radiation non-thermally affected subcutaneous adipocytes and bones in rats. It induced putative, non-thermal damage of bone marrow, which was mediated by apoptosis. However, it increased subcutaneous and bone marrow adipocytes, CD34-positive hematopoietic stem cells in bone marrow, and cortical bone mass. Acknowledgments {#S5} =============== We thank Ikuo Matsuyama for histological staining and the members of Cutera Inc for helpful comments. **Disclosure**: None of the authors of this study have a conflict of interest. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### Hematoxylin and eosin staining (a-f). The dermis, subcutaneous tissues, and panniculus carnosus are shown above, and the spinous process is shown below for each experimental point. The control at day 0 (a), and experimental samples at day 7 (b), day 30 (c), day 90 (d), day 180 (e), and the control at day 180 (f) are indicated. Scale bars = 500 µm. An asterisk indicates a layer of subcutaneous tissue where adipocytes have been induced, 2 asterisks indicate the panniculus carnosus, and 3 asterisks indicate the evaluated spinous process. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Immunohistochemical staining with an anti-CD34 antibody (a-f). The dermis, subcutaneous tissues, and panniculus carnosus are shown above, and the spinous process is shown below for each experimental time point. CD34-positive cells (stained brown) in the control at day 0 (a), experimental samples at day 7 (b), day 30 (c), day 90 (d), day 180 (e), and the control at day 180 (f) are indicated. Scale bars = 100 µm. An asterisk indicates a layer of subcutaneous adipocytes and subcutaneous tissue, 2 asterisks indicate the degenerated panniculus carnosus and 3 asterisks indicate an enlarged part of the evaluated spinous process. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### \(a) Chronological changes in the percentage of the area occupied by subcutaneous adipocytes in an area 0.5 mm high × 2 mm wide on the panniculus carnosus, which was located in the middle of the spinous process, are indicated in yellow. Chronological changes in the percentage of the area occupied by the cortical bone mass in an area of superficial depth (1.5 mm) in the spinous process are indicated in red. (b) Chronological changes in the hematopoietic bone marrow cell counts are indicated in blue, the percentage of the area occupied by bone marrow adipocytes in yellow, and the CD34-positive stem cell counts in brown. All were located in an area of superficial depth (1.5 mm) in the spinous process. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### Transferase-mediated dUTP nick-end labeling (TUNEL) staining of apoptotic cells (a-d). TUNEL staining of bone marrow was almost completely negative in controls at day 0 (a) and experimental samples at day 60 (d), but positive at day 7 (b) and day 30 (c). Representative TUNEL positive cells are indicated by black arrows. :::  :::

PubMed Central

2024-06-05T04:04:19.424150

2011-3-9

PMC3053139

In the Funding section, one of the grant numbers from the National Cancer Institute was listed incorrectly. The grant number \"CA 132706\" should be \"CA 134304.\" **Competing Interests:**No competing interests declared.

PubMed Central

2024-06-05T04:04:19.425170

2011-2-25

PMC3053160

Introduction {#sec1} ============ 1. Cell adhesion molecules are the primary mediators for various types of cell--cell junctions and play essential roles in various cellular processes, including morphogenesis, differentiation, proliferation and migration (Harris & Tepass, 2010[@bb4]; Ogita & Takai, 2006[@bb7]). In polarized epithelial cells, cell--cell junctions comprise several adhesive apparatuses including tight junctions (TJs), adherens junctions (AJs), desmosomes and gap junctions. AJs regulate TJ formation and the establishment of the apical--basal polarity at cell--cell adhesion sites (Takai, Ikeda *et al.*, 2008[@bb10]). Nectins, which are immunoglobulin-like cell adhesion molecules, and cadherins, which are Ca^2+^-dependent cell adhesion molecules, localize at AJs and have essential cooperative roles in AJ formation. Nectins initiate AJ formation before cadherins form cell--cell adhesions. Initial cell--cell contacts are formed between two neighbouring cells by nectins, and cadherins are then recruited to the nectin-based adhesion sites to form strong cell--cell adhesions. TJs are formed at the apical side of AJs. Together, nectins and cadherins mediate TJ formation by recruiting junctional adhesion molecules (JAMs), followed by claudins and occludins, to the apical side of AJs. Nectins comprise a family of four members, nectin-1, nectin-2, nectin-3 and nectin-4, each with multiple isoforms (Takai *et al.*, 2003[@bb11]; Ogita *et al.*, 2010[@bb6]). Most nectin-family members are membrane glycoproteins with an extracellular N-terminal variable region-like domain, two extracellular constant region-like domains, a transmembrane region and a cytoplasmic tail. All nectins, except for nectin-1β, nectin-1γ, nectin-3γ and nectin-4, contain a PDZ-binding motif (E/A-*X*-Y-V) in their cytoplasmic tail. Through this sequence the nectins can bind afadin, which itself binds to an actin filament and α-­catenin. Two nectin molecules on the surface of the same cell first form *cis*-­dimers, which is followed by the formation of *trans*-dimers of the *cis*-­dimers on apposing cells, resulting in the formation of cell--cell adhesions (Takai, Miyoshi *et al.*, 2008[@bb12]). Heterophilic *trans*-inter­actions have been detected between nectin-1 and nectin-3, between nectin-2 and nectin-3 and between nectin-1 and nectin-4. In addition to their cell adhesion activity, nectin-1 and nectin-2, but not nectin-3 or nectin-4, serve as entry receptors for α-herpesviruses by binding the virus-envelope glycoprotein gD with different specificities (Cocchi *et al.*, 1998[@bb1]; Sakisaka *et al.*, 2001[@bb9]; Lopez *et al.*, 2000[@bb5]; Geraghty *et al.*, 1998[@bb3]; Warner *et al.*, 1998[@bb13]). While nectin-1 shows activity as a receptor for herpes simplex virus type 1 (HSV-1), HSV-2, pseudo­rabies virus (PRV) and bovine herpesvirus type 1, nectin-2 mediates the entry of HSV-2 and PRV. To further understand how nectin-family members form *cis*-interactions and *trans*-interactions, structural information on the extracellular region of nectin-family members is fundamental. Here, we report the refolding, crystallization and preliminary X-ray crystallo­graphic analyses of the extracellular regions of nectin-1 and nectin-2. To our knowledge, this is the first report of the crystallization of the whole extracellular regions of nectins. Materials and methods {#sec2} ===================== 2. Protein expression and purification of inclusion bodies {#sec2.1} ------------------------------------------------------- 2.1. Genes encoding the extracellular region of human nectin-1α (nectin-1-EC; residues 30--335) or mouse nectin-2α (nectin-2-EC; residues 32--339) were PCR-amplified using the Expand High Fidelity PCR system (Roche) from respective full-length cDNAs using primer A (5′-TA**GGATCC**GTCCCAGGCGTCCACTCC-3′; nectin-1α, forward) and primer B (5′-TA**GCGGCCGC**TTCTGTGATATTGACCTCCACC-3′; nectin-1α, reverse) or primer C (5′-TA**GGAT** **CCGC**AGGATGTGCGAGTTCAAGTGC-3′; nectin-2α, forward) and primer D (5′-TA**GCGGCCGC**GTCTCGCACCAGGATGACCT-3′; nectin-2α, reverse). The PCR products were inserted into the pGEM-T Easy vector (Promega), which contains 3′-T overhangs at the insertion site. Target genes were isolated by digestion of the plasmids with the restriction enzymes *Bam*HI and *Not*I and were confirmed by DNA sequencing. This was followed by ligation into the T7 promoter expression vector pET21b (Novagen) in frame with a C-­terminal 6×His tag. Each recombinant plasmid was transformed into *Escherichia coli* BL21 (DE3) cells (Novagen). The cells were cultured in Luria--Bertani broth containing 100 µg ml^−1^ ampicillin at 310 K until the OD~600~ reached 0.5--0.7. Protein expression was induced at 298 K by the addition of isopropyl β-[d]{.smallcaps}-1-thiogalactopyranoside to a final concentration of 0.4 m*M*. At 16 h post-induction the cells were harvested by centrifugation, suspended in phosphate-buffered saline containing 40 µg ml^−1^ hen egg-white lysozyme, subjected to two cycles of freezing and thawing and sonicated until the lysate was homogeneous. Centrifugation of these lysates at 10 000*g* for 20 min yielded inclusion bodies. The inclusion bodies were washed four times with a buffer con­sisting of 20 m*M* Tris--HCl pH 7.5, 300 m*M* NaCl, 1 m*M* EDTA, 0.5% Triton X-100 and 1 m*M* DTT and were dissolved in a buffer con­sisting of 50 m*M* MES--NaOH pH 6.0, 8 *M* urea, 1 m*M* EDTA and 1 m*M* DTT. This mixture was slowly rotated overnight at 277 K before centrifugation at 10 000*g* for 20 min to remove insoluble materials. The yields of inclusion bodies were ∼0.1 g for nectin-1-EC and ∼0.2 g for nectin-2-EC per litre of culture. Unfolded proteins were refolded by 300-fold dilution into refolding solution *A* \[500 m*M* [l]{.smallcaps}-arginine, 100 m*M* Tris--HCl pH 9.0, 2 m*M* oxidized glutathione (GSSG) and 1 m*M* reduced glutathione (GSH)\] in the case of nectin-1-EC or refolding solution *B* (500 m*M* [l]{.smallcaps}-­arginine, 100 m*M* Tris--HCl pH 9.0, 10 m*M* GSSG, 0.1 m*M* GSH) in the case of nectin-2-EC, followed by incubation for 48 h at 277 K. After concentration using a 10 000 molecular-weight cutoff ultrafiltration membrane (GE Healthcare), the samples were sub­jected to size-exclusion chromatography on a HiLoad 16/60 Superdex 200 pg column (GE Healthcare) to separate correctly folded proteins from aggregated forms. These fractions were dialyzed against 20 m*M* MES pH 6.0 to precipitate almost-misfolded proteins, filtered using an Ultrafree-MC GV 0.22 µm (Millipore) and applied onto a HiTrap SP HP column (5 ml; GE Healthcare) followed by a Mono Q column (1 ml; GE Healthcare). The protein yields were ∼0.5 mg for nectin-1-EC and ∼10 mg for nectin-2-EC from 100 mg inclusion bodies. Optimization of refolding conditions and protein purification {#sec2.2} ------------------------------------------------------------- 2.2. The standard conditions for refolding nectin-1-EC and nectin-2-EC were as follows: 400 m*M* [l]{.smallcaps}-arginine, 100 m*M* Tris--HCl pH 9.0, 2.5 m*M* GSH, 2.5 m*M* GSSG and 100 µg ml^−1^ unfolded protein. Small-scale refolding assays (1 ml) were performed to investigate the effects of changing the [l]{.smallcaps}-arginine concentration from 100 to 600 m*M* (for nectin-1-EC), the pH from 7.0 to 9.0 (for nectin-1-EC), the GSSG:GSH ratio from 10.0:0.1 m*M* to 0.1:10.0 m*M* (for both nectin-1-EC and nectin-2-EC) and the concentration of unfolded protein from 25 to 200 µg ml^−1^ (for nectin-1-EC). In each case, the unfolded protein solutions were diluted at least 300-fold into each of the refolding solutions such that only one parameter was varied while the other parameters were kept at the standard conditions. The solutions were incubated at 277 K for 48 h and then subjected to size-exclusion chromatography on a Superdex 200 10/300 GL column using an ÄKTA FPLC system (GE Healthcare; Figs. 1*a*--1*d* [▶](#fig1){ref-type="fig"}). A peak at an elution volume of ∼15.5 ml corresponded to the correctly folded protein containing native intramolecular disulfide bonds. Aggregates containing intermolecular disulfide bonds eluted in the void volume (∼8.5 ml). Crystallization {#sec2.3} --------------- 2.3. Purified nectin-1-EC and nectin-2-EC were dialyzed against solutions *C* (20 m*M* Tris--HCl pH 7.5, 150 m*M* NaCl) and *D* (20 m*M* Tris--HCl pH 9.0, 150 m*M* NaCl), respectively, and then concentrated to 5 and 4 mg ml^−1^, respectively, using a Vivaspin 6 10k (GE Healthcare). The homogeneous proteins were analyzed by screening additives using dynamic light scattering with a Zetasizer Nano ZS (Malvern Instruments) to determine their suitability for crystallization. In the presence of 0.2 *M* NDSB201 nectin-1-EC and nectin-2-EC were monodisperse. Initial crystallization trials were performed using a Phoenix liquid-handling system (Art Robbins Instruments) at 296 K using SaltRx 1, SaltRx 2, 50%(*v*/*v*) (half concentration) PEGRx 1, 50%(*v*/*v*) PEGRx 2, 50%(*v*/*v*) PEG/Ion 1 and 50%(*v*/*v*) PEG/Ion 2. The volume of the reservoir solution was 60 µl. The drops consisted of 0.2 µl of both the protein and reservoir solution. In the presence of 0.4 *M* NDSB201, nectin-1-EC and nectin-2-EC crystals were obtained using both polyethylene glycol and salt conditions as the reservoir. The initial crystallization conditions for nectin-1-EC and nectin-2-EC were further refined by changing the pH, precipitant concentration and additives. The most promising crystals of nectin-1-EC were observed in drops comprised of equal volumes of nectin-1-EC solution \[20 m*M* Tris--HCl pH 7.5, 150 m*M* NaCl and 6%(*w*/*v*) 1,6-hexanediol\] and precipitant solution \[50 m*M* citric acid, 50 m*M* bis-Tris propane and 1--3%(*v*/*v*) PEG 3350\] at 296 K (Fig. 2[▶](#fig2){ref-type="fig"} *a*). Crystals of nectin-1-EC could be obtained even if the crystallization conditions contained no NDSB201, and the NDSB201 did not influence the diffraction quality of the crystals. The most promising crystals of nectin-2-EC were observed in drops comprised of equal volumes of nectin-2-EC solution (20 m*M* Tris-HCl pH 9.0, 150 m*M* NaCl and 0.35 *M* NDSD201) and precipitant solution (45 m*M* citric acid, 55 m*M* bis-Tris propane and 3.6 *M* sodium nitrate) at 296 K (Fig. 2[▶](#fig2){ref-type="fig"} *b*). Data collection {#sec2.4} --------------- 2.4. To improve the diffraction quality of the nectin-1-EC crystals, the crystals were subjected to dehydration with increasing concentrations of PEG 300. The crystals were transferred in a large number of steps from harvesting buffer \[20 m*M* Tris--HCl pH 7.5, 150 m*M* NaCl, 6%(*w*/*v*) 1,6-hexanediol, 50 m*M* citric acid, 50 m*M* bis-Tris propane and 5%(*v*/*v*) PEG 3350\] to harvesting buffer including 25%(*v*/*v*) PEG 300 at 277 K. Before freezing with liquid nitrogen, the crystals were equilibrated in the final buffer for 3 d. This procedure markedly improved the resolution limit of the crystals from ∼5 to ∼2.8 Å. The nectin-2-EC crystals were soaked in a cryoprotection solution consisting of 20 m*M* Tris--HCl pH 9.0, 150 m*M* NaCl, 0.4 *M* NDSB201, 45 m*M* citric acid, 55 m*M* bis-Tris propane, 4.0 *M* sodium nitrate and 14%(*v*/*v*) ethylene glycol by stepwise transfer at room temperature. Diffraction data sets were collected from nectin-1-EC and nectin-2-EC crystals on the BL44XU beamline at the SPring-8 synchrotron facility (Harima, Hyogo, Japan) at 100 K using a DIP6040 imaging-plate detector (MAC Science/Bruker AXS). A total of 60 frames of data were collected for nectin-1-EC in three runs with a translation of 70 µm along the rotation axis. The nectin-1-EC data-collection parameters included a crystal-to-detector distance of 540 mm, an oscillation angle of 0.5° and an exposure time of 20 s per frame at a wavelength of 0.9000 Å (Fig. 3*a* [▶](#fig3){ref-type="fig"}). For nectin-2-EC, a total of 70 frames of data were collected with a crystal-to-detector distance of 400 mm, an oscillation angle of 0.5° and an exposure time of 2 s per frame at a wavelength of 0.9000 Å (Fig. 3*b* [▶](#fig3){ref-type="fig"}). All data sets were processed and scaled with the *HKL*-2000 program package (Otwinowski & Minor, 1997[@bb8]). Data-collection statistics are summarized in Table 1[▶](#table1){ref-type="table"}. Results and discussion {#sec3} ====================== 3. The extracellular regions of nectin-1 and nectin-2 fused with a C-­terminal 6×His tag were expressed as inclusion bodies in *E. coli* BL21 (DE3). After solubilizing the inclusion bodies in 8 *M* urea, nectin-1-EC and nectin-2-EC proteins were successfully refolded by rapid dilution with a glutathione redox couple. To increase the yields of refolded nectin-1-EC protein, the refolding conditions (*i.e.* pH, GSSG:GSH ratio, [l]{.smallcaps}-arginine concentration and nectin-1-EC con­centration) were optimized in a series of small reactions (1 ml). Correct folding was assessed by size-exclusion chromatography on a Superdex 200 10/300 GL column using an ÄKTA FPLC system (GE Healthcare; Figs. 1*a*--1*d* [▶](#fig1){ref-type="fig"}). Notably, a lack of the C-terminal 6×His tag significantly decreased the yield of correctly folded protein; we could not obtain crystals of nectin-1-EC without the tag. Based on the results for nectin-1-EC, only the GSSG:GSH ratio was optimized for nectin-2-EC (Fig. 2[▶](#fig2){ref-type="fig"} *e*). The optimized refolding solutions for nectin-1-­EC and nectin-2-EC are described in §[](#sec2){ref-type="sec"}2. Nectin-1-EC crystals belonged to the cubic space group *P*2~1~3, with unit-cell parameters *a* = *b* = *c* = 164.9 Å. Nectin-2-EC crystals belonged to the hexagonal space group *P*6~1~22 or *P*6~5~22, with unit-cell parameters *a* = *b* = 79.3, *c* = 235.4 Å. Molecular-replacement calculations with *MOLREP* and *Phaser* (Collaborative Computational Project, Number 4, 1994[@bb2]) using the structure of a homologous protein \[CD155, which has the maximum sequence identity to nectin-1-EC (48.4%) and nectin-2-EC (25.3%); PDB code [3eow](http://scripts.iucr.org/cgi-bin/cr.cgi?rm=pdb&pdbId=3eow); Zhang *et al.*, 2008[@bb14]\] as a search model were unsuccessful. Therefore, heavy-atom derivatives of nectin-1-EC and nectin-2-EC crystals have been prepared for phase determination. Diffraction data sets for the heavy-atom derivatives are currently being collected on BL44XU at the SPring-8 synchrotron facility for phase determination. We thank Dr Eiki Yamashita for help with diffraction data collection on the BL44XU beamline at SPring-8. This work was supported in part by a Targeted Proteins Research Program (TPRP) grant from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan (to MS). ::: {#fig1 .fig} Figure 1 ::: {.caption} ###### Optimization of refolding conditions for nectin-1-EC and nectin-2-EC. (*a*--*d*) Effects of varying (*a*) the [l]{.smallcaps}-arginine concentration, (*b*) the pH, (*c*) the GSSG:GSH ratio and (*d*) the protein concentration on the recovery of nectin-1-EC. (*e*) Effects of varying the GSSG:GSH ratio on the recovery of nectin-2-EC. Samples from small-scale refolding reactions were subjected to size-exclusion chromatography on a Superdex 200 10/300 GL column. Correctly folded proteins eluted between 15 and 16.5 ml. Aggregates eluted between 8 and 9.5 ml. The insets show close-up views of the elution profiles between 14 and 17 ml. :::  ::: ::: {#fig2 .fig} Figure 2 ::: {.caption} ###### Crystals of the whole extracellular regions of nectins. (*a*) Nectin-1-EC crystal. (*b*) Nectin-2-EC crystal. The scale bars are 100 µm in length. :::  ::: ::: {#fig3 .fig} Figure 3 ::: {.caption} ###### X-ray diffraction images from nectin crystals. (*a*) 0.5° oscillation image from a nectin-1-EC crystal. (*b*) 1.0° oscillation image from a nectin-2-EC crystal. :::  ::: ::: {#table1 .table-wrap} Table 1 ::: {.caption} ###### Data-collection statistics Values in parentheses are for the highest resolution shell. :::   Human nectin-1α extracellular region Mouse nectin-2α extracellular region -------------------------------------------- -------------------------------------- -------------------------------------- Beamline BL44XU, SPring-8 BL44XU, SPring-8 Space group *P*2~1~3 *P*6~1~22 or *P*6~5~22 Unit-cell parameters (Å) *a* = *b* = *c* = 164.9 *a* = *b* = 79.3, *c* = 235.4 Wavelength (Å) 0.90000 0.90000 Detector DIP6040 DIP6040 Crystal-to-detector distance (mm) 540 400 Rotation range per image (°) 0.5 1.0 Total rotation range (°) 60 70 Exposure time per image (s) 20 2 Resolution range (Å) 50.00--2.80 (2.90--2.80) 50.00--2.55 (2.59--2.55) Total no. of observations 137039 (13666) 61011 (3039) No. of unique reflections 36109 (3504) 14654 (707) Completeness (%) 97.9 (96.8) 98.2 (96.8) 〈*I*〉/〈σ(*I*)〉 19.4 (4.3) 21.8 (3.6) Multiplicity 3.8 (3.9) 4.2 (4.3) *R*~merge~[†](#tfn1){ref-type="table-fn"} 0.068 (0.358) 0.061 (0.396) Overall *B* factor from Wilson plot (Å^2^) 85.5 58.1 † *R* ~merge~ = , where 〈*I*(*hkl*)〉 is the mean intensity of symmetry-equivalent reflections. :::

PubMed Central

2024-06-05T04:04:19.425629

2011-2-23

PMC3053210

Background ========== In recent years, Gram-negative pathogens such as *Pseudomonas aeruginosa*, *Acinetobacter baumannii*, and *Klebsiella pneumoniae*have become increasingly resistant to antibiotics. The widespread dissemination of bacterial resistance genes is mediated by horizontal transfer and many of these genes are integrated and expressed as operons in DNA elements called integrons. Integrons are genetic elements that can integrate and disseminate genes as cassettes by a site-specific recombination mechanism \[[@B1]\]. They contain an integrase gene (*intI*), a recombination site (*attI*), and a promoter region (P~c~) that directs the expression of captured genes (Figure [1](#F1){ref-type="fig"}) \[[@B2]\]. Cassettes located within the variable region of integrons all share certain characteristics. First, the integrated cassettes are composed of a gene and an imperfect inverted repeat, called an *attC*site, located downstream of the gene (Figure [1](#F1){ref-type="fig"}) \[[@B3]-[@B6]\]. Second, the boundaries of each integrated cassette are defined by two GTTRRRY sequences that are targets for recombination events mediated by integron integrases (IntIs). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **General structure of class 1 integrons**. Cassettes are inserted in the variable region of integrons by a site-specific recombination mechanism. The *attI1*and *attC*sites are shown by a vertical rectangle and oval, respectively, and promoters are denoted by Pint, Pc and P. Integrated cassettes are composed of a gene and an *attC*recombination site. Genes are as follows: *intI1*, integrase gene; *qacEΔ1*, antiseptic resistance gene; *sul1*, sulphonamide resistance gene; orf5, gene of unknown function. :::  ::: Studies on the site-specific recombination mechanism mediated by IntIs have demonstrated that IntIs form a separate subfamily, characterized by the presence of an additional domain required for their activity, within the larger family of tyrosine recombinases \[[@B7],[@B8]\]. IntIs can share as little as 35% sequence identity, indicating a long evolutionary history for these enzymes. Their catalytic domain is similar to that of other members of the tyrosine recombinase family and contains the conserved residues: Arg~146~-Lys~171~-His~277~-Arg~280~-His/Trp~303~and the nucleophilic tyrosine, Tyr~312~(coordinates are those of IntI1). Unlike other members of the family, the IntI recombinases can exchange DNA using two sites with different structures, the non-palindromic *attI*and palindromic *attC*. Integration of cassettes occurs preferentially by recombination of the *attC*site in a closed-circular cassette with the *attI*site of an integron \[[@B3]\] while excision of a cassette, generating a circular form, occurs preferentially by recombination between two *attC*sites, one of them associated with the upstream cassette \[[@B4]\]. The *attI*and *attC*sequences are complex attachment sites that include the crossover site and additional binding sites (Figure [2](#F2){ref-type="fig"}), suggesting that integrase monomers act as accessory factors at these additional sites \[[@B9]-[@B11]\]. *attI*sites are located at the end of the 5\' conserved region of integrons and their sequences vary considerably. Unlike the *attI*sites, *attC*sites share a common set of characteristics that enable them to be identified despite the diversity of their sequence and size \[[@B6],[@B12]\]. They are characterized by a palindrome of variable length and sequence between the RYYYAAC inverse core site and the GTTRRRY core site \[[@B12]\]. The size of these recombination sites (57 to 141 bp) is currently the main criterion for classification of *attCs*\[[@B13],[@B14]\]. They consist of two pairs of binding sites in opposite orientation (1L-2L and 2R-1R), each pair forming a simple site (LH and RH), separated by a segment of variable length and sequence but including an inverted repeat (Figure [2](#F2){ref-type="fig"}) \[[@B12]\]. These features are generally well recognized by IntI enzymes since many *attC*sites can act as recombination sites for IntIs sharing less than 50% amino acid sequence identity \[[@B6],[@B12],[@B15]-[@B20]\]. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Integron recombination sites**. (a) Sequence of the double strand (ds) *attI1*site. (b) Sequence of the ds *attC~ant(3\'\')-Ia~*site. (c) Secondary structure of the folded bottom strand of the *attC~ant(3\'\')-Ia~*site, according to MFOLD (<http://mfold.bioinfo.rpi.edu/cgi-bin/dna-form1.cgi>). The inverted repeats L, 1L, and 2L, R, 1R and 2R are shown by horizontal black arrows. The *attI1*direct repeats bound by IntI1 are indicated by horizontal lines with an empty arrowhead. The crossover positions are indicated by vertical arrows and the extrahelical bases are identified by asterisks. :::  ::: As members of the tyrosine recombinase family, IntIs use a topoisomerase I type mechanism of cleavage \[[@B8],[@B21],[@B22]\]. Four integrase monomers are involved in the site-specific recombination reaction in which the exchange of one DNA strand contributes to the formation of a Holliday junction \[[@B23],[@B24]\]. For most tyrosine recombinases, this intermediate is resolved by the exchange of the second strand \[[@B22],[@B25]\]. IntIs differ from other tyrosine recombinases by their use of a folded single-stranded *attC*site \[[@B26],[@B27]\] and by the exchange of one strand, with the intermediate possibly resolved by DNA replication (Figure [3](#F3){ref-type="fig"}) \[[@B28],[@B29]\]. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Model of cassette integration mediated by integron integrases**. The site-specific recombination reaction is carried out between the folded bottom strand of an *attC*site and the *attI*site of an integron. (a) Two integrase monomers bind each DNA molecule. (b) Attacking monomer on each DNA molecule cuts one of the two strands. (c) Strand exchange of cut strands. (d) Ligation of exchanged strands forms a Holliday junction. (e) This intermediate structure could be resolved by DNA replication, generating three products, one of which contains the inserted cassette. The attacking monomers are shown by gray circles and the non-attacking monomers are shown by white circles. Y: catalytic tyrosine residues. Adapted from \[[@B28]\]. :::  ::: IntI recombinases bind specifically to the bottom strand (bs) of *attC*\[[@B26]\] and the extrahelical bases resulting from folding of the *attC*bs are important for recognition by IntIs \[[@B27],[@B29],[@B30]\]. The VchIntIA-*Vibrio cholerae*repeat (VCR)~bs~three-dimensional structure showed that the β-4,5 strands from the non-attacking subunits interact with the extrahelical base T12\" (first extrahelical base of the folded *attC*bs) while the α-I~2~helix from the attacking subunits forms several important contacts in *trans*with DNA in the region of the extrahelical base G20\" (second extrahelical base of the folded *attC*bs; Figure [4A](#F4){ref-type="fig"} and [4B](#F4){ref-type="fig"}) \[[@B29]\]. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### ***cis*and *trans*extrahelical base interactions**. (a) *cis*interactions of the VchIntIA non-attacking subunit made by P232, H240 and H241 (in blue) with the extrahelical base T12\'\'. (b) *trans*interactions of the VchIntIA attacking subunit made by Q145, W157, K209, Y210 and W219 (in magenta) with the extrahelical base G20\'\'. The non-attacking subunit is in yellow, the attacking subunit in green, DNA in orange, and the extrahelical bases T12\'\' and G20\'\' are in red. (Based on the structure of the VchIntIA-VCR~bs~complex (PDB:2A3V) \[[@B29]\]. :::  ::: To date, more than 100 IntIs have been reported in the literature and databases and it is estimated that about 10% of partially or completely sequenced bacterial genomes carry genes coding for these enzymes \[[@B31]\]. Activity for cassette excision and integration has been demonstrated for IntI1 \[[@B5],[@B16],[@B32]\], IntI2\*179E \[[@B18]\], IntI3 \[[@B15]\], SonIntIA \[[@B17]\], NeuIntIA \[[@B19]\] and VchIntIA \[[@B32]\]. Moreover, it has been shown that IntIs can excise cassettes containing a variety of *attC*sites \[[@B17],[@B19],[@B33]\]. However, it is not well understood why these enzymes can easily recognize and excise some cassettes, while others are poorly (or not) excised. We compared the ability of several IntIs to excise cassettes flanked by different *attC*sites. Preliminary results of these excision tests combined with molecular modelling based on the structure of the VchIntIA integrase leads us to suggest that IntIs prefer certain *attC*sites to others and that these preferences could be related to the recognition of the extrahelical bases. In this study, we used the *attC~dfrA1~*site upstream of the *sat2*cassette as a template to alter nucleotide sequence and spacing between the extrahelical bases in order to determine how these modifications influence the efficiency of cassette excision by IntI1, IntI2\*179E, IntI3 and SonIntIA. Results ======= Comparative excision activities of IntI1, IntI3, IntI2\*179E, SonIntIA and VchIntIA on cassettes containing different *attC*sites --------------------------------------------------------------------------------------------------------------------------------- In order to determine why some cassettes are excised by several IntIs while others are poorly (or not) excised, we compared the efficiency of five IntIs in excision of cassettes flanked by different *attI*and *attC*sites. Nineteen clones (pLQ423 to pLQ431 and pLQ437 to pLQ446 (Table [1](#T1){ref-type="table"}) containing various resistance gene cassettes cloned into pACYC184 were used to compare the recombination activity of IntI1, IntI3, IntI2\*179E, SonIntIA and VchIntIA by qualitative excision tests (QL-ETs). The results showed a pronounced effect of the identity and spacing of the extrahelical bases in the *attC*sites on the efficiency of cassette excision. All integrases efficiently excised cassettes flanked by *attC*sites whose extrahelical bases are T and G separated by a distance of six nucleotides, with some exceptions for VchIntIA. IntI1 and IntI2\*179E also efficiently excised cassettes with their homologous *attI*site upstream and this same *attC*site downstream. IntI1 was also able to recognize *attI2*but IntI2\*179E was unable to recognize *attI1*. Notably, IntI2\*179E and SonIntIA could easily recognize and excise cassettes with the *attC~dfrA1~*site located upstream of the cassette, whereas IntI1 and IntI3 had only a weak excision activity for the same cassettes and the VchIntIA integrase did not excise any of them. The *attC~dfrA1~*folded bs has the extrahelical bases A71 or A72 (either of these adenines could pair with the thymine at position 23) and C80 separated by a distance of 7 or 8 nucleotides (Figure [5](#F5){ref-type="fig"}). The unusual specificity shown by IntI2\*179E and SonIntIA led us to choose clone pLQ430, with *attC~dfrA1~*upstream of the *sat2*cassette (with its T-N~6~-G-containing *attC*site downstream), for tests of the effect of changes of the upstream *attC~dfrA1~*site on efficiency of cassette excision by the various integrases. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Results obtained from the qualitative excision tests ::: Clone pLQ Upstream *att*site Cassette Integron integrases (IntIs) ----------- ------------------------------------ ------------------------------------ ----------------------------- ----- ----- ----- ----- 423 *attI1* *ant(3\'\')-Ia*T-N~6~-G ++ \- \- \- \- 424 *attI2* *dfrA1*A-N~7-8~-C \- \- \+ \- \- 425 *attI3* *bla~IMP-1~*T-N~6~-G\* \- \- \- \- \- 426 *attC~dfrA1~*A-N~7-8~-C *ant(3\'\')-Ia*T-N~6~-G \+ \- \+ \+ \- 427 *attI1*N/A *dfrA1*A-N~7-8~-C \- \- \- \- \- 428 *attC~ant(3\'\')-Ic~*T-N~6~-G *aac(6\')-Ia*-orfG + orfH T-N~6~-G +++ +++ +++ +++ +++ 429 *attI1*N/A *ant(3\'\')-Ic*T-N~6~-G ++ \- \- \- \- 430 *attC~dfrA1~*A-N~7-8~-C *sat2*T-N~6~-G \+ \+ ++ ++ \- 431 *attC~aac(6\')-Ib~*TC-N~6~-G *bla~oxa10~*T-N~6~-G \- \- \- \- \- 437 *attC~ant(3\'\')-Ic~*T-N~6~-G *dfrA1*A-N~7-8~-C \- \- \+ \- \- 438 *attC~ant(3\'\')-Ic~*T-N~6~-G *bla~IMP-1~*T-N~6~-G\* +++ +++ +++ +++ \- 439 *attI1*N/A *bla~IMP-1~*T-N~6~-G\* \+ \- \- \- \- 440 *attI1*N/A *aac(6\')-Ia*-orfG + orfH T-N~6~-G +++ \- \- \- \- 441 *attI2*N/A *aac(6\')-Ia*-orfG + orfH T-N~6~-G ++ \- +++ \+ \- 442 *attI3*N/A *aac(6\')-Ia*-orfG + orfH T-N~6~-G \+ \- \- \- \+ 443 *attC~ant(3\'\')-Ic~*T-N~6~-G *ant(3\'\')-Ia*T-N~6~-G +++ +++ +++ +++ +++ 444 *attC~dfrA1~*A-N~7-8~-C *dfrA1*A-N~7-8~-C \+ \+ +++ +++ \- 445 *attC~aac(6\')-Ia~*~-orfG~T-N~6~-G *ant(3\'\')-Ia*T-N~6~-G +++ +++ ++ ++ ++ 446 *attC~blaIMP-1~*T-N~6~-G *aac(6\')-Ia*-orfG + orfH T-N~6~-G +++ +++ +++ +++ \- -, No excision; +, weak excision (\<20%); ++, moderate excision (20 to 75%); +++, strong excision (\>75%); \*, additional secondary structure between the extrahelical bases. ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Secondary structure of the folded bottom strand of the *attC~ant(3\'\')-Ia~*, *attC~sat2~*, *attC~aac(6\')-Ia-~*~orfG~and *attC~dfrA1~*sites**. The extrahelical bases are identified by arrows. :::  ::: Comparative excision activities of IntI1, IntI3, IntI2\*179E and SonIntIA on cassettes with an upstream *attC~dfrA1~*site or mutant *attC~dfrA1~*sites ------------------------------------------------------------------------------------------------------------------------------------------------------ We then determined the effect of different *attC*structures on excision by IntI1, IntI3, IntI2\*179E and SonIntIA. We made several mutants of the *attC~dfrA1~*site, upstream of the *sat2*cassette in pLQ430, with altered extrahelical base identity and spacing, and used QN-ETs. The first set of mutants was made using various substitutions to determine whether the presence of a cytosine or a guanine at position 80 of the *attC~dfrA1~*bottom strand (bs) could alter recognition and excision by IntI1, IntI3, IntI2\*179E and SonIntIA (Figure [6A](#F6){ref-type="fig"}). First, we compared the ability of the four IntIs to carry out cassette excision on clones pLQ430 (A-N~7-8~-C *attC~dfrA1~*+ *sat2*) and pAL4316 (A-N~7-8~-G *attC~dfrA1~*+ *sat2*). The results of our QN-ETs showed that the excision activity of IntI1 remained very weak on the *sat2*cassette when we changed the cytosine at position 80 to guanine, keeping adenines located at positions 71 and 72 of the *attC~dfrA1~*site and the distance between the extrahelical bases at seven or eight nucleotides (Figure [7A](#F7){ref-type="fig"}). The efficiency of recognition and excision of this mutant cassette by IntI3 was only slightly decreased. However, the excision by IntI2\*179E and SonIntIA were decreased from 51% to 21% and from 25% to 12%, respectively, by the C80G substitution of the *attC~dfrA1~*site. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Secondary structure of the folded bottom strand of several mutants of the *attC***~***dfrA1***~**site**. (a) First set of mutants. (b) Second set of mutants. The extrahelical bases are identified by arrows. The ‡ indicates that the first extrahelical base is located on the opposite side of the folded bottom strand structure. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Excision percentage of *sat*cassette by the IntI1, IntI3, IntI2\*179E and SonIntIA integron integrases**. For each integrase, the bars indicate the excision percentage determined in the *in vivo*quantitative excision tests. Excision percentages correspond to the average of two independent assays. We tested the excision percentage of the gene cassette *sat2*\[pLQ430 and several mutants (A: first set of mutants and B: second set of mutants) created from the clone pLQ430 carrying the *attC~dfrA1~*+ *sat2*cassette\] coding for streptothricin resistance. The ‡ indicates that the first extrahelical base is located on the opposite side of the folded bottom strand structure. Error bars show standard error. :::  ::: We also compared the excision activity of IntI1, IntI3, IntI2\*179E, and SonIntIA on clones pAL4318 (T-N~6~-C *attC~dfrA1~*+ *sat2*) and pAL4319 (T-N~6~-G *attC~dfrA1~*+ *sat2*). These mutants were made by A22T substitution, and A22T and C80G substitutions, respectively, on *attC~dfrA1~*. When the first extrahelical base is a thymine (at position 73) and the distance between the extrahelical bases is six nucleotides, the C80G substitution slightly decreased excision by IntI1 and IntI3. The decrease was more pronounced with IntI2\*179E and SonIntIA, from 71% to 51% and from 54% to 33%, respectively. Together, these results suggest that the presence of a cytosine as the second extrahelical base favors cassette excision by IntI, in particular by IntI2\*179E and SonIntIA. These QN-ETs therefore supported our hypothesis that IntIs differ in their preferences for the extrahelical bases. We then tested several other mutants of the *attC~dfrA1~*site, made in order to eliminate ambiguities in the spacing of the extrahelical bases. The second set of mutants was based upon double deletions of bases A22 and T23 and A72 and T73 from the wild-type bs of the *attC~dfrA1~*site. This forced the spacing to be six bases and made these *attC*s more comparable to T-N~6~-G *attC*s (Figure [6B](#F6){ref-type="fig"}). Several other substitutions were added and tested to determine whether the presence of an adenine or a thymine at position 69 or the presence of a cytosine or a guanine at position 76, of the *attC~dfrA1~*ΔAT22-ΔAT72 site (corresponding to positions 72 and 80 of the wild type site) could alter its recognition by IntI1, IntI3, IntI2\*179E, and SonIntIA (Figure [6B](#F6){ref-type="fig"}). The results of our QN-ETs showed that a reduction of the distance between the extrahelical bases of the *attC~dfrA1~*site (A-N~6~-C instead of A-N~7-8~-C) led to a significant increase of excision by IntI1, IntI2\*179E, and SonIntIA, while the activity of IntI3 was unchanged (Figure [7B](#F7){ref-type="fig"}). This preference for a shorter distance between the extrahelical bases was particularly marked for the IntI1 integrase. The excision activity of IntI1 increased from 3% to 33.5% while that of IntI2\*179E increased from 51% to 79% and that of SonIntIA increased from 26% to 45% when the distance between the extrahelical bases is six nucleotides rather than seven or eight. Mutations of the extrahelical bases to those of the consensus of the *attC*s most easily excised by IntI1 (A69T and C76G) were tested individually and in combination on the *attC~dfrA1~*∆AT22-∆AT72 site (clone pAL4305: A-N~6~-C). The A69T substitution increased excision of the *sat2*cassette by IntI1 and IntI3, from 33.5% to 41.5% and from 17% to 38.5%, respectively, when a cytosine is present at position 76, whereas excision by IntI2\*179E and SonIntIA was unchanged. The C76G substitution significantly decreased excision of this cassette by IntI1 (33.5% to 14%), IntI3 (17% to 4%), IntI2\*179E (79% to 45.5%) and SonIntIA (45% to 30%) when an adenine is located at position 69. The combination of the A69T and C76G substitutions (clone pAL4311) increased excision of the *sat2*cassette by IntI1 (33.5% to 43%) and IntI3 (17% to 24%) while it decreased excision by IntI2\*179E and SonIntIA, from 79% to 58% and from 45% to 35%, respectively. Comparison of the excision activity of the four IntIs on clones pAL4310 (A-N~6~-G *attC~dfrA1~*+ *sat2*) and pAL4311 (T-N~6~-G *attC~dfrA1~*+ *sat2*) showed that the efficiency of IntI1, IntI3, IntI2\*179E, and SonIntIA in excision of the *sat2*cassette was increased by the A69T substitution when the second extrahelical base is a guanine and the distance between the extrahelical bases is six nucleotides. Comparison of excision on clones pAL4308 (T-N~6~-C *attC~dfrA1~*+ *sat2*) and pAL4311 (T-N~6~-G *attC~dfrA1~*+ *sat2*) showed that the activity of IntI1 was unchanged by the C76G substitution when the first extrahelical base is a thymine and the distance between the extrahelical bases is six nucleotides. However, the C76G substitution significantly decreased excision by IntI3, IntI2\*179E and SonIntIA, from 38.5% to 24%, 83% to 58% and 44% to 35%, respectively, when a thymine is at position 69 of the *attC~dfrA1~*site. The VchIntIA-VCR~bs~three-dimensional structure shows that IntIs interact closely with the extrahelical bases T12\" and G20\" \[[@B29]\]. We, therefore, tested three other mutant cassettes to determine whether variation of the distance between the extrahelical bases could alter interaction between IntIs and their substrates. The results of our QN-ETs using clones pAL4307 (A-N~5~-C *attC~dfrA1~*+ *sat2*) and pAL4309 (T-N~5~-C *attC~dfrA1~*+ *sat2*) showed that IntI1 can excise the *sat2*cassette (15% excision) when the *attC~dfrA1~*site located upstream contains the extrahelical bases T and C separated by five nucleotides (pAL4309) but not when the extrahelical bases are A and C (pAL4307) separated by the same distance. However, the IntI3, IntI2\*179E and SonIntIA integrases were very inefficient in excision of this cassette with either of these clones. Folded bottom strand *attC*s found in integrons are characterized by two extrahelical bases located on the same side of the structure. In order to determine if IntIs can recognize and excise cassettes containing an *attC*site characterized by two extrahelical bases located on either side of the folded bs structure, we changed the position of the first extrahelical base and tested the excision activity of IntI1, IntI2\*179E, IntI3 and SonIntIA. Johansson *et al.*\[[@B27]\] previously reported that binding to the *attC~aadA1~*bs by IntI1 was decreased by deletion of the T32 extrahelical base and by insertion of a T or an A between positions 16 and 17 to generate a bulge on the opposite side of a potential stem-loop. They showed that the presence of an adenine between positions 16 and 17 weakly affects the binding by IntI1 while the presence of a thymine significantly decreases its binding \[[@B27]\]. However, they did not test the excision activity of IntI1 on a cassette containing this mutant *attC*site. Our QN-ETs using clone pAL4304 (A^‡^-N~6~-C *attC~dfrA1~*+ *sat2*: the ‡ indicates that the first extrahelical base is located on the opposite side of the folded bottom strand structure; the two extrahelical bases are located on either side of the folded bs structure) showed that IntI1 was not able to excise the *sat2*cassette when the altered A^‡^-N~6~-C *attC~dfrA1~*site is located upstream, while IntI3 showed an excision activity of only 3%. However, the IntI2\*179E integrase showed an excision activity of 24% on this very unusual *attC*site, while SonIntIA showed a level of excision of 9%. Discussion ========== The dissemination of antibiotic resistance by mobilization of resistance gene cassettes is a key factor affecting the clinical usefulness of antibiotics. The recruitment of these cassettes by integrons is carried out by IntIs and the efficiency of recombination by these enzymes varies greatly. In this work, we studied some parameters that affect the specificity of these recombinases for *attC*sites. Influence of extrahelical base identity, spacing and position on specificity of IntIs ------------------------------------------------------------------------------------- ### Extrahelical base identity The data presented in this study showed that the IntI1 was efficient in cassette excision using T-N~6~-G or T-N~6~-C *attC*sites, while IntI3 recognized a limited range of *attC*s and recombined mainly cassettes with T-N~6~-C *attC*sites (Figure [7](#F7){ref-type="fig"}). For their part, IntI2\*179E and SonIntIA tolerated changes to the identity of extrahelical bases, as they efficiently excised cassettes with *attC*s characterized by most of the extrahelical base combinations tested (A-N~6~-C, A-N~6~-G, T-N~6~-C and T-N~6~-G). In their IntI binding study, Johansson and colleagues \[[@B27]\] showed that substitution of the first extrahelical base (T) with a cytosine or an adenine, and substitutions of the second extrahelical base (G) with any of the alternative bases, does not affect binding of IntI1 to the *attC~ant(3\")-Ia~*site. However, they did not test the excision efficiency of IntI1 on these mutant cassettes. Taken together, these results suggest that the extrahelical base identities do not affect the binding of IntIs but do influence the recombination reaction. They confirm our hypothesis that *attC*preferences of IntIs are related to the recognition of the extrahelical bases. Surprisingly, our results showed that the presence of a cytosine rather than a guanine at the second extrahelical base position increased cassette excision by IntI3, IntI2\*179E and SonIntIA, whether the first extrahelical base is a thymine or an adenine. The excision activity of IntI1 was increased by the presence of a cytosine at the second extrahelical position only when the first extrahelical base is an adenine. These results were unexpected since most *attC*sites have a guanine at the second extrahelical position. They suggest that the emergence of cassettes with increased mobility is possible. The only example of a cassette containing an *attC*site with a cytosine as the second extrahelical base is *bla~VIM-2~*, but its mobility remains to be evaluated. Results recently obtained by Bouvier *et al.*\[[@B34]\] show that the *attC*x *attC*recombination (*attC~aadA7~*x VCR) carried out by IntI1 is slightly increased by the guanine (G) to cytosine (C) substitution of the second extrahelical base of the VCR when the first extrahelical base is a thymine. In their study, mutations were made on the downstream *attC*partner whereas they were made on the upstream *attC*partner in this study. Our QL-ETs and results obtained by Bouvier *et al.*\[[@B34]\] suggest that changes to the downstream *attC*partner may have a greater impact on the ability of IntIs to excise cassettes. ### Distance between the extrahelical bases Since the VchIntIA-VCR~bs~three-dimensional structure shows that IntIs interact closely with the extrahelical bases T12\" and G20\" \[[@B29]\], we tested the effect of the distance between these bases on this interaction. Our data showed that IntI1, IntI3, IntI2\*179E and SonIntIA most efficiently excised cassettes containing *attC*s when the spacing between the extrahelical bases was six nucleotides. They also showed that IntI2\*179E and SonIntIA were more tolerant than IntI1 and IntI3 to changes in spacing between the extrahelical bases. Johansson *et al.*\[[@B27]\] showed that increasing the distance between the two extrahelical bases (from six to eight or 10) does not affect binding of IntI1 to the *attC~ant(3\")-Ia~*site, but they did not test the recombination activity of IntI1 on these mutant *attC*sites. It appears that the distance between the extrahelical bases is important for the excision reaction but not for bs *attC*binding by IntIs. ### Position of the first extrahelical base As mentioned above, Johansson *et al.*\[[@B27]\] reported that the presence of an adenine between positions 16 and 17 on the opposite side of a potential stem-loop combined with the deletion of the T32 extrahelical base decreased binding to the *attC~aadA1~*bs by IntI1. In our study, we observed that, despite the fact that their excision activity is decreased, IntI2\*179E and SonIntIA can excise the *sat2*cassette when the altered A^‡^-N~6~-C *attC~dfrA1~*site, characterized by an extrahelical base at position 22 and another at position 78, is located upstream. IntI1 and IntI3 have no apparent activity on this very atypical site. The influence of the first extrahelical base position (corresponding to T32) was also tested by Bouvier *et al.*\[[@B34]\] and they showed that the re-localization of this base at the corresponding location on the opposite strand leads to a decrease of VCR~bs~excision by IntI1. Together, these results show that, in addition to being more tolerant to changes in the identity and spacing between the extrahelical bases, IntI2\*179E and SonIntIA are more tolerant than IntI1 and IntI3 with respect to the position of the first extrahelical base. They also suggest that changing the first extrahelical base to a thymine decreases the binding by IntIs and probably affects the excision activity. However, the re-localization of the first extrahelical base as an adenine does not affect the binding but the excision activity is decreased. QN-ETs using the *attC~dfrA1~*site: IntI1 versus IntI2\*179E ------------------------------------------------------------ The results of our QN-ETs using IntI1 and IntI2\*179E with the T-N~6~-G *attC~dfrA1~*site raised an important issue. It is not clear why the excision percentage observed with IntI1 on cassettes containing the T-N~6~-G *attC~dfrA1~*site is not higher than that observed with IntI2\*179E on the same substrates. Our QL-ETs showed that IntI1 is generally more effective than IntI2\*179E in excision of cassettes containing T-N~6~-G *attC*sites (for example, *aac(6\')-Ia*-orfG and *ant(3\'\')-Ia*). It is possible that these differences are explained by the presence of different nucleotides near the extrahelical bases of the *attC~dfrA1~*mutants used for our QN-ETs and those of *attC*sites used in our QL-ETs. It has been shown that the identity of the bases located near the extrahelical bases influences the binding of IntI1 \[[@B27]\]. Structural elements of IntI1, IntI2\*179E, IntI3, SonIntIA and VchIntIA involved in *attC*recognition ----------------------------------------------------------------------------------------------------- IntIs bind specifically to the *attC*bs \[[@B26]\] and the extrahelical bases resulting from its folding are important for recognition by these enzymes \[[@B27],[@B29]\]. The three-dimensional structure of the VchIntIA-VCR~bs~complex reveals that the extrahelical base T12\'\' is stabilized by *cis*interactions with the β-4,5 strands from the non-attacking subunits by becoming inserted between two stacked histidines (H240 and H241 in VchIntIA (Figure [4A](#F4){ref-type="fig"}); H250 and H251 in IntI1) and a highly conserved proline \[P232 in VchIntIA (Figure [4A](#F4){ref-type="fig"}); P242 in IntI1\] \[[@B29]\]. The attacking subunits make important DNA contacts in *trans*with the extrahelical base G20\'\' through interactions with Q145, W157, K209, Y210 and W219 in VchIntIA (Figure [4B](#F4){ref-type="fig"}) that correspond to K156, R168, K219, Y220 and W229 in IntI1 \[[@B29]\]. The protein-DNA interactions are otherwise essentially nonspecific \[[@B29]\]. We compared the region located between the αI~2~helix and the β-4,5 strands of IntI1, IntI3, IntI2\*179E, SonIntIA and VchIntIA and observed that many residues are conserved among these enzymes (Figure [8](#F8){ref-type="fig"}). This reflects the importance of this region in *attC*recognition by IntIs. However, we identified some differences between IntI2\*179E and SonIntIA versus IntI1 and IntI3 sequences that could be responsible for the greater versatility of IntI2\*179E and SonIntIA in excision of cassettes containing non-T-N~6~-G *attC*s. One interesting difference is the presence of two cysteine residues in the β-4 and β-5 strands of IntI2\*179E and SonIntIA. The same positions are occupied by a serine and an arginine in IntI1 and IntI3. We previously found that the cysteine residue in the β-5 strand is essential to the excision activity of *Shewanella*-type integrases while the cysteine in the β-4 strand is less important \[[@B33]\]. Mutagenesis of the two cysteines studied in SonIntIA suggests that there is no disulfide bridge between the β-4 and β-5 strands of these integrases \[[@B33]\]. However, we do not know if these cysteine residues play a role in the ability of IntI2\*179E and SonIntIA to tolerate changes to the extrahelical base identity and spacing. Other differences (indicated by arrows in Figure [8](#F8){ref-type="fig"}) are located at various positions between the αI~2~helix and the β-4,5 strands of IntIs and could also contribute to the greater versatility of IntI2\*179E and SonIntIA in excision of cassettes. ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Sequence alignment of the αI**~**2**~**helix and the β-4,5 strands region of some integron integrases (IntIs)**. IntI2\*, class 2 IntI from Tn*7*; SonIntIA, IntI from *Shewanella oneidensis*MR1; VchIntIA, IntI from the *Vibrio cholerae*chromosomal integron; IntI1, class 1 IntI from plasmid pVS1; IntI3, class 3 IntI from a *S. marcescens*plasmid. Residues potentially related to the difference in specificity of IntI2 and SonIntIA versus IntI1 and IntI3 are identified by arrows. Positions of the αI~2~helix and the β-4,5 strands in this figure are based on the VchIntIA structure (Protein Data Bank accession no. 2A3V) \[[@B29]\]. :::  ::: Some residues and motifs located outside the αI~2~helix and the β-4,5 strands could also be related to the different preferences of IntIs for extrahelical base identity and spacing. For example, Demarre *et al.*\[[@B35]\] found mutations of IntI1, with higher activity on wild type and mutant *attC*sites, in the loop located between the β-1 and β-2 strands. Interestingly, the β-2 strand contains one of the two residues, R168 (W157 in the sequence of VchIntIA), that interact with the extrahelical base G20\" \[[@B29]\]. Also, Johansson *et al.*\[[@B30]\] showed that substitution of the tryptophan residue at position 199 of IntI1 with alanine, which is aliphatic, small, and uncharged, decreases DNA binding. Interestingly, when it was replaced by an aromatic residue (W199Y), it regained its affinity for *attC~bs~*\[[@B30]\]. These results suggest that the presence of an aromatic or a bulky amino acid residue at this position is important. The authors propose that the decreased binding of the IntI1W199A mutant could be explained by structural changes of the αI~2~helix and the β-4,5 strands region \[[@B30]\]. Thus, in addition to residues that interact directly with the extrahelical bases, we must also consider the residues that are important to protein structure in our attempt to identify the factors that may affect the ability of IntIs to excise cassettes. The alignment of IntIs used in our study show that the position 199 is occupied by a tryptophan residue in IntI1 and IntI3 while it is occupied by a glutamine in IntI2\*179E and SonIntIA. This difference may contribute to explain the different ability to excise cassettes of these IntIs. In summary, it seems that IntI2 and SonIntIA have an evolutionary path that is different from IntI1 and IntI3, in their ability to recognize and excise cassettes. IntI2\*179E and SonIntIA, although generally less efficient in cassette excision, tolerate a wider variety of configurations of the extrahelical bases of *attC*sites. We believe that the ability of IntI2\*179E and SonIntIA to excise cassettes containing *attC*s characterized by a broader range of extrahelical base identity and spacing could be related to a greater flexibility of their αI~2~helix and their β-4,5 strand domains. Acquisition of new cassettes ---------------------------- Analysis of the variable region of class 1 integrons showed that these multiresistance integrons contain a large number of different cassettes and those containing T-N~6~-G *attC*s are found at various positions within the variable region. This can be explained by our finding that these cassettes can be easily recognized and excised by IntI1. The IntI1 integrase is efficient in cassette integration and, since cassettes are preferentially integrated by *attI*x *attC*recombination \[[@B3]\], cassette order tends to reflect the order of introduction of antibiotics, with the most recently acquired cassettes closest to the promoter. Class 2 integrons, carrying the IntI2\* or IntI2 integrases, as well as class 3 integrons, carrying the IntI3 integrase, have only a limited range of cassettes. Among the arrangements of cassettes associated with the class 2 integrons are *dfrA1*-*sat2*-*aadA1*-*orfX*, *estX*-*sat2*-*aadA1*-*orfX*and *sat2*-*aadA1*-*orfX*\[[@B18]\]. The position of cassettes within this class of multiresistance integrons is conserved since most class 2 integrons carry the inactive IntI2\* integrase. Recently, class 2 integrons with active integrases have been found \[[@B36],[@B37]\] but they still have a relatively limited number of cassette arrangements. The *dfrA1*and *estX*cassettes contain non-T-N~6~-G *attC*s while the *sat2*, *aadA1*and *orfX*cassettes contain T-N~6~-G *attC*s. Two different arrangements of resistance gene cassettes were found to be associated with class 3 integrons: *bla~IMP-1~*-*aac(6\')-Ib*and *bla~GES-1~*-*bla~OXA~*/*aac(6\')-Ib*\[[@B38],[@B39]\]. The *bla~GES~*~-1~cassette contains a T-N~6~-G *attC*site while *bla~IMP-1~*, *aac(6\')-Ib*, and *bla~OXA~*/*aac(6\')-Ib*cassettes contain non-T-N~6~-G *attC*s. The cases of the *dfrA1*, *bla~IMP-1~*and *bla~GES-1~*cassettes are particularly interesting. Although first found in class 2 and class 3 integrons, these cassettes are now more frequently disseminated by class 1 integrons. This may reflect the greater versatility of IntI1 in cassette rearrangement. Moreover, the *dfrA1*cassette is nearly always located in first position in class 1 integrons. As we have shown, cassettes containing non-T-N~6~-G *attC*s are weakly excised by IntI1, which suggests that integration of cassettes containing *attC*sites like that of *dfrA1*may hinder their own subsequent excision and that of their downstream neighbors (*attC*x *attC*excision). In the case of the *bla~IMP-1~*cassette, the additional secondary structure located between the extrahelical bases of the *attC*bs does not interfere with the ability of IntIs to excise cassettes when this recombination site is located either upstream or downstream of the gene (Table [1](#T1){ref-type="table"}). The *bla~GES-1~*cassette is associated with a T-N~6~-G *attC*site that would facilitate its acquisition by integrons (in particular, class 1 integrons) and its dissemination among bacteria. Leon and Roy \[[@B40]\] have shown that there is no relationship between a cassette structural gene and its associated *attC*site. According to their new model for gene cassette formation, group IIC-*attC*introns can target separately a transcriptional terminator adjoining a gene and an isolated *attC*. Thereafter, the gene and the *attC*can be joined by homologous recombination between the introns, followed by transcription, RNA splicing, and reverse transcription to lead to the formation of a cassette \[[@B40]\]. The characteristics of the structure of the bottom strand of the *attC*site would determine the subsequent mobility of the cassette. Conclusions =========== In conclusion, this work and previous studies \[[@B27],[@B29],[@B32],[@B34]\] clearly show that the *attC*structure is an important factor that facilitates the integration of new cassettes into integrons. In our study, we carried out excision tests with several IntIs on cassettes containing a wide variety of *attC*sites. This work could aid the development of a site-specific recombination system using the IntIs. In contrast to the Cre recombinase of the Cre-*lox*system, IntIs have a more relaxed specificity for their recognition sites *attI*and *attC*. The main advantage of a site-specific recombination system using an IntI is that it would allow insertion of genes (cassettes) in tandem. The results presented in this article could be used to optimize such a system. Methods ======= Bacterial strains and growth media ---------------------------------- *Escherichia coli*strains were cultured at 37°C in Luria-Bertani (LB) broth or on LB agar supplemented with ampicillin (100 μg/mL; Sigma, MO, USA), chloramphenicol (50 μg/mL; Sigma) or streptothricin (3 μg/mL). DH5α cells \[F^-^*endA1 glnV44 thi-1 recA1 relA1 gyrA96 deoR nupG*Φ80d*lacZ∆*M15 ∆(*lacZYA-argF*)U169 *hsdR17*(*r~K~*^-^*m~K~^+^*) λ^-^\] were used as a host for construction and maintenance of all plasmid clones and for QL-ETs, while HB101 cells \[F^-^*mcrB mrr hsdS20*(r~B~^-^m~B~^-^) *recA13 leuB6 ara-14 proA2 lacY1 galK2 xyl-5 mtl-1 rpsL20*(Sm^R^) *glnV44 λ*^-^\] were used for QN-ETs. Bioinformatic analysis ---------------------- Sequence analysis was done using the Genetics Computer Group programs (Wisconsin Package version 10.3; Accelrys). Folding of *attC*bottom strands was done using the MFOLD software (<http://mfold.bioinfo.rpi.edu/cgi-bin/dna-form1.cgi>). Mutagenesis method ------------------ Several mutations were introduced within the *attC~dfrA1~*site located upstream of the *sat2*cassette cloned into pACYC184 (clone pLQ430). Specific mutations were introduced into pLQ430 using the QuickChange site-directed mutagenesis system including *Pfu*Turbo (Stratagene, CA, USA) DNA polymerase. Primer pairs, designed with the OLIGO software package (version 4.1; National Biosciences, MN, USA), were used to create 16 mutants of the *attC~dfrA1~*site. The forward primers are shown in Additional File [1](#S1){ref-type="supplementary-material"}. Mutagenesis products were digested with DpnI, transformed into *E. coli*DH5α, grown in LB medium for 1h and selected for chloramphenicol resistance by plating on LB agar plates containing chloramphenicol. DNAs from several colonies were purified using a QIAprep spin miniprep kit (Qiagen, Düsseldorf, Germany) and sequenced to confirm the presence of desired mutations and the integrity of surrounding sequences. The isolateswere maintained as glycerol stock cultures at -80°C. Qualitative excision tests (QL-ETs) ----------------------------------- IntI1, IntI3, IntI2\*179E, SonIntIA and VchIntIA clones (see Additional File [2](#S2){ref-type="supplementary-material"}) were introduced by transformation into *E. coli*DH5α containing various cassettes cloned into pACYC184 (Table [1](#T1){ref-type="table"}). *E. coli*was grown in LB medium at 37°C to an optical density at 600 nm of 0.5. Cassette excision was induced by the overexpression of the integrase gene using 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG; Sigma) and incubation at 37°C overnight. Cell cultures were done in the presence of ampicillin and chloramphenicol. Plasmid DNA was subsequently extracted from 5-ml cultures with a QIAprep spin miniprep kit (Qiagen). In order to determine the ability of IntIs to excise cassettes, we used polymerase chain reaction (PCR) primers pACYC184-5\' and pACYC184-3\' (See Additional File [1](#S1){ref-type="supplementary-material"}) to detect reductions in length of cassette clones. PCR conditions were 5 min at 95°C, 30 cycles consisting of 30 s at 95°C, 30 s at 62°C and 3 min 30 s at 68°C, and a final elongation step of 5 min at 68°C. Quantitative excision tests (QN-ETs) ------------------------------------ Cells containing integrase clones were transformed by various plasmids containing gene cassettes cloned into pACYC184 (See Additional File [2](#S2){ref-type="supplementary-material"}). One colony of each double transformant was used to inoculate 5 mL of LB medium and grown at 37°C to an optical density at 600 nm of 0.5. Cell cultures were done in the presence of ampicillin and chloramphenicol. Isopropyl-β-thio-galactoside (IPTG) was then added to a final concentration of 1 mM to induce cassette excision and cultures were incubated at 37°C overnight; plasmid DNA extractions (Qiagen) were done on each culture. DNA was incubated at 37°C 1h with PstI to digest the integrase clone and prevent its co-transformation into the *E. coli*strain used to determine, by replica plating, the antibiotic resistance cassettes that were excised. Thereafter, cassette clones were transformed into *E. coli*HB101 and colonies selected for chloramphenicol resistance. One hundred colonies of each transformation were replicated on LB chloramphenicol + streptothricin plates and incubated at 37°C overnight. The proportion of transformants that could not grow indicated the excision percentage of the *sat2*cassette for each integrase and each upstream *attC*tested. Abbreviations ============= A: adenine; Arg: arginine; bs: bottom strand; C: cytosine; G: guanine; His: histidine; IntI: integron integrase; LB: Luria-Bertani; Lys: lysine; N: nucleotide; QL-ETs: qualitative excision tests; QN-ETs: quantitative ETs; R: arginine (amino acid symbol) or purine (nucleotide symbol); T: thymine; Tn: transposon; Trp: tryptophan; Tyr: tyrosine; VCR: *Vibrio cholerae*repetitive DNA sequence; W: tryptophan; Y: pyrimidine. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= AL performed the experiments and wrote the article. PHR supervised the work and participated in writing the article. Supplementary Material ====================== ::: {.caption} ###### Additional File 1 **Primers used in this study**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional File 2 **Integrase clones and mutant cassette clones used in this study**. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was supported by Canadian Institutes for Health Research (CIHR) grant MT-13564 to PHR.

PubMed Central

2024-06-05T04:04:19.427355

2011-2-18

PMC3053211

Introduction ============ The choice to be active or sedentary depends in part on individual differences in the motivation to be active or sedentary, as well as constraints on access to sedentary or active alternatives \[[@B1],[@B2]\]. The motivation to be active or sedentary can be operationalized by providing children with the choice to be active or sedentary, varying the behavioral costs to obtain access to the alternatives, and quantifying the amount of work or effort the children will do to gain access to the alternatives. This provides an index of the relative reinforcing value of being active (RRV~ACT~) or sedentary (RRV~SED~). The relative reinforcing value of physical activity has been associated with physical activity levels, with children who find physical activity more reinforcing also being the most active \[[@B3]-[@B5]\]. In addition, there are strong individual differences in the reinforcing value of physical activity, as obese children find physical activity less reinforcing than leaner children \[[@B6]\]. Reported time spent watching television, as one component of being sedentary, is cross-sectionally correlated with obesity in children and adults \[[@B7]-[@B10]\], as well as being a risk factor for the development of obesity in children \[[@B11],[@B12]\]. Given the role of a sedentary lifestyle in weight gain and the development of obesity, research suggests that reducing sedentary behavior may be a valuable tool in prevention \[[@B13],[@B14]\] and treatment of pediatric obesity \[[@B15],[@B16]\]. There are two potential ways in which reducing sedentary behaviors can be associated with weight changes. As sedentary behaviors are reduced, complementary reductions in energy intake may occur, or as sedentary behaviors are reduced, children may substitute physical activity for sedentary behaviors. Despite the importance of the motivation to be sedentary or active as a predictor of a child\'s lifestyle choices, there has been no research on how the motivation to be active or sedentary is associated with weight change when sedentary behaviors are reduced. The purpose of this study is to report on how individual differences in the RRV~SED~or RRV~ACT~are correlated with weight loss during an intervention when sedentary behaviors are reduced. Methods ======= Participants ------------ Participants were 56 overweight/obese, 8-12 year old American children, recruited from flyers, a direct mailing, and a pre-existing database. All of the children were considered to be overweight or at risk for overweight, defined as a Body Mass Index (BMI) percentile adjusted for age and sex at or above the 85^th^percentile \[[@B17]\]. Criteria for participation included the following; at least one parent agreed to help their child reduce targeted sedentary behaviors, and measure usual physical activity and dietary intake; the participating child must have engaged in at least 18 hours of targeted sedentary behaviors per week; could not participate in swimming and/or weight training for greater than 5 total combined hours per week; no activity restrictions or physical limitations that could interfere with changes in physical activity, such as developmental disability or injury; no psychopathology or developmental disabilities that would limit participation. All procedures and measures were approved by the University at Buffalo Children and Youth Institutional Review Board. Design and Procedure -------------------- After completing the phone screen families were scheduled for an orientation. During the orientation parents and children completed consent and assent forms, child height and weight were measured, and families were oriented on the TV allowance device, the physical activity monitor, and activity diaries. Interested families were fitted with an accelerometer which was worn on two weekdays and one weekend day. Families were scheduled for two laboratory sessions. The child\'s RRV~SED~and RRV~ACT~, were measured during the first session, and the accelerometers were calibrated during the second session using a progressive treadmill test. After laboratory testing, families were scheduled for 5 home visits throughout the nine week intervention, and children were scheduled to wear the accelerometer on three randomly selected days, two weekdays and one weekend day. Children were also instructed to self-monitor time on each sedentary and active behavior to ensure adherence with the experimental manipulation in a seven day diary during the last week of each phase. Activity devices were downloaded at home visits 3-5 and self report diaries were checked for accuracy. Weight was measured at each home visit and height was measured at the last home visit. Reminder phone calls were made to ensure the child wore the activity device on their scheduled day. Parent and child manuals were provided to each family explaining the study goals as well as to provide techniques for praise and reducing sedentary behaviors. During the first of five home visits, TV Allowance™ devices were connected to each TV and computer in the home, families were trained on using the devices and asked to maintain their usual pattern of sedentary behaviors, physical activity and dietary intake through the baseline phase; activity devices were fitted to the child, the child was trained on recording in the weekly diary; and weight was measured. During the second home visit each device was checked and TV and computer hours were recorded. At the third home visit, eligibility was determined, amount of screen time was calculated from the allowance devices, and the devices were programmed to decrease TV and computer use by 25% for the next three weeks. During the fourth home visit, devices were programmed to decrease TV and computer use by 50% for the next three weeks. Devices were removed at home visit 5. Two positive reinforcement techniques were used to facilitate adherence to the experimental protocol, praise and monetary reinforcement. Parents were instructed to praise their children when they observed behavior changes in the appropriate direction, and to be very specific in stating what the praise is for and to be consistent in using praise. Families earned up to \$325.00 for participation in the 9 week-study. Children earned up to \$15/week during the 25% and 50% reduction phases for making the reductions in targeted sedentary behavior (\$90), with the amount proportional to the degree of change, with \$10 for reaching the decrease goals and an additional 1\$ for every hour under their goal up to \$5. During the baseline phase families could earn up to \$25/week (\$75) for completing measurements, and up to \$10 per week (\$60) for completing measurements during the 25% and 50% reduction phases. Families also earned \$100 for completing the study. Families could distribute the family money as they chose. Measurement ----------- ### Demographic variables and medical history Family size, family income, parent educational level and racial/ethnic background were obtained using a standardized questionnaire. Current medical problems, including psychiatric diagnoses and eating disorders were assessed at baseline by parent interview. ### Weight, height, BMI Child weight was assessed by use of a Tanita BWB-800P digital scale. Height was assessed using a Digi-Kit digital stadiometer. On the basis of the height and weight data Body Mass Index (BMI) is calculated according to the following formula: (BMI = kg/m^2^). Children were considered overweight if they were at or above the 85^th^BMI percentile for their age and sex \[[@B17]\]. ### Liking of activities, food, and videos/computer games Liking of the activities, videos or computer games was measured on 7 point Likert-type scales anchored by 1 (Do not like) to 7 (Like very much) \[[@B18]\]. ### The relative reinforcing value of sedentary behavior (RRV~SED~) and physical activity (RRV~ACT~) RRV~SED~or RRV~ACT~is assessed by evaluating how hard a participant will work to obtain access to physical versus sedentary activities \[[@B1]\]. The child first sampled each of the four physical activities and four sedentary behaviors for at least two minutes and then rated them on a scale from 1-7. Children were asked to rank the activities, and the highest rated physical activity and sedentary behavior were chosen for the task. The physical activity alternatives included a balance board, a stationary youth mountain bike, a stepper, and a skipping game, while the sedentary alternative included magazines, puzzles, movies, and Playstation™ 2 video games. The child was instructed how to use the computer-generated task to earn points toward their favorite physical or sedentary activity. The computer displayed three squares where shapes rotated and changed color within each square every time a mouse button was pressed. When all of the shapes matched, the participant earned one point. The child worked on one of two computer monitors, one monitor had the physical activity alternative the other had the sedentary alternative. The reinforcement schedules for both components were initially set at FR4 (fixed ratio 4, which means the participant will earn one point after 4 responses). The schedule increased on a progressive ratio schedule that doubled after 5-points were earned on each schedule. (FR4, FR8, FR16, FR32, FR64, FR128, FR256, FR512, FR 1024 and FR2048). For every five points earned, the participant would receive 2-minutes of time to engage in the activity for which they were playing. The child was able to end the session at any time, they were instructed to tell the experimenter they were all finished when they did not want to earn points any longer. The computer recorded the participants\' points earned throughout the session. After instructions were given, the experimenter left the room. RRV~SED~and RRV~ACT~were quantified by the OMAX~SED~and the OMAX~ACT~, which is the maximal amount of responding at the highest reinforcement schedule completed. An intercom and a video camera were in the room so that the experimenter could hear and see into the experimental room from an adjoining room. ### Measurement of television, video and computer game playing at home Television, VCR/DVD, video game playing, and computer use was measured using the TV Allowance™. The device has a memory which recorded the amount of time that the targeted child and each family member used since the unit was installed. The device has been used in ongoing research in our laboratory, and was used as an important component of a previous study that successfully reduced television watching to prevent the development of obesity in youth \[[@B13],[@B19]\]. At baseline, unlimited TV and computer hours were set on each device, so that study staff could access the total number of hours for television and computer use for each family member. During the 25 and 50% decrease reduction phase the TV Allowance™ was programmed for the sedentary budget for that phase based on baseline amounts. In addition to the TV Allowance™, self-monitoring of sedentary behavior was recorded in a daily habit book which assessed reading, homework and use of hand-held computer games and targeted sedentary behaviors that cannot be quantified in this objective way. Recording was part of the intervention methodology to facilitate the child meeting their behavioral goals during the reduction phases. ### Physical activity The objective measure of physical activity was the Actigraph™ activity monitor, a small, unobtrusive unidirectional accelerometer with extensive validation in youth as a measure of physical activity \[[@B20]-[@B23]\]. The activity monitor was set to record minute by minute measures of physical activity. The activity monitor was worn during school and non-school waking hours on three days (two weekdays and one weekend day) during the last week of each three week period. If a child did not wear their activity monitor on the scheduled day, a day similar to the missed day was rescheduled. Weekly diaries were used in combination with the activity monitor to indicate what physically active and targeted sedentary behaviors the child was engaging in during the last week of each phase of the experiment. The activity monitors were downloaded to a computer at each home visit and weekly diaries were reviewed during the weekly home visit with the family. To determine levels of physical activity based on energy expenditure (METS or metabolic equivalents) we individually calibrated each accelerometer based on a progressive treadmill test. The VO~2~(mL/kg/min) and accelerometer counts/minute were sampled each minute, and the accelerometer values were regressed against VO~2~values to estimate energy expenditure for different intensities of physical activity. Based on the regression line, rates of accelerometer counts were determined for each participant that corresponded to rest (0 counts), 2 METS (7 mL/kg/min), 3 METS (MVPA, 10.5 mL/kg/min) and 6 METS (VPA, 21 mL/kg/min). Because the accelerometer was only worn for waking hours, and did not include time spent sleeping, we estimated energy expenditure for the remainder of non-accelerometer sampled minutes as expending 1 MET per minute, and computed calories per minute using estimated resting metabolic rate (RMR) using previously published equations for children \[[@B24]\]. Daily RMR calories were converted to RMR calories per minute, then multiplied by number of minutes not sampled by the accelerometer. Energy expenditure estimates while the accelerometer was being worn included RMR component of energy expenditure. Based on the estimated total daily energy expenditure, we estimated energy intake and changes in energy balance by considering total daily energy expenditure in respect to weight change. If weight was stable over the nine weeks, it was assumed that energy intake = energy expenditure. If children lost weight, it was assumed that one pound of weight loss was equivalent to negative energy balance of 3500 kcals, or 55.6 kcal/day. Similarly, a gain of one pound over the nine weeks would be equivalent to positive energy balance of 55.6 kcal/day. Based on the estimated energy expenditure and observed weight change, estimated energy intake and changes in energy balance were calculated. Analytic Plan ------------- Repeated measures analysis of variance was used to assess whether changes in targeted sedentary behavior were established to ensure that the targeted behavior had been manipulated by the intervention. Repeated measures analysis of variance was also used to assess changes in body weight, and physical activity. Pearson product moment correlation coefficients were used to assess predictors of weight change, as well as the relationship between changes in targeted sedentary behaviors and physical activity, as well as the relationship between RRV~ACT~and RRV~SED~, RRV~ACT~and measured physical activity, and RRV~SED~and total sedentary behavior. Significant factors were then studied in regression models controlling for age, sex and minority status. Significant relationships between weight change and RRV~SED~were explored by median splits dividing children into those who decreased or maintained weight (N = 20) versus increased (N = 41) their weight over the nine weeks of observation. Results ======= The average child was 10.7 ± 1.2 years of age, with a height of 57.4 ± 3.7 in, weight of 118.8 ± 30.5 lbs, BMI of 25.0 ± 4.2, and zBMI of 1.8 ± 0.4. Twenty seven (48.2%) of the children were male, and 14 (25%) were non-Caucasian or minority (Table [1](#T1){ref-type="table"}). Changes in weight, targeted sedentary behaviors and physical activity are shown across the three phases in Table [2](#T2){ref-type="table"}. The average child had a reduction in targeted sedentary behavior from baseline to 25% and 50% reduction phases (F(2, 110) = 285.00, p \< 0.0001), with significant changes from baseline to 25% (p \< 0.001), with a further significant reduction from 25% to 50% (p \< 0.001). There was a reduction in targeted sedentary behavior of 67.6%, with the majority of change (53%) occurring during the initial reduction phase. The changes in sedentary behavior included significant reductions in television watching (F(2,110) = 197.00, p \< 0.001) and computer use (F(2,110) = 40.07, p \< 0.001). Small, but significant increases in body weight were observed (F(2,110) = 4.84, p \< 0.001), with significant changes from baseline to 50% (p = 0.003) phases, but no differences between the 25% and 50% phases (p = 0.20). There was little change in physical activity accelerometer counts (F(2,110) = 0.49, p = 0.61) over phases. In addition, there were no significant changes in average METS (F(2,94) = 0.60, p = 0.55), or in the percentage of time below 2 METS (F(2,94) = 1.61, p = 0.21), above 2 METS (F(2,94) = 1.61, p = 0.21). Amount of time above 3 METS showed a significant decrease over phases (F(2,94) = 4.78, p = 0.01), with decreases from baseline to the 25% (p = 0.03) and 50% (p = 0.02) reduction phases. Estimated energy intake did not significantly change over phases (F(2,110) = 0.49, p = 0.61). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Characteristics of the sample (N = 56) ::: Variable \% ----------------- -------------- Child sex (M/F) 48.2 (27/56) \% Minority 25.0 (14/56) Mean (SD) Age 10.7 ± 1.2 Height (in) 57.4 ± 3.7 Weight (lb) 118.8 ± 30.5 BMI (kg/m^2^) 25.0 ± 4.2 zBMI 1.8 ± 0.4 ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Behavior and weight changes during the baseline, 25 and 50% reduction phases (Mean ± SEM) ::: Sedentary Behavior Reductions Phases --------------------------------- -------------------------------------- --------------- --------------- --------------- **Variable** **0** **25%** **50%** **p** Sedentary behavior (hrs/week) 34.9 ± 1.7 16.4 ± 1.0 11.3 ± 0.7 0 \> 25 \> 50  Television watching (hrs/week) 29.4 ± 1.7 13.9 ± 0.9 9.5 ± 0.6 0 \> 25 \> 50  Computer games (hrs/week) 5.5 ± 0.8 2.5 ± 0.4 1.8 ± 0.3 0 \> 25 \> 50 Weight (lbs.) 118.8 ± 4.1 119.2 ± 4.1 119.6 ± 4.1 0 \< 50 Physical activity (counts/min) 576.9 ± 23.8 572.4 ± 25.6 555.7 ± 24.3 NS Average METs 1.60 ± 0.05 1.58 ± 0.04 1.60 ± 0.05 NS  % time below 2 METS 83.3 ± 2.4 84.6 ± 2.4 84.5 ± 2.4 NS  % time above 2 METS 16.7 ± 2.4 15.4 ± 2.4 15.5 ± 2.4 NS  % time above 3 METS 4.2 ± 0.7 3.5 ± 0.6 3.5 ± 0.6 0 \> 25,50 Estimated energy intake 2106.4 ± 65.7 2173.5 ± 65.2 2144.3 ± 74.6 NS ::: Variables related to weight change included only OMAX~SED~(r = 0.31, p = 0.022). Child age (p = .71), sex (p = .64), minority status (p = .67), income (p = .38), reinforcing value of physical activity (p = .72) or baseline values of weight (p = .45), changes in targeted sedentary behavior (p = .44) or changes in physical activity (p = .15) were not related to weight change. Multiple regression controlling for child age, sex and minority status did not reduce the impact of RRV~SED~on weight change (p = 0.035). Differences in the motivation to be active or sedentary, and changes in sedentary and active behaviors, estimated energy intake and body weight for children who gained (N = 39) or lost (N = 19) weight are shown in Table [3](#T3){ref-type="table"}. There was a significant difference in OMAX~SED~between children who lost or maintained versus those who gained weight during the study (F(1,54 = 4.79, p = 0.03). Figure [1](#F1){ref-type="fig"} shows differences in OMAX (left graphs) and the pattern of responding (right graphs), while the top and bottom graphs show motivated responding for sedentary behaviors or physical activity, respectively. As shown in Figure [1](#F1){ref-type="fig"}, children who maintained or lost weight had lower OMAX~SED~and responding over progressive ratio schedules for access to sedentary behaviors than children who gained weight, who worked much harder for sedentary behaviors. There were no differences in OMAX~ACT~(F(1,54) = 0.11, p = 0.74) as a function of whether children lost or maintained versus gained weight during the study. As shown in Table [3](#T3){ref-type="table"}, there were no differences in the alterations in any activity variable for children who gained or lost weight over the nine weeks of the study. Children who lost weight reduced energy intake by an estimated 223 kcal/day calories when sedentary behaviors were reduced, while children who gained weight when sedentary behaviors were reduced increased their estimated energy intake by 172 kcal/day (F(1,54) = 13.13, p = 0.0006). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Differences in behavior for children who lost or gained weight after reduction of sedentary behaviors (Mean ± SEM) ::: Weight change groups --------------------------------- ---------------------- --------------------- ---------- **Variable** **Lost (N = 19)** **Gained (N = 37)** **p** OMAX~SED~ 540.4 ± 143.5 1189.0 ± 198.4 .03 OMAX~ACT~ 258.9 ± 89.4 226.6 ± 52.3 .74 Sedentary behavior (hrs/week) -26.3 ± 2.0 -22.2 ± 1.6 .13  Television watching (hrs/week) -22.8 ± 2.1 -18.4 ± 1.6 .12  Computer games (hrs/week) -3.5 ± 0.8 -3.8 ± 0.8 .83 Weight (lbs.) -1.2 ± 0.2 1.8 ± 0.2 \< .0001 Physical activity (counts/min) 20.6 ± 37.4 -42.5 ± 30.6 .22 Average METs -0.02 ± 0.03 -.01 ± 0.03 .78  % time below 2 METS 1.45 ± 1.42 1.10 ± 1.09 .85  % time above 2 METS -1.45 ± 1.42 -1.10 ± 1.09 .85  % time above 3 METS -0.81 ± 0.6 -0.78 ± 0.4 .96 Estimated energy intake (kcal) -223.2 ± 105.4 172.0 ± 56.6 .0006 ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Differences (mean ± SEM) in the OMAX (left graphs) and reinforcing value (right graphs) of sedentary behaviors (top graphs) or physical activity (bottom graphs) for children who lost or maintained weight versus gained weight during the 9 week study**. :::  ::: OMAX~ACT~was not significantly related to OMAX~SED~(r = 0.07, p = 0.63). Similarly, OMAX~ACT~was not correlated with activity counts at baseline (r = 0.12, p = 0.38), and OMAX~SED~was not correlated with total sedentary behavior at baseline (r = 0.07, p = 0.63). Discussion ========== The results show that sedentary behavior was successfully manipulated by over 50%, and that children made the majority of the changes during the 25% reduction phase. These changes were associated with no significant increases in physical activity. We have previously observed minimal changes in physical activity when sedentary behaviors were reduced \[[@B25]\]. Weight change was not associated with changes in physical activity when sedentary behaviors are reduced, suggesting that for the average overweight child, reducing sedentary behaviors does not result in greater physical activity and weight loss. This is consistent with cross-sectional research arguing that the effects of sedentary behavior on body weight are not due to changes in activity energy expenditure \[[@B26]\]. In other research we have shown that increasing sedentary behavior results in a reduction in physical activity, suggesting that the relationship between sedentary behavior and physical activity is not symmetric, and the association between these behaviors may only be present in one direction \[[@B25]\]. The variable that was associated with weight change over time was the motivation to be sedentary, which represents the reinforcing value of sedentary behaviors such as television watching, watching videos and playing on the computer. These are activities for which obese children allocate a great deal of their time, and time being sedentary may be independent of time being active \[[@B27],[@B28]\], as the motivation to be sedentary is independent of the motivation to be active. If they were direct substitutes for each other, then they would be significantly negatively correlated, and reducing sedentary behaviors would result in an increase in physical activity, which is generally not observed. In the current study the motivation to be active and the motivation to be sedentary were not significantly related. It is important that only a subset of sedentary behaviors was targeted, to allow the child to choose how to reallocate time that had been allocated for watching television or playing computer games. Since no increases in physical activity were observed, it is likely that children substituted other sedentary behaviors for the targeted sedentary behaviors. Unfortunately, we did not have children record all sedentary behaviors during the reduction phases, so it is not possible to know what sedentary behaviors they engaged in as substitutes. The question is how is the reinforcing value of sedentary behaviors related to weight change? Weight change is due to change in energy balance, which must be due to either reductions in energy intake or increases in energy expenditure. The absence of increases in physical activity reduces the likelihood that physical activity is a substitute for sedentary behaviors, such that children spontaneously increase their physical activity when sedentary behaviors are reduced. This suggests that changes in energy intake are the component of energy balance that is promoting the weight change. The estimated energy intake data showed that children who lost weight reduced their estimated daily energy intake by 223 kcal/day, while those who gained weight increased their estimated daily energy intake by 172 kcal/day. Other investigators have argued that while sedentary behaviors are correlated with weight, the relationship is not mediated by changes in measured physical activity, but are likely to be mediated by changes in eating and energy intake \[[@B26]\]. We have previously shown in non-overweight children that energy intake is a reliable complement to shifts in sedentary behavior \[[@B19]\], such that reductions in sedentary behavior paired with eating result in reductions in energy intake. This may be due to the strength of the relationship between eating and engaging in sedentary behaviors when children enter the study. While eating in association with sedentary behaviors is common, and experimental research has shown that increasing television watching increases energy intake \[[@B29]-[@B31]\], there is variability in this relationship. If a child never eats in association with television watching, then reducing television watching cannot result in a reduction in energy intake. On the other hand, if a child consumes food often in association with watching television, then reducing television watching may have a large effect on energy intake and body weight. The motivation to be sedentary is a behavioral phenotype that may lead to a better understanding of factors related to how changing sedentary behavior may relate to changes in energy balance behaviors and weight loss. Given the potential relationship between changes in television watching and energy intake, it is also possible that the motivation to eat is an important factor to predict how reducing sedentary behaviors influences eating. For example, it may be that children who find food more reinforcing would have a harder time reducing food intake when sedentary behaviors are reduced, and they may compensate by increasing intake at other times, or they may resist changes to reduce television watching that is associated with eating, since this would reduce access to powerful reinforcers they want to obtain. This would be an interesting set of studies for future research. One surprising result was the failure to show that the motivation to be active was related to physical activity, or those high in the motivation to be active were more likely to become more active. We have shown in previous research that the motivation to be active is related to more physical activity \[[@B3],[@B4]\], but these studies included children with a wide variety of motivation to be active as well as a variety of levels of physical activity. In the present study of overweight and obese, sedentary children, the range of motivation to be active and of activity levels was constrained, which can lead to lower relationships if there is little variability in the predictor and/or outcome. There are limitations to this study in the measurement of activity and diet. While we had objective measurements of physical activity, we did not collect detailed self-reported information on what types of behaviors people engaged in and what types of behaviors people used to substitute for reduced targeted sedentary behaviors. It would have been interesting to know if specific classes of active or sedentary behaviors were changed. For example, recent data suggests that standing rather than sitting may confer health benefits in adults \[[@B32]\], and it would be interesting to know if children stood more or engaged in light physical activity to replace sedentary behaviors and if these changes would be associated with weight loss or improvement in health. Likewise, it would have been interesting to know if other popular sedentary behaviors, such as talking on the phone or texting, replaced television watching or computer game playing. The TV Allowance™ is a useful behavioral engineering approach to reducing television watching, however, a limitation of using the device is that it is possible to overestimate television watching since children may turn on the television and become engaged in an alternative activity and not watch it. This should be minimized since children are reinforced for reducing television watching, but there may be instances in which television watching is overestimated, and the degree of reduction in television watching is underestimated. Physical activity was measured for three days during each phase, and it may have been useful to collect more extended samples of physical activity during each phase \[[@B23]\]. The results point to changes in energy intake, rather than physical activity, as the mechanism for changes in body weight as people reduce their television watching or computer game playing. Estimates of energy intake were consistent with this hypothesis, and accelerometer based activity counts can provide valid information about energy expenditure \[[@B20],[@B33]\]. In addition, self-reports of energy intake are notoriously inaccurate \[[@B34]\], and previous studies using the current methods for reducing sedentary behavior have shown consistent underreporting of energy intake in obese children and adolescents \[[@B19]\]. Despite the challenges in collecting valid dietary intake data, it would be useful to have dietary information that includes dietary intake as well as macronutrient intake. In summary, the present study replicates previous research that suggests that reducing sedentary behaviors is not associated with an increase in physical activity. The motivation to be sedentary is related to short term weight change when sedentary behaviors are reduced, and this effect may be mediated by changes in energy intake. Thus, one predictor of the effectiveness of programs to reduce sedentary behavior for child weight change may be the motivation to be sedentary. Competing interests =================== Dr. Epstein is a consultant to Kraft foods and NuVal. The other authors do not have any potential conflict of interests. Authors\' contributions ======================= LHE and JNR designed the study, and LHE obtained the research funding. MDC obtained IRB approval, and supervised study implementation and data collection. LHE and RAP conducted data analysis. LHE wrote the initial draft of the manuscript, and all authors contributed to the interpretation of data and the writing of the manuscript. All authors read and approved the final manuscript. Acknowledgements ================ Appreciation is expressed to Dana D. Winicwicz and Barbara F. Jedrzejck for assisting in the implementation of protocol. This research was funded in part by a grant from the National Institute of Child Health and Human Development, R01 HD39778 awarded to Dr. Epstein.

PubMed Central

2024-06-05T04:04:19.431819

2011-2-22

PMC3053213

Background ========== Cardiovascular disease is the leading cause of death in western countries and a major cause of morbidity and mortality world-wide. Studying coronary atherosclerotic disease (CAD) is challenging for several reasons, since it has substantial environmental and genetic components. Furthermore, despite the nearly universal presence of coronary atherosclerosis, particularly as individuals age, cardiovascular events such as acute coronary syndromes, sudden death and the need for revascularization therapy only occur in some individuals, highlighting the difficulty in precisely defining atherosclerosis phenotypes. In symptomatic patients, revascularization therapy is often required, and percutaneous intervention with balloon angioplasty and stent implantation is a cornerstone of therapy. In-stent restenosis (ISR) is a late complication of stent implantation in which inflammatory and proliferative responses to the vascular injury caused by angioplasty and stenting lead to neointimal hyperplasia within the stent and at its edges over the following weeks and months. Many of the same inflammatory and proliferative processes are activated in the development of atherosclerosis but occur over years or decades. ISR is characterized by proliferative responses to the vascular wound incurred as a result of stent implantation\[[@B1]\]. Therefore, ISR may be viewed as a model phenotype of vascular wound repair for which the mechanisms represent part of the pathologic picture of atherosclerosis, with relatively accelerated wound repair responses operative in the vascular wall and in peripheral blood leukocytes. In the analysis reported here, we apply a novel method to analyze time-course gene expression data to gene expression profiles of peripheral blood mononuclear cells (PBMCs) of patients enrolled in our study of ISR. The results of the discovery transcriptome analysis of the CardioGene Study were tested for replication in an independent sample of Icelandic patients. We ultimately identified and validated a set of 32 genes that are temporally differentially expressed in the blood of patients who develop ISR after stenting, compared to those who do not develop ISR, highlighting the importance of cellular growth pathways and identifying several biologic candidates for further mechanistic investigation. Results ======= Microarray data quality control and filtering --------------------------------------------- In the CardioGene Study, 312 patients were included (52 with ISR, and 260 who did not develop ISR, as defined by criteria for clinical restenosis\[[@B2]\]) after quality control filtering (Table [1](#T1){ref-type="table"}). All 312 patients had baseline gene expression profile data (Figure [1a](#F1){ref-type="fig"}). Of these, 203 had a blood sample and high quality gene expression profile data at early follow-up (2-4 weeks post-stent), and 166 had high quality gene expression profile data at both early follow-up and late follow-up (6 months post-stent). A total of 681 samples were included in the time course analysis. Box plots showing the distributions of the early and late follow-up times are presented in Figure [1b](#F1){ref-type="fig"}. From the deCODE sample of patients in Iceland, 97 patients were enrolled and had high quality gene expression profiles at baseline (28 with ISR, 69 without ISR). Of these, 86 patients had a follow-up blood sample and acceptable gene expression data at 6 months post-stent. Thus, a total of 183 samples were included in the replication analysis. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Clinical characteristics of the CardioGene and deCODE cohorts ::: CardioGene N = 312 deCODE N = 97 ------------------------------------------------------------- --------------------- ---------------------- Number of female (% female) 100 (32.0%) 19 (19.6%) Age (mean ± sd) 65.6 (10.4) 64.8 +/- 10.0 Diabetes 91 (29.2%) 12 (12.5%) Hypertension 213 (68.2) 58 (64%) Ever smoked 194 (62.1%) 23 (48%) Hyperlipidemia 241 (77.2%) 47 (48.4%) Prior ISR (N, %) 10 (4.3%) \-\-\-- ISR 52 (16.7%) 28 (28.9%) Acute coronary syndrome at time of index PCI 64 (20.5%) \-\-\-- Reference diameter (mean ± sd) 2.74 mm +/- 0.74 mm 2.71 mm +/- 0.050 mm Lesion length (mean ± sd) 9.4 mm +/- 5.3 mm 12.7 mm +/- 8.14 mm LAD location (% with stent in LAD) 126 (40.6%) 52 (53.6%) Post-stent % stenosis in patients with ISR 10.96 mm +/- 6.2 mm 6.0 mm +/- 12.7 mm Number of timepoints at which PBMC gene expression measured 3 2 ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **The CardioGene Study Design and intervals for blood sampling**. (a) Times of blood sampling and clinical follow-up. Blood draws occurred at times T0 (pre-stent), T1 (2-4 weeks post-stent) and T2 (six months post-stent) (b) Distribution of the intervals for blood sampling from baseline to early follow-up (2-4 weeks post-stent) and late follow-up (6-12 months) time points, are shown for the CardioGene Study. :::  ::: CardioGene discovery analysis of time-course RNA expression in PBMCs -------------------------------------------------------------------- Applying the time varying intercept model, 46 probes, corresponding to 42 distinct genes, were found to be significant at FDR adjusted q-value less than 0.05 out of 12467 probe sets that passed the quality control (Additional File [1](#S1){ref-type="supplementary-material"}). Two patterns of gene expression were observed, with some genes showing divergent gene expression patterns across the time course (Figure [2a](#F2){ref-type="fig"}) and others showing consistently differential gene expression over the time course (Figure [2b](#F2){ref-type="fig"}). Descriptions of the genes to which these probe sets map are provided in Additional File [1](#S1){ref-type="supplementary-material"}. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Patterns of time course results for three probe sets in the CardioGene and deCODE studies**. The color of the lines plotted is red for ISR, blue for non-ISR, and black for the model fitted to all patients, under the null hypothesis. Probeset 201222\_s\_at (a) shows a divergent pattern of gene expression between individuals who develop ISR and those who do not, whereas 203605\_at (b) shows consistently differential expression across the timecourse. Gene expression for probe set 218047\_at shows similar gene expression patterns in both the CardioGene (c) and deCODE (d) cohorts. :::  ::: Replication analysis using deCODE time course samples ----------------------------------------------------- Among 46 significant probe sets identified in the CardioGene time-course analysis, 36 probe sets had FDR adjusted q-value less than 0.05 in the deCODE replication analysis (Figure [3](#F3){ref-type="fig"}). Gene expression patterns were largely consistent with the patterns observed in the CardioGene discovery analysis (Figure [2c](#F2){ref-type="fig"} and [2d](#F2){ref-type="fig"}). The 36 probe sets mapped to 32 unique genes (Table [2](#T2){ref-type="table"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Genes identified in the CardioGene discovery analysis and replicated in the deCODE analysis**. A Venn diagram shows the genes identified in the CardioGene discovery analysis and those replicated in the deCODE analysis. :::  ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Validated genes x(32 genes, 36 probesets) with annotation information and gene expression patterns, as shown in Figure 2 ::: Affymetrix probesets Gene Symbol Gene Name Gene expression pattern ---------------------- ------------- -------------------------------------------------------------- ---------------------------- 202502\_at ACADM acyl-Coenzyme A dehydrogenase, C-4 to C-12 straight chain Consistent (ISR \> No-ISR) 221504\_s\_at ATP6V1H ATPase, H+ transporting, lysosomal 50/57 kDa, V1 subunit H Consistent (ISR \> No-ISR) 202654\_x\_at AXOT axotrophin Consistent (ISR \> No-ISR) 212460\_at C14orf147 chromosome 14 open reading frame 147 Divergent (ISR \> No-ISR) 203994\_s\_at C21orf2 chromosome 21 open reading frame 2 Consistent (No-ISR \> ISR) 203996\_s\_at C21orf2 chromosome 21 open reading frame 2 Consistent (No-ISR \> ISR) 208374\_s\_at CAPZA1 capping protein (actin filament) muscle Z-line, alpha 1 Consistent (ISR \> No-ISR) 217886\_at EPS15 epidermal growth factor receptor pathway substrate 15 Consistent (ISR \> No-ISR) 218646\_at FLJ20534 hypothetical protein FLJ20534 Consistent (ISR \> No-ISR) 204829\_s\_at FOLR2 folate receptor 2 (fetal) Consistent (ISR \> No-ISR) 220131\_at FXYD7 FXYD domain containing ion transport regulator 7 Consistent (No-ISR \> ISR) 217814\_at GK001 GK001 protein Consistent (ISR \> No-ISR) 218092\_s\_at HRB HIV-1 Rev binding protein Consistent (ISR \> No-ISR) 212411\_at IMP4 U3 snoRNP protein 4 homolog Divergent (No-ISR \> ISR) 200821\_at LAMP2 lysosomal-associated membrane protein 2 Consistent (ISR \> No-ISR) 203276\_at LMNB1 lamin B1 Consistent (ISR \> No-ISR) 214773\_x\_at MGC3794 putative MAPK activating protein Consistent (ISR \> No-ISR) 212803\_at NAB2 NGFI-A binding protein 2 (EGR1 binding protein 2) Consistent (No-ISR \> ISR) 203371\_s\_at NDUFB3 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 3, 12 kDa Consistent (ISR \> No-ISR) 218047\_at OSBPL9 oxysterol binding protein-like 9 Consistent (ISR \> No-ISR) 221123\_x\_at PBF papillomavirus regulatory factor PRF-1 Consistent (No-ISR \> ISR) 208857\_s\_at PCMT1 protein-L-isoaspartate (D-aspartate) O-methyltransferase Consistent (ISR \> No-ISR) 205202\_at PCMT1 protein-L-isoaspartate (D-aspartate) O-methyltransferase Consistent (ISR \> No-ISR) 209345\_s\_at PI4KII phosphatidylinositol 4-kinase type II Consistent (No-ISR \> ISR) 201222\_s\_at RAD23B RAD23 homolog B (S. cerevisiae) Divergent (ISR \> No-ISR) 201223\_s\_at RAD23B RAD23 homolog B (S. cerevisiae) Consistent (ISR \> No-ISR) 209207\_s\_at SEC22L1 SEC22 vesicle trafficking protein-like 1 (S. cerevisiae) Consistent (ISR \> No-ISR) 217758\_s\_at SMBP SM-11044 binding protein Consistent (ISR \> No-ISR) 203605\_at SRP54 signal recognition particle 54 kDa Consistent (ISR \> No-ISR) 212513\_s\_at USP33 ubiquitin specific protease 33 Consistent (ISR \> No-ISR) 201807\_at VPS26 vacuolar protein sorting 26 (yeast) Consistent (ISR \> No-ISR) 210849\_s\_at VPS41 vacuolar protein sorting 41 (yeast) Divergent (ISR \> No-ISR) 205126\_at VRK2 vaccinia related kinase 2 Divergent (ISR \> No-ISR) 216304\_x\_at YME1L1 YME1-like 1 (S. cerevisiae) Consistent (ISR \> No-ISR) 201351\_s\_at YME1L1 YME1-like 1 (S. cerevisiae) Consistent (ISR \> No-ISR) 222294\_s\_at RAB27A Consistent (ISR \> No-ISR) ::: Annotation of Genes Identified ------------------------------ Using the DAVID/EASE annotation tool, several categories of genes were identified, falling into multiple gene ontology classifications. No single category achieved a significance score indicating statistically significant over-representation of a specific gene ontology, and several gene ontology classes showed multiple genes (Additional File [2](#S2){ref-type="supplementary-material"}). The categories with the highest representation (i.e. the largest number of genes mapping to these ontologies) included cell growth and/or maintenance, cellular metabolism, catalytic activity, nucleotide and nucleic acid metabolism, transport, protein binding and cellular binding. Genes with prior evidence of expression in the arterial wall after vascular injury included the *NAB2*and *LAMP2*genes, with data in the prior literature suggestive of functional relevance of these genes, by mediating proliferative responses and recruitment of PBMCs to injured endothelium, respectively\[[@B3]-[@B13]\] (Figure [4](#F4){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Gene expression for the top eight genes, *NAB2*and *LAMP2***. Gene expression plots are shown for the top eight ranked genes, and *NAB2*and *LAMP2*genes are shown together, as these genes have known functions relevant to ISR. Gene expression data is shown for both cohorts across the time course of samples in each study. For each gene, expression levels are shown on the y-axis as the log 2 of normalized gene expression (GE), and the timepoints shown on the x-axis of each plot correspond to baseline (BL), early follow-up (EF), and late follow-up (LF) timepoints, although data was analyzed across the continuous timeframe of follow-up blood sampling, rather than categorically as is depicted here for illustration purposes. The box plots are plotted with the whiskers at 1.5 times the interquartile range. :::  ::: Sensitivity Analysis -------------------- In a sensitivity analysis, we found that the median number of significant probe sets using the same criteria of FDR 0.05 was only 2, the mean was 3.04, and the maximum number was 14 in one iteration, demonstrating a significantly low likelihood that the probesets we identified and replicated were identified due to chance. Hypothesis-driven analysis of gene expression --------------------------------------------- We compared our results to published studies examining gene expression in ISR, chiefly in human vascular tissue samples \[[@B14],[@B15]\]. In one study, peripheral blood total leukocytes were also examined \[[@B14]\]. In the CardioGene baseline blood sample, genes reported by Ashley et al were examined \[[@B15]\], and none were significantly differentially expressed in the baseline blood sample or in our timecourse results. In another study by Zohlnhofer et al \[[@B14]\], gene expression data from neointimal vascular tissue and peripheral blood leukocytes were reported. We examined those genes meeting a Bonferroni multiple testing correction from this discovery report and did not find overlap with our gene expression findings. Discussion ========== Using a repeated sampling study design, we investigated PBMC transcript expression in patients undergoing stent implantation, using a novel time-course analysis method. We identified a set of 42 genes with differential temporal expression among patients with and without ISR at one year follow-up in a discovery analysis of the CardioGene Study. Independent replication testing in an Icelandic sample confirmed differential expression of 36 probesets mapped to 32 genes. The gene expression patterns over time may be of interest as well, with consistently expressed genes representing gene expression data that may be able to predict ISR and differentially expressed genes over the time course representing genes with possible direct functional roles in the development of ISR, both of which require further investigation to explore more fully. Gene expression profiling with DNA microarray technology is a popular tool to monitor the expression level of thousands of genes simultaneously and has been applied in cardiovascular research, to detect patterns of gene expression indicative of underlying disease states\[[@B16]-[@B21]\]. Since the data generated represent the temporal abundance of mRNA levels in the sample, measurements of the change of this abundance over the course of disease progression (or any biological process) is therefore both possible and of great scientific interest using this technology. In fact, Yuan and Kendziorski\[[@B22]\] reported that more than one-third of the experiments catalogued in the Gene Expression Omnibus (National Center for Biotechnology Information; <http://www.ncbi.nlm.nih.gov/geo/>) are from experiments that measure gene expression over time. Early time-course RNA expression studies have focused on identifying clusters of genes with a similar pattern over time\[[@B14],[@B23],[@B24]\]. More recently, detecting a differential gene expression pattern over time between several biological groups has become an interesting goal of time-course gene expression data. To detect differential gene expression pattern over time, we used a time-varying intercept model which can account for differences in sample intervals between patients. The term \"differential gene expression pattern over time\" can be interpreted in several ways to form suitable questions and, thus, the hypotheses are dependent on the particular experiment under consideration. We considered two related questions where each can roughly correspond to the main effect of the group, and the interaction between group and time points. Consider, for example, a gene with exactly the same expression pattern over time in both groups, but the first group has a higher expression than the second group, consistently over all time points. This is a gene that shows the main effect of the group. On the other hand, consider another gene with similar expression level at two time points, 1 and 2, but this gene\'s expression increases from time 1 to 2 in one group but decreases in another group. This group-by-time interaction cannot be examined by methods that only test the main effect of the group. The time-varying intercept model we used can detect both the main effect and group-by-time interaction. This method, however, requires a large number of bootstrap resampling to evaluate the significance level of the difference between groups, which can be computationally challenging especially when a large number of genes are tested. Of the genes we identified, the most extensive prior literature in vascular disease was found for the *NAB2*gene, which is also known as EGR1 binding protein 2. Early growth response (EGR) genes, which are transcription factors that are implicated in a wide variety of proliferative and differentiative processes\[[@B15]\]. Nab proteins are necessary for Schwann cells to exit the cell cycle and generate a myelin-specific gene profile and are key regulators of the myelination process of peripheral nerves\[[@B25]\]. *NAB2*is expressed in vascular smooth muscle cells in response to injury\[[@B3]-[@B12],[@B26]\] and EGR1 has been identified in a microarray study of in vitro smooth muscle cell proliferation\[[@B27]\]. The *LAMP2*gene product protects the lysosomal membrane from proteolytic enzymes within lysosomes and also functions as a receptor for proteins to be imported into lysosomes\[[@B28]\]. Mutations in the *LAMP2*gene have been identified in patients with hypertrophic cardiomyopathy\[[@B29]\], and the gene product mediates adherence of PBMCs to the vascular endothelium\[[@B13]\]. Cellular adhesion to the vascular endothelium has been well-described in animal models and post-mortem human examinations, in atherogenesis and acute vascular injury\[[@B30],[@B31]\]. In the latter, the extent of leukocyte adhesion is predictive of the degree of subsequent neointimal hyperplasia, which is the key lesion of ISR. Some genes identified have no apparent role in vascular biology, such as *VPS26, VPS41*, *SRP54*, and *RAD23B*, and comparison to previously published reports in the literature do not show differential gene expression in other studies of restenosis, although these investigations were conducted primarily on vascular tissues or culture vascular smooth muscle cells rather than peripheral blood\[[@B17],[@B18],[@B27],[@B32]\]. The *VPS26 and VPS4*genes belongs to a group of vacuolar protein sorting genes and may have a role in lysosome maintenance\[[@B33]\], and *SRP54*is a protein in the signal recognition particle, which directs secretory proteins to membranes as they emerge from the ribosome \[[@B34]\]. Specific vascular or inflammatory cell function has not been described, but derangement of these basic cellular processes may impact vascular and other physiologic functions adversely. *RAD23B*has a role in DNA (nucleotide excision) repair, and genetic variants in *RAD23B*have been associated with several solid tumors\[[@B35]-[@B38]\]. The association with cancer would suggest a possible link to excess proliferative mechanisms in vascular wound repair, as has been described for many other cell cycle regulatory genes\[[@B39]\]. Genetic variants in other genes we identified are also associated with human diseases. Several genetic variants in *ACADM*have been associated with medium-chain acyl-CoA dehydrogenase activity, but there is no known vascular implication of this disorder\[[@B40]\]. Variants in *PCMT1*and *FOLR2*have been associated with neural tube defects\[[@B41]-[@B43]\], with no known vascular phenotype in these cases. Overall the findings of this study are hypothesis-generating and can be used to support the rational for investigating the function of specific genes and pathways in adverse vascular remodeling, which is relevant to both ISR and more general CAD phenotypes. We compared the results of our study to previously published reports of transcriptome analysis in ISR. Our results were negative for replication of these studies which focused primarily on vascular tissue samples, in relatively small sample sets. In one study, peripheral blood total leukocyte gene expression was studied in 10 patients with ISR and atherectomy specimens \[[@B23]\]. While a high degree of correlation between peripheral blood leukocyte and arterial neointima tissue gene expression was identified in a subset of genes, these findings were based upon single measurements in a small sample size and were not replicated in the original report. These prior reports highlight the major difficulty of studying vascular tissues, since access to these tissues in adequately large sample sizes is limited. Vascular biopsy and atherectomy are performed infrequently as part of routine clinical care and would not support well-powered studies of vascular tissues. Tissue sampling over a time course is not clinically indicated or possible. Additionally, a large degree of intra-individual variability in gene expression was noted in these prior studies of vascular tissues, making replication testing critical, yet this cannot be done without access to additional tissues samples. In our study, we analyzed peripheral blood leukocytes, specifically focusing on the mononuclear fraction which contains primarily B and T lymphocytes and monocytes. Although the analysis of peripheral blood cells would ideally be complemented by similar studies in vascular tissues, studying gene expression profiles in blood leukocytes is biologically relevant due to well-defined interactions with the arterial wall, particularly in the setting of vascular injury and repair as in the setting of ISR\[[@B17]\]. The overlap between vascular and blood gene expression in one prior transcriptome analysis of ISR was supportive of our rationale to study PBMCs. For these reasons, and with the additional prior knowledge that inflammation plays a significant role in the development of ISR, we undertook a study of PBMCs in several hundred patients, with adequately powered replication testing for our top discovery findings. Additionally, we use a time course analysis method that improved our ability to detect gene expression signals between the two comparison groups, overcoming some of the difficulty of substantial variability in single point microarray gene expression data. To address the possibility of false positives identified with our statistical methods, we conducted replication analyses in the independent sample of deCODE samples and we conducted bootstrap resampling to assess significance of the findings. Through this sensitivity analysis, we demonstrated that the validation of 36 probe sets is not likely to be due to chance. Additional potential limitations of this study of ISR are the use of a clinical restenosis outcome, rather than an angiographic outcome, in which clinically silent ISR may have been missed, and the choice of tissue analyzed, as discussed. The CardioGene and deCODE cohorts differ in the incidence of ISR (16.7% in the CardioGene Study and 28.8% in deCODE) with the patients in the deCODE sample showing overall lower residual percent stenosis in the treated lesion after stent implantation. Also, the proportions with hyperlipidemia and diabetes differ. However, despite the differences in the cohorts, we find replication of a substantial proportion of the discovery findings. Conclusions =========== In summary, we have used a method to analyze gene expression in serial blood samples and identified a set of genes that show differential expression in the blood of patients who develop ISR after BMS implantation, compared to those who do not. These gene expression patterns of adverse vascular remodeling suggest possible hypotheses for the mechanisms of injury-induced remodeling observed in both ISR and CAD, since ISR is a niche phenotype occurring in a subset of patients with CAD. Further studies are needed to investigate the functional relevance of these genes and are warranted based upon the findings of this study. Methods ======= Study samples ------------- The CardioGene Study was an IRB-approved, prospective cohort study of 358 patients enrolled at the time of bare metal stent (BMS) implantation to treat de novo, previously untreated native coronary artery lesions at William Beaumont Hospital (Royal Oak, Michigan, USA) and the Mayo Clinic (Rochester, Minnesota, USA). Patients were followed for one year to determine ISR outcomes (Figure [1](#F1){ref-type="fig"}). Enrollment began in February 2002 and was closed in September 2003, prior to the approval and clinical use of drug-eluting stents (DES) in the United States. The protocol was approved by the NHLBI IRB as well as the IRB at each of the clinical enrollment sites. Informed consent was provided by each patient. Standardized case report forms were used to collect baseline clinical data and outcome information in follow-up\[[@B2]\]. Since gene expression discovery studies may suffer from false positives, despite the use of statistical corrections for multiple hypothesis testing, we sought out independent validation of the primary findings of the CardioGene Study in a separate cohort. For independent replication testing, 97 patients undergoing stent implantation with BMS were enrolled in Iceland by collaborators at Landispitali University Hospital and University of Iceland, and RNA samples were provided through a contract with deCODE Genetics, Inc. The study was approved by the National Bioethics Committee and the Data Protection Authority of Iceland, and each patient provided informed consent for participation in the study. Clinical characteristics of both study cohorts are summarized in Table [1](#T1){ref-type="table"}. Clinical phenotype ------------------ Consecutive patients presenting to the cardiac catheterization laboratories of the clinical enrollment sites were approached for participation in the study. Follow-up clinical evaluation was performed via patient interview and review of all available medical records at 6 months and 12 months post-stent (Figure [1](#F1){ref-type="fig"}). ISR was defined as clinical restenosis\[[@B2]\], which was defined by ischemic symptoms after stent implantation and evidence of flow limitation in the treated vessel by either invasive or non-invasive testing. Follow-up angiography was not specifically performed for the CardioGene Study. Any available angiographic data performed as part of each patient\'s clinical care was recorded. The Icelandic subset had follow-up angiography of all patients at 6 months. Quantitative coronary angiography (QCA) in the CardioGene Study was performed by a single QCA reader, and by a TIMI core QCA center for the Icelandic subset. PBMC Microarray Analysis ------------------------ Whole blood was sampled immediately prior to stenting, 2-4 weeks after stenting and six months post-stent. Blood samples were collected into EDTA-containing tubes at three times from patients enrolled in the CardioGene Study (Figure [1](#F1){ref-type="fig"}), and two time from patients enrolled in Iceland. Samples from each time point were handled according to the standardized blood handling and RNA isolation protocols\[[@B2]\], for consistency across time points and clinical enrollment sites. PBMCs were harvested by technicians at each clinical enrollment site following standardized protocols for Ficoll separation of whole blood and osmostic lysis of red blood cells, to avoid technical variation and to minimize globin mRNA contamination. All blood processing was completed within 4 hours of blood draw, at room temperature. Snap-frozen PBMCs samples were shipped to the NIH laboratory, and RNA isolation was performed using the Qiagen RNA isolation protocol, with DNase treatment. RNA labeling and microarray image analysis was according to the Affymetrix U133A protocol. To minimize the impact of batch-to-batch variation, the multiple samples from each patient were processed together through all stages, and the order of RNA labeling and microarray hybridization was randomized. The same methods were used in the CardioGene and deCODE study samples, and microarray analysis was conducted on all samples at the NIH. In both studies, technicians performing the assays were blinded to the clinical status of patients. Statistical Analysis -------------------- Raw image data were normalized using RMA (Robust multichip average) normalization\[[@B44]\]. Several steps of quality control were performed. These included the investigation of the residual plots using affyPLM, gender match using a linear discriminant analysis and shrunken centroid method\[[@B45]-[@B47]\]. Probes with values less than 6 in log 2 scale in more than two thirds of samples were excluded, and age and gender were adjusted before the analysis\[[@B48]\]. In the study design of the time course of blood gene expression profiling, the follow-up times were set at 2 weeks and 6 months for the early and late follow-up. A window was set at each of the time points, with the early follow-up designated as any time 2-4 weeks post-stent and the late follow-up designated as any time between 5-7 months for patients seen at 6 months. If patients did not return for follow-up at the 6 month time point, an attempt was made to have the patient provide the late follow-up blood sample at 12 months post-stent, at the time when final clinical ascertainment for ISR was made. As a result of the use of time windows, we were able to increase ascertainment of follow-up blood samples, but the actual sample collection intervals after stent implantation were not precisely spaced. To analyze such data, we used the time varying intercept model using a B-spline basis of dimension 2 based on only one knot at the median follow-up time (which was 14 days) (Additional File [3](#S3){ref-type="supplementary-material"})\[[@B49]\]. This is because the majority of the early follow-up is centered around 14 days. The model was fit under the null hypothesis using combined samples of ISR and non-ISR patients, and under the alternative hypothesis where separate models were fitted for ISR and non-ISR patients, respectively. The test statistic is the improvement in the fitted model under the alternative hypothesis from that under the null hypothesis, as measured by the relative difference in the residual sum of squares. The bootstrap method was used to estimate the p-value of this test statistic\[[@B47],[@B49]\]. To obtain accuracy in estimating the p-values, 10 million bootstrap samples were generated. Significant probe sets were selected based on an FDR-adjusted q-value less than 0.05\[[@B50]\]. The replication analysis was performed in an independent set of samples, using deCODE samples from patients enrolled in Iceland, to validate probe sets identified by the time course analysis of the CardioGene samples. We performed the same analysis using the time varying intercept model. P-values were estimated based on 100,000 bootstrap samples, and FDR adjusted q-value of 0.05 was, again, used to select significant probe sets. Here the adjustment was based on the selected set of genes since it was a replication analysis. As a sensitivity analysis and to evaluate whether any random set of 46 probe sets would result in significant results as we saw in the analysis of deCODE samples, we randomly selected 46 probe sets, and repeated the validation analysis. This procedure was repeated 100 times, and the number of probe sets yielding an FDR value less than 0.05 was observed. Data analyses were performed based on the R language utilizing libraries freely available via the Bioconductor project. Box plots were generated using the default R software options. The upper whisker is the data point which is 1.5 times the difference between 75th percentile and median from the upper bound of the box, and lower whisker is 1.5 times the difference between median and 25th percentile from the lower bound of the box. Overall, the whisker represents the data point which is 1.5 times the interquartile range from the box. Finally, to evaluate genes identified in a prior studies of ISR \[[@B23],[@B24]\], we analyzed the reported genes in our data, at the baseline timepoint comparing normalized gene expression values between ISR cases and no-ISR controls and also against our timecourse analysis results. Pathway analysis ---------------- Annotation of probe sets was done using the NetAffx (Affymetrix U133A library set) and the DAVID/EASE software, which provides gene name and functional annotation of probe sets\[[@B51],[@B52]\]. Analysis of over-representation, using a Fisher\'s exact test of over-representation of genes from different functional categories based on gene ontology classifications, was performed. List of abbreviations ===================== BMS: bare metal stents; CAD: coronary atherosclerotic disease; DES: drug-eluting stents; ISR: in-stent restenosis; PBMC: peripheral blood mononuclear cells Conflict of Interest Disclosures ================================ The authors declare that they have no competing interests. Authors\' contributions ======================= Study design (EGN, JJ, KAS, NLG, SKG, WON), CardioGene clinical enrollment and blood and data collection (DRH, KAS, KON, LM, RDS, SKG, WON), deCODE clinical enrollment and data collection (AH, KA, GT, TG), RNA expression data generation (EMB, KON, SKG), Data analysis (JJ, SKG), Manuscript preparation (AH, DRH, EGN, JJ, SKG, WON). All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1755-8794/4/20/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Supplementary Table 1**. Initial 46 probesets identified in the timecourse analysis of the CardioGene samples. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Supplementary Table 2**. Gene ontology annotation of the final 32 genes identified and validated. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Supporting Materials**. Statistical methods and lists of probe IDs discovered and replicated. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was supported by the Division of Intramural Research of the National Heart, Lung and Blood Institute, National Institutes of Health.

PubMed Central

2024-06-05T04:04:19.435228

2011-2-28

PMC3053214

Background ========== Centromeres are the regions of chromosomes where kinetochores form and microtubules attach to guide chromosomes to opposite poles during cell division. A wide variety of centromere forms are observed in the diverse cellular contexts of different organisms, from the small point centromeres of *Saccharomyces cerevisiae*to the diffuse, holocentric centromeres of *Caenorhabditis elegans*. Even within a single individual, centromeres must meet the distinct requirements of chromosome dynamics in both meiotic and mitotic cell division. From the outset, one would guess that centromere DNA elements would specify binding sites for structural and regulatory proteins. Indeed, specific, functionally important DNA sequences have been identified, such as the CENP-B (centromere protein B) box site in vertebrates \[[@B1]-[@B3]\] and the CP1 binding site in *S. cerevisiae*\[[@B4],[@B5]\]. Intriguingly, however, multiple lines of evidence indicate that core kinetochore proteins can have a large degree of sequence independence (for recent review, see \[[@B6]\]). Foremost among these arguments is the unexpectedly low level of DNA conservation in centromeres. General structural features are common, most notably large arrays of repeated units (called centromere tandem repeats or satellite DNA) and retrotransposons. In addition, the core centromeres, where the kinetochores form, are often flanked by pericentromeres that may contain similar repetitive DNA. Core centromeres are often megabases in length, as are the pericentromeres. The defining distinction between functional core centromeres and all other chromosomal regions lies in its chromatin composition: specifically, histone H3 of canonical nucleosomes is replaced by a variant called centromeric histone H3 (CENH3; CENP-A in mammals). The precise structure of CENH3 nucleosomes and higher dimensional chromatin is a matter of much investigation and may differ considerably in different cellular contexts (for recent review, see \[[@B6]\]). Given the ambiguous role of DNA sequence in specifying centromere identity and function, it is of considerable importance to understand what contribution DNA sequence makes to chromatin structure. A central question related to both CENH3 and canonical histone H3 nucleosomes concerns interactions between histones and DNA. Both *in vitro*and *in vivo*studies have demonstrated preferences for nucleosomes to bind DNA containing an approximate 10-base periodicity of AA or TT dinucleotides \[[@B7]-[@B9]\]. This nucleotide distribution, along with others, favours the tight bending and twisting of the DNA double helix required for wrapping around a histone octamer \[[@B10],[@B11]\]. A propensity for tight bending is a conserved feature of satellite DNA, and it is thought to contribute to the structural stability and compaction of centromere chromatin \[[@B12]-[@B15]\]. In addition, satellite DNA in humans, in conjunction with CENP-B, favours a phased arrangement of nucleosomes, where nucleosomes often reproducibly adopt the same positions on the DNA (relative to satellite dimers) \[[@B1],[@B3],[@B16]\]. Nucleosomes of African green monkeys are also preferentially positioned but do not depend on the canonical CENP-B binding site, suggesting that histone-DNA sequence affinities or other proteins may influence the placement of nucleosomes \[[@B17]-[@B22]\]. In these species the satellite DNA unit is about 171 bp - just long enough to accommodate a single nucleosome core (histone octomer) and linker (the DNA between nucleosomes cores, bound by histone H1). However, in other species such a relationship is probably impossible because the satellite unit is too small. For example, in mouse the basic satellite unit is 120 bp, in sugarcane it is 140 bp, and in soybean it is two separate units of 92 and 411-bp \[[@B23],[@B24]\]. Factors such as transcription are also known to play important roles in nucleosome positioning \[[@B25]-[@B27]\]. Centromere elements such as satellites are thought to be transcriptionally active, although it is unclear how centromeric transcripts are initiated (for a recent review, see \[[@B28]\]). Questions related to transcription and nucleosome positioning are of general interest in centromere regions because of the potential contribution to both kinetochore assembly and large-scale chromosome dynamics. In addition, the contribution of other centromere DNA elements such as retrotransposons to centromere chromatin structure has received much less attention than satellites but is, nevertheless, of importance given the abundance of these elements in centromeres. The availability of deep sequencing technology prompted us to explore the relation between DNA sequence and *in vivo*CENH3 chromatin structure on centromeres in maize (*Zea mays*). Through analysis of the CENH3 nucleosome core footprints at single-molecule resolution provided by 454 sequencing and through other approaches, we found a tendency for nucleosome spacing intervals to average about 190 bp, which was maintained across the three major DNA components of maize centromeres. These three different DNA elements, however, were sharply distinguished from each other by modes of nucleosome positioning. On arrays of the 156-bp satellite (tandem repeat) *CentC*we found no evidence for nucleosome phasing but, instead, found that the nucleosome positioning was in conformity with an underlying \~10-bp periodity of AA/TT dimers. Nucleosomes on the two centromeric retrotransposons *CRM1*and *CRM2*lacked a detectable correlation with AA/TT periodicity but *CRM2*showed strong nucleosome phasing while *CRM1*did not. We conclude from these data that DNA elements within centromeres can have distinct effects on chromatin structure in terms of nucleosome positioning (and hence conformational rigidity of the local chromatin). Results and discussion ====================== Derivation and validation of CENH3 nucleosome core sequences ------------------------------------------------------------ A dataset of DNA sequences associated with maize CENH3 chromatin was described previously as a resource for defining functional centromere regions (core centromeres) \[[@B29],[@B30]\]; GenBank Sequence Read Archive SRA009397). This dataset also provides a unique opportunity to characterize CENH3 nucleosome footprints at the nucleotide level due to both the method of preparing DNA fragments and the sequencing technology used. In particular, digestion of isolated chromatin with micrococcal nuclease eliminated unprotected DNA while preserving the stretches that were tightly bound by protein. Chromatin immunoprecipitation (ChIP) with antibodies against CENH3 then enriched for DNA/protein complexes containing CENH3. 454 sequencing with read lengths of up to 300 bases allowed for precise mapping of DNA fragment endpoints. (The normal steps to remove small fragments were omitted in order to avoid biasing towards large DNA fragments). In order to define the DNA fragments at high resolution, we first filtered through the resulting 149,756 raw reads to remove the ones where the 3\' end of the DNA fragment could not be identified due to low sequence quality or fragment lengths in excess of the read length capability. We aligned the resulting set of 110,678 reads to the maize genome to filter out junk DNA fragments (for example, artefacts of library preparation and contamination by non-maize DNA), to produce a set of 105,284 reads corresponding to putative CENH3 nucleosome cores. We examined the lengths and nucleotide composition of these reads and found two characteristics similar to those of general nucleosome cores: First, a predominant length of \~155 bp, slightly larger than the 147 bp typical of canonical H3 nucleosome core DNA in other organisms (Figure [1A](#F1){ref-type="fig"}; also see Additional File [1](#S1){ref-type="supplementary-material"}). Second, a tendency to favour either AA or TT dinucleotides every 10 bp. We calculated the number of AA or TT dimers at each position along the length of the aggregated, complete set of reads, aligned at their starts, and we observed a \~10-bp periodicity consistent with nucleosome structure and previous deep sequencing analyses \[[@B7],[@B10]\] (Figure [1B](#F1){ref-type="fig"}; Additional File [1](#S1){ref-type="supplementary-material"}). Furthermore, in cases where the MNase products were substantially longer than an expected nucleosome core length, a subtle change in AA/TT content was visible (near bp 155; Figure [1B](#F1){ref-type="fig"}), providing further support for these reads being derived from nucleosome cores. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **CENH3 nucleosome core lengths and AA/TT contents**. (A) Number of reads per read length (bp). Genome-matching reads with unambiguous termini favour a length of around 155 bp. \'total genome\' includes everything that aligned to the maize genome, version 1. The three centromere elements are shown separately. See also Additional File [1](#S1){ref-type="supplementary-material"}, part B. (B) The frequency of AA or TT dimers at each position along the read lengths. The reads were aligned at their starts and the number of AA or TT dimers counted at each position for each aggregated set. We arbitrarily cut off the analysis at position 200 in order to avoid statistical noise from the low number of reads with lengths greater than 200 bases. In the total genomic reads, the change in AA/TT content visible near position 155 suggests a subtle distinction in nucleotide content between the nucleosome core DNA and linker DNA. The *CentC*, *CRM1*, and *CRM2*reads were identified by blastall alignments. See also Additional File [1](#S1){ref-type="supplementary-material"}, part C. :::  ::: AA/TT periodicity is a prominent feature of nucleosome positioning on *CentC*arrays, but not on *CRM1*and *CRM2*retrotransposons -------------------------------------------------------------------------------------------------------------------------------- The tendency for AA/TT periodicity in the underlying DNA indicates that centromere DNA sequence contributes to centromere nucleosome positioning. In order to further examine this, we compared results for nucleosomes from different centromere DNA elements. The three most substantial centromere components identified in maize DNA are the \~156-bp *CentC*elements which exists in long tandem arrays, the \~7.5-kbp retrotransposon *CRM2*, and the \~7-kbp retrotransposon *CRM1*\[[@B31]-[@B34]\]. All three of these are highly enriched for centromere core regions and are conserved in other cereals. We aligned the total genomic reads to these elements and measured read counts, dinucleotide periodicities, and nucleosome position and spacing preferences. With our alignment parameters, 7.6% of the reads corresponded to *CentC*, 7.0% to *CRM2*and 4.2% to *CRM1*(Figure [1A](#F1){ref-type="fig"}; Additional File [1](#S1){ref-type="supplementary-material"}). While neither *CRM1*nor *CRM2*reads showed detectable AA/TT periodicities, *CentC*reads showed a dramatic \~10-bp AA/TT periodicity (Figure [1B](#F1){ref-type="fig"}). A control dataset of 454 reads unenriched for CENH3 and which was not subjected to micrococcal nuclease (MNase) digest lacked this periodicity, which indicated that it was not an artefact of library preparation or sequencing technology (Additional File [1](#S1){ref-type="supplementary-material"}). Distinct nucleosome phasing constraints on *CRM2*and *CRM1*retrotransposons --------------------------------------------------------------------------- Visual analysis of the alignments for each element revealed strong constraints on nucleosome positioning, or phasing, on *CRM2*, especially in its highly conserved long terminal repeat (LTR) regions \[[@B33]\] (Figure [2A](#F2){ref-type="fig"}). In particular, peaks in nucleosome occupancy occurred on \~190-bp intervals. Alignment positions on *CRM1*, in contrast, showed very little constraint, suggesting a lack of nucleosome phasing (Figure [2B](#F2){ref-type="fig"}). In the case of *CentC*, the alignments showed a diversity of allowed positions but with an enrichment for a single position at the ends of the sequence shown in Figure [3A](#F3){ref-type="fig"}. One concern is that inherent MNase cleavage biases could explain the enrichment for such a single position \[[@B35]-[@B37]\], since any site favoured by MNase on *CentC*would be repeated every 156 bp in long arrays. In order to test whether the favoured position on *CentC*is a result of nucleosome occupancy or MNase bias, we measured the relative enrichment of MNase fragments from both naked DNA and chromatin by selective amplification and Sanger sequencing. The results of this analysis clearly indicate that the enriched position can be explained as a consequence of MNase sequence preference (Figure [3B](#F3){ref-type="fig"}). It remains possible that an actual nucleosome position overlaps with a site of inherent MNase bias or that phasing could involve a complex relation with higher order *CentC*polymers. However, a one-nucleosome/one-repeat scenario would require unusually compact chromatin and would be hard to reconcile with the existence of even shorter tandem repeats of soybean (92 bp), mouse (120 bp) and sugarcane (140 bp). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Alignments of CENH3-ChIP reads to *CRM1*and *CRM2***. Number of CENH3-ChIP reads aligning to each position on *CRM2*(A) and *CRM1 (B)*. Vertical lines are spaced every 190 bp. In the case of *CRM2*, but not *CRM1*, alignments show strong tendency to peak every 190 bp, particularly toward its long terminal repeats (LTRs), indicating nucleosome phasing. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### ***CentC*alignments and micrococcal nuclease (MNase) sequence preference**. (A) The number of alignment edges at each position on *CentC*from CENH3-ChIP reads. Due to the short and repetitive character of *Cent*, just the alignment edges are depicted rather than the entire alignments. An apparent preferred position is defined by right edges positioned over the bracketed section (but see B). Alignments to *CentC*were split into two categories, forward and reverse orientation; then the 5\'-most position of the forward alignments were combined with the 3\'-most position of the reverse alignments to count the total number of left edge positions along the length of *CentC*. Likewise, the 5\'-most position of the reverse alignments and 3\'-most position of the forward alignments were combined to count the total number of right edge positions. Since this analysis uses only information from alignment edges, the alignments were first filtered so as to include only the ones with approximate nucleosome lengths (145 to 175 bp). (B) Sequence preference in MNase digestion of *CentC*. DNA fragments produced by MNase digest of naked DNA or chromatin were captured by ligation to adapters then amplified with primers that selected for *CentC*-adapter junctions (corresponding to the right edges shown in A). The bar chart indicates the number of cloned fragments (per 100) whose right edges were within 3 bp of the enriched site found in the CENH3-ChIP 454 dataset (the bracketed section in A, centred on the last base in GGGTGTCGGGGTG). Errors bars are standard errors of the means, calculated based on the sample sizes (74 sequences for \'Naked DNA No.1\', 37 for \'Naked DNA No.2\', 46 for \'Naked DNA No.3\', 46 for \'Naked DNA No.4\', 71 for \'Chromatin No.1\', 41 for \'Chromatin No.2\' and 47 for \'Chromatin No.3\'. :::  ::: Uniform nucleosome spacing on *CentC*arrays and on *CRM1*and *CRM2*retrotransposons ----------------------------------------------------------------------------------- Spacing between nucleosomes is known to involve factors other than DNA sequence \[[@B38]-[@B40]\]. We wondered whether CENH3 nucleosomes would exist in arrays with regularly spaced nucleosomes across the whole centromere core or whether spacing patterns would vary within different centromere regions depending on the underlying DNA sequence. The alignments to *CRM2*(Figure [2A](#F2){ref-type="fig"}) showed a tendency for spacing of nucleosomes on 190-bp intervals. In order to quantify this, we calculated distances between alignment start positions from each strand of DNA. As depicted in Figure [4A](#F4){ref-type="fig"}, a strongly positioned nucleosome would be predicted to produce start-to-start distances that match the size of one nucleosome core at roughly 155 bp. Over longer distances, we expect this distance to reflect the overall spacing of the nucleosomes including linkers. If we first align all CENH3 ChIP reads to the *CRM2*consensus sequence and, beginning with a start site demarcating one end of a nucleosome core region, plot the distance to a start site on the opposite strand and the start site after that, and so on, we will reveal the spacing parameters typical of the *CRM2*sequence. As can be seen in Figures [4B](#F4){ref-type="fig"} and [4C](#F4){ref-type="fig"}, these data show an unequivocal 190 base periodicity that spans the entire element (though less evident in the intermediate distances, consistent with the weaker positioning seen in the central, genic area of the retrotransposon; Figure [2A](#F2){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Start-to-start distance analysis of nucleosome spacing on *CRM1*and *CRM2***. (A) A schematic description of the start-to-start distance analysis. Three 155-bp nucleosomes (represented by purple circles) starting 190 bp apart give rise to 155-bp reads with alignments of both orientations\--forward in blue, reverse in red. Distances between the start of the first forward alignment and start of the first reverse alignment correspond to the length of one nucleosome core. Distances between start of the first forward alignment and subsequent reverse alignments increase by 190 bp, corresponding to an additional nucleosome core and linker region. (B) A short range start-to-start distance analysis for *CRM1*and *CRM2*nucleosomes. The number of starts separated by each possible distance was quantified in terms of a coincidence index value. We arbitrarily cut off this plot at 650 bp to emphasize the presence of a peak at \~155 bases (as well as two subsequent peaks separated by \~190 bp). Vertical lines are spaced every 20 bp. (C) A long range start-to-start distance analysis for *CRM2*CENH3 nucleosomes. Distances between start positions from alignments of opposite orientations were calculated and the number of starts separated by each possible distance between -7577 and +7577 bp quantified in terms of a coincidence index value. The length of the *CRM2*reference sequence is 7577 bp, precluding any longer distances. For a random sampling of alignment pairs, the probability is lowest in order for them to be separated by the longest possible distance (the element\'s length), as there is only one possible combination for either the positive or negative extreme. In contrast, the number of combinations that allow for a distance of zero is almost as large as the length of the element. Hence, the triangular shape of the plot, with a peak centred near zero and tapering off in both directions. Vertical lines are spaced every 190 bp. :::  ::: In contrast to *CRM2*, *CRM1*lacked both the single-nucleosome peak at start-to-start distance 155 as well as the subsequent peaks every 190 bp (Figure [4B](#F4){ref-type="fig"} and [4C](#F4){ref-type="fig"}), consistent with the lack of phasing seen in Figure [2B](#F2){ref-type="fig"}. This result does not rule out uniform spacing: Even a perfectly uniform spacing arrangement would be undetectable without phasing constraint. Similarly, since *CentC*is only long enough to accommodate a single nucleosome and longer arrays consist of highly similar repeats, we could not perform meaningful start-to-start distance analyses to measure nucleosome spacing. Instead, we used Southern blotting to compare nucleosome polymer lengths from MNase-treated chromatin with probes for *CentC*, *CRM1*and *CRM2*. Intriguingly, the greater part of the signals from all three probes was indistinguishable and yielded DNA fragment sizes corresponding to \~190-bp nucleosome spacing (Figure [5A](#F5){ref-type="fig"}). Combined with the lack of measurable phasing on *CentC*(not attributable to MNase sequence preference), this result argues against any substantial level of one-nucleosome/one-repeat phasing on *CentC*. 190-bp spacing was also the predominant arrangement for nucleosome polymers genome-wide, as evidenced by ethidium bromide staining of total chromatin after MNase digest (Additional File [2](#S2){ref-type="supplementary-material"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **A Southern analysis of nucleosome spacing on *CRM1*, *CRM2*and *CentC***. Southern blots of micrococcal nuclease (MNase) digested CRM1, *CRM2*and *CentC*nucleosome cores. (A) Total chromatin was lightly digested with MNase and isolated DNA fragments corresponding to nucleosome monomers, dimers and higher order polymers were separated by gel electrophoresis and probed for *CentC*and for *CRM1*and *CRM2*LTRs. (B) A rare fragment, \~30-bp shorter than the majority of the nucleosome dimers, was often visible, especially for *CentC*(blue arrow). In some cases a shorter fragment could also be seen in the trimer range. :::  ::: A second, minor potential spacing pattern was also suggested, apparently derived from adjacent nucleosomes with short or absent linkers (Figure [5B](#F5){ref-type="fig"}). This was manifest as a \~30-bp shorter fragment in the nucleosome dimer range, particularly for *CentC*. In some cases we could also distinguish a shorter band in the nucleosome trimer range, but higher order polymers did not give sufficient resolution on the gel. (The shorter fragments did not appear to be centromere specific, as they were also present in arrays of the maize 180-bp *Knob*repeat; data not shown). Another piece of evidence for deviations from perfect, absolute 190-bp spacing is the 156-bp length of *CentC*and the tendency for nucleosome positions to conform to the underlying AA/TT distribution. Periodic deviations of a few bases would be required in order to retain the expected conformation between the histones and the DNA (though an average 190-bp spacing could be maintained). Since the chromatin used for Southerns was not enriched for CENH3 by ChIP, a substantial portion of the signal from each element resulted from canonical H3 nucleosomes. Even the analyses from CENH3 ChIP cannot fully distinguish between canonical H3 and CENH3 effects if both types of histones coexist in nucleosomes arrays (which is unknown at this time, but would be consistent with the unusual timing of CENH3 deposition compared to canonical H3 deposition \[[@B41]-[@B43]\]). Nonetheless, these data suggest that the dominant theme for nucleosomes on *CRM1*, *CRM2*and *CentC*elements (as well as the genome as a whole) is spacing intervals of \~190-bp on average - though the mode of positioning and rigidity of the chromatin depends on the local DNA. Conclusions =========== We conclude from these data that structural features of centromere chromatin can be linked to the underlying DNA and that this can result in non-uniform chromatin environments within CENH3 regions. In addition, factors independent of local DNA impose a generally similar nucleosome spacing arrangement. It has long been known that centromere chromatin is more conformationally rigid than other chromatin and this has been proposed to stem, in part, from intrinsic DNA bending/nucleosome interactions \[[@B12]-[@B15]\]. We further support this idea with our finding that nucleosome positioning on tandem repeats adapts to the underlying AA/TT distribution. The absence of a nucleosome phasing signal on the tandem repeat in maize contrasts with both the phasing reported on tandem repeats in primates and with the phasing on the maize retrotransposon *CRM2*. Interestingly, for *CRM2*, the dominant positioning frame is specified without a recognizable AA/TT pattern. However, a strong sequence motif may be present and through enforcing the exact position of one nucleosome on *CRM2*, could set the phase for the entire array of nucleosomes. Such a motif need not interact directly with the nucleosome. For example, binding of the transcription factor CTCF sets the positions of large nucleosome arrays \[[@B44]\]. The retrotransposon *CRM1*presents yet a different situation, with neither a detectable correlation with AA/TT dimers nor any strongly favoured positioning frame. Centromere DNA presents a puzzle in that although it tends to have less conservation than the genome as whole and is largely made up of unstable repetitive elements, it forms the basis for a highly conserved and essential structure. If the DNA makeup were important for kinetochore function, one would expect to find mechanisms for preserving centromere DNA integrity, mechanisms for adapting to a fluid DNA landscape, or both \[[@B45]\]. Whatever the forces in play, our data suggest that centromere sequences are limited by constraints on AA/TT periodicity and other features affecting chromatin structure. An important consideration in interpreting these results is the structure of CENH3-containing nucleosomes. Recent studies of animal centromeres suggest that CENH3 exists in a specialized smaller nucleosome made up of a tetramer of H2A, H2B, H4 and CENH3 with the DNA wrapped around the nucleosome in a right-handed configuration rather than the canonical histone octamer with the DNA wrapped around in a left-handed configuration \[[@B46]-[@B48]\]. In contrast, other studies have shown that CENH3 is capable of forming octameric nucleosomes, and a recent DNA topological analysis revealed DNA in a left-handed configuration \[[@B49]\]. Consequently, a lively debate is taking place as to the relative importance of both nucleosome structures (and other alternative ones). Since the data supporting the presence of tetrameric nucleosomes in animals is compelling and CENH3 is highly conserved, we expect that functionally important tetramers will also be found in maize. As such, we might have also expected to see a substantial amount of smaller nucleosome core particles corresponding to tetrameric nucleosomes in the 454 sequencing dataset and Southern blots (Figures [1A](#F1){ref-type="fig"} and [4](#F4){ref-type="fig"}). Nevertheless, we did not see such structures despite the fact that the sample was highly enriched for the genetically-defined centromeric regions and were not only enriched for *CentC*(33.5 fold enrichment), *CRM1*(16.3 fold) and *CRM2*(51.1 fold) but for hundreds of other unique sites within the centromere core regions \[[@B29]\]. In fact, these reads were used to precisely define centromere cores \[[@B29],[@B30]\]. Therefore, our data suggest that the 155-bp, CENH3-containing nucleosomes are a major component of the functional centromere chromatin. It remains possible that specificities of the antibodies or the extent of MNase digestion may have biased the analysis towards octameric structures. Perhaps more important is the fact that most of the cells in young leaves have ceased division and are, presumably, at the G1 phase of the cell cycle when CENH3 nucleosomes are stable and fully assembled. One way to explain the inconsistencies between our data and prior work in animals would be to postulate that tetrameric nucleosomes are more common in dividing cells as an assembly intermediate following DNA replication but before CENH3 deposition. In summary, these results indicate that centromere DNA elements influence chromatin structure but leave open the question of whether one or more of these distinct modes of nucleosome positioning have consequences in terms of centromere specification and kinetochore function. The identification of apparently stable neocentromeres that lack canonical centromere elements suggests limits to the importance of centromere DNA sequence (for recent review, see \[[@B50]\]). However, in maize, after a centromere has been epigenetically deactivated (when CENH3 and other kinetochore markers are lost), the same region can regain activity, hinting that centromere elements present a particularly permissive environment for centromere formation \[[@B51]\]. Similarly, satellite DNA is known to be important for human artificial chromosomes \[[@B2],[@B52]\] and the non-standard geometry of neocentromeres is associated with kinetochore defects \[[@B53]\]. These data, and the fact that kinetochore positions are remarkably stable over long periods of time, make a strong argument for the functional consequences of centromere DNA. Our data suggest that this favourable environment in maize includes multiple DNA elements including *CRM2*- which can strongly phase CENH3 nucleosomes - and *CentC*, whose sequence makeup readily conforms to the tight bending of a nucleosome core structure. Methods ======= 454 read processing and aligning -------------------------------- The 454 outputs in SFF format were first converted to fasta format using sff\_extract \[<http://bioinf.comav.upv.es/sff_extract/>\]. The reads were then scanned from the 3\' end for the first instance of \"CTGAGACAC\" (the first nine bases of the 454 adapter) and trimmed back to that point. Reads that lacked this sequence or were shorter than 39 bases were discarded. The resulting set of trimmed reads was aligned to the maize genome, version 1, using a local installation of BLAT \[[@B54]\]. The following parameters were changed from default settings: -fastMap, -minIdentity = 99, -maxIntron = 5, and -minScore = 30 (87% of the CENH3 reads and 82% of the control reads could be aligned with the much more stringent setting -minScore = 300). The resulting reads with genomic matches were then aligned to a reference set of DNA sequences consisting of a *CentC*trimer, a *Knob180*trimer and *CRM1*-*CRM4*. For this much smaller reference, a local installation of blastall was used, with default DNA parameters except that expectation value (-e) was set to 1 × 10^-15^. In cases where a single read produced multiple equally good alignments, only one alignment was kept. Start-to-start distance analysis -------------------------------- Alignments were separated based on orientation relative to the reference sequence and the distance between forward- and reverse-oriented alignment start positions calculated using a similar method as Gent *et al.*\[[@B55]\]. In brief, this analysis quantified the tendency for alignment start positions, \'starts\', to occur at particular distances from each other, \'start-to-start distances\'. The number of starts at each position was counted and then the distances between all possible pairwise combinations of starts were measured and assigned a \'coincidence index\' value related to the cube root of the number of starts separated by that distance. Control for MNase bias in digestion of *CentC* ---------------------------------------------- ### Chromatin isolation Whole maize seedlings were flash frozen in liquid nitrogen and ground to a powder with a prechilled mortar and pestle; 10 mL ground tissue was homogenized by vigorous shaking in a 50 mL volume of TBS (Tris-buffered saline) plus 0.5% Tween 40, filtered through two successive layers of 35-μm-pore Nitex bolting cloth (Wildlife Supply Co, FL, USA), then the nuclei pelleted by centrifugation at 600 g for 10 min. After removing the supernatant, the nuclei were gently suspended in 5 mL of 25% sucrose in TBS, overlaid onto 2.5 mL of 50% sucrose in TBS, then centrifuged at 1500 g for 20 min. At all times, including during centrifugation, the material was kept near 4°C. ### Micrococcal nuclease digestion Isolated chromatin and naked DNA samples were suspended in digestion buffer (0.32 M sucrose, 50 mM Tris, 4 mM MgCl~2~, 1 mM CaCl~2~). MNase (2 × 10^6^gel units/mL; New England Biolabs, MA, USA, No. M0247S) was added at about 1 μL per 500 μL of digestion buffer solution (on ice). The MNase and DNA mixtures were then separated into several smaller volumes for incubations at 37°C for different time periods (30 s to 3 min). The reactions were stopped and the DNA purified as described previously \[[@B7]\]. ### Sequencing library preparation The purified DNA fragments were run on 3.5% NuSieve^®^GTG^®^Agarose (Lonza, Basel, Switzerland) gels, and the nucleosome monomer size range, 140-190 bases, excised and extracted from the gels with a QIAquick^®^Gel Extraction Kit (QIAGEN Sciences, Maryland, USA). The samples were kept at room temperature during the entire procedure to avoid denaturation of the complex pools of DNA fragments. The fragments were then inserted between 5\' and 3\' adapters and amplified using a method similar to that of Gu and Fire \[[@B56]\]. A key difference in our method was that, rather than using the published primers that bind to both the 5\' and 3\' adapters and amplify independently of the DNA insert sequence, we selected for *CentC*inserts with a *CentC*-specific primer. The primer sequences were: GTGGTTTCGCGCAATTTCGTTGTC (KD-JIG-72, binds *CentC*) and AAGACGGCATACGAGCTCTTCCGAT (KD-JIG-98, binds the 3\' adapter). Hence, the junctions of the 3\' adapter and 5\' portions of *CentC*inserts were amplified selectively. Amplified products of \~30 to 200 bp were again gel purified (3.5% NuSieve agarose) and extracted with a QIAquick kit without heating, then cloned and Sanger sequenced. Southern blotting ----------------- In general, all steps were as described in the *DIG Application Manual for Filter Hybridization*by Roche (Basel, Switzerland) <http://www.roche-applied-science.com/PROD_INF/MANUALS/DIG_MAN/dig_toc.htm>; specific details were as follows: purified MNase digested DNA fragments were produced as described above and were then separated on a 4% NuSeive agarose gel for at least 3 h at 100 V and then transferred to a Amersham HyBond-N^+^(GE Healthcare, NJ, USA) membrane by capillary action overnight without prior depurination of the DNA. Membranes were baked at 80°C for 2 h after transfer. Probes were allowed to hybridize overnight in DIG Easy Hyb (Roche No. 11 603 558 001) at 42°C for *CRM1*and *CRM2*and 44°C for *CentC*. All three probes were \~150 bases in length and were produced with the PCR DIG Probe Synthesis Kit (Roche No. 11 636 090 910). Primers were as follows: For *CentC*, GCCACCGGAACCATTTCTTCGTTT (KD-JIG-63) and TCGTGCTTTGTATGCACCC (KD-JIG-91); for *CRM1*, ATCTCCACTCACCGAAAGATTGGG (KD-JIG-82) and AAGGGTGCTGGGATAAGGTCTAAC (KD-JIG-83); and for *CRM2*, TTGTAAGCGCGCGTGCTAGTTCA (KD-JIG-86) and ACTCGTCCTGCAAGCAATCGAAGA (KD-JIG-87). For C*RM1*and *CRM2*detection, the high stringency buffer was 0.5% SDS (sodium dodecyl sulphate) at a temperature of 65°C; for *CentC*it was 0.1% SDS at 68°C. Detection was carried out with the DIG Luminescent Kit for Nucleic Acids (Roche No. 11 363 514 910) Preparation of a control 454 library ------------------------------------ Total chromatin was extracted from maize (inbred B73 stock) immature ears, chemically crosslinked and fragmented by sonication \[[@B57]\]. After an, at best, marginally successful attempt at enrichment for centromeric sequences by ChIP (immunoprecipitation with an antibody against maize Centromere Protein C), the DNA was processed and 454 sequenced. Abbreviations ============= ChIP: chromatin immunoprecipitation; CENH3: centromeric histone H3; CENP: centromere protein; MNase: micrococcal nuclease; LTR: long terminal repeat. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= JIG, CNT and CR performed the experiments. JIG, KLS and GGP performed the bioinformatic analyses. RKD, JIG and GGP designed the study. JIG and RKD wrote the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **454 control reads for centromere enrichment, length distribution and AA/TT content**. (A) Enrichment of reads for centromere DNA. Genome-matching, CENH3-chromatin immunoprecipitation (ChIP) and control reads were aligned to a set of reference DNA sequences consisting of *CentC*, *CRM1*, *CRM2*, *CRM3*, *CRM4*and *Knob180*by blastall. The number of reads that map to each alignment relative to the number of reads that map to the genome is shown. In order to allow for the complete alignment of any theoretical perfect matching read up to 313 bp, a trimer of *CentC*was used rather than a single unit. *Knob180*was also trimerized. In cases where a single read produced an alignment to multiple elements, only the longest alignment was counted. While *Knob180*is an extremely abundant repeat, it is not a component of centromeres; neither have we observed CENH3 localized to *knob*repeats by immunolocalization. We do not know the reason for the strong *knob*signal in the CENH3 reads but suspect some level of background non-CENH3 nucleosome cross-reactivity with the antibody \[[@B30]\]. The control reads provide a comparison for the potential biases introduced during 454 sequencing or library preparation. (B) The number of control reads per read length (bp). Genome-matching reads with unambiguous termini do not strongly favour a particular length. (C) The frequency of AA or TT dimers at each position in first 200 bp of each control read. The reads were aligned at their starts and the number of AA or TT dimers counted at each position for each aggregated set. We arbitrarily cut off the analysis at position 200 in order to avoid statistical noise from the low number of reads with lengths greater than 200 bp. In addition to a lack of detectable AA/TT periodicity, no change in AA/TT content is visible near position 155, in contrast with the CENH3 ChIP reads (see Figure [1](#F1){ref-type="fig"}). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **General genomic nucleosome spacing**. (A) A plot of gel migration versus DNA fragment length. Total chromatin was lightly digested with micrococcal nuclease (MNase) and run on a 4% NuSieve agarose gel along with a 100 bp DNA ladder. The known lengths of the ladder DNA were used to plot a standard curve of DNA length as a function of migration distance (dashed line and diamonds). The gel migration of each discernable band produced in the MNase digest, corresponding to successively larger polymers of nucleosome cores, was then mapped onto the standard curve to estimate its length. (B) Summary table of nucleosome spacing. Since a polymer of N nucleosomes contains N-1 linkers, the spacing between the start of each nucleosome in a nucleosome core N-mer should be (total length + one linker length)/N. Using a linker length of 35 bp gives a consistent nucleosome spacing distance of \~190 bp for the polymers examined (dimers to hexamers). The monomer length of \~175 bp is larger than expected due to the incompleteness of the digest (required for retention of larger polymers). ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was supported by grants from: the National Science Foundation (0421671 and 092270) to RKD; from the National Institutes of Health (NIGMS-F32-GM095223) to JIG; and by a University of Georgia Graduate School Dissertation Completion Award to CNT. We thank Richard Meagher for helping to inspire the project and Sarah Wolf for technical help.

PubMed Central

2024-06-05T04:04:19.439234

2011-2-25

PMC3053215

Background ========== Adhesions develop in over 90% of patients after abdominal operations \[[@B3],[@B7]\] and can lead to significant postsurgical complications, including small bowel obstruction, infertility, chronic pelvic pain and difficult re-operative surgeries \[[@B2],[@B8]\]. Adhesions formation is a dynamic and complex process, which involves a cascade of reactions of cellular, biochemical, immunological and biomechanical factors \[[@B9]\]. Unfortunately, there is no available marker to predict the occurrence or severity of adhesions preoperatively \[[@B10]\] and therapeutic prevention still remains a challenge. At present, the prevention of adhesions formation after surgery has focused on minimizing peritoneal trauma and reducing the implantation of foreign materials into the peritoneal cavity, as they may aggravate the inflammatory response \[[@B11]-[@B14]\]. Numerous approaches have been attempted, including profibrinolytic agents and physical barriers \[[@B3],[@B15]\], such as Dextran (Dex) \[[@B2]\] and sodium hyaluronate (SH)\[[@B3]\]. While the barrier products have been proven the most clinically successful \[[@B4]-[@B6]\], there is no effective method of preventing adhesions formation currently \[[@B1]\] Previously, penicillamine was reported to prevent collagen fibers from crossing into non-soluble collagen tissue and inhibit the maturation of dissoluble collagen. Recent studies indicated the possibility of oral D-penicillamine-induced prevention on peritoneal adhesions band formation \[[@B16]-[@B18]\]. We therefore hypothesized that it can prevent the fibrin from converting into permanent fiber adhesions tissue. Hereby, we developed a novel membrane, which is composed of two regents- penicillamine and hyaluronic acid, and then applied this penicillamine-bound membrane to treat abdominal adhesions in an animal model, in order to identify its preventive and therapeutic potential for adhesions formation. Methods ======= Method for manufacturing novel penicillamine-bound membrane ----------------------------------------------------------- Chitosan \[[@B2]\](Shanghai Qisheng Biologic Agent Company), or polylactic acid or hyaluronic acid \[[@B10]\](Center for New Drug Evaluation, Shandong University) was individually dissolved into saline at the concentration mentioned in previous literatures. Penicillamine (Catalog number: 000108, Shanghai, PR China) was dissolved into three different solutions. The solutions were drained into the flat bottom plastic container and dried thoroughly until polymerized. The thickness and dissolve time for the three different kinds of polymerized membranes were measured in order to select the best substrate of penicillamine-bound membrane. The release of penicillamine was defined by dissolving the membranes into saline solution. Eventually, hyaluronic acid was chosen for the substrate of penicillamine-bound membrane. Hyaluronic acid and aluminum chloride (at the concentration of 5%) were dissolved into autoclaved PBS to make solution 1. Carboxymethyl Cellulose was dissolved in double-distilled water (ddH~2~O) to make solution 2. And then solution 1 and 2 were 1:1 mixed thoroughly. 2.5 ml of 10% penicillamine was pipetted into 50 ml mixed solution in order to lead to cross-linking between penicillamine and substrate. The solutions were drained into the flat bottom plastic container and dried thoroughly for 4-7 days until fully polymerized in air. The thickness of penicillamine-bound membrane was about 0.1 mm, and the degradation of the membrane cost 5 days in corpore. We tested the concentration of penicillamine of the solution after the membrane was dissolved into saline at a different time. Animal model of abdominal adhesions ----------------------------------- Total 150 rats (Wistar rats of both genders from animals facility of Shandong University) at 9 weeks of age, weighing 200\~230 g, were involved in the present study with 25 animals per group in order to calculate the occurrence of abdominal adhesions. The rats were treated under the animal use guidelines of Institutional Animal Care and Use Committee (IACUC) at Qilu Hospital & College of Medicine, Shandong University. Rats were allowed to adapt to the new environment for 1-2 days prior to experimental study. The study was approved by the ethics committee of Qilu Hospital, Shandong University. All animals were randomly divided into six groups (25 per group), including the vehicle group (A) and five treated groups (B, C, D, E and F). All groups were anesthetized with 10% chloral hydrate at the dosage of 300\~350 mg/kg by intraperitoneal injection, and then underwent abdominal surgery through midline incision 1.5 cm in length. The caecum serosa was scratched with dry gauze at 2 cm × 2 cm \[[@B19]\] (Figure [1](#F1){ref-type="fig"}). One milliliter saline was put into the rats\' abdominal cavity in group A, while 1 ml 40% Dextran (Dex), 0.5 ml sodium hyaluronate (SH) (Shandong Zhengda Freda Tragacanth Company) and 1 ml 3% penicillamine for group B, C and D. The scratched area on caecum serosa in group E was covered by penicillamine-bound membrane (Figure [1](#F1){ref-type="fig"}), and group F was covered by non-penicillamine-bound membrane. The membrane was not fixed and adhered to the scratched area naturally. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Animal model of abdominal adhesion**. Arrow in (A) shows the caecum of normal animals. (B) Shows that the caecum is scratched with dry gauze at 2 cm × 2 cm, which is treated as the sham group. While the arrow in (C) shows that, the scratched caecum is covered with penicillamine-bound membrane. :::  ::: Tissue preparation ------------------ Half of the animals in each group (about 10 animals for every group and time point) were sacrificed at post-surgical day 7, while another half was sacrificed at post-surgical day 14. The adhesions tissue inside the abdominal cavity in different groups was removed and stored in 4% paraformalhyde, then subjected to immunohistochemistry staining. The incision, associated with the lateral skin tissue in different groups, was sheared at 4×0.5 cm to test its breaking strength. Measurement of adhesions degree and breaking strength of incision ----------------------------------------------------------------- The occurrence of adhesions was calculated as the ratio between animals with adhesions tissue and total animals within that group (Table [1](#T1){ref-type="table"}). The adhesions grade and score in different groups were defined by Bigatti\'s method \[[@B20]\](Table [2](#T2){ref-type="table"}). The breaking strength of incision was measured by a strength-tester (See Additional file [1](#S1){ref-type="supplementary-material"}). After removing the stitches from the incision tissue, it was connected with a water container by a pulley and the breaking strength was defined by the gravity of the water (The Unit was gram), which was drained into the container when the incision tissue was broken. All of these measurements were performed by a blinded observer. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Comparison of the occurrence of adhesion, adhesion score and breaking strength of incision between control (A) and treated groups (B, C, D). ::: ------------------------------------------------------------------------------------------------------------------------------------------------------- Group (s) Occurrence of adhesion Adhesion score Breaking strength of incision(unit: g) ---------------- ------------------------ ---------------- ---------------------------------------- -------------- ----------------- ------------------ Control\ 100% 100% 7.625 ± 2.92 9.25 ± 1.91 245.1 ± 16.51 323.13 ± 43.77 (A) Dextran\ 48% 80% 2.9 ± 1.19\* 4.625 ± 2.92 100.6 ± 20.02\* 190.45 ± 38.33\* (B) SH\ 56% 80% 3.5 ± 1.6\* 5.25 ± 1.91 115.0 ± 15.5\* 183.6 ± 20.0\* (C) Penicillamine\ 40% 48% 2.7 ± 3.19\* 2.5 ± 1.82\* 198.0 ± 12.35 287.8 ± 11.09 (D) ------------------------------------------------------------------------------------------------------------------------------------------------------- (\* means p \< 0.05, compared to control level) ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Adhesion Score (Bigatti\'s method): ::: Characteristic Adhesion Score ----------------------------------------------------------- ---------------- Tenacity None 0 Adhesions essentially fell apart 1 Adhesions lysed with traction 2 Adhesions required sharp dissection 3 Type None 0 Filmy, no vessels (transparent) 1 Dense, no vessels (translucent) 2 Dense, vascular, small vessels (diameter 50 μm) 3 Dense, vascular, large vessels (diameter 50-110 μm) 4 Extent (% of SILASTIC patch surface covered by adhesions) 0 0 \< 25 1 25-50 2 50-75 3 \> 75 4 ::: Immunohistochemistry -------------------- Slices (40 μm) were made from 4% paraformalhyde-fixed adhesions tissue with a microtome, then transferred into 0.1 M phosphate buffer (PB) (pH = 7.4). The slices were incubated in 1:500-diluted polyclonal rabbit anti-collagen type I (Beijing Biosynthesis Biotechnology Co., Ltd., China) at 4°C overnight, and then washed in 0.1 M PB three times. Slices were then transferred into avidin-biotin-peroxidase complex and incubated for 20 min at 37°C, then washed with 0.1 M PB, incubated with 3, 3-diaminobenzidine (DAB) tetrahydrochloride for 5-15 min and washed three times in 0.1 M PB. Slices were then mounted onto gelatin-treated slides, dried overnight, and dehydrated serially with 50%, 70%, 95% ethanol once, and 100% ethanol and xylene twice. Slides were then coverslipped using the mounting solution and viewed under the microscope. Negative control experiment was performed by applying 0.1 M PB solution as the primary antibody. Data analysis and statistics ---------------------------- The occurrence of adhesion was compared by Chi-square test. The adhesion score, the breaking strength of incision were compared between different groups by one-way ANOVA with post hoc Tukey\'s test. Data was shown as percentage or mean ± SD. *P*\< 0.05 was considered statistically significant. Data analyses were performed using SPSS statistical program version 16.0. Results ======= Penicillamine prevents the abdominal adhesions formation significantly ---------------------------------------------------------------------- An animal model of abdominal adhesions was achieved with a success rate of 88.33%. About 24 animals in total died during surgery, about 4 animals per group for reasons such as bleeding or overdose of anesthesia. Animal death occurred across all groups, which could indicate no potential toxicity of any of the compounds used. Dead animals were discarded from our study. Additionally, our experiments showed no significant difference in adhesion occurrence and scores relating to the sex of the animal. The occurrence of adhesions in group A and D was summarized in Table [1](#T1){ref-type="table"}. Compared to the control group, the occurrence in group D was significantly lower at postsurgical day 7 and 14 (P = 0.0326, P \< 0.05). Further, the adhesions score in group D was significantly decreased 7 days or 14 days after the surgery, compared to control level (P = 0.0064, P \< 0.01). Immunohistochemical staining showed that more collagen fibers (Figure [2](#F2){ref-type="fig"}) and blood vessel hyperplasia (Figure [3](#F3){ref-type="fig"}) were observed in group A than in group D (Figure [2](#F2){ref-type="fig"} at 10×, and Figure [3](#F3){ref-type="fig"} at 40×magnification). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Comparison of adhesion formation in different groups by immunochemistry staining**. Arrows in (A) and (B) show the stained collagen I. Significantly less collagen I and fibers are found in penicillamine-treated group (B) than in the sham group (A). :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Comparison of wound healing in different groups by immunochemistry staining**. Blood vessels hyperplasia is observed in the sham group (A), while not in the penicillamine-treated group (B). :::  ::: Penicillamine has more preventive effects on abdominal adhesions than dextran (Dex) and sodium hyaluronate (SH) --------------------------------------------------------------------------------------------------------------- Compared to group B and C, penicillamine decreased the adhesions score in group D most significantly (P = 0.0326, P \< 0.05). The occurrence of adhesions in group D was significantly lower at postsurgical day 7 (40% for D, 48% for B, 56% for C) and day 14 (48% for D, 80% for B, 80% for C). Penicillamine-bound membrane shows greater benefits in therapeutic prevention of abdominal adhesions than penicillamine ----------------------------------------------------------------------------------------------------------------------- The carrier for the penicillamine must be stable, non-toxic, non-irritating and not react with the drug. The Chitosan dissolve in acid solution, which has irritation. Nevertheless, the polylactic acid dissolve in organic solvents (chloroform, acetone, e.g.) in which penicillamine can\'t dissolve. Penicillamine-bound membrane was developed by using hyaluronic acid as the substrate. The membrane was made by only natural hyaluronic acid, which dissolved into the saline in 15 minutes, and so penicillamine was released thoroughly from the membrane. The new membrane was made by hyaluronic acid, aluminum chloride and carboxymethyl cellulose, which dissolved into the saline in 5 days. Penicillamine has a more prolonged period of action. The concentration of penicillamine in the membrane is 1.501 ± 0.023 mg/cm^2^. The occurrence of abdominal adhesions in the penicillamine-bound membrane-treated group (40%) was significantly lower than control (100%), penicillamine-treated groups (48%) and non-penicillamine-treated groups (78%) (Table [1](#T1){ref-type="table"}) on postsurgical day 14. Comparison of adhesions score (Table [1](#T1){ref-type="table"}) showed a significance between control and treated groups, indicating that both penicillamine and penicillamine-bound membrane successfully prevented abdominal adhesions formation, which was confirmed by morphological observation (Figure [4](#F4){ref-type="fig"}). Moreover, penicillamine-bound membrane showed better effects than penicillamine itself and non-penicillamine-bound membrane (P = 0.0046, P \< 0.01) (Table [3](#T3){ref-type="table"}). Penicillamine-bound membrane could be loaded directly and locally onto the traumatic area, contributing to its advantages in clinical administration than penicillamine. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Comparison of adhesion formation in the sham and penicillamine-bound membrane-treated groups**. Arrows in (A) and (B) show that abdominal adhesion tissue is found in the sham group, while the adhesion is not found in the penicillamine-bound membrane- treated group(C) and (D). Arrows in (C) and (D) show that the scratched caecum is recovered. :::  ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Comparison of occurrence of adhesion, adhesion score and breaking strength of incision between membranes with/without penicillamine- treated groups. ::: --------------------------------------------------------------------------------------------------------------------------------------------------------------------- Group (s) Occurrence of adhesion Adhesion score Breaking strength of incision(unit: g) ------------------------------------ ------------------------ ---------------- ---------------------------------------- ------------- --------------- --------------- Penicillamine -bound Membrane\ 40% 40% 1.7 ± 1.45\* 2.0 ± 1.6\* 228.6 ± 19.97 331.7 ± 15.17 (E) Non-Penicillamine -bound Membrane\ 80% 92% 7.1 ± 1.3 8.91 ± 2.32 259.4 ± 18.32 376.4 ± 23.43 (F) --------------------------------------------------------------------------------------------------------------------------------------------------------------------- (\* means p \< 0.05, compared to non-penicillamine -bound membrane- treated group) ::: Penicillamine and penicillamine-bound membrane did not influence the incision healing, although they insignificantly decreased the breaking strength of incision ---------------------------------------------------------------------------------------------------------------------------------------------------------------- Incision healing occurred very well in all groups. Compared to postsurgical day 7, the breaking strength of incision of each group was higher at postsurgical day 14. Compared to group A, the breaking strength of incision in group D, E and F were lower at postsurgical day 7 or day 14. The possibility exists that tensile strength of the abdominal wound might have been more affected if the membrane had been placed immediately deep to the laparotomy incision. Penicillamine and penicillamine-bound membrane did not influence the incision healing, although they insignificantly decreased the breaking strength of incision (Table [1](#T1){ref-type="table"}, [3](#T3){ref-type="table"}) (P \> 0.05). Discussion ========== Postsurgical abdominal adhesions have a great impact on the quality of life of millions of people worldwide. Small bowel obstruction and others complications of adhesions are serious, causing not only morbidity but also mortality \[[@B8],[@B21]\]. Adhesions are non-anatomic connections of fibrous tissue within normal peritoneal surfaces \[[@B7]\]. It may have a potential benefit, including neovascularization of ischaemic structures such anastomoses, but it is also responsible for various clinical problems \[[@B22]\]. The abdominal formation of fibrin is a common pathophysiological pathway for adhesions. Fibrin is formed after peritoneal injury, which can cause fibrinous adhesion. If the fibrinolytic system, which results in lysis of abdominal fibrin, is not activated, the adhesions will become fibrous \[[@B23]\]. This can be explained when the equilibrium between coagulation and fibrinolysis is disturbed \[[@B24]-[@B26]\]. Our present study confirmed this by the evidence of more collagen I fibers observed in the abdominal adhesions animal model than in the vehicle group. Apart from the formation of fibrin, a complex interaction of biochemical components, including inflammation, fibrinolysis and wound healing, is involved in the pathological process of abdominal adhesions \[[@B27]\]. For instance, initially the deposition of fibrin is regulated and maintained by growth factors and cytokines \[[@B28]\]. After the first week and up to a month, the matrix is remodeled and replaced by persistent proteins, such as collagen, and revascularization occurs. Fibrinolysis stimulators, such as tissue plasminogen factor (t-PA), and urokinase and fibrinolysis inhibitors, such as plasminogen activator inhibitor type I (PAI-1), transforming growth factor (TGF)-β, a key molecular mediator of pathological fibrosis, have also been shown to play a role in adhesions pathogenesis \[[@B29],[@B30]\]. The interrelationship between all the factors remains largely unknown, therefore, identifying the effective treatment or prevention for abdominal adhesions remains a big challenge. Numerous approaches have been used to prevent adhesions \[[@B15]\]. The three main principle pathways are: (1) decreasing the trauma to the peritoneum; (2) medical intervention in the fibrin formation/degradation balance, and (3) barriers (including fluid barrier and membranes) preventing organs from bridging over to other structures in the abdomen and thereby forming adhesions. Unfortunately none of these measures have proven uniformly effective under all surgical conditions. Barrier products, including hyaluronic acid-carboxymethyl cellulose membrane have been the most clinically successful in reducing adhesions formation by preventing the close apposition of injured tissues. However, treatment with these products induced many side effects, such as postponing wound healing. Furthermore, many treated models have a high standard deviation, which makes the relevance of results with only moderate effects questionable. Penicillamine can decrease the permeability of vessels by inhibiting aggregation of platelets, stabilizing lysosome and inhibiting releasing of lysomal enzymes. It can also attenuate immune reaction and decrease blood disk effusion and fibrin deposition by inhibiting generation of IgG and IgM and decreasing antigen-antibody complex in blood-serum, which blocks the first stage of abdominal adhesions. It is reported that D-penicillamine administration markedly reduces severe adhesions band formation without severe side effects \[[@B18]\]. Therefore, we hypothesized, based on these results, that a combination of penicillamine and barrier products may be a better treatment for adhesions. In the present study, we developed a novel penicillamine- bound membrane, which used hyaluronic acid as the ideal substrate, and then found that both penicillamine and penicillamine-bound membrane have better therapeutic effects on preventing abdominal adhesions than Dextran (Dex), sodium hyaluronate (SH) and non-penicillamine-bound membrane. Both of them did not affect wound healing. Although they decreased the breaking strength of incision insignificantly at postsurgical day 7 but, this decrease was ameliorated at postsurgical day 14. Penicillamine-bound membrane showed greater benefits than penicillamine itself in preventive effects and local administration. Recent studies investigated that penicillamine can inhibit blood vessel hyperplasia\[31\], which plays an important role in adhesions generation, by inhibiting the proliferation of endangium and smooth muscle cell, and this was confirmed by our results. Conclusions =========== The present research indicated that penicillamine-bound membrane can be applied as an effective therapeutic intervention for abdominal adhesion with inconsequential side effects, but further studies on the detailed mechanisms for treating abdominal adhesion are still warranted. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= Q-YZ carried out the preparation of penicillamine-bound membrane, created the animal model of abdominal adhesions and drafted the manuscript. SM carried out the tissue preparation and measurement of adhesions degree and breaking strength of incision. DX carried out the Immunohistochemistry. W-TZ participated in the design of the study and performed the statistical analysis. A-WL conceived the study, and participated in its design and coordination. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2482/11/5/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Test for breaking strength of incision**. Step 1. Connecting the incision with an empty water container. Step 2. Draining the water gradually into the container until the incision is broken, then calculating the breaking strength by this formula (Breaking strength = the gravity of total water). ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We would like to thank The Key Laboratory of Cardiovascular Remodeling and Function Research, Shandong University, China for technical support. We also thank Dr. Melissa Snyder from Drexel University College of Medicine, PA, US and Dr. Bassil Kublaoui, Mr. Nilay Gandhi from The Children\'s Hospital of Philadelphia, PA, US for helping us copyedit this manuscript. This work was supported by the Natural Science Foundation of Shandong Province (Y2003C18), China.

PubMed Central

2024-06-05T04:04:19.443236

2011-2-25

PMC3053216

Background ========== Healing of full-thickness injury to the gastrointestinal tract remains an unresolved topic. It begins with a surgical reapposition of the bowel ends, which is most often the initial step in the repair process. Failure of healing results in dehiscence, leaks, and fistulas, which carry significant morbidity and mortality. The myofibroblasts play a central role in the process of wound healing \[[@B1]-[@B5]\]. They contain smooth muscle myosin isoforms in addition to α-smooth muscle actin (α-SMA), the requisite machinery for contraction and/or motility, respond to proinflammatory cytokines with elaboration of matrix proteins and additional growth factors and then disappear by apoptosis following repair or scar formation \[[@B6]-[@B11]\]. The purpose of this study is to assess the events of colonic anastomosis healing with an emphasis in the role of myofibroblasts. Methods ======= Experimental Design ------------------- The study protocol was approved by the Research Ethics Committee of the Aristotle University of Thessaloniki and conducted in accordance with the Declaration of Helsinki. One hundred male Wistar rats weighing 274 ± 9,1 g (mean age: 3.5 months) were used, which is the least acceptable number for statistical comparison of the groups. The rats were housed two per cage in a standard animal room in the animal laboratory of AHEPA University Hospital, and were allowed free access to food and water before experimentation, so that the integrity of the bowel mucosa was maintained. Rats were anesthetized by intramuscular administration of midazolam (2 mg/kg) (Dormicum Roche K Pharma) and fentanyl (300 μg/kg) (FENTANYL/JANSSEN). Laparotomy was performed through a midline 4 cm incision. A left colonic segment, 1 cm in length, 4 cm proximal to the peritoneal reflection was transected. The colon was re-anastomosed end-to-end using 7-0 propylene sutures (Prolene ETHICON) in single-layer interrupted fashion. Propylene was used as a material because it is a monofilament, nonabsorbable suture. Its advantages include high tensile strength, minimal tissue reactivity, and slipperiness (allowing easy removal from tissues). To facilitate the suturing, an intraluminal metal tube was inserted transanally at the level of the anastomosis offering absolute apposition of the cut ends. About 10-12 sutures were placed symmetrically for each anastomosis to secure an inverted anastomosis without mucosal protrusion, which is a major cause of perianastomotic adhesions. By this process the operative time was reduced, given that the placement of sutures was expedited and accelerated, while the risk of catching the mucosa of the opposite side was nullified. What is more, the technical uniformity and perfection was ensured, so that the factor of operative technique affected the healing of anastomoses equally. The tube was removed transanally immediately after the suturing was completed. The abdominal muscle wall was then closed with 4-0 silk sutures, followed by skin closure with 4-0 silk sutures (Mersilk ETHICON). Animals were randomly divided into two groups. The first group experimental animals (n = 50) were sacrificed on postoperative day 3, while the second group rats (n = 50) were sacrificed on postoperative day 7 with an overdose of ether. The previous abdominal incision was reopened, and the anastomotic site identified and inspected for possible adhesions and leakage. A 5 cm segment of the colon with the anastomosis in the middle was resected. Care was taken not to detach adhesions from the anastomosis, but to dissect the surrounding tissues. The resected specimen was gently irrigated with saline to remove feces and was mounted on a table. Healing of colonic anastomosis was studied in terms of anastomotic bursting pressure, as well as myofibroblastic reaction and expression of α-SMA, adhesion formation, inflammatory reaction and neovascularization. The anastomotic bursting pressure was studied by applying an internal hydraulic pressure (ANNE, Anesthesia Infuser, ABBOTT, USA). As a result, a stress state equivalent to the so-called biaxial tension appeared on the surface of the concave cylindrical specimen (reduced to mmHg, Monitor Minimon 7132 Kontron Instruments, Ltd, England). The level of adhesion formation was determined according to the Van der Ham scale score \[[@B12]\]. Histopathological and immunohistochemical assessment ---------------------------------------------------- A 2 cm-long segment of colon, including the anastomotic site, was resected from each animal, was fixed in 10% buffered formalin for 48 hours, and was submitted for histological examination. Four-cm thick paraffin sections were prepared from each paraffin block using a standard microtome. Sections from each block were stained with hematoxylin-eosin (H&E) and with the streptavidin-biotin immunohistochemical procedure for alpha-smooth muscle actin (α-SMA). The antiserum was obtained from Dakopatts, Glostrup, Denmark. In the bowel, α-SMA is expressed in smooth muscle fibers of the muscularis mucosae and the muscularis propria, as well as smooth muscle fibers and pericytes of vascular walls. In areas of wound healing, myofibroblasts also show α-SMA immunoreactivity. The histological and immunohistochemical assessment were performed by a \"blinded\" observer, a pathologist who was unaware of the study groups. The following histological findings were assessed semiquantitatively as mild = +, moderate = ++ or severe = +++: density of inflammatory reaction, intensity of myofibroblastic reaction and density of neovascularization. The density of the inflammatory infiltrates, as well as the density of the new vessels were assessed in H&E-stained sections. The intensity of the myofibroblastic reaction was evaluated in immunohistochemical stains for α-SMA as follows: +: loosely arranged myofibroblasts; ++: moderately dense myofibroblasts; +++: densely arranged myofibroblasts. All examinations were performed at an Olympus BH2 microscope, at magnifications of ×100 and ×400. Statistical Analyses -------------------- All data were recorded using Statistical Package for the Social Sciences (SPSS) 16 for Windows. An independent-samples t-test was conducted to compare the bursting pressure for the two groups of experimental animals. The Chi-square test for independence was used to find out whether there was a relationship between group of rats and adhesion formation, inflammatory reaction, myofibroblastic reaction and neovascularization. *P*values below 0.05 were considered significant. Results ======= The mean anastomotic bursting pressure increased from 20.6 ± 3,5 mmHg on the 3^rd^postoperative day to 148.8 ± 9.6 mmHg on the 7^th^postoperative day. There was statistically significant difference in bursting pressure between the two groups of rats (p \< 0.001). The magnitude of the differences in the means was very large (eta squared = 0.987). Interestingly, wound rupture was located at the anastomotic site, when rats were sacrificed on the 3^rd^postoperative day. On the other hand, wound breakdown took place at the wound margin, when animals were sacrificed on the 7^th^postoperative day. Adhesion formation according to Van der Ham scale score is illustrated in Table [1](#T1){ref-type="table"}. There was a statistically significant relationship between group of rats and adhesion formation (p = 0.007). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### van der Ham scale for adhesion formation score ::: Adhesion formation Total ----------- ------------- ------------- -------------------- ------- -------- ------- **group** **3rd day** Count 5 30 15 50 \% of Total 5.0% 30.0% 15.0% 50.0% **7th day** Count 5 15 30 50 \% of Total 5.0% 15.0% 30.0% 50.0% Total Count 10 45 45 100 \% of Total 10.0% 45.0% 45.0% 100.0% Van der Ham adhesion scoring system: \- No adhesion \+ Adhesion towards the anastomosis line from the omentum ++ Adhesion of bowel and omentum towards the anastomosis line +++ Extensive adhesions towards the anastomosis line and contemporary presence of abscess ::: The granulation tissue at the anastomotic area on the 3^rd^postoperative day was rich in polymorphonuclear leukocytes, lymphocytes and histiocytes. The stroma was edematous and contained many capillary vessels. The intensity of the inflammatory reaction detected was less on the 7^th^postoperative day, as compared to the 3^rd^postoperative day (Table [2](#T2){ref-type="table"}). On the 7^th^postoperative day, lymphocytes and histiocytes were the predominant cells in the granulation tissue, which also contained prominent myofibroblasts. The Chi-Square test showed that there was no statistically significant relationship between group of rats and inflammatory reaction (p = 0.356). The neovascularization was less intense on the 7^th^postoperative day as compared with the 3^rd^postoperative day (Table [3](#T3){ref-type="table"}). There was no statistically significant relationship between group of rats and neovascularization (p = 0.108). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Inflammatory reaction at the anastomosis site. ::: Inflammatory reaction Total ----------- ------------- ------------- ----------------------- ------- -------- **Group** **3rd day** Count 10 40 50 \% of Total 10.0% 40.0% 50.0% **7th day** Count 15 35 50 \% of Total 15.0% 35.0% 50.0% Total Count 25 75 100 \% of Total 25.0% 75.0% 100.0% ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Neovascularization at the anastomosis site. ::: neovascularization Total ----------- ------------- ------------- -------------------- ------- ------- -------- **group** **3rd day** Count 5 40 5 50 \% of Total 5.0% 40.0% 5.0% 50.0% **7th day** Count 13 32 5 50 \% of Total 13.0% 32.0% 5.0% 50.0% Total Count 18 72 10 100 \% of Total 18.0% 72.0% 10.0% 100.0% ::: The myofibroblastic reaction was more profound on the 7^th^postoperative day in comparison with the 3^rd^postoperative day (Table [4](#T4){ref-type="table"}). There was a statistically significant relationship between group of rats and myofibroblastic reaction (p \< 0,01). In addition to myofibroblasts, α-SMA was expressed in pericytic and smooth muscle cells of old and newly formed vessels, as well as muscle fibers of the muscularis mucosae and the muscularis propria. The staining intensity of the myofibroblasts increased from the 3rd to the 7th postoperative day. On the 7^th^day, the α-SMA content of the myofibroblasts reached the level of the muscular layer cells (Figures [1](#F1){ref-type="fig"}, [2](#F2){ref-type="fig"}, [3](#F3){ref-type="fig"}, [4](#F4){ref-type="fig"}). Anastomotic leakage or other complications that might correlate to the histological and immunohistological findings were not observed. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Myofibroblastic reaction at the anastomosis site. ::: myofibroblasts Total ----------- ------------- ------------- ---------------- ------- ------- -------- **group** **3rd day** Count 35 15 0 50 \% of Total 35.0% 15.0% .0% 50.0% **7th day** Count 15 10 25 50 \% of Total 15.0% 10.0% 25.0% 50.0% Total Count 50 25 25 100 \% of Total 50.0% 25.0% 25.0% 100.0% ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Expression of α-SMA in anastomotic sites: The antigen is expressed in myofibroblasts, pericytes and smooth muscle cells (3^rd^postoperative day)**. The staining intensity of myofibroblasts is weaker than that of muscularis propria fibers (immunohistochemical stains for α-SMA, ×400). :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Bigger magnification of the anastomotic site (3^rd^postoperative day) (immunohistochemical stains for α-SMA, ×400)**. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **On the 7^th^postoperative day, the α-SMA content of the myofibroblasts is similar to that of muscular layer cells (immunohistochemical stains for α-SMA, ×400)**. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Bigger magnification of the anastomotic site (7^th^postoperative day) (immunohistochemical stains for α-SMA, ×400)**. :::  ::: Discussion ========== Dehiscence of colonic anastomosis is a common, serious and potentially life-threatening complication after colorectal operation. The early integrity of the anastomosis is dependent on formation of a fibrin seal on the serosal side, which achieves watertightness, and on the suture-holding capacity of the intestinal wall, particularly the submucosal layer \[[@B13]\]. A key event in the process of wound repair is contraction of the gastrointestinal lamina propria \[[@B4],[@B6],[@B14]-[@B17]\]. Ultimate anastomotic strength in the gastrointestinal tract is not always related to the absolute amount of collagen, and the structure and arrangement of the collagen matrix may be more important \[[@B18]\]. Myofibroblasts appear to be key cells in the process of wound healing. They are the major responsible cells for contraction and are also involved in the formation and repair of the extracellular matrix and proliferation and differentiation of epithelial, vascular and neurogenic elements \[[@B13],[@B19]-[@B22]\]. The integrity of the anastomosis represents equilibrium between collagen lysis, which occurs early in the healing process, and collagen synthesis, which takes a few days to initiate. During the first 3 to 5 days collagen breakdown far exceeds collagen synthesis. There is a significant decrease in marginal strength during the first week due to an early and marked collagenolysis \[[@B13],[@B18]\]. Local infection, which often occurs near colonic anastomoses, promotes lysis and delays synthesis, thus increasing the likelihood of perforation. The danger of leakage is most from the fourth to seventh days, when tensile strength normally would rise rapidly but is prevented from doing so by increased lysis or compromised collagen deposition \[[@B23]\]. Leakage is about as likely to occur a few millimeters from the anastomosis as it is in the anastomosis itself. By 1 week intestinal anastomosIs resists bursting more strongly than the more normal surrounding intestine. This has been attributed to the fact that the surrounding intestine also participates in the reaction to injury and loses a large part of its collagen by lysis \[[@B23]\]. However, one could argue that although collagen lysis occurs both at the anastomosis and the surrounding intestine, the wound margin is intact, while the anastomotic site is cut and forming anew. Our study demonstrated that wound rupture was located at the anastomotic site on the 3^rd^postoperative day while wound breakdown took place at the wound margin on the 7^th^postoperative day. The immunohistochemical findings our study expand those of Darby et al. \[[@B6]\] on experimental wound healing. Darby et al examined skin wounds in rats by immunofluoresence and found that α-SMA started to become detectable in myofibroblasts on day 6, and then was increasingly expressed for the following 15 days of wound healing \[[@B6]\]. On the other hand, the contemporary immunohistochemical procedure that was utilized in our study showed that α-SMA immunoreactive myofibroblasts were evident at the anastomosis already on the 3rd postoperative day. By the 7th day, the number of cells with positive α-SMA expression increased. Actually, the immunohistochemical study of specimens showed that the content of myofibroblasts in α-SMA was remarkably increased by the 7^th^day, in fact, reaching the level of the muscular layer cells. This finding provides an explanation for the reduction in the incidence of wound dehiscence after the 7th postoperative day. Moreover, it could explain why by 1 week the proportion of wound dehiscence that occurs at the anastomotic site is not expectedly more than that at the wound margin. Our results are supported by the fact that wound contraction begins immediately after injury. This had been considered to be a puzzling point, since it did not correspond to the day (6^th^) myofibroblasts were identified in the wound. It had been attributed to the in vitro findings of Ehrlich that fibroblasts placed in a collagen lattice actively move in the lattice and contract it without expressing stress fibers, postulating that the movement of cells with concomitant reorganization of the cytoskeleton is responsible for contraction \[[@B6],[@B13],[@B24]\]. Conclusions =========== The potential to minimize failure of wound healing depends on the surgeon\'s knowledge of the events responsible for this phenomenon. The results of our study emphasize the pivotal role of myofibroblasts in the process of colonic anastomosis healing and provide an explanation for the reduction in the incidence of wound dehiscence after the 7th postoperative day. Further understanding of the molecular mechanisms of normal and pathologic wound healing may provide valuable insights into future therapies that can control dehiscence, leaks, and fistulas. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= CK carried out the experimental studies and drafted the manuscript, made substantial contributions to conception and design, acquisition of data, analysis and interpretation of data. CE participated in the experimental studies. GA performed the statistical analysis involved in revising critically the manuscript for important intellectual content. GB participated in the design of the study and made substantial contributions to interpretation of data. SA made substantial contributions to acquisition of data and the study of the anastomotic bursting pressure. PH performed the histological and immunohistochemical assessment involved in revising critically the manuscript for important intellectual content. KV made substantial contributions to collecting references, designing and revising data of the paper and resubmitting it. JP conceived of the study, and participated in its design and coordination. EF conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2482/11/6/prepub> Acknowledgements ================ There are no acknowledgments to be mentioned.

PubMed Central

2024-06-05T04:04:19.445404

2011-3-2

PMC3053217

Background ========== Nonbacterial thrombotic endocarditis (NBTE) is a rare condition associated with cancer and other illnesses with hypercoagulable states, including septicemia and autoimmune disease \[[@B1]\]. It causes aseptic masses of fibrin and platelets usually along the line of previously undamaged heart valves. The vegetations in NBTE are friable and tend to detach and cause extensive infarction more readily than the vegetations observed in infective endocarditis. Nearly half of patients present clinically with systemic emboli, with cerebral emboli occurring most commonly \[[@B2]\]. A diagnosis of NBTE is most commonly made at autopsy and is rarely made during life. This report concerns a case of NBTE associated with cancer complicated with thrombus in the left auricular appendage and multiple cerebral infarctions. Case presentation ================= A 63-year-old man was admitted to our hospital with a complaint of right lateroabdominal pain. He had a history including hypertension, hyperlipidemia and chronic rheumatoid arthritis. Physical examination found his blood pressure was 130/74 mmHg and pulse was regular at 72 beats/min. No heart murmur was heard and the lungs were clear on auscultation. Chest X-ray and 12-lead electrocardiography findings were normal. Laboratory test results showed D-dimer of 22.8 μg/ml and serum carcinoembryonic antigen of 1073 ng/ml. Colon contrast enema and colonoscopy revealed the stenosis of the rectum and transverse colon. Endoscopic ultrasonography disclosed that a tumor in the muscular layer of the rectum had invaded from the peritoneal cavity, but that the mucosa and submucous layer were intact. These findings led to a diagnosed of metastatic colon cancer. Seven days after admission, aphasia occurred, and magnetic resonance imaging of the brain confirmed the presence of multiple embolic infarctions. Transesophageal echocardiography to evaluate the neurological symptoms indicated that two mobile, echo-dense masses measuring 3 × 7 mm and 6 × 4 mm, were attached to the anterior and posterior mitral valve leaflet, respectively (Figure [1](#F1){ref-type="fig"}, [2](#F2){ref-type="fig"} and [3](#F3){ref-type="fig"}, Additional file [1](#S1){ref-type="supplementary-material"}; Video 1). However, no significant mitral stenosis or regurgitation was observed (Additional file [2](#S2){ref-type="supplementary-material"}; Video 2). Furthermore, a thrombus (12 × 8 mm) was detected in the left auricular appendage (Figure [4](#F4){ref-type="fig"}, Additional file [3](#S3){ref-type="supplementary-material"}; Video 3). Other sites of thomboembolic dissemination were not documented. The patient was therefore suspected of having infectious endocarditis with subsequent multiple brain infarctions. However, we also considered the possibility of aseptic endocarditis because serial blood cultures were sterile and there were no clinical signs of infection. Anticoagulation therapy with heparin was administrated, but the patient\'s general condition worsened and he died 13 days after admission. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Masses of the mitral valve**. **Ao, aorta; LA, left atrium; LV, left ventricle**. Transesophageal echocardiography showed mobile, echo-dense masses measuring 3 × 7 mm and 6 × 4 mm attached, respectively, to the anterior (white arrow) and posterior (dotted white arrow) mitral valve leaflet. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Masses of the mitral valve**. All abbreviations and descriptions as in Figure 1. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Masses of the mitral valve**. All abbreviations and descriptions as in Figure 1. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **A thrombus in the left autricular appendage**. **LA, left atrium; LAA, left auricular appendage**. A thrombus (12 × 8 mm) was observed in the left auricular appendage (white arrow). :::  ::: The autopsy showed poorly differentiated cancer in multiple organs, including the serous side of the transverse colon, the sigmoid colon, rectum, pancreas head, bladder, mediastinum, peritoneum and epicardium. However, no primary tumor could be identified. The masses of the mitral valve were histologically composed mainly of fibrin without bacteria or oncocytes (Figure [5](#F5){ref-type="fig"}). This patient was therefore ultimately diagnosed with NBTE associated with cancer of unknown origin. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Pathological findings**. **(A)**Gross examination of the vegetation removed from the mitral valve leaflet. **(B)**Histological examination showed the vegetation is composed mainly of fibrin without bacteria or oncocytes (hematoxylin-eosin staining × 100). :::  ::: Discussion ========== The case presented here demonstrates an uncommon paraneoplastic disease that a patient with NBTE associated with cancer of unknown origin can be complicated with thrombus in the left auricular appendage and multiple cerebral infarctions. NBTE, characterized by the presence of platelet-fibrin deposits on cardiac valves is known to be associated with hypercoagulable states, particularly in patients with malignant tumors \[[@B1]\]. The aseptic masses on cardiac valves of NBTE are friable and tend to detach and cause extensive infarction more readily than the vegetations observed in infective endocarditis. Nearly half of all NBTE patients present clinically with systemic emboli, most commonly cerebral emboli \[[@B2]\]. All complications of NBTE lead to severe damage to multiple organs and result in a poor outcome. The fact that a diagnosis of NBTE is usually made at autopsy indicates that making a definitive antemorterm diagnosis is very difficult. Although the antemorterm diagnosis is typically made by means of echocardiography, it can be difficult to distinguish between infective endocarditis and NBTE on the basis of echocardiographic features alone. However, such a distinction is of paramount importance in terms of selecting of the most appropriate therapeutic strategy. Various characteristics of NBTE have been reported before. The most commonly affected valves are the aortic valve, the mitral valve, and a combination of the two valves \[[@B3]\]. A large-scale study reported a mitral valve predominance of 64%, followed by aortic valve (24%) and both valves (9%) \[[@B4]\]. Platelet-fibrin deposits are commonly located on the atrial surface of the mitral and tricuspid valves as well as on the ventricular surface of the aortic and pulmonic valves. Valvular vegetations along coaptation lines without leaflet destruction (regurgitation), simultaneous occurrences of venous thromboembolism, negative blood cultures, and absence of clinical signs of infection are all suggestive of NBTE \[[@B5]\]. In the present case, thrombus in the left auricular appendage was complicated. A left atrial thrombus is usually associated with atrial fibrillation or mitral stenosis but is very infrequently detected in the presence of sinus rhythm \[[@B6],[@B7]\]. In addition to above-mentioned characteristics of NBTE, the simultaneous presence of intracardiac thrombus may also be suggestive of NBTE. NBTE has been reported in 4% of patients with all types of end-stage cancer, and the aseptic masses on cardiac valves of NBTE are very friable and therefore embolize easily. NBTE must be taken into account for patients with a background of cancer, with cerebral embolism due to unknown causes, and with negative blood cultures and absence of clinical signs of infection \[[@B8]\]. Conclusions =========== In summary, NBTE is associated with hypercoagulable states, particularly in patients with malignant tumors, and it causes systemic embolism readily. NBTE must be taken into account for patients with a background of cancer, with cerebral embolism due to unknown causes, and with negative blood cultures and absence of clinical signs of infection. Abbreviations ============= NBTE: nonbacterial thrombotic endocarditis; Consent ======= Written informed consent was obtained from the kin of the patient for publication of this case report and any accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= KN designed the study, carried out subject recruitment, performed echocardiography, analysed the data, and wrote the manuscript. HT, TO, AK, TT, KY, KR KT, KM, NM, JS, YM, SH, HK and KH assisted recruitment and manuscript revision. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Video 1. Transesophageal echocardiographic view of the mitral valve**. Transesophageal echocardiography showed mobile, echo-dense masses measuring 3 × 7 mm and 6 × 4 mm attached, respectively, to the anterior and posterior mitral valve leaflet. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Video 2. Transesophageal color Doppler echocardiographic image of the mitral valve**. Transesophageal echocardiography showed that mitral regurgitation was trivial. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Video 3. Transesophageal echocardiographic view of the left auricular appendage**. A thrombus (12 × 8 mm) was observed in the left auricular appendage. ::: ::: {.caption} ###### Click here for file :::

PubMed Central

2024-06-05T04:04:19.447414

2011-2-28

PMC3053218

Background ========== In the age of genomics, the genetics of common chronic disorders, pharmacogenetics and large-scale applications in screening are becoming increasingly important. Primary care providers (e.g. general practitioners and midwives) will have to discuss these issues with their patients, who are becoming increasingly aware of genetic contributions to disease and also have high expectations of genetic testing \[[@B1]\]. Consequently, primary care providers need to be educated to meet the needs of their patients that are created by rapid advances in genomics \[[@B2]\]. Genetics literacy among primary care providers needs to be improved to enable their participation in the debate on the hopes and hypes of genomic medicine and to distinguish between useful and useless practical applications in health care \[[@B3]\]. Currently, genetics and genomics are rather underrepresented in postgraduate (physician) training programmes in general practice (here, the terms *general practice*and *family medicine*(commonly used terms in the Dutch health care system) are considered synonymous to the more commonly used term *family practice*in the U.S. healthcare system) as well as in master programmes in midwifery and public health \[[@B4],[@B5]\]. It is widely recognised that medical professionals and medical students should be educated about genetics \[[@B2],[@B5]-[@B7]\]. Research into the perspectives of general practitioners and midwives on the educational priorities and attitudes in relation to genetics \[[@B8]-[@B16]\] revealed a need for genetics education for primary care providers in areas like psychosocial issues and screening, assessment of the risk of genetic malformations and basic genetics. However, the educational needs of primary care providers and their views on the role of genetics in family practice are still under investigation, and international efforts to translate these needs into education programmes are still in their early stages \[[@B17],[@B18]\]. Primary care providers have a unique role in the Dutch health care system and general practitioners are easily accessible to all patients for any complaint, request or question. Midwives provide obstetric and perinatal care and give advice and guidance to patients on pregnancy and childbirth. Genetics could have an effect on daily primary care practice if basic and clinical science advances in genomics of common chronic diseases in practice and midwifery care are successfully translated \[[@B2],[@B3],[@B19]\]. Changes will only be effective, however, if they fit well into practice routines. It is therefore important to understand how these key professional groups conceive of their responsibilities and experiences in relation to genetics. For the implementation of genetics training in daily practice to be successful, it is important to identify factors that can enhance or inhibit effective genetic primary care. Up till now, most studies of educational needs concerning genetics have been limited to the perspectives of target groups \[[@B1],[@B4]-[@B14]\]. Professional training needs can be derived from those studies and from challenges posed by new applications as foreseen by experts. This study explored the views of general practitioners, midwives, patient advocacy groups and others involved in genetics in health care and education regarding their need for genetics education and the role of genetics in primary care. This information was collected to help develop effective genetics education and training as well as effective integration of genetics in primary care. Method ====== Design ------ We used a qualitative study design with focus groups because this enables the exploration of the meaning and significance of the role of genetics and the need for education in that area as perceived by different stakeholders. A discussion with members of the research team and primary care providers revealed that general practitioners and midwives were the primary care providers most likely to be confronted with issues of genetics and genomics in their practices. Participants ------------ We used purposive sampling to recruit specific groups of professionals for focus group interviews in order to obtain rich, relevant and diverse data. The participants were expected to provide complete and possibly complementary perspectives on genetics in primary care practice and education. Potential participants were named by key persons and network contacts at academic departments of general practice and the midwifery academies at Amsterdam and Maastricht, the Netherlands. We convened three types of focus groups, (1) two groups of general practitioners, (2) two groups of midwives, and (3) three multidisciplinary groups composed of clinical genetics professionals (clinical geneticists or genetic counsellors), primary care educators, and representatives of patient advocacy groups. The participants in the multidisciplinary groups were considered experts who were expected to have a broad view of the role of genetics in primary care and the need for genetics education i.e. what is needed, what works and what does not work. For each focus group, ten to fifteen professionals from one region were invited by email or telephone. Those who responded positively received an invitational letter, informing them that the purpose of the study was to explore their perceptions of the role of genetics and genomics in primary care and the related education needs. The term \'genetics\' was commonly used during the focus group discussions, and the term \'genomics\' was used to denote a broad definition (e.g. common complex disorders and rapid technological developments). Focus groups ------------ Additional file [1](#S1){ref-type="supplementary-material"} provides an overview of the participants. The interviews lasted approximately two hours and were held between March and August 2009. The discussions were facilitated by an independent and experienced moderator (SL), who encouraged the participants to participate actively and to openly state their viewpoints and engage in discussion. An assistant (IH) took notes and all the sessions were attended by one observer (LH). Participation was voluntary and participants received €100 plus travel expenses. The study was approved by the Medical Ethics Committee of VU University Medical Center, Amsterdam and Maastricht University. All participants gave informed consent at the start of their focus group session. Interview guide --------------- An interview guide with open-ended questions was developed to ensure coverage of the major topics (Additional file [2](#S2){ref-type="supplementary-material"}). The moderator opened each focus group interview with an introduction and a round-robin open question: \"What are your experiences with genetics/genomics in primary care?\" Further probing by the moderator was rarely required, since each group spontaneously talked about genetics education needs and the role of genetics in primary care. When needed, asking for clarification of the answers was sufficient to elicit ample additional information. Data analysis ------------- The focus groups were audio recorded and transcribed verbatim. Member checking entailed sending a summary of the sessions to all participants and inviting their comments, which were then incorporated in the transcripts. Using Atlas.ti5.2 for data analysis, two of the researchers (IH and LH) independently coded the themes that emerged from the transcripts. They compared their coding for reliability and reached consensus on differences through discussion. Through a process of discussion and deliberation, connections between the codes were identified and categories and themes developed. The transcripts were read repeatedly to check the accuracy and completeness of the themes and subthemes. In the results section, representative quotes from the focus groups, translated from the Dutch, are presented to illustrate the themes. Results ======= All participants acknowledged the importance of genetics education for primary care providers. A need for education was expressed more urgently by the general practitioners than by the midwives, while members of the multidisciplinary group generally indicated that both groups were deficient in genetics knowledge and skills. The extent and focus of the discussions differed between the general practitioners and the midwives, because the general practitioners saw themselves as generalists while perinatal care was the primary focus for the midwives. This difference was reflected in their priorities for education. Four distinct themes emerged: (1) the need for genetics knowledge, (2) taking a family history, (3) ethical dilemmas and psychosocial effects related to genetics, and (4) insight into the organisation and role of clinical genetics services (Additional file [3](#S3){ref-type="supplementary-material"}). 1\. The need for genetics knowledge General practitioners perceived deficiencies in their basic understanding of genetics, since they had never been taught this or because their knowledge had faded. They experienced their lack of knowledge as a barrier to the use of genetics in diagnosis, treatment and in consulting clinical geneticists, and they expressed a strong need for this knowledge, as reflected in the words of one general practitioner (male, 37 years): \"I think it\'s essential to know the basics \[of genetics\]\.... You should know what a gene is, that there are deviant genes and that genes can be turned on and off\... These are basics that clarify genetics and make it understandable so that it can be translated to patients in relation to diagnosis or treatment or to advise them to refrain from something, smoking, for example.\" Other participants acknowledged this need and argued that general practitioners should have more knowledge in order to provide general genetic information. Primary care providers wondered how much they really needed to know about such a complex field. All groups shared the view that primary care providers cannot be expected to know everything. Since many genetic diseases are rarely seen in primary care, there is no urgent need for them to be included in training programmes. General practitioners appeared to be generally aware of their lack of knowledge, which made them ill equipped to identify genetic problems in their patients. In response to the question about experiences with genetics/genomics, one general practitioner (male, 52 years) said: \"I desperately hope the midwife or obstetrician will think of \[prenatal diagnosis for women of advanced maternal age\], because I don\'t always think of bringing it up. Also, I tend to think the specialist will consider all the genetic aspects of a clinical problem, but this often is not the case.\" Midwives did not perceive a lack of genetics knowledge and said the midwifery master programme provided sufficient education on this topic. Multidisciplinary group members, however, said that both general practitioners and midwives needed education to increase their knowledge, as they observed a lack of knowledge in both groups. There was a discrepancy in the content of the required knowledge as perceived by the participants. General practitioners mentioned a need for knowledge about prevention of common genetic diseases, whereas the midwives were primarily interested in prevention of perinatal diseases. To meet these educational needs, general practitioners and multidisciplinary experts believed that genetics education should address family history and inheritance patterns. Both general practitioners and midwives were interested to know where they could find more genetic information. There appeared to be a general need for easily accessible sources of information such as web-based education or websites with short and easy to understand information that could be applied in daily practice. 2\. Taking a family history Primary care providers believed that taking a family history was extremely important as it allows for familial risk stratification and identification of hereditary conditions. For midwives the importance of family history was limited to perinatal disease. The need to increase awareness of familial diseases in primary care was discussed by a midwifery educator (female, medical doctor, 56 years): \"We agree on when to think of some familial diseases. I mean, it\'s clear when speaking of colon carcinoma or breast cancer that you realise \[as a general practitioner\] that it can be inherited. There should be a clinical guideline to help one decide when to consult a clinical geneticist for further diagnosis. Today, more general practitioners should realise this \[colon carcinoma or breast cancer\] could be familial.\" An educator, allied to a postgraduate general practice programme, thought that taking a family history was well covered during postgraduate training, but the general practitioners were not quite so sure. Although family history was part of some consultations, most general practitioners said that family history and pedigree drawing were not part of their daily routine. The midwives said that family history was something they did every day, but they lacked the skills for pedigree drawing. They thought this was not important, however, because they did not do it often enough to maintain this skill. They thought the main thing was to be able to recognise high-risk factors in a family history \"Midwives generally lack sufficient knowledge to draw a pedigree. I think the same applies for general practitioners, but it is more important for them \[general practitioners\] to detect high-risk family history criteria. How and when is something important? The signalling function, that is important and it should be taught. Counselling is important for both general practitioners and midwives. But above all it is important to be clear on when something is important, how important it is for you, what you want to do about it, and once this is clear you can take the next step\" (a female GP and midwifery educator, 39 years). General practitioners were uncertain about recording information from family history in the electronic patient record. Participants said that measures should be taken to improve the electronic patient record to include information from family history. 3\. Ethical dilemmas and psychosocial effects related to genetics Most primary care providers expressed concern about the surge of genetic testing which confronted them with ethical dilemmas and more profound psychosocial effects of genetics in their daily practice. One participant said he was faced with \"an increasing amount of vague, worrying and inexplicable genetic information\". He referred to information about genetic risks provided to consumers by commercially available Personal Genome Services. Primary care providers felt unqualified to deal with these issues and thought that genetic ethical dilemmas should be part of genetics education. General practitioners also wondered whether it was beneficial to their patients and themselves to know everything about a patient\'s genetic background. They voiced concern about the possibility of unauthorised dissemination of genetic information and related privacy issues if it were to become obligatory for them to take a family history and record the information derived from it. They asked: \"Who should you inform about genetic information and who should you not inform if you want to keep your patient\'s best interest at heart, for example when a patient wants to take out life insurance or needs a mortgage to buy a new house?\" Midwives thought that developments in genetics were moving at a very rapid pace, giving rise to feelings of insecurity both in midwives and their patients. They discussed whether following the protocol for perinatal screening might be inappropriate and therefore not uniformly applicable. They preferred to adapt the application of genetic protocols to individual patients, because test results could have important genetic and personal consequences. One midwife, (female, 40 years) explained: \"Surely because this child is already in the uterus, the basic question is rather what do you really want to know? Because what are you going to do with this \[genetic\] information?\" Clear guidelines as to when a general practitioner should be proactive and bring up the subject of familial disease to patients and their families were non-existent but considered necessary. General practitioners were unclear about how to guide patients in their decisions around prenatal and genetic testing. As a result they perceived non-directive counselling as difficult because it was influenced by their personal opinions and sense of urgency. One general practitioner (male, 52 years) explained: \"I used to discuss prenatal screening with patients at first when it was a hot item. Everybody wanted the triple test, but this seems to change when you explain that it gives a probability that doesn\'t offer any certainty and things can also happen after the baby is born or you cannot always see things on the outside and people then pull back automatically. I try to counsel nondirectively, because it promotes shared responsibility. I like to hold on to this nice shared responsibility, because it doesn\'t make me feel that I am solely accountable.\" General practitioners and midwives alike mentioned consanguinity as a complex issue to discuss with patients and they raised the problem of how to deal with a potentially increased risk of congenital disease. One general practitioner (female, 52 years) said: \"When cousins get married, they are blissfully happy when presenting this news to their doctor. As a general practitioner I find it complicated, when there is a disease in the family, to confront these people with this problem in today\'s society.\" Inter-cultural differences were considered a source of difficulties in discussing prenatal or preconceptional screening. Midwives sensed urgency early on in pregnancy when recommending this type of screening to members of ethnic minorities in order to prevent congenital disease. This feeling of urgency was sometimes enhanced by language barriers and time constraints. 4\. Insight into the organisation and role of clinical genetics services General practitioners expressed a need for education with regard to indications for referral to clinical genetics services. Some general practitioners preferred to first gather information from other sources (such as online websites) before turning to a clinical geneticist. Other general practitioners said it was better to refer than to do it all yourself. Midwives, on the other hand, said it was easy for them to consult clinical geneticists, whom they regularly telephoned for advice. Some general practitioners and midwives said it was not clear to them what clinical geneticists do and what the clinical trajectory would be once a patient was referred to such a service. General practitioners mentioned their lack of familiarity with this type of service as a cause of inappropriate consultation strategies, which could result in untimely referrals. Geneticists argued that general practitioners should consult them more often, and that this should be stimulated by education. \"The relation between primary and secondary care is sometimes difficult. Once the clinical geneticist has diagnosed a certain genetic disease, which means more specialised information, the patient\'s family members find themselves in a pickle together with the GP and then the whole process (of consulting) starts all over again. \[\...\] I hope education of primary care providers will result in accessible consultation services. Because I don\'t think the general practitioner should know everything about everything \... but they should at least know that help is easily available.\" (Clinical geneticist, female, 45 years). The role of genetics in primary care ------------------------------------ The role of genetics in primary care was perceived to be unduly limited as a result of care providers\' inadequate genetics knowledge and skills. Although care providers might show some interest in improving their knowledge, representatives of patient advocacy groups indicated that primary care providers were \"not sufficiently proactive\" in this area. They perceived an urgent need for inclusion of genetics in primary care guidelines in order to make genetics a \"hot item\". General practitioners and midwives said they were unsure about their responsibilities in relation to genetics, perhaps because they lacked insight into the genetic background of diseases and its possible consequences. A representative of a patient advocacy group (male, 59 years) said: \"No knowledge (of genetics) and no interest (in genetics). It\'s not a hot item. It seems as if general practitioners are not interested in identifying patients with familial hypercholesterolemia (FH). We have an excellent screening programme for this in the Netherlands, which threatens to go under because too few index patients are put forward by general practitioners. General practitioners do not alert patients that they might have FH when there is a positive family history and high cholesterol levels, and patients are not advised to take part in a brief screening programme \[\...\]. General practitioners often say they don\'t have the time or they\'re not interested or they see no benefit.\" Primary care providers noticed a change in their experiences with and views of the role of genetics in primary care, which led to an increased need for basic knowledge of genetics and family history taking. General practitioners felt their current knowledge was insufficient to meet these needs. Participants also said they noticed an increase in patients\' questions about genetic issues. They perceived a change in the responsibilities of primary care providers that prompted increased attention for genetics. They also saw an urgent need for a description of the responsibilities of different disciplines in relation to genetic issues. A clinical geneticist (female, 40 years) put it as follows: \"When I think of genetics I think of monogenetic disorders, but of course disorders (seen in primary care) are often complex disorders or multigene or gene environment interaction disorders. Of course, clinical genetics cannot deal with all those problems, it\'s simply impossible. \[\...\] I think it is the task of the general practitioner, but I find this difficult\... is it really the task of the general practitioner to deal with such complicated problems?\" Overall, participants were positive about the changed role of genetics in primary care. They said this change emphasised their role as an easily accessible source of information. However, there was also some criticism. General practitioners saw even more work coming their way, which caused some concern. Taking a family history, non-directive counselling, and unfamiliarity with recording information from family history in their electronic patient record were said to take up a great deal of time. General practitioners indicated that they were aware of rapid developments in genetics and the subsequent lag in its application in primary care. They regarded this as important and pointed to two aspects of this change. Firstly, they said it was urgent for limits to be set in relation to required genetics-related knowledge and responsibilities in primary care. Secondly, education should include the clinical application of genetic developments and ways to communicate genetic information. A midwifery educator (female, medical doctor, 56 years) clarified these aspects: \"Highly educated people develop national genetics guidelines. Even the ethical issues involved in these problems and how to deal with them are prescribed. In primary care you are in close contact with patients and it can be difficult sometimes to apply theory-based guidelines in a way that can be understood by patients, it is difficult to do this appropriately.\" Suggestions for strategies for effective genetic education ---------------------------------------------------------- At the end of each focus group interview, the participants were asked to briefly consider effective strategies for teaching genetics in primary care education (Additional file [4](#S4){ref-type="supplementary-material"}). The following general considerations emerged: programmes should be relevant to primary care practice, participants in the multidisciplinary group emphasised the importance of assuring the quality of educational strategies and suggested that programmes should range in duration from brief sessions to ten-day programmes. Finally, strategies should be added to existing programmes or could be integrated with other topics, such as cardiovascular risk management or familial breast cancer as examples of common diseases. Discussion ========== The results of this study indicate that Dutch primary care providers need, and would welcome, more extensive education in genetics. Four major themes emerged in relation to the role of genetics in primary care and the related educational needs: lack of basic knowledge, need for education on family history taking and the potential clinical consequences, ethical dilemmas and psychosocial effects related to genetics and insight into the organisation of regional genetics services and the referral system. There was general agreement that increased genetics knowledge and family history taking by primary care providers would require a better understanding of the organisation of genetics services in order to promote more appropriate and timely referrals. In summary, the results point to a need for courses in genetics for master programmes in midwifery and postgraduate programmes in family medicine. A similar need for genetics education in primary care was also found in other studies. The identified needs are in line with the learning outcomes and core competencies in genetics proposed by genetics experts for non-genetic health care professionals \[[@B17],[@B18]\]. Since there is little published research on the extent to which the need for genetics education matches the core competencies, we used a qualitative approach to explore the views of the target group. In this way we gained insight into the educational needs of this group with regard to genetics; general practitioners indicated that a paucity of knowledge can lead to poor recognition of and unresponsivess to genetic problems in daily patient care. The results of this study are in line with some studies and differ from others with regard to the need for increased genetics knowledge among midwives \[[@B8],[@B13]\] and general practitioners \[[@B5],[@B10],[@B12],[@B14],[@B16],[@B19],[@B20]\]. The midwives in our study seemed more confident of their basic knowledge and did not perceive as strong a need to adapt existing educational programmes as was expressed by midwives in studies by Benjamin *et al*. and Metcalfe *et al*. \[[@B8],[@B13]\]. This difference may be due to differences between master programmes in midwifery or between health care systems. Our results support the outcomes of the afore mentioned studies regarding deficiency in skills (e.g. taking a family history, referral to appropriate regional genetics services and non-directive counselling) \[[@B8],[@B13]\]. It may be problematic for primary care providers to take appropriate steps in response to the perceived shift in the importance of genetics in primary care, such as taking enough time to discuss the family history or non-directive counselling. Another step to take would be to improve the electronic patient record in order to achieve accurate documentation of family history information. Martin and Wilikofsky reported on general practitioners\' perceptions of their role in genetic counselling and their unwillingness to accept this role due to time and organisational constraints \[[@B21]\]. Representatives of patient advocacy groups and genetic counsellors in our study emphasised the need to increase acceptance of the importance of genetics and genetic counselling in primary care. The responsibility, on the part of the patient or the doctor, to report data from the family history remains a topic of debate, however, even though the importance is clear and primary care seems well suited to include this role in daily practice routines\[[@B22]\]. Perhaps a joint effort by all stakeholders would be realistic and useful. General practitioners and other participants in our focus groups recognised the important role of genetics in primary care. This is in contrast to a study conducted by Fetters *et al*. in 1999 \[[@B12]\], which found general practitioners reluctant to invest in self-education in genetics, because they felt genetic problems were not clinically relevant. Our study suggests that today\'s primary care providers are aware of a progressive impact of genetics on primary care and therefore increasingly conscious of what they don\'t know. They recognise the need for attention to genetics in educational programmes. Perhaps this is a reflection of family medicine finally becoming aware that genetics and genomics are an integral part of primary care. Clinicians were seen to be uncomfortable in applying genetics in their daily practice, which resulted in difficulties in referring adult patients for genetic counselling\[[@B7]\]. Our study showed similar results. Some general practitioners were reluctant to consult a clinical geneticist, whereas midwives seemed to be more comfortable with this. Representatives from patient organisations were also aware of this barrier and urged more genetic education for primary care providers, general practitioners in particular. Taylor *et al*. also suggested that insurance coverage of genetic consultation can be a problem. There is currently a paucity of published research on the clinical value of genetic evaluation in primary care \[[@B23]-[@B25]\]. Genetic counselling could be of greater value and might be integrated in periodical check-ups more often if its results had greater practical applicability. The educational strategies suggested by general practitioners and midwives in this study appear to be supported by Gaff *et al*. \[[@B26]\], who concluded that \"Program logic, adult learning theory, and evaluation theory together provide a useful and relevant theoretic framework for the development of genetics education programs for health professionals.\" Limitations ----------- The use of focus groups has engaged primary care providers of a potential genetics education programme in the Netherlands. A variation in concepts is possible, because it is unknown how far the themes reach in their contribution and interaction in real practice. The aim of this study was intended to yield results regarding the participants\' particular views on knowledge, skills and attitudes in relation to genetics education in primary care. Apart from homogenous groups of general practitioners and midwives, we included participants from a variety of backgrounds to obtain input on broader and future developments in genetics in primary care. However, it remains to be investigated if the results have relevance beyond the Dutch health care system, since the nature of the sample was drawn from this particular health care system. Together with previously published studies on various aspects of genetics in primary care education, our study offers a broad perspective on genetics education. We believe this information can be used to develop genetics education programmes in the near future. The inclusion of multidisciplinary focus groups which could provide meta views can be considered a strength but also a weakness of this study because of the unequal representation of different fields of expertise in these groups. Another limitation is that purposive sampling can result in self-selection, which can introduce bias. Our study revealed four major themes concerning the role of genetics in primary care. In order to ensure that our picture is complete and usable for educational purposes, and possibly for policy makers as well, consensus has to be sought, for example by means of a Delphi procedure. Conclusions =========== The results of this study suggest that postgraduate training in primary care could be enhanced by incorporating additional training in basic clinical genetics. For midwives and general practitioners there should be more emphasis on counselling using strategies that are clinically feasible and on ethical issues relating to genetic conditions. Insight into the organisation of regional genetics services and the referral system should be enhanced to promote interdisciplinary collaboration. There is an urgent need for a clear description of responsibilities and guidelines to enable effective use of developments in genetics in primary care. Especially descriptions of the genetic responsibilities of primary care providers and their specific role in this area will have to be addressed by future research. Useful and effective application of genetics knowledge can only become a reality when genetics education is improved. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= All authors made a substantial contribution to the concept and design of the study. IH organised and attended all the focus groups, wrote the summaries for member checking, carried out the analysis together with LH and wrote the first draft of the paper. SJvL was moderator of the focus groups. LH attended the focus groups to take notes and observe. SJvL and LH participated in the design of the study and the development of the focus group scenario. CvdV, GJD and MC were involved in coordination of the study and design. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2296/12/5/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Table 1 Characteristics of the participants in the focus groups**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Table 2 Interview guide for the focus group discussions**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Table 3 Themes and subthemes identified regarding the genetic educational needs of general practitioners and midwives**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **Table 4 Suggested strategies, including details, for effective ways to incorporate genetics into primary care education**. CME continuing medical education. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The study is part of the research programme of the Centre for Society and Genomics & Centre for Medical Systems Biology, and is financially supported by the Netherlands Genomics Initiative. We thank all focus group participants for their contribution.

PubMed Central

2024-06-05T04:04:19.448529

2011-2-17

PMC3053219

Background ========== Accurate survival or outcome information is important for many clinical studies, clinical routine and epidemiology. The standard method for estimating survival is the Kaplan-Meier plot (KM-plot) \[[@B1]\] with high relevance in medical research. Particularly in oncology the Kaplan-Meier technique is used to compare survival information between different therapy strategies or stages of the disease \[[@B2]\]. Survival time and follow-up status are used to compute an estimate of a survival curve for censored data. In the German healthcare system a distinction can be made between the routine documentation which is commonly performed in hospital information systems or still paper based and the research documentation mainly performed in electronic data capture systems (EDC) or on paper based case report forms (CRF). There are several systems in healthcare resulting in separate documentation of medical routine, research and quality management data. In the context of clinical studies survival information is currently captured on paper based CRF and occasionally on electronic CRF (eCRF), but generally separated from HIS. However, some patients already have at least a basic electronic medical record \[[@B3]\]. Currently survival analysis is not possible within the HIS as the required data is not available. A problem of these external databases consists in the resulting difficult multiple usage of information. Data from a single patient can be relevant for several studies (e.g. therapy study, biomarker discovery study, epidemiological studies) in addition to clinical routine. Redundant documentation is common but inefficient regarding a resource limited setting and carries the danger of inconsistent information. In this setting it would be attractive to use health data outside of direct care delivery. A secondary use of documented data for research and quality management \[[@B4]\] may be of potential benefit for those physicians who already use electronic documentation \[[@B5]\]. For instance, the idea of the REUSE project implies that clinical data should be based in the electronic health record (EHR) independent from the context, in which data is captured \[[@B6]\]. The secondary use of clinical data has enormous potential for improving quality of care \[[@B7],[@B8]\]. Furthermore, there is heterogeneity in the required follow-up documentation depending on the disease and the department in which the data is obtained. Physicians need further information to interpret the status in combination with clinical data. Also, the parameter values for the follow-up status may be different between oncological diseases. Therefore, an efficient implementation should be based on a generic data model which is suitable for several diseases. Studies of health services research, epidemiological studies and phase III/IV studies often consist of a high number of cases resulting in laborious documentation and high data management costs. Using routine documentation may reduce the documentation workload and reduce costs. To reach these goals the respective source data have to be \"accurate, legible, contemporaneous, original, attributable, complete and consistent\" \[[@B9]\]. However, data quality of follow-up documentation is often unsatisfactory \[[@B10]\] and requires adaption before it can be used for research. The follow-up documentation needs to be complete to obtain meaningful KM-plots as incomplete information can bias the analysis of the results \[[@B11]\]. One strategy to increase completeness and achieve higher data quality is to use an electronic documentation tool \[[@B12],[@B13]\]. This approach could be extended to use the HIS for clinical and research documentation which would also result in high data quality \[[@B14]\]. Commercial HIS usually cover only events during hospitalisation so there is a need for a special follow-up module to allow the survival documentation for inpatients and outpatients. With respect to clinical quality management it is important to obtain timely KM-plots of all patients and not just from those in clinical studies. Thus, it would be desirable to integrate the follow-up documentation into clinical routine and document it in the EHR within the HIS. Using this method, it will be made available for all treating physicians and clinical research projects can be combined with routine documentation by reusing the EHR \[[@B15]\]. Because of the high relevance, especially in oncology, we analyse whether an integrated follow-up documentation is feasible and focus on the following objectives: 1\. Is it feasible to design a follow-up documentation system in the HIS which is suitable for several oncological diseases and provides a secondary use of data? 2\. Is it possible to extract survival information from routine HIS documentation so that physicians can obtain KM-plots in a timely manner? 3\. What level of data quality can be achieved with respect to completeness of forms and completeness of items per form? Methods ======= Process analysis ---------------- We analysed the current follow-up documentation process in the urology and haematology department at the University Hospital Muenster. Physicians were interviewed to identify weak points and requirements during the follow-up appointments. Form analysis ------------- During the form analysis we investigated the attributes which characterise the overall survival (OS) and the event-free survival (EFS). We compared follow-up documentation of two tumours regarding type, complexity and concordances. Concept, data model and implementation -------------------------------------- We developed a concept for comprehensive survival documentation based on a generic data model which is described in the CDISC Operational Data Model (ODM v1.3.1.) \[[@B16]\] using SNOMED CT V3 Codes \[[@B17]\]. Based on this concept an electronic survival form was implemented within the local HIS containing specialised parts for leukaemia and for prostate cancer. We used the integrated tools of the HIS (ORBIS^®^from Agfa Healthcare) \[[@B18]\] to create a form which contains all relevant survival data for KM-plots and integrated it into the clinical workflow. Data export and survival analysis --------------------------------- We created a report of the HIS form to extract follow-up information (initial diagnosis date, initial therapy date, date of last contact, status at the last contact, etc.) from the HIS and to transfer pseudonymized data sets to statistical programs. This report was integrated in the HIS in such a way that it can be used by the physicians to view summarized data of their patients in the HIS or to export pseudonymized survival data as comma separated values (CSV). Survival analyses were implemented in R (version 2.10.1) and for the KM-plots we used the \"*survfit function*\" of the R survival library \[[@B19]\]. Differences between subgroups of patients were assessed using log-rank test, implemented in the \"*survdiff function*\". A batch script was written to execute the R survival function in such a way that survival curves and data quality information are directly accessible in a PDF-file. Analysing data quality ---------------------- We analysed completeness to describe data quality according to Chan et al \[[@B20]\]. Therefore, we designed a report to query missing and incomplete forms. Based on these export data, completeness of forms (reference: all prostate cancer patients in urology) and completeness of documented items per form were analysed. The study was performed in compliance with the World Medical Association Declaration of Helsinki on Ethical Principles for Medical Research Involving Human Subjects. HIS data access was approved by the responsible data protection officer, only de-identified data items were exported. Results ======= Process analysis of the paper based follow-up documentation ----------------------------------------------------------- Follow-up documentation was paper based for AML patients in haematology and spreadsheet based for prostate cancer patients in urology. In both departments, forms were completed when patients were appointed to follow-up examinations. If the patient lacked a follow-up date the treating physicians could obtain the information from general practitioners, from the epidemiology cancer registry or from the registration office (in case of fatality). For statistical analysis and survival curves the follow-up documentation was entered manually into spreadsheets or statistical programs which were then used to generate KM-plots. Form analysis ------------- Each department has its own follow-up form. While analysing both forms we identified 13 common attributes. Particularly the relevant survival information (diagnosis date, therapy date, follow-up status and follow-up date) are common to both departments so that a generic form for both diseases is possible. The remaining attributes of both departments were not included in the generic data model but belong to the disease specific documentation which is considered in a specialised part of the form. Table [1](#T1){ref-type="table"} shows the results of the form analysis. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Results of the form analysis ::: *Department* *\#Pages* *\#Attributes (total)* *\#Attributes (common)* -------------- ----------- ------------------------ ------------------------- Urology 1 35 13 Haematology 1 30 13 Follow-up forms from both departments contain more than 30 attributes to document survival information, 13 of which are common (e.g. diagnosis date, therapy date, follow-up date, follow-up status). ::: Two of the common attributes (study, status) have different parameter values. For example, the values of the status lists differed between the two departments considering the different stages of the diseases. Therefore we implemented a catalogue with variable parameter values for each department. For the KM-plots it is only relevant to distinguish between overall survival (OS \--\> yes/no) and event free survival (EFS \--\> yes/no). Therefore, we mapped the more detailed elements of the status list unambiguously to these two parameters. The resulting status list with the mapping and coding is presented in table [2](#T2){ref-type="table"}. It is now possible to use this encoding in both departments for Kaplan-Meier plots. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Status values in the department of haematology and urology ::: *Status in the haematology department (AML)* *Status in the urology department (prostate cancer)* *OS* *EFS* ---------------------------------------------- ------------------------------------------------------ ------ ------- Initial diagnosis Initial diagnosis 0 0 \- Relapse free 0 0 Aplasia \- 0 0 First remission \- 0 0 Complete remission unconfirmed \- 0 0 Second remission \- 0 0 Relapse Relapse (PSA) 0 1 \- Relapse (imaging) 0 1 Persistent AML \- 0 1 Death of AML Death of prostate cancer 1 1 Death independent from AML Death independent from prostate cancer 1 1 Death of unknown cause Death of unknown cause 1 1 Both departments have different status list values to be considered in the survival analysis. Concerning overall survival (OS) analysis all death events are coded as 1, regarding event-free survival (EFS) analysis all except for initial diagnosis and relapse free are coded as 1. ::: Concept and data model ---------------------- Based on the results of the form analysis we developed a generic data model which includes the 13 common attributes. These attributes were determined through data types and also tagged with SNOMED CT codes. For the specification of the data model we used CDISC ODM because it is designed to facilitate interchange of metadata and data for clinical research \[[@B16]\]. The ODM follow-up form consists of 5 item groups (identity, diagnosis, therapy, study data and follow-up data). An extract of this form is shown in figure [1](#F1){ref-type="fig"}. The complete ODM example of the follow-up form can be found in the supplement (see Additional file [1](#S1){ref-type="supplementary-material"}). All items are specified by name, id and data type. For all attributes we added SNOMED CT codes so that the used concepts are well-defined. Regarding the different status values and the different study lists in each department we modelled code tables for the parameter values of the follow-up status. Finally we obtained a system independent specification for the follow-up form, which also allows semantic interoperability through the SNOMED codes. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Extract of the follow-up representation in CDISC ODM**. The extract shows the definition of diagnosis and therapy data with their data type. The complete ODM example of the follow-up form can be found in the supplement (see Additional file [1](#S1){ref-type="supplementary-material"}). :::  ::: Implementation -------------- Based on the generic data model we implemented a follow-up form within our HIS (ORBIS^®^) and introduced the electronic follow-up documentation for prostate cancer patients in the urology department and for AML patients in the haematology department. An extract of the implemented HIS form, used for routine documentation, is shown in figure [2](#F2){ref-type="fig"}. For each patient the initial diagnosis and initial therapy can be specified with date, text and classification. Hence, both points of time can be used as start date of the KM-plots. The lower part of the form contains the date of the last visit with the associated status and provides the possibility to document the source of the survival information (e.g. general practitioner, registration office). In addition the participation in clinical studies can be documented. We also added disease specific sub-forms for prostate cancer and AML containing the attributes which were not common in both departments. These forms complete the follow-up documentation in both departments. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Extract of the follow-up form**. Initial diagnosis and initial therapy can be documented with date, text and classification in the upper part of the form. The lower part contains the date of the last visit with the associated status in addition to the studies. This figure shows the electronic version of the AML follow-up documentation. All attributes are similar for prostate cancer but status lists and studies have different parameter values. In addition to this generic form, disease specific form components can be provided. :::  ::: In routine documentation the follow-up form should be completed for every cancer patient in the respective department. Patients with prostate cancer or leukaemia have regular follow-up examination and during those visits the survival information is documented. If patients do not appear the information has to be collected from the current practitioner. In this case physicians and study nurses enquire relevant information from general practitioners as well as the registration office (if there is a fatality reported) and enter this information into the system. We also added the attribute \"source\" so that the documenting physician can document the enquired institution. In this manner information from outpatients can be also documented. In addition to this generic form, disease and department specific attributes can be documented in specific form components which were provided for prostate cancer and leukaemia. To analyse the feasibility of KM-plots based on structured HIS data the form was integrated in the clinical workflow in the urology department and during a short time frame data from follow-up patients was documented. To achieve a high number of cases which allow for comprehensive analyses and to reuse previously existing survival data from paper based records or spreadsheet files in the urology department this information was retrospectively transferred to the HIS. In the haematology department up until now the electronic follow-up documentation was only used in a pilot installation with relatively few patients. Therefore, the following analyses are based on urology data. In our approach we provide one form per follow-up date so that each patient has several forms during the course of his disease. The current survival status is always taken from the most current form. HIS-based Kaplan-Meier ---------------------- The documented survival data were extracted from the EHR and analysed with the statistic software R. We considered a follow-up period from 03.06.1992 to 31.05.2010 in which patients where documented and completed follow-up forms also retrospectively to obtain a relevant data basis for the KM-plots. In total follow-up forms were entered from 23^rd^February to 01^st^July 2010 for 1029 patients; 223 of them were completed in the routine documentation process, 806 were entered retrospectively. Using the R survival library we implemented survival analyses and KM-plots based on exported CSV-data. First we removed all duplicated forms from all patients and kept only the data from the most current form. Observation time was defined on basis of therapy start date and follow-up date, overall survival and event-free survival were computed concerning the following status encodings (table [2](#T2){ref-type="table"}). We created data frames for the EFS and OS for the patient collective as well as data frames for grouped analyses. Prostate cancer patients were divided into two groups. KM-plots with numbers at risk were generated from all four data frames. The complete R code can be found in the supplement (see Additional file [2](#S2){ref-type="supplementary-material"}). Survival information was available for 881 of the 965 patients. The median overall survival time was 16.4 years, the median event-free survival time was 7.7 years. The probability of 5-year overall survival was 98.2% (EFS: 82.8%), the probability of a 10-year overall survival was 89.9% (EFS: 32.6%). Table [3](#T3){ref-type="table"} shows the basic information of patients in the urology department. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Survival information of patients in the urology department ::: *Department of urology* *Analysis of overall survival* *Analysis of event-free survival* ------------------------------------------- -------------------------------- ----------------------------------- Number of patients 965 965 Survival information available 881 881 Median survival time (years) 16.4 7.7 Probability of 5-year survival (percent) 98.2 82.8 Probability of 10-year survival (percent) 89.9 32.6 Total number of events 31 267 -\>Death 31 31 -\>Relapse \- 236 Survival information was available for 881 patients covering a time interval of nearly 17 years. ::: Based on these survival data from HIS the following KM-plots (figure [3](#F3){ref-type="fig"}+[4](#F4){ref-type="fig"}) were created. Figure [3](#F3){ref-type="fig"} shows the overall and the event free survival with a 95% confidence interval of all prostate cancer patients documented in the HIS. Figure [4](#F4){ref-type="fig"} illustrates that is also possible to obtain group analyses. To differentiate between the patients it is possible to combine survival data with other routine information (e.g. biopsy information, lab results, body weight, height, ultra sound findings) available in the EHR. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Kaplan-Meier Plots**. These figures show the survival with 95% confidence interval and numbers at risk. The left figure represents overall survival and the right figure event-free survival of patients with prostate cancer after radical prostatectomy. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Kaplan-Meier Plots (Group analysis)**. These figures show the survival of the prostate cancer patients divided into two groups with numbers at risk. The left figure represents overall survival with no significant difference (p = 0.552) and the right figure event-free survival which shows a significant difference between both patient groups (p \< 0.001). :::  ::: We wrote a batch script to automate the process from the export file to the final PDF with the KM-plots. During several discussions with health professionals we discovered that there are patients for whom no accurate date of relapse or freedom of relapse can be specified. To handle this, the generic data model was adapted to allow for imprecise inputs (e.g. Oct. 2008, 2007). Data quality ------------ Based on the exported data sets we assessed completeness of the documentation to describe data quality. With a focus on the mandatory items for the survival analyses we used R to identify missing items of the survival parameters used for the KM-plots (therapy start date, follow-up status and follow-up date). These reports are permanently installed so that data quality may be obtained also for future analyses and interpretation of the KM-plots. Concerning the documented items per form, we distinguished between routine cases and retrospective cases and measured also the completeness of the whole data set. During the routine documentation the completeness of all items was 86.6% due to a few missing follow-up dates. Considering only retrospectively entered data we reached a completeness of 92.4%. In total, the completeness in the follow-up date was high (\~94.3%) while completeness of therapy start date and follow-up status was very high (\> 97%). KM-plots based on all three attributes were available for 881 patients (91.3%). Table [4](#T4){ref-type="table"} shows the results. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Completeness of survival data ::: *Attribute* *Therapy start date* *Follow-up Status* *Follow-up date* *All three available* --------------------- ---------------------- -------------------- ------------------ ----------------------- Routine cases 207 Available 204 207 183 180 Completeness 98.5% 100% 88.4% 86.6% Retrospective Cases 758 Available 738 743 727 701 Completeness 97.3% 98% 95.9% 92.4% Total Cases 965 Available 942 950 910 881 Completeness 97.6% 98.5% 94.3% 91.3% Completeness was measured for routine cases (207) and retrospective cases (758) separately. In total therapy date and follow-up status is relatively complete for all patients (\> 97%), follow-up date was available for 94.3% of the patients. Imprecise date items (follow-up month and follow-up year available) are considered. Survival information was available for 91.3% of the patients. ::: To analyse the acceptance of the electronic forms we measured how many patients with prostate cancer diagnosis already have an electronic follow-up form. For the analysis of the completeness of forms, we considered patients with a main diagnosis of prostate cancer from 01.01.2009 to 31.12.2009 in the department of urology. We chose this time range because patients with recent diagnoses are currently not appointed to follow-up examination and could bias the result. In total 115 of the 191 patients have at least one follow-up form with survival information so that completeness of forms is 60.2%. Discussion ========== With our implementation it is now possible to generate KM-plots from routine data. After exporting follow-up data sets in a pseudonymized format from the HIS the physician can start the batch script with the R survival functions to obtain a resulting PDF-file with the KM-plots. Both actions can be done by physicians within a minute and therefore it is a feasible method to get timely curves from the current data. Up to now, this procedure was not possible for physicians because data from routine documentation has to be transferred into statistical programs for survival analyses to be performed. The idea to enhance the primary information system is not a new one. Previously, in 1996, Balas et al. state that \"to manage care and improve quality, primary care computer systems should incorporate these effective information services\" \[[@B21]\]. However, the implementation of single source systems as described by Kush et al. \[[@B22]\] is still rare. During our literature search, we failed to find similar approaches of integrating this kind of follow-up documentation in the HIS in such a way that timely survival curves can be generated. Ene-Iordache et al. analysed regulatory-compliant eCRF \[[@B23]\] and Embi et al. identified in 2009 a lack of tools for clinical research activities \[[@B24]\]. By using electronic point-of-care documentation to generate KM-plots clinical research activities can be supported and the integrated documentation contributes to the single source approach. Data Quality Aspects -------------------- The measured data quality, especially the completeness of documented items per form, was high but most of information was transferred from retrospective data collections (paper based documentation and spreadsheets). The completeness of the electronic forms in the HIS (retrospective and current documentation) was only 60.2% so we assume that there are still patients with missing or paper based documentation. However, to analyse the completeness of forms more data is needed. Further analyses will show the differences in completeness of forms between the retrospective collection and the routine documentation. Chan et al. reviewed data quality in EHRs of recent studies and reported that data completeness varied substantially across studies and that even in the same organisation the amount of missing data is varying \[[@B21]\]. To measure follow-up completeness Clark et al. introduced the ratio of the total observed person-time of follow-up as a percentage of the potential time of follow-up in a study \[[@B11]\]. Therefore we measured follow-up completeness of C = 76.9% using this approach for all prostate cancer patients who underwent radical prostatectomy in the urology department. The follow-up documentation is heterogeneous and therefore the implementation in a local HIS is complex and time consuming. There is currently no possibility to reuse captured routine data in electronic study documentation systems and therefore the advantages of HIS based documentation are limited. Especially studies with a high number of study cases (e.g. epidemiological studies, phase IV studies) are attractive for data re-use. The follow-up documentation is usually scheduled in defined intervals (depending on the disease). The HIS-based approach adds the functionality to notify the treating physicians of required documentation activities. We plan to integrate work lists which show all patients without a follow-up form in the last year. This approach could be extended by an automated creation of forms in the HIS related to the follow-up intervals. Survival information in the HIS can be re-used for physician letters. Our approach of follow-up documentation is generic and we intend to extend it to other departments and diseases. If the follow-up information is available for many patients it can also be used for patient recruitment for clinical trials \[[@B25]-[@B28]\] as attributes like survival status are now documented in a structured way within the HIS and can be used as inclusion or exclusion criteria. In a short time frame 223 cases were documented during routine documentation which indicates a good clinical acceptance. In addition 806 cases were completed retrospectively in order to have the entire follow-up documentation electronically available showing the need for HIS based follow-up documentation. Structured documentation of follow-up items should therefore be a standard functionality in HIS. Conclusion ========== Different follow-up forms can be united in a comprehensive module which is suitable for oncological diseases. The integration in the local HIS is feasible and provides possibilities to reuse this information for quality control and clinical research (single source) by generating timely survival curves from routine data. Abbreviations ============= AML: acute myeloid leukaemia; CDISC: Clinical Data Interchange Standards Consortium; CRF: case report form; eCRF: electronic case report form; CSV: comma separated values; EDC: electronic data capture; EFS: event-free survival; EHR: electronic health record; HIS: hospital information system; ODM: Operational data model; OS: overall survival; PDF: portable document format; PSA: prostate specific antigen; SNOMED CT: Systematized Nomenclature of Medicine - Clinical Terms; Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= BB implemented the HIS forms and reports, created the survival curves, analysed data quality and wrote the manuscript. AS and CMT provided clinical data and survival information. FF supported implementation of forms and MD contributed the data quality analysis. FF and MD critically revised the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-6947/11/11/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **ODM Follow-up form**. This xml-file contains the follow-up form in the ODM format (v1.3.1). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **R-Code**. This file contains the R-Code for the survival analysis. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Regina Weis supported retrospective data collection. Kathleen O\'Hagan supported English writing.

PubMed Central

2024-06-05T04:04:19.451824

2011-2-16

PMC3053220

Background ========== Mitotic progression and assembly of the bipolar mitotic spindle are regulated by several serine/threonine protein kinases, including members of the cyclin-dependent kinase (Cdk), Polo-like kinase (Plk), Aurora, and NIMA-related kinase (Nek) families \[[@B1]-[@B4]\]. The founding member of Nek family, the NIMA kinase of *Aspergillus nidulans*, contributes to multiple aspects of mitotic progression including the timing of mitotic entry, chromatin condensation, spindle organization and cytokinesis. Mammals contain a large family of eleven Neks, the catalytic domain of which is evolutionarily related to that of NIMA \[[@B4]\]. Nek2 has a central role in centrosome maturation and disjunction \[[@B5]\], whereas Nek1 and Nek8 have been proposed to contribute to ciliary function \[[@B6],[@B7]\]. Besides Nek2, Nek1, -6, -7 and -9 were also described to participate in centrosomal regulation \[[@B7]-[@B11]\]. Nek6, Nek7 \[[@B9]\] and Nek9 \[[@B12]\] are involved in the control of mitotic spindle formation, acting together in a mitotic kinase cascade, with Nek9 being upstream of Nek6 and Nek7 \[[@B13]\]. Nek kinases are also described as related to pathologies, particularly cancer, presenting thereby interesting potential chemotherapeutic targets \[[@B14]-[@B20]\]. Recently, hNek6 was described to have its transcript, protein, and/or kinase activity levels highly elevated in a number of tumors and human cancer cell lines, indicating an important role for hNek6 in tumorigenesis \[[@B21]-[@B24]\]. Structurally, Neks in general are characterized by having a conserved N-terminal catalytic domain, followed by a nonconserved C-terminal regulatory domain that varies in size and structure. However, Nek6 and Nek7 are significant exceptions to this, in that they are the smallest of the kinases and consist only of a catalytic domain with a relatively short N-terminal extension \[[@B4]\]. Although they share significant similarity with each other, being \~86% identical within their catalytic domains, the N-terminal extensions are not conserved, and it has been suggested that they may play a role in differential regulation of the kinases \[[@B25]\]. The mechanisms of regulation of hNek6, -7, and -9 kinases are currently unknown, and elucidating this pathway would provide relevant knowledge on early mitotic events as well as new hints for drug design and cancer therapy. However, hNek2 and hNek7 are the only NIMA-related kinases for which structures have been reported \[[@B26]-[@B28]\]. In this context, we present here the first low resolution three-dimensional structure of hNek6 protein in solution. SAXS experiments, together with SEC-MALS and comparative molecular modeling revealed a monomeric mostly globular, though slightly elongated conformation for hNek6, with a flexible disordered N-terminal domain. Results and Discussion ====================== Human Nek6 is predicted to be phosphorylated at various sites and has an unfolded short N-terminal domain --------------------------------------------------------------------------------------------------------- Human Nek6 amino acid sequence was analyzed considering its secondary structure, disordered regions, conserved motifs and putative phosphorylation sites by upstream kinases, resulting in a linear representation of its main structure predictions (Figure [1A](#F1){ref-type="fig"}). These analyses were also performed for hNek7, for which the crystallographic structure was recently determined \[[@B28]\], in order to validate our results (Figure [1B](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Human Nek6 and Nek7 structure predictions and sequence aligment**. Consensus of predicted secondary structure for Nek6 (A) and Nek7 (B), is represented by helices (red waves), strands (yellow arrows) and coils (black lines); it was generated from 5 different predictions and the consensus score for each amino acid (1-5) is shown (scores 1-2 represent ambiguous predictions). The consensus of predicted disordered regions was generated from 9 predictions; it is represented by blue bars when the consensus score (0-9) is from 5 to 9. The N-terminal regulatory and C-terminal kinase domains are depicted and two key signatures within the kinase domain are in colored boxes as follows: ATP-binding region, green; Ser/Thr kinases active-site, orange. The conserved glycine-rich sequence, HRD and DLG motifs are in purple, and the conserved residues K^74^(β3 strand) and E^93^(αC helix) are in red. The putative nuclear export signal LGDLGL based on la Cour et al., 2004 \[[@B63]\], is underlined and the putative WW domain binding motifs PY and pSP based on Ingham et al., 2005 \[[@B64]\], as well as the PPLP motif experimentally described for hNek6 by Lee et al., 2007 \[[@B65]\], are indicated by dotted lines. Putative phosphorylation sites were predicted by NetPhosK 1.0 software and are indicated in blue; S^206^in hNek6 is also indicated because of previously described phosphorylation of this residue by hNek9 \[[@B13]\]. These motifs are also present in Nek7 amino acid sequence. The author-approved secondary structure of hNek7 in PDB ([2WQM](2WQM)) is also depicted. (C) Primary sequence alignment between human Nek6 (NP\_055212.2) and Nek7 (NP\_598001.1) using Clustal W2. The N-terminal regulatory domain is highlighted in yellow. :::  ::: Our consensus of secondary structure was scored by the number of times (one to five times) the predicted secondary structure element (α-helices, β-strands or coils) scored positive from five predictions using different databases: **PredictProtein/Prof**\[[@B29]\], **PSIPRED**\[[@B30]\], **SSpro**\[[@B31]\], **SOPMA**\[[@B32]\] and **GOR4**\[[@B33]\]. In summary, the secondary structure analysis suggested that hNek6 was composed of approximately 34% α-helices, 12% β-strands and 54% coils. Our hNek7 consensus of predicted secondary structure is 80% identical to the author-approved secondary structure in PDB ([2WQM](2WQM)) (Figure [1B](#F1){ref-type="fig"}). In the case of the disordered regions predictions, our consensus was obtained following the same criteria as for the secondary structure predictions, except that we used here nine different databases: **FoldIndex**\[[@B34]\], **GlobPlot Russell/Linding**\[[@B35]\], **PONDR VL-XT**\[[@B36]\], **DISpro**\[[@B37]\], **IUPred**\[[@B38]\], **DisEMBL Hot-loops**, **DisEMBL Remark-465**, **DisEMBL Loops/coils**\[[@B39]\], and **VSL2B**\[[@B40]\]. From this analysis, we were able to identify a short high scored segment of disorder covering the majority of hNek6 N-terminal extension before its catalytic domain, which we are calling here the regulatory domain. This characteristic is also present in our hNek7 consensus of disorder predictions and in its crystal structure \[[@B28]\], where amino acids 1-19 are missing residues (due to a flexible region) and amino acids 20-23 are coils. Notably, we found that hNek6 unfolded short N-terminal region is important to mediate interactions with diverse partners \[[@B41]\] and, since hNek6 and hNek7 are similar in their catalytic domain sequences (\~86% identity), but different in their N-terminal extensions (\~20% identity) (Figure [1C](#F1){ref-type="fig"}), it is possible that both proteins depend on their disordered N-terminal domain to regulate the interactions with specific/different partners. For phosphorylation analysis, NetPhosK \[[@B42]\] and NetPhos \[[@B43]\] databases were used to predict phosphorylation sites (Table [1](#T1){ref-type="table"}), and together with *in vitro*and *in vivo*phosphorylation data \[[@B13]\], they were used to assign tyrosine, threonine and serine residues as putative phosphorylation sites for hNek6 and hNek7 (Figure [1](#F1){ref-type="fig"}). This analysis shows a variety of high score phosphorylation predictions for hNek6. Interestingly, there are four predicted sites localized in hNek6 N-terminal domain (serine residues 13, 14, 37 and 41), and one of them (Ser^37^) also described to be phosphorylated *in vivo*\[[@B13]\], suggesting that this is an important phosphorylation-regulated region. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Prediction of putative phosphorylation sites in human Nek6. ::: **Residue**^**a**^ **Putative Upstream Kinase**^**b**^ Predictor (Score) -------------------- ------------------------------------- -------------------------------- S^13^ PKA NetPhosK (0.63) S^14^ PKA NetPhosK (0.52) S^14^ Cdc2 NetPhosK (0.53) S^37^ PKC NetPhosK (0.82) S^41^ PKC NetPhosK (0.56)/NetPhos (0.72) S^41^ PKA NetPhosK (0.65)/NetPhos (0.72) T^70^ PKC NetPhosK (0.79) S^111^ CKII NetPhosK (0.64) S^131^ DNAPK NetPhosK (0.62) S^131^ ATM NetPhosK (0.68) S^158^ CKII NetPhosK (0.53) S^158^ Cdc2 NetPhosK (0.51) T^183^ PKC NetPhosK (0.65) S^198^ PKA NetPhosK (0.56)/NetPhos (0.79) S^198^ Cdc2 NetPhosK (0.50)/NetPhos (0.79) S^199^ CKII NetPhosK (0.50) S^199^ Cdc2 NetPhosK (0.55) T^201^ PKC NetPhosK (0.73) S^206^ CKI NetPhosK (0.53)/NetPhos (0.99) T^210^ GSK3 NetPhosK (0.51) T^210^ CDK5 NetPhosK (0.55) S^215^ GSK3 NetPhosK (0.50)/NetPhos (0.99) S^215^ Cdk5 NetPhosK (0.59)/NetPhos (0.99) Y^224^ EGFR NetPhosK (0.55)/NetPhos (0.72) S^232^ PKA NetPhosK (0.67) S^245^ p38MAPK NetPhosK (0.50) S^245^ Cdk5 NetPhosK (0.55) S^256^ CKI NetPhosK (0.54) S^256^ PKC NetPhosK (0.50) S^275^ CKII NetPhosK (0.52) S^311^ PKC NetPhosK (0.66) ^a^Putative phosphorylated residues predicted with the highest scores by NetPhosK, which may additionally be predicted by NetPhos server. ^b^Kinases as predicted by NetPhosK to phosphorylate the corresponded residue in hNek6 sequence \[Protein Kinase A, C or G (PKA, PKC and PKG); Cell division cycle 2 (Cdc2); Casein Kinase I or II (CKI and CKII); DNA-dependent Protein Kinase (DNAPK); Ataxia Telangiectasia Mutated (ATM); Ciclin-dependent kinase 5 (Cdk5); Glycogen synthase kinase 3 (GSK3); Epidermal Growth Factor Receptor (EGFR); Mitogen-Activated Protein Kinase p38 (p38MAPK)\]. ::: Secondary structure analysis ---------------------------- The secondary structure content of hNek6 was analyzed by Circular Dichroism (CD) spectroscopy. Figure [2](#F2){ref-type="fig"} shows the CD spectra of recombinant wild-type hNek6 fused to a 6xHis-tag recorded at 4°C. Purified protein presents negative ellipticity in the far-UV, with minima at 208 (-15567 deg cm^2^dmol^-1^) and 222 nm (-12053 deg cm^2^dmol^-1^). This spectrum is typical of many globular proteins \[[@B44]\] and suggests a high content of α-helices, since this secondary structure is characterized by minima around 208 and 222 nm. Deconvolution of the CD spectrum using the CDNN software \[[@B45]\] indicated approximately 41.7% of α-helices, 13.2% of β-strands, 15.7% of beta-turns and 25.8% of random coils. Deconvolution using another software, K2d \[[@B46]\], indicated a similar amount of secondary structure elements: approximately 41% of α-helices, 17% of β-strands and 42% of coils. We also estimated the quantity of α-helix structure by the evaluation of the CD spectrum signal at 222 nm, according to Corrêa and Ramos, 2009 \[[@B47]\], resulting in 38.6% of α-helices. Compared to our predictions, wild-type hNek6 showed a very similar content of α-helices (\~34%), β-strands (12%) and coils (54%). In conclusion, both the *in silico*prediction and the experimentally derived data are in reasonable agreement, since they demonstrate an expected high α-helical content for hNek6. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Far-UV Circular Dichroism spectra of recombinant hNek6wt fused to a 6xHis tag**. Residual molar ellipticity was measured from 200 to 260 nm at a protein concentration of 5.8 μM, in 50 mM Phosphate buffer, pH 7,5 with 300 mM NaCl at 4°C using a Jasco J-810 spectropolarimeter. :::  ::: Comparative molecular modeling of human Nek6 -------------------------------------------- The absence of a three-dimensional structure of hNek6 and the increasing interest in studying Nek proteins within the context of drug design strategies prompted us to construct a homology model for the aforementioned protein. Here, we are interested in the activation loop mutant hNek6(S206A). The activation loop is a centrally located loop, typically 20-30 residues in length, that provides a platform for the peptide substrate. Activation of most protein kinases usually requires phosphorylation of a residue in this loop, which leads to a rearrangement of the loop and increase in enzymatic activity \[[@B48]\]. In hNek6, S206 is an important residue, which phosphorylation leads to an increase in the activation status of the kinase \[[@B13],[@B41]\]. The activation loop has the capacity to undergo large conformational changes when the kinase switches between inactive and active states, adopting distinct conformations in different kinases when they are inactive (unphosphorylated activation loop), a fact that has recently been exploited to great medical benefit \[[@B49]\] and which makes our hNek6 mutant an interesting target to be studied. To obtain a homology model of the hNek6 protein, the crystallographic structure of hNek7, available in the Protein Data Bank (PDB: [2WQM](2WQM)) \[[@B28]\], was used as a template. Both proteins share about 77% identity in primary sequence alignment, being \~86% identical in their catalytic domain sequences (Figure [1C](#F1){ref-type="fig"}). Several homology/comparative modeling tools were used as described in the \"Methods\" section of this article. In order to choose the best predicted model, stereochemistry quality analyses were done to check for φ and ψ torsion angles using the Ramachandran plots. A comparison of the results indicated that the model generated by SWISS-MODEL \[[@B50]\] is more acceptable than those generated by the other programs (more amino acids in the most favourable regions and less in the disallowed regions). The SWISS-MODEL homology model is shown in Figure [3](#F3){ref-type="fig"}. The Ramachandran plot (Additional file [1](#S1){ref-type="supplementary-material"}, Figure S1A) showed 86.6% residues in the most favourable regions, 12.1% in additional allowed regions, 0.8% in generously allowed regions and 0.4% (only 1 amino acid) in a disallowed region. As compared to the 2WQM template, these values were 90.8%, 8.8%, 0.4% and 0.0%, respectively. It is important to keep in mind that the template has relatively long regions (one of them consists of 18 amino acids) where the phases could not be solved by X-ray crystallography. Consequently, the homology modeling may not be so accurate in these regions, although the high identity of the target-template sequences makes the whole model plausible. The results revealed that the majority of the amino acids are in a φ-ψ distribution consistent with right handed α-helix and reliable to be a good quality model (Additional file [1](#S1){ref-type="supplementary-material"}, Figure S1A). More details of the validation of the predicted structure results and its quality assessment using PROSA \[[@B51],[@B52]\] are displayed in the Additional Material section (Additional file [1](#S1){ref-type="supplementary-material"}, Figure S1B and S1C). No abnormalities were observed in the validation process, which indicated a good model for the protein. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Comparative molecular modeling of hNek6(S206A) based on hNek7 crystal structure**. Resulting homology model obtained using the structure deposited in the protein data bank under the code [2WQM](2WQM). The N and C lobes are colored blue and green, respectively, and the predicted α-helix following the conserved DLG motif is shown in red. The hinge region and the activation loop (T-loop) are depicted. :::  ::: Our hNek6(S206A) model generated by SWISS-MODEL \[[@B50]\] shows a short region of α-helix composed of twelve residues (G^192^LGRFFSSETTA^203^) following the conserved DLG motif, with high score (Figure [3](#F3){ref-type="fig"}). A helical structure following the DFG/DLG motif is also present in hNek2(T175A) structures (PDB: [2JAV](2JAV), [2W5H](2W5H) and [2W5B](2W5B)) \[[@B26],[@B27]\] and in other kinase families, such as inactive forms of both the EGFR kinase \[[@B53]\] and Src/Hck \[[@B54]\]. Therefore, although the activation loop is missing in the electron density map of hNek7, a short helical structure is possibly present in hNek6(S206A), which was predicted in the model generated by SWISS-MODEL \[[@B50]\]. Human Nek6 is a monomeric mostly globular, though slightly elongated protein in solution, as revealed by SAXS ------------------------------------------------------------------------------------------------------------- To study hNek6 molecular structure, in addition to our homology modeling, we also performed SAXS (Small Angle X-ray Scattering) experiments for the recombinant 6xHis-hNek6(S206A). SAXS is a very useful technique for the determination of overall size, shape and oligomerization status of the macromolecules in solution \[[@B55]-[@B58]\]. Figure [4A](#F4){ref-type="fig"} shows the corrected and normalized experimental scattering curve and theoretical fitting of data by using the program GNOM \[[@B59]\]. The Guinier region providing an *R*~g~value of 32.0 ± 1.0 Å is shown in the inset. The *p*(r) function resulting from these calculations is shown in Figure [4B](#F4){ref-type="fig"}, with an inset showing the Kratky representation of the intensity curve. The Kratky plot indicates a slightly globular conformation for 6xHis-hNek6(S206A) in solution, although, as expected, the same plot also indicated the presence of flexible regions in the structure, possibly the N-terminal region and the activation loop. The *p(r)*function shows that the protein conformation is slightly elongated. The maximum dimension (*D*~max~) value obtained was approximately 110 Å and the *R*~g~value calculated from the *p*(r) function was 32.4 ± 0.8 Å, in close agreement with that calculated from the Guinier approximation. Using BSA as a standard sample, the molecular mass of 6xHis-hNek6(S206A) estimated from the SAXS data was \~42 kDa. This value indicates that the protein is a monomer in solution, since the theoretically calculated molecular mass of the monomer was \~38 kDa (calculated from the amino acid sequence using ProtParam tool \[[@B60]\]). The molecular mass and consequently the monomeric nature were also confirmed by analytical size-exclusion chromatography coupled to multi-angle light scattering (SEC-MALS). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Small Angle X-Ray Scattering (SAXS) of recombinant hNek6(S206A)**. (A) Experimental SAXS curve from hNek6(S206A) (open circles) and the theoretical fitting (solid line) calculated using the program GNOM. *Inset:*Guinier region and linear fit. (B) Pair distance distribution function *p(r)*. *Inset:*Kratky plot (*q*^*2*^*I(q) vs. q*) of the intensity curve. :::  ::: The low-resolution models obtained from the SAXS data by the combination of *ab initio*calculation and rigid body modeling methods are presented in Figure [5](#F5){ref-type="fig"}. The calculated homology model of the mutant hNek6(S206A) was used in the rigid body calculation. We displayed two typical models of the set of results (Figure [5A](#F5){ref-type="fig"} and [5B](#F5){ref-type="fig"}) and a superposition of all 10 models (Figure [5C](#F5){ref-type="fig"}) obtained in different and independent runs of the program BUNCH \[[@B61]\]. In spite of the flexibility of the N-terminal region, the NSD values of the pairwise comparison of the models obtained ranged from 0.96 to 1.20, which shows the stability of the independent calculations. In order to further compare the resulting models with the information contained in the SAXS curve, we also calculated one average molecular envelope of the 10 models (Figure [5D](#F5){ref-type="fig"}) and compared with the filtered average models obtained from the two sets of 10 purely *ab initio*model calculations (Figure [5E](#F5){ref-type="fig"} and [5F](#F5){ref-type="fig"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Low-resolution models of hNek6(S206A) obtained from the SAXS data using a combination of *ab initio*calculations and rigid body modeling**. (A and B) Two typical models selected from the set of 10 resulting models. (C) Superposition of all 10 models obtained in different and independent runs of the program BUNCH. (D) Three orthogonal views of the average molecular envelope of the 10 low resolution BUNCH models. (E) Three orthogonal views of the *ab initio*dummy atoms model. (F) Three orthogonal views of the *ab initio*dummy residues model. :::  ::: Analytical size-exclusion chromatography reveals variations in human Nek6 conformation dependent on its phosphorylation status ------------------------------------------------------------------------------------------------------------------------------ Analytical [S]{.underline}ize-[E]{.underline}xclusion [C]{.underline}hromatography (SEC) was performed for five variants of recombinant hNek6 fused to a 6xHis tag: wild-type hNek6 (6xHis-hNek6wt), activation loop mutant hNek6(S206A) (6xHis-hNek6(S206A), kinase domain of wild-type hNek6 (6xHis-hNek6(Δ1-44)), and dephosphorylated wild-type and mutant hNek6 (6xHis-hNek6wtD and 6xHis-hNek6(S206A)D). Interestingly, the dephosphorylated wild-type and mutant proteins, which were co-expressed with lambda phosphatase, were eluted at equal elution volumes (Figure [6A](#F6){ref-type="fig"}), showing the same Stokes radius of \~2.1 nm, while the more phosphorylated wild-type hNek6 showed a larger radius of \~2.6 nm and the partially phosphorylated mutant hNek6(S206A) showed an intermediate radius of \~2.4 nm (Figure [6B](#F6){ref-type="fig"}, Table [2](#T2){ref-type="table"}). As expected, 6xHis-hNek6(Δ1-44) showed the smallest radius of \~1.8 nm. Wild-type hNek6 elution curve also showed a smaller peak corresponding to a population of higher molecular weight, possibly due to aggregates. It is interesting to compare the Stokes radius estimated by SEC for 6xHis-hNek6(S206A) (\~2.4 nm) with the radius of gyration determined by SAXS (\~3.2 nm). The resulting ratio *R*~g~/*R*~s~for this protein is \~1.3 and, as *R*~g~/*R*~s~ratios are reported to vary from 0.78 for homogeneous spheres, up to values nearing 2 for extended coils and prolate ellipsoids \[[@B62]\]. This reinforces our results that hNek6 has a slightly elongated conformation, with a flexible unfolded N-terminal domain contributing to this shape. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Analytical size-exclusion chromatography of recombinant hNek6 variants fused to a 6xHis tag**. (A) Chromatographic profile of hNek6 variants (colored lines) and a mixture of standard proteins (dashed line). (B) Linear fit of known Stokes radii (*R*~s~) as a function of the measured partition coefficient (*K*~av~). The colored circles show the partition coefficient of hNek6 variants, and the black squares, the partition coefficients of the standards. The circles colors correspond to the same colors depicted in (A). The proteins were eluted in 50 mM HEPES pH 7.5, 5 mM sodium phosphate, 600 mM NaCl, 5% glycerol, as homogeneous species at different elution volumes (V~e~) using an analytical Superdex 200 10/300 GL column. Conalbumin, ovalbumin, carbonic anhydrase and ribonuclease were used as Stokes radii standards. (wt: wild-type, D: dephosphorylated). :::  ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Recombinant hNek6 Stokes radii (*R*~s~) determined by analytical size-exclusion chromatography. ::: Protein ***R***~**s**~**(nm)** **V**~**e**~**(mL)**^**a**^ **Apparent *R***~**s**~**(nm)**^**a**^ -------------------- ------------------------ ----------------------------- ---------------------------------------- Conalbumin 3.6 14.9 ± 0.1 3.5 ± 0.2 Ovalbumin 3.0 15.8 ± 0.1 2.9 ± 0.0 Carbonic Anhydrase 2.0 16.9 ± 0.1 2.2 ± 0.0 Ribonuclease 1.6 18.2 ± 0.0 1.4 ± 0.1 Nek6wt 16.4 ± 0.3 2.6 ± 0.1 Nek6wtD 17.1 ± 0.0 2.1 ± 0.0 Nek6(S206A) 16.7 ± 0.2 2.4 ± 0.0 Nek6(S206A)D 17.1 ± 0.0 2.1 ± 0.0 Nek6(Δ1-44) 17.6 ± 0.3 1.8 ± 0.1 ^a^S.E. of two measurements at different buffers. ::: SEC was also coupled to MALS ([M]{.underline}ulti-[A]{.underline}ngle [L]{.underline}ight [S]{.underline}cattering), which is an useful technique to measure the weight average molecular masses (*M*~w~) of the eluted proteins, as described in the Methods section. As expected, wild-type and mutant hNek6, dephosphorylated or not, showed a *M*~w~of \~38 kDa, while the kinase domain showed a *M*~w~of \~33 kDa (Table [3](#T3){ref-type="table"}, Additional file [1](#S1){ref-type="supplementary-material"}, Figure S2). This corroborates our experimental data from SAXS for 6xHis-hNek6(S206A) and the theoretically calculated molecular masses for all hNek6 variants, showing that hNek6 is a monomer in solution. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Recombinant hNek6 weight average molecular masses (*M*~w~) determined by SEC-MALS and melting temperatures (*T*~m~) during thermal shift denaturation. ::: Protein **Pred. *M*(kDa)**^**a**^ ***M***~**w**~**(kDa)**^**b**^ **Apparent *T***~**m**~**(°C)**^**c**^ -------------- --------------------------- -------------------------------- ---------------------------------------- Nek6wt 37.7 38.4 ± 0.6 39.5 ± 0.1 Nek6wtD 37.7 37.8 ± 2.2 40.8 ± 0.1 Nek6(S206A) 37.7 38.1 ± 2.1 38.0 ± 0.1 Nek6(S206A)D 37.7 38.4 ± 3.3 36.8 ± 0.3 Nek6(Δ1-44) 33.2 33.3 ± 1.6 41.0 ± 0.2 ^a^Molecular masses predicted by ProtParam tool \[[@B60]\]. ^b^S.E. of two measurements at different buffers. ^c^S.E. of three measurements. ::: These results suggest that, although having the same molecular mass of \~38 kDa, wild-type hNek6 is purified from bacteria more phosphorylated than its mutant variant, mainly because of their different activation/autophosphorylation status, as described by Meirelles et al. \[[@B41]\], and these different phosphorylation degrees may cause changes in protein conformation and compactness, resulting in changes in their Stokes radii. This was better visualized for both proteins when dephosphorylated by lambda phosphatase, which promoted smaller radii and, possibly, more compact or less hydrated conformations. It seems that an increase in phosphorylation induces a structural change that increases the apparent size or shape of hNek6. In fact, in most kinases, the activation loop is phosphorylated when the kinase is active, which stabilizes it in an open and extended conformation that is permissive for substrate binding \[[@B49]\]. This phosphorylated extended conformation may therefore contribute to the increase in hNek6 Stokes radius. All hNek6 variants were submitted to SEC-MALS twice, using two different buffers (the same one used for SAXS and another one containing 600 mM NaCl in order to avoid any unspecific binding to the gel filtration column resin), and the same Stokes radii for each protein were obtained in both measurements. Figure [6](#F6){ref-type="fig"} shows the results from SEC-MALS using the buffer containing 600 mM NaCl, and Tables [2](#T2){ref-type="table"} and [3](#T3){ref-type="table"} show all the results obtained from both measurements. Thermal denaturation shift assays were also performed for the five recombinant hNek6 variants described above. The results revealed a slightly higher stability of wild-type hNek6 compared to the activation loop mutant (Figure [7](#F7){ref-type="fig"}, Table [3](#T3){ref-type="table"}). This may be explained by the fact that, in many kinases, like PKA, phosphorylation of the activation loop cause global stabilization of the active site \[[@B63]\], and molecular dynamics simulation of Cdk2 has demonstrated a decrease of B-factors throughout the molecule upon phosphorylation \[[@B64]\]. A model was proposed that dephosphorylation of the activation loop leads to mutual repulsion of the positive charges that were bound to the phosphate, which leads to the destabilization of the magnesium-binding loop, movement of the αC-helix out of the active site, disturbance of the hydrophobic spine, and loosening of the N-lobe, thereby providing an explanation of the protein kinase stabilization induced by phosphorylation \[[@B48]\]. This may therefore also reflect a higher stability for the wild-type hNek6 compared to its activation loop mutant in thermal melting measurements. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Thermal shift denaturation of recombinant hNek6 variants fused to a 6xHis tag**. The graph shows representative thermal shift data (markers) and fitted curves (lines). Proteins were buffered in 10 mM HEPES pH 7.5, 150 mM NaCl and assayed at a final concentration of 2.0 μM in 25 μL volume. SYPRO-Orange (Molecular Probes) was added as a fluorescence probe at a dilution of 1 in 1000. Temperature was raised with a step of 1°C per 1.0 min from 10°C to 85°C and fluorescence readings were taken at each interval. (wt: wild-type, D: dephosphorylated). :::  ::: Conclusions =========== Our data presents the first low resolution 3D structure of hNek6 protein in solution. SAXS experiments show that hNek6 is a monomer of a mostly globular, though slightly elongated shape, which was also confirmed by analytical SEC-MALS experiments. These also showed that hNek6 conformation is dependent on its activation/phosphorylation status, a higher phosphorylation degree corresponding to a bigger Stokes radius. Thermal denaturation shift assays revealed a slightly higher stability of wild-type hNek6 compared to the activation loop mutant hNek6(S206A). Methods ======= *In silico*sequence analysis ---------------------------- Human Nek6 and Nek7 amino acid sequences were used as queries in five different secondary structure prediction databases: PredictProtein/Prof \[[@B29]\], PSIPRED \[[@B30]\], SSpro \[[@B31]\], SOPMA \[[@B32]\] and GOR4 \[[@B33]\]. Comparison of their outputs resulted in a consensus of predicted secondary structure, where each amino acid was assigned a score ranging from 1 to 5. Our Nek7 consensus of predicted secondary structure was compared to the author-approved secondary structure in PDB ([2WQM](2WQM)) as a measure to validate our analysis. We also performed disordered regions analysis for both protein sequences using nine different predictors: FoldIndex \[[@B34]\], GlobPlot Russell/Linding \[[@B35]\], PONDR VL-XT \[[@B36]\], DISpro \[[@B37]\], IUPred \[[@B38]\], DisEMBL Hot-loops, DisEMBL Remark-465, DisEMBL Loops/coils \[[@B39]\], and VSL2B \[[@B40]\]. From this, a consensus of predicted disordered regions was generated with a consensus score ranging from 0 to 9, where a score above 4 represents disorder. Additionally, NetPhosK \[[@B42]\] and NetPhos \[[@B43]\] databases were used to predict phosphorylation sites for human Nek6 and Nek7. The conserved glycine-rich sequence, the HRD and DLG motifs, the conserved residues K^74^(β3 strand) and E^93^(αC helix), the putative nuclear export signal LGDLGL based on la Cour et al., 2004 \[[@B65]\], the putative WW domain binding motifs PY and pSP based on Ingham et al., 2005 \[[@B66]\], as long as the PPLP motif, experimentally described for hNek6 by Lee et al., 2007 \[[@B67]\], were also assigned to both protein sequences. Plasmid Constructions --------------------- All plasmid constructions were developed accordingly to Meirelles et al., 2010 \[[@B41]\]. Site-directed Mutagenesis ------------------------- The hNek6 activation loop mutation S206A was introduced by PCR-based mutagenesis accordingly to Meirelles et al., 2010 \[[@B41]\]. Protein Expression and Purification ----------------------------------- Soluble full-length hNek6 wild-type - 6xHis-hNek6wt - and mutant - 6xHis-hNek6(S206A) - or truncated hNek6 wild-type kinase domain - 6xHis-hNek6(Δ1-44) - fused to a 6xHis tag were expressed and purified accordingly to Meirelles et al., 2010 \[[@B41]\]. Protein Dephosphorylation ------------------------- In order to obtain dephosphorylated wild-type and mutant hNek6, plasmids encoding either 6xHis-hNek6wt or 6xHis-hNek6(S206A) and λ phosphatase were transformed into *E. coli*BL21 (DE3/pRARE) cells that were induced and purified as described by Meirelles et al., 2010 \[[@B41]\]. Lambda phosphatase cloned into pCDF-Duet (Novagen) was kindly provided by Dr. Richard Bayliss (Section of Structural Biology, Institute of Cancer Research, London, UK). Circular dichroism ------------------ Circular dichroism (CD) spectra were recorded in a JASCO model J-810 CD spectropolarimeter equipped with Peltier-type system PFD 425S. Data were collected from 260 to 200 nm at 4°C using a quartz cuvette of 1 mm pathlength. Thirty-two spectra of purified 6xHis-hNek6wt at 5.8 μM, in 50 mM Phosphate buffer pH 7.5, 300 mM NaCl, were averaged and corrected from the baseline for buffer solvent contribution. Experimental data were analyzed using CDNN \[[@B45]\] and K2d \[[@B46]\] softwares. Comparative/Homology Molecular Modeling --------------------------------------- The comparative/homology molecular modeling and model validation were performed in a similar way to that described in Bodade et al., 2010 \[[@B68]\]. Briefly, several comparative/homology modeling tools were used: I-TASSER \[[@B69]-[@B71]\], Geno3D \[[@B72]\], 3D-JIGSAW \[[@B73]-[@B75]\], SWISS-MODEL \[[@B50]\] and MODELLER 9v8 \[[@B76]\]. The NCBI Basic Local Alignment Search Tool (BLAST, <http://www.ncbi.nlm.nih.gov/BLAST/>) was used to search the crystal structure of the closest homologue available in the Protein Data Bank (PDB, <http://www.rcsb.org/pdb/>). The input was the amino acid sequence of hNek6(S206A). The NCBI results revealed that the structure of hNek7, deposited under the code [2WQM](2WQM) in the PDB, was a very suitable template (identity score of 81% and E-value 3 × 10^-141^). This structure was used as a single template for the modeling approach. The overall stereochemical quality of the models was assessed by PROCHECK software \[[@B77]\]. The quality of the models was also evaluated by PROSA \[[@B51],[@B52]\] and by the standard validation procedures included in the automated mode of the SWISS-MODEL server \[[@B50]\]. Small Angle X-Ray Scattering Analysis ------------------------------------- The sample was first inspected by dynamic light scattering (DLS) to test for its monodispersity and then ultracentrifuged at 200.000 × g for 40 min at 4°C to remove any possible aggregates. The SAXS experiments were performed at the D02A-SAXS2 beam line at the LNLS, and data treatment and analyses were done following standard procedures similar to those described in Trindade et al., 2009 \[[@B57]\]. Briefly, the measurements were performed at 4°C and the sample-to-detector distance was 902 mm, covering a scattering vector range of 0.015Å^-1^\< q \<0.25 Å^-1^(q is the magnitude of the **q**-vector defined by q = (4π/λ)sinθ and 2θ is the scattering angle) using a wavelength of λ = 1.488 Å. The measurements were performed using two different protein concentrations in HEPES buffer (50 mM HEPES pH 7.5, 5 mM sodium phosphate, 300 mM NaCl, 5% glycerol, 200 mM imidazole): 0.5 and 1.0 mg/mL. A 8 mg/ml BSA (66 kDa) solution in the same sample buffer was used as a standard sample to estimate the molecular mass of 6xHis-hNek6(S206A) making use of the ratio of the extrapolated values of the intensity at the origin, *I*(0) \[[@B78],[@B79]\]. The radius of gyration (*R*~g~) was calculated from the Guinier approximation (valid for q*R*~g~\< 1.3) \[[@B80]-[@B82]\] and also from the pair distance distribution function, *p*(r), which was obtained using the program GNOM \[[@B56]\]. The maximum dimension (*D*~max~) of the molecule was obtained from the *p(r)*function. The Kratky plot (*q*^2^I(*q*) vs. *q*) \[[@B81],[@B82]\] was used to analyze the compactness of the protein conformation. Low resolution SAXS-based modeling ---------------------------------- The low resolution model of 6xHis-hNek6(S206A) was obtained from the SAXS data using a combination of *ab initio*calculation and rigid body modeling methods. Taking advantage of the homology model obtained, we used the program BUNCH \[[@B61]\] to model the protein. No symmetry restraints were used in the calculation. We would like to mention that no unique solution can be obtained from these calculations. For this reason, 10 independent calculations were run for each sample data. The multiple solutions were analyzed and the reliability and stability of the set of models were estimated. A pairwise comparison and the normalized spatial discrepancy (NSD) evaluation was performed using the DAMAVER program suite \[[@B83]\] complemented by the SUPCOMB \[[@B84]\] routine. Analyzing the NSD values (which describe the dissimilarity between pairs of models of the several calculations), the models with common features led to the selection of a representative, low resolution conformation for hNek6(S206A) protein. Models were displayed by the PyMOL program \[[@B85]\]. For comparison purposes, two other low resolution models were also obtained by using two different *ab initio*approaches: the *dummy atoms*method implemented in the program DAMMIN \[[@B86]\] and the *dummy residues*method implemented in GASBOR \[[@B87]\]. The procedures were similar to those described in Trindade et al. \[[@B57]\]. SEC-MALS Analysis ----------------- We used Analytical [S]{.underline}ize-[E]{.underline}xclusion [C]{.underline}hromatography coupled to [M]{.underline}ulti-[A]{.underline}ngle [L]{.underline}ight [S]{.underline}cattering (SEC-MALS) to estimate the hydrodynamic or Stokes radii (*R*~s~) of recombinant hNek6wt, hNek6(S206A), hNek6(Δ1-44) and dephosphorylated hNek6wt and hNek6(S206A), all fused to a 6xHis tag. SEC was performed with an analytical Superdex 200 10/300 GL column using an ÄKTA FPLC system (GE Healthcare) equilibrated with two column volumes of 50 mM HEPES pH 7.5, 5 mM sodium phosphate, 600 mM NaCl, 5% glycerol, at a flow rate of 0.5 ml/min, at 20°C. Recombinant hNek6 variants at concentrations ranging from 0.2 to 0.7 mg/ml and a mixture of standard proteins with known Stokes radii (conalbumin: 3.64 nm, 3.2 mg/ml; ovalbumin: 3.05 nm, 4.2 mg/ml; carbonic anhydrase: 2.30 nm, 3.0 mg/ml; and ribonuclease: 1.64 nm, 3.4 mg/ml) (Sigma) were loaded onto the column and their elution profiles were monitored by absorbance at 280 nm. The Stokes radius of each hNek6 variant was estimated by a linear fit of the Stokes radii of the standard proteins versus the partition coefficient *K*~av~\[[@B88],[@B89]\] as described by the equation: *K*~av~= *V*~e~- *V*~o~/*V*~t~- *V*~o~, where *V*~e~is the elution volume of the protein, *V*~o~the void volume and *V*~t~is the total volume of the column. The SEC was also coupled to a DAWN TREOS™ MALS instrument (Wyatt Technology). The on-line measurement of the intensity of the Rayleigh scattering as a function of the angle of the eluting peaks in SEC was used to determine the weight average molecular masses (*M*~w~) of the eluted proteins \[[@B90]\], using the ASTRA™ (Wyatt Technologies) software. SEC-MALS measurements were performed using two different buffers: the first one described above for SEC and a second one also used in SAXS experiments (50 mM HEPES pH 7.5, 5 mM sodium phosphate, 300 mM NaCl, 5% glycerol, 200 mM imidazole). The chromatographic profile of the recombinant hNek6 variants were the same in both measurements and the mean and standard errors of their *M*~w~and *M*~n~were calculated. Thermal Shift Assays -------------------- Thermal shift assays were performed based on a protocol devised by the Structural Genomics Consortium \[[@B91]\] using a real time PCR machine 7300 (Applied Biosystems). Proteins were buffered in 10 mM HEPES pH 7.5, 150 mM NaCl and assayed at a final concentration of 2.0 μM in 25 μL volume. SYPRO-Orange (Molecular Probes) was added as a fluorescence probe at a dilution of 1 in 1000. The emission filter for the SYPRO-Orange dye was set to 580 nm. Temperature was raised with a step of 1°C per 1.0 min from 10°C to 85°C and fluorescence readings were taken at each interval. OriginPro 8 software was used to fit data to the Boltzmann equation, y = LL+(UL-LL)/1+exp((*T*~m~-x)/a), where LL and UL are the slopes of the native and denatured baselines, *T*~m~is the apparent melting temperature and a describes the slope of the denaturation. *T*~m~values were calculated by determination of the maximum of the first derivative. Authors\' contributions ======================= GVM and JK conceived and designed the experiments, analyzed the data and wrote the manuscript. GVM performed the experiments. JCS performed comparative molecular modeling and SAXS experiments and interpreted them together with ICLT. YAM helped in SEC-MALS and CD experiments which were analyzed together with CHIR. JK supervised the project. All authors read and approved the final version of the manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Supplemental Figures S1 and S2, PROCHECK and PROSA analysis results of the validation procedures of the hNek6(S206A) homology/comparative model and Analytical SEC-MALS of the five recombinant protein variants of Nek6**. Figure S1: PROCHECK and PROSA analysis results of the validation procedures of the hNek6(S206A) homology/comparative model. (A) Ramachandran Plot calculated using the program PROCHECK. (B) Plot of the residue score showing the local model quality by plotting energies as a function of the residue sequence position using PROSA. In general, positive values correspond to problematic or erroneous parts of the structure. Here, the plots were smoothed by calculation the average energy over 10- and 40-residues. This average is needed because of the large fluctuation in a plot of single residue energies. (C) The Z-score indicated overall model quality using PROSA. The Z-score of the hNek6(S206A) model was -7.14 (black point). The plot contains the Z-scores of all experimentally determined protein chains in the current PDB. The structure determined by X-ray and NMR are distinguished by different colors. Figure S2: Analytical SEC-MALS of recombinant (A) hNek6wt, (B) hNek6wtD, (C) hNek6(S206A), (D) hNek6(S206A)D and (E) hNek6(Δ1-44). The Mw determined by MALS correspond to a monomer in all five cases. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Financially supported by: Fundação de Amparo à Pesquisa do Estado São Paulo, the Conselho Nacional de Pesquisa e Desenvolvimento and the LNLS. We thank Maria Eugenia R. Camargo for technical assistance, Rodrigo Martinez for the technical support at the SAXS2 beamline, and Dr. Richard Bayliss (Section of Structural Biology, Institute of Cancer Research, London, UK) for providing the pCDF-Duet lambda phosphatase construct.

PubMed Central

2024-06-05T04:04:19.454860

2011-2-14

PMC3053221

Background ========== Developing a patient safety culture was one of the recommendations made by the Institute of Medicine to assist hospitals in improving patient safety \[[@B1],[@B2]\]. Assessing the organization\'s existing safety culture is the first stage of developing a safety culture \[[@B3]\]. Patient safety culture assessments, required by international accreditation organizations, allow healthcare organizations to obtain a clear view of the patient safety aspects requiring urgent attention, identify the strengths and weaknesses of their safety culture \[[@B4]\], help care giving units identify their existing patient safety problems \[[@B5]\], and benchmark their scores with other hospitals \[[@B6]\]. According to literature, the major predictors of a positive patient safety culture in healthcare organizations specifically hospitals include communication founded on mutual trust, good information flow, shared perception of the importance of safety, organizational learning, commitment from management and leadership, and the presence of a non-punitive approach to incident and error reporting \[[@B7]\]. Patient safety culture outcomes include the staff members\' perception of safety, the willingness of staff members to report events, the number of events reported, and an overall patient safety grade given by staff members to their units \[[@B8]\]. A multitude of evidence has been published in the area of patient safety culture in recent years. Some of the available evidence tackles patient safety culture issues that require attention, factors affecting incident reporting by hospital staff, the role of workplace environment in shaping safety, and steps that can be followed to improve safety. Despite the wealth of evidence on patient safety culture, limited evidence still exists about the linkage between predictors and outcomes of patient safety culture especially in countries of the Eastern Mediterranean Region. One of the first efforts to assess the culture of safety in hospitals in the region was conducted in Lebanon by El-Jardali et al. \[[@B9]\]. Lebanese Context ---------------- The study by El-Jardali et al. (2010) entitled \"The Current state of Patient Safety Culture in Lebanese Hospitals: A study at Baseline\" \[[@B9]\] utilized an Arabic translated version of the *Hospital Survey on Patient Safety Culture*(HSOPSC) \[[@B8]\]. It aimed at identifying the most critical issues related to patient safety culture and potential strategies to implement the patient safety accreditation standards in the light of a newly added chapter to the Lebanese handbook of hospital accreditation \[[@B10]\]. The HSOPSC measures 12 patient safety culture composites representing several patient safety culture predictors (See Box 1). The HSOPSC also requires respondents to give their work area/unit a patient safety grade and to answer a question on the number of events reported in the past 12 months \[[@B8]\]. Calculating the percentage of positive responses for each composite revealed that the composites with the highest positive ratings were teamwork within units, hospital management support for patient safety, and organizational learning and continuous improvement. However, composites with the lowest ratings were teamwork across hospital units, hospital handoffs and transitions, staffing, and non-punitive response to error \[[@B9]\]. Approximately 60% of respondents reported not completing any event reports in the past 12 months and over 70% gave their units an \"Excellent/Very Good\" patient safety grade. Bivariate and multivariate analyses revealed significant differences across hospitals of different size and accreditation status \[[@B9]\]. Study findings outlined above represent the first component of data analysis for that stage \[[@B9]\]. Findings provided evidence that communication across units, staffing, event reporting, and the culture of response to error were major patient safety culture issues \[[@B9]\]. This paper, though, will further explore the association between patient safety culture predictors and outcomes, taking into consideration respondent and hospital characteristics. In addition, it will examine the correlation between the patient safety culture composites. Thus, the objective of this paper is to address the afore-mentioned objectives using bivariate as well as multivariate analyses. Methods ======= Study Design, Setting, and Sample --------------------------------- The study adopted a cross sectional research design utilizing a customized version of the HSOPSC. A pilot testing phase preceded data collection in order to ensure the validity and reliability of the Arabic translated version of the questionnaire. Of the 126 hospitals registered in the Lebanese Syndicate of Private Hospitals that were contacted and asked to participate, 68 consented \[[@B9]\]. The survey targeted hospital employees including physicians, nurses, clinical and non-clinical staff, pharmacy and laboratory staff, dietary and radiology staff, supervisors, and hospital managers. The questionnaire was distributed to 12,250 hospital employees and 6807 were returned yielding an overall response rate of 55.56%. Additional details on the study methodology can be found in El-Jardali et al. \[[@B9]\]. Survey Measures and Outcome Variables ------------------------------------- The HSOPSC is composed of 42 items that measure 12 composites of patient safety culture (See Box1, Additional file[1](#S1){ref-type="supplementary-material"} ). Items were scored using a five-point scale reflecting the agreement rate on a five-point frequency scale. The percentage of positive responses for each item was calculated; negatively worded items were reversed when computing percent positive response. Composite level scores were computed by summation of the items within the composite scales and dividing by the number of items with non-missing values \[[@B9]\]. Two of the composites (frequency of events reported and overall perception of safety) are two of the four patient safety culture outcome variables \[[@B8]\]. The remaining two outcome variables are the patient safety grade and the number of events reported \[[@B8]\]. Bivariate Analysis ------------------ Bivariate analyses were used to examine the associations between patient safety culture composites and differences across hospitals of different size and accreditation status. Pearson correlations were used to examine the association between the patient safety culture composites. ANOVA f-test with multiple comparison corrected using the Bonferroni method was used to examine the association between the two outcome variables (patient safety grade and the number of events reported) with the remaining patient safety culture composites. Student T-Test and ANOVA f-test with multiple comparison corrected using the Bonferroni method were then used to examine how trends in the outcome variables (frequency of events reported and overall perception of safety) differ across hospital and respondent characteristics. Finally, cross tables were constructed and chi-square tests were used to assess how trends in the outcome variables (patient safety grade and the number of events reported) differed across respondent and hospital characteristics. Multivariate Regression Analysis -------------------------------- The four outcome variables were regressed against the 10 composite scores, respondent\'s position in the hospital, accreditation status, and hospital size. Since the data was clustered by hospital, we used appropriate statistical techniques to control for this effect. Four regression models were constructed, two adopted Generalized Estimating Equations (the two categorical outcome variables: number of events reported and patient safety grade) and the other two models followed a linear mixed regression model (the two composites for frequency of events reported and overall perception of safety). Results ======= Comparison of Means for the Frequency of Events Reported and the Overall Perception of Safety across Respondent and Hospital Characteristics -------------------------------------------------------------------------------------------------------------------------------------------- Significant differences were observed between units and positions when comparing results across respondent and hospital characteristics. For both the frequency of events reported and the overall perception of safety, significantly higher means were observed for diagnostics (mean = 3.97 ± 1.01; mean = 3.96 ± 0.68) as compared to surgical (mean = 3.78 ± 1.05; mean = 3.82 ± 0.67) and medical units (mean = 3.93 ± 0.99; mean = 3.83 ± 0.68) (See Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Comparison of means for two outcome composite scores across hospital and respondent characteristics *(identical letters represent significance between indicated groups)* ::: Frequency of Events Reported Overall Perception of Safety ---------------------------------------------------- ------------------------------ ------------------------------ --------------- ------------- ------------ --------------- **Unit**  Many different hospital units/no specific unit 3.89 (0.96) a, f ***\<0.001*** 3.72 (0.66) a, d, e, f ***\<0.001***  Administration 3.78 (1.16) b, f 3.72 (0.75) b, e,  Medical 3.93 (0.99) c, d 3.83 (0.68) a, c, e  Surgical 3.78 (1.05) c, d, e, f 3.82 (0.67) a, d, e,  Diagnostics 3.97 (1.01) b, d, e 3.96 (0.68) a, b, c, e  Other 4.06 (0.98) a, b, d, f 3.91 (0.64) a, b, f  N 5707 5201 **Position**  Nurse 3.89 (1.00) a, e, f ***\<0.001*** 3.80 (0.66) a, c ***\<0.001***  Physician 3.78 (0.92) b, f 3.69 (0.75) b, c, d  Pharmacist 3.87 (1.20) 3.83 (0.90)  Other health professions 3.95 (0.93) c, e, 3.90 (0.77)  Unit assistant/clerk/secretary/Technician 3.92 (1.05) d, e, 3.92 (0.68) a, b  Administration 3.92 (1.06) 3.75 (0.85)  Quality and Safety 3.49 (1.06) a, c, d, e 3.86 (0.63)  Other 4.04 (0.99) a, b, e, f 3.90 (0.71) b, d  N 5925 5407 **Experience at hospital**  Less than 1 year 3.78 (1.08) a, b, c, d ***0.001*** 3.81 (0.72) ***0.003***  1 to 5 years 3.88 (1.00) 3.82 (0.67)  6 to 10 years 3.94 (1.00) a, b 3.83 (0.67)  11 to 15 years 3.98 (0.98) a, c 3.88 (0.65) a  16 to 20 years 3.91 (1.04) 3.82 (0.69)  21 years or more 4.00 (1.02) a, d 3.74 (0.79) a  N 6132 5573 **Hospital Size (El-Jardali et al., 2010)**  Small (\<100 beds) 3.95 (1.00) a, b, c ***0.001*** 3.84 (0.67) a ***0.012***  Medium (100-199 beds) 3.87 (1.03) a, b 3.80 (0.69)  Large (\>=200 beds) 3.81 (1.08) a, c 3.76 (0.71) a  N 6307 5742 **Interaction with patients**  Yes 3.90 (1.00) 0.517 3.82 (0.68) 0.092  No 3.93 (1.05) 3.87 (0.73)  N 5707 5201 **Accreditation Status (El-Jardali et al., 2010)**  Yes 3.91 (1.03) ***0.047*** 3.84 (0.69) ***\<0.001***  No 3.84 (1.03) 3.71 (0.68)  N 6307 5742 ::: Significantly higher means were observed for nurses (mean = 3.89 ± 1.00; mean = 3.80 ± 0.66) than physicians (mean = 3.78 ± 0.92; mean = 3.69 ± 0.75) on the frequency of events reported and the overall perception of safety respectively (See Table [1](#T1){ref-type="table"}). For frequency of events reported, higher means (and thus more events) were observed for increasing years of experience whereas the opposite trend prevailed for overall perception of safety (See Table [1](#T1){ref-type="table"}). Small hospitals (mean = 3.95 ± 1.00) had a significantly higher frequency of events reported in comparison to medium (mean = 3.87 ± 1.03) and large hospitals (mean = 3.81 ± 1.08). Small hospitals (mean = 3.84 ± 0.67) also had a better overall perception of safety than large hospitals (mean = 3.76 ± 0.71). Accredited hospitals (mean = 3.91 ± 1.03) had a significantly higher frequency of events reported than non-accredited hospitals (mean = 3.84 ± 1.03). Accredited hospitals (mean = 3.84 ± 0.69) also had a better overall perception of safety than non-accredited hospitals (mean = 3.71 ± 0.68) (See Table [1](#T1){ref-type="table"}). Correlations between Patient Safety Culture Composites ------------------------------------------------------ While significant correlations were observed among all patient safety culture composites, differences in the strength of the correlation were observed (See Table [2](#T2){ref-type="table"}). The composite measuring staffing showed the weakest correlation with the outcome on frequency of event reporting (Pearson r = 0.107). The composite measuring feedback and communication about errors had the strongest correlation with this outcome variable (Pearson r = 0.378) (See Table [2](#T2){ref-type="table"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Correlation between patient safety culture composites ::: ------------------------------------------------------------------------------------------------------------------------------ Frequency of Events\ Overall Perception of\ Reported Safety -------------------------------------------------------------- ---------------------- ------------------------ -------- ------ Supervisor/manager expectations and actions promoting safety .206\* 5878 .371\* 5467 Organizational Learning-Continuous Improvement .221\* 5884 .303\* 5542 Teamwork Within Hospital Units .195\* 6008 .308\* 5531 Communication Openness .267\* 6017 .276\* 5539 Feedback and Communication About Errors .378\* 6056 .288\* 5555 Non-punitive Response to Error .116\* 5975 .179\* 5560 Staffing .107\* 5168 .310\* 4959 Hospital Management Support for Patient Safety .243\* 6160 .358\* 5614 Hospital Handoffs and Transitions .159\* 5139 .355\* 4832 Teamwork Across Hospital Units .250\* 5988 .297\* 5498 ------------------------------------------------------------------------------------------------------------------------------ \*Correlation is significant at the 0.01 level (2-tailed). ::: The composite with the weakest correlation with the outcome variable measuring perceptions of patient safety was non-punitive response to error (Pearson r = 0.179) while that with the strongest correlation was the composite measuring supervisor/manager expectations and actions promoting safety (Pearson r = 0.371) (See Table [2](#T2){ref-type="table"}). Comparison of Means between Outcome Variables and Patient Safety Composites --------------------------------------------------------------------------- Significantly different means were reported for patient safety grade with all patient safety composite scores. Respondents who gave \"Excellent/Very Good\" patient safety grades had the highest mean scores for patient safety composites (See Table [3](#T3){ref-type="table"}). The number of events reported was significantly associated with the composites measuring communication openness, feedback and communication about errors, non-punitive response to error, hospital handoffs and transitions, and teamwork across hospital units. The highest means were observed when reporting more than 5 events for the composites on communication openness, feedback and communication about errors, and non-punitive response to error. The opposite was observed for the composites measuring hospital handoffs and transitions and teamwork across hospital units where the highest means were observed when no events were reported (see Table [3](#T3){ref-type="table"}). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Comparison of means between patient safety grade and number of events reported with patient safety culture composite scores ::: Patient Safety Grade Number of Events Reported -------------------------------------------------------------- ---------------------- --------------------------- ------------- ------------- ------- ------------- ------------- ------------- Supervisor/manager expectations and actions promoting safety a,b,c 3.00 (0.84) 3.39 (0.69) 3.82 (0.62) 3.71 (0.67) 3.70 (0.67) 3.72 (0.72) Organizational Learning-Continuous Improvement a,b,c 3.26 (0.93) 3.72 (0.63) 4.04 (0.56) c 3.93 (0.62) 3.96 (0.58) 4.02 (0.62) Teamwork Within Hospital Units a,b,c 3.37 (0.88) 3.75 (0.66) 4.14 (0.57) 4.02 (0.65) 4.04 (0.59) 4.04 (0.68) Communication Openness a,b,c 2.88 (0.92) 3.33 (0.86) 3.76 (0.80) b,c 3.62 (0.87) 3.65 (0.81) 3.77 (0.87) Feedback and Communication About Errors a,b,c 2.88 (1.09) 3.51 (0.87) 4.12 (0.76) a,b,c 3.90 (0.88) 3.97 (0.81) 4.08 (0.81) Non-punitive Response to Error b,c 2.42 (0.91) 2.46 (0.76) 2.66 (0.79) b,c 2.58 (0.78) 2.62 (0.78) 2.74 (0.89) Staffing a,b,c 2.41 (0.80) 2.65 (0.68) 2.97 (0.75) 2.89 (0.77) 2.91 (0.73) 2.89 (0.76) Hospital Management Support for Patient Safety a,b,c 3.04 (0.97) 3.62 (0.71) 4.14 (0.65) 4.00 (0.72) 4.00 (0.71) 3.97 (0.83) Hospital Handoffs and Transitions a,b,c 2.80 (0.76) 3.03 (0.76) 3.49 (0.76) b,c 3.40 (0.80) 3.34 (0.76) 3.19 (0.82) Teamwork Across Hospital Units a,b,c 2.64 (0.69) 3.09 (0.67) 3.56 (0.68) b 3.44 (0.73) 3.42 (0.68) 3.34 (0.77) Patient Safety Grade a\. Significant difference between \"Poor or Failing\" and \"Acceptable\" b\. Significant difference between \"Poor or Failing\" and \"Excellent/Very Good\" c\. Significant difference between \"Acceptable\" and \"Excellent/Very Good\" Number of Events Reported a\. Significant difference between \"No events reported\" and \"1 to 5 events reported\" b\. Significant difference between \"No events reported\" and \"\> 5 events reported\" c\. Significant difference between \"1 to 5 events reported\" and \"\> 5 events reported\" ::: Distribution of the Patient Safety Grade and the Number of Events Reported across Respondent and Hospital Characteristics ------------------------------------------------------------------------------------------------------------------------- ### Patient Safety Grade The highest percentage of respondents reporting \"Excellent/Very Good\" patient safety grades were those working in diagnostics (83.6%). The highest percentage of respondents reporting a \"Poor/Failing\" patient safety grade (4.2%) worked across many different hospital units. Although the majority of respondents gave their units an \"Excellent/Very Good\" patient safety grade, physicians (63.3%) and administrators (63.3%) were the least to report an \"Excellent/Very Good\" patient safety grade (See Table [4](#T4){ref-type="table"}). ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Distribution of two outcome variables across hospital and respondent characteristics ::: Patient Safety Grade Number of Events Reported -------------------------------------------------------------------- ---------------------- --------------------------- ------ ------- ----- ------ --------------- ------ ------- ------ ------- ----- ------- --------------- **Work area/unit where respondents spend most of their work time** Many different hospital units/no specific unit 458 64.2% 225 31.6% 30 4.2% ***\<0.001*** 323 49.9% 266 41.1% 58 9.0% ***\<0.001*** Administration 353 68.8% 142 27.7% 18 3.5% 311 63.5% 116 23.7% 63 12.9% Medical 1181 72.5% 414 25.4% 35 2.1% 831 57.6% 525 36.4% 86 6.0% Surgical 1049 72.3% 359 24.7% 43 3.0% 769 59.2% 457 35.2% 74 5.7% Diagnostics 602 83.6% 112 15.6% 6 0.8% 390 60.4% 202 31.3% 54 8.4% Other 480 81.8% 99 16.9% 8 1.4% 323 60.8% 156 29.4% 52 9.8% **Position at the hospital** Nurse 2581 71.3% 941 26.0% 99 2.7% ***\<0.001*** 1845 57.2% 1190 36.9% 191 5.9% ***\<0.001*** Physician 148 63.3% 80 34.2% 6 2.6% 117 57.1% 71 34.6% 17 8.3% Pharmacist 53 77.9% 15 22.1% 0 0.0% 23 37.7% 22 36.1% 16 26.2% Other health professions 83 73.5% 25 22.1% 5 4.4% 61 55.5% 44 40.0% 5 4.5% Unit assistant/clerk/secretary/Technician 676 79.6% 162 19.1% 11 1.3% 521 68.4% 195 25.6% 46 6.0% Administration 107 63.3% 53 31.4% 9 5.3% 81 48.5% 59 35.3% 27 16.2% Quality and Safety 72 67.9% 30 28.3% 4 3.8% 41 41.8% 28 28.6% 29 29.6% Other 538 81.6% 115 17.5% 6 0.9% 374 61.0% 187 30.5% 52 8.5% **Years of experience at hospital** Less than 1 year 556 70.8% 203 25.9% 26 3.3% ***0.001*** 476 68.7% 180 26.0% 37 5.3% ***\<0.001*** 1 to 5 years 1734 72.4% 586 24.5% 72 3.0% 1229 56.3% 788 36.1% 165 7.6% 6 to 10 years 882 74.2% 289 24.3% 18 1.5% 599 55.5% 386 35.8% 94 8.7% 11 to 15 years 625 78.2% 160 20.0% 14 1.8% 414 58.2% 246 34.6% 51 7.2% 16 to 20 years 277 73.5% 96 25.5% 4 1.1% 197 60.6% 107 32.9% 21 6.5% 21 years or more 343 73.1% 120 25.6% 6 1.3% 244 58.9% 136 32.9% 34 8.2% **Hospital Size (El-Jardali et al., 2010)** Small (\<100 beds) 2214 73.1% 748 24.7% 66 2.2% 0.563 1600 58.2% 948 34.5% 199 7.2% 0.179 Medium (100-199 beds) 1751 74.0% 553 23.4% 61 2.6% 1235 59.1% 684 32.7% 172 8.2% Large (\>=200 beds) 550 72.5% 187 24.6% 22 2.9% 389 58.2% 240 35.9% 39 5.8% **Interaction with patients** Yes 3562 73.1% 1196 24.5% 119 2.4% 0.153 2486 56.9% 1574 36.0% 310 7.1% ***\<0.001*** No 805 75.6% 239 22.4% 21 2.0% 648 65.1% 266 26.7% 81 8.1% **Hospital Accreditation Status (El-Jardali et al., 2010)** Yes 3871 74.2% 1224 23.5% 121 2.3% ***\<0.001*** 2687 57.5% 1621 34.7% 361 7.7% ***0.001*** No 644 68.8% 264 28.2% 28 3.0% 537 64.2% 251 30.0% 49 5.9% ::: The association between years of experience and patient safety grade is similar to an inverted J-shaped association. Respondents with less than 1 year of experience were less likely to give \"Excellent/Very Good\" patient safety grade but this increased as years of experience gradually increased. However, after experience exceeded 16 years, the percentage of respondents giving \"Excellent/Very Good\" patient safety grade started to decrease gradually (See Table [4](#T4){ref-type="table"}). As detailed in El-Jardali et al. (2010), respondents working at accredited hospitals were more likely to give their units an \"Excellent/Very Good\" patient safety grade than respondents working at non-accredited hospitals (74.2% vs. 68.8%) (See Table [4](#T4){ref-type="table"}). ### Number of Events Reported Respondents working in administrative units were most likely to report more than 5 events over the past year (12.9%) while respondents working in surgical units were the least likely to report more than 5 events over the past year (5.7%) (See Table [4](#T4){ref-type="table"}). Quality and safety officers had the highest percent reporting of more than 5 events (29.6%) followed by pharmacists (26.2%). It is worth noting that 57.1% of physicians and 57.2% of nurses reported no events over the past year (See Table [4](#T4){ref-type="table"}). Respondents with less than 1 year of experience were the largest group of respondents to report no events over the past 12 months (68.7%). This percentage dropped as years of experience increased and then rose again after experience exceeded 10 years (See Table [4](#T4){ref-type="table"}). Interaction with patients was significantly associated with number of events reported where respondents having no direct contact with patients were more likely to report more than 5 events over the past 12 months (8.1%). Respondents working at accredited hospitals were more likely to report more than 5 events (7.7%) over the past 12 months as compared to respondents working at non-accredited hospitals (5.9%) (See Table [4](#T4){ref-type="table"}). Generalized Estimating Equations for the Patient Safety Composite Scores and Respondent and Hospital Characteristics against the Patient Safety Grade and the Number of Events Reported (See Table [5](#T5){ref-type="table"}) ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Generalized Estimating Equations ::: --------------- -------------------------------------------------------------- -------------------------- ------------------------------- -------------------- --------------- **Patient Safety Grade** **Number of Events Reported** **OR (95% CI)** **P-value** **OR (95% CI)** **P-value** Composite Supervisor/manager expectations and actions promoting safety 1.23 (1.03 - 1.47) ***0.024*** 0.93 (0.83 - 1.03) 0.152 Organizational Learning-Continuous Improvement 1.30 (1.03 - 1.65) ***0.029*** 1.11 (0.94 - 1.32) 0.227 Teamwork Within Hospital Units 1.64 (1.40 - 1.93) ***\<0.001*** 0.97 (0.85 - 1.12) 0.708 Communication Openness 1.09 (0.94 - 1.26) 0.256 1.03 (0.92 - 1.15) 0.579 Feedback and Communication About Errors 1.54 (1.33 - 1.78) ***\<0.001*** 1.17 (1.03 - 1.32) ***0.013*** Non-punitive Response to Error 1.02 (0.91 - 1.15) 0.768 1.10 (0.99 - 1.22) 0.066 Staffing 1.31 (1.16 - 1.48) ***\<0.001*** 1.05 (0.93 - 1.19) 0.380 Hospital Management Support for Patient Safety 1.85 (1.53 - 2.28) ***\<0.001*** 1.02 (0.90 - 1.16) 0.703 Hospital Handoffs and Transitions 1.36 (1.15 - 1.61) ***\<0.001*** 0.79 (0.68 - 0.91) ***0.003*** Teamwork Across Hospital Units 1.14 (0.92 - 1.40) 0.226 0.94 (0.82 - 1.08) 0.409 Position Physician 0.79 (0.50 - 1.23) 0.292 1.30 (0.83 - 2.03) 0.258 Pharmacist 1.07 (0.26 - 4.33) 0.925 2.14 (0.77 - 5.97) 0.146 Other health professions 0.73 (0.67 - 1.49) 0.389 1.22 (0.60 - 2.46) 0.579 Unit assistant/Clerk/Secretary/Technician 1.79 (1.17 - 2.74) ***0.008*** 0.78 (0.55 - 1.13) 0.190 Administration 0.39 (0.27 - 0.57) ***\<0.001*** 3.57 (1.99 - 6.40) ***\<0.001*** Quality and Safety 0.93 (0.39 - 2.58) 0.887 3.14 (1.24 - 7.91) ***0.015*** Other 1.66 (1.12 - 2.48) ***0.012*** 1.28 (0.95 - 1.72) 0.104 Nurse 1 1 Accredited Yes 1.13 (0.69 - 1.46) 0.365 1.23 (0.88 - 1.73) 0.232 No 1 1 Hospital Size Small 0.83 (0.88 - 1.14) 0.246 1.29 (0.67 - 2.46) 0.450 Medium 1.03 (0.70 - 1.42) 0.859 1.22 (0.64 - 2.32) 0.542 Large 1 1 N 3653 3207 --------------- -------------------------------------------------------------- -------------------------- ------------------------------- -------------------- --------------- ::: ### Patient Safety Grade Results showed that a one unit increase in the composite score for supervisor/manager expectations and actions to promote patient safety increased the odds of reporting a better patient safety grade by 1.23 (95%CI = 1.03-1.47; P-Value = 0.024). Similarly, the odds of reporting a higher patient safety grade increased by 1.30 (95% CI = 1.03-1.65; P-Value = 0.029) for a one unit increase in the composite score for organizational learning and continuous improvement. Moreover, a one unit increase in the composite score on teamwork within hospital units increased the odds of reporting a higher patient safety grade by 1.64 (95%CI = 1.40-1.93; P-Value \< 0.001). The odds of reporting a higher patient safety grade increased by 1.54 (95% CI = 1.33-1.78; P-Value \< 0.001) for a one unit increase in the composite score for feedback and communication about errors. The odds of reporting a higher patient safety grade increased by 1.31 (95%CI = 1.16-1.48; P-Value \< 0.001) for a one unit increase in the composite score for staffing. A one unit increase in the composites score for hospital management support for patient safety and hospital handoffs and transitions increased the odds of reporting a higher patient safety grade by 1.85 (95% CI = 1.53-2.28; P-Value \< 0.001) and 1.36 (95%CI = 1.15-1. 61; P-Value \< 0.001) respectively (See Table [5](#T5){ref-type="table"}). Respondents who held positions as unit assistants and clerks had 1.79 higher odds (95%CI = 1.16-2.74; P-Value = 0.008) of reporting a higher patient safety grade as compared to the nurses while those working in the administration had 0.39 lower odds (95%CI = 0.27-0.57; P-Value \< 0.001) of reporting a higher patient safety grade. Respondents working in \"other\" hospital departments had 1.66 higher odds (95% CI = 1.12-2.48, P-Value = 0.012) of reporting a higher patient safety grade. ### Number of Events Reported A one unit increase in the composite score for feedback and communication about errors increased the odds of reporting a higher number of events by 1.17 (95%CI = 1.03-1.32; P-Value = 0.013). Similarly, the odds of reporting a higher number of events decreased by 0.79 (95% CI = 0.68-0.91; P-Value = 0.003) for a one unit increase in the composite score for hospital handoffs and transition (See Table [5](#T5){ref-type="table"}). Respondents who reported working in administration had 3.57 higher odds (95%CI = 1.99-6.40; P-Value \< 0.001) of reporting a higher number of events compared to nurses. Moreover, respondents working in Quality and Safety had 3.14 higher odds (95%CI = 1.24-7.91) of reporting a higher number of events compared to nurses (See Table [5](#T5){ref-type="table"}). Linear Mixed Regression for the Patient Safety Composite Scores and Respondent and Hospital Characteristics against the Frequency of Events Reported and the Overall Perception of Safety (See Table [6](#T6){ref-type="table"}) -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ### Frequency of Events Reported Mixed linear regression analysis showed that a one unit increase in the score on organizational learning and continuous improvement increased the frequency of events reported by 0.159 (P-Value \< 0.001). An increase of 0.044 (P-Value = 0.05) in the composite measuring communication and openness was observed for one unit increase in frequency of events reported. A one unit increase in the score on feedback and communication about errors increased the frequency of events reported by 0.366 (P-Value \< 0.001). An increase of 0.040 (P-Value = 0.06) in the frequency of events reported was observed for a one unit increase in the score on non-punitive response to errors. Furthermore, an increase of 0.050 (P-Value = 0.05) in the frequency of events reported was observed for a one unit increase in the score on hospital management support for patient safety. A one unit increase in the score on teamwork across hospital units was found to increase the frequency of events reported by 0.100 (P-Value = 0.001) (See Table [6](#T6){ref-type="table"}). ::: {#T6 .table-wrap} Table 6 ::: {.caption} ###### Linear Mixed Model Regression ::: --------------- -------------------------------------------------------------- ---------------------------------- ---------------------------------- --------------------------- --------------- **Frequency of Events Reported** **Perception of Patient Safety** **Beta (Standard Error)** **P-value** **Beta (Standard Error)** **P-value** Composite Supervisor/manager expectations and actions promoting safety -0.019 (0.029) 0.52 0.094 (0.016) ***\<0.001*** Organizational Learning-Continuous Improvement 0.159 (0.029) ***\<0.001*** 0.129 (0.016) ***\<0.001*** Teamwork Within Hospital Units 0.000042 (0.028) 1.00 0.111 (0.016) ***\<0.001*** Communication Openness 0.044 (0.023) ***0.05*** 0.039 (0.013) ***\<0.001*** Feedback and Communication About Errors 0.366 (0.023) ***\<0.001*** 0.021 (0.013) 0.10 Non-punitive Response to Error 0.040 (0.021) ***0.06*** 0.031 (0.012) ***0.01*** Staffing -0.017 (0.023) 0.46 0.120 (0.013) ***\<0.001*** Hospital Management Support for Patient Safety 0.050 (0.026) ***0.05*** 0.099 (0.015) ***\<0.001*** Hospital Handoffs and Transitions 0.028 (0.025) 0.26 0.163 (0.014) ***\<0.001*** Teamwork Across Hospital Units 0.100 (0.029) ***0.001*** -0.045 (0.017) ***\<0.001*** Position Physician -0.042 (0.076) 0.58 -0.075 (0.046) 0.10 Pharmacist -0.214 (0.195) 0.27 -0.079 (0.118) 0.50 Other health professions 0.079 (0.129) 0.54 0.073 (0.078) 0.35 Unit assistant/Clerk/Secretary/Technician 0.016 (0.053) 0.76 0.054 (0.031) 0.08 Administration -0.012 (0.108) 0.91 -0.035 (0.062) 0.57 Quality and Safety -0.346 (0.153) ***0.02*** 0.045 (0.094) 0.63 Other -0.001 (0.061) 0.98 0.041 (0.034) 0.22 Nurse 0 0 Accredited Yes 0.010 (0.064) 0.87 0.113 (0.043) ***0.01*** No 0 0 Hospital Size Small 0.069 (0.082) 0.41 -0.035 (0.058) 0.54 Medium 0.026 (0.086) 0.76 0.0001 (0.061) 1.00 Large 0 0 N 4046 3971 --------------- -------------------------------------------------------------- ---------------------------------- ---------------------------------- --------------------------- --------------- ::: Respondents working in the quality and safety departments had lower frequency of events reported (beta = -0.364, P-Value = 0.020) as compared to nurses (See Table [6](#T6){ref-type="table"}). ### Overall Perception of Safety Perception of patient safety improved by 0.094 (P-Value \< 0.001) for a one unit increase in the score on supervisor/manager expectations and actions promoting safety, by 0.129 (P-Value \< 0.001) a one unit increase in the score on organizational learning and continuous improvement, and by 0.111 (P-Value \< 0.001) for a one unit increase in the score on teamwork within hospital units. Similarly, perception of patient safety also improved by 0.031 for a one unit increase in the score on non-punitive response to error (P-Value = 0.01), 0.120 for staffing (p-Value \< 0.001), 0.099 for hospital management support for patient safety (p-Value \< 0.001) and 0.163 for hospital handoffs and transitions (p-Value \< 0.001). However, perception of patient safety decreased by -0.045 (P-Value \< 0.001) for a one unit increase in the score on teamwork across hospital units (See Table [6](#T6){ref-type="table"}). Respondents working as unit assistant/clerk/secretary/technicians had 0.054 (P-Value = 0.08) better perception of patient safety as compared to nurses (See Table [6](#T6){ref-type="table"}). Accredited hospitals were found to have better perception of patient safety as compared to non-accredited hospitals (beta = 0.113, p-value = 0.01) (See Table [6](#T6){ref-type="table"}). Discussion ========== The HSOPSC is one of the most common tools being used to assess the culture of safety in hospitals. Studies that utilize this tool usually report the 12 composite scores and the scores on the patient safety grade and the number of events reported. However, exploring the association between the patient safety composite scores and the hospital and respondent characteristics with the patient safety culture outcomes are not common in the literature. To our knowledge, this is one of the few studies that explore such an association. The analysis of results identified that patient safety culture predictors such as event reporting, proper communication, patient safety leadership and management, hospital size, and accreditation status are associated with the patient safety culture outcomes. Event Reporting --------------- A safety culture includes three major components a just culture, a reporting culture, and a learning culture \[[@B11]\]. Event reporting, an essential component for achieving a learning culture, can only happen in a non-punitive environment where events can be reported without people being blamed \[[@B12]\]. The non-punitive response to error composite received one of the lowest scores revealing that Lebanese hospital employees are not at ease when it comes to reporting errors. The majority of physicians and nurses in our sample reported no events over the past 12 months. Training opportunities that empower physicians to improve patient safety are limited thus investing in the training of physicians is important because they can play a major role in improving patient safety initiatives primarily by improving patient safety outcomes and reducing hassles and wasted time \[[@B13]\]. According to Leape et al. (1995), errors will be more frequent if nurses did not intercept 86% of all potential errors \[[@B14]\]. However, the majority of nurses in our sample reported no events over the past 12 months. Many errors in health care go unreported for many reasons including fear, humiliation, the presence of a punitive response to error, and the fact that reporting will not usually result in actual change \[[@B15]\]. Encouraging health professionals, specifically nurses, to report events in a non-punitive environment is crucial for improving patient safety. In our study, the fewest number of respondents to report more than 5 events over the past 12 months worked in surgical units. Errors in operating rooms are not uncommon and can sometimes be catastrophic \[[@B16]\]. Creating a patient safety culture in surgical units by improving communication and reporting more events is a high priority for operating room staff and hospitals \[[@B16]\]. Employees who do not deal directly with the patient are more at ease when it comes to reporting errors. As mentioned in Jones et al. (2008), work in laboratory units is considered as more organized than other units since it is controlled by more professional standards and because errors investigated in these units are done as a group \[[@B17]\]. On the contrary, when an error is performed by a nurse, the nurse is investigated as an individual rather a member of a medical team \[[@B17]\]. Work experience at the hospital also had some impact on the frequency and number of events reported. Frequency of events reported was found to increase with increasing years of experience. On the other hand, the score for perception of patient safety decreased as experience in the hospital increased. The perception of safety is defined as the extent to which procedures and systems are good at preventing errors and the lack of patient safety problems. As people become more experienced, they become more aware of the safety practices undertaken in the institutions they work in. When the perception of safety decreases with the increase in the years of experience, it means that the staff members do not agree that the patient safety practices, systems, and procedures in the hospitals act as barriers to errors and problems. A study by Bodur and Feliz (2009) showed that patient safety culture scores decreased as seniority increased \[[@B18]\]. The observation may be the result of an increase in medical errors done by the respondent due to frustration with hospital regulations or increasing staff awareness of safety problems and thus additional reporting. Communication ------------- Proper communication within and across healthcare teams is essential to remove any threats to safety of patients. Communication problems have been identified as major contributing factors to adverse events \[[@B7]\]. An analysis of 2,455 sentinel events reported to the Joint Commission on the Accreditation of Healthcare Organizations showed that 70% of the cases were a result of failure in communication \[[@B19]\]. In the absence of proper communication between the different hospital units, patient safety might be jeopardized. Our results revealed that higher scores on teamwork across hospital units increase the frequency of events reported. Moreover, higher scores on hospital handoffs and transitions increased the likelihood of having a better perception of safety among respondents and the likelihood of respondents to report a higher patient safety grade. Patient safety leadership and management ---------------------------------------- In order for a patient safety program to be successful, strong leadership is needed. Senior leaders are the only ones who are able to create the culture and commitment needed to solve underlying system causes of medical errors and harm to patients. When leadership and management is committed to a culture of safety, the whole organization will follow and thus disclosing adverse events and finding their root causes will become an organizational process. The focus should be in emergency rooms, operating rooms, and intensive care units \[[@B20]\]. Our results showed that more support from hospital management for patient safety increased the frequency of events reported. It also increased the likelihood of having a better overall perception of safety among respondents and the likelihood of respondents to report a higher patient safety grade. Staffing -------- Staffing, a major predictor of patient safety, was one of the composites that received a very low score. Having a strong, capable, and motivated workforce is one of the biggest challenges for hospitals today \[[@B21]\]. Evidence has shown that a strong link exists between the availability of health care providers and population health outcomes \[[@B21]\]. According to Sandars & Cook (2007), major catastrophes have occurred in organizations with insufficient staffing. Medical personnel in under-staffed hospitals are overworked \[[@B7]\]. They also suffer from stress and sleeplessness which might lead to lapses in performance thus leading to errors affecting quality and performance \[[@B22]\]. Our results showed that a more positive score on staffing increased the likelihood of having a more positive perception of safety among respondents and the likelihood of respondents to report a higher patient safety grade. Hospital size and accreditation ------------------------------- Hospital size and accreditation status were also factors affecting the culture of safety in hospitals. Small hospitals scored higher than larger hospitals on both the frequency of events reported and on the overall perception of safety. Large hospitals usually face challenges when it comes to implementing quality work especially because of bureaucracy. On the contrary, small hospitals have a more homogenous culture where staff members are more likely to share the same values \[[@B23]\]. Accredited hospitals also scored higher on both frequency of events reported and on the overall perception of safety in comparison to non-accredited hospitals. Respondents working at accredited hospitals had an increased likelihood of having a more positive perception of safety. Since the development of hospital accreditation programs that made accreditation a requirement for financial reimbursement by the MOPH, quality improvement initiatives have gained increased attention in Lebanese hospitals \[[@B24]\]. A study by El-Jardali et al. (2008) showed that Lebanese nurses perceived that accreditation created an improvement in the quality of health care \[[@B25]\]. Limitations ----------- A methodological limitation pertaining to this study should be acknowledged. In the literature, it is argued that social and behavioral research, particularly those that include self-reports such as surveys, is prone to Common method variance (CMV). CMV is believed to pose a threat to the validity of the data as it may either inflate or deflate the correlations among research variables. Although this issue can be attended to during survey development, evidence in the literature stipulate that although such research methods are more prone to CMV, they should not be considered weak or inappropriate if the researchers follow rigorous research conventions in research design, data collection and analysis \[[@B26]\]. In our study, CMV was avoided by the following \[[@B26]\]: \- Assuring that the composite scores were derived based on a combination of items; \- Counterbalancing the order of the questions; \- Ensuring the confidentiality and anonymity of respondents; \- Using clearly written scale items to make responses less subject to bias; and \- Informing respondents in the consent form that there is no preferred or correct answer, and that their individual responses would not be viewed by their management and survey completion would not affect their status at their hospital so they would be encouraged to honestly assess and respond freely Conclusion ========== Investing in patient safety practices in Lebanese hospitals is crucial. While patient safety is everyone\'s concern, it is not easy for everyone who works in health care organizations to understand this concept. In Lebanon, most health workers have had different training, and often hold a value system that is specific to their professional group. To be truly effective, patient safety needs to be incorporated into the education of health professionals across the spectrum of health care. Our results demonstrate that patient safety should be a top strategic priority for the health care organizations and its leaders. There should be blame-free system for identifying threats to patient safety, sharing information and learning from events. In addition, there should be a collaborative environment so that all health workers in the healthcare organization can share and exchange information about patient safety To facilitate change in cultural behaviors, hospital management should assess and redesign their current patient safety system including governance and reporting structures. In addition, they should provide their health professionals with comprehensive training on patient safety concepts, tools, and implementations. Our results also show that the move to prioritize patient safety by healthcare systems through accreditation is important. However, there is still a good deal skepticism on the part of health professionals who feel that by sharing highly sensitive information they may still be subject to blame and penalties. Progress in this respect will need far more enlightened policies, where health professionals are actively encouraged to report errors for the purpose of learning and improvement. In Lebanon, it is important to strengthen the new accreditation chapter on patient safety by supporting hospitals in training their staff, especially the less experienced ones, on patient safety competencies and about effective implementation of the new standards. Further research is needed to study the association between patient safety culture and clinical outcomes. It is hoped that the outcome of this study will help conduct a policy dialogue meeting for policy makers and stakeholders in Lebanon to discuss the findings and make deliberation about potential next steps. Senior policy makers, managers and leaders are the only ones who are able to create the culture and commitment needed to identify and resolve underlying systemic causes related to patient safety. List of abbreviations used ========================== IOM: Institute of Medicine; HSOPSC: Hospital Survey on Patient Safety Culture; MOPH: Ministry of Public Health; Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= FE was the principal investigator on this study and contributed to the conception, design, as well as analysis and interpretation of results. HD was involved in conceptualizing and implementing the data analysis plan as well as overseeing drafting of research results. DJ made substantial contributions to project management, including data collection and analysis as well as in drafting the manuscript. MJ was in charge of project management including data collection, she also contributed to data analysis and write-up of the final version of the paper. NH contributed to data entry and analysis and also made major contributions to the write up of the paper. All authors have read and approved the final version of the manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1472-6963/11/45/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Description of Patient Safety Culture Composites and Cronbach\'s Alpha**. File contains a definition of each of the Patient Safety Culture Composites in addition to Cronach\'s Alpha for each of the composites. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Authors would like to thank the Eastern Mediterranean Regional Office of the World Health Organization for funding this project. Thanks are also due to the Syndicate of Hospitals and the hospitals that participated in this national study. Authors would also like to thank Adele Semaan, Rana Hamdan and Hiba AbuSweid for data entry.

PubMed Central

2024-06-05T04:04:19.458668

2011-2-24

PMC3053222

Background ========== After it infects host *E. coli*cells, bacteriophage λ follows either of two fates, lytic or lysogenic. How the virus decides which pathway to follow after infection depends upon a complex genetic circuit. An increase in the number of infecting phages converts the decision making process from a deterministic to a stochastic one, with the cell fate depending on the number of phages deciding in favour of lysogeny \[[@B1],[@B2]\]. There are phage coded proteins and transcription factors \[[@B3]-[@B5]\] dedicated for this decision making process, but host factors are also involved \[[@B6]-[@B9]\]. Mutations in the *cI, cII*and *cIII*genes of λ \[[@B10]\] enhances the lytic frequency (leading to clear plaque formation, hence the names) and therefore the products of these genes were thought to be responsible for the establishment of lysogeny. CII, the key tetrameric transcription factor for lysogenic establishment, is a very unstable protein \[[@B7],[@B11],[@B12]\] and its presence in sufficient amounts is crucial for the lysogenic choice \[[@B13]-[@B15]\]. Other factors such as λCIII and the host hfl proteins that influence the lysis-lysogeny switching affect the stability of CII in one way or the other. λCIII promotes lysogeny by acting as a general inhibitor of *E. coli*HflB that degrades CII \[[@B16]\]. Mutations in the host *hfl*loci cause an infecting λ particle to follow the lysogenic mode. These genes therefore encode factors that somehow destabilize CII. Primarily from mutational studies, two such loci, *hflA*and *hflB*, were initially identified. The product of the latter gene, HflB, is an ATP-dependent metalloprotease known as a \'quality control\' protease that removes misfolded proteins produced due to rapid translation during good nutrient conditions \[[@B17],[@B18]\]. CII is also a substrate of HflB \[[@B7]\] and thus acts as a sensor for cellular nutrient conditions of the host. Rapid degradation of CII in cells growing in rich media thus favors the lytic development \[[@B13],[@B14]\]. The *hflA*locus consists of the genes *hflX*, *hflK*and *hflC*that are under the control of the same promoter \[[@B19]-[@B22]\]. Of these, *hflX*has been demonstrated to have no role in lambda lysogeny \[[@B23]\]. The products of the other two, HflK and HflC, are tightly associated with each other and copurify as the \'HflKC\' complex, which was earlier thought to be a protease \[[@B24]\]. Subsequently, HflKC was found only to act as a \'modulator\' of HflB by forming a complex with the latter \[[@B25]-[@B27]\]. The only other known *E. coli*factor in this process, HflD \[[@B9]\], has been shown to inhibit CII-mediated activation of transcription by impairing the DNA-binding ability of CII \[[@B28]\]. HflKC antagonizes the action of HflB towards the membrane associated substrates of the latter \[[@B18],[@B25]\]. The behavior of HflKC with respect to the cytosolic substrates of HflB (such as λCII), however, remains unclear. Likewise, the role of HflKC in the lysis-lysogeny decision of λ is not well understood. Though an \'hfl\' protein, mutations in whose gene(s) causes an increase in the lysogenic frequency of λ \[[@B6]\], the deletion of these genes has little effect on the *in vivo*stability of exogenous CII \[[@B26]\]. CII expressed from a plasmid is found to be stabilized in an *hflKC-*deleted cell, only if the host is simultaneously infected with a lambda phage \[[@B26]\]. On the other hand, *E. coli*cells overexpressing HflKC exhibit an enhanced frequency of lysogenization \[[@B26]\]. These results lead to a paradox: lysogeny increases both upon deletion or overexpression of HflKC. Therefore, the role of HflKC in the λ lysis-lysogeny switch merits further investigations. Methods ======= Plasmids, bacterial strains and phages -------------------------------------- Plasmid pQKC was constructed by PCR cloning of the *hflK*and *hflC*ORFs (not fused, because the genomic region between these two contains the stop codon for *hflK*and the RBS for *hflC*) between the *Bam*HI and *Sal*I sites of pQE30 (purchased from Qiagen, contains the phage T~5~promoter under the control of a Lac operator). Construction of pKP219 (which contains the *cII*gene under the lac promoter LacP and a P15A replication origin) has been described earlier \[[@B28]\]. Plasmid pC2C3 (containing the *cII*and *cIII*genes) was constructed in three steps. First, the *Nde*I-*Bam*HI fragment of pAB905 containing the *cIII*gene \[[@B29]\] was cloned into pKP07 \[[@B28]\] and was named pLaCIII (containing the *cIII*gene under LacP). Then the *Bgl*II-*Xho*I fragment of pLaCIII (i.e. the cIII gene along with the LacP) was cloned into the compatible *BamH*I-*Xho*I sites of pKP106 (which already contained the cII gene under LacP) \[[@B28]\]. The resulting plasmid was named pLaC2C3. In the final step the *Bam*HI-*Bgl*II fragment of pLaC2C3 (containing both *cII*and *cIII*under individual LacP promoters) was cloned into the linearized arm of pK109 (having a P15A origin of replication) \[[@B30]\] at the *Bgl*II site. For wild type *E. coli*, the strain MG1655 (*F*^-^*λ*^-^*ilvG rfb-*50 *rph-1*) was used. The strain AK990 \[[@B26]\] (*ΔhflKC:: Kan*) served as cells with mutant *hflKC*. The phage strain λ*cIII*^67^was used as the CIII-defective phage. In this strain, a G to T mutation in the 23rd nucleotide of the *cIII*ORF leads to an alternative structure of the cIII mRNA that is incapable of translation \[[@B31]\]. This is one of the most effective cIII mutants \[[@B32]\] and has been used as cIII^-^by many workers. Purification of proteins ------------------------ For the purification of the HflKC complex, XL1Blue cells carrying pQKC was used and 100 μg/ml of ampicillin was used for selection. 7.5 ml of the overnight saturated culture was inoculated into 750 ml of fresh M9 medium with the appropriate antibiotic and allowed to grow on a 37°C shaker incubator till the culture O.D. (at 600 nm) was 0.4-0.5. The culture was then cooled to 18°C and induced by 500 μM IPTG, followed by further growth at 18°C with constant shaking (at 100 rpm) for 20 hours. After induction, bacterial cells were recovered by centrifugation at 3000 g for 10 minutes in Sorvall RC5C, using an SA600 rotor, at 4°C. The medium was decanted out and the pellet was washed with 0.9% NaCl and dissolved in 20 ml of lysis buffer (20 mM TRIS-HCl, pH 8.0, 100 mM KCl, 10% glycerol, 5 mM imidazole, 0.5% NP40, bacterial protease inhibitor cocktail (MBI Fermentas) and 200 μg/ml lysozyme). Cells were then lysed by sonication with 5 pulses (at a pulse rate of 10 mV/30 seconds), followed by centrifugation at 26000 g for 30 minutes at 4°C. The supernatant was collected into a fresh tube and loaded on to a Ni^+2^-NTA column, pre-equilibrated with the lysis buffer. After loading, the column was washed with wash buffer (20 mM TRIS-HCl, pH 8.0, 600 mM KCl, 10% glycerol, 15 mM imidazole). Proteins were eluted from the column using the elution buffer (20 mM TRIS-HCl, pH 8.0, 100 mM KCl, 10% glycerol, 0.1% NP40, 300 mM imidazole). Imidazole was removed by dialysis in 20 mM TRIS-HCl, pH 8.0, 100 mM KCl, 10% glycerol, 0.1% NP40). Native CII \[[@B33]\] and GST-HflB \[[@B29]\] were purified as described earlier. *In vitro*proteolysis of CII ---------------------------- HflB mediated proteolysis of CII was carried out in buffer P (50 mM Tris-acetate, 100 mM NaCl, 5 mM MgCl~2~, 25 μM Zn-acetate, 1.4 mM β-ME; pH 7.2). ATP was added to a concentration of 5 mM in all the reaction mixtures. 8 μM of CII was taken with 1 μM of purified GST-HflB in a 30 μl reaction mix. The reactions were incubated at 37°C for the specified time intervals followed by the addition of SDS-PAGE loading buffer and heating in a boiling water bath for 5 minutes. The samples were analyzed on a 15% SDS-PAGE. The effect of HflKC on the proteolysis of CII was observed by the addition of His-HflKC (up to 2 μM) to GST-HflB prior to the addition of CII. The band corresponding to CII was quantitated by volume analysis (software used: Versadoc (Bio rad) Quantity-1) and used as the amount of CII remaining (expressed as the percentage of the amount of CII at zero time) after the specified time. *In vivo*proteolysis of CII --------------------------- *In vivo*proteolysis of CII was carried out in *E. coli*MG1655 cells (having wild type HflB) transformed with pKP219 or pC2C3, both of which contained *cII*under Lac promoter. In addition, pC2C3 contained *cIII*under a second Lac promoter. Cells carrying pKP219 or pC2C3 were inoculated in 10 ml of LB medium supplemented with 50 μg/ml kanamycin. Expression of CII was induced by 1 mM IPTG after the O.D. of the culture (at 600 nm) had reached 0.6. The culture was further grown at 37°C for another 30 minutes, followed by the addition of 10 μg/ml spectinomycin to arrest further protein synthesis. Samples were taken out at regular intervals after spectinomycin addition, and immediately centrifuged to pellet the cells. 30 μl of sterile water and 8 μl of SDS gel loading dye were added to each sample, followed by immediate boiling and loading onto a 15% SDS-PAGE. The gel was transferred to a PVDF membrane (Pierce Biotech) and was blotted with anti-CII antibody. Each CII band was quantitated by volumetric analysis as described above. The effect of overexpression of *hflKC*was observed by transformation of MG1655 cells by plasmid pQKC (plus pKP219 or pC2C3). The transformed cells were grown in the presence of both kanamycin and ampicillin. Promoters in both the plasmids are inducible with IPTG. The effect of deletion of *hflKC*was observed by transformation of AK9990 cells by pKP219 or pC2C3. For measurement of the stability of CII under conditions of infection by λ*cIII*^67^, MG1655 or AK990 cells carrying pKP219 were grown in Luria broth supplemented with 0.4% maltose and were infected with the phage (at an MOI of 10 to ensure that all the cells were infected), 20 minutes after the addition of IPTG. Spectinomycin was added after another 25 minutes to ensure the entry of phage DNA and the expression of phage factors. Samples were then taken out at regular intervals and analyzed as described above. Assay of plaque morphology -------------------------- The plaque morphology of λ*cIII*^67^was assayed in *E. coli*MG1655 (wild type), in MG1655 cells carrying pQKC, and in strain AK990 (*ΔhflKC::Kan*). Cells were grown up to an O.D. (at 600 nm) of 0.6 in Luria broth supplemented with 0.4% maltose, and were induced with 500 μM IPTG. A bacterial lawn was made by pouring 5 ml of soft top agar (0.5% Luria agar supplemented with 0.4% maltose) mixed with 300 μl of these cells onto a 2% Luria agar plate. Another 100 μl of the above liquid culture was infected with λ*cIII*^67^at an M.O.I. of 0.1. It was further incubated at 32°C for 10 minutes to allow adsorption of the phage. Appropriate dilutions were then plated onto the prepared bacterial lawn and the plates were incubated overnight at 32°C. The turbidity of plaques formed in AK990 cells or in cells overexpressing HflKC were compared with the clear plaques formed in wild type cells upon infection by λ*cIII*^67^. Results and Discussion ====================== Role of HflKC on the proteolysis of CII *in vivo* ------------------------------------------------- *E. coli*HflKC inhibits the proteolysis of all the membranous substrates of HflB (e.g., SecY, YccA) \[[@B18]\]. However, the behaviour of HflKC toward λCII, a cytosolic substrate, is perplexing. The deletion of *hflKC*as well as its overexpression causes an increase in the lysogenic frequency of λ \[[@B26]\]. The *hflKC*genes were first identified as mutants that caused turbid plaques of λ on a bacterial lawn \[[@B6]\]. It is therefore expected that CII would be stabilized in an *hflKC-*deleted host cell. Kihara *et al.*\[[@B26]\], however, showed that the deletion of *hflKC*has little effect on the stability of CII cloned under an AraBAD promoter. We obtained similar results when the effect of *hflKC*deletion (strain AK990) on the stability of CII (cloned under *lac*promoter) was tested (Figure [1](#F1){ref-type="fig"}). Here we measured the stability of CII expressed from the plasmid pKP219 in wild type and in AK990 (Δ*hflKC*) cells. In both cases, CII was unstable. We also tested the effect of overexpression of HflKC from a second plasmid (pQKC), and found that in this case, CII expressed from pKP219 was stabilized (Figure [1](#F1){ref-type="fig"}). This data is consistent with *in vitro*results that showed that purified HflKC \[[@B26],[@B34]\] inhibits the proteolysis of CII. The inhibitory activity is an intrinsic property of HflK and HflC, since HflK or HflC can individually inhibit the proteolysis of CII \[[@B34]\]. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Role of HflKC on *in vivo*proteolysis of CII**. Left panel shows the proteolytic pattern of exogenous CII (expressed from pKP219) in wild type cells (open circles), AK990 (Δ*hflKC*, squares) or wild type cells carrying plasmid pQKC (triangles). In each case, the expression of CII was induced with IPTG and translation was stopped 30 minutes later with spectinomycin. Relative amount of CII was measured after regular intervals (0, 5, 10, 15, 20 minutes) by western blotting followed by quantification using densitometric analysis. Corresponding western blots showing the stability of CII in different host strains are shown in the right panel. :::  ::: These results pose an intriguing question. Why does the deletion of an inhibitor of CII proteolysis promote lysogeny? One can think of the following possibilities: (i) A proper assembly of HflB that is necessary for its activity against cytosolic substrates, may require HflKC; or (ii) In the absence of HflKC, HflB is guided towards its membrane-associated substrates \[[@B26]\], indirectly stabilizing the cytosolic substrate CII. However, from *in vivo*proteolysis experiments we found that in AK990 cells (Δ*hflKC*), exogenous CII was not stabilized (Figure [1](#F1){ref-type="fig"}), confirming that HflB was active against CII even in the absence of *hflKC*. This result rules out both the possibilities mentioned above. It may be noted that similar results were also obtained by Kihara *et al*\[[@B26]\]. Therefore, an increase in lambda lysogeny upon overexpression of host HflKC \[[@B26]\] is not at all surprising, since HflKC inhibits the proteolysis of CII. Effect of increasing concentrations of HflKC on the proteolysis of CII *in vitro* --------------------------------------------------------------------------------- The paradoxical effect of an increase in the lysogenic frequency of λ upon deletion as well as overexpression of *hflKC*has been reported \[[@B26]\]. A possible reason behind this paradox could be that a critical molar ratio between HflB and HflKC, believed to be 1:1 in wild type cells \[[@B35]\], is necessary for a proper proteolysis of CII by HflB. Both the increase or decrease of HflKC would offset this critical ratio and could lead to a stabilization of CII, promoting lysogeny. To examine this possibility, we carried out the proteolysis of CII by HflB *in vitro*, in the presence of three different concentrations of HflKC (Figure [2](#F2){ref-type="fig"}). In the first case, when HflKC was absent (mimicking the deletion of HflKC), CII (8 μM) was rapidly cleaved by HflB. The rate of proteolysis was much slower when HflKC was added in a molar ratio of HflKC:HflB = 1:1. The proteolysis was inhibited further when HflKC was added in excess (HflKC:HflB = 2:1). If the above hypothesis was true, proteolysis of CII should have been maximum at a molar ratio of 1:1. Therefore we conclude that HflKC acts as a simple inhibitor of CII proteolysis and the stabilization of CII in the absence of HflKC may involve other factors. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Effect of varying concentrations of HflKC on *in vitro*proteolysis of CII**. CII (8 μM) was treated with GST-HflB (1 μM), in the presence of His-HflKC in various concentrations: 0 (open circles), 1 μM (squares) and 2 μM (triangles). Samples were taken out at various time points, run on a 15% SDS-PAGE, and the CII bands were quantitated by densitometry. :::  ::: Role of HflKC on *in vivo*proteolysis of CII: the effect of CIII ---------------------------------------------------------------- Deletion of *hflKC*genes lead to an increase in the lysogenic frequency of λ. Nevertheless, the stability of cloned CII remained unaffected in Δ*hflKC*cells. An interesting phenomenon, however, was observed in Δ*hflKC*cells that were infected by λ. CII expressed from a plasmid was stabilized in these cells \[[@B26]\]. Thus it appears that some additional factors, supplied by the infecting phage, caused a stabilization of CII in the absence of HflKC. The only known phage factor that favors lysogeny by inhibiting the proteolysis of CII by HflB, is CIII \[[@B29],[@B36]\]. We therefore tested the possible involvement of CIII as the λ factor responsible for the above result, *viz.*stabilization of CII in λ-infected Δ*hflKC*cells. We sought to supply λCIII instead of the whole phage in an *hflKC-*deleted host and investigate its effect on the proteolysis of cloned CII. For this purpose, we cloned *cIII*in tandem behind *cII*in the same plasmid and inserted it in a host with deleted (AK990) or overexpressed *hflKC*. CII was indeed stabilized in these cells, even without simultaneous infection by λ (Figure [3](#F3){ref-type="fig"}). Therefore it appears that infection by λ stabilized CII in Δ*hflKC*cells because it supplied CIII. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Role of HflKC on *in vivo*proteolysis of CII in the presence of CIII**. Proteolytic pattern of exogenous CII (expressed from pC2C3) in wild type cells (open circles), AK990 (Δ*hflKC*, squares) or wild type cells carrying plasmid pQKC (triangles). Experimental conditions were similar to those used in Figure 1. :::  ::: CIII is a general inhibitor of CII proteolysis \[[@B29],[@B36],[@B37]\]. It is therefore expected that between a wild type strain alone and one that carries CIII, CII would exhibit a greater stability in the latter. A comparison of figures [1](#F1){ref-type="fig"} and [3](#F3){ref-type="fig"} (open circles) shows that this is indeed the case. Nonetheless, a greater stability of CII in Δ*hflKC*cells compared to wild type (both carrying the CIII-expressing plasmid) is surprising, since the absence of *hflKC*does not affect the stability of CII. CIII is itself a substrate of HflB \[[@B38]\]. If HflKC facilitated the proteolysis of CIII, the above effect could be explained by the preferential stabilization of CIII in Δ*hflKC*cells. However, there was no difference in the *in vitro*proteolysis of CIII by HflB in the presence or absence of purified HflKC (data not shown). Therefore the role of CIII in this paradoxical effect is indirect. Are there additional λ factors that influence the lysis-lysogeny decision? -------------------------------------------------------------------------- If CIII was the only factor responsible for the stabilization of CII in Δ*hflKC*cells, infection with a *cIII*-defective phage would produce clear plaques in a Δ*hflKC*host. We tested this possibility by infecting both AK990 (Δ*hflKC*) cells and *hflKC*-overexpressing cells with lambda *cIII*^*67*^\[[@B31],[@B39]\]. Interestingly, turbid plaques were obtained in each case, unlike the clear plaques produced in wild type *E. coli*(Table [1](#T1){ref-type="table"}). This result is really surprising as *cIII*^-^phage always produces clear plaques. Since CIII and HflKC both inhibit the proteolysis of CII, it is also surprising that the absence of both leads to increased lysogeny. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Plaque morphology upon infection with λ*cIII*^67^ ::: Genotype of host *E. coli*cell Plaque morphology -------------------------------- ------------------- Wild Type Clear Wild Type + pQKC Turbid AK990 (Δ*hflKC::Kan*) Turbid ::: Is it then possible that enhancement of lysogeny can occur through a different mechanism that does not involve the stabilization of CII? Increase in lambda lysogeny is invariably linked to the stability of CII in all published reports to date. Can the two phenomena be delinked in some special case such as a Δ*hflKC*host? We tested this possibility by measuring the stability of cloned CII in wild type and Δ*hflKC*cells, both infected with λ*cIII*^*67*^. A greater stabilization of CII occurred in Δ*hflKC*cells (Figure [4](#F4){ref-type="fig"}). Therefore, an increase in the lysogenic frequency indeed requires the stabilization of CII. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Effect of infection by *cIII*-mutant lambda on *in vivo*proteolysis of CII**. The proteolysis of CII was visualized in wild type (open circles) or AK990 (diamonds) cells infected with λ*cIII*^*67*^. The expression of CII was induced with IPTG, and the cells were infected with the phage after 20 minutes. Protein synthesis was stopped 25 minutes later with spectinomycin. The relative amount of CII was measured at regular intervals by western blotting followed by quantification using densitometric analysis. :::  ::: This enhanced stabilization of CII is observed only under conditions of phage infection, even when CIII is nonfunctional. Therefore in addition to CIII, there could be another as yet unidentified factor in λ that increases the stability of CII and hence, promotes lysogeny (see Figure [5A](#F5){ref-type="fig"}). The presence of such a CII-stabilizing factor (CSF) can only be demonstrated in HflKC-deleted cells. Therefore, the activities of CSF and HflKC must have some connections (Figure [5B](#F5){ref-type="fig"}). Likewise, CIII and HflKC are likely to be connected as well. The different outcomes for deletion or overexpression of *hflKC*on lysogeny as well as on the stability of CII under various conditions are summarized in Figure [5A](#F5){ref-type="fig"}. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **The effect of deletion or overexpression of hflKC on λ lysogeny and on the stability of CII: A summary of results and possible mechanisms**. (A) A summary of results published previously as well as reported in this study is shown schematically. Some unanswered questions that remain are highlighted in the boxes. (B) Mechanisms for the stability of CII and the lysogenic outcome under various conditions are shown. HflB acts upon CII to digest CII, as indicated by the arrow. This digestion is inhibited by HflKC, by CIII or by the postulated CII-stabilizing factor CSF. The levels of inhibition are denoted by the lengths of the blunt lines. Possible crosstalk between HflKC and CIII or CSF are indicated by curved arrows. Dashed arrows denote lack of crosstalk. HflKC, CIII or CSF inhibits the digestion of CII. In wild type E. coli cells, this inhibition is unable to sufficiently stabilize CII, leading to normal plaques (left panel). When HflKC is overexpressed, CII is stabilized better by the action of HflKC, and turbid plaques are produced (middle panel), while in Δ hflKC cells, CIII and/or CSF act better to stabilize CII, giving rise to turbid plaques (right panel). :::  ::: The unknown factor CSF could have been a non-protein factor (i.e, DNA) and lambda DNA would have been a good candidate for the same, since CII may be stabilized by binding to its cognate promoter. However, in our *in vivo*experiments, the plasmid pKP219 (used for the expression of exogenous CII) contained the promoter sequence P~E~, ruling out such a possibility. Stabilization of CII in cells overexpressing *hflKC*is not surprising since HflKC is an inhibitor of CII-proteolysis. It is worthwhile to note that the effect of HflKC deletion is epistatic over the effect of *cIII*deletion, since even the absence of CIII cannot produce clear plaques in a Δ*hflKC*host. It is possible that CIII (and the hypothesized CIII-like factor CSF) works better in the absence of HflKC (Figure [5B](#F5){ref-type="fig"}). Therefore CII is better stabilized under these conditions and produces turbid plaques in *ΔhflKC*cells. *cI*, *cII*and *cIII*were first described as phage mutations which led to clear plaques in a wild type host. On the other hand, λ gives very turbid plaques in a Δ*hflKC*host. Our study thereby raises the possibility of finding novel phage mutations that would give clear plaques in an *hflKC*-deleted host. Conclusions =========== 1\. *E. coli*HflKC inhibits the proteolysis of λCII by HflB and hence the overexpression of the former results in an increase in the lysogenic frequency. 2\. In the absence of HflKC, λCII is stabilized upon infection by cIII-defective λ, suggesting the presence of a yet unidentified phage factor CSF (CII-stabilizing factor). Authors\' contributions ======================= KB and PP designed the experiments, KB performed the experiments and analysed the results of the HflKC-based *in vitro*and *in vivo*experiments. PKP designed and constructed the vector pKP219 and designed the method to determine the stability of CII *in vivo*. ABD helped in designing experiments and drawing inferences from the experimental results. PP designed research and supervised all the work. KB and PP wrote the manuscript and all authors approved the final version. Acknowledgements ================ We thank S. Adhya (NIH, Bethesda) for λ*cIII*^*67*^and for his comments on the manuscript, K. Ito and Y. Akiyama (Kyoto University) for *E. coli*AK990 strain, S. K. Dasgupta (Bose Institute) for the plasmid pSD5b and K. Shearwin (University of Adelaide) for anti-CII antibody. This work was funded by Institutional Project 5 (Microbial Genomics) of Bose Institute. KB was supported by CSIR, India (F. No. 9/15 (302)/2004-EMR-I).

PubMed Central

2024-06-05T04:04:19.467262

2011-2-17

PMC3053223

Background ========== Transition metals play an essential role in all organisms as they are used as structural or catalytic cofactor in a very large number of proteins \[[@B1]\]. Among these elements, zinc is the one which is found in the largest number of enzymes with known three-dimensional structure \[[@B2]\] and recent bioinformatics investigations have established that zinc-binding proteins constitute about 5% of bacterial proteomes \[[@B3]\]. Despite its abundant employment in proteins, the intracellular concentration of zinc must be accurately controlled to prevent its potential toxicity. To this aim bacteria have developed effective systems to regulate the balance between uptake and export of zinc and maintain an optimal intracellular level of this metal \[[@B4]-[@B6]\]. In *Escherichia coli*K12, for example, zinc efflux is achieved through the two transporters ZitB, a member of the cation diffusion facilitator family \[[@B7]\], and ZntA, a P-type ATPase \[[@B8]\]. ZntA synthesis is regulated by ZntR \[[@B9]\], a zinc-responsive Mer-like transcriptional regulator that activates *znt*A transcription by binding to zinc, thus favoring the efflux from the cell of the metal in excess. Zinc uptake is ensured by a few transporters characterized by different affinity for the metal. Under conditions of moderate zinc availability, metal uptake is carried out by the low affinity permease ZupT, a member of the ZIP family of transporters \[[@B10]\]. In contrast, when bacteria grow in environments characterized by very low zinc availability, zinc import is ensured by the high affinity zinc transporter ZnuABC \[[@B4],[@B11]\], whose synthesis is tightly controlled by the binding of this metal to the promoter of *zur*gene \[[@B12]\]. Studies carried out in different bacterial species have established that ZnuABC is strictly required to promote an efficient microbial growth in media deficient in zinc and to ensure bacterial virulence, indicating that zinc availability in the infected host is very limited and that several bacteria strictly rely on this specific transporter to compete with their host for zinc binding \[[@B13]-[@B20]\]. It has been recently shown that in some bacterial species the fine-tuning of zinc uptake involves another protein, ZinT (formerly known as YodA), which was initially identified in *E. coli*as a cadmium stress stimulated protein \[[@B21]-[@B23]\]. Subsequent investigations have demonstrated that ZinT is involved in periplasmic zinc binding under zinc-limiting conditions \[[@B24],[@B25]\] and it has been hypothesized that it could play a zinc-chaperone role by delivering metal ions to apo-proteins in need of their cofactor \[[@B12]\]. More recently, studies carried out in *Salmonella enterica*serovar Typhimurium have suggested that ZinT participates to the zinc uptake process mediated by ZnuABC, through a mechanism involving its direct interaction with ZnuA \[[@B18]\]. Such a role, however, appears to be dispensable, as many bacteria expressing ZnuABC do not possess ZinT \[[@B18]\]. To strengthen our knowledge on the relevance of zinc import in the host-pathogen interaction, we analyzed the role of ZnuABC and ZinT in the enterohemorrhagic *E. coli*O157:H7 strain. This pathogen is able to colonize the large intestine mucosa of humans, where it causes characteristic attaching and effacing lesions on intestinal epithelial cells which are responsible for the major symptoms of hemorrhagic colitis and Haemolytic Uremic Syndrome (HUS) \[[@B26]\]. Our results highlight the central importance of this zinc uptake pathway in *E. coli*O157:H7 and confirm the participation of ZinT to the mechanisms of metal import mediated by the high affinity zinc transporter ZnuABC. Methods ======= Reagents -------- Antibiotics, bovine serum albumin and D-MEM, were purchased from Sigma-Aldrich. Restriction endonuclases, DNA-modifying enzymes and DNA polymerase High-Fidelity Expand were obtained from Roche, while Euro*Taq*and *Pfu*DNA polymerases were obtained from EuroClone and Promega, respectively. All other chemicals were purchased from BDH and were of the highest available grade. The oligonucleotides were synthesized by Primm (Milan, Italy). Strains and growth conditions ----------------------------- All strains used in this work are listed in Table [1](#T1){ref-type="table"}. *E. coli*O157:H7 ED597 is a clinical human isolate associated to a HUS case \[[@B27]\]. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Bacterial strains ::: ----------------------------------------------------------------------------------------------------- Strains Relevant genotype or characteristic Reference or source --------------------------------- ---------------------------------------- -------------------------- [*E. coli*O157:H7]{.underline}\ ED597 Wild type D\'Orazio *et al*., 2008 RG112 Δ*zinT*::*kan* this study RG113 Δ*znuA*:: *kan* this study RG114 Δ*znuA*::*cat*Δ*zin*::*kan* this study RG115 Δ*etpC*::*cat* this study RG-F116 *zinT*::3xFLAG-*kan* this study RG-F117 *znuA*::3xFLAG-*kan* this study RG-F118 Δ*zur*::*cat zinT*::3xFLAG-*kan* this study RG-F119 Δ*zur*::*cat znuA*:: 3xFLAG-*kan* this study RG-F120 Δ*zinT*::*cat znuA*::3xFLAG- *kan* this study RG-F121 Δ*znuA*::*cat zinT*::3xFLAG- *kan* this study RG-F122 Δ*etpC*::*cat zinT*::3xFLAG- *kan* this study RG-F123 Δ*etpD*::*cat zinT*::3xFLAG- *kan* this study [*E. coli*]{.underline} BL21 Wild type laboratory collection DH 5α Wild type laboratory collection [*S. enterica*]{.underline}\ PP134 *zinT*::3xFLAG- *kan* Petrarca *et al*., 2010 SA140 *znuA*::3xFLAG- *kan ilv*I::Tn10dTac-\ Ammendola *et al*.,2007 *ca*t:: 3xFLAG- *kan* ----------------------------------------------------------------------------------------------------- ::: Bacteria were grown at 37°C in Luria-Bertani (LB) liquid medium (1% bacto tryptone w/v, 0.5% yeast extract w/v, 1% NaCl w/v) or in LB medium solidified with 1.5% (w/v) agar. For growth under metal limiting conditions a modified M9 minimal medium, hereafter named modM9 (43 mM Na~2~HPO~4~, 22 mM KH~2~PO~4~, 19 mM NH~4~Cl, 1 mM MgSO~4~, 0.1 mM CaCl~2~and 0.2% glucose) was used. To prepare the modM9, as well as other zinc-free solutions, we used ultra-pure water produced by a reverse osmosis system characterized by conductivity lower than 0.03 μS/cm. Moreover, bacterial culture and all solutions used with modM9 were prepared and incubated using zinc-free polypropylene plasticware (Falcon 50 and 10 ml tubes, Gilson tips and Eppendorf microtubes) avoiding glassware and other uncontrolled materials, except the 96-well plates used for the growth curves in modM9 which were in polystyrene. In this case, to remove metal contaminants of microtiter plates were treated overnight with 10 μM EDTA and then washed three times with fresh modM9 to eliminate EDTA traces. The effective ability of this procedure in removing zinc traces was evaluated by measuring the emission spectra of the final washing solution after the addition of 25 μM Zinquin, a highly specific Zn-fluorophore \[[@B17]\]. When required, the culture media were supplemented with the appropriate antibiotics (ampicillin 100 μg/ml, kanamycin 50 μg/ml, chloramphenicol 15 μg/ml). Mutant strains construction --------------------------- All *E. coli*O157:H7 knockout mutants and the 3xFLAG strains were obtained following the protocol described by Datsenko and Wanner \[[@B28]\] and the epitope tagging method described by Uzzau *et al*. \[[@B29]\], respectively. The plasmids and the oligonucleotides used for mutants\' construction are listed in Table [2](#T2){ref-type="table"} and [3](#T3){ref-type="table"}, respectively. Recombinant strains were selected on chloramphenicol or kanamycin LB plates and confirmed by PCR using oligonucleotides internal to the chloramphenicol or kanamycin resistance cassettes in combination with primers specific for each gene. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Plasmids ::: Plasmid Relevant genotype or characteristic Reference or source ------------------ ----------------------------------------------- --------------------------- pKD46 lambda red recombinase function Datsenko and Wanner, 2000 pKD3 chloramphenicol resistance cassette template Datsenko and Wanner, 2000 pKD4 kanamycin resistance cassette template Datsenko and Wanner, 2000 pSUB11 3xFLAG-kanamycin resistance cassette template Uzzau *et al*., 2001 p18ZnuAO157 ZnuA of E. *coli*O157:H7 cloned in pEMBL18 This work p18ZnuA*E. coli* ZnuA of E. *coli*K12 cloned in pEMBL18 This work ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Oligonucleotides used ::: Primer Sequence(5\'-3\') ----------------------- ------------------------------------------------------------------------------- ZinT-HP1 TTTAGGTGTCTTTATTGTTAGCGCTCCTGCCTTTTCGCAT[TGTAGGCTGGAGCTGCTTCG]{.underline} ZinT-HP2 CCACTTCTTCGCTACTCAACTGATATGGATAATACGTTGGCC[CATATGAATATCCTCCTTAG]{.underline} ZnuA-HP1 ATTATTCGCCGCTCTCTGGGGCGGTGCAACACAGGCCGC[TGTAGGCTGGAGCTGCTTCG]{.underline} ZnuA-HP2 CCTTTCAGGCAGCTCGCATACTGGTTGGCTAATTGACTCAGG[CATATGAATATCCTCCTTAG]{.underline} Zur-HP1 AAATCTGCGCGCAGCGTAATGTGCGCCTGACCCCACAGCT[GTAGGCTGGAGCTGCTTCG]{.underline} Zur-HP2 GCACAGAGTGATCATGCTGGCACTGTTCAGGATGACGACACGC[CATATGAATATCCTCCTTAG]{.underline} EtpC-HP1 ATAAAAGATATCGTACTTAAAATGCTGACGCCAAACCGGC[TGTAGGCTGGAGCTGCTTCG]{.underline} EtpC-HP2 CAGCAATAAATGCATCATATAACTGACCATCACGCTCGAC[CATATGAATATCCTCCTTAG]{.underline} EtpD-HP1 TTTATGGTTTTGCTGTGGTCGATATGCACAACGGTATACT[TGTAGGCTGGAGCTGCTTCG]{.underline} EtpD-HP2 GCATAAAACGCAGCAATCGCCGCTTTCACCTTCCGGAAAG[CATATGAATATCCTCCTTAG]{.underline} ZinT-3xFLAGF GTTGAGTAGCGAAGAAGTGGTCGAGGAAATGATGTCTCAT[GACTACAAAGACCATGACGG]{.underline} ZinT-3xFLAGR CTTTCTCTGTTGGCCGTATTGTGTATGGAATCCGTTATTGG[CATATGAATATCCTCCTTAG]{.underline} ZnuA-3xFLAGF TCAATTAGCCAACCAGTATGCGAGCTGCCTGAAAGGAGAT[GACTACAAAGACCATGACGG]{.underline} ZnuA-3xFLAGR TGACAATTGGCGTGGCATCGCGGTGATAAACATAGGGCCG[CATATGAATATCCTCCTTAG]{.underline} ZnuA~O157~-Pst-F AACTGCAGTGTCGACTTACCTGCG ZnuA~O157~-Xba-R GCTCTAGATTATTAAACGCCAGGGCGA ZnuA ~E\ *coli*~Kpn-F GGGGTACCGTTACATAAAAAAACGCTTC ZnuA ~E\ *coli*~Xba-R GCTCTAGATTAATCTCCTTTCAGGCAG Clo-Int^a^ CTGGATATACCACCGTTGAT Stop Clo-Int^a^ CACTCATCGCAGTACTGTT Kan-F^a^ TGAACAAGATGGATTGCACG Kan-R^a^ AAGAACTCGTCAAGAAGGC P1, P2 and 3xFLAG homolog sequences are underlined **a**primers used for screening of mutant clones or in Southern experiments ::: The double mutants (strains bearing double gene knockout or an epitope-flagged gene and a null mutation simultaneously) were constructed by a previously described procedure \[[@B27]\], electroporating the products of PCR reaction with primers specific for second mutation, in cells with the chromosome bearing the previous mutation. The resulting strains RG114, RG-F118, RG-F119, RG-F120, RG-F121, RG-F122, and RG-F123 were selected as mentioned above. To further verify the modification of the targeted genes, all mutant strains were checked also by Southern-blot procedure (data not shown). Plasmids used for complementation assays were obtained by cloning the *znu*A gene from *E. coli*O157:H7 and *E. coli*K12 in pEMBL 18. The *znu*A sequences, including their promoter regions, were amplified by PCR using specific oligonucleotides (Table [3](#T3){ref-type="table"}) and inserted in the XbaI and PstI (*E. coli*O157:H7) or XbaI and KpnI (*E. coli*K12) restriction sites of pEMBL 18. The resulting plasmids were called p18ZnuAO157 and p18ZnuA*E. coli*(Table [2](#T2){ref-type="table"}). Growth curves ------------- Each bacterial strain was grown overnight in LB broth at 37°C and then diluted 1:1000 in fresh LB, supplemented or not with 0.5 mM or 2 mM EDTA and 0.2, 0.5 and 1 mM ZnSO~4~. Aliquots of 300 μl of these dilutions were inoculated in 96-well plate (Becton-Dickinson) and incubated at 37°C with shaking. Growths in modM9 of each strain, including the RG113 bearing plasmid p18ZnuA O157 or p18ZnuA*E. coli*, were carried out by diluting preinocula 1:500 in fresh medium supplemented or not with 0.25, 0.5, 1 or 5 μM ZnSO~4~. Bacterial growth was monitored at 595 nm every hour for 15 hours using a microtiter-plate reader (Biotek instrument mod. ELX808). Assays were performed in triplicate and each strain was tested in three independent experiments. Complementation assay --------------------- Wild type, *znu*A deleted strain (RG113) and RG113 harbouring plasmids p18ZnuAO157 or p18ZnuAE. *coli*were grown overnight at 37°C in LB broth supplemented with the appropriate antibiotics, diluted to 1 OD~600~and then streaked on LB plates containing 0, 0.5, 1 and 2 mM EDTA with or without antibiotics. Bacterial ability to form visible colonies on these plates was analyzed after 24 hours of incubation at 37°C. Western blot analysis --------------------- The expression of *zin*T and *znu*A was indirectly analyzed by measuring the intracellular accumulation of the epitope-tagged proteins. Strains carrying the epitope-tagged genes were grown at 37°C in LB or in modM9 in presence or absence of EDTA or transition metals. Bacteria cultivated in LB were exposed to 0.5 mM EDTA and 0.2 mM ZnSO~4~, or 0.25 mM CdSO~4~, whereas bacteria in modM9 were grown in presence or not of 5 μM EDTA and of 5 μM ZnSO~4~, FeSO~4~, CuSO~4~or MnCl~2~. After 4 h of growth in LB and 6 h or 16 h in modM9, aliquots of 2×10^8^cells were harvested by centrifugation, lysed in sample buffer containing sodium dodecyl sulphate (SDS) and β-mercaptoethanol and boiled for 8 min at 100°C. Extracellular ZinT was prepared by filtering through a 22 μm-pore size filter (Millex, Millipore) the supernatant from a volume of culture containing 5×10^8^cells. Extracellular proteins were concentrated to 100 μl by Amicon ultra centrifugal filter devices (10,000 NMWL-Millipore) and incubated overnight at -20°C in 1 ml ice-cold acetone. Each pellet, obtained after 10 min centrifugation at 13,000 × g at 4°C, was resuspended in 10 μl of Lysis Buffer (1 mM EDTA, 100 mM NaCl, 50 mM Tris-HCl, pH 8.0). Proteins were separated by 12% SDS-PAGE and blotted onto nitrocellulose membranes (Hybond C, Amersham). The epitope-flagged proteins were revealed by anti-FLAG M2 monoclonal antibody (Sigma-Aldrich) as primary antibody and anti-mouse HRP-conjugated IgG (Bio-Rad) as secondary antibody. Native ZinT was revealed by rabbit anti ZinT polyclonal antibody (produced by AnaSpec using the synthetic peptide CDYDGYKILTYKSGK) as primary antibody, and goat anti-rabbit HRP-conjugated IgG (Bio-Rad) as secondary antibody. Detection was performed by enhanced chemiluminescence (ECL Advance, Amersham). Studies on ZinT import and preparation of apo and zinc containing-ZinT ---------------------------------------------------------------------- A deleted *zin*T strain (RG-F120) was grown overnight in LB and diluted 1:500 in fresh broth and incubated at 37°C until to OD~600~= 0.5. Subsequently, 25 or 0.25 μg of extracellular tagged-ZinT, derived from the supernatant culture of RG-F116 strain (grown in modM9 for 6 h as described in Western-blot analysis), were mixed to 5×10^8^cells and incubated in LB or LB supplemented with 0.5 mM EDTA at 37°C without shaking. At starting point or after 4 h of incubation the cells were washed three times in PBS to remove external ZinT. Total extracts were analyzed by Western blot. In order to prepare the apo or the holo form of ZinT, extracellular ZinT was isolated from the culture supernatants of the RG-F116 strain grown in modM9 for 6 h at 37°C. Zinc was removed from ZinT by dialysis against 2 mM EDTA, 50 mM acetate buffer, pH 5.4, for 24 h. Subsequently, the protein was dialyzed for 24 h against 100 mM NaCl, 50 mM acetate buffer, pH 5.4 to remove excess EDTA and finally against 50 mM Tris-HCl, pH 6.0. The solutions used for the dialyses of apo-ZinT were prepared with ultra-pure water (0.03 μS/cm) in nitric acid-treated glassware. To prepare holo-ZinT, the apo-ZinT protein was dialyzed for 24 h against 1 mM ZnSO~4~, 50 mM Tris-HCl, pH 7.5, and then extensively dialyzed against 50 mM Tris-HCl, pH 7.5. Protein concentration was evaluated by the method of Lowry \[[@B30]\]. Cell cultures and competition assay ----------------------------------- Human epithelial colorectal adenocarcinoma cells (Caco-2) were cultured at 37°C in humidified air with CO~2~. Caco-2 cell line was maintained in Dulbecco\'s modified Eagle\'s medium (D-MEM) containing 1 g/l glucose, 100 μg/ml penicillin, 100 μg/ml streptomycin, 4 mM L-glutamine and 10% fetal calf serum. For adhesion experiments *E. coli*O157:H7 wild type and mutant strains were grown in LB broth supplemented with 2 mM EDTA. Overnight cultures were diluted in D-MEM to a final concentration of 10^6^cells/ml and then 1 ml of this dilution was used to infect Caco-2 cells previously seeded on a 24-well plate. After two hours of infection each well was washed three times with phosphate buffered saline (PBS), to remove non adherent bacteria, and then lysed with cold Triton X-100 solution (0.5% in PBS). Serial dilutions of the cellular lysates were plated on LB containing kanamycin or chloramphenicol (see Table [4](#T4){ref-type="table"}) to enumerate adherent bacteria. The same approach was used to carry out competitive infections. In this case, the 10^6^cells/ml bacterial suspensions in D-MEM were mixed in pairs in a 1:1 ratio and 1 ml of these mixtures was used to infect Caco-2 cells. Each competition experiment was performed in five different wells and repeated tree times. The infected cells were treated as described above and, after plating of the adherent bacteria, 200 colonies were individually picked on selective plates. The competitive index (CI) was calculated by the formula CI = output (Strain A/Strain B)/inoculum (Strain A/Strain B). Statistical differences between outputs and inputs were determined by the Student\'s *t*-test. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Competition assays in CaCo-2 cells ::: Strain A (relevant genotype) Strain B (relevant genotype) Median CI^a^ P^b^ ------------------------------ ------------------------------ -------------- ---------- Wild type *znuA::cam\** 6.833 0.034 Wild type *zinT::kan\** 0.980 NS Wild type *zinT:: kan znuA:: cam\** 3.899 0.004 *zinT::kan* *zinT:: kan znuA:: cam\** 2.788 \< 0.001 *znuA::cam* *zinT:: kan\* znuA:: cam* 0.697 0.004 **a**. Competitive index = output (Strain A/Strain B)/inoculum (Strain A/Strain B). **b**. Statistical differences between output and inocula (the P-values) were determined by the Students *t*test. NS, not significant. \* Antibiotic used for strains selection ::: To analyse the expression of ZnuA and ZinT during infections, Caco-2 cells infected with the RG-F116 or the RG-F117 strains (which express epitope-tagged ZnuA and ZinT, respectively) were lysed 2 h post-infection, and the lysates were harvested and analysed by Western blot. Results ======= Influence of *zin*T and *znu*A on *E. coli*O157:H7 growth --------------------------------------------------------- We compared the growth of the wild type strain to that of mutant strains deleted of *zin*T (RG112), *znu*A (RG113) or both the genes (RG114). No differences in growth were observed when bacteria were cultivated in LB, whereas the growth of all mutant strains decreased with 0.5 mM EDTA (Figure [1](#F1){ref-type="fig"}, panel A, data not shown for RG114) and even more with 2 mM EDTA treatment (data not shown). A recovery in growth of all mutant strains was observed upon supplementation of ZnSO~4~to the LB containing EDTA. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Growth curves**. [Panel A]{.underline}: growth curves of wild type (squares), Δ*zin*T::*kan*(triangles) and Δ*znu*A::*kan*(circles) in LB medium (close symbols), in LB supplemented with 0.5 mM EDTA (open symbols) and 0.2 mM ZnSO~4~(dotted lines). [Panel B]{.underline}: growth curves of the same strains in modM9 (close symbols) and in modM9 supplemented with 5 μM ZnSO~4~(open symbols). :::  ::: In modM9 all mutant strains displayed a clear growth defect with respect to the wild type strain (Figure [1](#F1){ref-type="fig"}, panel B), with a major impairment of the growth of strains lacking *znu*A (RG114 data not shown) than that of the strain lacking only *zin*T. In this case, however, the addition of ZnSO~4~to the culture medium significantly reduced the rate of growth of the wild type (Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1, panel A) and *zin*T mutant strains, likely due to toxic effects of the extracellular metal. In contrast, a clear improvement in the growth of the strains lacking *znu*A was observed upon the addition of zinc to the medium (Figure [1](#F1){ref-type="fig"}, panel B and Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1, panel B). The growth defect of the *znu*A mutant strain was complemented by a multicopy plasmid overexpressing *E. coli*ZnuA, indicating that disruption of *znu*A does not abolish the functionality of the other genes of the *znu*ABC operon (Table [5](#T5){ref-type="table"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S2). The reduced rate of growth of the complemented strains is likely due to gene dosage effects, as previously described \[[@B17]\]. ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Growth on LB plates ::: Strains^a^ EDTA concentration -------------------------- -------------------- -------- ------ ------ 0 0.5 mM 1 mM 2 mM WT ++ ++ ++ ++ RG113 (Δ*znuA*:: *kan*) ++ +/- +/- \- RG113 + p18ZnuAO157 ++ \+ \+ \+ RG113 + p18ZnuA*E. coli* ++ \+ \+ \+ **a**The strains were grown overnight in LB medium and then streaked on LB plates containing the indicated amounts of EDTA. Growth on agar plates was not modified by the presence or absence of antibiotics. **Symbols**: ++ growth, + weak growth, +/- weak growth of very small colonies, - no growth. ::: ZinT and ZnuA expression studies -------------------------------- The expression of *zin*T and *znu*A was indirectly analyzed by monitoring the proteins accumulation in strains which were modified by introducing the sequence encoding the 3xFLAG epitope at the 3\'end of each gene (Figure [2](#F2){ref-type="fig"}). In agreement with previous studies \[[@B18],[@B21]\], also in *E. coli*O157:H7 cadmium and EDTA were able to induce the expression of ZinT and ZnuA. Moreover, ZnuA accumulation drastically decreased when bacteria were grown in 0.5 mM EDTA in presence of 0.2 mM ZnSO~4~, a quantity unable to saturate the binding ability of the chelator, whereas ZinT accumulation was only moderately affected. Higher zinc concentrations, however, were able to completely abolish ZinT and ZnuA accumulation (data not shown). In addition, the tagged proteins accumulated both in standard LB and in LB supplemented with zinc in *zur*deleted strains, confirming that *zin*T and *znu*A are negatively regulated by Zur, as already observed in other bacteria in previous studies \[[@B4],[@B12],[@B18],[@B31],[@B32]\]. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **ZinT and ZnuA accumulation in *zur*wild type and in *zur*deleted strains**. RG-F116 (*zin*T::3xFLAG-*kan*), RG-F117 (*znu*A::3xFLAG-*kan*), RG-F118 (Δ*zur*::*cat zin*T::3xFLAG-*kan*) and RG-F119 (Δ*zur*::*cat znu*A::3xFLAG-*kan*) strains were grown for 4 h in LB medium in presence or absence of 0.2 mM ZnSO~4~, 0.5 mM EDTA or 0.2 mM CdSO~4~as indicated. The extracts were analyzed by Western blot. :::  ::: To evaluate the specificity of the response of *zin*T and *znu*A to metal ions, the accumulation of the two proteins was analyzed in modM9 supplemented with 5 μM ZnSO~4~, FeSO~4~, CuSO~4~or MnCl~2~. The expression of both genes was repressed by zinc (Figure [3](#F3){ref-type="fig"}) whereas, in contrast to the results obtained with *S. enterica*\[[@B17]\], *znu*A and, to a lesser extent, *zin*T expression was partially inhibited by copper. Small differences in the regulation of the Zur-regulated genes between *E. coli*O157:H7 and *S. enterica*(PP134 and SA140) were also suggested by a titration of protein accumulation in response to external zinc (Figure [4](#F4){ref-type="fig"}). In *E. coli*O157:H7 strains the two genes were similarly expressed, with a slightly higher ZinT accumulation in presence of 0.5 μM ZnSO~4~. In contrast, in *S. enterica*only ZnuA was detectable at this zinc concentration. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Influence of metals on ZinT and ZnuA accumulation**. RG-F116 (*zin*T::3xFLAG-*kan*) and RG-F117 (*znu*A::3xFLAG-*kan*) strains were grown for 16 h in modM9 (lanes 1 and 6) in presence of ZnSO~4~(lanes 2 and 7), FeSO~4~(lanes 3 and 8), CuSO~4~(lanes 4 and 9) or MnCl~2~(lanes 5 and 10). Metal concentration was 5 μM. The extracts were analyzed by Western blot. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Zinc-dependent ZinT and ZnuA accumulation in *E. coli*O157:H7 and *S. enterica*strains**. RG-F116 (*zin*T::3xFLAG-*kan*), RG-F117 (*znu*A::3xFLAG-*kan*) *E. coli*O157:H7 strains or PP134 (*zin*T::3xFLAG-*kan*) and SA140 (*znu*A::3xFLAG- *kan ilv*I::Tn10dTac-*ca*t:: 3xFLAG- *kan*) *S. enterica*strains were grown for 16 h in modM9 supplemented or not with various concentrations of ZnSO~4~, as indicated. The extracts were analyzed by Western blot. In SA140 strain the chloramphenicol acetyltransferase (CAT) was used as an internal standard. :::  ::: The accumulation of the tagged-proteins was analyzed also in mutant strains deleted of *zin*T (RG-F120) or of *znu*A (RG-F121). Figure [5](#F5){ref-type="fig"} shows that ZnuA accumulation in the strain lacking a functional *zin*T was comparable to that observed in the wild type strain in the same conditions (see Figure [2](#F2){ref-type="fig"}). In contrast, ZinT was expressed by the RG-F121 strain either in LB, where it was normally absent (Figure [5](#F5){ref-type="fig"}), or in modM9 supplemented with zinc (Figure [6](#F6){ref-type="fig"}). These observations support the hypothesis that the role of ZinT in zinc homeostasis is dependent on the presence of ZnuA and that this protein is not able to directly deliver zinc to ZnuB. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Different accumulation of ZinT and ZnuA in the deleted strains in LB medium**. RG-F120 (Δ*zin*T::*cat znu*A::3xFLAG-*kan*) and RG-F121 (Δ*znu*A::*cat zin*T::3xFLAG-*kan*) strains were grown for 4 h in LB medium in presence or absence of 0.2 mM ZnSO~4~, 0.5 mM EDTA or 0.25 mM CdSO~4~, as indicated. The extracts were analyzed by Western blot. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Different accumulation of ZinT and ZnuA in the deleted strains in modM9 medium**. The wild type strains RG-F116 (*zin*T::3xFLAG-*kan*) and RG-F117 (*znu*A::3xFLAG-*kan*), and the deleted strains RG-F120 (Δ*zin*T::*cat znu*A::3xFLAG-*kan*) and RG-F121(Δ*znu*A::*cat zin*T::3xFLAG-*kan*) were grown for 16 h in modM9 in presence or absence of 5 μM ZnSO~4~or 5 μM EDTA, as indicated. The extracts were analyzed by Western blot. :::  ::: Extracellular ZinT ------------------ In a previous work ZinT was identified in the culture supernatant of *E. coli*O157:H7 strain and suggested to be a substrate of the type 2 secretion system (T2SS) \[[@B23]\], whereas no studies have yet examined the possibility that ZnuA could be secreted. To investigate this possibility and better characterize ZinT export, total or extracellular extracts from RG-F116 and RG-F117 strains were analyzed. Strains were grown in LB supplemented with 0.5 mM EDTA or 0.25 mM CdSO~4~for only 4 h to prevent the possible release of proteins in the culture medium by lysis of starved bacterial cells. In none of the tested conditions could ZnuA be detected in the culture supernatant (data not shown). In contrast, as shown in Figure [7](#F7){ref-type="fig"} panel A ZinT was detectable in the extracellular fraction of bacteria grown in presence of EDTA but not in that of bacteria cultivated in presence of cadmium, suggesting that the secretion was not possible for Cd-containing ZinT while the sequestration of metals by EDTA likely produced an apo-form able to be secreted outside the cell. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Extracellular ZinT accumulation**. [Panel A]{.underline}: RG-F116 (*zin*T::3xFLAG-*kan*) strain was grown in LB medium supplemented with 0.5 mM EDTA (lanes 1 and 3) or with 0.25 mM CdSO~4~(lanes 2 and 4). After 4 h of growth, total (lanes 1 and 2) or extracellular extracts (lanes 3 and 4) were loaded on SDS-PAGE and analyzed by Western blot. [Panel B]{.underline}: RG-F116 (lanes 1 and 2) and RG-F121 (Δ*znu*A::*cat zin*T::3xFLAG-*kan*) strains (lanes 3, 4, 5 and 6) were grown in modM9 (lanes 1, 2, 3 and 4) or supplemented with 5 μM of ZnSO~4~(lanes 5 and 6). After 6 h of growth, total (lanes 1, 3 and 5) or extracellular extracts (lanes 2, 4 and 6) were loaded on SDS-PAGE and analyzed by Western blot. :::  ::: To verify if protein secretion was prevented by metal binding, ZinT was produced in the RG-F121 strain grown in modM9, supplemented or not with 5 μM ZnSO~4~(Figure [7](#F7){ref-type="fig"}, panel B). This strain was chosen because the absence of *znu*A allows the expression of *zin*T in modM9 also in presence of zinc, an essential condition to carry out the proposed experiment. As expected, an expression band was not visible in the supernatant obtained in presence of zinc whereas this band was observable in absence of the metal for RG-F116 and RG-F121 strains. Additional file [3](#S3){ref-type="supplementary-material"}: Figure S3 shows that *E. coli*O157:H7 secretes only a very limited number of proteins in modM9 and that there is not an evident release of intracellular proteins. In an attempt to identify a role for extracellular ZinT, we investigated the possibility that secreted ZinT could rebind to the bacterial cell. Cultures of RG-F120 strain, bearing a gene encoding a tagged-ZnuA and a deletion in *zin*T, were incubated for 4 h with extracellular tagged-ZinT obtained from the supernatant culture of RG-F116 strain grown in modM9 for 6 h. Subsequently, cellular extracts were analyzed by Western blot to examine the fate of ZinT, using tagged ZnuA as positive control. As shown in Figure [8](#F8){ref-type="fig"}, when RG-F120 was grown in LB or in LB with 0.5 mM EDTA in presence of 25 μg of extracellular ZinT the protein was not found in association with the bacterial cell. Unexpectedly, we observed that extracellular ZinT induced the accumulation of ZnuA in LB (Figure [8](#F8){ref-type="fig"} lane 3), where this protein was hardly detectable (see Figure [2](#F2){ref-type="fig"}). Such induction of *znu*A gene was not observed (Figure [8](#F8){ref-type="fig"} lane 6) in bacteria incubated in presence of a hundredfold lower amount of extracellular ZinT (0.25 μg), suggesting that ZnuA accumulation could be due to the ability of extracellular ZinT to sequester external zinc. To verify this possibility, the experiment was repeated using either apo- or zinc-containing ZinT. ZnuA accumulation appeared in LB only when the apo-form (data not shown) was used, showing the similar expression pattern obtained with the extracellular ZinT produced in modM9. These results indicated that apo-ZinT sequesters environmental zinc thus inducing the *zur*regulon, and that extracellular ZinT released by bacteria grown in modM9 is mainly in the zinc-free form, as already indicated by results described in Figure [7](#F7){ref-type="fig"}. ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Influence of extracellular ZinT on z*nu*A expression**. RG-F120 (Δ*zin*T::*cat znu*A::3xFLAG-*kan*) strain was grown in LB medium (lanes 2, 3, 5 and 6) or LB supplemented with 0.5 mM EDTA (lanes 4 and 7) in presence of 25 μg (lanes 2, 3 and 4) or 0.25 μg (lanes 5, 6 and 7) extracellular ZinT. The extracts, analyzed by Western blot, were prepared after a 4 h growth (lanes 3, 4, 6 and 7), or immediately after the addition of extracellular ZinT (lanes 2 and 5), as negative control. 25 μg of extracellular ZinT was loaded in lane 1 as positive control. :::  ::: In order to obtain strains unable to secrete ZinT we used the RG-F116 strain to delete *etp*C (RG-F122) or *etp*D (RG-F123), the first two genes of the operon of T2SS \[[@B33]\]. Contrary to our expectations, tagged-ZinT was detected in the supernatant of these mutants grown in LB supplemented with 0.5 mM EDTA and its accumulation was comparable to that observed in the wild type strain (data not shown). To exclude that the FLAG-epitope tail could interfere with the export of the protein, we have grown the *etp*C null mutant strain (RG115), where the *zin*T gene was unmodified, under the same experimental conditions. The observation of a band in extracellular extracts, revealed by anti ZinT polyclonal antibody as primary antibody, suggested that T2SS was not the main secretion system for the export of the protein encoded by chromosomal *zin*T (data not shown). Extracellular ZinT was also revealed in the culture supernatant of *E. coli*K12 (DH5α) and B (BL21) strains, by using the same anti ZinT polyclonal antibody (data not shown). This result supports the hypothesis that ZinT is not secreted by T2SS, as in the laboratory strains of *E. coli*the T2SS is transcriptionally silenced by the histone-like nucleoid-structuring protein H-NS \[[@B34],[@B35]\]. Effects of *zin*T and *znu*A deletion on *E. coli*O157:H7 adhesion to Caco-2 cells ---------------------------------------------------------------------------------- It has previously been reported that inactivation of *zin*T has a dramatic effect on the ability of *E. coli*O157:H7 to adhere to HeLa cells \[[@B23]\]. To investigate the relevance of the zinc import apparatus in the *E. coli*O157:H7 interaction with host cells, we have initially analyzed ZnuA and ZinT accumulation in bacteria (RG-F116 and RG-F117) adhering to Caco-2 epithelial cells. Results reported in Figure [9](#F9){ref-type="fig"} indicate that in presence of Caco-2 cells both proteins were expressed at levels that were significantly higher than those observed in bacteria grown in D-MEM. This observation suggests that Caco-2 cells deplete the medium of zinc or that the cell surface microenvironment is poor of zinc. Despite this finding and unlike the results obtained by Ho *et al*. \[[@B23]\] with HeLa cells under slightly different experimental conditions, we were unable to demonstrate that inactivation of *znu*A or *zin*T significantly decreases the ability of *E. coli*O157:H7 to adhere to Caco-2 epithelial cells with respect to the wild type strain (data not show). However, as the number of adherent bacteria was highly variable in different experiments, to better appreciate the contribution of ZnuA and ZinT to the interaction of *E. coli*O157:H7 with Caco-2 cells, we carried out adhesion experiments using mixtures of different strains (Table [4](#T4){ref-type="table"}). These competition experiments revealed that mutant strains lacking *znu*A (RG113 and RG114) were significantly disadvantaged compared to the wild type strain but failed to identify an adherence defect in the strain lacking only *zin*T (RG112). It is worth nothing that the loss of adherence ability of the *znu*A mutant strain is not trivially due to a reduced ability to grow in D-MEM. In fact, co-cultures of the wild type and of the *znu*A mutant revealed that the two strain grow equally well in this medium, indicating that it is likely rich in zinc (data not shown). Moreover, the results reported in Additional file [4](#S4){ref-type="supplementary-material"}: Table S1 show that during co-infection experiments the *znu*A mutant strain replicated more efficiently than the wild type strain, despite it was less able to adhere to the epithelial monolayer. In addition, we observed that the *zin*T/*znu*A mutant strain (RG114) was more able to adhere to epithelial cells than the single *znu*A mutant. This result, which replicates a comparable finding in Salmonella \[[@B17]\], could be tentatively explained by a toxic effect of ZinT in the absence of ZnuA, due to its ability to sequester zinc without being able to transfer the metal to the ZnuB permease. ::: {#F9 .fig} Figure 9 ::: {.caption} ###### **ZinT and ZnuA accumulation in *E. coli*O157:H7 adherent to epithelial cells**. ZinT and ZnuA accumulation of RG-F116 (*zin*T::3xFLAG-*kan*) and RG-F117 (*znu*A::3xFLAG-*kan*) strains, grown overnight in D-MEM (lanes 1 and 4), was compared to accumulation of proteins in bacteria recovered from infected Caco-2 cells (lanes 2 and 3). :::  ::: Discussion ========== The results reported in this work confirm the central importance of the ZnuABC transporter in the process of zinc uptake also in *E. coli*O157:H7. In fact, growth of strains lacking *znu*A, the gene encoding for the periplasmic component of the transporter, is severely impaired in media poor of zinc (LB supplemented with EDTA or modM9), but is identical to that of the wild type strain in LB medium where zinc is abundantly available (Figure [1](#F1){ref-type="fig"}). The growth impairment of *znu*A mutant strains is clearly attributable to the lacking of this gene because it is complemented by plasmids harbouring the *znu*A copy (Table [5](#T5){ref-type="table"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S2). In line with these observations, ZnuA accumulates in bacteria grown in zinc-limiting conditions but is hardly detectable in bacteria recovered from LB (Figures [2](#F2){ref-type="fig"} and [5](#F5){ref-type="fig"}). Accumulation of ZnuA is regulated by zinc and not by manganese or iron as shown in Figure [3](#F3){ref-type="fig"}. However, in line with previous observation by the group of Kershaw \[[@B36]\] on *E. coli*K12 and in contrast to results obtained on *S. enterica*\[[@B17]\], it is somehow modulated by copper. We believe that it is unlikely that ZnuABC participates to the mechanisms of copper homeostasis and we suggest that this effect could be explained by the very similar properties of the copper and zinc atoms which likely allow the accommodation of copper in the zinc binding site of Zur. The results reported in this work provide further evidences that also ZinT participates in the mechanisms of zinc uptake, in line with recent studies \[[@B18],[@B24],[@B25]\]. We have verified that also in *E. coli*O157:H7 *zin*T is regulated by Zur and that it is induced under conditions of zinc deficiency. The absence of *zin*T has no discernable effects on bacterial replication in rich media, but significantly affects growth either in presence of chelating agents or in modM9 (Figure [1](#F1){ref-type="fig"}). However, unlike what observed for the *znu*A mutant, zinc supply does not clearly improve the growth of the *zin*T mutant in modM9 and we could not observe an additive effect of the double mutation *zin*T*/znu*A. These observations corroborate the suggestion that the role of ZinT in zinc uptake is subordinated to that of ZnuA and that zinc ions bound by ZinT are subsequently transferred to ZnuA, which ensures zinc import in the cytoplasm \[[@B18]\]. This consideration is in agreement with the observation that *zin*T is constitutively expressed in a *znu*A mutant strain, but that ZnuA accumulation is not significantly modulated by the absence of *zin*T (Figure [5](#F5){ref-type="fig"}). This is likely explained by a decrease of the zinc concentration in the cytoplasm in the absence of ZnuA, but not of ZinT, with the consequent derepression of *zin*T by Zur. It should be highlighted that the *zin*T mutant strain exhibits a sharp growth defect either in LB supplemented with 0.5 mM EDTA or in defined medium. This behaviour was not observed in a *zin*T mutant of *S. enterica*\[[@B18]\], which showed a clear impairment of growth in LB only in presence of 2 mM EDTA, a concentration at which the *E. coli*O157:H7 mutant is hardly able to grow. Furthermore, our results indicate that there are differences between *E. coli*O157:H7 and *S. enterica*in the regulation of *znu*A and *zin*T in response to low zinc availability (Figure [4](#F4){ref-type="fig"}). In particular, in *E. coli*O157:H7 ZinT can be easily detected in bacteria growing in a medium supplemented with up to 1 μM zinc, whereas in *S. enterica*this protein accumulates only in media completely devoid of the metal. This observation, which is in agreement with the different effect of *zin*T disruption in the two bacterial species, may suggest that the relative role of ZnuA and ZinT could be slightly different in the two microorganisms. Although several of the bacteria which rely on the ZnuABC transporter to import zinc do not possess ZinT \[[@B18]\], our study suggests that, despite the role of ZinT is clearly dependent on the presence of ZnuA, its contribution to metal recruitment within the periplasmic space is considerable. The exact involvement of ZinT in zinc uptake is yet to be determined, but it is possible to hypothesize that ZinT and ZnuA display a diverse ability to sequester metal ions from different molecules within the periplasm or that the binding of ZinT to ZnuA accelerates the rate of metal transfer to ZnuB \[[@B18]\]. We have also analyzed the involvement of the zinc uptake system in the interaction between *E. coli*O157:H7 and epithelial Caco-2 cells. Both ZnuA and ZinT accumulates at high levels in bacteria adhering to the cell monolayer, but not in bacteria cultivated in D-MEM without cells (Figure [9](#F9){ref-type="fig"}). This finding expands previous observations showing that bacterial pathogens have to face with a problem of zinc paucity within the host \[[@B17]\] and specifically suggests that the host cell surface microenvironment is poor of zinc, possibly due to active metal sequestration mechanism implemented by eukaryotic cells. In line with this observation strains lacking *znu*A display a reduced ability to adhere to epithelial cells (Table [4](#T4){ref-type="table"}). We could not observe significant alterations in the adhesion ability of the *zin*T mutant strain nor an additive effect of the *zin*T*/znu*A mutations, confirming the subordinate role of ZinT already revealed by the analysis of growth curves *in vitro*(Figure [1](#F1){ref-type="fig"}). This last finding is in contrast with the recent results reported by Ho and colleagues \[[@B23]\] who analyzed the role of YodA (ZinT) in the *E. coli*O157:H7 strain EDL933, observing that the *zin*T mutant strain exhibits a dramatic reduction in its ability to adhere to HeLa cells and to colonize the infant rabbit intestine \[[@B23]\]. Furthermore, they observed a reduction in growth of the *zin*T mutant also in LB medium. In principle, divergences between these two studies could due to genotypic differences between the strains employed or to differences in the *E. coli*ability to interact with different eukaryotic cell lines. However, it is worth nothing that the reduction in growth of the *zinT*mutant in LB medium observed by Ho *et al*. is unexpected on the basis of the presumed role of ZinT in zinc import and that, in line with the here reported results, *zin*T mutants of *S. enterica*\[[@B18]\] and *E. coli*K12 \[[@B24],[@B25]\] grow as well as the wild type parental strains in zinc replete media. Moreover, Ho and colleagues identified ZinT even in the culture supernatants of *E. coli*O157:H7 strain and suggested that it is a substrate of the type 2 secretion system (T2SS) \[[@B23]\]. We have confirmed that a fraction of ZinT is actually exported selectively (ZnuA is not secreted) in the culture medium (Figure [7](#F7){ref-type="fig"}), but we failed to validate the suggestion that the secretion of this protein is facilitated by T2SS. In fact, ZinT is exported with comparable efficiency by the wild type strain or by mutant strains lacking *etp*C or *etp*D genes which encode for two different components of the T2SS gene cluster \[[@B33]\]. Moreover, we observed that ZinT is secreted also in *E. coli*K12 and B strains. This observation strongly argues against the involvement of T2SS in the export of ZinT because the genes encoding for the T2SS system are not expressed in *E. coli*K12 due to the repression by the histone-like nucleoid-structuring protein H-NS \[[@B34],[@B35]\]. We hypothesize that the different result obtained by Ho et *al*. could be explained by their choice to analyze the secretion of ZinT in a strain overexpressing a V5-tagged ZinT. The T2SS might be involved in the recognition of this specific tag or in the secretion of proteins when overexpressed \[[@B37]\]. In any case, the T2SS system seems not to participate in the secretion of chromosomally encoded ZinT. We have demonstrated that ZinT can be exported in the extracellular environment only in the metal free form. In fact, when ZinT is constitutively expressed in bacteria grown in media containing cadmium or zinc, it can not be identified in the culture supernatants, despite it is present in the periplasmic space (Figure [7](#F7){ref-type="fig"}). The release of metal-free ZinT in the extracellular environment may influence properties of the bacterial or host cells. This possibility is partially supported by the experiment showing that apo-ZinT, unlike the zinc containing protein, is able to influence *znu*A expression when provided externally to bacterial cells (Figure [8](#F8){ref-type="fig"}). The observed accumulation of ZnuA is likely due to the ability of ZinT to sequester the free zinc present in the culture medium, inducing a condition of zinc starvation. Although we have analyzed the effects of extracellular ZinT only on the bacterial cell, we hypothesize that the sequestration of extracellular zinc may have effects also on the host cells. In this view, it is interesting to note that several bacteria produce metal binding proteins located on the cell surface which mediates the microbial attachment to the human extracellular matrix. Proteins of this class include, for example, the laminin binding proteins (LBP) from *Streptococcus agalactiae*or *Streptococcus pyogenes*, which are structurally related to ZnuA \[[@B38],[@B39]\]. Although the details of the interaction of LBP with laminin are still to be clarified, it is likely that LBP acts as an adhesin which binds to the zinc containing laminin in a metal-mediated manner. By analogy, we suggest that extracellular ZinT may interact with zinc-containing proteins in the intestinal epithelia, thus favouring *E. coli*O157:H7 colonization, or that its capability to sequester zinc ions from the environment may damage epithelial cells ability to neutralize bacterial adhesion. Conclusions =========== This study demonstrates that the high affinity ZnuABC uptake system plays a key role in zinc uptake in *E. coli*O157:H7 and that ZinT is an additional component of this metal transport system which significantly enhances the rate of metal uptake. In addition, our data indicate that the functionality of this transporter may influence the adhesion of bacteria to epithelial cells. These findings improve our knowledge about the importance of zinc in bacterial physiology and its role in the host-microbe interaction. Authors\' contributions ======================= RG and RS coordinated the study, participated to the manuscript preparation, carried out *E. coli*O157:H7 mutants construction, performed growth curves, complementation assay and *in vitro*expression studies, PP carried out studies with cultured cells, SA collaborated in the preparation of strains and to the set up of zinc free media, AB and LN participated in the design of the study and in the writing of the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Figure S1: Influence of zinc on modM9 growth curve**. The figure shows the growth curves of wild type and D*znu*A::*kan*strains in modM9 supplemented with various concentrations of ZnSO~4~(0.25 mM, 0.5 mM, 1 mM and 5 mM). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Figure S2: Growth curve of the complemented D*znu*A::*kan*strain in modM9**. The figure shows as the growth curves of D*znu*A::*kan*containing the plasmid p18ZnuAO157 or p18ZnuAE. *coli*are improved respect to that of D*znu*A::*kan*. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Figure S3: Expression pattern of *zin*T in SDS-PAGE**. The figure shows the total extracellular extracts of *zin*T::3xFLAG-*kan*analysed by SDS-PAGE and stained by Coomassie- Blue or revealed by Western blot. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **Table S1: Competition assays in CaCo-2 cells**. The table shows as during co-infection experiments the *znu*A mutant strain replicated more efficiently than the wild type strain. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was partially supported by ISS grant to RG

PubMed Central

2024-06-05T04:04:19.469674

2011-2-21

PMC3053224

Background ========== *Yersinia enterocolitica*(YE) is an enteropathogenic bacterium transmitted via food or water and may cause sporadic infections as well as foodborne outbreaks of yersiniosis \[[@B1]-[@B5]\]. The symptoms of yersiniosis range from mild diarrhea to severe clinical manifestations and post-infectious complications such as reactive arthritis, myocarditis, glomerulonephritis, and erythema nodosum \[[@B6]\]. *Y. enterocolitica*can be divided into six biotypes, of which biotypes 1B and 2-5 are known to be pathogenic to humans. At present, pulsed-field gel electrophoresis (PFGE) is commonly used to discriminate between *Y. enterocolitica*strains. However, there are no standard PFGE procedures or databases similar to those, *e.g.*, for *Escherichia coli*O157:H7, *Salmonella*, and *Shigella*standardized by PulseNet \[[@B7]\]. Most of the restriction enzymes used in PFGE for *Y. enterocolitica*produce patterns with a high number of bands that are not ideal for analysis. Furthermore, the global homogeneity of the pulsotypes among *Y. enterocolitica*4/O:3 is high and different pulsotypes often display only minor differences \[[@B8]-[@B11]\]. However, the discriminatory power of PFGE has been improved by using more than one restriction enzyme \[[@B12]\]. Most bacterial genomes contain repeats of DNA sequences called \"variable-number tandem repeats\" (VNTR). These VNTR regions can be applied in the PCR-based subtyping of strains by multilocus variable-number tandem-repeat analysis (MLVA). MLVA is increasingly used for typing, surveillance and epidemiological investigations of pathogenic bacteria \[[@B13]\]. A study investigating the development of an MLVA subtyping method to be used for *Y. enterocolitica*4/O:3, based on six loci, was reported recently \[[@B14]\]. Although yersiniosis is seldom treated with antimicrobials, medication may be required, for example in the case of immuno-compromised patients. *Y. enterocolitica*is a known ß-lactamase producer and thus is resistant to ß-lactam antibiotics such as ampicillin, carbenicillin, penicillin, and first-generation cephalosporins \[[@B15]-[@B20]\]. In recent studies done in Switzerland, the USA, Germany, and Austria, *Y. enterocolitica*strains have shown high susceptibility to antimicrobials other than ß-lactams \[[@B21]-[@B24]\]. However, multiresistant *Y. enterocolitica*strains have also been reported, *e.g.*, from Spain and Brazil \[[@B16],[@B25],[@B26]\]. The antimicrobial resistance of *Y. enterocolitica*has not been monitored regularly in Finland although the surveillance of antimicrobial resistance would be useful for epidemiological studies. Over 20 years ago, 186 Finnish *Y. enterocolitica*strains were studied and found to be resistant only to ampicillin and susceptible to ceftriaxone, tetracycline, sulpha-trimethoprim, and ciprofloxacin \[[@B27]\]. The aim of the present study was to determine how MLVA using fluorescently labeled primers and fragment analysis compares to PFGE in its discriminatory power with regard to the sporadic and outbreak-related strains of YE bio/serotypes 4/O:3. We included traditional antimicrobial susceptibility testing in our study to see whether it provides additional information for the genotypic analysis concerning, *e.g.*, the geographical source of infection. We therefore used MLVA and PFGE to type 104 sporadic and outbreak-associated *Y. enterocolitica*strains and determined the sensitivity of the strains to 12 antimicrobial agents. In addition, we studied the genetic basis of the antimicrobial resistance that was detected. Results ======= Sporadic strains ---------------- The strains isolated in 2006 (n = 82) were discriminated into 77 types by MLVA (Figure [1](#F1){ref-type="fig"}) and into 23 pulsotypes by PFGE (Figure [2](#F2){ref-type="fig"}). There were two YE 4/O:3 strains with identical MLVA types in only five cases. In two of these cases, the identical strains had been isolated from one patient 7 days apart and from another patient 19 days apart. The discriminatory index for sporadic strains was 0.862 for PFGE and 0.999 for MLVA. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **MLVA tree**. UPGMA clustering of the MLVA results, with Pearson\'s correlation similarity coefficients, was performed using Bionumerics version 5.10. The key column provides the strain ID. Information on bio/serotype, travel abroad or place of domicile (PoD), MLVA types named as a string of six numbers showing the actual number of repeat units in each of the six loci, PFGE pulsotype, and antimicrobial resistance are presented in the columns. \*Strains isolated from a 1-year old children in the case of a suspected outbreak with PFGE pulsotype 5NotI\_ye\_a. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **PFGE types of the studied strains**. All 24 representative PFGE types of 104 strains in the present study. \* The strain number includes the outbreak types. :::  ::: The six loci used in MLVA V2A exhibited the highest discriminatory power (DI = 92%), resolving 17 different alleles. The least variation was observed for locus V9 (DI = 62%), which yielded only six different alleles, *i.e.*, 2-7 repeats of a repetitive sequence 12 bp in length. The discriminatory indexes of loci V4, V5, V6, and V7 were 71, 89, 91, and 90%, respectively. The fragment sizes defined by the capillary electrophoresis of the six VNTR loci and the number of repeats confirmed by DNA sequencing are shown in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Diversity of VNTR alleles. ::: Number of the repeats V2A TCTCAC (bp) n† V4 CGGCAAC (bp) n V5 GGTGCA (bp) n V6 GACTCA (bp) n V7 GTGCTG (bp) n V9 ATGTCGGTAGAA (bp) n ----------------------- ----------------- ---- ----------------- ---- ---------------- ---- ---------------- ---- ---------------- ---- ---------------------- ---- 2 \- 119\* 49 \- \- 108 2 3 246\* 2 126\* 26 182 1 \- 120\* 52 4 252 5 133\* 9 199 2 188\* 5 195\* 4 132 8 5 258 6 140 0 205 4 194 5 201\* 8 144\* 40 6 264 10 147 15 211 3 200\* 11 207 19 156 2 7 270 6 154\* 4 217 6 206\* 21 213 13 168\* 3 8 276 6 161 4 223\* 25 212\* 13 219 12 \- 9 282\* 7 \- 229 17 218 2 225 9 \- 10 288 10 \- 235 15 224 12 231 9 \- 11 294 6 \- 241 8 230 9 237 7 \- 12 300 10 \- 247 6 236 10 243 1 \- 13 306 20 \- 253 6 242 4 249 4 \- 14 312 4 \- 259 5 248 2 255\* 16 \- 15 318 3 \- 265\* 5 254 3 261 3 \- 16 324 1 \- 271 3 260 3 267 \- \- 17 330\* 7 \- 277 1 266 2 273 \- \- 18 336 3 \- 283 \- 272 \- 279 \- \- 19 342 1 \- 290 1 278 \- 285 \- \- 20 \- \- \- 284 2 291 1 \- 21 \- \- \- 300 2 297 \- \- 22 \- \- \- 303\* 1 \- Fragment sizes (bp) defined by capillary electrophoresis of VNTR alleles with different number of repeats and their diversity in 107 studied *Y. enterocolitica*strains. **\***Alleles were sequenced to confirm the number of repeats. **†**n = number of strains ::: Suspected outbreak strains -------------------------- The suspected outbreak strains (n = 22) from December 2003 to January 2004 represented four PFGE pulsotypes, which were further typed into 11 MLVA types. Twelve of the strains were identical in MLVA type. Eleven of these strains with identical MLVA types were isolated from the patients with an epidemiological connection to the disease outbreak. The 12 strains with identical MLVA type represented 2 slightly different (only one band difference) PFGE pulsotypes (Figure [2](#F2){ref-type="fig"}) and were multiresistant to antimicrobials (Figure [1](#F1){ref-type="fig"}). Among these strains, eleven were resistant to AMP, CHL, STR SUL, and TET; one strain was susceptible to TET. The suspected outbreak strains with different MLVA types did not have a proved connection to the city of Kotka, Finland. Nine of these strains were susceptible to all the tested antimicrobials except AMP and eight of them shared the same PFGE type. One of the strains (IH250258) had an antimicrobial resistance profile and a PFGE pulsotype identical to those of the outbreak strains. However, the different MLVA type and the lack of epidemiological connection distinguished this particular case from the outbreak-associated cases (Figure [2](#F2){ref-type="fig"}). Suspected YE 4/O:3 outbreak strains isolated in 2006 from six 1-year-old children displayed the same PFGE pulsotype (5NotI\_ye a). However, the MLVA discriminated all six strains. Association between the antimicrobial resistance and travel ----------------------------------------------------------- All the *Y. enterocolitica*strains studied here were resistant to ampicillin. Fifteen (19%) of 80 sporadic strains isolated in 2006 from 80 patients were resistant to four or five of the antimicrobials tested (Table [2](#T2){ref-type="table"}). The multiresistant strains belonged to certain PFGE pulsotypes (1NotI\_ye, 3NotI\_ye, 7NotI\_ye, 15NotI\_ye) that did not contain any susceptible strains. The travel history of 70 of the 80 patients was known. Of these patients, 46% (32/70) had traveled abroad before the onset of symptoms. Travel abroad was significantly (p = 0.002) associated with the antimicrobial multiresistance of *Y. enterocolitica*: 34% (11/32) of the patients with and 5% (2/38) of the patients without a trip abroad had a multiresistant *Y. enterocolitica*strain. Three strains resistant to nalidixic acid had decreased susceptibility (0.25, 0.25, or 0.5 mg/L) to ciprofloxacin in MIC determination. Sequencing of these three nalidixic acid resistant strains revealed amino acid changes due to the point mutations in the *gyrA*gene; *i.e.*, Ser83 to Arg or Asp87 to Asn or Asp87 to Tyr. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Antimicrobial resistance and travelling. ::: Resistance profile Suspected outbreak YE 4/O:3 (n = 22) Sporadic YE 4/O:3 (n = 75) Sporadic YE 3/O:3 (n = 2) Sporadic YE 2/O:9 (n = 5) ------------------------- -------------------------------------- ---------------------------- --------------------------- --------------------------- AMP CHL STR SUL NAL \- 3 (100%)\* \- \- AMP CHL STR SUL TET 12 (0%) 5 (40%) \- \- AMP CHL STR SUL TET MEC \- \- 1 (100%) \- AMP CHL STR SUL 1 (0%) 5 (80%) \- \- AMP CHL STR SUL MEC \- \- 1 (100%) \- AMP SUL \- 1 (100%) \- \- AMP STR \- 1 \- \- AMP MEC \- \- \- \- AMP 9 (0%) 61 (28%) \- 5 (0%) Antimicrobial resistance profiles of the sporadic *Y. enterocolitica*strains isolated in Finland in 2006 and suspected outbreak strains from 2003-2004 and related travel information. \* The percentage of the patients who had reported having traveled abroad before getting ill is in the parenthesis. ::: Conjugation of resistance plasmid --------------------------------- In the conjugation experiment, a sporadic YE 4/O:3 strain FE81008 (resistant to AMP, CHL, STR, SUL, and NAL) was able to transfer the CHL, STR, and SUL resistances to strain YeO3-U by conjugation. The conjugation frequency was 10^-5^-10^-6^. This indicated that the genes encoding resistance to CHL, STR, and SUL were carried on a conjugative plasmid. Indeed, plasmid isolation demonstrated that the recipient strain had received a large 30-40 kb plasmid. Discussion ========== In our study, MLVA typing using fluorescently labeled primers and fragment analysis was shown to be a high-resolution discriminatory method for epidemiological investigations of *Y. enterocolitica*. In the present study, the discriminatory power of MLVA was 99.9% while that of *Not*I PFGE was 87.9%. Our results were in agreement to those obtained by Gierczyński and colleagues \[[@B14]\] who demonstrated that the used MLVA markers are highly discriminatory and added the evidence that this method can successfully be applied for the outbreak strains of *Y. enterocolitica*ssp. *palearctica*biotypes 2 and 4. In the present study, only the VNTR loci V2A, V4 and V5 were found in six BT 1A strains tested with the MLVA method (data not shown). Another MLVA method designed using *Y. enterocolitica*ssp. *enterocolitica*strain 8081 whole genome and with four loci was introduced recently \[[@B28]\]. The method showed potential for the epidemiological investigation for YE biotype 1A strains with DI of 87% and worked also for six tested BT 2 and BT4 strains \[[@B28]\]. The discriminatory power of PFGE can be improved by using more than one restriction enzyme. For instance, the discriminatory index of 74% achieved with *Not*I PFGE was increased to 93% by using further characterization with *Apa*I and *Xho*I enzymes of 128 YE 4/O:3 strains \[[@B29]\]. However, both the time required and the costs of PFGE rise considerably when several restriction enzymes are used. The amount of working time needed for the PFGE protocol with one enzyme is two to three days, MLVA using fragment analysis can be done in one day. In December 2003, authorities from the city of Kotka, Finland reported an outbreak of gastroenteritis. Investigations revealed that it was caused by *Y. enterocolitica*4/O:3 \[[@B30]\]. Approximately 30 people fell ill; 12 patients had culture-confirmed, multiresistant YE 4/O:3 infection. Three of them had appendectomies before the disease was recognized as yersiniosis. Most of the patients had abdominal pain (94%), fever (78%), and diarrhea (72%). Most of the patients had eaten in the same cafeteria in the Port of Kotka between November 25 and December 15, 2003. However, the contaminated food source was not discovered, as the food served in that cafeteria was no longer available for microbiological investigations. In addition, a cohort study among cafeteria users did not show a significant association between any food and illness. During a microbiological sampling of the cafeteria\'s kitchen a month later, in January 2004, hygienists noticed some shortcomings in food handling and hygiene practices that increased the possibility of cross-contamination in the cafeteria. While no YE 4/O:3 strains were found in the specimens collected from the cafeteria, YE biotype 1A strains were isolated from iceberg lettuce imported from Spain and from domestic carrots. Unfortunately, the antimicrobial susceptibilities of these strains are not known. At the time of the outbreak in Kotka, there were around 20 confirmed YE 4/O:3 cases in other locations in Finland, mainly in the Turku area. The cases were suspected to be linked with the larger outbreak, but no epidemiological evidence for this was found. MLVA played a key role in confirming that the cases which occurred in the city of Kotka in 2003 belonged to a single outbreak: 12 isolates representing the Kotka outbreak were clonal by MLVA, and differed distinctly from those of epidemiologically unrelated strains that shared the same PFGE pulsotype. Another suspicion of outbreak was refuted by MLVA: six 1-year-old children had been infected in 2006 by YE 4/O:3 strains that shared the same PFGE pulsotype (5NotI\_ye a). Interviews, however, revealed no epidemiological connection between the cases. All of these strains which shared the same PFGE pulsotype were found to be of different types in MLVA. We also detected some evidence that the MLVA method can be as useful with YE 2/O:9 outbreaks as it was with YE 4/O:3. In a household outbreak in 2009, a mother and two children had YE 2/O:9 strains found to be identical in MLVA (data not shown here). MLVA also identified identical YE 2/O:9 strains in a school/day care center outbreak that occurred in Finland in 2010 (data not shown here). Support was obtained for genetic stability among sporadic cases, since two MLVA-typed strains were isolated twice from the same patient at intervals of 7 or 19 days. In both cases, the MLVA and PFGE types remained identical. Similar observations of the stability of the *Y. enterocolitica*MLVA markers\' loci *in vivo*had also been reported earlier \[[@B14]\]. Genetic events will eventually alter the MLVA patterns, but the rate of alteration is not known. However, previous studies confirmed that the MLVA type remained the same after as many as 20 serial passages of colony plating \[[@B14]\]. Our previous case-control study revealed that travel abroad was a risk factor for *Y. enterocolitica*infection in Finland \[[@B31]\]. In the present study, we found a statistically significant association between the antimicrobial multiresistance of YE strains and travel. The results indicate that a considerable number of multiresistant *Y. enterocolitica*infections are actually imported, and that domestic *Y. enterocolitica*strains in Finland are rather susceptible to antimicrobials. For instance, all of the nalidixic acid-resistant strains were isolated from patients who had been infected while on vacation in Spain or Brazil, countries where multiresistant *Y. enterocolitica*strains have been described previously \[[@B16],[@B25],[@B26]\]. The multiresistant strains belonged to certain PFGE pulsotypes, which were not found among susceptible strains. This is perhaps due to the DNA of the resistance plasmid. The MLVA types were so varied that no hint of the origin of the strains could be obtained on that basis. In the outbreak that occurred in Kotka, the patients had not been abroad before falling ill. However, the antimicrobial multiresistance of the outbreak strain nevertheless suggests that the strain originated from abroad. Spanish iceberg lettuce, at least, had been used in the cafeteria. In 2005 *Salmonella enterica*serotype Typhimurium, with a resistance profile identical to that detected now for the *Y. enterocolitica*outbreak strain, was isolated in an outbreak situation in Finland and traced to iceberg lettuce imported from Spain \[[@B32]\]. The resistance of *Y. enterocolitica*to NAL is based on point mutations in the fluoroquinolone resistance-determining regions of *gyrA*\[[@B26],[@B33]\]. In our study, the strains resistant to NAL had amino acid changes stemming from point mutations in the *gyrA*gene: *i.e.*, either Ser83Arg, Asp87Tyr, or Asp87Asn. Two of these mutations are identical to those reported previously for fluoroquinolone-resistant *Y. enterocolitica*strains \[[@B33]\]. Conjugation experiments confirmed that in *Y. enterocolitica*, the antibiotic resistance to CHL, STR, and SUL, at least, is encoded on a large conjugative plasmid and can easily be transferred to a susceptible *Y. enterocolitica*strain. Conjugative plasmids that carry antibiotic resistance genes have been isolated from a variety of clinical strains, but reports of this for *Y. enterocolitica*are rare. Hundreds of different antibiotic resistance cassettes have been identified as residing on mobile resistance integrons \[[@B34]\]; owing to the cassette nature of the resistance genes, they can easily change the resistance repertoire. In fact, one of the outbreak strains in our study had altered antimicrobial resistance and lacked resistance to TET. A study on the persistence of TET-resistant *E. coli*in colonic microbiota observed that three out of 13 strains lost TET resistance during intestinal colonization \[[@B35]\]. Conclusions =========== MLVA was less labor-intensive than PFGE and the results were easier to analyze, especially because they were independent of subjective interpretation. PFGE can still be useful for surveillance of the sources and transmission routes of sporadic *Y. enterocolitica*strains in future. However, for outbreak investigations, MLVA offers a powerful tool for the discrimination of *Y. enterocolitica*strains. More sporadic and outbreak *Y. enterocolitica*strains should be subjected to MLVA typing in order to determine whether this method could be considered for use as a new gold standard for outbreak investigations of *Y. enterocolitica*. This study revealed that multiresistant *Y. enterocolitica*strains do appear in Finland, but that the multiresistance was mainly associated with travel. All three nalidixic acid resistant strains were associated with travel to Spain or Brazil. Interestingly, all outbreak strains studied here were also multiresistant. Thus, traditional susceptibility testing provides additional information useful for genetic typing methods in epidemiological investigations. Methods ======= Bacterial strains ----------------- Sporadic *Y. enterocolitica*strains (n = 82) of bio/serotype 4/O:3 (n = 75), 3/O:3 (n = 2), 2/O:9 (n = 5) isolated in 2006 from fecal samples of 80 Finnish patients in ten regional clinical microbiology laboratories were used in the study. The patients\' mean age was 34 years (range 0.6-80); 55% of them were men. Isolation and identification of the strains were described previously \[[@B36]\]. In addition, 22 clinical *Y. enterocolitica*strains isolated between December 2003 and January 2004, and suspected of being associated with a *Y. enterocolitica*outbreak in Kotka, were studied. MLVA ---- For MLVA, we had three additional reference strains: NCTC 1176 (4/O:3); NCTC 11174 (2/O:9); and NCTC 10563 (3/O:5,27). DNA was extracted from the strains using the Jet Flex Extraction Kit (Genomed; Löhne, Germany) according to the instructions provided by the manufacturer and eluated in 100 μL TE-buffer. In the MLVA analysis, six known VNTR loci of the strains were amplified in two multiplex PCRs. Previously described primers \[[@B14]\] were labeled with ABI PRISM^®^fluorescent dyes, PET, NED, 6-FAM, or VIC (Applied Biosystems, Foster City, CA). Primers were used in two separate multiplex PCRs with the VNTR loci of V2A (PET), V4 (NED), and V6 (6-FAM), as well as V5 (NED), V7 (VIC), and V9 (PET). Multiplex PCRs were performed with QIAGEN Multiplex PCR kit (Qiagen, Hilden, Germany) according to the manufacturer\'s instructions in a total volume of 25 μl. The primer concentrations were 0.2 μM (V2A), 0.16 μM (V4), and 0.2 μM (V6) in the first PCR and 0.2 μM (V5), 0.2 μM (V7), 0.12 μM (V9) in the second PCR. The template DNA concentration was approx. 10 ng. Touchdown PCR was performed with 15 min initial denaturation at 95°C, followed by 9 cycles 30 s denaturation at 95°C, 30 s annealing at 63°C-55°C (decreasing by 1°C with every cycle), and elongation at 72°C with an additional 25 cycles with annealing 30 s at 58°C. The two PCR products of each strain were mixed, diluted to 1/500 in sterile water, and run in capillary electrophoresis with an ABI 3730xl DNA Analyzer (Applied Biosystems, Foster City, CA) using G5 (DS-33) fragment analysis chemistry according to the manufacturer\'s instructions. The GeneScan™ 600 LIZ^®^(Applied Biosystems) was used as an internal size standard and the data were analyzed using GeneMapper v4.0 software (Applied Biosystems). MLVA types were named as a string of six numbers showing the actual number of repeat units in each of the six loci. DNA sequencing -------------- The selected repeats (Table [1](#T1){ref-type="table"}) were sequenced in both directions with MLVA primers \[[@B14]\]. The *gyrA*gene PCR was performed for 77 sporadic *Y. enterocolitica*strains of bio/serotypes 4/O:3 and 3/O:3 with primers gyrAY1 (5\'-CGC GTA CTG TTT GCG ATG AA-3\') and gyrAY2 (5\'-CGG AGT CAC CAT CGA CGG AA-3\') as earlier described (35) (GenBank/EMBL/DDBJ accession numbers FN821873-FN821949). Sequencing was done in both directions with a Big Dye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems) with an ABI 3730xl DNA Analyzer (Applied Biosystems). PGFE ---- PFGE was performed using the previously described protocol for *Salmonella*\[[@B7],[@B37]\] with modifications: Strains were cultured overnight at 30°C on R1-agar and suspended in CBS-buffer (100 mM Tris:100 MM EDTA, pH 8.0) to a final turbidity of 0.38-0.39 at A~480~. Lysozyme (Roche Diagnostics GmbH, Mannheim, Germany) was added to the 400 μl bacterial suspensions to reach a final concentration of 1 mg/ml. The tubes were mixed and incubated for 15 min at 37°C and then heated to 50°C, after which 400 μl of 1% agarose (SeaKem Gold Agarose, Cambrex Bio Science Rockland, Inc, USA) and proteinase K (at a final concentration of 0.24 mg/ml, Roche Diagnostics GmbH, Mannheim, Germany) were added. The tube contents were cast into plugs, which were transferred into 3 ml of lysis buffer (50 mM Tris:50 mM EDTA, pH8.0 + 1% Sarcosyl) containing 1 mg/ml of proteinase K. The plugs were incubated at 54°C for 2 h and rinsed three times in sterile water and three times in TE buffer at 50°C. The plugs were then stored in 1 × TE buffer at 4°C. The released genomic DNA in the plugs was digested overnight at 37°C with 8 U of the restriction enzyme *Not*I (New England Biolabs, Ipswich, MA, USA). Electrophoresis was carried out in a 1% agarose gel in 0.5 × TBE buffer at 14°C with a switching time of 1 to 18 s for 40 h at 14°C with CHEF Mapper system (Bio-Rad Laboratories, Richmond, California). DNA of the *Salmonella enterica*serotype Braenderup strain H9812, digested with *Xba*I (Roche GmbH, Mannheim, Germany), was used as a size marker. The PFGE types were analyzed with Bionumerics v. 5.10 software (Applied Maths, Sint-Martens-Latem, Belgium). DNA bands smaller than 54.7 kb were excluded from the analysis. Discriminatory index of PFGE and MLVA ------------------------------------- Simpson\'s Index of diversity was used to calculate the discriminatory index (DI) of PFGE and MLVA \[[@B38]\]. In addition, the DIs of each MLVA locus was calculated. Susceptibility testing ---------------------- The antimicrobial susceptibility of the *Y. enterocolitica*isolates was determined using a set of 12 antimicrobials: ampicillin (AMP); chloramphenicol (CHL); streptomycin (STR); gentamicin (GEN); sulfonamide (SUL); tetracycline (TET); trimethoprim (TMP); ciprofloxacin (CIP); nalidixic acid (NAL); cefotaxime (CEF); mecillinam (MEC); and imipenem (IMI). The susceptibility tests were done using the agar diffusion technique on Mueller-Hinton agar according to the CLSI guidelines \[[@B39]\]. A strain resistant to at least four antimicrobials was called multiresistant. The minimal inhibitory concentration (MIC) for ciprofloxacin (CIP) was determined by the E-test (AB Biodisk, Solna, Sweden) for the isolates resistant to nalidixic acid, following the recommended MIC breakpoints S ≤1 mg/L and R ≥4 mg/L \[[@B39]\]. MIC 0.125-1.0 mg/L was considered to indicate reduced susceptibility to ciprofloxacin \[[@B40]\]. Conjugation experiments ----------------------- In conjugation experiments, the multiresistant (AMP, CHL, STR, SUL, NAL) strain YE 4/O:3 FE81008 was used as a donor strain and the kanamycin (KAN) resistant strain YeO3-U \[[@B41]\] as a recipient strain. Briefly, the donor strain and the recipient strain were grown overnight at room temperature shaking in 5 ml of Luria broth (LB). The cultures were refreshed by diluting them 1:10 in LB and grown for 2-3 h to get them into the exponential phase. The donor strain was grown in static culture. The bacteria were then pelleted by centrifugation and resuspended in 1 ml of PBS. After the OD600 were determined, the suspensions were mixed 1:1 and small droplets of the mixture were pipetted onto a Luria-agar plate and incubated overnight at room temperature. Only the donor or the recipient bacteria was pipetted onto the control plates. The plates were incubated overnight after which the bacteria were collected from the plates into ca. 1 ml of PBS. Several dilutions were spread on selective plates containing CHL, KAN, or both CHL and KAN. The conjugation frequency was calculated on the basis of the proportion of CHL KAN double-resistant colonies among the CHL-resistant colonies. The resistance of the CHL KAN double-resistant colonies to the other antimicrobials was tested as described above. Plasmid isolation from 100 ml cultures of the strains was performed using the E.Z.N.A plasmid midiprep kit (Omega Bio-Tek Inc., Norcross, GA, USA) according to the protocol provided by the manufacturer, and the plasmids were detected by running in a 1% w/v agarose gel. Travel information and statistical method ----------------------------------------- Data on the patients\' travel abroad were collected from the National Infectious Disease Register and from the notes of the laboratories sending the *Yersinia*strains for further typing. The association between travel and multiresistance was analyzed by using the chi-square method with the EpiInfo™ version 3.4.3. A p-value below 0.05 was considered to indicate statistical significance. The study was approved by the Ethics Committee of National Institute for Health and Welfare, THL. For this study informed consents were not required as only the isolated bacterial strains of the fecal samples were studied and not the individuals themselves. Authors\' contributions ======================= LMS participated in the design of the study, did or supervised the MLVA, PFGE, DNA sequencing, and antimicrobial susceptibility testing, carried out the data analysis, and drafted the manuscript. ST performed the conjugation experiment. KH participated in the design of the study and drafting of the manuscript. MK did the epidemiological investigations of the study and edited the manuscript. MS designed the conjugation experiment and participated in drafting of the manuscript. AS obtained the funding, conceived the study, and edited the manuscript. All of the authors have read and approved the final manuscript. Acknowledgements ================ We wish to acknowledge the excellent technical assistance of Tarja Heiskanen, Kaisa Jalkanen, and Heini Flinck. Susanna Lukinmaa is acknowledged for advising with PFGE and Taru Kauko with MLVA. Elisa Huovinen and Sari Jaakola are acknowledged for collecting of the travel information. This work was supported by a grant (4850/501/2004) from the Finnish Ministry of Agriculture and Forestry.

PubMed Central

2024-06-05T04:04:19.474278

2011-2-25

PMC3053225

Background ========== Recombinant antibodies and antibody fragments are indispensable tools for research, diagnostics and therapy \[[@B1]-[@B5]\]. Complete and natively glycolsylated antibodies, like IgGs, needed for therapeutic purposes must thus far be produced in mammalian cells. However, although yields from mammalian cells tend to be good, production times and costs are high \[[@B6]\]. However, for many non-therapeutic applications, and also some therapeutic applications for which effector functions are not necessary, antibody modules, such as single chain fragment variable (scFv) and fragment antigen binding (Fab) are sufficient \[[@B7],[@B8]\]. Because of their lower costs and faster production cycles, microbial systems are more attractive than mammalian cell systems for the production of antibody fragments. The folding and export of scFvs in Gram-negative systems is usually more efficient than that of Fab fragments \[[@B9]\]. One problem with prokaryotic production systems is that expression levels of fragments of different antibodies tend to differ markedly \[[@B10],[@B11]\] and, in many instances, only synthetic, codon-usage adapted, genes provide significant yields. The availability of a core suite of distinct efficient host-broad host range expression cloning vector systems, differing in their expression specificities, should enable determination of optimal production systems for different proteins. *Pseudomonas putida*strain KT2440 is a metabolically versatile soil bacterium with considerable potential in a broad range of diverse industrial and environmental applications \[[@B12]\]. Its certification as a biosafety strain \[[@B13],[@B14]\], its ability to express a broad spectrum of foreign proteins at high levels and the availability of powerful customized tools for genetic analysis and manipulation \[[@B15]\], make KT2440 an important prokaryotic cell factory. These features suggest that it might be a useful production system for antibody fragments. In this study, we have assessed the potential of KT2440 for the soluble production of different recombinant scFvs namely, the model murine anti-hen egg-white lysozyme scFv, D1.3 \[[@B11],[@B16]-[@B19]\], and two phage display-selected human scFvs: TOB5-D4 \[\[[@B11]\], Al-Halabi *et al.*in preparation\], directed against C-reactive protein (CRP) \[[@B20]\], an inflammation indicator in human blood, and HT186-D11 \[[@B21]\], directed against mucin1 (MUC1), a diagnostic marker and potential therapeutic target of cancer \[[@B22],[@B23]\]. Results and Discussion ====================== Expression plasmids and synthetic genes --------------------------------------- The key features of the antibody expression plasmids constructed in this study are shown in Table [1](#T1){ref-type="table"}. They were generated using new synthetic RK2 broad host range plasmid-based chassis developed by the group of Victor de Lorenzo (in preparation) carrying either the inducible *Ptac*\[[@B13],[@B24]\] or TOL plasmid *xyl*operon *Pm*\[[@B25],[@B26]\] promoters. We further modified the chassis by equipping it with the G10L ribosome binding site (RBS) with epsilon enhancer \[[@B27]\] and the *Erwinia carotovora pelB*leader sequence \[[@B28]\] to effect the export of expressed polypeptides to the periplasm. The scFv format consists of variable regions of the antibody heavy and light chains (VH and VL, respectively), joined by a 15-25 amino acid linker \[[@B29]\]. Polynucleotide sequences encoding Myc-HIS~6~-Strep-Tag^®^II or HIS~6~-Myc-Strep-Tag^®^II affinity tags were placed downstream of the VH-linker-VL coding sequences (Table [1](#T1){ref-type="table"}) to facilitate purification and detection of the scFvs in western blots and enzyme-linked immunosorbent assays (ELISA). In addition, synthetic scFv genes were generated to assess the influence of codon usage adaptation on expression levels in *P. putida*(Table [1](#T1){ref-type="table"}). The original scFv gene constructs were expressed from the *Ptac*promoter, whereas the synthetic gene constructs were expressed from either *Ptac*or *Pm*promoters (Table [1](#T1){ref-type="table"}). Although synthetic codon usage adapted gene constructs of TOB5-D4 (anti-CRP) and HT186-D11 (anti-mucin1) scFvs were readily generated, for unknown reasons the equivalent version of D1.3 (anti-lysozyme) scFv could not be synthezised by the commercial supplier. For generation of the native sequence constructs, the RBS and Strep-tag^®^II were added using primer overhangs. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Properties of the RK2-based broad-host range plasmid constructs with the *Ptac/lacIq*(pSEVAlac) or *Pm/xylS*(pSEVAxyl) promoter/regulator gene systems used in this study. ::: Construct scFv specificity Promoter Affinity-tags codon optimization^1^ \% GC^2^ -------------------- ------------------------ ---------- ------------------------- -------------------------- ---------- pSEVAlacD1.3n hen egg-white lysozyme *Ptac* HIS6/Myc/Strep-tag^®^II \_ 52,3 pSEVAlacTOB5-D4n human CRP *Ptac* Myc/HIS6/Strep-tag^®^II \_ 55,0 pSEVAlacTOB5-D4s human CRP *Ptac* Myc/HIS6/Strep-tag^®^II Jcat/Eurofins MWG Operon 68,3 pSEVAxylTOB5-D4s human CRP *Pm* Myc/HIS6/Strep-tag^®^II Jcat/Eurofins MWG Operon 68,3 pSEVAlacHT186-D11n human mucin1 *Ptac* HIS6/Myc/Strep-tag^®^II \_ 57,8 pSEVAlacHT186-D11s human mucin1 *Ptac* HIS6/Myc/Strep-tag^®^II GeneOptimizer^®^/GENEART 67,6 pSEVAxylHT186-D11s human mucin1 *Pm* HIS6/Myc/Strep-tag^®^II GeneOptimizer^®^/GENEART 67,6 ^1^Synthetic scFv determinants with optimized codon usage: - indicates the original scFv sequence, Jcat/Eurofins MWG Operon and GeneOptimizer^®^/GENEART indicates a codon adaptation tool and a supplier; ^2^%GC content of the scFv coding sequence. ::: Affinity isolation of scFvs --------------------------- Different combinations of cell lysis and affinity purification strategies were tested in order to determine a rapid, simple isolation/purification protocol. Purification of hexahistidine-tagged scFvs from clarified whole cell extracts by immobilized metal ion affinity chromatography (IMAC) resulted in a protein mixture containing three major species between 30-50 kDa, of which the mid band was confirmed by N-terminal sequencing to be the scFv. The addition of a second affinity chromatography on Strep-Tactin Superflow in series (tandem) resulted in scFv fragments of high purity with no visible background in sensitively Coomassie Blue-stained SDS-PAGE (data not shown). Single step purification on Strep-Tactin Superflow resin in gravity flow columns, however, proved sufficient for most purposes, including yield determination. This single-step protocol yielded scFv fragments with a purity of \~95%. Incubation of the purified fragments for 16-24 h on ice, followed by at least one freeze thaw cycle in elution buffer, did not cause significant degradation of the scFvs (Figure [1](#F1){ref-type="fig"}). Lysis of cells by sonication or use of the bacterial protein extraction reagent (B-Per, Thermo fisher scientific), or both in combination, gave similar results, which is perhaps not surprising, given that the scFvs are targeted to the periplasmic space. Moreover, purification of selectively released periplasmic proteins (see below) by single step Strep-Tactin Superflow chromatography resulted in a purity comparable to those obtained by the tandem strategy. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Purity of single-step affinity-purified scFvs from KT2440**. Proteins were isolated, affinity purified, incubated and subjected to freeze:thaw cycles to assess their stability in elution buffer, as described above and run on a 12.5% SDS polyacrylamide gel. M = PAGE Ruler unstained protein molecular weight marker (Fermentas); 1, D1.3; 2, TOB5-D4 (anti-CRP), 3, HT186-D11 (anti-MUC1). :::  ::: scFv yield optimization ----------------------- ### Induction Two transcriptional expression systems based on the *Ptac*and *Pm*promoters, that are widely used to express heterologous genes, were assessed for their efficacy in scFv production by KT2440 in otherwise identical RK2 plasmid-based expression constructs (Table [1](#T1){ref-type="table"}). In the case of the TOB5-D4s and HT186-D11s scFvs, yields from the *Ptac*promoter constructs were more than twice those from the *Pm*promoter constructs. Optimal yield conditions determined for *Ptac*constructs were 3 h induction by 1 mM IPTG at 30°C in LB-medium, which gave yields of 1.5 mg/l for the D1.3 scFv, 2.9 mg/l for the anti-CRP scFv, and 3.6 mg/l for the HT186-D11 scFv (Figure [2](#F2){ref-type="fig"}; yield means of 3-6 individual expression experiments involving 100 ml cultures agitated at 180 rpm, and isolation by single step affinity purification on 1 ml Strep-Tactin Superflow gravity columns). The influence of medium composition on yield was also assessed and it was found that yields decreased in the order: Terrific Broth (TB) \> LB \> M9 (15 mM succinate) \> R2A (Figure [3](#F3){ref-type="fig"}), indicating that enriched, buffered media favour scFv production and that medium optimization has considerable potential for yield enhancement. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Yields of soluble scFvs from KT2440 constructs**. Mean values and standard deviations are given of yields from single-step purified soluble scFvs obtained from 3-6 independent expression and purification experiments. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Effect of different culture media on HT186-D11 production**. SDS-PAGE of single step purified scFvs from periplasmic extractions of *P. putida*KT2440 cultures expressing scFvs from construct HT186-D11n on different media. :::  ::: ### Codon usage optimization KT2440 is an expression host with a high average GC content (\~61.5%) and a codon usage preference distinct from those of the scFv determinants used in this study \[[@B14]\]. To assess to which extent adaptation of the scFv codons to those preferred by KT2440 could influence scFv yields, synthetic scFv determinants composed of the preferred codons were designed *in silico*, synthesized and tested in the expression systems. Although synthetic versions of the TOB5-D4 (anti-CRP) and HT186-D11 (anti-MUC1) scFv sequences were made, it was not possible to obtain a synthetic version of the D1.3 (anti-lysozyme) scFv (Table [1](#T1){ref-type="table"}). For the two former scFvs, it was found that the yields - 2.4 mg/l for TOB5-D4s (anti-CRP) and 2.1 mg/l for HT186-D11s (anti-MUC1) (Figure [2](#F2){ref-type="fig"}) - were lower than those from the original constructs, so in these cases, codon usage optimization did not increase expression levels. This might reflect the ability of KT2440 to efficiently express a broad range of foreign genes, or limitation of higher levels of expression otherwise attainable due to the higher than normal (for KT2440) GC contents of the synthetic genes (Table [1](#T1){ref-type="table"}), or sub-optimal mRNA secondary structures created by the changes \[[@B30]\]. The key question arising from these results is how the scFv yields obtained compare with those obtainable from existing expression systems. Unfortunately, pertinent information could not be found in the literature, and values that are available are not usable due to differences in antibody and/or experimental procedures followed. To obtain relevant information on this issue, we introduced the broad range constructs into the *Escherichia coli*expression strain BL21 (DE3). Yields of the TOB5-D4n and HT186-D11n scFvs from *E. coli*were 2.5 to 4-fold lower than those from KT2440, and, as might be predicted, even lower in the case of the high GC synthetic genes (data not shown). Literature yields in *E. coli*for an anti-CRP scFv antibody of 0.55 mg/l \[[@B31]\] and the D1.3 scFv antibody of 0.29 mg/l \[[@B10]\] lie within the same low range. Periplasmic export ------------------ All genetic constructs were designed to create precise amino-terminal translational fusions of the scFv-coding sequences with the *E. carotovora pelB*signal sequence to target the scFvs for secretion by the Sec pathway. For assessment of the efficiency of processing of the scFv fragments produced in KT2440, and their translocation to the periplasm, periplasmic fractions were isolated. This procedure generally gave scFv yields in the same range as, though with greater variability than, those obtained by whole cell lysis. To ascertain whether the increased variation in yield is due to incomplete release of periplasmic proteins during the isolation procedure or incomplete translocation to the periplasm, perhaps resulting from overloading of the secretion apparatus, Sec-mediated proteolytic cleavage of the *pelB*leader sequences of the scFvs was analyzed by electrophoretic separation followed by western blot detection of the C-terminal Strep-tag^®^II (Figure [4A](#F4){ref-type="fig"}). As expected, only fully processed scFvs were detected in periplasmic preparations, indicating that periplasmic scFvs are completely processed. In the case of proteins extracted from whole cells, only fully processed mature anti-CRPs were detectable, although small amounts of unprocessed HT186-D11n and HT186-D11s scFvs were identified (Figure [4A](#F4){ref-type="fig"}, compare Figure [1](#F1){ref-type="fig"}), indicating that at the moment of sampling small amounts of untransported protein were still in the cytoplasm. The ratio of processed:unprocessed polypeptide did not seem to correlate with expression levels, since unprocessed protein was also observed in cells carrying the lower expressing HT186-D11s construct. More likely is the possibility that the VH-coding amino acid sequence affects the secondary structure around the cleavage junction and negatively impacts on the efficiency of processing. The variable yields in periplasmic extracts probably reflect incomplete disruption of the outer membrane by the procedure we used, but might also result from partial non-specific leakage of periplasmic proteins to the medium, a phenomenon previously observed for antibody fragments exported by means of the PelB leader peptide in *E. coli*\[[@B32]\]. It should, nevertheless, be emphasized that this procedure for selective isolation of periplasmic proteins has the major advantage of yielding active, soluble antibody fragments uncontaminated by cytoplasmic proteins. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Periplasmic export of scFvs and signal peptide cleavage**. A.) Western-blot of purified antibodies extracted from whole cells (WCE) and periplasmic preparations (PPP) and chromogenic detection of the Strep-tag^®^II at the C-terminus. A minor fraction of incomplete N-terminal processing (cleavage of the 22 amino acid *pelB*-leader sequence) is visible only in the case of HT186-D11 whole cell extracts (arrows). B.) N-terminal sequencing by Edman degradation confirmed the predicted cleavage sites following the A-M-A motif. X, amino acid position with ambiguous identification. :::  ::: The signal peptide cleavage sites of the recombinant scFvs were predicted by means of the *in silico*tool Predisi \[[@B33]\] to be located between amino acids 22 and 23, directly after the A-M-A motif, which is also the preferred sequence motif (A-X-A) for interaction with signal peptidase I \[[@B34]\], and experimentally confirmed by N-terminal sequencing of isolated antibody (Figure [4B](#F4){ref-type="fig"}). Other sites of cleavage were not detected, which indicates that the KT2440 LepB peptidase (PP\_1432), which shares 39% amino acid identity and 52% similarity with the *E. coli*K-12 peptidase, precisely recognizes and cleaves the PelB processing site. Antibody activity ----------------- In order to assess whether the extracted periplasmic scFv polypeptides were correctly folded and had acquired correct binding activity, the binding of the TOB5-D4 and HT186-D11 antibodies to their cognate antigens were measured by ELISA assays using plates coated with 100 ng per well of either CRP antigen (BiosPacific, Emeryville, USA), MUC1 32 aa peptide with a C-terminal cysteine (APDTRPAPGSTAPPAHGVTSAPDTRPAPGSTA-C)\[[@B21]\], or bovine serum albumin (BSA; controls). A primary mice anti-Myc-tag IgG, combined with a secondary Fab-specific goat anti-mice IgG horseradish peroxidase (HRP) conjugate were used for detection of bound antibody, as previously described \[[@B35]\]. As can be seen in Figure [5](#F5){ref-type="fig"}, high specific antigen binding was observed for both recombinant scFvs, with significant signals recorded till antibody concentrations of 10 nM, which corresponds to dilutions of 1:256, for an antibody having an initial concentration of 2.5 μM. Recombinant anti-CRP antibody preparations obtained from both whole cells and periplasmic extracts had similar binding activities (data not shown), which is consistent with the findings shown in Figure [4A](#F4){ref-type="fig"} and the correct folding of periplasmic antibody. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Specific binding activities of the scFvs produced in KT2440**. The binding activities of the scFvs to cognate antigens was measured by ELISA. The scFv concentration detection limits were \~10 nM. Absorbance was measured at 450 nm and corrected for absorbance at 620 nm. :::  ::: Conclusions =========== We have presented here an assessment of the potential of *Pseudomonas putida*KT2440 as a cell factory for the production of soluble recombinant antibody fragments that bind to antigens of high interest for diagnostics and therapy. With the described construct design, good yields of soluble, active scFvs are obtained through simple extraction of periplasmic polypeptides and single step affinity purification. These yields considerably exceeded those obtained from equivalent constructs in *E. coli*K-12, so KT2440 would seem to be a promising cell factory for recombinant antibody fragment production. It is likely that high density fed-batch cultivation of KT2440 \[[@B36]\], which are typically characterized by cell densities of up to more than 100 g/l, will allow achievement of much higher yields by this host. A common means of optimizing expression levels of recombinant proteins is the use of synthetic genes with host-optimized codon usage. In our study, this approach did not achieve increased expression, possibly due to the expression versatility of this host for foreign genes. However, the possibility that a less rigorous adaptation that maintains the *P. putida*KT2440 natural % G+C value might give better expression would seem worth exploring in future, especially in conjunction with other optimization efforts involving exploration of media and cultivation conditions as well as genetic engineering and genome streamlining approaches. The broad host range feature of the constructs presented here will of course allow their testing in other host systems and potential discovery of even more powerful cell factories for antibody production. Methods ======= Vector construction ------------------- Expression construct pSEVAlacTOB5-D4n was generated by amplification of the scFv sequence from antibody expression vector pOE101-TOB5-D4 \[\[[@B11]\], Al-Halabi *et al.*in preparation\], using primers A (fw 5\'-GC[GAATTC]{.underline}TTAACTTTAAGAAGGAGATATATCC ATG A AATACCTATTGCCTACGGC-3\') and B (rev 5\'-GC[TCTAGA]{.underline} TTACTTTTTCGAACTGCGGGTGGCTCCA ATGATGATGGTGATGATGGGATAGATCTTC-3\') and cloning via *Eco*RI and *Xba*I to pSEVA-RK2-Sm-lac (pSEVA424). Vector pSEVAlacHT186-D11n was generated amplifying HT186-D11 from phage display vector pHAL14-HT186-D11 \[[@B11],[@B21]\] using primers A and C (5\'-GC[TCTAGA]{.underline} TTACTTTTCGAACTGCGGGTGCGACCATGCGGCCCCATTCAGATCC TCT TCTGAGATG-3\'), and cloning via EcoRI and XbaI into pSEVA-RK2-Sm-lac. Plasmid pSEVAlacD1.3n was constructed using the same primers and restriction sites for the amplification product of pHAL14-D1.3 \[[@B11]\]. Sequences for ribosome binding sites \[[@B27]\], restriction sites and the Strep-tag^®^II \[[@B37]\] were added with the corresponding primer overlaps. Vector pSEVAlacTOB5-D4s and pSEVAxylTOB5-D4s were generated using an *in silico*-designed synthetic construct incorporating the RBS and Strep-tag^®^II. The sequence was codon-usage adapted to that of *P. putida*KT2440, using Jcat \[[@B38]\], and synthesized by Eurofins MWG (Ebersberg, Germany). The construct was cloned into pSEVA-RK2-Sm-lac and pSEVA-RK2-Sm-XylPm via *MfeI*/*Eco*RI and *XbaI*. A synthetic construct was also generated from HT186-D11, codon usage adapted to *P. putida*, and synthesized by GENEART (Regensburg, Germany). Cloning in pSEVA-RK2-Sm-lac and pSEVA-RK2-Sm-XylPm *via*EcoRI/HindIII resulted in pSEVAlacHT186-D11s and pSEVAxylHT186-D11s respectively. All constructs contained the 66 bp *Erwinia carotovora pelB*-leader sequence for periplasmic export and are summarized in Table [1](#T1){ref-type="table"}. *E. coli*strain DH5α, chemicals and enzymes for PCR and cloning were purchased from Fermentas (St. Leon-Rot, Germany) and New England Biolabs (Ipswich, MA, USA). Integrity of the constructs was verified by sequencing the synthetic construct prior to cloning, and restriction digestion after cloning. Native sequences were amplified from sequence validated templates \[[@B11],[@B21]\], using high fidelity proof-reading polymerase (Phusion™) and confirmed by restriction digestion after cloning. Transformation of *Pseudomonas putida*KT2440 -------------------------------------------- Competent cells were prepared using buffer containing 300 mM sucrose as described before \[[@B39]\]. 30 ng of plasmid constructs were used to transform 40 μl of competent *P. putida*KT2440 (DSM 6125) cells by electroporation, which was carried out in prechilled 2 mm cuvettes using a Gene Pulser II with pulse controller plus and capacitance extender plus (Bio-Rad, Hempel Hempstead, UK). Cell:DNA mixes were pulsed at 2.5 kV, 25 μF and 200-500Ω resistance, and subsequently plated on selection medium containing 100 μg/ml of streptomycin and spectinomycin. Production of scFvs with *Pseudomonas putida*KT2440 --------------------------------------------------- An 50 ml overnight liquid culture of the *P. putida*KT2440 clone freshly transformed with the respective construct (Table [1](#T1){ref-type="table"}) was used to inoculate the production culture 1:100. Production test-cultures were grown in 200 ml cultures at 150 rpm at RT, 30°C and 37°C. Luria Bertani (LB) (5 g/l yeast extract, 10 g/l tryptone, 10 g/l NaCl), rich Luria Bertani Broth (10 g/l yeast extract, 10 g/l tryptone, 10 g/l NaCl), M9 containing 15 mM succinate, TB, and R2A \[[@B40],[@B41]\] containing 50 μg/ml streptomycin were used in 500 ml baffled Erlenmeyer flasks. In the mid logarithmic phase (at OD~595~of \~0.5-0.6), production was induced by addition of isopropyl-ß-D-thiogalactopyranoside (IPTG) at concentrations ranging between 100 μM and 1 mM, or 15 mM 3-methylbenzoate, for pSEVAxyl-TOB5-D4s and pSEVAxyl-HT186-D11s. Final yields reported in Figure [2](#F2){ref-type="fig"} are from lysis of cells cultured in 100 ml LB at 180 rpm and 30°C. The cells were harvested by centrifugation. Production of scFvs in *E. coli* -------------------------------- To compare yields of recombinant antibody fragments from *P. putida*KT2440 with those from *E. coli*K-12, we introduced the pSEVAlacTOB5-D4 and pSEVAlacHT186-D11 constructs containing the *tac*promoter for expression into *E. coli*strain BL21 (λDE3) (Stratagene). K-12 and KT2440 constructs were cultured at RT for HT186-D11n and 30°C for TOB5-D4n with other parameters being as described for KT2440, which in initial experiments also gave the best antibody yields in *E. coli.* Periplasmic export and signal peptide cleavage ---------------------------------------------- To obtain periplasmic extracts, harvested cells were resuspended in PE buffer (20% (w/v) sucrose, 50 mM Tris, 1 mM EDTA, pH 8), incubated for 30 min on ice with brief vortexing every 5 minutes, and centrifuged at 20,000 rcf for 30 min. Completeness of signal peptide cleavage was assessed by electrophoresis of antibody fragments on 12.5% SDS-PAGE, semi-dry transfer to PVDF membranes (Peqlab), and anti-Strep-tag detection using Strep-Tactin alkaline phosphatase (AP) conjugate (IBA, Göttingen, Germany) and chromogenic BCIP (5-Bromo-4-Chloro-3\'-Indolyphosphate p-Toluidine Salt), NBT (Nitro-Blue Tetrazolium Chloride) detection (AP Blue Membrane Substrate Solution, Sigma). The cleavage position was determined by N-terminal sequencing by Edman degradation and compared to the *in silico*predicted sites obtained with the Predisi tool \[[@B33]\]. Affinity purification of antibody fragments ------------------------------------------- The proteins were obtained by whole cell lysis by sonication in sonication buffer (50 mM Tris, 100 mM NaCl, 5 mM MgCl~2~, 0,05% (v/v) Triton X-100, pH 8,0), bacterial protein extraction reagent B-Per (Thermo Fischer scientific) according to the manufacturer\'s instructions, sonication in B-Per or by periplasmic extraction (see below). Whole cell extracts were centrifuged for 15 min at 15,000 rcf to obtain cell-free extracts which, like periplasmic preparations were applied directly to affinity chromatography columns. We used IMAC on Ni-TED (Machery-Nagel), for HIS-tag-based affinity purification, or Strep-Tactin Superflow (IBA) columns for Strep-tag^®^II-based affinity purification, or both in sequence. The washing and elution procedures were carried out according to the manufacturers\' recommendations. Protein concentrations of purified scFv solutions were determined by the method by Gill and van Hippel \[[@B42]\], using the individually calculated molar extinction coefficients for the processed amino acid sequences without signal peptides (Protein Calculator v3.3, <http://www.scripps.edu/~cdputnam/protcalc.html>), and the UV absorbance at 280 nm (protein) and 260 nm (as control for possible nucleic acid contamination), determined with the NanoDrop^®^ND-1000 spectrophotometer (Peqlab) or lab spectrometer (Eppendorf). Antigen binding ELISA --------------------- The antigens, in PBS, were used to coat wells of 96-well microtitre plates (Maxisorp, Nunc). After coating overnight at 4°C, the wells were washed three times with PBST, and blocked with 2% (w/v) skim milk powder in PBST (2% M-PBST) for 1.5 h at room temperature, followed by three washes with PBST. For the ELISA, soluble scFvs were diluted in 100 μl 2% M-PBST and incubated in the antigen-coated plates for 1.5 h at room temperature, followed by three PBST washes. Bound scFvs were detected with the murine mAb (9E10, Sigma), which recognises the C-terminal c-myc tag, and a goat anti-mouse serum, conjugated with horseradish peroxidase (HRP) (Sigma; 1:10,000). For visualisation the chromogenic substrate 3,3\',5,5\'-tetramethylbenzidine (TMB) was added. The staining reaction was stopped by addition of 100 μl 1 N sulphuric acid and absorbance at 450 nm and 620 nm was measured using a SUNRISE™ microtitre plate reader (Tecan, Crailsheim, Germany). Abbreviations used ================== aa: amino acid; AP: alkaline phosphatase; RBS: ribosome binding site; CRP: human C-reactive protein; EDTA: ethylenediaminetetraacetic acid; ELISA: enzyme-linked immunosorbent assay; Fab: fragment antigen binding; HRP: horseradish peroxidase; IgG: immunglobulin G; IMAC: immobilized metal ion affinity chromatography; IPTG: isopropyl-β-D-thiogalactopyranoside; MUC1: human mucin1; PBS(T): phosphate-buffered saline (tween); PVDF: polyvinylidene fluoride, RPM: rotations per minute; rcf, relative centrifugal force; scFv: single chain fragment variable; VH: antibody variable domain of the heavy chain; VL: antibody variable domain of the light chain. Competing interests =================== MH and SD are inventors on a patent application regarding anti-MUC1 antibodies (PCT/EP2009/005218). This does not alter our adherence to all the Microbial Cell Factories policies on sharing data and materials. The other authors declare that they have no competing interests. Authors\' contributions ======================= TD designed and coordinated the study, drafted the manuscript, performed the experiments with lab support of SCK and analyzed the data. MS conducted the antigen ELISA, MH and SD provided the scFv sequences, MH, SD and KNT helped to draft the manuscript and to design the study. All authors read and approved the final manuscript. Acknowledgements ================ We are deeply indebted to Prof. V. de Lorenzo (CNB-CSIC Madrid) for providing prototypes of RK2 based pSEVA plasmids. Thanks are also due to R. Getzlaff (HZI) for conducting N-terminal protein sequencing and B. Jung (HZI) for technical assistance. We greatly acknowledge the financial support by the Federal Ministry of Education and Research to KNT (BMBF, \"Psysmo\" in the EraNet) and by the German Research Foundation (DFG, SFB 578) to SD.

PubMed Central

2024-06-05T04:04:19.477557

2011-2-21

PMC3053226

Background ========== Substance use among college and university students remains an important area of research due to the implications of early substance dependence on the future of the youth. Prior studies from various settings indicate relatively high rates of alcohol and other substance use among high school students and those in higher educational institutions \[[@B1]-[@B7]\]. Among the few studies from universities and colleges in Kenya, Odek-Ogunde et al reported high rates of substance use among students at a Kenyan private university, with rates as high as 84% for alcohol use and 54.7% for tobacco use \[[@B5]\]. A majority of the studies from this region focus on primary and secondary schools, and little information is available from college and university populations \[[@B2]-[@B4],[@B8]\]. Most of these studies still find high rates of substance use among school-children, suggesting that the rates would continue to rise among students in institutions of higher learning. For instance, Kuria \[[@B2]\] found alcohol use prevalence rates of upto 15% among secondary school students, while Kwamanga et al \[[@B3]\] found lifetime cigarette smoking rates of 32%. In a study among primary school pupils in Kenya, Ogwell et al \[[@B8]\] found a lifetime cigarette smoking rate of 31%. Among the very few similar studies from African countries other than Kenya, a South African study found an alcohol use prevalence rate of 39.1% and a cigarette use prevalence rate of 10.6% among high school adolescents \[[@B4]\]. Other drugs that are commonly used in these settings include cannabis, inhalants, tranquilisers, heroin and cocaine, among others \[[@B2],[@B5]\]. The importance of these findings lies in the fact that early exposure to substance use often predicts future substance use and other psychiatric disorders \[[@B9],[@B10]\]. Rohde and colleagues demonstrated that adolescent substance use disorder is associated with numerous functioning difficulties at age 30, some of which appear to be related to recurrent substance use disorder, co-morbid adolescent disorders, or functioning problems already evident in adolescence \[[@B11]\]. Some of the most studied risk factors for substance use among adolescents and young adults include low grade point average, lack of religiosity, early alcohol use, low self-esteem, psychopathology, poor relationship with parents, lack of social conformity (deviance), sensation seeking, perceived peer drug use, and perceived adult drug use \[[@B12]\]. No systematic study has been attempted to validate these factors in this region, and these factors are often cited as having universal application. The paucity of data on substance use among college and university students in Kenya and other low income countries has serious implications on the success of any interventions aimed at reducing this problem. The importance of documenting this problem cannot therefore be overstated. In this study, the main objective was to determine the prevalence and factors associated with substance use among students in tertiary institutions in Eldoret Municipality in Kenya, as well as to document any effects attributed to the use of various substances. Methods ======= Site ---- The study was carried out at four tertiary institutions of learning within Eldoret Municipality in the then Uasin Gishu District, Kenya. They consisted of one privately-run diploma level college, a government-run Technical Training Institute, a government-run polytechnic and the Law campus of Moi University. At the time of the study, the Education office in the municipality had 12 registered tertiary institutions of learning, including 4 campuses of Moi University, the government-run polytechnic, the government-run technical training institute and 6 private colleges offering various certificate, diploma and higher diploma courses. Information about the student populations in these institutions is not available in the public domain, so it was not possible to compare them for various socio-demographic variables before selection for inclusion into the study. Four of these institutions with a total student population of 11,304 were selected through simple random sampling for participation in this study. Using Fisher\'s formula with an expected substance use prevalence rate of 84% (based on alcohol use prevalence rate from the study by Odek-Ogunde et al) and a confidence interval of 95%, a minimum sample size of 206 was calculated. However, we decided to recruit a higher total sample of 500 students in order to be able to make meaningful analyses, and a sample of 125 students per institution was allocated for selection through stratified sampling. Participants ------------ Five hundred students were selected from the four institutions through a process of stratified sampling, with the sample allocation being proportionate to the number of students in each class in an institution. Within each class, simple random sampling was used to select the participants based on the student registration numbers obtained from the institutions\' heads. Design ------ The study used a cross-sectional descriptive survey design involving the administration of the World Health Organisation self-administered Model Core Questionnaire to collect information on the use of various drugs including alcohol, tobacco, stimulants, marijuana, cocaine and heroin, among others \[[@B13]\]. The WHO model core questionnaire has been standardized and used elsewhere in Africa to collect data on substance use among the youth \[[@B14],[@B15]\]. An additional questionnaire was used to gather data on other aspects of drug use. It contained three questions: 1\. Who introduced you to drug use? Possible responses for this question were: a) A friend; b) A parent; c) A sibling; d) Other relative; e) Other (specify). 2\. Which reason(s) explain why you use the substance you use? Possible responses for this question were: a) To relax; b) To relieve stress; c) To be accepted by peers (peer pressure); d) Excess pocket money; e) Desire to experiment; f) They are easily available; g) To enhance academic performance; g) To cope with problems; h) Other (specify). 3\. What problems have you had that you attribute to your use of the substance(s)? Possible responses for this question were: a) Quarrel or argument; b) Scuffle or fight; c) Accident or injury; d) Loss of money or other valuable items; e) Damage to objects or clothing; f) Problems in your relationships with your parents; g) Problems in your relationship with your friends; h) Problems in your relationship with teachers; i) Performed poorly at school or work; j) Victimized by robbery or theft; k) Trouble with police; l) Hospitalised or admitted to an emergency room; m) Engaged in sex you regretted the next day; n) Engaged in unprotected sex; o) Blackouts or flashbacks; p) Medical problems e.g. memory loss, hepatitis, head injury, bleeding etc; q) Other (specify). Both questionnaires were administered in the English language, given that this is the main language of instruction at secondary and tertiary educational institutions in Kenya. Lifetime substance use in this study refers to respondents who admitted to having ever used at least one of the substances listed in the questionnaire. Procedure --------- The study was carried out in the period between May and September 2009. Ethical review was conducted by the Institutional Research and Ethics Committee of Moi University and Moi Teaching and Referral Hospital, and administrative approvals were sought and granted by the Uasin Gishu District Education Officer and the heads of the respective institutions. Questionnaires were given to the selected participants by members of the research team and collected on the same day after being completed. No names or identifying information were indicated on the questionnaires, and all participants were assured of absolute confidentiality. Data storage and analysis ------------------------- Collected data was entered into a Microsoft Excel database and analysed using SPSS version 12.0. Descriptive statistics were used to compute means and standard deviations for numerical variables as well as frequencies for nominal and ordinal variables. The Chi square test statistic (χ^2^) was used to test the significance of association between various factors and the substance use measures, while the t-test was used to compare means. A value of *p*\< 0.05 was considered statistically significant. Results ======= Socio-demographic characteristics --------------------------------- Among the respondents, 261 (52.2%) were male and the mean age was 22.9 years (18-32, s.d. 2.5). Only 13.1% were married, and majority (59.7%) were staying at the institutional hostels. A minority of the respondents reported being employed, with jobs identified including business, accountancy, hairdressing and teaching. A majority of them (56.5%) depended on a government loan programme for higher education. Table [1](#T1){ref-type="table"} shows the socio-demographic characteristics of the study population. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Socio-Demographic Variables by Gender ::: VARIABLE MALE (%) FEMALE (%) TOTAL (%) ----------------------- ------------ ------------ ------------ **Housing type** Hostel 131 (51.4) 159 (68.8) 290 (59.7) Non-hostel 124 (48.6) 72 (31.2) 196 (40.3) **Working** Yes 226 (90) 210 (91.7) 436 (90.8) No 25 (10) 19 (8.3) 44 (9.2) **Residence setting** Rural 76 (29.5) 52 (22.1) 128 (26) Urban 182 (70.5) 183 (77.9) 365 (74) **Marital status** Married 43 (16.9) 21 (9) 64 (13.1) Single 212 (83.1) 212 (91) 424 (86.9) **Government Loan** Yes 151 (59.7) 122 (53) 273 (56.5) No 102 (40.3) 108 (47) 210 (43.5) ::: Lifetime substance use ---------------------- Among the participants, 349 (69.8%) reported having used at least one substance in their lifetime. Table [2](#T2){ref-type="table"} shows the association of lifetime substance use with selected socio-demographic variables. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Lifetime Substance use Variation with Socio-Demographic Variables ::: VARIABLE Lifetime substance use % **X**^**2**^ p value ----------------------- -------------------------- -------------- --------- **Gender** Male 72.8 2.327 0.144 Female 66.5 **Housing type** Hostel 68.3 0.744 0.421 Non-hostel 71.9 **Working** Yes 70.5 0.015 0.863 No 71.3 **Residence setting** Rural 70.3 0.001 1.000 Urban 70.1 **Marital status** Married 70.3 0.000 1.000 Single 70.3 **Government Loan** Yes 70.3 0.001 1.000 No 70.5 ::: There was no statistically significant association between lifetime prevalence of substance use and any of the selected socio-demographic variables. Specific Substance use ---------------------- Among the 478 participants who answered the questions on alcohol use, 248 (51.9%) reported lifetime alcohol use in this study. Of these, 97.6% had had a drink in the preceding week, translating to a current alcohol use prevalence rate of 50.7% in this sample. There was no statistically significant variation of alcohol use with gender or any of the other socio-demographic variables. Table [3](#T3){ref-type="table"} shows the relationship between alcohol use and these variables. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Lifetime Alcohol use Variation with Socio-Demographic Variables ::: VARIABLE Lifetime Alcohol use % **X**^**2**^ P value ----------------------- ------------------------ -------------- --------- **Gender** Male 53.6 0.609 0.464 Female 50.0 **Housing type** Hostel 51.1 0.012 0.925 Non-hostel 51.6 **Working** Yes 51.2 0.004 1.000 No 51.7 **Residence setting** Rural 51.6 0.000 1.000 Urban 51.6 **Marital status** Married 55.9 0.586 0.488 Single 50.6 **Government Loan** Yes 50.4 0.251 0.642 No 52.7 ::: The average age at which the respondents started using alcohol was 17.5 years (11-23, s.d. 2.0), with males starting at 17.8 years (11-22, s.d. 2.0) and females starting at 17.3 years (12-23, s.d. 2.0). There was no statistically significant difference between male and female mean age of onset of alcohol use (t = 1.909, p = 0.06). Over the month preceding the study, the mean number of days of alcohol use was 6.1 (1-20, s.d. 3.9). Respondents living in hostels drank on more days than those that were living outside college hostels (t = 3.058, p = 0.002). Those receiving government assistance also drank on more days than those not on government assistance programs (t = 3.516, p = 0.001). There was no statistically significant association between the mean number of drinking days per month and the other socio-demographic variables. The mean number of drinks per drinking day over the preceding month was 2.9 (0-9, s.d. 1.8). As shown in Table [4](#T4){ref-type="table"}, those living in hostels (t = 2.362, p = 0.019), those residing in urban areas (t = 3.203, p = 0.002) and those receiving government loan (t = 4.564, p = 0.000) reported statistically significantly higher rates than respondents without those attributes. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Variation of Drinking Pattern with Socio-Demographic Variables ::: VARIABLE Drinking days in the last month (Mean = 6.1 ± 3.9) Mean number of drinks per drinking day (Mean = 2.9 ± 1.8) ----------------------- ---------------------------------------------------- ----------------------------------------------------------- **Gender** t = 0.125, p = 0.901 t = -1.179, p = 0.239 Male 6.1 ± 3.7 2.8 ± 1.6 Female 6.1 ± 4.2 3.1 ± 2.1 **Housing type** t = 3.058, p = 0.002\* t = 2.362, p = 0.019\* Hostel 6.8 ± 4.0 3.2 ± 2.0 Non-hostel 5.3 ± 3.7 2.6 ± 1.6 **Working** t = -1.023, p = 0.307 t = -0.519, p = 0.604 Yes 7.0 ± 4.6 3.2 ± 1.4 No 6.1 ± 3.9 3.0 ± 1.9 **Residence setting** t = -1.921, p = 0.056 t = -3.203, p = 0.002\* Rural 5.3 ± 4.0 2.3 ± 1.3 Urban 6.4 ± 3.9 3.2 ± 1.9 **Marital status** t = 0.820, p = 0.413 t = -0.092, p = 0.927 Married 6.7 ± 4.2 2.9 ± 1.5 Single 6.1 ± 3.9 3.0 ± 1.9 **Government Loan** t = 3.516, p = 0.001\* t = 4.564, p = 0.000\* Yes 7.0 ± 4.0 3.4 ± 2.1 No 5.2 ± 3.7 2.4 ± 1.2 \*Statistically significant ::: Table [4](#T4){ref-type="table"} summarises the variation of drinking patterns with various socio-demographic variables. Among the respondents using alcohol, 50.4% admitted having taken 5 or more drinks per day on one or two days in the preceding month, while 9.2% had done so on 3 or more days. The rest reported taking less than five drinks on each drinking day. A total of 203 (42.8%) of the 474 participants who answered the question on smoking admitted to having smoked a cigarette at least once in their lifetime. The lifetime cigarette smoking prevalence rate was statistically significantly higher for the males than for the females (47.6% vs. 37.5%, X^2^= 4.922, p = 0.032). There was no statistically significant association between prevalence rate of cigarette smoking and any of the other variables. Table [5](#T5){ref-type="table"} shows the relationship between cigarette smoking and various variables. ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Lifetime Cigarette use ::: VARIABLE Lifetime Cigarette use % **X**^**2**^ P value ----------------------- -------------------------- -------------- --------- **Gender** Male 47.6 4.922 0.032\* Female 37.5 **Housing type** Hostel 43.1 0.018 0.925 Non-hostel 42.5 **Working** Yes 45.2 0.081 0.871 No 43.0 **Residence setting** Rural 40.5 0.418 0.530 Urban 43.8 **Marital status** Married 41.4 0.059 0.888 Single 43.1 **Government Loan** Yes 44.3 0.663 0.453 No 40.6 \*Statistically significant ::: The mean age at first cigarette smoking was 15.7 years (10-21, s.d 1.8). Females started smoking at the mean age of 16.0 years (s.d. 1.8) while males started at the age of 15.5 years (s.d. 1.7). This difference was statistically significant (t = -2.151, p = 0.033). When asked how many cigarettes a day they had smoked \'in the past 30 days or month\', 62.5% admitted to smoking \'less than one cigarette per day\', while 25.6% smoked 1-5 cigarettes per day. A small proportion (4.4%) smoked 6-15 cigarettes a day, while 7.5% smoked 16 or more cigarettes each day. Ten (2%) respondents admitted to using cannabis. Of these, 8 were male and 2 were female. Out of these, 6 admitted to having used cannabis in the past 12 months. Only 3 respondents admitted to ever having used cocaine. Who introduced respondent to substance use ------------------------------------------ Among the 349 participants who admitted to having ever used any of the above substances, 75.1% were introduced by a friend while 23.5% were introduced by a relative other than the nuclear family. A very small percentage of respondents in this sample (1.4%) were introduced to drug use by a member of the nuclear family. Reasons for substance use ------------------------- Reasons given for substance use were: To relax (62.2%), to relieve stress (60.8%), desire to experiment (41.9%), peer pressure (38.9%), and to cope with problems (38.9%). Negative effects attributed to substance use -------------------------------------------- Most of the negative effects attributed to substance use by the respondents were associated with alcohol use. Significantly, 55.2% of the participants using alcohol reported having experienced medical problems as a result of their alcohol use. Further, 60.5% engaged in unprotected sex and 62.5% engaged in sex they regretted the next day. Over 60% of the participants reported engaging in scuffles, loss and damage to property and quarrels. Table [6](#T6){ref-type="table"} shows the negative effects attributed to the use of alcohol. ::: {#T6 .table-wrap} Table 6 ::: {.caption} ###### Negative Effects Attributed to Alcohol Use ::: PROBLEM \% FREQUENCY (N = 248) ------------------------------------------- ------------------------ Quarrel/argument 82.7 Damage to objects 68.5 Problem in relationship with parents 68.1 Scuffle/Fight 68.1 Loss of money or other valuable item 68.1 Accidental injury 63.3 Engaged in sex you regretted the next day 62.5 Engaged in unprotected sex 60.5 Medical problems 55.2 Robbery or Theft 52.3 Problem in relationship with friends 52.0 Trouble with police 50.8 Poor performance at school 42.3 Problems with teachers 41.5 ::: Among cigarette smokers, the problems they attributed to cigarette smoking were in relationships with teachers (20.7%), parents (11.8%) and friends (6.9%). Some of the respondents attributed medical problems (8.4%) and poor school performance (4.4%) to cigarette smoking. None of the respondents attributed any problems to substance use other than alcohol or cigarette use. Discussion ========== This study found a lifetime substance use prevalence rate (69.8%) that is significantly higher than the 41% rate found among high school students in Kenya \[[@B2]\]. This implies that substance use rates, in general, increase with age and transition through the education system. This has major policy implications, including the need to focus substance use interventions on younger age-groups such as primary and high school students. Preventing early substance-related problems will reduce the risk of these problems in later adulthood when the magnitude of life stresses is greater. The prevalence rate of lifetime alcohol use in this study was 51.9%. This rate, though lower than what others have found in similar settings, is still significant. Odek-Ogunde et al \[[@B5]\] found rates of lifetime alcohol use as high as 84% among students at a private university in Nairobi, Kenya, while Othieno et al \[[@B16]\] in a more heterogeneous population consisting of outpatients attending several Nairobi primary health care facilities found a lifetime alcohol use prevalence rate of 62%. The paucity of literature on alcohol use among college students in Africa and other parts of the world was demonstrated by Karama et al \[[@B17]\] in a 2006 review article in which the authors argued for the need for more studies to be carried out in this area in order to inform interventions. However, studies among high school students in this region report similarly high prevalence rates of alcohol use, ranging from 15% to 57.9% \[[@B2],[@B4],[@B6]\]. Almost all (97.6%) of those who had ever used alcohol in this study could be regarded as current users (had an alcoholic drink in the week preceding the study), translating to a 50.7% current alcohol use prevalence rate. In contrast, a Kenyan general population study \[[@B18]\] found a current alcohol use prevalence rate of 13.4%. The huge difference implies that alcohol use is probably still regarded as being fashionable among college students and very few social sanctions exist to discourage this behaviour. Chances of moving from use to abuse and dependence are therefore heightened, especially considering other factors such as age of onset and frequency of use. Participants reported having used alcohol on an average of 6 days in the month preceding the study. The only factors significantly associated with this were housing type and government loan. Students residing in hostels drank on average 1.5 days more than those not residing in hostels, suggesting a significant role for peer influence in shaping drinking behaviour. Participants receiving government loan reported drinking almost 2 days more than those not receiving such loans, probably reflecting more readily available funds to sustain drinking behaviour. However, other factors may be responsible for this finding, and further research would be necessary to clarify this. Students in tertiary institutions in Kenya may apply for a Higher Education Loan from the government, and many of them are loaned various amounts of money depending on the course of study and the institution. The loan is recovered once the loanee completes their studies and secures employment or a steady source of income. The mean number of days the participants reported drinking suggests that they drink between one and two days each week, pointing towards a binge-drinking pattern rather than regular daily intake. In further support for the likelihood of a binge-drinking pattern of alcohol use, almost 60% of the participants in this study reported taking 5 or more drinks per day at least once in the preceding month. Previous research suggests that although it is difficult to generalise alcohol use patterns in many African countries, binge-drinking is a common phenomenon in the Kenyan general and even clinical population. For instance, a study by Saunders et al among patients at primary health care centres \[[@B19]\] reported that out of the six countries involved, the Kenyan sample had the highest rates of binge-drinking (episodic heavy drinking). The other countries in the study were the USA, Mexico, Australia, Norway and Bulgaria. A community based survey by Clausen et al \[[@B18]\] reported that 27.6% of the Kenyan current drinkers had had 5 or more drinks on at least one occasion in the preceding week, suggesting that binge-drinking was a common phenomenon in the Kenyan drinking population. This pattern of drinking would also explain the high prevalence of reported serious problems associated with alcohol use in our study, such as scuffles, loss and damage to property, medical problems and sexual indiscretions. An average of 3 drinks per drinking day was found among alcohol users in this study, with students living in hostels, in urban areas and receiving government loan taking more drinks on average than those without those attributes. These findings further illustrate the role of peer influence and urbanisation on alcohol use. The possibility of under-reporting cannot be discounted, given that the participants in this study were students who may be reluctant to reveal their drinking habits, even on an anonymised questionnaire. In the study by Saunders et al \[[@B19]\] which was carried out in a primary health care setting, the mean daily number of drinks taken by the \'drinking population\' was 9.6, while those classified as alcoholics were taking a mean of 23.4 drinks. Another study in a general population setting \[[@B18]\] had found a mean weekly drink rate of 9.9 drinks among Kenyan regular alcohol users, many of whom displayed a binge-drinking pattern of alcohol use. In this study, the mean age at first alcoholic drink was 17.5 years, with the youngest reported age being 11 years. There was no statistically significant difference in male and female age of onset of alcohol use. Otieno and Ofulla \[[@B6]\] similarly found the highest prevalence of alcohol use among young people aged 16-18 years. Several other studies have reported early age of onset of alcohol use among adolescents and the associated psychological problems in later life \[[@B9],[@B20],[@B21]\]. Despite the fact that the legal drinking age in Kenya has been set at 18 years for a long time, it is clear that alcohol is available for sale even to underage drinkers. As has been observed before \[[@B22]\], failure of enforcement may be part of the problem, as well as the domination of the alcohol market by small-scale often illicit producers and sellers. Partanen \[[@B22]\] suggests that 80-90% of total alcohol consumed in Kenya comes from \'small-scale production within the informal sector, licit or illicit, using traditional African methods of brewing and local skills of distilling\'. Among some of the problems associated with early age of onset of alcohol use is increased risk of later alcohol abuse and dependence as well as associated social and occupational difficulties \[[@B20],[@B21]\]. In one study, Grant and Dawson showed that the odds for dependence decreased by 14% while those for abuse decreased by 8% with each increasing year of onset of use \[[@B20]\]. The implication of the high rate of alcohol use and early age of onset in this study is that a large proportion of the respondents are at a high risk of developing alcohol-related disorders as adults. In the present study, negative effects attributed to alcohol use by the respondents included quarrels and fights, loss and damage to property, regretted sex, unprotected sex, and medical problems. As indicated earlier, most of these problems could be attributed to a binge-drinking pattern of behaviour, rather than regular light use of alcohol. The high risk sexual behaviour is particularly ominous due to the high prevalence of HIV and other sexually transmitted infections in Kenya. In 2007, Chersich et al \[[@B23]\] reported an association between binge-drinking and unsafe sex, sexual violence and sexually transmitted infections among Kenyan female sex workers, and the students in the present study are clearly exposed to the same risks. It is clear that unless substance use among adolescents and young adults is addressed, interventions targeting HIV/AIDS, violence and accidents will achieve less than optimal results. The lifetime prevalence rate of cigarette smoking in this sample was 42.8%, which is higher than what has been reported elsewhere \[[@B3],[@B4],[@B6],[@B8]\]. However, most of these prior reports focused on adolescents in high school, once again raising the possibility that the rates of cigarette use increase with age and academic progression. In this study, there was a statistically significant difference in the lifetime cigarette use prevalence rates between males and females, with males having a higher rate than females. This is consistent with what has been found in other studies, and may reflect a more tolerant social attitude to male than to female smoking \[[@B2]-[@B5],[@B16],[@B24]\]. The mean age at first cigarette use in this study was 15.7 years, and the youngest reported age was 10 years. Males had a significantly lower mean age of onset of cigarette use than females. The Kenyan Tobacco Control Act (2007) prescribes the minimum smoking age to be 18 years, yet tobacco products appear to be available to people as young as 10 years old. The same regulatory problems identified in relation to alcohol use are probably also operational with regard to tobacco use in Kenya. Elsewhere, Kwamanga et al \[[@B3]\] reported an age of cigarette smoking onset of 12 to 16 years, while Peltzer \[[@B24]\] reports a mean age of onset of 14.8 years, similar to the findings in this study. Early onset of cigarette use has obvious implications for these young people in social and academic spheres. As a matter of fact, the commonest smoking-attributed problems reported by participants in this study were associated with teachers, parents and physical health. Although it was beyond the scope of this study to identify exactly what sort of problems the smokers encountered with teachers and parents, it may be presumed that they were probably related to discipline and interpersonal conflict. Smoking is outlawed in learning institutions in Kenya, as well as in most public places. Although only a small proportion (7%) of those who smoked reported using more than 16 cigarettes a day, this is still significant in view of the health risks posed by this amount of exposure to cigarette smoke. It also suggests that this group is already dependent on nicotine, and this may cause problems in important areas of functioning including school, social and family functioning. The rates of other substance use in this study were low. This may be attributed to either under-reporting or a lack of availability of these substances. It is difficult to ascertain the true reason for the low prevalence rates since the questionnaire used was a self-report format and there was no follow-up of responses. However, this finding appears consistent with the previous studies in the same environment, where reports of illicit drug use have been low \[[@B2],[@B5],[@B16]\]. Only cannabis use has been found to be more common in the earlier studies compared to the present one. For instance, Odek-Ogunde \[[@B5]\] reported a 19.7% lifetime prevalence rate of cannabis use among university students, but found rates less than 5% for drugs such as heroin, cocaine and amphetamines. Majority of those reporting lifetime substance use indicated that they were influenced by a friend or relatives other than the nuclear family. Similar findings have been reported in other studies, confirming the role of peer pressure and social learning in initiation of substance use \[[@B8],[@B24]\]. The implication of this finding is that peers and older relatives would also serve as good role models for a substance use intervention program for young people in this setting. In this study, most respondents indicated that they used substances to relax and relieve stress. Madu and Matla found that majority of drug using adolescents do so when they are bored, tired or stressed up, or at parties \[[@B4]\]. A 1987 Medical Student survey in the US also found similar reasons for most of the drugs used, including to relax, to have a good time, to feel good and even to experiment \[[@B1]\]. This raises the possibility that interventions that help young people to use their time more productively would reduce the incidence of substance use. This study encountered a number of limitations, chief among them being the relatively small sample size and a study design that precludes generalisation of the results to students in other institutions. Due to the heterogeneity of the institutions and the courses offered in this region, it would take a larger study to fully describe the substance use patterns among college students in Kenya or even within Eldoret Municipality. The information generated from this study may however be useful in the design of such a study. Another key limitation is the fact that a self-administered questionnaire was used to collect data on current and lifetime substance use. Without any other validating measures it is not possible to conclude with certainty that the information generated is an accurate representation of substance use in this population. However, the WHO model questionnaire has been used elsewhere \[[@B2],[@B14]\] under similar circumstances, and therefore the results are at least comparable to those found in similar studies. Conclusions =========== This study has demonstrated a high prevalence of substance use among college and university students in a low income country, with the most commonly used substances being alcohol and cigarettes. Very serious effects have also been reported, with a significant proportion of those using alcohol attributing serious physical and psychosocial effects to its use. The risk of substance dependence and other mental and general medical conditions is therefore very high in this population, and targeted interventions are recommended to deal with this issue expeditiously. Due to the early age of onset of substance use found in this study, it is recommended that these interventions must target people as young as 10-12 years, and involvement of peers and role models would have a high probability of success. Competing interests =================== The authors declare that they have no competing interests. Author contributions ==================== LA was involved in the conceptualisation of the study idea, development of the study design and instruments, supervision of data collection and analysis, and preparation of the final manuscript. PAM, MNN, KCK and EMO were involved in conceptualisation of the study idea, design of the study, data collection and analysis. All authors were involved in the approval of the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-244X/11/34/prepub> Acknowledgements and Funding ============================ The authors would like to acknowledge the heads of the institutions that participated in the study for organizing their students to participate. This study was funded by Moi University School of Medicine as part of the Community Based Education and Service (COBES IV). However, the University did not have any influence on the study at any stage, apart from providing logistical support.

PubMed Central

2024-06-05T04:04:19.480387

2011-2-28

PMC3053227

Background ========== The postoperative hypothermia is a morbid condition that frequently affects patients recovering from general anesthesia \[[@B1]\]. Although the etiology of postoperative hypothermia is still under debate, it has long been associated with impaired metabolism, negative nitrogen balance, injury severity, hemorrhage, and multiple organ failure \[[@B2]-[@B4]\]. The occurrence of those dysfunctions depends on endogenous factors (endogenous hypothermia) or accidental factors (unintentional hypothermia) \[[@B4],[@B5]\]. Postoperative hypothermia, in turn, may lead to severe complications such as negative catabolism and nitrogen balance, reduced resistance to infections, delayed wound healing, impaired coagulation, myocardial ischemia, and cardiac morbidities \[[@B6]-[@B8]\]. In particular, the cardiac complication may result in an overall higher mortality especially in patients with coronary artery disease \[[@B9]\]. Physiologically, the thermal input is integrated at numerous levels within the nervous system, but the hypothalamus is the dominant controller \[[@B10]\]. The inhibition of the thermoregulatory center in the hypothalamus by general anesthetics and sedatives is one of the causes of postoperative hypothermia, via mechanisms not completely known \[[@B11]\]. In contrast, the consequence of postoperative hypothermia on electrical properties especially of brain, heart, and neuromuscular activity is more characterized. Clinical evidence shows that hypothermia decreases the passive cable properties (membrane resting potential and amplitude, resistance) and impairs neuromuscular activity in such a magnitude that it can be recorded by noninvasive electrophysiological techniques \[[@B12]-[@B14]\]. Traditionally, bioelectrical impedance (BI) has been used as a noninvasive method to assess body-fluid compartments \[[@B15],[@B16]\] and body composition \[[@B17]\]. The resistance, reactance, and impedance of each arm, leg, and trunk can be easily quantified by using hand and foot contact electrodes \[[@B18]-[@B20]\]. In the clinic, BI analysis has been successfully used to determine the cellular uncoupling triggered by an increase of intracellular Ca2+ and-or dephosphorylation of connexins during ischemia \[[@B21]\]. The cellular uncoupling is an important factor for the highest incidence of ventricular arrhythmia and cardiac failure \[[@B21],[@B22]\]. Because the ischemia and coagulopathy are complications of prolonged hypothermia, mainly in patients with trauma or coronary artery disease \[[@B23]-[@B26]\], the routine monitoring of BI in hypothermic postoperative patients would be invaluable to determine the adequate treatment. Recently, the active core rewarming has demonstrated benefits of rapid rewarming of critically injured patients \[[@B27]\]. Therefore, we used active core rewarming in patients who had undergone laparoscopic cholecystectomy under balanced inhalation anesthesia to determine whether BI depends on the core temperature. Methods ======= The study included 60 patients (52 female and 8 male) age 40 to 80 years with an ASA I-II classification. All patients had laparoscopic cholecystectomy at the \"Hospital Regional Primero de Octubre\" (ISSSTE, Mexico City, Mexico) and were divided into two groups. The control group (30 patients) was treated with room-temperature parenteral solutions and a cotton blanket, which is referred to as passive external rewarming. The experimental group (30 patients) was treated with warmed parenteral solutions (38.5°C) and a cotton blanket, which is referred to as active core rewarming. This protocol was approved by the Institutional Review Board of \"Hospital Regional Primero de Octubre\" (ISSSTE, Mexico City, Mexico) and written, informed consent was obtained from all patients. Anesthesia Technique -------------------- The anesthesia procedures were done using a Dameca siesta iTS anesthesia machine (Dameca A.S., Roedovre, Denmark). All patients were treated with balanced inhalation anesthesia and complementary medication as follows; a) In the premedication phase, Midazolam (benzodiazepine); b) in the induction phase, Atropine (anticholinergic), Fentanyl (opioid), Cisatracurium (neuromuscular relaxant), and Propofol (inductors); c) in the maintenance phase, Fentanyl and Sevoflurane (volatile anesthetic agent). The dosage of all drugs was based on the body weight. The hemodynamic variables, respiratory rate, blood-gas levels, and biochemical variables were continuously measured by using a Cardiocap TM/5 monitor (Datex-Ohmeda Corp. Helsinki, Finland). During the surgical procedure, the heart rate, blood pressure, electrocardiographic trace, arterial gasometry, breathing rate, and esophageal temperature were continuously monitored. The patients treated with passive external rewarming (control group) received room-temperature parenteral solutions (1250 ± 250 mL) during the entire study. The group of patients treated with active core rewarming received their parenteral solutions (1250 ± 250 mL) warmed to 38.2°C by using a fluid warmer (Hl-90-SP-115; Smiths Medical ASD; Rockland, MA, USA) during the trans- and postsurgical period. Additional rewarming by using a cotton blanket was provided to both groups during the postanesthesia recovery. Patients treated with passive external rewarming remained in the postanesthesia care unit longer. Physiological Evaluation ------------------------ During the trans- and postsurgical periods, homeostasis variables (blood pressure, arterial gasometry, hematic biometry, glycemia, and serum potassium levels) were evaluated three times. The first set of measurements was done before the surgical procedure; the second set, during the surgical procedure before the gallbladder was removed; and the third set, after 30 minutes of stay in the postanesthesia care unit. The body temperature (BT), respiratory rate, heart rate (HR), electrocardiography trace (EKG), mean arterial blood pressure (MABP), oxygen saturation (sO~2~), arterial carbon dioxide tension (PaCO~2~), and shivering were measured every 30 minutes for 2 hours during the stay in the postanesthesia care unit. Bioimpedance analysis --------------------- The BI was recorded by using a 4-contact electrode system to collect dual sets of segmental impedance measurements in the deltoid muscle. This apparatus was designed by Dr. E. Suaste-Gomez, and made by the staff of the Bioelectronics Section at CINVESTAV (Mexico City, Mexico). A constant current (0.8 mA) was applied to the lateral electrodes and the voltage drop was measured by the central electrodes. All assessments were made at 50 kHz with and 0.8 mA of constant skin current. The voltage output value was obtained by the addition of input impedance (Zo), as Zo + ΔZ (delta Z). The voltage output recorded in the two central electrodes was proportional to the input current and the impedance measured by the lateral electrode as a function of the distance between electrodes. BI measurements were made after the patient had rested 30 minutes in the postanesthesia care unit, which has a controlled temperature and relative humidity. The BI measurements were repeated every 30 minutes for 2 hours. Statistical analysis -------------------- All values were expressed as the mean (*s*). We determined the proper sample size (n) for BI measurements by using the equation for comparison of two means and the following values, α = 0.05; β = 0.20 with 80% of statistical power; Zα = 1.645; Zβ = 0.842; S = 30 Ω; d = 20 Ω. The Mann-Whitney unpaired *t*-test was used to determine differences in the number of patients who developed shivering between each group at each measurement period. Measurements of heart rate, mean arterial blood pressure, respiratory rate, body temperature, and BI were repeated for each group over time. Their differences were analyzed using the two-way repeated-measures ANOVA. All the statistical analysis was made using the PSPPIRE software version 0.6.0 for Linux (GNU General Public License). The null hypothesis was rejected at *P*\< 0.05 level. Results ======= The age, weight, height, gender ratio, ASA classification, and duration of surgery were not statistically different between the two tested groups. Effect of active core rewarming on body temperature and bioelectrical impedance ------------------------------------------------------------------------------- As expected, the body temperature in the control group remained as low as 35.5 ± 0.5°C, whereas the temperature increased gradually from 36°C to 37°C in patients treated with active-core rewarming. The temperature rise started 30 min after the onset of the treatment (Figure [1A](#F1){ref-type="fig"}). The patients treated with active core rewarming showed BI values of 68 ± 3 Ω, which remained stable up to the end of the testing and were significantly different to those of the patients treated with passive external rewarming (Figure [1B](#F1){ref-type="fig"}). In the latter group, the BI values increased from 76 ± 6 Ω to reach the highest values of 95 ± 11 Ω at 60 minutes postanesthesia and then decreased to 88 ± 9 Ω at the end of study period (Figure [1B](#F1){ref-type="fig"}). These values were 30% higher than those found in the group treated with active core rewarming. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Effect of active core rewarming on body temperature and bioelectrical impedance of postoperative patients**. Sixty adult patients underwent laparoscopic cholecystectomy under balanced inhalation anesthesia. A group of patients (*n*= 30) was treated with warmed parenteral solutions (38.5°C) and a cotton blanket. This group is referred to as Active Core Rewarming (•). The other group (*n*= 30) was treated with room-temperature parenteral solutions and a cotton blanket. This group is referred to as Passive External Rewarming (o). The values are expressed as the mean (*s*). All *P*values are \< 0.001. :::  ::: Effect of active core rewarming on physical and biochemical variables --------------------------------------------------------------------- The percentage of control patients showing shivering was 25% at the beginning of the study, increased to 63% after 30 min, and then gradually decreased to 25% up to the end of the study (Figure [2A](#F2){ref-type="fig"}). The counts of shivering frequency were significantly decreased (50%) throughout the testing period in patients treated with active core rewarming when compared with those of the controls (Figure [2A](#F2){ref-type="fig"}). The heart rate of the postsurgical patients treated with active core rewarming remained stable at 67 ± 1 beats/min throughout the period of study. In contrast, the heart rate of patients who received passive external rewarming increased gradually from 77 to 85 beats/min (Figure [2B](#F2){ref-type="fig"}). A statistical difference was also found in the arterial blood pressure of patients treated with active core rewarming when compared with the controls. The arterial blood pressure of patients treated with active core rewarming was 20 mmHg lower than the control (Figure [2C](#F2){ref-type="fig"}). The mean respiratory rate of patients treated with active core rewarming was statistically lower than the control values during the period of study (Figure [2D](#F2){ref-type="fig"}). No significant difference between groups was found in the measurements of pH, oxygen or carbon dioxide tension, arterial-oxygen tension to the fractional inspired-oxygen ratio, arterial-oxygen saturation (*P*\> 0.05), and in the biochemical variables (Table [1](#T1){ref-type="table"}). The postoperative patients treated with passive external rewarming showed a significant increase in the serum-potassium levels during their stay in the anesthesia care unit (Table [1](#T1){ref-type="table"}). The increase in the serum-potassium levels was not seen in the group treated with active core rewarming (Table [1](#T1){ref-type="table"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Effect of active core rewarming on physical variables of postoperative patients**. Sixty adult patients underwent laparoscopic cholecystectomy under balanced inhalation anesthesia. A group of patients (*n*= 30) was treated with warmed parenteral solutions (38.5°C) and a cotton blanket. This group is referred to as Active Core Rewarming (•). The other group (*n*= 30) was treated with room-temperature parenteral solutions and a cotton blanket. This group is referred to as Passive External Rewarming (o). The values are expressed as the mean (*s*). All *P*values are \< 0.001. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Hemodynamic and Biochemical data during pre-, trans-, and postanesthetic periods in patients after laparoscopic cholecystectomy and treated with two different rewarming techniques. ::: Variables Passive External Rewarming \[mean (*s)*, *n*= 30\] Active Core Rewarming \[mean (*s)*, *n*= 30\] ---------------------------- ---------------------------------------------------- ----------------------------------------------- ---------------- --------------- --------------- --------------- **HR (beats/min)** 78.8 ± 2.1 64.5 ± 1.9 78 ± 1\* 79.2 ± 2.1 67.9 ± 2 71 ± 1\* **RR (breaths/min)** 15.1 ± 0.4 11.5 ± 0.17 16 ± 0 17.5 ± 0.4 11.6 ± 0.2 17 ± 0 **T°C** 36.3 ± 0.05 35.9 ± 0.06 35 ± 0 36.38 ± 0.07 36.54 ± 0.57 36 ± 0 **SABP (mmHg)** 133.9 ± 3.5 104.1 ± 3.8 126 ± 2\* 124.8 ± 2.1 110.1 ± 3.1 116 ± 2\* **DABP (mmHg)** 80.7 ± 1.8 64.0 ± 1.9 77 ± 1 76.4 ± 1.8 62.3 ± 1.9 69 ± 1\* **sO~2~(%)** 92.1 ± 0.6 98.4 ± 0.1 95 ± 0.1 92 ± 0.52 98.5 ± 0.2 96 ± 0.1 **PaCO~2~(mmHg)** 34 ± 0.0 30.3 ± 0.6 40.2 ± 0.7 35 ± 0.4 31 ± 0.5 41.1 ± 0.4 **PaO~2~(mmHg)** 64.6 ± 1.4 296.9 ± 1.2 76.1 ± 1.5 65.4 ± 1.8 322.9 ± 9.1 66 ± 1.4 **HCO~3~(mmol/L)** 20.6 ± 0.2 17.78 ± 0.5 20.5 ± 0.3 21.6 ± 0.2 19.7 ± 0.3 21.13 ± 0.21 **Hgb (mg/dL)** 13.4 ± 0.3 12.8 ± 0.3 12.6 ± 0.3 13.5 ± 0.3 13.1 ± 0.3 12.9 ± 0.3 **Hct (%)** 42.9 ± 1 41 ± 1 39.9 ± 0.1 42.6 ± 0.3 40.4 ± 0.9 40 ± 1 **Leukocytes (10^3^/cm³)** 7.5 ± 2.8 7.3 ± 2.8 7.3 ± 2.8 7.9 ± 2.8 7.7 ± 3.1 7.8 ± 2.7 **Platelets (mm³)** 283567 ± 13030 274200 ± 12495 271200 ± 12818 326500 ± 9088 326467 ± 9007 326333 ± 9028 **pH** 7.41 ± 0.01 7.35 ± 0.01 7.30 ± 0.01 7.38 ± 0.01 7.38 ± 0.01 7 ± 0.01 **K (mEq/L)** 3.57 ± 0.04 3.96 ± 0.03\* 4.08 ± 0.03\* 3.58 ± 0.03 3.51 ± 0.03\* 3.45 ± 0.04\* **Glucose (mg/dL)** 95.87 ± 4.8 109.4 ± 5.4\* 116.5 ± 5.2 93.7 ± 3.5 88.7 ± 4.2\* 79.8 ± 2.7\* Data are expressed as mean (*s*) \*Significant differences between active-core rewarming and passive-external rewarming (*P*\< 0.05). HR, heart rate; HCO~3~, bicarbonate; SABP, systolic arterial blood pressure; DABP, diastolic blood pressure; sO~2~, oxygen saturation; PaCO~2~, arterial carbon dioxide tension; PaO~2~, arterial oxygen tension; RR; respiratory rate; Hgb, hemoglobin; Hct, hematocrit; K, potassium; T°C, body temperature in Celsius. ::: Discussion ========== Our results show for the first time that the BI is directly affected by the changes of core temperature in postoperative patients subjected to laparoscopic cholecystectomy. We showed that the BI values are inversely associated with the changes in the body temperature, but not with the temporal course of shivering. The gradual decrease in the body temperature led to the increase in the BI values. The highest BI value (95 ± 11 Ω) appears at 60 min when the lowest temperature (35.5 ± 0.5°C) is reached. Furthermore, the active core rewarming maintained stable physiological values of body temperature (over 36.5°C), which were associated with stable low values of the BI (68 ± 3 Ω) over time. An increasing number of works show that the BI analysis is a reliable method to determine the fluid and body composition, and the cellular uncoupling in pathological conditions \[[@B16],[@B19],[@B21]\]. On this basis, we suggest that the increase in BI in postoperative hypothermic patients reflects the development of physiological, biochemical and cellular alterations. The precise molecular composition in fluid and body and the physiological mechanism underpinning BI changes in hypothermic patients remain to be determined. Hypothermia is usually caused by a prolonged time of CO~2~insufflation during the laparoscopic surgery \[[@B28]\]. In addition, the anesthetics, sedatives, and muscle relaxants are other factors able to potentiate the effect of CO~2~insufflation on hypothermia during surgery \[[@B11]\]. These drugs are able to depress the hypothalamic thermoregulation and generate shivering as a compensatory mechanism \[[@B10]\]. In our study, the long surgical time (82 ± 5 min) as compared with others \[[@B29]\] can account for the sustained hypothermia, the increased hemodynamic values, and the shivering observed in patients who received passive external rewarming. The increased values of hemodynamic variables in the hypothermic patients of this study could result from the adrenergic response to hypothermia, which has been supported by others \[[@B30],[@B31]\]. It has long been known that the adrenergic response caused by the increased release of norepinephrine in response to cold causes vasoconstriction and secondary hypertension \[[@B32],[@B33]\]. In our study, the postoperative patients treated with passive external rewarming also showed a mild but significant increase in the serum-potassium levels during the trans- and postanesthetic period. This alteration is possibly caused by an overstimulation of the neuromuscular junction and ectopic contractions in shivering patients. Similar results have been found in electroconvulsive therapy, in which there is a fast increase of plasmatic potassium levels \[[@B34]\]. In agreement with previous results \[[@B35]\], the active core rewarming used in this study caused stable values of body temperature near the physiological range (over 36.5°C) along with a significant decrease in the hemodynamic values and shivering. These results suggest that the control of core temperature is able to prevent the adrenergic response and maintain the passive electrical properties within the physiological range. On this basis, we suggest that the BI changes depend on the core temperature changes. Because of this property, the BI may be a useful tool to monitor the recovery from hypothermia in postoperative patients, especially for those with cardiac antecedents \[[@B22]\]. In addition to routine monitoring, the BI data might lead to more timely interventions, resulting in clinical improvement and a shorter stay in the anesthesia care unit. Conclusions =========== The BI changed as a function of the core temperature and independently of the involuntary muscle activity. The physiological and molecular mechanisms underpinning the BI changes in hypothermic postoperative patients require further studies. In addition, our results support the beneficial use of active core rewarming to prevent accidental hypothermia and changes in BI. Competing interests =================== The authors have no financial and personal relationships with other people, or organizations that could inappropriately influence their work, all within 3 years of beginning the work submitted. Authors\' contributions ======================= ARG made the anesthesia procedure, and participated in the collection and analysis of the data. BSR participated in the data collection and analysis. BALCH participated in the data analysis. ESG designed the 4-contact electrode system recorder. DMF was the scientific advisor and prepared the final version of this manuscript. JAGB participated in the design of the study, in the data analysis, conceived the paper, and prepared the figures, table, and the final version of this manuscript. All authors read and approved the final version of the manuscript. Consent ======= Each patient for publication freely gave a written informed consent. A copy of the written consent is available to the Editor-in-Chief of this Journal. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2253/11/2/prepub> Acknowledgements ================ The authors thankfully acknowledge the ISSSTE for the financial support. We thank the General Director of ISSSTE (Miguel Angel Yunes-Linares) and the Director of \"Hospital Regional Primero de Octubre\" (Enrique Núñez González) for providing us the clinical and laboratory facilities. The invaluable support by the anesthesiology staff is also acknowledged. Thanks to Dr. Ellis Glazier for editing of the English-language text.

PubMed Central

2024-06-05T04:04:19.484399

2011-2-16

PMC3053228

Background ========== Cerebral small vessel disease (SVD) includes white matter lesions (WML) and lacunar infarcts and is a frequent finding on computer tomography (CT) and magnetic resonance imaging (MRI) scans of elderly people \[[@B1]\]. It is associated with vascular risk factors, such as hypertension, atherosclerosis, diabetes mellitus and atrial fibrillation \[[@B2]-[@B4]\]. In cerebral SVD symptoms are due to either complete (lacunar syndromes) or incomplete infarction (WML) of subcortical structures leading to accompanying complaints including the lacunar syndromes, cognitive, motor (gait) and/or mood disturbances \[[@B5]\]. The prevalence of WML and lacunar infarcts varies considerably across studies from 5-95% and 8-28% respectively, depending on the population studied and the imaging technique used \[[@B1],[@B6]\]. There is evidence of an increased risk of cognitive decline, dementia, gait and balance disturbances and parkinsonism among individuals with SVD, although prospective studies are scarce \[[@B7]-[@B10]\]. However, individuals with a virtually identical WML burden on conventional FLuid Attenuated Inversion Recovery (FLAIR) imaging present with a wide variance in cognitive and motor performance ranging from no complaints at all to subjective cognitive complaints and mild parkinsonian signs to dementia and parkinsonism. Apparently there are other factors that determine whether identical appearing WML on FLAIR lead to for example cognitive or motor decline in one person, while leaving others unaffected. One of the other factors could be the presence the coexisting manifestations of cerebral SVD on conventional MRI such as lacunar infarcts and cerebral microbleeds which might influence the cognitive and motor performance \[[@B11]\]. As identical appearing WML on conventional MRI are actually histopathologically heterogeneous \[[@B12]\], it could be that only the WML with a high loss of microstructural integrity are related to cognitive and motor impairment. It is also important to realize that only a small proportion of the white matter (usually less than a few percent) is affected by SVD, even among individuals with severe SVD \[[@B13]\]. As conventional MRI is not sensitive to early loss of microstructural integrity in the normal appearing white matter (NAWM), possible changes in this largest part of the white matter cannot be assessed \[[@B14],[@B15]\]. These limitations of conventional MRI can potentially be overcome with the use of Diffusion Tensor Imaging (DTI) which allows us to assess the microstructural integrity of the whole white matter \[[@B16]\]. DTI, amongst others, provides two parameters; mean diffusivity (MD), a measure of the magnitude of diffusion of water in the white matter, and fractional anisotropy (FA), which provides information about the directionality of water diffusion. Damage to the white matter is supposedly accompanied roughly by an increase in MD and a decrease in FA \[[@B17]\]. Another explanation for the clinical diversity due to WML could be the efficiency of compensation mechanisms that prevent further cognitive and motor (gait) deterioration. Support for the existence of compensatory mechanisms comes from a study among young carriers of a pre-senilin mutation (at risk for genetically determined Alzheimers\' disease (AD), but still without cognitive impairment) who showed altered functional connectivity (assessed with fMRI) compared with controls \[[@B18]\]. With innovative resting state fMRI techniques the strength of functional connectivity between brain regions can be investigated \[[@B19]\]. In that way it might be that these compensation mechanisms also play a role in the variety of clinical presentation of individuals with SVD. In the RUN DMC (Radboud University Nijmegen Diffusion tensor and Magnetic resonance imaging Cohort) study we prospectively investigate the effect of SVD on the transition from non-demented, independently living elderly people with cerebral SVD between 50 and 85 years towards cognitive and motor (gait) decline, and ultimately dementia and parkinsonism in a population with cerebral SVD. The primary objective of the RUN DMC study is to prospectively investigate the risk factors for and cognitive and motor (gait) consequences of longitudinal functional and structural changes in the integrity of the cerebral white matter as assessed by DTI, resting state fMRI and conventional structural MRI. To the best of our knowledge there are no other prospective cohort studies investigating the development of incident dementia and parkinsonism using these novel imaging techniques. Here we describe the study design and protocol of the RUN DMC study. Methods/Design ============== Study population ---------------- Cerebral SVD is characterized on neuroimaging by either WML or lacunar infarcts. Symptoms of SVD include acute symptoms, such as transient ischemic attack (TIA) or lacunar syndromes, but also subacute manifestations such as cognitive and motor (gait) disturbances \[[@B5]\]. As the onset of cerebral SVD is often insidious, clinically heterogeneous, and typically with mild symptoms, it has been suggested that the selection of subjects with cerebral SVD in clinical studies should be based on the more consistent brain imaging features \[[@B20]\]. Accordingly, in 2006, consecutive individuals referred to the Department of Neurology between October 2002 and November 2006, were selected for possible participation. Inclusion criteria were: (a) age between 50 and 85 years; (b) cerebral SVD on neuroimaging (WML and/or lacunar infarcts). Subsequently, the above mentioned acute and subacute clinical symptoms of SVD were assessed by standardized structured assessments (a questionnaire for TIA and stroke \[[@B21]\]; for cognition the Cognitive Failures Questionnaire \[[@B22]\]; for gait the Falls Questionnaire \[[@B23]\] and the Freezing of Gait Questionnaire \[[@B24]\]) Subjects who were eligible because of a lacunar syndrome were included only \> 6 months after the event to avoid acute effects on the outcomes. To be able to detect incident dementia and parkinsonism we applied the following exclusion criteria: (a) presence of dementia \[[@B25]\] and (b) parkinson(-ism)\[[@B26],[@B27]\]. In addition patients with (c)intracranial hemorrhage; (d) life expectancy of less than six months; (e) intracranial space occupying lesion; (f) (psychiatric) disease interfering with cognitive testing or follow-up; (g) recent or current use of acetylcholine-esterase inhibitors, neuroleptic agents, L-dopa or dopa-a(nta)gonists; (h) non-SVD related WML (e.g. multiple sclerosis); (i) prominent visual or hearing impairment; (j) language barrier; (k) MRI contraindications or known claustrophobia were excluded. All participants signed an informed consent form. The Medical Review Ethics Committee region Arnhem-Nijmegen approved the study. Follow-up --------- After 5 and 10 years all participants alive will be contacted for the prospective assessment of possible outcome events. This evaluation is currently being prepared for July 2011. Between 2006 and 2011 we contacted all participants every year by letter for an update on their address information and telephone number and for their survival status. In 2011 all participants alive will be invited by letter and subsequently contacted by telephone to visit our research centre. During their visit to the research centre a cognitive, gait, balance and parkinsonian signs assessment, a structured interview, physical examination, neurological examination, and an extensive MRI protocol, an electrocardiogram and an ultrasonography of the carotid arteries will be performed. All tests will be performed by the same two trained neurology residents and all MRI scans will take place on the same scanner. Outcome events -------------- Primary measures of outcome of the study are incident dementia and parkinsonism according to international diagnostic criteria\[[@B25],[@B27]\], as well as all-cause mortality and death from all vascular causes, non-fatal stroke, and non-fatal myocardial infarction. Secondary outcome measures are defined as change from baseline examination in cognitive function, gait and balance and parkinsonian signs. Incident outcome events are to be identified by three different approaches. 1\. During the follow-up a structured questionnaire on the possible occurrence on these outcome events is administered to each participant. When an incident event is suspected the treating physician will be contacted for the most recent information on that particular outcome event. 2\. When a participant died before follow-up, the general practitioner will be contacted for the most recent information on the cause of death and presence of primary outcome events. In case of presence of primary outcome events the treating physician will be contacted for the most recent information available. 3\. When during follow-up assessment participants\' test results are suggestive for incident dementia or parkinsonism, subjects will be referred to our outpatient clinic. In case the diagnosis is established according to the international criteria, this will be considered an incident case. All outcome events will be adjudicated independently by two specialised physicians, if the two classifications differ, the outcome event will be discussed and consensus will be made. Assessment of cognitive and motor outcomes ------------------------------------------ Two trained residents in neurology will administer the complete outcome assessment. ### Cognitive assessment We will use an extensive neuropsychological test battery that encompasses items from other large scale epidemiological studies that cover virtually all cognitive domains \[[@B10],[@B28]\]. A measurement of global cognitive function will be assessed by the Mini Mental State Examination (MMSE) \[[@B29]\]. The verbal memory function will be assessed by the three-trial version of the Rey Auditory Verbal Learning Test (RAVLT), a test used to evaluate the ability to acquire and retain new verbal information \[[@B30]\]. Visuospatial memory will be administered by the Rey\'s Complex Figure Test (RCFT), that consists of three subtasks: the copy trial, the immediate recall trial, within 3 minutes and the delayed recall trial, after 30 minutes \[[@B31]\]. To evaluate speed of mental processes four tests will be used; the Stroop test (three subtasks) \[[@B32]\], the Paper and Pencil Memory Scanning Task (four subtasks) \[[@B33]\], the Symbol-Digit Substitution Task, which is a modified version of the Symbol Digit Modalities Test \[[@B34]\] and a verbal fluency task in which as many animals as possible have to be named within 60 seconds, followed by as many professions within 60 seconds. To evaluate attention, the verbal series attention test (VSAT) will be used \[[@B35]\]. To register subjective cognitive failures we will administer the Cognitive failures questionnaire (CFQ) \[[@B22]\]. The tests will be carried out in quiet rooms and a stopwatch will be used in timed tests. ### Assessment of gait, balance and parkinsonian signs All participants will perform a tandem walk by walking ten steps heel to toe (registering: intact, one side step, more side steps, impossible). A quantitative gait analysis will be performed with a 5.6-meter long, 0.89-meter wide electronic walkway (GAITRite^®^MAP/CIR Inc., Havertown, PA) with sensor pads (12.7 mm apart from each other) connected to a computer. This system has strong concurrent validity and test-retest reliability, also in older people \[[@B36]\]. The participants walk twice at self-selected gait speed on low-heeled shoes. They start two meters before the carpet and walk until two meters behind it in order to measure steady-state walking. We will use a widely used modified version of the original Tinetti test with 17 items: 9 for body balance (score 0-16) and 8 for gait (score 0-12), with a maximum score of 28 \[[@B37]\]. It grades balance while sitting, standing with eyes open and closed, nudging and turning, gait initiation, stride length and width and symmetry. Functional mobility will be classified by using the widely-used TUG-test which is a timed test during which the participant is asked to rise from a standard armchair, walk 3 m, turn, walk back and sit down again \[[@B38]\]. Each participant will perform the test three times. To evaluate parkinsonian signs we apply the Unified Parkinson\'s Disease Rating Scale (UPDRS), the motor score \[[@B39]\]. Finally disease severity will be assessed with the Hoenhn and Yahr stage assessing \[[@B40]\]. For the evaluation of gait and balance we will also administer the Freezing of gait questionnaire (FOG), a questionnaire consisting of 16 items regarding gait and falls and the Falls questionnaire \[[@B23],[@B24]\]. ### Assessment of activities of daily living As a measure of disability the Barthel Index will be used \[[@B41]\]. The activities of daily living will be assessed by the instrumental activities of daily living questionnaire \[[@B42]\]. ### Structured interview #### Demographics and life style Standardized questionnaires on demographics, education (classified using 7 categories, 1 being less than primary school and 7 reflecting an academic degree)\[[@B43]\], marital status, living conditions, and life style habits (alcohol consumption, smoking, exercise) will be administered. Alcohol consumption is defined as units per day and the age at which alcohol consumption had started (and if stopped) was noted. Cigarette smoking behaviour is defined as the number of pack-years, calculated as the number of packs of cigarettes smoked per day multiplied by the number of years a participant had smoked. Exercise is expressed in the metabolic equivalent value (MET) according to accepted standards, where 1 MET is proportional to the energy expended while sitting quietly \[[@B44]\]. #### Vascular risk factors and cardiovascular disease With the aid a of structured, standardized questionnaire each participant will be asked for a history of: hypertension, diabetes mellitus, atrial fibrillation, TIA, stroke, myocardial infarction, coronary artery bypass graft, per-cutaneous transluminal coronary angiography, aortic prothesis, vascular prothesis, carotid endartectomy \[[@B2]-[@B4],[@B21]\] and migraine \[[@B45]\] The presence of a family history of myocardial infarction, cerebrovascular disease and diabetes mellitus in next of kin will be recorded. #### Current medication Current medication use will be noted and classified according to the Anatomical therapeutic chemical (ATC) classification system. (World Health Organization, WHO Collaborating Centre for drug statistics and methodology, <http://www.whocc.no/atcddd/>) Assessment of other variables ----------------------------- ### Depressive symptoms A standardized structured questionnaire used in previous large scale epidemiological studies will be used to assess for the history of depressive symptoms; normal reactions to stressful events or normal grief will carefully be excluded \[[@B46]\]. In case of a depressive episode, age of onset, the medical advice and medication use will be registered. We defined \'depression\' as those depressive episodes that have required attention of a general practitioner, psychologist, or psychiatrist. This definition includes minor depression, as well as more severe depression syndromes such as major depression and bipolar depression \[[@B46]\]. In addition participants will be screened for depressive symptoms by means of the Mini International Neuropsychiatric Interview (MINI), part A, which is a short diagnostic structured interview based on the DSM IV \[[@B47]\]. Additionally, presence of actual depressive symptoms will be assessed by two self report questionnaires, the Center of Epidemiologic Studies Depression Scale (CES-D) \[[@B48]\] and the Hospital Anxiety and Depression Scale (HADS) \[[@B49]\]. ### Additional Self-report questionnaires For the assessment of sleep disorders we will use the SCOPA-Sleep scale \[[@B50]\] and for fatigue the Checklist on Individual Strength (CIS20R) \[[@B51]\]. The overall health status (quality of life) will be assessed with the Short Form 36 (SF-36) \[[@B52],[@B53]\]. ### Physical Examination Height and weight will be measured without shoes in light clothing. The body mass index (BMI) is calculated as weight divided by height (in meters) squared. The maximal waist circumference will be measured without shirt, in standing position, between the lowest rib and the iliac crest, at the end of normal expiration \[[@B54]\]. Blood pressure and pulse rate will be measured in triplicate in supine position after 5 minutes rest. Subsequently one measurement is performed after 1 minute in upright position \[[@B3]\]. ### Neurological examination #### Primary reflexes The presence of the glabella, snout and grasp reflex, the applause sign \[[@B55]\] and the plantar response will be registered. #### Muscle strength The strength of the biceps, hand grip, iliopsoas, quadriceps and foot extensor muscles on both sides will be measured by the medical research council scale (MRC) and by a dynamometer. (Citec^®^hand-held dynamometer) \[[@B56]\]. *Sensory system*will be assessed by a quantitative measurement by vibration tuning fork (Rydel-Seiffer^®^) on both first toes and both medial malleolus, also registering ankle oedema and ankle jerks. Ancillary Investigation ----------------------- ### MRI protocol MRI scanning will be performed on a 1.5-Tesla Magnetom scanner (Siemens, Erlangen, Germany). The scanning protocol includes whole brain 3 D T1 magnetization-prepared rapid gradient-echo (MPRAGE) sequence (TR/TE/TI 2250/3.68/850ms; flip angle 15°; voxel size 1.0 × 1.0 × 1.0 mm); FLAIR pulse sequences (time repetition \[TR\]TE/TI 9000/84/2200 ms; voxel size 1.0 × 1.2 × 6.0 mm (including slice gap of 1 mm); transversal T2\* weighted gradient echo sequence (TR/TE 800/26 ms; voxel size 1.3 × 1.0 × 6.0 mm (including slice gap of 1.0 mm); DTI (TR/TE 10100/93ms; voxel size 2.5 × 2.5 × 2.5 mm; 4 unweighted scans, 30 diffusion weighted scans, with non co-linear orientation of the diffusion-weighting gradient, and b value 900 s/mm^2^) and resting state imaging using a gradient echo EPI (TR/TE 2400/40ms; voxel size 3.5 × 3.5 × 4.4 mm (including slice gap of 0.4 mm)). During resting state, subjects will be told not to concentrate on any particular subject, but just to relax with their eyes closed. The complete scanning protocol takes 31 minutes. #### White matter lesions All images will be evaluated without prior notice of any clinical parameter. WML are defined as hyperintense lesions on FLAIR MRI without corresponding cerebrospinal fluid like hypo-intense lesions on the T1 weighted image. Gliosis surrounding lacunar and territorial infarcts is not considered to be WML \[[@B57]\]. Total WML volume is calculated by an in-house developed, validated technique. #### Brain volumetry Normalization parameters to the ICBM152 linear template (as provided with SPM5; Wellcome Department of Cognitive Neurology, University College London, UK) and gray and white matter tissue and cerebrospinal fluid probability maps is computed by using SPM5 unified segmentation routines on the T1 MPRAGE images \[[@B58]\]. Total grey and white matter volumes are calculated by summing all voxel volumes that have a p \> 0.5 for belonging to the tissue class. Total brain volume is taken as the sum of total grey- and total white matter volume. Co-registration parameters of the FLAIR image to the T1 image are computed (SPM5 mutual information co-registration) and used to bring both the FLAIR and WML segmentation images into the subject\'s (anatomical) reference frame. Transformed images will visually be checked for co-registration errors. Subsequently, the WML segmentations are resampled to and combined with the white matter maps to yield to a WML map (the intersection of WML and white matter) and NAWM map (the complement of WML in white matter) in the T1 reference space. Total brain volume is taken as the sum of total gray and white matter. #### Lacunar and territorial infarcts Lacunar infarcts are defined as hypo-intense areas \> 2 mm and ≤ 15 mm on FLAIR and T1, ruling out enlarged perivascular spaces (≤ 2 mm, except around the anterior commissure, where perivascular spaces can be large) and infraputaminal pseudolacunes \[[@B57]\]. Territorial infarcts are defined as hyperintense lesions on FLAIR and hypointense lesions on T1 images \>15 mm \[[@B57]\]. #### Microbleeds Microbleeds are defined as small, homogeneous, round foci of low signal intensity on T2\* weighted images of less than 10 mm in diameter \[[@B59]\]. Microbleeds are counted per hemisphere separately. In addition they are classified as cortical/subcortical including the periventricular white matter and deep portions of the centrum semiovale (frontal, parietal, occipital and temporal separately); in the basal ganglia, including caudate nucleus, internal and external capsule, globus pallidus, thalamus and putamen; infratentorial including the cerebellar hemispheres, pons and medulla oblongata \[[@B59]\]. Lesions are not considered to be microbleeds when they are symmetric hypointensities in the globus pallidus, most likely calcifications or iron deposits, flow voids artifacts of the pial blood vessels or hyposignals in T2\* inside a lesion compatible with an infarct, likely to be hemorrhagic transformation \[[@B59]\]. #### Diffusion tensor imaging The diffusion weighted images of each participant are realigned on the unweighted image using mutual information based Matlab (The Mathworks, Inc.) routines from SPM5. Then, the diffusion tensor and its eigenvalues are computed using an SPM5 add-on <http://sourceforge.net/projects/spmtools>\[[@B60]\]. Unphysical spurious negative eigenvalues of the diffusion tensor were set to zero, after which the tensor derivatives the FA and MD are calculated \[[@B61]\]. The mean unweighted image is used to compute the co-registration parameters to the anatomical T1 image (SPM5 mutual information co-registration), which are then applied to all diffusion weighted images and results. All images are visually checked for motion artefacts and coregistration errors. Electrocardiogram ----------------- An electrocardiogram (ECG) will be performed and evaluated by a standardized assessment by an experienced cardiologist, registering frequency, cardiac rhythm, cardiac ectopias, cardiac axis, conduction time over the PQ, QRS and QTC intervals, conduction disturbances, left ventricle hypertrophy, pathologic Q\'s, infarction, repolarisation disturbances and acute ischemia. A final diagnosis is defined as normal, abnormal without clinical significance, abnormal with clinical consequences or pathologic ECG with immediate consultation of a cardiologist when necessary. Ultrasonography of the carotid arteries --------------------------------------- All ultrasound measurements will be performed by three experienced and specific trained clinical neurophysiology technicians. A carotid ultrasound assessment at which the intima media thickness (IMT) is measured in the distal left and right carotis communis, near the bulbus, will be performed. All measurements will be performed using a phased array real-time scanner (Philips i-u22, The Netherlands) with a 17-5 MHz broadband linear transducer. Two-dimensional ultrasound imaging of the carotid artery will be performed to measure the IMT. The IMT will be automatically measured by QLab^®^qualification software (V. 4.2.1.). An edge detection algorithm identified the lumen/intima and the media/adventitia interfaces within a region of interest over a 10 mm long segment and calculated the average thickness \[[@B62]\]. The same cognitive, motor, gait and balance assessment, structured interview and assessment of other variables and the same ancillary investigation were performed at baseline in 2006. Statistical analysis -------------------- ### Sample size calculation Based on the literature we expect about 60 incident dementia cases during the five year follow up (absolute risk 4-5%/year), as about half of our study population has a relatively high degree of WML \[[@B63]\]. We expect that each SD increase in MD increases this absolute risk of dementia by 2% per year. To detect this increased risk with a high probability of 90% at the 5% significance level we will need 380 participants at the end of the follow up, so therefore we included 500 participants at baseline and hope to end up with 400 participants at follow-up protocol (taking into account an expected loss to follow up of about 20%). ### Analysis of primary outcome measures We will analyze mean baseline MD and FA and change in MD and FA on follow up imaging in relation to incident dementia and parkinsonism by Cox proportional hazard models adjusted for age, sex, education, depressive symptoms, total brain volume, white matter lesion volume and lacunar infracts, where appropriate. Discussion ========== The RUN DMC study is a large prospective cohort study on causes and consequences of structural and functional changes in the integrity of the cerebral white matter (in both the WML and the NAWM) as assessed by conventional MRI as well as new techniques, such as DTI and resting state fMRI, among elderly with cerebral SVD, starting to include participants for the follow-up protocol in July 2011. Numerous studies have shown that WML observed on conventional MRI are related to vascular risk factors and have reported associations with cognitive and motor decline and found these relations to be rather weak \[[@B3],[@B4],[@B8],[@B10],[@B46]\]. To the best of our knowledge there are no prospective cohort studies on individuals with cerebral SVD investigating the development of incident dementia and parkinsonism in relation to white matter changes assessed by DTI and resting state fMRI. Strengths of the RUN DMC study include the prospective fashion of the study in which all vascular risk factors, clinical and imaging measures will be followed up after five years, and the large and well-established protocol used to explore demographics, vascular risk factors, and cognitive and motor function. The tests chosen are furthermore widely accepted and have been proven specific and sensitive in this population with structural brain changes. Another strength is the fact that it is a single centre study. Moreover, the complete study protocol will take place in one research centre with the use of a single scanner and only two investigators performing all investigations. In conclusion, the RUN DMC study has the potential to further unravel the causes and consequences of changes in white matter integrity in elderly with cerebral SVD by using new imaging techniques, DTI and resting state fMRI. When proven, changes in white matter integrity assessed by these techniques might function as a surrogate endpoint for cognitive and motor function in future therapeutic trials of vascular risk factors in SVD. The execution and completion of the follow-up of our study will ultimately unravel the role of SVD on the microstructural integrity of the white matter in the transition from \"normal\" aging to cognitive and motor decline and impairment and eventually to incident dementia and parkinsonism. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= AvN, KdL, FEdL: contribution to conception and design; acquisition of data; involvement in drafting the manuscript; final approval of the version to be published RG, IvU, LvO: acquisition of data; revising the manuscript critically; final approval of the version to be published EvD, RE, BB, BvE, MZ, IT, MOR, MvdV, MZ, DN: contribution to conception and design; revising the manuscript critically; final approval of the version to be published Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2377/11/29/prepub> Acknowledgements and Funding ============================ Dr De Leeuw received a personal fellowship of the Dutch Brain foundation (H04-12) and a clinical fellowship of the Netherlands Organization for Scientific Research (project number: 40-00703-97-07197). This study was supported by the \'Internationale Stichting Alzheimer Onderzoek\' (number 8510).

PubMed Central

2024-06-05T04:04:19.486348

2011-2-28

PMC3053229

Background ========== Polyploidisation is considered to be a major creative force in plant evolution since approximately 70% of angiosperm lineages underwent whole-genome duplications during their evolution \[[@B1]\]. In most cases genome doubling comes along with interspecific hybridization (allopolyploidy) and the genetic outcomes of these combined events are manifold and not easy to predict \[[@B1],[@B2]\]. In principle the evolutionary fate of duplicated genes, including homeologs generated by polyploidization, can result in 1) the retention and co-expression of all copies, 2) loss or silencing of some copies (non-functionalisation), 3) development of complementary copy-specific functions (sub-functionalisation) and 4) divergence between copies leading to acquisition of new functions (neo-functionalisation) \[[@B3],[@B4]\]. In case of co-expression of duplicated genes allopolyploids have to cope with negative effects of increased gene dosage, thus most genes are expressed at mid-parent levels \[[@B5],[@B6]\]. The potential for reprogramming of genetic systems increases the plasticity to react on changing environments, buffers the effect of deleterious mutations and is probably responsible for the evolutionary success of polyploids \[[@B7]\]. A disadvantageous effect of polyploidy is the possible disturbance of meiosis by doubled chromosomes which may prevent correct bivalent formation \[[@B7]\]. However, newly formed allopolyploids can maintain sexual reproduction in the majority of cases because stable bivalent formation during meiosis is enhanced by the divergence between homeologous chromosomes. Contrary, the establishment of anorthoploid (odd ploidy) hybrids is based on asexual reproduction, e. g. in *Crepis*L., *Rubus*L. and *Taraxacum*F.H. Wigg \[[@B8]\]. Peculiar exceptions among these anorthoploids are the mostly pentaploid sexual European dogroses (*Rosa*L. sect. *Caninae*(DC.) Ser.). Section *Caninae*originated by multiple hybridization events \[[@B9]\] and overcame the sterility bottleneck due to odd ploidy by the development of a unique meiosis mechanism regaining sexual reproduction \[[@B10]-[@B13]\]. This meiotic system is unique in plants, but other meiosis systems leading to comparable effects have been observed e.g. in the sexual triploid plant *Leucopogon juniperinus*R.Br. (Ericaceae) \[[@B8]\] and the triploid hybrid fish *Squalius alburnoides*\[[@B14]\]. High ploidy levels and sexuality have probably been the prerequisites for the evolutionary success of dogroses after the retreat of Pleistocenic ice shields, because dogroses are very widely spread in Central Europe and occur on a broad range of different habitats, whereas diploid and tetraploid species of other sections of *Rosa*are mainly found in glacial refugia \[[@B15]\]. The so-called canina-meiosis produces haploid pollen grains (n = x = 7) and tetraploid egg cells (n = 4x = 28) which merge to pentaploid zygotes (2n = 5x = 35; Figure [1](#F1){ref-type="fig"}). A very similar process is observed in tetraploid dogroses (2n = 4x = 28), which form also haploid pollen grains (n = x = 7) but triploid egg cells (n = 3x = 21). Bivalent formation and thus recombination occurs always between chromosomes of the *same*two highly homologous sets, one transmitted by the pollen grain and the other by the egg cell. The remaining chromosomes are exclusively transmitted by the egg cell and do not undergo chromosome pairing \[[@B16]-[@B18]\]. Thus, canina-meiosis unites intrinsically sexual reproduction (recombining bivalents) and apomixis (maternally transmitted unrecombined univalents). Previous studies demonstrated that the number of different nuclear ribosomal DNA families and microsatellite alleles was always lower than the maximum number expected from ploidy level of investigated plants, thus one allele is always present in at least two identical copies \[[@B9],[@B16]-[@B19]\]. Research on artificial hybrids revealed that alleles with identical copies are located on bivalent forming chromosomes and refer probably to an extinct diploid Proto-*Caninae*ancestor, whereas the copies located on univalents are more diverged between each other \[[@B9],[@B16],[@B17],[@B19]\]. Studying expression patterns of rDNA loci within five different dogrose species Khaitová et al. (2010) observed stable expression patterns of rDNA families on bivalent-forming genomes in contrast to frequent silencing of rDNAs from univalent-forming genomes \[[@B20]\]. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Diagram of canina meiosis**. Dogroses with a pentaploid somatic chromosome number (2n = 5x = 35) produce haploid pollen grains (1n = 1x = 7) during microsporogenesis in the anthers and tetraploid egg cells (1n = 4x = 28) during megasporogenesis in the carpels. Fertilization of haploid pollen grains and tetraploid egg cells restores the pentaploid somatic level of the next generation. Bivalent forming chromosomes are presented in red, univalent chromosomes are presented in white, grey and black. :::  ::: In this study we wanted to determine whether the differential behaviour of chromosomes during meiosis is mirrored in gene divergence and expression patterns of homeologs by the analysis of three marker genes in *Rosa canina L*. Therefore, we analysed the extent of molecular divergence between alleles of two single copy genes: *LEAFY*and cytosolic glyceraldehyde 3-phosphate dehydrogenase (*cGAPDH*); and between families of nuclear ribosomal internal transcribed spacers (*nrITS-1*). *LEAFY*encodes a transcription factor which controls floral meristem identity \[[@B21]\] and *cGAPDH*encodes an essential enzyme of glycolysis. Nuclear ribosomal ITS is part of the 18S-5.8S-26 S ribosomal DNA cluster, which is organized in long tandem arrays in one nucleolus organizer region (NOR) per genome in dogroses \[[@B22],[@B23]\]. The apparent absence of interlocus homogenization between NORs \[[@B19],[@B24]\] allows tracking different dogrose genomes by diagnostic ITS families \[[@B9],[@B19],[@B25]\]. The sequence information obtained from the homeologs of the three marker genes was then used for allele-specific transcription analyses using pyrosequencing. Results ======= Gene copy numbers ----------------- Southern hybridizations were performed to estimate the copy numbers of *LEAFY*and *cGAPDH*in *Rosa canina*(additional file [1](#S1){ref-type="supplementary-material"}). One to three fragments were detected in the digestions of genomic DNA by six different enzymes hybridized against probes of *LEAFY*or *cGAPDH*. The maximum number of three fragments within the digestions did not contradict against the expectation for *LEAFY*and *cGAPDH*to have one copy per each dogrose genome, because we expected a maximum number of five bands in pentaploids. Variation in the observed one to three bands result either from restriction sites of the enzymes *Hin*cII and *Hin*dIII within the range of the probe for some of the alleles or from variation of the number of cutting sites between dogrose genomes. Allelic variation ----------------- We sequenced approximately 1990 bp of *LEAFY*in seven individuals of *Rosa canina*; only the first about 50 bp downstream of the translation start codon and the last about 50 bp upstream of the stop codon were missing. We detected four different alleles of *LEAFY*termed *LEAFY-1, -2, -3*and *-4*(Figure [2](#F2){ref-type="fig"}). We did not sample the allele *LEAFY-4*directly by cloning analysis in the individuals H21, 194 and 378, but we detected it with the help of PCR using *LEAFY-4*-specific primers (data not shown). Genomic sequences of alleles differed between each other by 0.07% - 4.1%; their coding sequences contained no premature stop codons and 29 amino acid substitutions in total (Table [1](#T1){ref-type="table"}). The analysed plants were pentaploid implying that one of the *LEAFY*alleles had two copies, which was allele *LEAFY-3*determined by pyrosequencing of an allele-specific single nucleotide polymorphism (SNP) in genomic DNA (Figure [3](#F3){ref-type="fig"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Phylogeny of *LEAFY***. Bayesian inference of phylogeny for different alleles of *LEAFY*in *Rosa canina*based on an alignment of genomic sequences (alignment length = 2280 bp). Posterior probabilities are given above branches. The allele *LEAFY-3*marked with an asterisk has two copies in the plants H13, H19 and H20. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Number of synonymous and non-synonymous substitutions in the alignments of the coding region of *LEAFY*and *cGAPDH*and parameter estimates for the null hypothesis (H0) of the selection analyses (one ω for all alleles) employed to codeml within PAML. ::: ----------------------------------------------------------------------------------------------------------------------- Gene *LEAFY* *cGAPDH* --------------------------------------------------- --------------------------------------- --------------------------- Length of coding region\* 1119 789 No. of synonymous substitutions 21 4 No. of non-synonymous substitutions 8 1 Indels 2 \- Parameter estimate for H0 (one ω for all alleles) dS = 0.1126, dN = 0.0190, ω = 0.1688\ dS = 0.0277, dN = 0.0074\ lnL = -1715.88, k = 1.472 ω = 0.2666\ lnL = -1023.59, k = 1.559 ----------------------------------------------------------------------------------------------------------------------- \*Sequences of coding region are not complete, approximately 50 bp are missing in *LEAFY*and approximately 120 bp are missing in *cGAPDH*at both ends to start and stop codon, respectively. ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Allele-specific transcription of *LEAFY***. Frequency of allele-specific bases for five SNPs in PCR products from genomic DNA and from cDNA pools of small and large flower buds were obtained by pyrosequencing for the plants (H13, H19, H20) and are presented as boxplots consisting of sample minimum, lower quartile, median, upper quartile and sample maximum. Black boxes refer to genomic DNA, white boxes to cDNAs. Dotted lines represent the proposed frequency of an allele-specific base in genomic DNA and the null hypothesis of equal transcription for all alleles referring to their copy number: Alleles with one copy have an expected frequency 0.2; alleles with two copies have an expected frequency of 0.4 in pentaploids. Allele *LEAFY*-3 has two copies, alleles *LEAFY*-1 and *LEAFY*-4 have one copy. P-values of GLM statistics (additional file [2](#S2){ref-type="supplementary-material"}) comparing base frequencies of genomic and cDNA pools at a SNP are given above boxplots. Significant results are presented in bold. We did not find an allele-specific SNP for *LEAFY*-2 suitable for pyrosequencing analysis, but sampled this allele in all individuals from genomic DNA. :::  ::: We isolated approximately 2100 bp of the *cGAPDH*sequence in five individuals of *R. canina;*only the first about 120 bp downstream of the translation start codon and the last about 120 bp upstream of the stop codon were missing. We found four different alleles of *cGAPDH*in individual H20 and three different alleles in the other individuals (Figure [4](#F4){ref-type="fig"}). Using allele-specific primers the allele *cGAPDH-2*could be detected in all individuals but the allele *cGAPDH-4*only in individual H20 (data not shown). Genomic sequences of alleles were very similar to each other (0.08 - 2.42% sequence divergence) and we detected only five amino acid substitutions and no premature stop codons in the coding region (Table [1](#T1){ref-type="table"}). Allele frequency determination of genomic DNA indicated that allele *cGAPDH-1*has three copies in H13 and H19 and two copies in H20 (Figure [5](#F5){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Phylogeny of *cGAPDH***. Bayesian inference of phylogeny for different alleles of *cGAPDH*in *Rosa canina*based on an alignment of genomic sequences (2171 bp). Posterior probabilities are given above branches. Allele *cGAPDH-1*marked with an asterisk has three copies in plants H13 and H19 and two copies in H20. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Allele-specific transcription of *cGAPDH***. Frequency of allele-specific bases for three SNPs in PCR products from genomic DNA and from cDNA pools of small and large flower buds were obtained by pyrosequencing for the plants (H13, H19, H20) and are presented as boxplots consisting of sample minimum, lower quartile, median, upper quartile and sample maximum. Black boxes refer to genomic DNA, white boxes to cDNAs. Dotted lines represent the proposed frequency of an allele-specific base in genomic DNA and the null hypothesis of equal transcription for all alleles referring to their copy number: Alleles with one copy have an expected frequency of 0.2; alleles with two copies have an expected frequency of 0.4 and alleles with three copies have an expected frequency of 0.6 in pentaploids. Allele *cGAPDH*-1 has three copies in the plants H13 and H19 and two copies in H20, alleles *cGAPDH*-2 and *cGAPDH-3*have one copy in all sampled plants. P-values of GLM statistics (additional file [2](#S2){ref-type="supplementary-material"}) comparing base frequencies of genomic and cDNA pools at a SNP are given above boxplots. Significant results are presented in bold. We did not find an allele-specific SNP for *cGAPDH*-4 suitable for pyrosequencing analysis, but sampled this allele in plant H20 from genomic DNA. :::  ::: We identified three different alleles of *nrITS*in the plants H13 and H20 and four alleles in H19 (Figure [6](#F6){ref-type="fig"}). The alleles *Canina-1*, *Rugosa*and *Woodsii*were identical to sequences found in a previous study \[[@B9]\], but allele *Canina-2*was sampled for the first time. Whereas in case of *LEAFY*and *cGAPDH*the same allele was present in multiple copies in all plants we observed that the two closely related alleles *Canina-1*and *Canina-2*(Figure [6](#F6){ref-type="fig"}) had several copies. We determined three copies of the *Canina-1*allele in H13 and H20. We concluded from base frequencies at the SNPs measured in the genomic DNA samples of H19 that this individual had two copies of the *Canina-2*and one copy of the *Canina*-1 allele. However, base frequency at SNP 4 specific for the *Canina*-1 and *Canina*-2 allele is higher (0.778) than expected (0.6; Figure [7](#F7){ref-type="fig"}, additional file [2](#S2){ref-type="supplementary-material"}). ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Haplotype network of *nrITS-1***. Haplotype network of *nrITS-1*sequences of *Rosa canina*based on an alignment (254 bp) including consensus sequences of different *nrITS-1*types (bold font) taken from \[[@B9]\]. The allele *Canina-1*marked with an asterisk had three copies in H13 and H20, the allele *Canina-2*marked with two asterisks had two copies in H19. Pyrosequencing revealed that all individuals contained one *Rugosa*allele (Figure 7). :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Allele-specific transcription of *nrITS***. Frequency of allele-specific bases for three SNPs in PCR products from genomic DNA and from cDNA pools of small and large flower buds were obtained by pyrosequencing for the plants (H13, H19, H20) and are presented as boxplots consisting of sample minimum, lower quartile, median, upper quartile and sample maximum. Black boxes refer to genomic DNA, white boxes to cDNAs. Dotted lines represent the proposed frequency of an allele-specific base in genomic DNA and the null hypothesis of equal transcription for all alleles referring to their copy number: Alleles with one copy have an expected frequency of 0.2; alleles with two copies have an expected frequency of 0.4 and alleles with three copies have an expected frequency of 0.6 in pentaploids. SNP2 and SNP4 did not differentiate between the *Canina-1*and *Canina-2*allele, thus boxplots of genomic DNA summarize the frequency of both alleles. Because allelic composition in the genomic DNA varied between individuals H13, H20 and H19, results are presented separately. P-values of GLM statistics (additional file [2](#S2){ref-type="supplementary-material"}) comparing base frequencies of genomic and cDNA pools at a SNP are given above boxplots. Significant results are presented in bold. :::  ::: In all three marker genes we hardly observed any variation between sequences of one clade isolated from different individuals (referred as alleles, Figures [2](#F2){ref-type="fig"}, [4](#F4){ref-type="fig"}, [6](#F6){ref-type="fig"}). Within the *LEAFY*-2 and *LEAFY*-3 clade sequences of two individuals formed statistically supported sub-clades (Figure [2](#F2){ref-type="fig"}). Sequences of *LEAFY-3*H20 and H21 differed from the remaining *LEAFY*-3 sequences by one substitution in intron 3; sequences of *LEAFY*-2 H19 and 378 differed by one synonymous substitution in the coding region and three substitutions in the non-coding region. Following a strict definition these sequences have to be treated as different alleles. However, for pragmatic reasons we decided to summarize them as *LEAFY*-2 and *LEAFY*-3 alleles, respectively, because sequences were very closely related and the individuals contained only one of the respective alleles. Tree topologies based on genomic sequences (Figures [2](#F2){ref-type="fig"}, [4](#F4){ref-type="fig"}) were identical to those based on coding regions only, but posterior probabilities were higher using genomic sequences (data not shown). In order to investigate the differential evolution between alleles present in multiple copies and single copy alleles we estimated the relative rate of substitutions between different alleles of *LEAFY*and *cGAPDH*by Relative Rate Test (RRT), but no pair of sequences rejected the null hypothesis of equal branch lengths for all alleles (additional file [3](#S3){ref-type="supplementary-material"}). Selection analyses using codeml (PAML) revealed that alleles of *LEAFY*and *cGAPDH*evolved under purifying selection (Table [1](#T1){ref-type="table"}). In both genes the models assuming different selective regimes between alleles with multiple copies and singly copy alleles were not significantly better than the null hypothesis (same selective regime for all alleles; data not shown). Allele-specific transcription ----------------------------- We found five SNPs in the coding region of *LEAFY*, three SNPs of *cGAPDH*and five SNPs of *nrITS*which were specific for a certain allele and suitable for allele frequency determination by pyrosequencing (additional file [4](#S4){ref-type="supplementary-material"}). We compared the frequency of allele specific bases between samples from cDNA pools and genomic DNA to estimate the relative level of transcription for each allele. Base frequencies obtained from genomic DNA indicate the copy number of an allele and represent the null hypothesis (equal transcription for all alleles with regard to their copy number). The frequency of the allele-specific bases in cDNA-pools did not vary between plants (with regard to the copy number of this allele in a plant) and between small and large flower buds (data not shown). In *LEAFY*the frequency of the allele-specific bases of all investigated SNPs differed significantly from the null hypothesis (Figure [3](#F3){ref-type="fig"}, additional file [2](#S2){ref-type="supplementary-material"}). Transcription level of the allele *LEAFY-3*with two copies in all investigated plants was 2.3-fold lower, but transcription levels of single copy alleles *LEAFY-1*and *LEAFY-4*was approximately 2.9-fold higher than expected (Figure [3](#F3){ref-type="fig"}). We could not estimate the transcription level of *LEAFY-2*, because no suitable SNP was available. Contrary to the results of *LEAFY*, transcription of *cGAPDH-1*with three copies in plants H13 and H19 and two copies in H20 was 1.2-fold higher than expected under the null hypothesis (Figure [5](#F5){ref-type="fig"}, additional file [2](#S2){ref-type="supplementary-material"}). Base frequency of allele *cGAPDH-2*with presumably one genomic copy was slightly lower than expected, but the difference was only marginally significant. Transcription level of allele *cGAPDH-3*was significantly higher than expected from genomic DNA. Transcription of *cGAPDH-4*sampled only in plant H20 could not be analysed, because we detected no specific SNP in the coding region suitable for pyrosequencing. In *nrITS*we did not observe significant differences between the frequency of allele-specific bases of cDNA-pools and genomic DNA in any of the alleles, so that the null hypothesis of equal transcription was not rejected (Figure [7](#F7){ref-type="fig"}, additional file [2](#S2){ref-type="supplementary-material"}). Discussion ========== In this study we investigated by the analysis of two single copy genes and one ribosomal DNA locus, whether sequence divergence and transcription levels differ between homeologous nuclear genes in pentaploid *Rosa canina*. We were interested to determine whether the fate of a homeolog depends on its copy number and thus very likely on whether it is localized on bivalent forming chromosomes undergoing recombination, or on univalent chromosomes, which are transmitted \"apomictically\" (without recombination) to the offspring in dogroses. Sequence divergence between alleles ----------------------------------- We detected a maximum number of four different alleles in the analysed genes in pentaploid *Rosa canina*(Figures [2](#F2){ref-type="fig"}, [4](#F4){ref-type="fig"}, [6](#F6){ref-type="fig"}) suggesting that at least one allele has two or more identical copies, which is in accordance with previous research \[[@B9],[@B16]-[@B20]\]. These studies based on rDNA loci and microsatellites from different linkage groups demonstrated that the alleles with identical copies were always transmitted by pollen grains and egg cells and therefore must be located on bivalent forming chromosomes, whereas the remaining alleles are exclusively maternally inherited via univalent chromosomes. It is assumed that chromosome sets forming bivalents refer to a probably extinct diploid Proto-*Caninae*progenitor characterized by the *Canina*-ITS type (Figure [6](#F6){ref-type="fig"}) so far solely found in polyploid dogroses (referred to as β clade in \[[@B9],[@B19],[@B20]\]). However, this unique nrITS type might also have arisen by mutation as shown for the hybrid-specific rDNA units in *Nicotiana*allopolyploids \[[@B26]\]. The preservation of homeologs in dogroses is not exceptional and has often been used to track the hybridogenic origin of allopolyploids, e.g. \[[@B27]-[@B30]\]. However, loss of homeologs has been observed in other very recently evolved hybridogenic species \[[@B31]-[@B35]\]. These cases of massive gene loss are mainly documented in herbaceous plants, while dogroses are woody and have much longer generation times. Our data correspond with the situation found in allotetraploid cotton for which gene loss seems not to be a common phenomenon accompanying allopolyploidy \[[@B36]\]. The results found for the nrITS region are comparable but more complicated than those of the single copy genes *LEAFY*and *cGAPDH*, because nrITS is part of a gene family, large tandem repeats of ribosomal DNA loci, whose copies are normally homogenized by mechanisms of concerted evolution \[[@B37]\]. However, in dogroses homeologous rDNA clusters are also preserved, because sequences are mainly homogenized within one locus but not between loci \[[@B22],[@B23]\]. In contrast to this, some rDNA families were physically lost, degenerated or were overwritten by more dominant ones in other well studied allopolyploid systems \[[@B38]-[@B41]\]. During our analyses we found very few chimeric sequences (\< 5%) and all of them were unique, thus these sequences originated most likely by stochastic PCR recombination \[[@B42]\]. This apparent absence of recombinant alleles is concordant with the study of Khaitová et al. (2010) \[[@B20]\] in dogroses and corresponds with results from *Nicotiana*demonstrating that recombination between nuclear glutamine synthetase sequences occurred in diploid but not in allopolyploid *Nicotiana*hybrids \[[@B30]\]. In contrast, recombinant alleles between progenitor sequences were observed in allopolyploid *Gossypium*\[[@B43]\] and *Tragopogon*\[[@B44]\]. In none of the investigated genes we observed signs of loss of function for homeologs (e.g. premature stop codons, deviating GC content; see also \[[@B9]\]). Moreover, relative rate tests for *LEAFY*and *cGAPDH*did not detect differential rates of sequence evolution between alleles of one locus. Selection analyses revealed that all homeologs evolved under purifying selection and did not detect differential selective regimes between them (Table [1](#T1){ref-type="table"}). These results suggest that all homeologs of investigated loci are fully functional. However, only eight non-synonymous substitutions in *LEAFY*and only one non-synonymous in *cGAPDH*(Table [1](#T1){ref-type="table"}) were observed, so that sequence divergence might not suffice to detect different selective regimes. Differential transcription of homeologous alleles ------------------------------------------------- All homeologs of the marker genes investigated here were co-expressed, but transcription levels deviated from values expected from genomic copy number for many homeologs. Co-expression has been observed for the majority of homeologous genes in allopolyploid systems \[[@B5]\]. We found no evidence for complete epigenetic silencing of a homeolog, which has been reported for *cGAPDH*and ribosomal DNAs in allotetraploid *Tragopogon*\[[@B32]\] and for nrITS in several other pentaploid dogrose species \[[@B20]\]. Differences in transcription level were most strongly pronounced in *LEAFY*displaying a significantly lower transcription for *LEAFY-3*with two genomic copies and a higher transcription for homeologs with one copy than expected from the copy number (Figure [3](#F3){ref-type="fig"}). We observed contrary but less pronounced results for *cGAPDH*. Alleles with two or more copies were more strongly expressed than expected from genomic copy number (Figure [5](#F5){ref-type="fig"}). For nuclear ribosomal RNA we detected no deviation from the expected transcription level (Figure [7](#F7){ref-type="fig"}). These results demonstrate that transcription level is not directly related to copy number of alleles. Analogous to results from microsatellites and rDNA loci \[[@B9],[@B16]-[@B20]\] we assume that alleles with two copies are located on the bivalent forming chromosomes, even though alternative scenarios cannot be completely ruled out at the moment. Following this assumption our results suggest that there is no general evolutionary fate for a homeolog located on a bivalent- or univalent-forming chromosome. Comparable results were obtained in case of the triploid hybrid fish *Squalidus alburnoides*for which silencing patterns for dosage compensation were rather gene- than genome-specific \[[@B14]\]. According to the above cited studies we presume that *LEAFY*homeologs with one copy are located on the univalent chromosomes. The increased transcription of these *LEAFY*alleles (Figure [3](#F3){ref-type="fig"}) provides an example that genetic information from non-recombining genomes is functional and active. This contrasts findings from *Nicotiana*allopolyploids for which an inverse correlation between silencing and the intensity of inter-genomic recombination has been proposed \[[@B45]\]. It is a matter of speculation whether the pronounced transcription differences in *LEAFY*represents an exception because *LEAFY*is an transcription factor expressed in floral organs whereas the two other loci *cGAPDH*and nrITS are expressed in every tissue, but recent studies demonstrate that gene classification is not a strong predictor for differential expression patterns \[[@B46]\]. Contrary to our results Khaitová et al. (2010), who investigated six different dogrose species based on cleaved amplified polymorphism sequence (CAPS) analysis, concluded that *nrITS-1*copies located on univalent genomes are more frequently silenced than loci from bivalent forming genomes \[[@B20]\]. Using the same marker but pyrosequencing for transcription analysis we did not find any differential transcription of rDNA loci in *Rosa canina*. However, according to the results of Khaitová et al. 2010, differences in transcription level of rDNA alleles were less pronounced in *R. canina*compared to other dogrose species, e.g. *R. rubiginosa*L. \[[@B20]\]. Differences between the two studies might be caused by the origin of ribosomal RNA, which was extracted from leaves by Khaitová et al. (2010) and from two different stages of flower buds here \[[@B20]\]. Gene expression has shown to be organ-specific \[[@B47],[@B48]\] and varies strongly between leaves and floral tissues in allopolyploids \[[@B49]\]. Moreover, rRNA genes which were silenced in leaves were expressed in floral organs in *Brassica*\[[@B50]\]. We did not find differences in expression patterns between very young and elder flower buds, whereas such developmentally dependent expression patterns were shown in cotton \[[@B51]\]. Our results might also be influenced by the method of reverse transcription. We used oligo-dT primers, which are suited for RNA polymerase II products with polyadenylated 3\' ends but the poly(A) stretch is normally absent in functional rRNAs and present in intermediates of a RNA degradation pathway \[[@B52]\]. However, we do not expect a strong impact of these rare degradation products on our results because conditions of reverse transcription were not stringent and rRNAs were highly overrepresented in RNA templates. Conclusions =========== We analysed three marker genes to investigate homeolog-specific transcription levels in pentaploid dogroses. Based on previous research we assume that alleles located on bivalent-forming (recombining) chromosomes have identical copies \[[@B16],[@B17],[@B19],[@B20]\]. We could show that sequence divergence and transcription intensity is not always strongly correlated with the copy number of alleles. Thus we found no evidence that genetic information on non-recombining genomes is degraded or less functional than genes from recombining chromosomes. The absence of differential selection between dogrose genomes is surprising because it is assumed that sect. *Caninae*originated during Miocene to Pliocene (approximately 6 Mya) \[[@B53]\] and fossils of rose hips were found in deposits of the Lower Oligocene (approximately 25 Mya) \[[@B54]\]). Contrary, massive gene loss and heavily changed expression profiles have been observed in other very young allopolyploids even after a few generations \[[@B31]-[@B34]\]. Despite the preferential pairing of two homologous chromosome sets during meiosis, dogroses are functional diploids in terms of chromosome pairing as suggested by Grant (1971) \[[@B8]\]; however, they are no functional diploids in terms of genome activity, since they transcribe genes and thus use information from all involved genomes. This might be a selective advantage because polyploid dogroses dominate Central European rose populations and repel diploid wild roses towards more or less isolated habitats \[[@B15]\]. Their success could be caused by fixing heterozygosity on univalent genomes on one hand and escaping the evolutionary bottleneck of complete apomixis by maintaining recombination between bivalents on the other hand. The heterogeneous results from our analysis demand further research on the transcriptome of dogroses which considers a broader sampling of species and genes and accounts also for possible tissue-specific differences. Methods ======= Plant Material -------------- Five individuals of *R. canina*were sampled from a natural population \"Himmelreich\", Jena, Germany (plants: H13, H17, H19, H20, H21), and two individuals of *R. canina*were taken from the dogrose collection at the Botanical Garden Gießen, which were originally collected at the natural population \"Einzelberg\", Groß Schneen, Germany (plants 194, 378). Voucher specimens have been deposited at the Herbarium Gießen (GIE). Ploidy Determination -------------------- Flow-cytometry was conducted according to the method described in \[[@B55]\] using a Cell Counter Analyzer CCA II (Partec, Münster, Germany) and *Rosa arvensis*Huds. (2n = 2x = 14) as an internal diploid standard. A minimum of 10,000 nuclei giving peaks with a coefficient of variation of approximately 10% were counted. DNA and RNA Extraction ---------------------- DNA was extracted from young leaf material according to \[[@B56]\]. Total RNA was obtained from small and large floral buds using RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer\'s protocol and its modifications described by \[[@B57]\]. First strand cDNA was synthesized by RevertAid™ H Minus M-MulLV Reverse Transcriptase (Fermentas, St. Leon-Rot, Germany) using an oligo-dT primer. Sequence determination ---------------------- Sequences of *LEAFY*, *cGAPDH*and *nrITS-1*were obtained from genomic DNA to identify polymorphisms between alleles located on different chromosome sets. Primers for the amplification of *LEAFY*were designed from an alignment of cDNAs of *LEAFY*of different species of Rosaceae: LFYex1-fwd (5\'-CAAGTGGGACCTACGAGGCATGG-3\') and LFYex3-rev (5\'-TCGGCGTGACAAAGCTGACGAAG-3\'). Primers for the amplification of *cGAPDH*were designed from cDNAs of *Rosa chinensis*Jacq. and *Fragaria*× *ananassa*(Weston) Rozier taken from Genbank: GPDex2-fwd (5\'-GCCAAGATCAAGATCGGAATCAACG-3\') and GPDex11-rev (5\'-CTCGTTCAATGCAATTCCAGCCTTG-3\'). Primers for amplification of nrITS were taken from \[[@B58]\]. PCR was performed in 50 μl containing 2 μl of undiluted or diluted genomic DNA, 2 units *Taq*-Polymerase (Fermentas, St. Leon-Rot, Germany), 5.0 μl 10-fold polymerase buffer (Fermentas), 4.0 μl MgCl~2~(25 mM), 2 μl of each primer (10 μM), 5.0 μl dNTPs (2 mM). The following PCR protocol was performed: initial denaturation cycle of 150 s at 94°C, followed by 30 cycles of 30 s denaturation at 94°C, 60 s annealing \[annealing temperature (T~A~): T~A~= 58°C for *LEAFY*, T~A~= 51°C for *cGAPDH*and T~A~= 48°C for *nrITS-1*\], 180 s extension at 72° C and a final extension for 10 min at 72°C. Purified PCR-products (Wizard SV Gel and PCR clean up system, Promega, Mannheim, Germany) were cloned into the vectors pGEMT (Promega) or pJET1 (Fermentas). Ligation products were electroporated into *E. coli*JM109 or DH5α. Twenty positive clones of at least two PCR products were sequenced in both directions using the same primers as for amplification and additional internal primers for *LEAFY*(LFYex2-fwd: 5\'-CAAGAGAAGGAGATGGTTGGGAG-3\'and LFYex2-rev: 5\'-GCTGCTTGGCAATGTTCTGGAC-3\') and *cGAPDH*(GPDex6-fwd: 5\'-GTCAATGAGCATGAATACAAGTCC-3\' and GPDex6-rev: 5\'-GACTTGTATTCATGCTCATTGAC-3\'). Sequences of the alleles *LEAFY-4*, *cGAPDH-2*and *cGAPDH-4*were only sampled in some plants. To test for the presence of these alleles in the remaining plants we performed allele-specific PCRs according to the conditions described above using the forward primers (LFYin1-al4-fwd: 5\'-GGACATGTAAATAGGTCGAGAATATAT-3\', GPDin2-al2-fwd: 5\'-AGTTTTCGGATTTTGGTTTCGATC-3\' and GPDin3-al4-fwd: 5\'-ATCTTTGATGTTTTCGGAGTTATATG-3\', respectively) spanning over allele-specific indels in introns. Resulting sequences were assembled and aligned using Bioedit \[[@B59]\]. New sequence information generated within this study was deposited at the EMBL sequence archive under accession IDs FR725963 - FR725973. Southern Hybridizations ----------------------- To estimate the copy numbers of *LEAFY*and *cGAPDH*30 μg of genomic DNA of plant sample H20 was digested with either *Eco*RI, *Hin*cII, *Hin*dIII, *Kpn*I, *Pst*I or *Xba*I, separated on 1% agarose gels and blotted onto positively charged nylon membranes (VWR, Darmstadt, Germany). Membranes were hybridized with 32P-αdATP-labelled *LEAFY*or *cGAPDH*fragments according to NEBlot Kit (NEB, Frankfurt, Germany). Hybridization probes were prepared from pJET1 plasmids by PCR using the primers LFYex2-fwd, LFYex3-rev and GPDx7F \[[@B27]\], GPDex11-rev, respectively, under same conditions as above. Gene fragments of *LEAFY*produced under these conditions have an expected length of 1200 bp and those of *cGAPDH*a length of 850 bp. Phylogenetic analyses --------------------- The best fitting model according to the corrected Akaike Information Criterion for each alignment was estimated for exon and intron sequences separately with MrModeltest v. 2.3 \[[@B60]\]. The parameters of the best model for each partition were employed to reconstruct phylogenies of *LEAFY*and *cGAPDH*with MrBayes v.3.1.2 \[[@B61]\], additional file [5](#S5){ref-type="supplementary-material"}). We ran the analyses over 10,000,000 generations, sampling every 100^th^generation and discarding the 100,000 trees as burn-in resulting in a 50% majority rule consensus tree showing all compatible partitions supported by posterior probabilities (PP) for each node. The phylogeny of *LEAFY*was rooted with cDNA sequences of *Fragaria vesca*L. and other species of Rosaceae, phylogeny of *cGAPDH*with a cDNA sequence of *Fragaria × ananassa*and *Arabidopsis thaliana*L. (Heynh.). Alignments and phylogenies were deposited in Treebase \[<http://www.treebase.org> (study accession: TB2:S11025)\]. A phylogenetic network was calculated with TCS v. 1.2.1 \[[@B62]\] for the *nrITS-1*sequence data including also consensus sequences of different *Rosa nrITS*alleles detected in a former study \[[@B9]\] under 95% connection limit and gaps treated as missing data. Selection analyses ------------------ Previous studies on microsatellite alleles demonstrated that alleles with two or more copies are involved in bivalent formation \[[@B16],[@B17]\] and thus undergo recombination during meiosis. Therefore, we wanted to investigate whether alleles of *LEAFY*and *cGAPDH*with two or more copies evolve differentially from alleles with one copy. We conducted Maximum Likelihood pairwise Relative Rate Tests (RRT) implemented in the program HyPhy \[[@B63]\] using the Muse-Gaut model (MG94W9 in HyPhy \[[@B64]\] of codon substitution to estimate the relative rates of substitutions between different alleles of *LEAFY*and *cGAPDH*, respectively, and out-group sequences from *Fragaria*. The resulting parameter estimates were compared by a series of Likelihood Ratio Tests (LRT). To control for the False Discovery Rate we corrected original *P*-values with the Benjamini-Hochberg \[[@B65]\] formula as recommended by the HyPhy online discussion forum. To test whether coding regions of *LEAFY*and *cGAPDH*alleles with two or more genomic copies are under other selective regimes than alleles with one copy we estimated the ratio (ω) of the rate of non-synonymous substitutions at non-synonymous sites (dN) to synonymous substitutions at synonymous sites (dS). The estimates of ω indicate whether an allele is under purifying selection (ω \< 1), positive selection (ω \> 1) or evolves neutrally (ω = 1). We conducted the analyses based on an alignment of consensus sequences of the coding region of *LEAFY*and *cGAPDH*alleles and an unrooted topology of them using the program codeml from the PAML package \[[@B66],[@B67]\]. LRT was employed to test, whether the model assuming different ω\'s for the allele with two or more copies than alleles with one copy (alternative hypothesis) fits better to the data than the model assuming the same ω for all sequences (null hypothesis). Allele-specific transcription ----------------------------- Allele-specific single nucleotide polymorphisms (SNPs) in the coding region were used to estimate the frequency of the different alleles in cDNA pools by pyrosequencing as a measure for their specific transcription. Suitable pyrosequencing templates containing allele-specific SNPs were deduced from alignments of *LEAFY*, *cGAPDH*and *nrITS*-1, respectively. We analysed the same SNPs in genomic DNA to control for the copy number of alleles in the plants. The expected frequency in genomic DNA of an allele-specific base at a SNP is 0.2 for an allele with one copy, 0.4 for an allele with two copies and 0.6 for an allele with three copies in pentaploid individuals. These expected frequencies represent the null hypothesis of equal transcription of all alleles referring to their copy number. Three PCR products from cDNA pools of small and large flower buds of the individuals H13, H19 and H20 were amplified with primers presented in additional file [6](#S6){ref-type="supplementary-material"} according to the cycling programs mentioned above. To control for contamination of RNA extracts with genomic DNA we performed PCR reactions using RNA extracts directly. Additionally, two PCR products from genomic DNA of the same plants were generated. Template generation was done as described previously \[[@B68]\]. Briefly, purified PCR products were ligated into the vector pCR2.1-TOPO (Invitrogen, Karlsruhe, Germany). The recombinant DNA was used as template in a second PCR using universal biotin-labelled primers bt-f or bt-r and sequence specific pyrosequencing primers (additional file [7](#S7){ref-type="supplementary-material"}). Purification of biotin-labelled ssDNA was done using streptavidin Sepharose (Biotage, Uppsala, Sweden). Sequencing reaction and allele frequency determination was carried out on a PSQ96 MA machine (Biotage) following the manufacturer\'s instruction. Statistics ---------- Statistical tests were performed with SPSS v. 17.0. To test the influence of bud age and the investigated individual on transcription levels of alleles we performed Univariate ANOVA for each SNP in each locus. We detected a significant impact of the individuals but no significant impact of bud age on transcription level (data not shown). Thus we performed General Linear Model (GLM) analysis with \"individual\" as random factor to test whether allele frequency measured in genomic DNA differs significantly from allele frequency measured in cDNAs for each SNP in each locus. In cases where genomic allele composition differed between individuals we performed the tests separately. Authors\' contributions ======================= CMR, IK and MG carried out the molecular genetic studies, CMR, GT and VW participated in the design of the study. CMR drafted the manuscript. All authors read and approved the manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Southern hybridization experiments**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Pairwise comparison between allele frequencies of genomic DNA and cDNA applying General Linear Model (GLM)**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Relative Rate Test (RRT)**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **Single nucleotide polymorphisms (SNP) suitable for pyrosequencing in *LEAFY***. *cGAPDH*and *nrITS-1*. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 5 **Results of Akaike information criterion (AIC)**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 6 **Primer sequences for the amplification of primary PCR products from cDNA**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 7 **Primer sequences used for pyrosequencing analysis**. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We thank T. Krügel and D. Schnabelrauch (Max Planck Institute for Chemical Ecology, Jena, Germany) for help with flow cytometry and DNA sequencing, M. Platzer (Leibniz Institute for Age Research - Fritz Lipmann Institute, Jena, Germany) for help with pyrosequencing, M. Sandmann, F. H. Hellwig (Institute of Systematic Botany, Friedrich Schiller University, Jena, Germany), A. Härter, R. Melzer (Department of Genetics, Friedrich Schiller University), M. Krauss (student assistant, Friedrich Schiller University) for excellent technical assistance, M. Ritz (Senckenberg Museum of Natural History Görlitz, Germany) for help with statistics. The work was funded by a grant from the Deutsche Forschungsgemeinschaft (Wi 2028/1-3) under the DFG priority programme 1127 \"Radiations - Origins of Biological Diversity\". We thank the three anonymous reviewers for very valuable comments on the manuscript.

PubMed Central

2024-06-05T04:04:19.489208

2011-2-23

PMC3053230

Background ========== The union of sperm and egg pronuclei to form a new organism is the endpoint of a carefully choreographed process that includes gamete recognition, binding, and fusion of plasma membranes. Our understanding of the molecular processes underlying these phenomena has been shaped by studies in sea urchins \[[@B1]\], *Chlamydomonas*\[[@B2]-[@B4]\], *Drosophila*\[[@B5]\], mice \[[@B6]-[@B8]\] and, within the last decade, the nematode *Caenorhabditis elegans*\[[@B9]-[@B12]\]. *C. elegans*is particularly suited to the discovery of molecules necessary for spermatogenesis or fertilization because its hermaphroditic mode of reproduction is unique among genetic model organisms. Spermatogenesis defective (*spe*) mutants discovered in genetic screens using hermaphrodites can be easily recovered by mating wild type males into the sterile hermaphrodites. Mutations in the *C. elegans spe-42*gene result in the production of morphologically normal spermatozoa that are fertilization defective despite making direct contact with eggs in the spermatheca, the site of fertilization \[[@B13]\]. Six other *C. elegans*genes that have the same mutant phenotype as *spe-42*constitute the *spe-9*class \[[@B9]-[@B12],[@B14]\], named for the first of these genes to be cloned \[[@B15]\]. The five cloned genes in this class have been shown, or are predicted, to be sperm plasma membrane proteins \[[@B13],[@B16]-[@B19]\]. The phenotype of these mutants suggests that *spe-42*and the other *spe-9*class genes function at the moment that sperm and egg plasma membranes meet. The consensus of 11 topology prediction algorithms \[[@B20]-[@B30]\] suggests that SPE-42 is a six-pass transmembrane protein with its N- and C-termini facing the cytosol (Figure [1A](#F1){ref-type="fig"}). Amino acid sequence analysis showed 3 domains of potential importance for SPE-42 protein function: (1) a large extracellular domain between transmembrane helices 3 and 4 containing six conserved cysteines separated by 9-13 amino acids, (2) a DC-STAMP domain \[[@B31]\] that includes transmembrane helices 5 and 6 and (3) a predicted RING finger \[[@B32],[@B33]\] formed by 8 conserved cysteines in the C-terminal cytoplasmic domain. These 3 features are conserved in each of the 2 SPE-42 homologs that are present in many metazoan genomes including *Drosophila*, mice and humans \[5, our unpublished data\]. One of the two *Drosophila*SPE-42 homologs, Sneaky, is necessary for the plasma membrane breakdown (PMBD) event that occurs after a spermatozoon is engulfed by the egg during fertilization \[[@B5]\]. Mutation of the second cysteine in the large extracellular domain disrupts PMBD, indicating its importance for Sneaky function. Because it is predicted to be extracellular, this domain could potentially interact with proteins on the egg surface. The presence of a DC-STAMP domain, in both Sneaky and SPE-42, suggests SPE-42 may be involved in a membrane fusion event because the canonical DC-STAMP protein is required for cell fusion events in the mammalian monocyte cell lineage \[[@B34]-[@B36]\]. The mechanism through which DC-STAMP mediates membrane fusion is not presently clear. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **SPE-42 membrane topology and sequence alignment of putative RING finger domain with known RING finger proteins**. (A) SPE-42 is predicted to be a 6-pass integral membrane protein with both N- and C-termini in the cytoplasmic space. Domains of interest include a large extracellular domain between transmembrane segments 3 and 4 containing 6 conserved cysteine residues, a DC-STAMP domain and a C-terminal cytoplasmic domain with a predicted RING finger. The locations of the *eb5*and *tn1231*point mutations and the *tm2421*deletion are indicated. The *tm2421*mutant is completely sterile at all temperatures (our unpublished data). (B) Amino acid sequence alignment of the RING finger domains (dashed box from panel A) of SPE-42 and the four proteins used to build the structural model. Two Zn^++^ions are predicted to be coordinated through electron sharing with the boxed residues shown in each protein. Structurally determined and putative SPE-42 Zn^++^liganding residues are boxed. N- and C-terminal amino acids were removed to optimize the alignment in PROMALS3 D. The misaligned Zn^++^ligand is the result of topologically equivalent displacement in the known structures dependent on amino acid identity, *i.e*. cysteine or histidine. :::  ::: SPE-42 homologs in all species examined to date possess a predicted RING finger motif near the C-terminus. RING fingers are structural domains, held together by the coordination of two Zn^++^ions by the side chains of nearby cysteine, histidine or aspartic acid side chains \[[@B32],[@B37]\]. These motifs are most commonly found in E3 ubiquitin protein ligases \[[@B38],[@B39]\] where they facilitate ubiquitination of target proteins \[[@B40]\]. E3 RING finger proteins simultaneously interact with a substrate and a ubiquitin-conjugated E2 enzyme, allowing transfer of ubiquitin to the substrate by the E2. Unlike HECT domain E3 s that covalently bind ubiquitin and transfer it to the substrate directly \[[@B41]\], RING domain E3 s lack catalytic activity. Although the SPE-42 protein does not show significant sequence homology to E3 ligases outside the RING finger, this domain is the chief unifying feature in an otherwise diverse E3 ligase family. Therefore, we cannot confirm or rule out an E3 ligase-like function for SPE-42. Sequence analysis of SPE-42 family members predicts that in all cases the Zn^++^ions are coordinated by eight cysteines (C4C4 pattern), unlike most E3 RING fingers, which include a histidine in a C3HC4 arrangement \[[@B39]\]. In the experiments described here, we used a systematic mutagenesis approach to characterize the cytoplasmic C-terminal region of SPE-42 containing the putative RING finger. The ability of *spe-42*transgenes bearing single amino acid substitutions or larger scale changes to produce progeny in an otherwise self-sterile *spe-42*null mutant strain allowed us to determine the importance of individual amino acids for SPE-42 function. Our results showed that the 8 cysteine residues predicted to form a RING finger are critical for SPE-42 function *in vivo*. We used these data along with previously solved RING finger structures to develop a structural model of the SPE-42 RING domain including predictions of the specific amino acids that are most likely to participate in protein-protein interactions. Results and discussion ====================== A structural model of the SPE-42 RING finger -------------------------------------------- The RING finger at the C-terminus of SPE-42 was predicted by primary amino acid sequence analysis and closely matches the consensus Zn^++^ligand spacing of the Pfam RING finger consensus (<http://pfam.janelia.org/>; accession number PF00097) and the cloned SPE-42 homolog Sneaky \[[@B5]\]. We developed a structural model of SPE-42 using the structures of known RING finger containing proteins. The model is based on sequence conservation with the four most closely related RING finger sequences in the RSCB protein data bank (PDB) <http://www.rcsb.org/pdb/home/home.do>; c-Cbl proto-oncogene RING domain \[pbd:1FBV\] \[[@B42]\], tumor necrosis factor receptor-associated factor-6 (TRAF-6) RING domain \[pdb:[3HCT](3HCT)\] \[[@B40]\], CCR4-NOT transcription complex, subunit 4, (CNOT4) \[pdb:[1UR6](1UR6)\] \[[@B43]\], and cellulose synthase, catalytic subunit (IRX3), \[pdb:[1WEO](1WEO)\] (He, F., *et al*. personal communication). The criteria for selecting these four structures were overall sequence identity, number of contiguous homologous residues (SPE-42 sequence to known structure) and relative positioning/spacing of the zinc ligands. Multiple sequence alignment of the 5 RING domains resulted in 7 of 8 ligands to the two zinc centers aligning across all sequences represented (Figure [1B](#F1){ref-type="fig"}). Based on the known structures, we observed topological flexibility about ligands 3 and 4 leading to the misalignment of the last zinc ligand. Nonetheless, the overall RING finger domain fold was maintained (Additional file [1](#S1){ref-type="supplementary-material"}: panel A). Additional file [1](#S1){ref-type="supplementary-material"}: panel B shows the homology model of SPE-42 based on the four RING finger structures. Importantly, polar and charged residues were found in more solvent exposed positions while non-polar residues were less solvent exposed in our homology model. Because RING finger domains are typically protein-protein interaction sites, we identified the potential protein-protein interaction surface and residues on the SPE-42 model. The template structure 3HCT contains both the TRAF-6 RING finger domain and its interaction partner Ubc13, providing a means to assess critical points of contact between them. Figure [2](#F2){ref-type="fig"} shows the structural alignment of SPE-42 and TRAF-6 RING finger domains with respect to TRAF-6 binding partner Ubc13. All residues in SPE-42 within 5Å of Ubc13 resulting from the alignment are shown. Figure [3](#F3){ref-type="fig"} shows the RING finger sequence with these residues colored blue. The sequential location of identified residues with respect to the Zn^++^ligands is consistent with well-characterized E2 ubiquitin conjugating-E3 ubiquitin ligase interaction sites \[[@B39]\]. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Putative protein-protein contact residues in SPE-42 homology model**. (A) Interaction between TRAF6 (gray) and Ubc13 (red) including the SPE-42 RING domain (green) aligned for comparison. TRAF-6 was co-crystalized with Ubc13 in pdb:[3HCT](3HCT). Following structural alignment of SPE-42, all residues within 5Å of Ubc13 were determined and shown. (B) SPE-42 RING domain alone with putative protein-protein interacting residues on SPE-42 surface shown. Individual amino acids are labelled and numbered in accordance with the complete SPE-42 amino acid sequence. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **RING domain amino acids with protein-protein interaction potential**. The 8 cysteines shown in red are predicted to coordinate the two Zn^++^ions, stabilizing the RING finger structure. They are indispensable for SPE-42 function except for C681, loss of which can be partially compensated for by C678 (see Tables 1 and 2). Amino acids colored blue are predicted to be on the protein surface and within 5Å of a theoretical binding partner by the model shown in Fig. 2B. :::  ::: An additional outcome of the SPE-42/RING finger protein alignment and homology model generation was the exclusion of the first cysteine, C678, in this short sequence domain from participating in metal binding. Further, the proximity of cysteine 678 to cysteine 681, both in sequence space and in the SPE-42 model, led to the prediction that mutation of cysteine 681 would show an intermediate, non-null, phenotype. This would be expected if cysteine 678 is able to rearrange and weakly associate with the Zn^++^in place of cysteine 681. An extension of this prediction is that if both cysteine 678 and cysteine 681 are mutated simultaneously the phenotype will follow that shown by the other cysteine residues in which Zn^++^binding is irreconcilably disrupted. Contribution of RING finger cysteines to SPE-42 protein function ---------------------------------------------------------------- Transgenic constructs coding for cysteine-to-alanine substitutions at each predicted RING finger cysteine (Figure [3](#F3){ref-type="fig"}, red amino acids) were created. Individual transgenes were crossed into a *spe-42(tn1231)*or a *spe-42(tm2421)*genetic background and protein function was measured by counting live progeny. Worms homozygous for either of these *spe-42*mutations are completely sterile at the 25°C assay temperature \[13, Table [1](#T1){ref-type="table"}\], so any progeny produced result from transgenic rescue. The results of the crosses are reported in Table [1](#T1){ref-type="table"}. SPE-42 protein function was severely reduced for all cysteine to alanine mutations except for cysteine 681. Surprisingly, average broods for the cysteine 681 mutation were 45% of broods from lines with a wild type transgene, suggesting that either this amino acid is not essential for Zn^++^coordination or some other amino acid was partially compensating for its loss. As predicted by our SPE-42 homology model above, a good candidate for the compensating amino acid was cysteine 678. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Transgenic rescue with substitution mutations in RING finger cysteines. ::: Transgene Mutation Progeny n ---------------- ----------- ---------- ---- xyEx186 C681A 39 ± 6 10 xyEx187 C681A 51 ± 11 7 xyEx188 C681A 63 ± 12 5 xyEx69 C684A 6 ± 2 9 xyEx72 C684A \<1 6 xyEx77 C684A 0 9 xyEx48 C700A \<1 8 xyEx52 C700A 1 ± 0.68 8 xyEx63 C700A \<1 10 xyEx134 C703A 3 ± 1 12 xyEx135 C703A \<1 9 xyEx136 C703A 3 ± 1 10 xyEx73 C708A 0 5 xyEx78 C708A 0 5 xyEx80 C708A 0 5 xyEx74 C711A \<1 9 xyEx81 C711A 3 ± 1 10 xyEx129 C711A 2 ± 0.76 6 xyEx130 C718A 0 9 xyEx131 C718A \<1 9 xyEx132 C718A 0 9 xyEx85 C721A 2 ± 0.60 9 xyEx86 C721A 5 ± 2 10 xyEx87 C721A 11 ± 1 9 ebEx498 wild type 92 ± 16 11 xyEx175 wild type 123 ± 15 10 xyEx177 wild type 115 ± 17 11 N2 (wild type) \- 185 ± 5 10 spe-42(tn1231) \- 0 24 spe-42(tm2421) \- 0 15 Live progeny were counted for three independently-derived transgenes for each amino acid substitution. All transgenes are in a *spe-42(tn1231 or tm2421)*null genetic background. N2 and *spe-42*worms listed at the bottom do not carry transgenes. All counts were conducted until the first day each worm produce no live progeny. Counts shown are ±SEM and were rounded to the nearest whole number with the exception of SEM values less than one. *spe-42(tn1231)*data are from \[[@B13]\]. ::: To test this model, two additional transgenes were made: one in which cysteine 678 was substituted with alanine and a second in which both cysteines 678 and 681 were substituted (Table [2](#T2){ref-type="table"}). As predicted by the model, transgenes with the cysteine 678 mutation rescued at the same level as wild type transgenes, indicating that this mutation alone does not affect protein function. The cysteine 678/681 double mutant reduced brood counts to the same low level observed for the other 7 cysteine mutant transgenes described above. These results clearly demonstrate that, while cysteine 678 is not required for Zn^++^coordination under normal circumstances, it can partially compensate for the loss of cysteine 681, restoring protein function to a level sufficient for fertilization to take place. Taken together, our results make a strong case that the 8 red cysteines shown in Figure [3](#F3){ref-type="fig"} are actually involved in the Zn^++^coordination that holds together the SPE-42 RING finger structure and that our SPE-42 homology model, generated from known RING finger structures, is robust and predictive. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Cysteine 678 is not part of the RING finger. ::: Transgene Mutation Progeny n ------------------ -------------- ---------- ---- *xyEx208* C678A 117 ± 22 10 *xyEx209* C678A 101 ± 22 9 *xyEx212* C678A 113 ± 13 8 *xyEx186* C681A 39 ± 6 10 xyEx187 C681A 51 ± 11 7 xyEx188 C681A 63 ± 12 5 *xyEx207* C678A; C681A 3 ± 1 10 *xyEx213* C678A; C681A 2 ± 1 10 *xyEx220* C678A; C681A 1 ± 0.65 10 *ebEx498* wild type 92 ± 16 11 *xyEx175* wild type 123 ± 15 10 *xyEx177* wild type 115 ± 17 11 N2 (wild type) \- 185 ± 5 10 *spe-42(tn1231)* \- 0 24 *spe-42(tm2421)* \- 0 15 Counts were conducted as described for table 1. ::: Incompatibility of the *C. elegans*and *C. briggsae*SPE-42 proteins is not caused by the specific amino acid sequence C-terminal to the RING finger --------------------------------------------------------------------------------------------------------------------------------------------------- *C. briggsae*is in the closest sister taxon to *C. elegans*, the two species having diverged about 100 million years ago \[[@B44]\]. Although *C. elegans*and *C. briggsae*are difficult to distinguish by phenotype, they are reproductively isolated. Pair-wise comparison of *C. elegans*SPE-42 with its *C. briggsae*homolog revealed that the two proteins are 85% identical and 93.4% similar except for the C-terminal 29 and 30 amino acids, respectively, where they are only 25% identical and 37.5% similar \[13, Figure [4A](#F4){ref-type="fig"}\]. The divergent sequence is encoded by the last exon of each gene. The two most likely explanations for this divergence are first, that this part of the protein is not important for function and therefore is under very little evolutionary constraint or second, that the sequence changes define a key difference between the two proteins that may explain the observed reproductive isolation. A transgene containing a genomic DNA clone of the *C. briggsae spe-42*homolog under the control of the *C. elegans spe-42*promoter and 3\' UTR (Figure [4B](#F4){ref-type="fig"}) failed to produce more than a few progeny in *spe-42*null mutant worms, suggesting that some aspect of the *C. briggsae*protein is incompatible with the *C. elegans*fertilization machinery (Table [3](#T3){ref-type="table"}). Two chimeric constructs in which the last exon of *C. elegans spe-42*was swapped into the *C. briggsae*gene and vice versa were created to ask whether adding the *C. elegans*exon to the *C. briggsae*gene would improve transgenic rescue and adding the *C. briggsae*exon to the *C. elegans*gene would decrease progeny compared to a wild type *C. elegans spe-42*transgene (Figure [4B](#F4){ref-type="fig"}). The results of these crosses show that replacing the C-terminal 30 amino acids in the *C. briggsae*protein with the C-terminal 29 amino acids in the *C. elegans*protein increased average progeny counts by only a few worms, while the reciprocal swap of *C. briggsae*amino acids into the *C. elegans*protein did not change progeny counts at all (Table [3](#T3){ref-type="table"}). These results demonstrate that the incompatibility between the *C. elegans*and *C. briggsae*homologs therefore lies in one or more of the individual sequence changes scattered throughout the rest of the protein. These differences define a starting point for continuing our structure-function investigation of SPE-42 outside of the RING finger domain. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### *C. elegans*/*C. briggsae spe-42*final exon swap. ::: Construct Transgene Progeny n ----------- ------------------ ---------- ---- *ebEx498* 92 ± 16 11 *xyEx175* 123 ± 15 10 *xyEx177* 115 ± 17 11 *xyEx1* 2 ± 1 10 *xyEx2* 1 ± 0.62 9 *xyEx3* 6 ± 2 9 *xyEx27* 11 ± 4 10 *xyEx30* 8 ± 4 10 *xyEx34* 17 ± 4 10 *xyEx36* 83 ± 15 9 *xyEx37* 110 ± 10 8 *xyEx41* 123 ± 15 10 N2 (wild type) 185 ± 5 10 *spe-42(tn1231)* 0 24 *spe-42(tm2421)* 0 15 Transgenic rescue was quantified for three independently-derived transgenes for each construct. Exon shading scheme is the same as in Fig. 4B. Constructs were fused to the *C. elegans*promoter and 3\' UTR at the start and stop codons. All transgenes are in a *spe-42*null genetic background. N2 and *spe-42*worms listed at the bottom do not carry transgenes. All counts were conducted until the first day each worm produced no live progeny. Counts are ± SEM rounded to the nearest whole number except for values less than one. The *spe-42(tn1231)*data are from \[[@B13]\]. ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### ***C. elegans*/*C. briggsae*chimeric constructs**. (A) Pair-wise comparison of *C. elegans*and *C. briggsae*SPE-42 proteins showing the striking divergence between the two proteins at the C-terminus. The divergent sequence corresponds exactly to the 3\'-most exon of each gene. Solid lines above the sequence indicate membrane-spanning domains. (B) Chimeric constructs were made to determine the effect of the divergent sequence on protein function. All four constructs are under the control of the *C. elegans spe-42*promoter and 3\' UTR. :::  ::: Potential function of the RING domain ------------------------------------- RING fingers are multi-functional protein-protein interaction domains normally associated with E3 ubiquitin ligase activity. Although examples of RING fingers in proteins that are not E3 s may exist, functional data support an E3 ligase activity for almost half of the 300 predicted human RING finger proteins \[[@B39]\]; the rest either have not been examined or no alternative molecular mechanism of action has been assigned. Some RING finger proteins like Bard1 do not possess intrinsic E3 ligase activity but are instead part of a multi-protein complex that does, in this case the Brca1-Bard1 heterodimer \[[@B45]\]. Other RING fingers like TRAF6 must form homodimers in order to interact with E2 ligases \[[@B40]\]. The SPE-42 RING finger could potentially function in multiple ways during *C. elegans*fertilization. Regardless of whether SPE-42 is itself an E3 ubiquitin ligase, is part of an E3 complex with other proteins, or does not have E3 activity at all, the RING finger-like domain likely mediates protein-protein interactions. SPE-42 interaction with its binding partner(s) could serve as a checkpoint for sperm competence to fuse with the egg, as a signal to the egg that the sperm nucleus and other contents have been delivered, or even as a signal that fusion has occurred and membrane proteins important for fertilization should be degraded to prevent polyspermy. Recent evidence from *Chlamydomonas*showed that FUS1 and HAP2, proteins essential for gamete fusion, are rapidly degraded immediately following membrane fusion \[[@B2]\]. These researchers further demonstrated that it was the membrane fusion event and not merely gamete binding prior to fusion that initiated protein degradation. The SPE-42 RING finger could act while still tethered to the plasma membrane or it could be cleaved from the rest of the protein, releasing it into the cytoplasm to act elsewhere. The mouse RING finger protein 13 (RNF13) is an E3 ubiquitin ligase that is normally an integral endosomal membrane protein. RNF13 can be proteolytically cleaved, however, releasing the RING domain into the cytoplasm \[[@B46]\]. SPE-42 also has a DC-STAMP domain that may be involved in membrane fusion events and a large extracellular domain that could interact with egg surface proteins. SPE-42 therefore has potential to interact with egg surface proteins, help to mediate sperm-egg fusion and signal that fusion has occurred. Use of SPE-42 expressed from extrachromosomal transgenic arrays to analyze protein function ------------------------------------------------------------------------------------------- Extrachromosomal arrays composed entirely of simple repeats of the injected experimental and marker DNA are selectively targeted for silencing in the male germline \[[@B47],[@B48]\]. The presumably small amount of protein expressed from simple-repeat extrachromosomal arrays is not detectable by immunofluorescence or western blot using antibodies that can detect both wild type and mutant protein expressed from the endogenous gene locus \[[@B17]\]. Nevertheless, this small amount of protein is sufficient to produce robust rescue of every cloned *spe*and *fer*mutant described in the literature \[for individual references, see 9, 12\]. Hemagglutinin (HA) epitope-tagged SPE-42 expressed from a low-copy genomic insertion is not detectable by a commercial HA monoclonal antibody despite the fact that the transgene rescues the self-sterile Spe-42 mutant phenotype (H. Nishimura and S.W. L\'Hernault, personal communication). Attempts by the same group to detect a SPE-42::HA fusion expressed in CHO tissue culture cells have also been unsuccessful. Because the mutant SPE-42 protein expressed in our assays is not detectable, the possibility that the loss of fertility observed with the mutant transgenes is due to protein degradation or mis-localization instead of a simple loss of protein function cannot be excluded. The complete rescue observed in the C678A mutant, partial rescue seen in the C681A mutant and almost complete lack of rescue seen in the C678A; C681A double mutant strongly suggest that C678 can at least partially compensate for loss of C681 but that C681 is actually involved in Zn^++^coordination in wild type SPE-42. Our results conclusively demonstrate that enough functional protein is reaching the site of action to provide \~46% of wild type activity levels in the C681A mutant, suggesting that significant degradation is not occuring. It is therefore plausible that SPE-42 protein expressed from our mutant transgenes is correctly localized and simply lacks function. Conclusions =========== We built a structural model of the SPE-42 RING finger domain based on solved structures of other RING domain proteins. Our experimental results demonstrate that 8 cysteine residues predicted to form a RING finger in the C-terminal domain of SPE-42 by the model are critical for protein function during fertilization. We also showed that incompatibility of the *C. elegans*and *C. briggsae*SPE-42 homologs does not result from evolutionary divergence in amino acid sequence at the C-terminus of the proteins. Our data provide a starting point for further investigation of SPE-42 function during *C. elegans*fertilization and elucidation of the potential roles of SPE-42 homologs in other species. Methods ======= Worm strains and handling ------------------------- Worm culture and genetic crosses were performed according to standard methods \[[@B49]\]. Bristol N2 was used as the wild type strain, and all mutants are in an N2 genetic background. The following mutant alleles, markers and genetic balancers were used: *him-8(e1489)*IV \[[@B50]\], *nT1\[qIs51\]*(IV;V) \[51, K. Siegfried and J. Kimble, personal communication\], *spe-42(tn1231)*V \[[@B13]\], *spe-42(tm2421)*V (S. Mitani, personal communication), *mIs10*V (K. Liu and A. Fire, personal communication). DNA constructs -------------- A 4946 bp *spe-42*genomic DNA fragment of cosmid B0240, pTK15, was created by sequentially cloning PCR products made with primers TK306 and TK309 (5\'-GCGGGCCCTGAAACAATAAATCAGTGAATTAG-3\'; 5\'-GTCTCGAGTTGACTGAAATATTTTCAATTCCTCG-3\'; 944 bp; ApaI and XhoI restriction sites), TK310 and TK311 (5\'-AACTCGAGACCGGAAATTCCCATTTACC-3\'; 5\'-GTTCTAGATGAAAAGATGACAAAGTAAGTTG-3\'; 1635 bp; XhoI and XbaI restriction sites) and TK312 and TK313 (5\'-GGCCACCGCGGTGGCGAGTTTGTGGTTTC-3\'; 5\'-CATCTAGAACCAATAGCATTTTCTTGACC-3\'; 2367 bp; XbaI and BstXI sites) into pBluescript II SK(+) (Agilent Technologies). During PCR, the central 6 bases in the *spe-42*BstXI site were changed to match those in the plasmid. Pfu Turbo polymerase (Agilent Technologies) was used for PCR, and all products were sequenced. This construct (pTK15; \"wild type\" in Tables [1](#T1){ref-type="table"}, [2](#T2){ref-type="table"} and [3](#T3){ref-type="table"}; \"*C. elegans*\" in Figure [4B](#F4){ref-type="fig"}) includes 1409 bp of promoter sequence and 265 bp of 3\' UTR sequence in addition to the *spe-42*coding region and contains no genes other than *spe-42*(B0240.2). Mutations to create cysteine to alanine substitutions in the SPE-42 RING domain were made using an overlap PCR strategy and cloned into the ApaI and XhoI sites in pTK15, replacing the wild type sequence. One of the 2 *C. elegans*preferred alanine codons (GCT or GCC) was used for all mutations to ensure robust protein expression \[[@B52]\]. Primer sequences used to make mutations are available upon request. A 6078 bp genomic DNA fragment containing the *C. briggsae spe-42*homolog was amplified from *C. briggsae*genomic DNA using the Expand Long Template PCR System (Roche) with primers TK330 (5\'-GCCTCGAGTGAATGTTAATGAGCAGCCACC-3\') and TK331 (5\'-GCGGATCCGTGGGATGGTGGAAGCAGAAAGTTG-3\') and cloned into the BamHI and XhoI sites of pBluescript II SK(+) to create pTK14. Overlap PCR using amplimers made with primers T7 and TK361 (5\'-GTAATACGACTCACTATAGGGC-3\'; 5\'-GTTCTGCAGTTGTTGTTTGAATGATGTTTGAAAAATTAACTTTC-3\') and primers TK360 and TK412 (GAAAGTTAATTTTTCAAACATCATTCAAACAACAACTGCAGAAC-3\'; 5\'-GCCTCGAGAAATGCGGCAGACGATAC-3\') was used to join the *C. briggsae*gene 3\' exons to the *C. elegans*stop codon and 3\' UTR. The 1775 bp amplimer was digested with ApaI and XhoI and cloned into the corresponding sites of pTK15 to make pTK43. Overlap PCR using amplimers made with primers T3 and TK343 (5\'-CGCAATTAACCCTCACTAAAGGG-3\'; 5\'-CATAATCCTATCAACCCCATAGTTGATACAATTCAGTTAGATTTTA-3\') and primers TK342 and TK359 (5\'-TAAAATCTAACTGAATTGTATCAACTATGGGGTTGATAGGATTATG-3\'; 5\'-GCTCGAGGTTTCAAACATGATGATGTTCCGGA-3\') was used to join the *C. elegans spe-42*promoter to the *C. briggsae*gene at the start codon. A 1949 BstXI/XhoI fragment of the amplimer was cloned into pTK43 to create pTK44. The central part of the *C. briggsae spe-42*gene was cut from pTK14 with BspEI and ClaI and cloned into pTK44 to make pTK45 (\"*C. briggsae*\" in Figure [4B](#F4){ref-type="fig"}). The pTK45 construct contains the entire *C. briggsae spe-42*gene (including introns) under the control of *C. elegans*5\' and 3\' regulatory sequence. Overlap PCR using amplimers made with primers T7 (sequence above) and TK413 (5\'-TTTTGTGACTAAATTTAATTTCAGATATTTCATTGAGTGAACGC-3\') and primers TK414 and TK345 (5\'-GCGTTCACTCAATGAAATATCTGAAATTAAATTTAGTCACAAAA-3\' and TAGAGTTCGTGCTCTTAGCCTGTAC-3\') was used to join the last exon of *C. elegans spe-42*to the penultimate exon of the *C. briggsae*gene. This 882 bp chimeric fragment was cloned into the KpnI and XbaI sites of pTK45 to create pTK46 (\"*C. briggsae*with *C. elegans*last exon\" in Figure [4B](#F4){ref-type="fig"}). In the reciprocal swap, overlap PCR using amplimers made with primers T7 and TK415 (5\'-AACATTAATAATCATCTATTCCAGAAAGTTTATCTAGTGAAACAG-3\') primers TK309 (sequence above) and TK416 (5\'-CTGTTTCACTAGATAAACTTTCTGGAATAGATGATTATTAATGTT-3\') was used to join the *C. briggsae*last exon to the penultimate exon of the *C. elegans*gene. The 982 bp fragment was cloned into the ApaI and XhoI sites of pTK15 to make pTK47 (\"*C. elegans*with *C. briggsae*last exon\" in Figure [4B](#F4){ref-type="fig"}). Transgenes and fertility assays ------------------------------- Transgenes were created by co-injecting *spe-42*DNA constructs with the marker plasmid pPD118.20 (Fire Lab Vector Kit), which drives GFP expression in body wall muscle under the control of the *myo-3*promoter using standard protocols \[[@B53]\]. At least 3 independently derived transgenes were analyzed for each mutant construct to control for variations inherent to extrachromosomal arrays. Transgenes were crossed into a *spe-42(tn1231 or tm2421)*mutant background using *myo-3::GFP*to follow the transgenes and *mIs10*, an integrated *myo-2::GFP*transgene, to pick *spe-42(tn1231 or tm2421)*homozygotes. Crosses were conducted at 20°C to facilitate efficient mating until the last step when *spe-42(tn1231 or tm2421)*/*mIs10; xyEx*mutant transgene hermaphrodites were transferred to 25°C. Worms used for progeny counts were therefore raised exclusively at 25°C since they were embryos. Crossing over between *mIs10*and *spe-42*was occasionally observed. To control for this possibility, non-GFP worms were picked from suspected recombinants and tested for fertility. Lack of complete sterility of *spe-42(tn1231 or tm2421)*homozygotes at the 25°C assay temperature is an indicator for recombination. Broods from recombinant worms were excluded from analysis. Transgenic worms were picked to individual NGM plates at 25°C and transferred to new plates daily. Counts for each worm were continued until the first day in which no live progeny were produced. SPE-42 homology model generation -------------------------------- The relationship of the SPE-42 c-terminal domain to RING finger containing proteins was determined by BLAST search with a low stringency expect threshold \[[@B54]\]. To generate a homology model, all RING finger containing structures in the RSCB protein data bank (PBD) were extracted. Sequences were retrieved by common name, sequence similarity to known RING fingers, and structure search using the VAST algorithm \[[@B55]\]. A set of 108 sequences was retrieved and immediately reduced by removing duplicate sequences, false positives (*i.e*. non-zinc containing proteins), and NMR structures having minimal restraints and/or domain sizes. The remaining 46 RING finger-like sequences were aligned by multiple sequence alignment; the sequences are listed in Additional file [2](#S2){ref-type="supplementary-material"}. The PROMALS and PROMALS3 D programs were used for all sequence alignments \[[@B56],[@B57]\]. The four most closely related sequences, by sequence similarity and metal binding residue disposition, were used for homology modeling using MODELLER software \[[@B58]\]. PDB structures \[[1FBV](1FBV), [3HCT](3HCT), [1UR6](1UR6), and [1WEO](1WEO)\] were used to restrain the SPE42 model based on a PROMALS3 D seed alignment following N and C-terminal truncation to match sequence lengths. Functionally important residues were identified by structural alignment of the SPE-42 model with 3HCT (TRAF6 RING finger domain and its Ubc13 E2 domain partner) using Chimera \[[@B59]\]. Following alignment, all residues within 5Å of the E2 ubiquitin-conjugating enzyme were determined (Figure [2A](#F2){ref-type="fig"} and [2B](#F2){ref-type="fig"}). Authors\' contributions ======================= LDW and TLK made molecular constructs and performed worm microinjection. LDW, JMS and TLK derived transgenic worm lines. LDW, BDM, ALR, KT and TLK performed worm crosses and progeny counts. JNR performed *in silico*modelling and analysis of the RING domain and wrote the corresponding sections of the manuscript. TLK designed and supervised the experiments and wrote the remainder of the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **RING finger overlay and homology model of SPE-42 RING domain**. PNG image file showing (A) Structural alignment of known RING finger proteins 1FBV (blue), 3HCT (white), 1UR6 (red) and 1WEO (orange) with putative SPE-42 RING finger (green). (B) SPE-42 model resulting from the simultaneous comparison to 1FBV, 3HCT, 1UR6, and 1WEO using MODELLER. Overall backbone structure, location of the 2 Zn^++^ions, and position of critical cysteines are shown. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **RING finger-like sequences used for alignment with SPE-42**. Excel spreadsheet file containing accession numbers and descriptions for the 46 RING finger proteins initially considered for modeling the SPE-42 RING domain. The top 4 proteins in the list were used to build the SPE-42 structural model. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We thank Matt Andrews and Shannon Stevenson for critical comments on the manuscript. We also thank Julia Curry and Kayla York for performing preliminary analysis of some mutant transgenes. Some worm strains were provided by the *Caenorhabditis*Genetics Center, which is funded by the NIH National Center for Research Resources. The *spe-42(tm2421)*allele was provided by S. Mitani at the Japanese National BioResource Project <http://www.shigen.nig.ac.jp/c.elegans/index.jsp>. Funding to TLK was from the National Science Foundation (IOS-0918464) under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5), University of Minnesota Startup Funds, and a University of Minnesota Faculty Grant-in-Aid.

PubMed Central

2024-06-05T04:04:19.492893

2011-2-23

PMC3053231

Background ========== Football (soccer) matches traditionally take place on natural grass, although different playing surfaces are used including sand, dirt, clay, concrete, asphalt, and hardwood. In dray counties maintaining a natural grass surface is expensive; therefore in this area especially in parts of Asian and Africa most football fields are dirt fields (DFs) (bare earth) and many football players, especially amateur ones, play on DFs in these regions. Another substitute in this area could be artificial turf field (ATF). In recent years the playing surfaces of amateur football players have begun being replaced by artificial grass in some counties\[[@B1]\]. The forces transmitted to football player tissues are varied on different surfaces. Therefore, injury frequency and injury pattern in football might be varied among players who play on different surfaces. It has been suggested that changes in surfaces might have effects on the performance and injury pattern of the sport\[[@B2],[@B3]\]. Two main factors which might affect surface related football injuries reported to be the stiffness of a surface and the friction between surface and shoe\[[@B2],[@B3]\]. DFs are generally not only stiff but also high friction surfaces. Moreover these fields might be uneven. These might lead to a higher and different type of injuries on diet field compare to other football fields. Some studies have been evaluated the risk of football injuries on artificial turf (AT) and natural grass. These studies have shown conflicting results, although in general the range of outcomes has been of equivalent risk for injuries on AT and natural grass\[[@B4]-[@B7]\]. A few studies evaluated incidence of football injuries on other types of surfaces that show conflicting results. Engebretsen and Kase (1987) \[[@B2],[@B8]\] investigated 16 Norwegian football teams over a two year period. They reported an injury rate of 20 per 1000 hours on gravel surface that was lower than the reported injury rate on AT. McGrath and Ozanne-Smith also cited a study \[[@B2]\] in which incidence of injury on 230 football fields during 380,000 playing hours was investigated (1985). The injury incidence was reported as 2.6/1000 hours on grass, 1.8 on gravel and 0.4 on AT. Arnason, et al (1996) \[[@B9]\] reported that significantly more injuries occurred on AT than on grass or gravel in correlation to number of hours in games and practices among Icelandic elite football players. In contrast, Ekstrand and Nigg (1989) \[[@B3]\] in their review on football surfaces, cited a 2-year study on the first artificial football surface in Sweden. They reported no difference in the incidence of injury between artificial, gravel or grass surfaces. To our knowledge no study has evaluated the incidence, nature, severity and cause of match injuries sustained on dirt field (DF). The objectives of this study were to evaluate and to compare the incidence, nature, severity and cause of match injuries sustained on DF and ATF by amateur male football players. Methods ======= A prospective two-cohort design was employed in this study. Participants were 252 male non professional football players (mean age 27 years, range 18-43) in 14 teams who participated in a local (zone) championship carried on a DF during 13 weeks (91 matches) (first cohort), and 216 male non professional football players (mean age 28 years, range 17-40) in 12 teams who participated in a local championship carried on a ATF (Second Generation) during 11 weeks (66 matches) (second cohort). There are numerous football DFs in Tehran. Council of Tehran have recently started to cover some of these fields by AT. This made an opportunity to compare the injuries incidence rate of players who play on dirt and AT fields in two near zones which are in a same area in Tehran and are in similar athletic performance levels. Football dirt fields in Iran are football fields on bare earth that just have been leveled and not covered by any material such as clay or sand. Normally no maintenance activity is needed for these fields except for making re-flat the field every several years. The incidence, nature, severity and cause of match injuries were recorded prospectively during the leagues. Both leagues were conducted during four months from mid July to mid November 2008. Injury definitions and recording procedures were compliant with the international consensus statement for epidemiological studies of injuries in football\[[@B10]\]. Two physicians recorded details of the diagnosis, severity and cause of all match injuries. One physician was present in each match that was carried out on either ATF or DFs. Match exposures were recorded by the physicians. The number of days lost from training and match play was used to define the severity of an injury. Match exposures (player hours) of all subjects were recorded. An injury is defined as \" any physical complaint sustained by a player that results from a football match, irrespective of the need for medical attention or time loss from football activities\"\[[@B10]\]. We do not included problems other than physical complaints including illnesses and diseases. The study procedure was approved by the Tehran University of Medical Sciences Ethics Committee. Data analysis ------------- Incidences of match injuries on both artificial turf and dirt fields were reported as the number of injuries/1000 player match hours with 95% confidence interval (CI). Rate ratio was calculated to compare the ratio of injuries incidences that were occurred in two cohorts. The differences between the injuries incidences were considered significant when the p value of the two-sided z test for the comparison of rates was less than 0.05. Version 10 of \"Data Analysis and Statistical Software\" was used for data analysis. Results ======= Demographic characteristics of the subjects including player positions and leg dominancy are presented in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Demographic characteristic of the players who played on dirt field (DF) and artificial turf field (ATF). ::: -------------------------------------------------------------------------- DF\ ATF\ (95% CI) (95% CI) -------------------------- ----------------------- ----------------------- Number of cases 252 216 Age(years) 27.04(26.50-27.58) 27.96(27.36-28.56) Mean weight(kg) 72.79(72.01-73.58) 73.28(72.55-74.00) Mean height(cm) 171.85(171.27-172.43) 172.31(171.72-172.90) **% of dominant leg**  • left 13(9-18) 13(8-17)  • right 70(62-74) 68(61-74)  • bilateral 17(12-22) 19(14-24) **% of player position**  • goalkeeper 14(9-18) 15(10-19)  • defender 34(27-39) 27(20-32)  • midfielder 33(27-38) 35 (28-41)  • forward 19(14-24) 23(17-28) -------------------------------------------------------------------------- ::: The overall incidence of match injuries for subjects was 36.9 injuries/1000 player hours on DF and 19.5 on ATF (incidence rate ratio 1.88; 95% CI 1.19-3.05). This differences was statistically significant (P \< 0.004) (Table [2](#T2){ref-type="table"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Exposure, number of injuries and incidence of injuries on dirt field (DF) and artificial turf field (ATF). ::: -------------------------------------------------------------------------------------- Exposure, playing hours Number of injuries **Injury incidence**,\ n/1000 playing hours (95% CI) ------- ------------------------- -------------------- ------------------------------- DF 1897 70 36.9(28.7-46.6) \* ATF 1378 27 19.5(12.9-28.5) \* Total 3275 97 29.6(24.0-36.1) -------------------------------------------------------------------------------------- \*Differences P value \< 0.004 ::: In terms of severity of injuries in DF cohort most injuries were slight and minimal but in AT cohort the most injuries were mild (Table [3](#T3){ref-type="table"}). Incidence of slight and minimal injuries was significantly greater on DF than on ATF (Table [3](#T3){ref-type="table"}). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Incidence of match injuries sustained on dirt field (DF) and artificial turf field (ATF) a function of injury severity ::: ----------------------------------------------------------------------------------------- DF ATF ---------------------- ------------------- ------------------ ------------------ -------- slight (0 days) &\ 13.70(8.95-20.08) 2.90(0.79-7.43) 4.72(1.63-18.61) 0.0009 Minimal (1-3 days) Mild (4-7 days) 11.07(6.85-16.92) 9.43(5.02-16.13) 1.17(0.56-2.55) 0.66 Moderate (8-28 days) 8.43(4.82-13.69) 5.80(2.50-11.43) 1.45(0.58-3.92) 0.39 Severe (\>28 days) 2.10(0.57-5.39) 1.45(0.17-5.24) 1.45(0.38-22.07) 0.35 ----------------------------------------------------------------------------------------- ::: The most common injured part on ATF was ankle (25.9%) and on DF was knee (24.3%) (Table [4](#T4){ref-type="table"}). ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Location of injuries sustained by the subjects who played on dirt field (DF) and artificial turf field (ATF). ::: \% of total injuries ------------------- --------------------------- ---------------------- ------ **Head and Neck** Head/face/Neck 7.4 5.7 **Upper limbs** Shoulder/clavicle 3.7 2.9 Upper arm 3.7 0 Elbow, Forearm, Hand 3.7 10.0 **Trunk** Trunk 3.7 14.3 **Lower limbs** Hip/groin/thigh 14.8 15.7 Knee 18.5 24.3 Lower leg/Achilles tendon 14.8 10 Ankle 25.9 14.3 Foot/toe 3.7 2.9 **Total** 100 100 ::: The most common injury type in the DF cohort was skin injuries (e.g. abrasion and laceration) and in the AT cohort was sprain and ligament injury followed by haematoma/contusion/bruise.). Laceration and skin lesion was significantly greater on DF than on AT (P \< 0.0003) (Table [5](#T5){ref-type="table"}). Incidences of other injury types were not significantly greater on DF than on ATF (Table [5](#T5){ref-type="table"}). Most injuries sustained by all subjects were acute (ATF89%, DF 91%) and resulted from a player-to-player contact (ATF59.2%, DF 51.4%) (Table [6](#T6){ref-type="table"}). Incidence of acute injuries was significantly higher on DF than on ATF(Table [6](#T6){ref-type="table"}). ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Characteristics of injuries incurred during matches, Dirt field (DF), Artificial turf field (ATF) ::: DF ATF ------------------------------- ---------------------- -------------------- -------------------- -------- Fractures and bone stress 2.63\[0.85-6.15\] 1.45\[0.17-5.24\] 1.81\[0.29-19.07\] 0.5 Joint (non-bone) and ligament 8.96\[5.22-14.34\] 6.53\[2.98-12.39\] 1.37\[0.57-3.49\] 0.45 Muscle and tendon 2.63\[0.85-6.15\] 2.90\[0.79-7.43\] 0.90\[0.19-4.57\] 0.88 Contusions 5.79\[2.89-10.37\] 5.07\[2.04-10.46\] 1.14\[0.40-3.47\] 0.79 Laceration and skin lesion 16.34\[11.10-23.19\] 3.62\[1.17-8.46\] 4.5\[1.73-14.83\] 0.0003 Other injuries 0.52\[0.01-2.93\] 0\[0-2.67\] \* \-\-\-\-\-- Total 36.9(28.7-46.6) 19.5(12.9-28.5) 1.88 (1.19-3.05) (\*)one-sided ::: ::: {#T6 .table-wrap} Table 6 ::: {.caption} ###### The incidence of match injuries with different natures and causes on dirt field (DF) and artificial turf field (ATF). ::: DF ATF ------------------------------- ---------------------- ---------------------- ------------------- ------- **Nature of onset** • Acute 32.68\[25.05-41.89\] 16.69\[10.58-25.04\] 1.95\[1.19-3.31\] 0.004 • Gradual 4.21\[1.82-8.30\] 2.90\[0.79-7.43\] 1.45\[0.38-6.59\] 0.56 **Cause of onset** • Not any contact 15.8 (10.6-22.5) 7.2 (3.4-13.3) 2.17 1.03-4.99) 0.02 • Contact with another player 18.9 (13.2-26.2) 11.6 (6.6-18.8) 1.63 (0.88-3.15) 0.09 • Contact with the ball 0.5 (0.01-2.9) 0 (0-2.6)\* \-\-\-\-- • Contact with other objects 1.5 (0.3-4.6) 0.7 (0.01-4) 2.17(0.17-144) 0.55 (\*)one-sided ::: We did not find any significant differences between percentage of injury in different player positions among players who played on DF and those who played on ATF. In general 64.9% of the injured part was in the right side and 26.8% in the left side (8.2% not applicable). The differences between two groups in this regards was not significant. About 45.7% of the injuries in DF cohort occurred in the first halftime and 54.3% in the second halftime. On AT 37% of the injuries occurred in the first halftime and 63% in the second halftime. The rate of injury in the second halftime was higher than first halftime in both DF cohort and AT cohort however these differences were not statistically significant. Discussion ========== The major outcome of this study is that the incidence of match injuries of players who play on DF is higher (about two times) than this rate of players who play on ATF. However, this difference was mainly because of acute slight and minimal injuries and laceration and skin lesions. The overall incidence of match injuries on AT for men in this study was 19.5 that is consistent with values reported by Ekstrand, et al. (2006) \[[@B6]\] for injury rate on AT (19.6). However Ekstrand, et al. \[[@B6]\] only included traumatic match injuries. Match injury incidence rate on AT reported in this study is lower than those reported by Fuller, et al (2007) \[[@B4]\] and Fuller C.(2006) \[[@B11]\]. Fuller, et al (2007) \[[@B4]\] reported a higher rates of football match injuries on AT among American college and university football teams at 25.43. Similarly Fuller C. (2006) reported incidence rate on AT for the FIFA men\'s U-17 world cup tournaments as 26 \[[@B11]\]. This might be because our subjects were amateur and low level players. The incidence of football injuries reported to be higher in the professional players compare to the low-level players \[[@B12],[@B13]\]. Our reported rate is consistent with those injury rates reported for amateur and low level football players 11.7 to 21.7 \[[@B13]-[@B17]\]. We not found equivalent published data to compare results of this study to the incidence of football injury on DFs. The incidence rate of injuries on DF in this study (36.9) was about two-times more than this rate on ATF. This rate (36.9) is higher that reported football match injury rate on gravel at 20 per 1000 hours of game time \[[@B2],[@B8]\]. This rate also is higher than football injury incidence rate on ATF reported by other studies (19.6 - 26) \[[@B4],[@B6],[@B11]\]. Moreover, this rate is higher than football injury incidence rate reported in the literature that is ranged from 11.7 to 35.3 per 1000 hours \[[@B9],[@B12],[@B13],[@B18]\]. This difference was mainly because the rate of slight and minimal injuries was high on DF in the present study. Incidence of slight and minimal injuries was significantly greater on DF (13.70) than on AT (2.90). This rate on DF was higher that rates reported by other studies on AT (4.97) \[[@B6]\] and (8.34) \[[@B4]\]. The stiffness of a surface and the friction between surface and shoe have been reported as important football injury risk factors \[[@B2],[@B3]\], which might explain why the frequency of injury was significantly higher on DF than on ATF in this study. Uneven surface also might be a possible risk factor for football injuries \[[@B19]\]. However all football fields in this study were not uneven. Players wear shoes with no spikes at the bottom to play on dirt field but they wear special shoes with spikes to play on artificial turf. More study is needed to evaluate the effects of shoes on injury rates in different fields. The higher rate of injury and especially slight and minimal injury in this study might be partly because of the employed injury definition. New definition of injury presented by the international consensus statement for epidemiological studies of injuries in football \[[@B10]\] is not exclusively based on time loss or medical attention that might lead to record more of slight and probably minimal injuries. Some abrasion injuries may not lead to time loss and therefore, are not included in some studies. Generation of AT also might affect the incidence of abrasion injuries. Some may discuss that the higher incidence rate of injuries on dirt field was mainly because of laceration and skin lesions and slight and minimal injuries are not as important as more severe injuries. If we do not include these mild injuries the injury rates on dirt field and artificial filed are not significantly different. In present study the injury incidence on AT was peaked for injuries of mild (4-7 days) severity (9.43) that is in contrast with both Ekstrand et al\[[@B6]\] and Fuller, et al \[[@B4]\] studies. In Ekstrand et al\[[@B6]\] study incidence of injuries was peaked for injuries of moderate (8-28 days) severity (6.35). In Fuller, et al \[[@B4]\] study the incidence of injury decreased with increasing injury severity on both AT and grass. In present study the injury incidence on DF decreased with increasing injury severity. Joint (non-bone)/ligament injuries to the lower limb was reported as the most common combination of injury type and location sustained on AT in this study. This is similar to results reported in previous studies of football injuries on both grass and AT surfaces \[[@B4]\]. However, in present study on DF laceration and skin lesion was the most common injuries followed by Joint (non-bone)/ligament injuries. An important cause of football injuries is contact with another player that reported to be varied from 41% to74% \[[@B12]\]. The results of this study are consistent with this data in that most injuries sustained by all subjects were resulted from a player to player contact (AT 59.2%, DF 51.4%). As a limitation we only recorded match injuries. More study is needed to evaluate the incidence rate and natures of training injuries on DF. As another limitation we did not analysed the content of the DF. Quality of DFs are not the same. DF surface might be variable surface from field to field and climate to climate. It would be very difficult to find two similar player groups who play on DF and grass because those players who play on DF normally are amateur and in lower age and playing levels in terms of technique and tactic compare to those who play on grass. However as it was mentioned previously we believed the levels of players who played on DF and AT in this study was about the same. To prevent skin abrasions on dirt fields the skin area that may potentially receive trauma need to be protected. Protective equipments such as sliding pads, long-sleeve shirts, long socks, \"biker\" shorts might be used to protect the vulnerable exposed areas \[[@B20]\]. Dirt fields should regularly be made re-flat because uneven surface might be a possible risk factor for football injuries on these fields. Conclusions =========== Results of this study suggest that the incidence of match football injuries on DF are higher (about two times) than this rate on AT. This difference was mainly because of acute minimal injuries and skin lesions. Therefore, in regions in which maintaining natural gross is too expensive AT might be an alternative to reduce the rate of football injuries. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= RK and FH contributed to the study concept and, with VZ, the study design. RK, FH and HH were responsible for the acquisition of data. FH and HH contributed to the analysis and interpretation of the data. RK and VZ drafted the manuscript. RK, FH and HH critically revised the manuscript. All of the authors approved the final version of the manuscript submitted for publication. Acknowledgements ================ This study was funded by the Sports Medicine Research Centre, Faculty of Medicine, Tehran University of Medical Sciences

PubMed Central

2024-06-05T04:04:19.496152

2011-2-9

PMC3053232

Background ========== Chronic inflammations, such as periodontitis, are suggested to be risk factors for the development of cardiovascular diseases \[[@B1]\]. It has been suggested that the total pathogenic burden from the oral cavity, and possibly also from the gut, correlates with disease markers of atherosclerosis \[[@B2]\]. Periodontitis is a bacteria-induced inflammatory condition that causes degradation of the tooth-supporting tissues, bone and connective tissue \[[@B3],[@B4]\]. Bioactive molecules released from pathogenic microorganisms located in the subgingival biofilm cause imbalance in the inflammatory response, which results in loss of the tooth-supporting tissues \[[@B5]\]. For the host to maintain homeostasis within the periodontal tissues, the immune response system contributes to controlling the microbial colonization and invasion \[[@B6]\]. This immune response includes local and systemic production of antibodies induced by antigens from the microorganisms that are localized in this biofilm \[[@B7]\]. There are more than 700 different microbial species found in the oral cavity of humans \[[@B8]\]. A recent report using the pyrosequencing methodology to analyze the composition of the oral microbiota indicate a substantial increase in that number \[[@B9]\]. Among the different species found, *Aggregatibacter actinomycetemcomitans*is a bacterium associated with aggressive forms of periodontitis, and it produces a leukotoxin that specifically affects human leukocytes \[[@B10]\]. Individuals infected with a specific, highly leukotoxic clone (JP2) of this bacterium have a significantly increased risk for periodontitis \[[@B11]\]. The proinflammatory response induced by the leukotoxin is a cellular response associated with the pathogenesis of periodontitis \[[@B10],[@B12],[@B13]\] and atherosclerosis \[[@B14]\]. The proportion in a population that harbor *A. actinomycetemcomitans*varies depending on geographic origin and periodontal condition of the subjects \[[@B10]\]. It has been shown that systemic leukotoxin-neutralization is correlated to the presence of this bacterium in the oral subgingival biofilm \[[@B15]-[@B17]\]. Data from a previous study showed that women with systemic neutralizing capacity against *A. actinomycetemcomitans\'*leukotoxin had a significantly decreased incidence of stroke \[[@B18]\]. This systemic neutralizing capacity has been shown to correlate (p-value \< 0.001) to the presence of leukotoxin-specific antibodies, as well as to antibodies against whole *A. actinomyctemcomitans*bacteria \[[@B19]\]. We hypothesized that a virgin *A. actinomycetemcomitans*infection late in life might be a risk factor for stroke and contribute to the negative association between stroke and the presence of these neutralizing antibodies. The aim of the present study was to analyze if the presence of systemic immunoreactivity to *A. actinomycetemcomitans*leukotoxin also interferes with the incidence of a future myocardial infarction (MI). Methods ======= Study population ---------------- The study population was derived from the Northern Sweden Health and Disease Study (NSHDS), which consists of three sub-cohorts: The Västerbotten Intervention Programme (VIP) \[[@B20]\], the WHO\'s Multinational Monitoring of Trends and Determinants in Cardiovascular Disease (MONICA) study in northern Sweden \[[@B21]\] and the Mammography Screening Project (MSP) \[[@B22]\]. Both VIP and MONICA are health examination programmes for CVD and diabetes. Participation rates were 59 and 77%, respectively. The VIP was designed to be as similar as possible to the MONICA study. In order to increase the number of female cases, participants in the MSP were included in sex specific analyses. The participation rate in the MSP was 85% in the screening phase, of which 57% donated blood samples. By December 31, 1999 approximately 73,000 unique subjects had been screened in these 3 sub-cohorts in NSHDS. Incident cases and matching controls were identified through 13 years of follow-up (1985-1999) from the Västerbotten Intervention Program and the Multinational Monitoring of Trends and Determinants in Cardiovascular Disease (MONICA) study. Study participants were from the Västerbotten and Norrbotten counties in northern Sweden. Participants with a history of MI, stroke or cancer were excluded from this study. Participants were followed from baseline examination until first MI or death. There was an average time period of 4 years between the time of inclusion and the MI event. Fatal and nonfatal cases of MI occurring from October 1, 1994 to December 31, 1999 were identified through the Northern Sweden Monica Incidence Registry \[[@B23]\]. Throughout the follow-up period, 532 incident first events of MI (cases) were identified. For each case, two controls were individually matched for age, sex and ± 4 months of case occurrence. Thus, the study population consisted of 1,532 participants, 532 cases (382 men and 156 women) and 1,000 sex and age-matched controls (706 men and 294 women), aged 30-77 years at baseline. This study population has previously been described in detail \[[@B24]\]. The study was approved by the Ethics Committee of Umeå University and was conducted in accordance with the Helsinki Declaration. All participants gave informed consent. Leukotoxin-neutralization assay ------------------------------- The *A. actinomycetemcomitans*leukotoxin-neutralizing capacity in the plasma samples was detected as a reduction of leukocyte damage and subsequent leakage of lactate dehydrogenase (LDH) upon exposure to purified leukotoxin, as described previously \[[@B18]\]. This assay quantifies the activity of the LDH enzyme and does not allow freezing and thawing of the supernatants. The neutralization assay also limits the possible influence from LDH present in the different plasma samples from the study population. Briefly, human polymorphonuclear leukocytes (PMNs) were isolated from human peripheral blood as described previously \[[@B25]\]. The isolated PMNs were suspended in RPMI 1640 (Sigma-Aldrich, St Louis, MI, USA) with 10% fetal bovine serum (FBS) (Sigma-Aldrich) at a density of 3 × 10^6^cells/ml. The blood was taken from donors visiting the University Hospital blood bank in Umeå, Sweden. Informed written approval was given by all subjects, and authorization for the study was granted by the Human Studies Ethical Committee of Umeå University, Sweden (§67/3, dnr 03-019). For detection of leukotoxin-neutralizing capacity, purified leukotoxin (450 ng/mL) \[[@B26]\] was mixed with each plasma sample (10%) in RPMI 1640. One hundred μl portions of the plasma-leukotoxin mixtures were added in duplicate to a 96 well culture plate (Nunc, Roskilde, Denmark) and incubated for 15 min at room temperature. Then 50 μl of PMN was added and the mixtures were incubated for 60 min at 37°C in 5% CO~2~. Activity of the released LDH into the culture supernatant was quantified as described previously \[[@B25]\]. Plasma samples that inhibited the leukotoxin-induced LDH release by ≥50% were classified as positive and were further analyzed in the assay diluted to 1% of the final volume. Plasma without leukotoxin-neutralization capacity was classified as \"*null\"*, plasma that neutralized leukotoxin at 10% plasma concentration but not at 1% was classified as \"*low*\", and plasma that neutralized the leukotoxin at both 10% and 1% concentrations was classified as \"*high\"*. Statistical anaylses -------------------- The Mantel-Haenszel χ^2^-test for trend was used to analyze the association between age and antibodies. To investigate if the presence of leukotoxin antibodies (categorized into null, low or high) affects the risk of having an MI, conditional logistic regression appropriate for the matched design, was used. A multivariable model was used to adjust for smoking, self-reported diabetes, systolic blood pressure and apoB/apoA1. Results are presented as p-values, odds ratios (ORs) and corresponding 95% confidence intervals (CIs). No correction for multiple testing was performed. SAS version 9.1 was used for the statistical analyses. Results ======= Prevalence of leukotoxin-neutralizing capacity ---------------------------------------------- The study population was classified into 4 different age groups, and the distribution in relation to age and gender is shown in Figure [1](#F1){ref-type="fig"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Distribution of men and women in the age groups for the whole study cohort, i.e. including both cases and referents**. :::  ::: Among the 1,532 analyzed plasma samples, 817 (53.3%) could neutralize *A. actinomycetemcomitans*leukotoxicity in the neutralization assay. Further dilution of the plasma samples resulted in loss of the capacity to neutralize leukotoxin in 526 of these samples, and they were classified as samples with low neutralizing capacity. The 291 samples that neutralized leukotoxin also at the higher dilution were classified as high. The distribution of the study population in relation to their systemic capacity to neutralize leukotoxin was 46.7% negative, 34.3% low and 19.9% high. There were no significant differences between men and woman in capacity to neutralize the leukotoxin (Figure [2](#F2){ref-type="fig"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Proportion of men and women with different systemic capacity to neutralize *A. actinomycetemcomitans*leukotoxin**. :::  ::: Prevalence and age ------------------ The proportion of subjects with capacity to neutralize leukotoxin increased with increasing age (Figure [3](#F3){ref-type="fig"}). This age-related increase was significant for men (p-values ≤ 0.001) but not for women (p-values = 0.170). In order to avoid combinations with no or few observations, the two lowest age groups (29-44 and 45-54) were merged for women, and the two highest age groups (55-64 and 65-77) were merged for men in the formal analysis. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Proportion of men and women (low + high) with systemic *A. actinomycetemcomitans*leukotoxin-neutralizing capacity in relation to age**. The distributions of men and women in the different groups were: 29-44 yr, 179 men and 21 women; 45-54 yr, 397 men and 114 women; 55-64 yr, 503 men and 221 women; and 65-77 yr, 6 men and 94 women. :::  ::: Prevalence in relation to incidence of MI ----------------------------------------- Women with low capacity to neutralize leukotoxin had a significantly (p-value = 0.031) higher incidence of MI than women without the capacity to neutralize leukotoxin (Table [1](#T1){ref-type="table"}). The odds ratio (OR) of having an MI in this group was 1. 8 (95% CI: 1.13-2.8). No other significant differences were seen between the incidence of MI and the different groups classified in relation to systemic leukotoxin neutralization and gender. After adjustments for other known risk-factors for MI (smoking, diabetes, systolic blood pressure and ApoB/ApoA1) there were no significant differences (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Proportion of controls and cases with plasma that neutralized leukotoxic activity. ::: *[Men & Women]{.underline}* [Neg]{.underline} [Low]{.underline} [High]{.underline} ----------------------------- ----------------------- ----------------------- ------------------------ **Control** 47.1% (471) 33.3% (333) 19.6% (196) **Case** 45.9% (244) 36.3% (193) 17.9% (95) **OR** 1.1 (0.8-1.4) 0.9 (0.7-1.2) **adjusted OR** 1.0 (0.7-1.4) 1.3 (0.9-2.0) ***[Women]{.underline}*** **[Neg]{.underline}** **[Low]{.underline}** **[High]{.underline}** **Control** 51.7% (152) 28.6% (84) 19.7% (58) **Case** 42.9% (67) 39.7%\*(62) 17.3% (27) **OR** 1.8\* (1.1-2.8) 1.0 (0.8-1.3) **adjusted OR^\#^** 1.2 (0.5-2.9) 1.4 (0.5-4.3) ***[Men]{.underline}*** **[Neg]{.underline}** **[Low]{.underline}** **[High]{.underline}** **Control** 45.2% (319) 35.3% (249) 19.5% (138) **Case** 47.1% (177) 34.8% (131) 18.1% (68) **OR** 0.9 (0.7-1.2) 0.8 (0.6-1.2) **adjusted OR^\#^** 1.0 (0.7-1.4) 1.3 (0.9-2.1) Significant differences between controls and cases (\*), number of subjects (n) and the odds ratio (OR) of having an MI compared with the antibody-negative group (null) are indicated. ^\#^) Adjusted for smoking, self-reported diabetes, systolic blood pressure and ApoB/ApoA1 ::: Discussion ========== Results from the present study showed that 53.3% of the plasma samples from a Swedish adult cohort of 1,532 subjects had the capacity to neutralize *A. actinomycetemcomitans*leukotoxin. It has recently been demonstrated that this leukotoxin-neutralizing capacity correlates with the presence of leukotoxin-specific antibodies \[[@B19]\]. The high prevalence of subjects with this leukotoxin-neutralizing capacity was not expected, however, in line with results from some previous studies \[[@B19],[@B27]\]. Both of these studies were based on Swedish study populations from a similar age group as examined in the present study. In one of these reports the study population consisted of a total of 197 subjects from a case control study of myocardial infarction and matched healthy controls, and the prevalence of systemic leukotoxin-neutralizing capacity was 57% without significant differences between cases and controls \[[@B19],[@B28]\]. The other study consisted of 50 subjects with periodontitis and 41 healthy referents, and in this population the prevalence of systemic leukotoxin-neutralizing capacity was 45% without significant differences between the two groups \[[@B27]\]. Another study showed lower prevalence (15.2%) of leukotoxin-neutralizing capacity \[[@B18]\]. In this study \[[@B18]\] a target cell line (HL-60) with lower sensitivity to leukotoxin than the PMNs was used in the neutralization assay \[[@B29]\], which resulted in a need for enhanced leukotoxin concentration to obtain cell lyses and subsequently more antibodies for neutralization. This difference in leukotoxin sensitivity makes this assay with PMNs more sensitive to detect leukotoxin neutralization than the previous used assay with HL-60 cells \[[@B15]\]. The proportion of samples (19.9%) with high leukotoxin-neutralizing capacity in the present study might be comparable with the results from Johansson et al., 2005 \[[@B18]\]. We have previously shown that the systemic leukotoxin-neutralizing capacity correlated to a decreased incidence of stroke in women \[[@B18]\]. The mechanisms behind this phenomenon are not known. We speculate that a virgin infection with *A. actinomycetemcomitans*in middle-aged and elderly subjects might be a risk factor for stroke and that the capacity to neutralize leukotoxin might be protective. The leukotoxin has been shown to induce a rapid proinflammatory reaction in human macrophages, already at a ratio of 1 bacterium/macrophage \[[@B13]\], and therefore the leukotoxin is a possible risk factor in atherosclerosis. The common etiology of both stroke and MI with inflammatory processes and atherosclerosis \[[@B14]\], indicates that the presence of systemic leukotoxin-neutralizing capacity also might interfere with the incidence of MI. The results of the present study showed that systemic presence of leukotoxin-neutralizing capacity did not affect the incidence of MI, except for women classified as low for this neutralizing capacity. Our main finding refutes the hypothesis that systemic immunoreactivity to leukotoxin has a protective effect against MI. The significant association found for women might be an effect of multiple testing or a type-1 error, but further studies on this finding are warranted to confirm this association. The periodontal status of the analyzed subjects is not known, but it could be expected to be in line with a similar recently examined Swedish population \[[@B28]\]. In that population a correlation between MI and periodontitis was observed. In addition, both periodontitis and MI correlated to the presence of systemic immunoreactivity against *Porphyromonas gingivalis*but not against *A. actinomycetemcomitans*\[[@B19],[@B28]\]. The proportion of subjects with systemic neutralizing capacity to leukotoxin increased with increasing age, significantly for men but not for women. This age-related increase is in line with previous findings \[[@B19]\] and indicates that a virgin infection with *A. actinomycetemcomitans*can take place late in life. A virgin *A. actinomycetemcomitans*infection might decrease the risk for stroke in middle-aged and elderly subjects without systemic leukotoxin-neutralizing capacity \[[@B18]\]. Even though the role of leukotoxin-neutralizing antibodies in the pathogenesis of periodontal disease is unknown \[[@B7]\], the antibodies might help mitigate the systemic effects of *A. actinomycetemcomitans*infections. The leukotoxin produced by *A. actinomycetemcomitans*is a unique virulence factor with the capacity to cause a rapid proinflammatory reaction \[[@B10]\]. However, to fully investigate a potential role of leukotoxin in the pathogenesis of stroke, the presence of systemic leukotoxin antibodies has to be analyzed both before and after the disease incidence. We therefore still look at the results that showed a negative correlation between systemic leukotoxin antibodies and stroke as preliminary \[[@B18]\]. Conclusions =========== The results from the present study do not support the hypothesis that systemic leukotoxin-neutralizing capacity can decrease the risk for MI. In addition, the prevalence of systemic *A. actinomycetemcomitans*leukotoxin-neutralizing capacity is high (53.3%) in adults from northern Sweden. The prevalence of leukotoxin-neutralizing capacity increased with increasing age, significantly for men but not for women. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= AJ conceptualized the study, conducted the analyses, and wrote the first draft of the manuscript. A-MÅ and AJ performed the analyses and made a first draft of the result calculations. J-HJ and KB planned and supervised data collection. GH and IJ coordinated data collection, and ME performed the statistical calculations. All approved the final version of the submitted manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2334/11/55/prepub> Acknowledgements ================ This study was supported by the Research Fund (TUA) of Västerbotten County, Sweden. Thanks to the Research Unit at the Department of Medicine and Geriatrics, Skellefteå County Hospital, Sweden for supporting this study.

PubMed Central

2024-06-05T04:04:19.499359

2011-3-1

PMC3053233

Introduction ============ In dairy herds, paratuberculosis, a worldwide disease caused by *Mycobacterium avium*subspecies *paratuberculosis*(*Map*), provokes decreases in milk production, drops in carcass slaughter value, and premature culling. It is important to predict as early as possible after *Map*first introduction in a dairy cattle herd whether infection is likely to fade out or to persist. This prediction could then be used to inform the implementation of control methods. Ideally, a point of no return should be identified, after which *Map*will persist and spread in the herd without control, i.e. when control actions must ideally be implemented. However, due to the long incubation period \[[@B1]\] and the low sensitivity of available diagnostic tests \[[@B2]\], studying the infection dynamics in the field is nearly impossible. Therefore, modelling is used to better understand *Map*spread within a herd. Stochastic models are particularly suitable for investigating the likelihood of persistence versus fade-out of infection. Three stochastic models of *Map*transmission in dairy herds have been published \[[@B3]-[@B5]\]. However, these models neither take *Map*survival in the environment nor all relevant *Map*transmission routes into account, and therefore are not suitable for examining this persistence (see \[[@B6]\] for recent and detailed review of the models). Yet, the survival of *Map*in the environment can result in a delay between shedding by infectious animals and infection of susceptible animals. As a result of contamination of the farm environment, infection of susceptible animals can occur in the absence of infectious animals \[[@B7],[@B8]\]. To study fadeout and persistence of *Map*in a dairy herd, we propose a new stochastic model that includes transmission via the environment. Furthermore, we have included calf-to-calf transmission, which has been demonstrated recently \[[@B9]\]. Hence, transmission routes are: vertical, horizontal via the ingestion of contaminated colostrum or milk, or horizontal via the ingestion of adult or calf faeces. Our model accounts for all of these transmission routes, thus rendering it possible to identify which routes contribute the most to *Map*spread in the modelled dairy herd. In the model, we assume that no further infected animals are introduced to avoid the possibility that persistence of *Map*might be due to continuous reintroductions (i.e. no fadeout being possible). Such a situation will be typical for herds with very low yearly purchase rates (e.g. dairy herds in Brittany without any fattening activity; \[[@B10]\]) or in the context of certification, when only certified animals are purchased (with a very low risk of being infected; \[[@B11]\]). In Europe, control of *Map*introduction into cattle herds has indeed priority over control of within-herd *Map*spread. Materials and methods ===================== We develop a model of *Map*spread within a dairy cattle herd initially naive towards *Map*infection, following the introduction of a single infected cow. We use this model to predict *Map*spontaneous fadeout or persistence as early as possible after *Map*introduction, before any control measure is implemented. Model description ----------------- A discrete time compartmental model is developed to represent *Map*spread in a dairy cattle herd. We couple a model that simulates the population dynamics within a dairy herd and explicitly represents animal housing facilities with an epidemiological model of *Map*transmission. A time step of one week is chosen as the longest possible to allow the different transmission routes and calf exposure in housing facilities to be represented. A stochastic model is used in order to study the chance of fadeout of the disease *versus*persistence probability. Because of the slow progression of paratuberculosis, we choose to study the infection over a 25-year period. The model is implemented with Scilab 5.1 \[[@B12]\]. ### Population dynamics The population dynamics only considers characteristics related to *Map*transmission. Contacts between susceptible animals and any environment contaminated by shedding animals depends on the time spent by animals on farm, the time spent in individual and collective pens, and possible shared environments. An ageing process occurs before the infection process at each time step. An exit rate for mortality, sale, and culling is defined per age class (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Parameters for herd management and population dynamics used in a *Mycobacterium avium paratuberculosis*infection dynamics model within a structured dairy herd ::: Notation Value Definition Source ---------- -------------------- -------------------------------------------------------------------------------------------- --------------- *σ~B~* 0.07 Mortality rate of calves at birth a, \[[@B32]\] *σ~m~* 0.206 Exit rate of male calves, weeks 2 to 4 (per week) *σ~C1~* 0.015 Death rate of female calves, weeks 1 and 2 (individual housing facilities) (per week) \[[@B32]\] *σ~C2~* 0.0035 Death rate of female calves, weeks 3 to weaning (collective housing facilities) (per week) \[[@B33]\] *σ~C3~* 0.00019 Death rate of heifers from weaning to first calving (per week) b *σ~h~* 0.11 Sale rate of bred heifers 10 weeks before 1^st^calving b *σ~Ai~* 27, 25, 31, 31, 62 Yearly culling rate of cows in parity 1, 2, 3, 4 and above 5 respectively (%) a, \[[@B34]\] *m* 2 Maximal age in individual pen (weeks) \[[@B13]\] *w* 10 Weaning age (weeks) \[[@B13]\] *y* 52 Age when entering the young heifer group (weeks) *nb* 2 Number of neighbours for a calf in an individual pen b *h* 91 Age at first artificial insemination (weeks) a *cal* 130 Age at first calving (weeks) a,b *cci* 56.3 Calving-to-calving interval (weeks) a,b *b* 5 Quantity of colostrum fed to calves (L/day for 3 days) b *d* 7 Quantity of milk fed to calves after 3 days (L/day/calf) b *prop* 0.85 Proportion of lactating cows a *ε* 25 Quantity of milk or colostrum produced (L/day/cow) a *f~1~* 0.5 Quantity of faeces produced by a non-weaned calf (kg/day) b *f~2~* 5.5 Quantity of faeces produced by a weaned calf (kg/day) b *f~Y~* 10 Quantity of faeces produced by a heifer (kg/day) b *f~A~* 30 Quantity of faeces produced by a cow (kg/day) b *Graz* \[14-46\] Grazing period (1 being the first week of the year) b *K~c~* 110 Number of cows above which the heifer selling rate increases \- ^a^Agricultural statistics ^b^Expert opinions ::: In Europe, dairy herds generally are structured in groups, the younger animals being separated from the older ones \[[@B13]\]. Here, group definition accounts for animal housing and management, and the maximal age (*u*, Table [2](#T2){ref-type="table"}) at which an animal is susceptible (Figure [1](#F1){ref-type="fig"}). Therefore, contacts between susceptible animals and contaminated environments can be assessed. Calves younger than one year of age are either in individual pens (from birth to *m*), in collective pens before weaning (from *m*to *w*), or in collective pens after weaning (from *w*to *y*). Calves in individual pens have limited contacts with the faeces of calves from contiguous pens (*nb*). Such a calf housing facility management follows European recommendations concerning animal welfare and social contacts (Council Directive 97/2/EC of 20 January 1997 amending Directive 91/629/EEC laying down minimum standards for the protection of calves ) and reflects the most common calf management in Europe \[[@B13]\]. After 1 year of age, the heifers are divided into 2 groups: from 1 year of age to 1^st^artificial insemination (AI) at age *h*, and from 1^st^AI to 1^st^calving at age *cal*. Cows are all gathered in the same batch assuming they are not susceptible. Parities are considered as the culling rate is higher for older cows and to account for age in the progress of *Map*infection. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Parameters for infection and transmission used in a *Mycobacterium avium*subsp. *paratuberculosis (Map)*infection dynamics model within a structured dairy herd\* ::: Notation Value Definition Source ---------- ------------------ ------------------------------------------------------------------------------------------ ------------------- *p~X~* Probability of *in utero*transmission for cow in health state *X* \[[@B24],[@B35]\] *p~L~*= 0.149 *X*= latently infected (*L*) *p~Is~*= 0.149 *X*= subclinically infected (*I~S~*) *p~Ic~*= 0.65 *X*= clinically affected (*I~C~*) *u* 52 Maximal age in the susceptible compartment (weeks) \[[@B15],[@B36]\] *h* 0.1 Susceptibility follows an exponential decrease exp(-h(age-1))) \[[@B14]\] *v~X~* Mean time spent in health state *X*(weeks) *v~T~*= 25 *X*= transiently infectious (*T*) \[[@B9]\] *v~L~*= 52 *X*= latently infected (*L*) \[[@B2],[@B16]\] *v~Is~*= 104 *X*= subclinically infected (*I~S~*) \[[@B37]\] *v~Ic~*= 26 *X*= clinically affected (*I~C~*) a *sh~X~* Probability of shedding in colostrum or milk for a cow in health state *X* \[[@B38],[@B39]\] *sh~L~*= 0 *X*= latently infected (*L*) *sh~Is~*= 0.4 *X*= subclinically infected (*I~S~*) *sh~Ic~*= 0.9 *X*= clinically affected (*I~C~*) *α* 10^6^ *Map*infectious dose \[[@B40]\] *β~l~* 5 × 10^-4^× 7 Transmission rate if ingestion of an infectious dose (per week) b *β~c~* 5 × 10^-5^× 7 Transmission rate if one infectious dose is present in the local environment (per week) \[[@B9]\] *β~g~* 9.5 × 10^-7^× 7 Transmission rate if one infectious dose is present in the global environment (per week) \[[@B9]\] *β~o~* 5 × 10^-6^× 7 Transmission rate if one infectious dose is present on pasture (per week) b *g~X~* Decrease in milk production for cattle in health state *X*(per week) \[[@B41]\] *g~Is~*= 2.5 × 7 *X*= subclinically infected (*I~S~*) *g~Ic~*= 4 × 7 *X*= clinically affected (*I~C~*) *μ~k~* Removal rate of *Map*from environment *k* \[[@B7],[@B8]\] *μ~g~*= 0.4  all the environments (per week) *μ~ip~*= 0.67  individual pens (when empty) *μ~cp~*= 0.17  collective pens (when empty) \*The values of the parameters in the epidemiological model (Table 2) are estimates based on experimental data reported in the literature. ^a^Expert opinions ^b^Parameters\' values are assumed ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Population dynamics in a closed dairy herd and flow diagram of *Mycobacterium avium*subsp**. ***paratuberculosis (Map)*****infection dynamics model, representing infection states, transitions between states,and origin of contamination of the local and whole farm environments**. Host health states are: *S*= susceptible; *R*(in grey square) = resistant; *T*= transiently infectious; *L*= latently infected; *I~S~*= subclinically infected; *I~C~*= clinically affected. Environment states are: *E~l~*= indoor environment in housing *l*, with *l*= 1 to 6 (1 for calves in individual pens, 2 for calves in collective pens before weaning, 3 for calves in collective pens after weaning before 6 months of age or during winter season, 4 for young heifers during winter season, 5 for heifers during winter season, and 6 for adults during winter season); *E~g~*= environment of the whole farm; *E*^*out*^= outdoor environment of calves when they are grazing. The population dynamics has to be read vertically. Moreover, *n*= number of individual pens; **Z~1~**to **Z~3~**= transmission functions for horizontal infection; *t*= time; G = grazing season; P~i~= cows in parity *i*; dotted arrows: contribution to the environment contamination. Exit rates of each compartment are not represented. :::  ::: *X*(*a,t*) represents the number of animals in health state *X*and age *a*at time *t*. Age is given in weeks until first calving (*cal*) and in parities (*cal*+1 to *cal*+5) after calving. An individual-based model is used until age *m*, when calves move to collective pens. Then, a compartmental model is used. If *a*≤ *m*, an index *k*indicates in which individual pen the calf is: *X*(*a,t,k*) = 0 or 1 depending on the occupancy of pen *k*. The total number of calves of age *a*at time *t*is: $X(a,t) = {\sum\limits_{k = 1}^{n}{X(a,t,k)}}$, with *n*the number of individual pens. The herd model is calibrated by integrating knowledge from various sources, from published data to experts\' knowledge, to realistically represent a French dairy cattle herd (Table [1](#T1){ref-type="table"}). All male calves (half the calves) exit the herd during the 2^nd^to 4^th^week after birth (rate *σ~m~*). Closed herds are modelled: there is no purchase of heifers for replacement. All female calves are thus kept to give flexibility to regulate the number of cows. Herd size is assumed to be stable over time. Heifers can be sold but only 10 weeks before the first calving (rate *σ~h~*). Above a given number of cows (*K~c~*), the heifer sale rate increases. Under this threshold, the sale rate decreases. An all-year round calving is modelled with a mean calving-to-calving interval *cci*. Animals older than six months of age graze from April to November (*Graz*). ### Infection process and Map transmission The progression of individual animals through different *Map*infection states is a complex continuous process which is converted into discrete phases to enable the model parameterization based on current knowledge. Animals are classified into mutually exclusive health states: susceptible (*S*), resistant (*R*), transiently infectious (*T*) (infectious only for a limited period of time), latently infected (*L*) (infected not infectious), subclinically infected (*Is*) (infected and infectious but not affected), and clinically affected (*Ic*) (infected, infectious, and affected) \[[@B2]\]. Parameters are displayed in Tables 1 (herd dynamics), 2 (infection process), and 3 (shedding characteristics). Assumptions are based on current knowledge on *Map*. Vertical transmission occurs with probability *p~X~*(*T*calf born to an infected cow). Horizontal transmission occurs by ingestion of colostrum, milk, or faeces. It depends on animal susceptibility, varying with age (maximal the first week of age and decreasing exponentially (*h*) until one year of age (*u*)). Under field conditions, animals older than one year of age have a low susceptibility to *Map*infection \[[@B14],[@B15]\] and in the current model are therefore assumed to be resistant to infection. If infected, there is no possible recovery. We assume an exponential distribution of the durations in infection states *T*, *L*, *Is*, and *Ic*. A transiently infectious state is assumed as infected calves have been reported to shed *Map*\[[@B9]\]. The transition from *T*to *L*either is modelled using a binomial distribution of probability 1/*v~T~*, *v~T~*being the mean duration of the transiently infectious period, or occurs at the latest when the age at first calving (*cal*) is reached. A latent state is assumed because, if the absence of shedding has not been proven, the detection of infectious adults and heifers is hardly possible before animals reach one to two years of age, indicating at least quite a low level of shedding \[[@B16]-[@B18]\]. Latent animals are assumed not to shed *Map*, since shedding can be considered to be negligible compared with that of other infected adults. The transition from *L*to *Is*is possible only after the 1^st^AI (at age *h*). Subclinical animals are assumed to shed sufficient quantities of *Map*to be detectable and to contribute to *Map*spread within the herd, without having any obvious clinical signs. The transitions from *L*to *Is*, *Is*to *Ic*, and *Ic*to exit of the herd are modelled using binomial distributions of probabilities 1/*v~X~*(*X*= *L*, *Is*, or *Ic*), *v~X~*being the mean time spent in state *X*. There is no additional mortality for *Is*and *Ic*cattle, but *v~Ic~*accounts for additional culling. Depending on their age, *S*calves are not all exposed to the same transmission routes. Calves born to infected dams can be infected via colostrum ingestion in the first week of age. During the first two weeks, calves are housed in individual pens. They can be infected via milk ingestion, exposure to the environment of the whole farm (global environment), or indirect transmission from infected calves of neighbouring pens. Before weaning, calves housed collectively can be infected via milk ingestion, exposure to the local environment of their pens, or exposure to the global environment. Inside (during winter), weaned calves can be infected via exposure to the local or to the global environment. On pasture, they can only be infected via exposure to the local environment shared with young heifers. Colostrum and milk contamination occurs because of direct shedding or indirect faecal contamination. A calf ingests the colostrum of its dam. A calf *k*born to a cow in state *X*∈ {*Is*,*Ic*} ingests at time *t*the following amount of bacteria: $$q_{c}^{k} = Bernouilli(sh_{X})\left\lbrack {f(X,direct) + f(X,indirect)} \right\rbrack\, b$$ with *f*(*X*,*r*) the quantity of bacteria per litre of colostrum for an animal in state *X*through route *r*(*f*(*X*,*r*) \~ F(*X*,*r*)), *sh~X~*the probability of shedding in colostrum for cows in state *X*, and b the quantity of colostrum fed to calf. The number of calves infected via colostrum ingestion is then: $$inf(c,t) = {\sum_{k = 1}^{k = n}\left\lbrack {S(1,k,t)Bernouilli\left( {1 - \exp( - \frac{\beta_{l}q_{c}^{k}}{\alpha})} \right)} \right\rbrack}$$ with *S*(1,*k*,*t*) = 1 if there is a susceptible calf of one week of age in pen *k*at time *t*and 0 otherwise, *β~l~*the transmission rate if ingestion of an infectious dose, and *α*the infectious dose. Similarly, the number of calves infected via milk ingestion is: $$inf\text{(}m,t\text{)} = {\sum\limits_{a = 1}^{a = w}{\lbrack Bin(S(a,t),1 - \exp( - \exp\lbrack - h(a - 1)\rbrack\frac{\beta_{l}q_{l}}{\alpha}}}))\rbrack$$ with *S*(*a*,*t*) the number of susceptible calves of age *a*at time *t*, *q~l~*the quantity of bacteria ingested per calf via milk ingestion. *q~l~*depends on the quantity of milk drank per calf (*d*) and the quantity of bacteria in the tank, which depends on the proportion of *Ic*and *Is*lactating (*prop*) and shedding (*sh~X~*) cows, these cows either directly shedding in milk (*f*(*X*,*direct*)) or because of faecal contamination of the milk (*f*(*X*,*indirect*)), and the quantity of milk they produce (*ε - g~X~*). Faecal-oral transmission is indirect, occurring by ingestion of bacteria present in the environment. Two types of environment are modelled to differentiate indirect adult-to-calf from indirect calf-to-calf transmissions (Figure [1](#F1){ref-type="fig"}). *E~g~*is the quantity of *Map*in the global environment, contaminated by all of the shedding animals. It is assumed that all calves are equally exposed to the farm\'s environment, not accounting for possible variation in distribution of *Map*. *E~1~*to *E~3~*are the quantities of *Map*in the calves\' local environments, exclusively contaminated by *T*animals housed in the associated facilities (Figure [1](#F1){ref-type="fig"}). We assume a homogeneous distribution of calves\' faeces in a local environment or that all calves in a contaminated pen have the same probability of ingesting *Map*during a week. Susceptible animals are exposed to *Map*in the global and their local environments. The global environment is the sum of the local environments for calves and adults. All infectious cattle shed *Map*in their faeces. We assume shedding varies with the infection state, but also over time for a given infectious animal. We assume *T*animals shed on average almost as much bacteria per kg of faeces as *Is*animals but with a lower variability, *Ic*animals shedding much more (Figure [2](#F2){ref-type="fig"}). To model such a heterogeneity in shedding between animals and states, we fit distribution laws F(*X*,faeces) (Figure [2](#F2){ref-type="fig"}) of *Map*quantities shed at time *t*per kilogramme of faeces by a given animal of state *X*to published observed data (Table [3](#T3){ref-type="table"}). At time *t*, the quantity of *Map*per environment is updated, according to the removal rate *μ*(mortality of *Map*, cleaning of the barn, straw management) and *Map*shed by infectious animals. We assume no bacterium survives on pasture during winter; pastures are free of *Map*at next turn-out. In individual pen *k*, a susceptible calf of age *a*is infected at time *t*because of *Map*residuals in the pen with probability: $$P_{\inf}^{k}(a,t) = 1 - \exp( - \exp( - h(a - 1))\frac{\beta_{c}E_{l}(k,t)}{\alpha})$$ ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Distribution of the amount of *Mycobacterium avium*subsp**.***paratuberculosis (Map)*shed per kg of faeces of transiently infectious (F~T~), subclinically infected (F~Is~) and clinically affected (F~Ic~) animals used in the *Map*spread model within a dairy herd\***. \*Distributions are here in log(*Map*)/kg of faeces (and not in *Map*/animal/day). Transiently infectious animals produce from 0.5 to 10 kg of faeces per day during 25 weeks on average (*f~1~*, *f~2~*, *f~Y~*), whereas *Is*and *Ic*animals are cows producing 30 kg of faeces per day (*f~A~*) for a longer period of time (Tables 1 and 2). Adults\' contribution to total *Map*shed is thus more important than the one of transiently infectious animals. :::  ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Summary of published data and modelled distributions of the quantities of *Mycobacterium avium*subsp. *paratuberculosis (Map)*shed, depending on the health state (X) and the route of transmission (r) in a *Map*infection dynamics model within a structured dairy herd ::: Route of transmission (r) Health state (*X*) Literature Model ------------------------------------------------------------------------------- ------------------------ ------------- ------------- ------------- ------------------- ---------------------------------- *Map*direct shedding in milk and colostrum (*Map*/L) Subclinically infected 2.2 × 10^4^ 8.8 × 10^4^ 5 × 10^4^ \[[@B39]\] 10^5^× beta(8,8) Clinically affected \- \- 5 × 10^4^ \[[@B42]\] 10^5^× beta(8,8) *Map*indirect shedding in milk and colostrum (faecal contamination) (*Map*/L) Subclinically infected 0 2 × 10^10^ 40 \[[@B43],[@B44]\] 1 + 10^3^× beta(1,25) Clinically affected 700 2 × 10^10^ 14 × 10^4^ \[[@B43],[@B44]\] 10^(3\ +\ 10\ ×\ beta(50,200))^ *Map*shedding in faeces (*Map*/kg) Transiently infectious 6 × 10^4^ 6.3 × 10^5^ 3 × 10^5^ \[[@B9]\] 10^6^× beta(8.8,19) Subclinically infected 10^4^ 10^15^ 2.6 × 10^6^ \[[@B45]\] 10^(4\ +\ 10\ ×\ beta(2.65,17))^ Clinically affected 10^8^ 10^15^ 10^10^ \[[@B26],[@B46]\] 10^(8\ +\ 10\ ×\ beta(2,17))^ ::: with *β~c~*the indirect calf-to-calf transmission rate. Calves also can be infected because of their infectious neighbours (randomly sampled among calves). In collective pen *i*, susceptible calves of age *a*are infected at time *t*via calf-to-calf indirect transmission with probability: $$P_{\inf}^{i}\left( {a,t} \right) = 1 - \exp( - \exp( - h(a - 1))\frac{\beta_{c}E_{i}(t)}{\alpha N_{i}(t)})$$ with *N~i~*(*t*) the number of animals in local environment *i*at time *t*. Susceptible calves of age *a*are infected at time *t*via the global environment with probability: $$P_{\inf}^{g}\left( {a,t} \right) = 1 - \exp( - \exp( - h(a - 1))\frac{\beta_{g}E_{g}(t)}{\alpha N(t)})$$ with *β~g~*the indirect transmission rate from this environment and *N*(*t*) the herd size. ### Initial conditions All animals younger than *u*are initially susceptible, other animals being resistant to infection. A subclinically infected parity one cow is introduced once in the herd with no further introduction. For each run, the date of introduction corresponds to the first week of January, i.e. three months before grazing starts. No specific measure is implemented in the herd to prevent or control *Map*infection. No change in herd management is implemented over time. Initially, study herds are composed of 277 animals (118 calves and young heifers, 45 bred heifers, and 114 cows). ### Model outputs Results are obtained from 400 runs over 25 years. We monitored the stability of means and variances of model outputs with increasing number of runs. We stopped when these estimates changed by less than 5% due to the last 100 runs. Therefore, runs are numerous enough to obtain stable simulated results. The first output is the infection persistence over time, i.e. the percentage of runs with the infection still present. We can deduce from this output the proportion of runs ending with fadeout. Other outputs then are studied separately for runs with persistent infection or runs with fadeout. The second output is the prevalence of infected (*T*+*L*+*I~S~*+*I~C~*), infectious (*T*+*I~S~*+*I~C~*) and affected animals (*I~C~*) over time, these categories being defined by Nielsen & Toft \[[@B19]\]. For runs with persistent infection, the pseudo-equilibrium of the prevalence is estimated. Among the two types of runs, the proportion of animals that become *I~C~*or detected with a systematic test (sensitivity of 0.5 and specificity of 1) during the early infection dynamics is studied. The third output is the relative contribution of the transmission routes to the number of newly infected animals. ### Model evaluation First, model outputs are compared with published data and field data from infected herds \[[@B20],[@B21]\]. The simulated proportion of infected adults is compared to the estimated prevalence of infected adults on farms that voluntarily participated in a control program based in Brittany (France) \[[@B22]\]. Data corresponds to 59 herds enrolled in the program between 2002 and 2005 and in which more than 20 adults per herd were tested in the year of enrolment. All adults older than 24 months of age were tested annually using both ELISA and either PCR or faecal culture until 2007, and systematic ELISA and PCR in faeces of ELISA positive animals in 2008. Ziehl-Neelsen tests \[[@B1]\] were performed when suspect clinical signs were observed. Individual statuses of adults during the first year of the program implementation (i.e. before any control measure was introduced) are retrospectively attributed based on a maximum of three successive annual results. These statuses are defined as: clinically affected (Ziehl-Neelsen positive test in the first year), subclinically infected (PCR or faecal culture positive in the first year but Ziehl-Neelsen negative if performed), latently infected (seropositive in the first year but PCR or faecal culture negative or negative in all tests in the first year with a positive test later, whatever the test), and resistant (testing negative in all tests). For animals always testing negative but with only one or two tests (instead of three), we assume that they are either resistant (optimistic option which may under-estimate infection) or latently infected (pessimistic option which may over-estimate infection). Based on these optimistic and pessimistic distributions, we estimate the distribution of animals per infection state at the start of the program and the within-herd prevalence at enrolment. To compare model outputs with field data, we assume farmers usually detect the disease from 5 to 9 years after *Map*introduction (time needed for clinical cases to occur). We calculate the distribution of the mean simulated prevalence in infected adults in infected herds over this time period. Second, a hypothesis-testing approach is used to validate the model, assuming a constant herd structure. We verify that either our conclusions are robust to variation in model parameters, or that parameter variation induces unrealistic within-herd prevalence and therefore cannot be retained. A one-at-a-time sensitivity analysis is performed for uncertain parameters (*ν~T~, u, h, p~x~, sh~x~, β~l~*, *β~c~, β~g~, β~o~*, F(*T*,faeces)). Variations of ± 50% from reference values are tested where applicable (*ν~T~, u, h, p~x~, sh~x~, β~l~*, *β~c~*, *β~g~*, *β~o~*). For F(*T*,faeces), the worst plausible case is tested, *T*animals shedding (per kilogram of faeces) as much as *Is*animals, with the same variability. Third, to evaluate how the conclusions change with herd management, a model evaluation is performed as regards to variations of parameters managed on farm (*μ~k~, K~c~, ν~Ic~, Graz*). Variations of ± 50% from nominal values are tested for *μ~k~*, and *ν~Ic~*. For *K~c~*(closely related to herd size), limits of 50 vs. 500 cows are tested. Lastly, a delay in the start of grazing (same duration but starts in the week *Map*is introduced vs. ends in the week before *Map*is introduced) and a variation in its duration (same start but duration of 28 vs. 37 weeks) are tested. Results ======= Spontaneous fadeout of *Map*infection without control measure ------------------------------------------------------------- Spontaneous fadeout occurred in 66% of the runs (Figure [3](#F3){ref-type="fig"}). In 43% of the runs, it occurred within the first two years (early extinction), while it occurred less quickly in the remaining 23%. Herds still infected seven years after *Map*introduction had thereafter a fadeout probability less than 6%. When shedding animals were no longer present on the farm but the environment was still contaminated thus fadeout had not occurred yet, new infection of cattle from residual *Map*in the environment occurred with a mean weekly probability of 3%. Hence, once the environment has been contaminated, spontaneous fadeout was hardly possible. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Probability of persistence over time (proportion of runs where an infected animal is still present) of *Mycobacterium avium*subsp**. ***paratuberculosis (Map)*infection in a dairy cattle herd after a single Map introduction (*t*= 0) in the herd.** :::  ::: The probability of fadeout only slightly varied with uncertain parameters (from 62 to 71%). It decreased to 51% when the mean time spent in state *Ic*increased by 50%, and to 58% when *Map*removal from the global environment decreased by 50%. Other parameters relating to herd management only had little influence on the fadeout probability. It needs to be emphasized that yearly single introduction of *Map*would lead to a decrease in the cumulative probability of spontaneous fadeout, which can be calculated for *n*years using 0.66^*n*^(e.g. 66% the first year as in the present study, 44% the second year, 29% the third year, etc.). *Map*spread within persistently infected herds ---------------------------------------------- Prevalence of infection reached a pseudo-equilibrium (when accounting only for runs in which infection persisted) 23 years after *Map*introduction when no control measure was implemented (Figure [4](#F4){ref-type="fig"}). At the end of the simulation period, the prevalence of infected, infectious, and affected animals reached 88%, 44%, and 6%, respectively. In adults, prevalence of infected, infectious, and affected animals was 87%, 67%, and 15%, respectively. Annual incidence reached 15% (Figure [4](#F4){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### ***Mycobacterium avium*subsp**. ***paratuberculosis (Map)*spread in a persistently infected dairy cattle herd since *Map*introduction (*t*= 0).** A/Mean prevalence over time of infected (black), infectious (dark grey), and affected (light grey) adults (\> 30 months) and related confidence intervals. B/Mean annual incidence and related confidence interval. :::  ::: Comparing the simulated and the observed distributions of prevalence in infected herds indicated that the model over-estimated the cases when infected herds had a low prevalence (more than 40% of the infected runs had a prevalence in infected adults less than 5%; Figure [5](#F5){ref-type="fig"}). For other levels of prevalence, simulated and observed distributions were similar. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Comparison of the simulated and the observed distributions of the prevalence in *Mycobacterium avium*subsp. *paratuberculosis (Map)*infected adults in infected dairy cattle herds**. The simulated distribution corresponds to runs of a *Map*spread model within a dairy cattle herd, the mean prevalence from year 5 to year 9 since *Map*introduction in the herd (*t*= 0) being calculated for each run still infected. The observed distribution is based on individual life long determined statuses in 59 dairy herds at enrolment in a paratuberculosis control program in France, before any control measure is implemented. :::  ::: Varying uncertain parameters produced in most cases (*u, ν~T~, p~Ic~, sh~X~, β~c~, β~l~, β~o~*, F*(T,faeces*)) prevalence distributions similar to the reference scenario and therefore these parameters cannot be more precisely estimated from the sensitivity analysis. For others (*h, p~L~, β~g~*), a variation of ± 50% resulted in a simulated prevalence not consistent with the observed prevalence. Among infected adults, the model provided mean proportions of *L*, *I~S~*, and *I~C~*animals 25 years after *Map*introduction of 60, 32, and 8%, respectively (Figure [6A](#F6){ref-type="fig"}). These proportions slightly varied over time, except in the transient period when prevalence was very low. In field data (Figure [6B](#F6){ref-type="fig"}), the proportion of animals per infection state depended on the option: the pessimistic option resulted as expected in a large proportion of latently infected animals. The mean proportion of subclinically infected animals varied from 17 to 40% in the optimistic option, and from 3 to 22% in the pessimistic option. Simulation values were in between the two assumptions (Figure [6](#F6){ref-type="fig"}). ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Comparison of the simulated and observed distributions of *Mycobacterium avium*subsp**. ***paratuberculosis (Map)*****infectedadults perinfection state in infecteddairy cattle herds**. A/Simulated mean distribution over time in persistently infected herds as predicted by a *Map*spread model within a dairy cattle herd; B/Mean percentage of tested adults per infection states based on individual life long determined status in 59 herds at enrolment in a paratuberculosis control program in France, before any control measure is implemented, according to the range of the initial within-herd prevalence. Animals tested twice or less and having negative results are assumed to be either resistant (state not shown) (B1: optimistic option) or latently infected (B2: pessimistic option). :::  ::: At the herd level, the main transmission routes were indirect transmission via the contaminated global environment, then in utero transmission. Transmission via colostrum or milk ingestion and calf-to-calf indirect transmission appeared to be minor routes (Figure [7](#F7){ref-type="fig"}). For high within-herd prevalence, the two main transmission routes equally contributed to new infections (Figure [7B](#F7){ref-type="fig"}). For parameter variation resulting in plausible results, these conclusions remained unchanged. Even an increase of one log (\*10) of the indirect transmission rate in the calf environment barely changed the contribution of calf-to-calf indirect transmission, which slightly increased for a low within-herd prevalence. Assuming *T*animals shed as much as *Is*animals (per kg of faeces) resulted in calf-to-calf indirect transmission contributing as much as in utero transmission for a very low within-herd prevalence, the contribution decreasing for a prevalence higher than 5%. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Mean relative contributions of the 5 transmission routes of *Mycobacterium avium*subsp**. ***paratuberculosis (Map)*****infection in persistently infected dairy cattle herds(118 runs out of 400)**. A/over time since *Map*introduction in the herd; B/over prevalence of infectious adults. *Map* is introduced only once (*t*= 0). :::  ::: Characteristics of the runs ending in fadeout vs. persistent infection ---------------------------------------------------------------------- No secondary infection (on top of the first introduced case) was observed in 75% of the runs with fadeout, contrary to herds persistently infected. Only 3% of the runs ending with fadeout had at least two clinically affected animals (simultaneously or successively) over five years, compared to 80% of the persistently infected runs (Table [4](#T4){ref-type="table"}). When combining clinical surveillance and systematic testing of cows, more than 2 animals were detectable after 3 years in 18% of the runs with fadeout and in 68% of the runs with persistent infection (21% and 96%, respectively after 5 years). ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Proportion (%) of runs having 0 to more than 3 clinically affected and/or subclinically infected animals (*Is*) detected (sensitivity of 0.5 and specificity of 1 for the tests used for Is animals detection) after 1 to 5 years of simulation in herds with spontaneous fadeout or persistent infection ::: Cumulated number of animals (*nr*) \% of runs with *nr*clinically affected animals \% of runs with *nr*clinically affected & detected subclinically infected animals --------------------------------------------------------- ------------------------------------ ------------------------------------------------- ----------------------------------------------------------------------------------- ---- ---- ---- ---- ---- ---- ---- ---- Proportion among herds with fadeout (282 runs) 0 75 67 64 62 62 40 37 37 35 35 1 25 33 36 36 35 50 48 45 45 44 2 0 0 0 1 2 10 14 17 17 17 ≥ 3 0 0 0 1 1 0 0 1 3 4 Proportion among persistently infected herds (118 runs) 0 48 23 9 5 2 24 8 2 1 1 1 52 75 67 40 18 46 51 30 14 3 2 0 2 15 15 19 60 34 25 22 8 ≥ 3 0 0 9 40 61 0 7 43 63 88 ::: Based on the model outputs in Table [4](#T4){ref-type="table"}, we can predict at the herd level the probability of *Map*persistence for a situation under a given detection threshold vs. the probability of spontaneous fadeout for a situation over this threshold. If a control programme based on clinical surveillance is implemented when at least one affected animal is observed in five years, the programme is unnecessarily implemented (fadeout would have spontaneously occurred) in 48% of the cases (i.e. the number of runs over the threshold ending with fadeout over the total number of runs over the threshold). If no control programme is implemented (no affected animals in five years after *Map*introduction), a persistent infection occurs in 1% of the cases. For a threshold of two affected animals, these proportions are 9% and 8%, respectively. For a threshold of 3, they are 4% and 14%, respectively. However, only 24% of the persistently infected herds had at least 2 affected animals within 3 years after *Map*introduction, 80% within 5 years. If the control programme is based on both clinical surveillance and imperfect tests (assuming a sensitivity of 0.5 and a specificity of 1) targeting adults, the proportions become 61% and 2% for at least 1 detected animal in 3 years after *Map*introduction, 39% and 14% for a threshold of 2, and 5% and 19% for a threshold of 3. 68% of the persistently infected herds had at least 2 detected animals within 3 years after *Map*introduction, 96% within 5 years. Discussion ========== The results from model experimentation have improved the understanding of *Map*spread within a dairy herd. Fadeout could occur even without implementation of control measures in an infected herd. This demonstrates the usefulness of a modelling approach, since such fadeout cannot be observed in the field given the low prevalence of infection and low likelihood of detection using available diagnostic methods. Probability of fadeout was estimated at 66%, showing this probability can be high. This absolute value cannot be used directly as it cannot be validated with observed data since most fadeout events cannot be observed. It is likely to vary with model assumptions (including herd characteristics). Nevertheless, the sensitivity analysis demonstrated that fadeout is likely to be frequent in a wide range of situations. The economic assessment of paratuberculosis control programmes should therefore account for this high probability. Moreover, this model predicts changes in the fadeout probability when the delay to cull clinically affected animals varies and shows how important a control measure it is. This model can be used similarly to evaluate the relative impact of other interventions. The cumulated number of clinically affected animals appears to be a good indicator of the progression of *Map*infection dynamics towards persistence. Furthermore, it is very easy to use in the field. A threshold of two affected cows seems adequate to trigger control measures in a herd. However, a farmer may miss the 1^st^clinical case and be unaware that there already have been two cases in his herd. An earlier indicator would be useful. Combining clinical surveillance with an imperfect test implemented on all potentially infected adults could reduce the time needed for detection. In that case, a threshold of three detected animals seems adequate. To assess the economic advantage of such surveillance, both the costs and benefits of early detection need to be analyzed. In the absence of control measures, the simulated mean prevalence in infected cattle increased to 88% after 25 years in the model, as previously published models also have shown \[[@B3]-[@B5],[@B23]\]. These levels of prevalence are not expected with field data as control measures will be implemented long before such levels are reached. However, herds with high apparent prevalence are found, which corresponds to these levels of true prevalence (e.g. \[[@B17],[@B24],[@B25]\]). Moreover, simulated prevalence between 5 and 9 years after *Map*introduction was lower than levels observed on farms prior to enrolment in a control programme. This suggests that the range of observed prevalence at control programme enrolment typically corresponds to a more advanced stage of within-herd *Map*dynamics, when without any control measure fadeout would rarely occur. With this new model, it was possible to assess the relative importance of transmission routes on *Map*spread in a dairy herd. This model accounts not only for vertical transmission and horizontal transmission via the ingestion of *Map*in milk and colostrum, as has been done in previously published models \[[@B6]\], but also for indirect contacts between animals of different ages raised in different groups, and horizontal transmission via the ingestion of faeces present in the contaminated environment. Possible exposure of calves to adults or to other calves is modelled and the level of exposure varies depending on calf age and calf housing facilities. In persistently infected herds, contamination of the environment by adults was the main transmission route, in utero transmission being the second. Calf-to-calf transmission appeared to be a minor route of transmission. However, in this model, milk and colostrum routes of transmission correspond to liquid contamination by the dam (direct shedding or faecal contamination), not contamination through the environment. On the other hand, possible faecal contamination of buckets used to give milk to calves is considered to be an element of global environmental contamination, not the milk route of transmission. As a priority, exposure of calves to any environment contaminated by adult faeces should be reduced, particularly at and just after birth when calves are the most susceptible. The model has been evaluated and provides qualitative predictions such as ranking routes and the description of possible dynamics. The model validation has been performed by comparing model outputs with field data. A hypothesis-testing approach has been used allowing us to conclude that our findings are robust to variation in uncertain model parameters. For some of the uncertain parameters (*h, p~L~, β~g~*), the true value is likely to be within a smaller interval than ± 50% of their reference value as larger variations led to results inconsistent with observations. However, only a partial validation has been possible because the introduction date of *Map*into a herd was not known for the observed field data. Furthermore, we assumed here herds are closed (a single *Map*introduction), whereas data may concern open herds with multiple introduction of potentially infected cattle. Finally, in practice, when paratuberculosis is diagnosed, farmers are likely to change their routines to ensure their animals\' welfare and protect their economic interests. It would be unethical to recommend that they do nothing. In contrast, we can model herds in which no control measures are implemented. In the model, we neglected some processes and factors that may interfere with *Map*spread but that are not yet sufficiently described. First, we did not represent passive or intermittent shedding in the model. The intermittent shedding sometimes noticed \[[@B26]\] indeed could be explained by the low sensitivity of diagnostic tests or by heterogeneity of faeces or milk samplings \[[@B27]\] which lead to an intermittent detection of infectious animals. If such intermittent shedding were to be shown, a different modelling approach would have to be used, where a given probability of shedding in the latent state according to age or to the physiological status (in gestation, in lactation, etc) would have to be assumed. However, given the current knowledge such a model cannot be parameterized. Moreover, such intermittent shedders would not be directly in contact with susceptible calves but be raised together with other adults. Their contribution to the global environmental contamination thus would be very limited as it would be diluted by the quantity of *Map*shed by subclinically infected and clinically affected animals. Therefore, the environmental contamination would be only slightly higher assuming latently infected animals shed intermittently. Second, super-shedders have been described \[[@B28],[@B29]\] but it is unknown whether they are specific animals or if shedding of all infectious animals varies highly over time. Therefore, we assumed here any animal can shed a high amount of *Map*at random time. Third, experimental animal models suggest there could be genetic factors responsible for resistance or susceptibility to *Map*infection \[[@B30]\]. Several genes have been identified to date. However, current knowledge is insufficient to include such genetic factors in modelling. Lastly, the incubation period is inversely related to the challenge dose, clinical signs occurring sooner under experimental than natural conditions \[[@B31]\]. However, the mechanism of the dose-response effect, the potential cumulative exposure, and the minimum infection dose are still uncertain. Therefore, this has not been included in the model. The model could be adapted to open dairy herds and used to evaluate control measures in both open and closed herds. Furthermore, this model could be used for herds of different sizes having similar herd structure and management. Herd management is driven by a number of parameters which gives flexibility to the model. However, the model would need to be modified if the structure of the herd is markedly different as exposure to the contaminated environment would differ. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= CM and PE conceived of the study, carried out the model development and analysis, and drafted the manuscript. CF conceived of the study, participated in its design and coordination, and drafted the manuscript. HS conceived of the study and participated in its design. DP participated in the study design and coordination. All authors read, amended and approved the final manuscript. Acknowledgements ================ The authors gratefully acknowledge A Lehebel and D Taisne for data analysis. Financial support for this research was provided by the French National Institute for Agricultural Research, Cemagref and Basse-Normandie, Bretagne, Pays de le Loire et Poitou-Charentes Regional Councils under SANCRE project, in the framework of \"For and About Regional Development\" programs, and by the French Ministry of Agriculture.

PubMed Central

2024-06-05T04:04:19.501059

2011-2-15

PMC3053234

Introduction ============ Antimicrobial resistance is an important concern in animal health. Antimicrobial treatment causes selective pressure that can trigger the selection of resistant bacteria \[[@B1]\]. The selection, persistence and spread of such strains within a population is often puzzling and can lead to failures of the usual treatments against infections \[[@B2]\]. Moreover, administration of antimicrobial agents used in human medicine to production animals, such as fluoroquinolones, could lead to the emergence of public health issues due to the persistence and dissemination of resistant bacteria \[[@B3]\]. Although surveillance of resistance in zoonotic bacteria has a direct implication on public health policies, monitoring bacteria from the commensal flora, such as *Escherichia coli*strains, remains a good indicator of the resistance pattern within a population. Knowledge of the resistance pattern in such commensal bacteria is pivotal because of the risk of transmission of resistant genes from non-pathogenic to zoonotic bacteria \[[@B4]\]. Quinolones are widely used in pig production, mainly to control urinary tract infections in sows \[[@B5]\]. Belloc et al. \[[@B6]\] observed that quinolone treatment caused a strong selective pressure in the *E. coli*population of the treated sows and their piglets. In this study, the treatment was routinely administered to pregnant sows around farrowing-time. Even though the effects were transitory, the faecal flora of the piglets was clearly modified by the treatment of their dams, leading to a dramatic increase in the within-host proportion of resistant *E. coli*. The long-term impact of such routine practices of antimicrobial use is unknown, but enrichment of the resistant flora with duration of use is highly suspected \[[@B7]\]. Studies of the long term ecological evolution within a population require a mathematical modelling approach \[[@B8],[@B9]\]. To date only a few epidemiological models have been developed to study the spread of infectious agents within pig-production units \[[@B1],[@B10]-[@B12]\]. To the best of our knowledge, only one of these modelling approaches focussed on the impact of antimicrobial consumption on the emergence and/or transmission of resistant strains within a finisher pig herd \[[@B13]\]. This study highlighted the need for a modelling framework which incorporates both current knowledge of the epidemiology of antimicrobial resistance in animals and the key factors which can be used to characterize within-herd dynamics of resistant bacteria under the influence of antimicrobial pressure. Knowledge of the transmission dynamics of resistant bacteria would improve the understanding of the influence of on-site farming practices (drug usage, infection control) on the dynamics of resistant bacteria in animals and could also form the basis for a release assessment (component of a risk assessment). Abatih et al. \[[@B13]\] used a deterministic SIS (Susceptible-Infectious-Susceptible) model in which infectious individuals were divided into two sub-groups (*I*~*I*~and *I*~*R*~) according to their levels of susceptibility to antimicrobial agents. The authors concluded that the persistence of resistant bacteria within the finishers was strongly related to the transmission parameter values \[[@B13]\]. Transmission rates (*β*) are key parameters of epidemiological models. However, estimation of these parameters often presents a challenge since, in many infectious diseases, only the onset of clinical symptoms or the final outcome of the disease can be observed in field conditions. Experimental transmission trials provide a useful tool for assessing infection dynamics in a controlled environment \[[@B14]\]. Different estimation methods have been used to analyse results from transmission experiments: maximum likelihood estimation (MLE, \[[@B15]-[@B19]\]), generalised linear model (GLM, \[[@B20]-[@B22]\]) or Becker\'s martingale formulae based on the final size of the epidemics \[[@B14],[@B23],[@B24]\]. The aim in these studies was to quantify the transmission of pathogenic infectious agents (viral or bacterial). However, to the best of our knowledge, none of these experimental studies focussed on the transmission of resistant bacterial strains. The aim in the present work was to study the horizontal transmission of a fluoroquinolone-resistant *E. coli*strain between pigs. For this purpose, transmission experiments were carried out by focussing on the transmission of an in vitro selected *E. coli*from inoculated to naïve specific pathogen free (SPF) piglets. Transmission parameters were estimated by applying a maximum likelihood method which accounted for different transmission levels related to the amounts of resistant bacteria recovered from the individual faeces of infected animals. Materials and methods ===================== Experimental design ------------------- The experiment was conducted in our air-filtered level 3 biosecurity-facilities. Five independent rooms were used (R1 to R5) with one or two pens per room (Figure [1](#F1){ref-type="fig"}). Pigs were housed on flat decks, with a fully slatted floor one metre above the floor of the facility. Sixty-four Large-White SPF pigs, aged seven weeks at the beginning of the experiment, were used in this study. Animals were weaned three weeks before the experiment and received non-supplemented feed. No antimicrobial treatment was administered either before the start or during the course of the experiment. Six pigs were kept as negative controls and 7 as positive inoculated controls (R5). The 51 remaining pigs were randomly assigned to 5 groups of 7 pigs (3 inoculated and 4 naïve SPF pigs) and 2 groups of 8 pigs (4 inoculated and 4 naïve SPF pigs). Each group was conducted separately and between-group transmission was avoided by having a solid partition between pens. Inoculations were performed orally, by successively administering 10 mL of a bacterial suspension (10^9^CFU/mL), on the day prior to contact with SPF pigs and on the contact-day. Contact was instigated by mixing the inoculated and naïve pigs about four hours after the second inoculation. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Experimental design for quantification of fluoroquinolone-resistant *E. coli*transmission from inoculated (circles) to naïve specific pathogen free (SPF) pigs (triangles)**. :::  ::: The bacterial suspension consisted of a ciprofloxacin-resistant *E. coli*in vitro selected from the faecal flora of a fully susceptible *E. coli*strain isolated from Specific Pathogen Free (SPF) pigs of the same herd. The selection of *E. coli*with chromosomal mutations was performed by spreading a large amount of inoculum (about 10^9^CFU) onto plates supplemented with increasing concentrations of ciprofloxacin (Sigma/Fluka, St-Quentin-Fallavier, France). The minimum inhibitory concentration of ciprofloxacin for the selected resistant strain was 1.5 mg/L (E-test, bioMérieux, Marcy-l\'Étoile, France). The suspension was administered using a soft sterile catheter which was gently but deeply introduced into the oesophagus to prevent the inoculum from flowing back and further contaminating the environment. All pigs were monitored daily by taking faecal samples from the day preceding inoculation until 4 days post-inoculation and then every 2 days until day 10. Clinical signs were also recorded daily and the pigs were weighed once a week. Bacterial counts were measured on 1 g of faeces using serial 10-fold dilutions. One hundred microliters of each dilution were plated on chromogenic agar (Chromocult Agar ES, Merck/VWR, Strasbourg, France) supplemented with 0.5 mg/L of ciprofloxacin (Sigma/Fluka, St-Quentin-Fallavier, France). The plates were incubated at 37°C for 24 h then the colonies were counted. This method allowed quantification of the concentrations of inoculated *E. coli*greater than or equal to 10^2^CFU/g of faeces. The experiment was ended two weeks after inoculation. Euthanasia was carried out by anaesthesia with an intravenous injection of 1g/50 kg liveweight of Nesdonal^®^(Merial, Lyon, France) followed by exsanguination. The following organs were examined macroscopically immediately after killing: lungs, tonsils, heart, kidneys, thymus, lymph nodes (inguinal, mesenteric, tracheo-bronchial) and the different parts of the digestive tract (oesophagus, stomach, duodenum, jejunum, ileum, colon and caecum). When macroscopic lesions were suspected, samples were taken from the tissues for histopathological examination. The experiment was performed in accordance with EU and French regulations on animal welfare in experimentation. The protocol was approved by the AFSSA/ENVA/UPEC ethical committee. Comparison of bacterial counts between groups --------------------------------------------- An analysis of variance was used to compare bacterial counts according to animal status (contact or inoculated pigs) and group size (7 or 8 pigs per pen). These comparisons were performed on the bacteriological results obtained at the start (D~1~) and end (D~10~) of the experiment (*lm*function, R software \[[@B25]\]). Quantification of transmission ------------------------------ Based on the bacteriological results, a stochastic SIS (Susceptible-Infectious-Susceptible) model was used to estimate the transmission parameters. As in a classical SIR (Susceptible-Infectious-Removed) model the rate at which susceptible animals (S) were infected is given by *βSI/N*, where *β*is the transmission parameter, defined as the mean number of new infections caused by a typical infectious individual per unit of time \[[@B26],[@B27]\], while *I*and *N*represent the number of infectious and the total number of individuals, respectively. Pigs were classified as susceptible or infectious according to the quantities of resistant bacteria recorded in individual faecal samples. Changes in individual status observed between two sampling dates, either from susceptible to infectious or reciprocally, were considered to occur at the mid-time interval. Individuals were considered susceptible when the quantity of fluoroquinolone-resistant *E. coli*was below a threshold level (*T*~1~), fixed at 5\*10^3^CFU/g in accordance with literature data \[[@B28]\]. Individuals with bacterial counts above this threshold quantity were considered as infectious, i.e. were considered able to transmit resistant bacteria to susceptible pigs. Three hypotheses (H1, H2, and H3) were proposed regarding the ability of infectious animals to transmit resistant bacteria to susceptible ones, depending on the amount of resistant bacteria in individual faecal samples: \- H1. Equal transmission rate for all shedding pigs with bacterial counts above the threshold value *T*~1~(5\*10^3^CFU/g). \- H2. Two transmission levels according to the bacterial counts: Pigs shedding between 5\*10^3^and 10^6^CFU/g of resistant *E. coli*were considered as low shedders (*I*~*L*~) as compared with pigs shedding more than 10^6^CFU/g of faeces (*I*~*H*~). \- H3. Three transmission levels according to the bacterial counts. An intermediate shedding level was included: pigs shedding between 5\*10^3^and 10^5^CFU/g of resistant *E. coli*were considered as low shedders (*I*~*L*~); but if the bacterial counts ranged between 10^5^and 10^6^CFU/g, the pigs were classified as moderate shedders (*I*~*M*~), whereas, in high shedders, as with hypothesis 2, the bacterial counts exceeded 10^6^CFU/g (*I*~*H*~). These hypotheses led to different models of increasing complexity which could be used to study the relationship between the shedding pattern and transmission potential of infectious pigs. However, the model obtained under the third hypothesis (H3) could be considered as the full model with three transmission parameters: *β*~*L*~, *β*~*M*~and *β*~*H*~, the second hypothesis (H2) being a simplification of this model in which *β*~*L*~= *β*~*M*~, and the first hypothesis (H1) derived from the latter by assigning equal values to all the transmission parameters (*β*~*L*~= *β*~*M*~= *β*~*H*~). The maximum likelihood method, based on a stochastic SIS (Susceptible-Infectious-Susceptible) model, was used to estimate the transmission parameters. It follows from the model that the probability *q*that a single susceptible pig escapes infection during a time interval of *d*is equal to *q*= exp(-*d*(∑~*i*~*β*~*i*~π~*i*~)), where *i*represents the transmission level of infectious pigs (*i*∈{*L*, *M*, *H*} for Low-, Moderate- and High-shedders). In this equation, π~*i*~are the proportions of infectious animals in each shedding class, and *β*~*i*~the corresponding transmission parameter. *p*= 1 - *q*is then the probability that a susceptible individual will be infected and the number of new infections follows a Binomial distribution *C \~ Bin*(*S, p*) with *S*the number of susceptible animals in contact. The log-likelihood for this Binomial distribution is $\log\left( {L\left( {\beta_{L},\beta_{M},\beta_{H}} \right)} \right) = {\sum_{j}{\left\lbrack {C_{j}\log\left( {\text{exp}\left\{ {d_{j}\left( {\sum_{i}{\beta_{i}\pi_{i}}} \right)} \right\}\ - \ 1} \right)\ - \ S_{j}d_{j}\left( {\sum_{i}{\beta_{i}\pi_{i}}} \right)} \right\rbrack}}$ with *S~j~*and *C~j~*the number of susceptible animals and cases at each sampling interval *j*respectively, *d~j~*the duration of the interval between two sampling times, and where $\log\left( \begin{array}{l} S_{j} \\ C_{j} \\ \end{array} \right)$ is omitted because it plays no role in the calculations. The different *β*parameters were estimated using maximization of the log-likelihood with *nlm*function in *R*software (Version 2.6.0., \[[@B25]\]). Confidence intervals were determined using the inverse of the Hessian matrix (variance/covariance matrix) of parameter estimates, also provided by *nlm*function in *R*software. Results ======= Transmission results -------------------- No clinical sign was observed during the experiment and the daily weight gain (DWG) was similar in challenged groups (Figure [1](#F1){ref-type="fig"}, R1 to R4), control pigs (R5) and in individuals from the contact groups (data not shown). No resistant *E. coli*was isolated either before inoculation of pigs in the contact groups or in the negative control group throughout the entire experiment. All animals in the inoculated control group had positive resistant bacterial counts on the contact-day (D~0~) after which these counts declined steadily from D~1~to the end of the experiment (D~10~) (Figure [2](#F2){ref-type="fig"}). All inoculated pigs in the contact groups had positive resistant bacterial counts on D~0~, i.e. one day after their first inoculation. Individual bacterial counts in inoculated pigs increased until day 1 post-contact, ranging between 4.66 and 8.96 log~10~(CFU/g). The bacterial counts of the inoculated pigs then decreased over a period of 3 to 5 days. An inverse trend, with an increase in bacterial counts, was observed in 13 of the 23 inoculated pigs between day 4 and day 10 post-contact. This was probably due to the infectious pressure exerted by contaminated contact pigs as such a tendency was not observed in the inoculated control group. A similar pattern was observed in the contact groups with a lag of 1 to 2 days. Nineteen out of 28 contact pigs had positive bacterial counts on day 1 post-contact and all pigs in the contact groups were found positive on day 2 post-contact. An overall decrease in bacterial counts was observed until the 6^th^day post-contact. Similar transmission behaviours were observed in all the pens. No significant difference was found in individual bacterial counts according to the number of pigs per pen (5 and 2 trials involving 3 or 4 inoculated pigs respectively, in contact with 4 naïve pigs) either in inoculated or in contact pigs (Table [1](#T1){ref-type="table"}). Equivalent bacterial counts were recorded on days 8 and 10 post-contact in both the inoculated and contact groups (*p*= 0.66) (Figure [2](#F2){ref-type="fig"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Mean number and standard deviation (error bars) of fluoroquinolone-resistant *E. coli*in faeces of inoculated (*n*= 23), inoculated control (*n*= 7) and contact pigs (*n*= 28)**. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Comparison of mean bacterial counts at days 1 and 10 post contact between inoculated (I) and contact (C) pigs and between experimental settings (3I × 4C versus 4I × 4C trials). ::: Mean Bacterial Count (Log~10~(CFU/g)) ------------------------------------------- --------------------------------------- -------- --------------- -------------- -------- --------------- **Time of comparison (day post-contact)** **1** **10** **Estimate** **SD** ***P*-value** **Estimate** **SD** ***P*-value** Intercept 2.45 0.37 \< 0.001 3.84 0.38 \< 0.001 Group size\* (*n*= 8/*n*= 7) -0.27 0.70 0.70 -0.11 0.70 0.87 Status (inoculated/contact) 4.34 0.57 \< 0.001 -0.08 0.57 0.89 Interaction (Group size\*Status) 1.22 1.01 0.24 0.84 1.01 0.41 \*n: number of animals per pen in the experimental settings. *n*= 7: 3 inoculated and 4 naïve pigs, *n*= 8: 4 inoculated pigs in contact with 4 naïve pigs. ::: During the experiment, the resistant bacterial counts in at least one sample from 12 contact and 8 inoculated pigs was below the detection limit of the method, but the resistant strain could not be detected on D~10~in only three pigs. All but 5 inoculated pigs (18/23) had bacterial counts over 10^6^CFU/g on D~0~and D~1~. Only 7 of these high-shedders still showed high shedding levels on day 2 post-contact. The moderate shedding level, between 10^5^and 10^6^CFU/g, was found to be a transient phase between low- and high-shedding. Transmission parameter estimates -------------------------------- Transmission parameters differed according to the assumption made about the shedding pattern (Table [2](#T2){ref-type="table"}). As all three models were nested, the Akaike information criterion (AIC) was applied for model selection. The AIC values were very close whatever the underlying hypothesis (H1, H2 or H3). The lowest AIC was obtained with the second hypothesis (H2, *AIC*= 117.85) in which two transmission rates were considered. The estimated transmission of low-shedders, in this model, was 2.5 times less efficient than that of high-shedders (*β*~*L*~= 0.41 \[0.27; 0.62\], *β*~*H*~= 0.98 \[0.59; 1.62\]). These estimates were not modified by introducing an intermediate transmission level (H3), corresponding to pigs shedding more than 10^5^and less than 10^6^CFU/g. *β*~*M*~was very close to the *β*~*L*~estimate with a larger confidence interval due to the limited number of animals in this class (*β*~*M*~= 0.42 \[0.15; 1.19\]). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Transmission parameter estimates (and their 95% confidence intervals). ::: Hypothesis\* Parameters AIC -------------- ------------------- ------------------- ------------------- -------- H1 0.55 (0.40; 0.74) \- \- 121.00 H2 0.41 (0.27; 0.62) \- 0.98 (0.59; 1.62) 117.85 H3 0.41 (0.23; 0.71) 0.42 (0.15; 1.19) 0.98 (0.59; 1.62) 119.85 \* Hypothesis on the relationship between bacterial counts and transmission efficiency: H1. Equal transmission parameters (*β*~*L*~) for all shedding pigs with bacterial counts above the threshold value (5\*10^3^CFU/g). H2. Two transmission parameters according to bacterial counts: *β*~*L*~for low-shedders (bacterial counts between 5\*10^3^and 10^6^CFU/g) and *β*~*H*~for high-shedders (bacterial counts above 10^6^CFU/g). H3. Three transmission parameters according to bacterial counts: *β*~*L*~for low-shedders (bacterial counts between 5\*10^3^and 10^5^CFU/g), *β*~*M*~for moderate-shedders (bacterial counts between 10^5^and 10^6^CFU/g), and *β*~*H*~for high-shedders (bacterial counts above 10^6^CFU/g). ::: Discussion ========== The aim of this study was to assess the horizontal transmission of a fluoroquinolone-resistant *E. coli*within a swine population. Although the emergence and selection of antimicrobial-resistance can be linked with antimicrobial consumption \[[@B29]\], the transmission of resistant bacteria between individuals could be a pivotal factor governing persistence within a pig herd. Such transmission may occur in swine herds when mixing pigs which have been treated or not, or litters derived from treated and non-treated sows. As very few observations and no parameter estimates were available for the passive horizontal transmission of resistant commensal bacteria between animals \[[@B13]\], an experimental trial was designed to study the transmission of fluoroquinolone-resistant *E. coli*from inoculated to naïve pigs, in the absence of selective pressure. When studying a natural infection, the use of contact animals, rather than artificially inoculated animals with a presumed higher infectiousness, has been recommended to start the infection chain under study \[[@B30]\]. Only artificially inoculated pigs were used in the present experiment, in order to limit the number of pigs involved. However, the amounts of resistant bacteria shed by inoculated pigs on the day of contact were within the range observed in treated animals \[[@B31]\] or in piglets born to treated sows \[[@B6]\]. In addition, no significant difference was found between the amounts of bacteria shed by inoculated and contact pigs. The present results can therefore be considered as realistic and representative of the transmission pattern occurring under field conditions. Moreover, to date no plasmid-mediated quinolone resistance has been reported in *E. coli*isolates from pigs in France, so this observation can be considered as consistent with the current epidemiological on-farm situation. The experimental design was developed to mimic a situation found under field conditions and represented the mixing of piglets from different litters with heterogeneous resistance patterns. As suggested by Kroese and de Jong \[[@B14]\], the numbers of inoculated and contact pigs were equivalent in each experimental trial, with slightly fewer inoculated pigs in 5 pens (3 inoculated versus 4 contact animals) which did not affect transmission effectiveness. All naïve contact pigs had positive counts of fluoroquinolone-resistant *E. coli*after only two days of contact, demonstrating the strong infectious pressure exerted by inoculated animals. After a 4- to 5-day decline in bacterial counts in the faeces of inoculated pigs, an overall increase was observed from day 6 until the end of the experiment. These results evidenced the possible re-infection of inoculated pigs by newly contaminated animals. This assumption is also supported by the fact that control inoculated pigs which were not mixed with naïve animals did not exhibit the same behaviour, i.e. their bacterial counts declined steadily from day 2 until the end of the experiment. Such re-infections could favour the within-group persistence of resistant bacteria. A maximum likelihood method was used for parameter estimation which took into account the time-course of the transmission process, as the re-infections of inoculated animals precluded the use of other algorithms such as final size algorithm \[[@B24]\], as well as individual bacterial counts. The threshold transmission level (5\*10^3^CFU/g) was fixed according to literature data \[[@B28]\] and additional specific transmission experiments would be needed to establish the infectious dose for the resistant *E. coli*strain. However, the method used in the present study is based on the classical hypotheses of the SIR epidemiological model, accounting for horizontal transmission between individuals. The roles of environmental and airborne transmissions were not considered and would need further investigation. Three shedding-levels were considered in the present study, resulting in 3 nested models to allow for heterogeneous transmission rates. Transmission by low-shedders was indeed 2.5 times less efficient than that of high-shedders. The role of variations in individual infectiousness on the course of infection has been highlighted in several studies. Indeed, transmission rates could be influenced by several factors such as time-since-infection \[[@B16],[@B17],[@B32]\] or individual bacterial counts \[[@B33]\]. The role of super-spreaders was clearly identified for different infectious agents (e.g. MRSA \[[@B34]\], *E. coli*O157:H7 \[[@B35]\], *Salmonella*\[[@B36]\]). As transmission rates were found to be related to the individual bacterial counts, modelling the dynamics of drug-resistant bacteria within a pig herd would require representing the within-host dynamics of the bacterial population accounting for different susceptibility levels \[[@B37],[@B38]\]. Such a modelling approach should take into account the resistance mechanisms of bacteria to targeted antibacterial agents (plasmidic or chromosomal genes) and the impact of selective pressure exerted by antimicrobial exposure. To this end, the within-host bacterial population model, accounting for the pharmacokinetic-pharmacodynamic aspect, could be coupled with an epidemiological model representing the spread of resistance within the population. Such a model would allow the identification of key management and treatment strategies that could limit or prevent the spread of resistant strains between individuals. The present study, focussing on the transmission of an *E. coli*strain harbouring chromosomal resistance to fluoroquinolones, could be a cornerstone for the development of this pharmaco-epidemiological model. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= AM analysed the data and drafted the manuscript, ALR and EJ performed bacteriological studies, RC supervised the experiment, EJ, NR, PS, ML, and CC conceived of the study, participated in its design and coordination. All authors read and approved the final manuscript.

PubMed Central

2024-06-05T04:04:19.506230

2011-3-2

PMC3053235

Background ========== Cleidocranial dysplasia (CCD), also known as cleidocranial dysostosis or Marie-Sainton syndrome, is a disorder that affects most prominently those bones derived from endochondral and intramembranous ossification and it\'s characterized by defective development of the cranial bones and by the complete or partial absence of the clavicles. Diagnosis is based on clinical and radiographic findings, that include imaging of the cranium, thorax, pelvis and hands. Frequently these patients presents a delayed ossification of the skull fontanels and a premature closing of the coronal suture that leads to a frontal, parietal and occipital bossing of the skull; a short stature, occasionally accompanied by a spinal scoliosis; a wide and flat nasal bridge due to hypertelorism; different anomalies of pubis and hipbone, with flat feet and knock knees; a brachycephaly with an high arched palate and sometimes cleft palate; a prolonged retention of deciduous teeth and several impacted permanent successors and supernumerary elements, sometimes accompanied by follicular cysts and eruptive pseudocysts \[[@B1],[@B2]\]. This pathology is transmitted as an autosomal dominant trait or it\'s caused by a spontaneous genetic mutation and is present at a frequency of one in one million individuals. To date, RUNX2(CBFA1) is the only gene known to be associated with CCD; although not all cases clinically diagnosed have mutations in RUNX2, there is little additional evidence for locus heterogeneity. Mutations in RUNX2 have a high penetrance and extreme variability. CCD affects all ethnic groups \[[@B3],[@B4]\]. Although the spectrum of phenotypic variability in CCD ranges from primary dental anomalies to all CCD clinical features plus osteoporosis, no clear phenotype-genotype correlation has been established \[[@B5]\]. Children with CCD should be monitored for orthopedic complications, dental abnormalities, upper airway obstruction, sinus and ear infections, hearing loss, and osteoporosis. Intelligence is normal in individuals with classic CCD ------------------------------------------------------ The most important dental problem associated with this syndrome is the malocclusion and the crowding of the dental arches caused by the retention of multiple deciduous teeth and the presence of several supernumerary. These supernumerary, associated with a diminished alveolar bone resorption, also lead to the impaction or the ectopic location of the permanent teeth \[[@B6],[@B7]\]. The previous approach to the dental problems of these patients consisted in no treatment or in the extraction of the impacted or malformed teeth and their prosthetic replacement \[[@B8],[@B9]\], with a consequent important bone loss. In the last years a more conservative approach has been developed, combining orthodontics and maxillofacial surgery. Orthodontic treatment consist in the extraction of the supernumerary teeth and the deciduous with delayed exfoliation, followed by the surgical exposure of impacted permanent teeth and their orthodontic guided eruption. Extractions are not accomplished in one time: there must be a staged approach in order to maintain the vertical occlusal dimension while the different groups of unerupted teeth are exposed and pulled in their ideal position. If there isn\'t any, or just a mild, skeletal discrepancy between maxilla, mandible and cranium, the treatment is finished with the alignment of all permanent teeth, obtaining a correct occlusion and an agreeable smile aesthetics (\[[@B10]\]. In presence of an important skeletal discrepancy, most commonly a mandibular prognathism, that preclude the possibility to achieve an acceptable orthodontic camouflage, it\'s necessary to wait until the completion of skeletal growth and then restoring a correct bone position through orthognatic surgery, followed by the orthodontic finishing \[[@B11]-[@B14]\]. Traditional dental radiographs are very useful tools for the diagnosis of CCD, permitting to observe two features of the classical triad considered pathognomonic for diagnosis of this syndrome: multiple supernumerary teeth and open suture and fontanels of the skull (the third sign is the partial or complete absence of the clavicles). They also show other features helping in the diagnostic process like the presence of impacted teeth, the underdevelopment of maxillary sinuses and the parallelism of mandibular ramus, with an upward and posteriorly pointing coronoid process \[[@B15]\]. Unfortunately, especially when there are a lot of supernumerary teeth, traditional dental radiographs are not enough rich in details to allow correct planning the orthodontic treatment of patients in late mixed dentition. In these cases it\'s appropriate to use a multi-slice computed tomography (MSCT) scanner with an accurate tridimensional information regarding the anatomy of every single tooth, the spatial relation between adjacent teeth and face to the surrounding anatomical structures. These data are of crucial importance in order to perform the best orthodontic treatment for different reasons: they permit to surely identify the supernumeraries teeth, often with some hidden malformations like dilacerations or dental invagination; they give the exact position of the impacted teeth in relation to the roots of the adjacent erupted teeth, preventing damage of these roots during the forced orthodontic eruption of the impacted teeth; and they also define the position of every tooth inside the alveolar bone and near some important vascular vessels and nerve fibers, avoiding unexpected complications during the extraction or the surgical exposition of impacted teeth, like bone fracture, local hemorrhage or denervation. Unfortunately a conventional MSCT exam expose the patient to an high dose of x-ray, thereby limiting the application of this techniques only to the most complex cases. Recently a relatively new technique, the CBCT, reducing the dose of radiation adsorbed by the patient, had been improved by different manufacturer, obtaining good quality images. The principal difference between MSCT and CBCT is that conventional CT uses a fan of x-rays and a narrow detector, so multiple slices are stacked to obtain a complete image, whereas CBCT use a cone of x-rays and a two-dimensional square detector allowing a single rotation of the radiation source to capture an entire region of interest: hereby the total radiation is less important than with conventional CT \[[@B16]\]. Case Report =========== A 15-year-old man, with a diagnosis of Cleidocranial dysplasia, was referred to our department for orthodontic treatment. The patient was previously treated only with deciduous and supernumerary teeth extraction and four panoramic x-ray were taken in the last two years in order to check teeth development and eruptive direction (Figure [1](#F1){ref-type="fig"}, [2](#F2){ref-type="fig"}, [3](#F3){ref-type="fig"}, [4](#F4){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Orthopantomography at 13 years 2 months of age**. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Orthopantomography at 13 years 9 months of age**. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Orthopantomography at 14 years 1 months of age**. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Orthopantomography at 14 years 6 months of age**. :::  ::: Oral examination revealed a bilateral cross bite with marked teeth misalignment ------------------------------------------------------------------------------- Old radiograms examination shows the presence of several supernumeraries and impacted teeth: we were able to estimate the number of impacted teeth but it was not possible to exactly evaluate their morphology, in order to detect possible root reabsorption areas and to establish which were supernumeraries teeth and which not, and their anatomical relationship with other teeth and with some critical structures like mandibular nerve and foramen. On the basis of these findings a three-dimensional CBCT scan was obtained, in order to exactly recognize teeth anatomical anomalies, to decide which of them are to be extracted and to plan the surgical access. We decided to use the NEWTOM 3G (QR, Verona - Italy) scanner and we choose the spherical field of view (FOV) of 9\" (medium FOV) in order to obtain an image of only maxillary, mandibular and TMJ region with a unique acquisition. The scan was completed within 36 seconds and the time of exposure was of 3,6 seconds. The tube voltage was of 110 kV and the tube current was of 4,70 mA. We used the 2.11 version of the QR NNT program to visualize the most interesting axial sections showing the impacted and the supernumerary teeth (Figure [5](#F5){ref-type="fig"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **An axial section of the maxilla showing the relationship between the upper central left incisor and a supernumerary tooth**. :::  ::: Afterward we selected a limited area of the volume acquired (Figure [6](#F6){ref-type="fig"}, [7](#F7){ref-type="fig"}) and we utilized the study reconstruction function to prepare a 3D image of this volume, that can be rotated in all the directions to better visualize every possible perspective (Figure [8](#F8){ref-type="fig"}). ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Selection of the area to be studied - Frontal view**. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Selection of the area to be studied - Sagittal view**. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Sagittal section of the maxillary arch displayed in MIP mode: bottom left side view**. :::  ::: We also utilized the dynamic 3D function in order to obtain a panoramic overview of the intermaxillary dental relationship (Figure [9](#F9){ref-type="fig"}). ::: {#F9 .fig} Figure 9 ::: {.caption} ###### **3D dynamic setup of the maxilla-mandibular complex: front right side view**. :::  ::: Images analysis showed the presence of three supernumeraries - between 3.2 and 3.3 (Figure [10](#F10){ref-type="fig"}, [11](#F11){ref-type="fig"}), 3.3 and 3.4 (Figure [10](#F10){ref-type="fig"}, [12](#F12){ref-type="fig"}), 4.4 and 4.2 (Figure [13](#F13){ref-type="fig"}) - and one impacted tooth - 4.3 (Figure [14](#F14){ref-type="fig"}) - in the mandible, one supernumerary in the maxilla in 1.1 position (Figure [15](#F15){ref-type="fig"}). ::: {#F10 .fig} Figure 10 ::: {.caption} ###### **Panoramic view of left mandible**. :::  ::: ::: {#F11 .fig} Figure 11 ::: {.caption} ###### **Cross sections at 3.3 supernumerary level**. :::  ::: ::: {#F12 .fig} Figure 12 ::: {.caption} ###### **Cross sections at 3.4 supernumerary level**. :::  ::: ::: {#F13 .fig} Figure 13 ::: {.caption} ###### **4.3 impacted and supernumerary lingual view**. :::  ::: ::: {#F14 .fig} Figure 14 ::: {.caption} ###### **4.3 impacted and supernumerary vestibular view**. :::  ::: ::: {#F15 .fig} Figure 15 ::: {.caption} ###### **Cross sections at 1.1 supernumerary level**. :::  ::: Sections from 6 to 14 in figure nr. [7](#F7){ref-type="fig"} show that 3.3 supernumerary is positioned lingually between tooth 3.3 and tooth 3.2, that are inclined due to the presence of this supernumerary but with no sign of root reabsorption: surgical approach will be from the lingual side, paying attention to respect roots integrity of teeth 3.3 and 3.2, that are very close to the supernumerary. Sections from 21 to 24 in figure nr. [8](#F8){ref-type="fig"} show the relationship between supernumerary 3.4, tooth 3.4 and mental foramen. The supernumerary is situated lingually to tooth 3.4 and there is no contiguity with mental foramen, so its extraction could be performed safely with a lingual surgical approach; tooth 3.4 is buccally displaced, resulting in a thin cortical buccal bone, and there is a root reabsorption at the limit between the medial and the coronal third of the root: considering these findings treatment decision will consist in supernumerary 3.4 extraction and lingual repositioning of tooth 3.4 root, keeping in mind that, depending on root reabsorption, long term prognosis of this tooth is uncertain. Figures nr. [9](#F9){ref-type="fig"} and [10](#F10){ref-type="fig"} analysis shows that impacted 4.3 root is curved at the apical half: this could make more difficult or maybe even to block its orthodontic forced eruption. Figure nr. [11](#F11){ref-type="fig"} shows that supernumerary 1.1 extraction could be performed easily with no risk of damaging tooth 1.1. Discussion ========== The use of 3D computer assisted tomography is the best, and probably the only one, method permitting to elaborate a real individual orthodontic treatment plan for each CCD patient. It allows to precisely locate the impacted or ectopic teeth and therefore to perform a minimally invasive surgery and to plan the most effective orthodontic strategies \[[@B17],[@B18]\]. Therefore CT images permit to safely place titanium screw, that have been suggested by Kuroda \[[@B19]\] to be very useful as an absolute anchorage during the forced orthodontic traction of impacted teeth in order to reduce the patient\'s treatment time and psychological stress, in both maxilla and mandible, avoiding the risk of damaging during the screw insertion some important surrounding anatomical structures like dental roots, nerves and blood vessels. The shortcoming of the routinely use of this technique was related to the high radiation exposure of the patient, limiting the application only at complex cases. All other cases were studied with traditional dental radiology using the tube shift method (parallax technique), that taking two conventional radiographs permit to locate the impacted tooth location comparing the movement of this tooth respectively to the way in which the radiograph was taken. Rosenstein \[[@B20]\], using a technique suggested by Dado \[[@B21]\] for the study of the bone support of a tooth consisting on the visual inspection of a series of CT slices perpendicular to the root axis and the method of Bland and Altman \[[@B22]\] for assessing agreement of two different measures, found that traditional radiology is reliable in the cases where bone support is good or poor, but it\'s more inaccurate than a CT scan in the intermediate cases. Ericson and Kurol \[[@B23]\] documented that, even if traditional radiology permit in most cases to locate an impacted tooth, it frequently underestimate the presence and the extension of root resorption of the adjacent teeth. This dichotomy between conventional radiology and computed tomography can be overcome utilizing a CBCT system, that permit to obtain an accurate 3D reconstruction and several sagittal, frontal and axial view of the impacted tooth, with a radiation exposure level that lay between multi slice CT (MSCT) and conventional radiography \[[@B24]\]. Of course the radiation exposure of the patient always depends on the setting (kV, mA and seconds of administration) used during the radiological exam: therefore every comparison between \"general systems\", like CBCT vs conventional radiology or CBCT vs MSCT, can produce only relative and approximate results. In order to conduct a precise comparison it\'s necessary to specify all the setting machine data. For example Ludlow \[[@B25]\] compared several CBCT scanners, with different FOV, with an average panoramic dose that they found using a Planmeca Promax digital panoramic device: using a different panoramic device, especially if using a non digital one, would have led to different results. Therefore we utilized in this case report the medium NewTom 3G FOV of 9\" that, with an effective dose lower than the large FOV of 12\" (indicated by Ludlow as administering a lower effective dose face to the most of the others CBCT scanners included in his study), permits to perform a three-dimensional study of: the upper airways \[[@B26]\]; the temporomandibular joint (TMJ) morphology (Figure [16](#F16){ref-type="fig"}), that is useful especially in these patients that are candidate to a an orthognatic surgery because of the risk of consequent condylar resorption \[[@B27]\]; the ramus (Figure [17](#F17){ref-type="fig"}), that has been described by McNamara \[[@B15]\] as typical with nearly parallel borders in CCD patients. ::: {#F16 .fig} Figure 16 ::: {.caption} ###### **Coronal, sagittal and axial view of the left condyle**. :::  ::: ::: {#F17 .fig} Figure 17 ::: {.caption} ###### **Coronal, sagittal and axial view of the two mandibular ramus**. :::  ::: Conclusion ========== In CCD patients the use of reconstructed 3D images obtained by a CBCT exam for diagnosis and treatment planning has only scarcely been documented until now, so no evidence-based conclusion can be made based on the current literature. CCD diagnosis is frequently made during the early childhood or even at birth, consequently it\'s incorrect to state that CBCT could be useful in identifying CCD. On the other hand, there is a general agreement that 3D images allows to obtain a more accurate reconstruction of the real anatomy than traditional 2D radiologic images, really useful especially in patients with impacted and supernumeraries teeth like CCD ones, even if there is no agreement about an extensive use of CT exams. We present this case report to support the use of a less invasive CT exam, the low dose CBCT technology, in CCD patients in late mixed dentition undergoing orthodontic treatment and to promote the collection of sufficient data to come to a common agreement on the use of 3D radiological exam in these patients. Consent ======= Written informed consent was obtained from the patient for publication of this case report and accompanying images. A copy of the written consent is available for review by the Editor-in-Chief of this journal. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= All authors read and approved the final manuscript. DD conceived of the study, and participated in its design and coordination and helped to draft the manuscript. LL has been involved in drafting the manuscript and to collect the results from follow-up examinations. MP has been involved in revising the manuscript critically for important intellectual content. CP and PB have done substantial contributions to conception and design and interpretation of data.

PubMed Central

2024-06-05T04:04:19.509161

2011-2-27

PMC3053236

Background ========== The recent arthropods comprise four classes: the insects, the crustaceans, the myriapods and the chelicerates. In some phylogenies pycnogonids are suggested to comprise a fifth class of arthropods, in some other phylogenies they are closely grouped with the chelicerates \[[@B1]\]. The sister-group of the arthropods is represented by the onychophorans that lack the most characteristic feature of the arthropods - segmentation of the appendages (arthropodization) (e.g. \[[@B2]\]). Body segmentation, tagmosis, and arthropodization are thought to be among the main causes why the arthropods became the dominating metazoan group in species number, number of individuals and morphological diversity, on our planet. A segmented body, often in combination with tagmosis, probably allowed the arthropods to adapt to new environmental situations quickly by modification of single segments and their often-specialized appendages without disturbing their general bodyplan \[[@B3]\]. Despite the biological importance of the arthropods and the enormous number of published phylogenies, the relationships of the arthropod classes remain controversial. In particular, the position of the myriapods has changed often and dramatically during the last century (reviewed in e.g. \[[@B4],[@B5]\]). The myriapods were traditionally thought to represent the sister-group of the hexapods (Atelocerata or Tracheata theory) (e.g. \[[@B6],[@B7]\]). This hypothesis is exclusively based on morphological data such as the presence of tracheae and Malpighian tubules or the appendage-less tritocerebral segment (reviewed in e.g. \[[@B8],[@B9]\]). Myriapods were even placed with onychophorans and insects (Uniramia theory), suggesting that arthropods are polyphyletic \[[@B10],[@B11]\]. This latter theory appears however to have lost its credibility (e.g. \[[@B12]\]). Another current theory places myriapods and chelicerates as closely related sister-groups (Myriochelata or Parodoxopoda theory). This theory finds support in morphological as well as in molecular studies (e.g. \[[@B2],[@B13]-[@B16]\]). Nevertheless, most molecular and a number of morphological phylogenetic analyses argue strongly in favor of a close relationship of crustaceans and insects (either Tetraconata or Pancrustacea theory) (e.g. \[[@B17]-[@B23]\]). Note that it is important to distinguish a true sister-group relationship of insects and crustaceans (= Tetraconata) and an in-group relationship of insects and crustaceans (= Pancrustacea). Morphological features supporting the Atelocerata are now often considered to have convergently evolved. Tracheae, Malpighian tubules and the loss of the tritocerebral appendage for example are thought to represent independent adaptations in insects and myriapods necessary for a life on land \[[@B4],[@B5],[@B24]-[@B26]\]. A number of genes involved in the formation of the head segments have been identified in *Drosophila*(e.g. \[[@B27]-[@B30]\]) and subsequent studies suggested that these factors may play widely conserved roles in insects (e.g. \[[@B31]-[@B33]\]). One of the key players in anterior head development is the COE-family HLH transcription factor *collier*(aka *knot*) \[[@B29]\]. Flies deficient for *collier*(*col*) function lack ectodermal structures of the intercalary segment, and the expression of segment defining genes like *engrailed*and *wingless*is disturbed \[[@B34]\]. Very recently a study on function and expression of the orthologs of *col*in insects, a crustacean and a chelicerate suggested that early *col*function in the development of the intercalary segment is only present in insects \[[@B33]\]. In their paper Schaeper and colleagues conclude that the early function of *col*in head segmentation is most probably an insect novelty. In accord with the Pancrustacea hypothesis, the development of the limbless tritocerebral segment in myriapods is most likely convergent and thus likely based on a different genetic mechanism \[[@B33]\]. Our data on *col*expression in two myriapod species, the millipede *Glomeris marginata*and the distantly related centipede *Lithobius forficatus*now show that the early expression of *col*is present in both the insects and the myriapods. This finding may be seen as support for the traditional Atelocerata hypothesis, and thus arguing against a true in-group relationship of insects and crustaceans in the sense of the widely accepted Pancrustacea concept, or alternatively that the early expression of *col*in the tritocerebral segment of insects and myriapods may represent a case of convergence in gene deployment. Methods ======= Species husbandry and embryo treatment -------------------------------------- The handling of *Glomerismarginata*, *Lithobius forficatus*and *Euperipatoides kanangrensis*specimens is described in \[[@B35]\], \[[@B36]\] and \[[@B2]\] respectively. After oviposition embryos of both myriapod species were allowed to develop at room temperature. Staging was done after \[[@B35]\] for *Glomeris*, after \[[@B37]\] for *Lithobius*and after \[[@B38]\] for *Euperipatoides*. The developmental stage of all embryos was determined by using the dye DAPI (4\'-6-Diamidino-2-phenylindole). Gene cloning ------------ A fragment of the *collier*gene was isolated from *Glomeris*, *Euperipatoides*and *Lithobius*each with degenerate primers from cDNA (SuperScript First Strand kit, Invitrogen). The primers *col\_fw1*(GCN CAY TTY GAR AAR CAR CC) and *col\_bw1*(TTR TTR TGN ACR AAC ATR TTR TC) for the initial PCR and *col\_fw1*and *col\_bw2*(GAT RTC NCK NGG RTT NCC NGC) for a semi-nested PCR were used to isolate the *Glomeris*fragment. The *Euperipatoides*fragment was isolated using primers *col\_fw1*and *col\_bw1*in a single PCR reaction. The *Lithobius*fragment was isolated using primers *col\_fw1*and *col\_bw1*in a first and *col\_fw2*(CAR GGC CAR CCN GTN GAR ATH GAR) and *col\_bw1*in a semi-nested PCR. Sequences of the fragments were determined from both strands by means of Big Dye chemistry on an ABI3730XL analyser by a commercial sequencing service (Macrogen, Korea). Sequences are available in GenBank under the accession numbers [AM279685](AM279685) (*Gm-col*), [FN827160](FN827160) (*Lf-col*), [FN827161](FN827161) (*Ek-col*). In situ hybridization and nuclei staining ----------------------------------------- Whole mount in situ hybridization for all species was performed as described for *Glomeris*in \[[@B39]\]. The inner membrane of *Lithobius*embryos is (or becomes) very fragile after fixation. As a consequence it is often hard to remove the membrane completely. Unlike the case for *Glomeris*, however, this membrane does not disturb the in-situ hybridization procedure; it does not stain unspecifically or inhibit detection of specific staining. Embryos were analyzed under a Leica dissection microscope equipped with either an Axiocam (Zeiss) or a Leica DC100 digital camera. Brightness, contrast, and colour values were corrected in all images using the image processing software Adobe Photoshop CS2 (Version 9.0.1 for Apple Macintosh). Results ======= collier ------- cDNA fragments of the ortholog of the *Drosophila*gene *collier (col)*have been amplified by RT-PCR from the myriapods *Glomeris marginata*(millipede) and *Lithobius forficatus*(centipede) and the onychophoran *Euperipatoides kanangrensis*. Orthology of the gene fragments has been assessed by comparison with published *collier*sequences from various metazoan species. There appears to be no risk of mistaking the isolated fragments with genes other than *collier*; no other similar sequences or indeed paralog of *col*is present in the published genomes of any protostome species \[[@B40]\]. We therefore designate the corresponding genes as *Gm-collier, Lf-collier*and *Ek-collier*respectively. collier expression in Glomeris ------------------------------ Expression of *Glomeris collier*is already detectable at the blastoderm stage (stage 0) as a broad closed ring surrounding the egg perpendicularly to the anterior-posterior axis of the embryo. This ring is situated in the anterior of the future head. Expression of the ring is weaker but broader in the dorsal of the embryo (this expression later disappears completely) (Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1A, B, D-F). Expression of *col*is possibly already present at earlier stages, but in situ hybridization experiments for such stages are not workable. The dorsal extra-embryonic tissue does no longer express *col*(Figure [1A, a](#F1){ref-type="fig"}). At stage 1 the stripe starts fading from ventral tissue (Figure [1B](#F1){ref-type="fig"}), and at later stages up to stage 2 *col*is only visible as two separated shorter stripes (Figure [1C, C\'](#F1){ref-type="fig"}). Finally, at stage 3, this expression disappears completely and at this point the embryos do not express *col*at all. At late stage 4, expression appears *de novo*in form of dots in the trunk lateral to the appendages. Soon after, at stage 5, stripes of expression extend from there to a position anterior to the limb buds (Figure [1D, D\'](#F1){ref-type="fig"}). At this point expression in the central nervous system of the trunk also appears (Figure [1D, D\'](#F1){ref-type="fig"}). In stage 6 embryos this expression intensifies; in addition the nervous system of the head segments also expresses *col*in a complex pattern (Figure [1E](#F1){ref-type="fig"}), and two thin stripes of *col*appear at the anterior rim of the head (Figure [1E](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Expression of *Glomeris marginata collier***. Expression of *collier*(*col*) (A-E) and combined expression of *col*and *engrailed*(*en*) (F-H). All embryos except the embryo shown in (a) are in ventral view. (A) Stage 0.2 embryo. *col*is expressed in a broad stripe in the anterior of the developing embryo posterior to the ocular field (cf. A\' which shows a DAPI staining of the same embryo). (a) Same embryo as in A; lateral view. (B) Stage 1 embryo. Expression of *col*fades from the ventral-most tissue (asterisk) (cf. B\', DAPI counterstaining). (C) Stage 2 embryo. Last remnants of *col*-expression at the junction between mandibular and premandibular/intercalary segment. (D) Stage 5 embryo. Expression appears in the trunk segments in dorso-lateral patches (asterisks) and thin stripes in the anterior of the segments (double-arrowhead). Weak expression appears in the ventral nervous system (arrow) (cf. D\', DAPI counterstaining). (E) Stage 6 embryo. Expression in dorso-lateral patches and segmental stripes in the trunk remains (asterisks and double-arrowhead). Expression in the ventral nervous systems enhances (filled circle). Expression appears at the anterior rim of the head (arrowheads). (F) Stage 0.2 embryo. Expression of *col*virtually abuts expression of *en*in mandibular segment. The distance to the *en*-stripe in the antennal segment is small (arrow). Arrowheads point to *en*expression in anntennal and mandibular segments. (G) Stage 0.4 embryo. Expression of *col*does not abut the mandibular *en*stripe any longer; the distance to the antennal *en*-stripe is increased (arrow). Arrowheads as in F. (H) Stage 0.4 embryo. *col*-stripe is further narrowing; distance to mandibular *en*-stripe is increased. Arrow and arrowheads as in F. Note that at these early developmental stages *en*is not yet expressed in the premandibular/intercalary segment. For details on *en*expression in young stages see \[[@B35]\]. Abbreviations: an, antennal segment; cu, cumulus; md, mandibular segment; oc, ocular field; PS0, parasegment 0; saz, segment addition zone; T1, first trunk segment. :::  ::: The early stripe of *col*expression is situated in the anterior part of the mandibular (md) segment and the posterior part of the premandibular (pmd) segment (intercalary segment in insects) (Figure [1A](#F1){ref-type="fig"}). This is clear from the position of the *col*-stripe at later stages when the intersegmental indentations form (Figure [1B](#F1){ref-type="fig"}). We also provided a one-colour double staining using the segmental marker *engrailed*(*en*) in a series of early stage embryos (Figure [1F-H](#F1){ref-type="fig"}). Note that at this stage the *en*-stripe of the pmd segment has not yet formed (cf. \[[@B35]\]). Therefore it is clear that in the shown embryos the stripe between the antennal and md *en*-stripes represents expression of *col*. The area expressing early *col*(Figure [1A](#F1){ref-type="fig"}) is homologous to parasegment 0 of *Drosophila*(e.g. \[[@B34]\]). At subsequent stages the anterior-most and posterior-most expression of *col*disappears, so that as a consequence *col*expression does not abut *en*expression in the md segment any longer (Figure [1F-H](#F1){ref-type="fig"}). Instead, a clear gap is seen between the expression of *en*in these two segments and the expression of *col*covering the pmd/md boundary (Figure [1F-H](#F1){ref-type="fig"}). collier expression in Lithobius ------------------------------- As in the millipede *Glomeris*, an early stripe of *collier*is also detectable in the centipede *Lithobius*at the blastoderm stage (Figures [2A, A\'](#F2){ref-type="fig"} and Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1C, G-I). In the embryo shown (Figure [2A](#F2){ref-type="fig"}), expression is somewhat weaker in the ventral part of the future germband, similar to the situation in *Glomeris*(Figure [2A, A\'](#F2){ref-type="fig"}). We find that as in *Glomeris*this stripe forms a closed ring at the early blastoderm stage and, similar as in *Glomeris*, is weaker but broader in the future dorsal tissue (Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1J-L). Because of the limited number of available young *Lithobius*embryos, we were not able to trace this early expression to unambiguously determine in which segment it lies. However, some aspects suggest that the expression in *Glomeris*and *Lithobius*at early developmental stages is indeed in a homologous position. First, the stripe/ring of expression lies in the anterior of the blastoderm stage embryo (Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1). Second, the expression is weaker in dorsal tissue (Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1). Third, the expression is broader in dorsal tissue compared to ventral tissue (Additional file [1](#S1){ref-type="supplementary-material"}: Figure S1). Fourth, the most ventral expression starts disappearing at an early stage in both myriapod species. At later stages the stripe of *col*expression disappears from the anterior of the future head. In stage 1 embryos *col*is not expressed and in stage 2 embryos expression is only seen faintly anterior to the labrum. The latter expression differs from the situation in *Glomeris*, where *col*is not expressed anterior to the developing labrum (Figure [2B, B\'](#F2){ref-type="fig"}). Later, with beginning of blastokinesis and at least up to stage 6, expression appears and persists in the developing brain, the central nervous system and in segmental patches between the developing limbs (Figure [2C, C\', D](#F2){ref-type="fig"}). Altogether the expression profile of *Lithobius col*is very similar to that of *Glomeris col*with the one exception of dominant expression anterior to the labrum in *Lithobius*. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Expression of *Lithobius forficatus collier***. (A) Blastoderm stage (stage 0) embryo. Transversal stripe of *col*expression in the anterior of the developing embryo. (A\') DAPI counterstaining of embryo shown in A. (B) Stage 3.2 embryo. Expression anterior to the outgrowing labrum (arrowhead). Expression has completely disappeared from the intercalary segment (arrow). (B\') DAPI counterstaining of the embryo shown in B. Arrow and arrowhead as in B. (C) Stage 5 embryo (flat-mounted). Expression persists anterior to the labrum. Expression extends from there towards the anterior rim of the germ band (arrowhead). De novo expression in the intercalary segment (ic) (large arrow). Note that the ic is now situated ventral to the base of the antennae. Segmental expression in the trunk at level of the developing legs (small arrow). Expression in the central nervous system along the ventral midline (asterisk). (C\') DAPI counterstaining of the embryo shown in C. (D) Stage 6 embryo. Expression as shown for a stage 5 embryo persists at later developmental stages. :::  ::: Collier expression in Euperipatoides ------------------------------------ In contrast to *col*expression in myriapods, onychophoran *collier*is not expressed at early blastoderm stages in form of an anterior transversal stripe. However *col*is expressed in a fuzzy domain around the blastopore in an early gastrulation embryo (Figure [3A, A\'](#F3){ref-type="fig"}) (here referred to as stage 0 embryo; cf. e.g. \[[@B41]\] for early stages). The earliest expression in the head appears much later as two lateral domains in the future brain anlage of the early stage II embryo (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S2). Soon after in late stage II embryos this pattern transforms into a complex pattern in the developing brain (Figure [3B](#F3){ref-type="fig"}). This expression profile is similar to the expression in *Glomeris*and *Lithobius*at later developmental stages (cf. Figures [1E](#F1){ref-type="fig"} and [2C](#F2){ref-type="fig"}). In early stage III embryos *col*is also expressed in the anterior of the developing limbs (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S2). In late stage III embryos *col*is expressed in spots in the limb buds and along the trunk ventral to the limbs (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S2). The latter expression may be associated with the developing neuropil \[[@B42]\]. In early stage IV embryos this expression profile persists, with exception of the dots in the limbs that disappear again (Figure [3C](#F3){ref-type="fig"}). At late stage IV the initially continuous expression along the trunk disappears from the position of the limbs (Figure [3C, D](#F3){ref-type="fig"}). At this point expression in the head transforms into broad domains in the developing brain (Figure [3D](#F3){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Expression of *Euperipatoides kanangrensis collier***. (A) Stage 0 embryo. The blastopore begins to form (arrow). Expression of *col*is around the blastopore. (B) Stage II embryo, lateral view. Diffuse expression (or possibly background) in the head and the posterior segment addition zone (saz) (term introduced in \[[@B60]\]; also see \[[@B61]\]). Well-defined expression profile in the head forms two stripes at the anterior ridge of the embryo (arrow). Two clear dots of expression anteriodorsal to the stripes (arrowheads). (C) Early stage IV embryo; ventral view. Expression along the complete trunk is on the same level as the limbs. Note that this expression is discontinuous at the base of the limbs that do not express *col*(asterisks). (D) Late stage IV embryo; ventral view; flatmounted. Complex strong expression in the brain (large arrow), with a dominant spot-like domain in each hemisphere (small arrow) and faint dot-like expression at the anterior rim of the brain lobes (open arrowhead). Expression between the bases of the limbs persists, but still no expression in the base of the limbs (asterisks). Faint expression appears in the tips of all appendages and the two anterior-most lip bulges (filled circles). Expression begins in the central nervous system associated with the anterior trunk segments (or ventral organs; discussed in \[[@B42]\] (closed arrowhead). Note that unspecific expression appears on the surface of the antenna, possibly because of beginning cuticle formation. Abbreviations: bl, brain bulb; fap, frontal appendage; L1, first walking leg; L15, 15^th^walking leg; sp, slime papilla; saz, segment addition zone. :::  ::: Faint staining also appears at this stage in the tips of the legs, the slime papillae, the jaws, and in the ventral nervous system (or ventral organs; for a discussion on the contribution of this tissue to the nervous system see e.g. \[[@B42],[@B43]\]) (Figure [1D](#F1){ref-type="fig"}). Note that this staining as well as the staining in the antennae may be unspecific due to the beginning of cuticle development. Discussion ========== Conserved and derived expression patterns of collier in arthropods ------------------------------------------------------------------ Data on *collier*expression and function are now available from a wide range of metazoan animals. These data suggest that the unifying theme, the ancestral function of *col*, is associated with the development of the nervous system \[[@B29],[@B32],[@B33],[@B44]-[@B50]\]. Including this study, *collier*orthologs have been examined in representatives of all extant arthropod classes \[[@B29],[@B32],[@B33]\]. Function of *col*in muscle differentiation and wing patterning appears to be arthropod or even only *Drosophila*specific \[[@B47],[@B51]\]. In addition the function of *col*in the patterning of the head segments was argued to be an insect-specific feature \[[@B33]\]. In order to gain information on the ancestral expression patterns of *col*in arthropods we examined its expression in the onychophoran *Euperipatoides kanangrensis*. The onychophorans represent the sister-group to the arthropods and can therefore serve as outgroup to distinguish ancestral from derived features in arthropods (e.g. \[[@B2],[@B16],[@B21],[@B52]\]). Most of the observed expression patterns of *col*in the onychophoran *Euperipatoides*may be associated with the development of the nervous system. We can however not totally exclude the possibility that some of the *col*-expressing cells are involved in the development of other tissues than the nervous system, for example the mesoderm. Overall we find that most aspects of *col*expression seem to be conserved among arthropods and onychophorans, for example, expression in: 1) the anterior rim of the head lobes; 2) the developing brain; 3) the central nervous system of the trunk and 4) dorsolateral patches of the trunk. An obvious exception is the prominent expression of *col*anterior to the labrum in the centipede *Lithobius*. However the observed expression patterns in arthropods + onychophorans suggest at least partially conserved functions of *col*in this group. The involvement of *col*in head segmentation in insects and myriapods represents a novelty, since the expression of *col*is absent from the crustacean, the chelicerate and the onychophoran. The question now is how likely it is that such novelty would have evolved independently in these two assumed rather distantly related arthropod groups, i.e. is due to convergent evolution (discussed below). Early expression of collier in insects and myriapods: Support for the traditional Atelocerata concept? ------------------------------------------------------------------------------------------------------ It has long been known from manipulation studies in the fly *Drosophila melanogaster*that *collier*plays a crucial role in anterior head patterning and that a loss of *col*-function causes the loss of the head regions expressing *col*\[[@B29],[@B34],[@B53]\]. The recruitment of *col*expression in patterning the anterior head and the coincident formation of the limb-less intercalary segment was recently argued to represent a developmental novelty in insects \[[@B33]\]. This idea was supported by the finding that *col*has no early expression and consequently also no early function in head development in a chelicerate and a crustacean that both have retained their tritocerebral appendage, the pedipalp and the second antenna respectively \[[@B33]\]. Our findings in two distantly related myriapods, the millipede *Glomeris*and the *centipede Lithobius*(note that the *Lithobius*data are less well worked-out than the *Glomeris*data), contradict this assumption and instead argue in favour of a conserved expression of *col*in head patterning in both, hexapods and myriapods. Together with the data provided by \[[@B33]\] on a chelicerate and a crustacean, our onychophoran data further support the idea that such early expression of *col*is not a plesiomorphic character for arthropods but a derived character. Though the unique absence of the tritocerebral appendages in insects and myriapods is a long discussed common feature of these two arthropod classes, it was often considered as a mere convergence and not as a synapomorphy (e.g. \[[@B13],[@B24]-[@B26]\]). One of the strongest arguments for this assumption was that the \"simple\" loss of an appendage could easily be caused by any disturbance or mutation of the underlying genetic network needed for limb development \[[@B2],[@B39]\]. That arthropods lose or modify appendages is indeed frequent; in millipedes for example - but not in centipedes - the second maxilla is also missing. Consequently the lack of the tritocerebral appendage as possible synapomorphy for insects and myriapods was often, and obviously with some justification, understated in phylogenetic discussions (e.g. \[[@B54],[@B55]\]). Now however it appears that the *col*gene may be involved in the development of the limb-less tritocerebral segment in both insects and myriapods. The finding that the same genetic factor(s)/mechanism(s) are possibly involved in the formation of the tritocerebral segment elevates this feature from a likely convergence to a possible synapomorphic character. This may therefore add a molecular perspective to the body of hitherto exclusively morphological data supporting the Atelocerata. In fact the most parsimonious scenario in terms of requiring the fewest number of evolutionary events affecting *col*would be the single recruitment of *col*in the formation of the tritocerebral segment (Figure [4](#F4){ref-type="fig"}). Since this would argue against an insect-crustacean relationship in the sense of the now widely accepted Pancrustacea concept, a case of gene convergence must still be considered likely for the expression of *col*in the tritocerebral segment of insects and myriapods (Figure [4D](#F4){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Lack of the tritocerebral appendage and early *col*-expression plotted on current arthropod phylogenies**. A The Atelocerata concept: myriapods and insects are allied, crustaceans are basally branching mandibulates; only [one]{.underline} evolutionary event: gain of early *col*-function (ecf). B The Tetraconata concept: crustaceans and insects are allied sister-groups, myriapods are basally branching mandibulates; [two]{.underline} evolutionary events: 1) gain of ecf in the stem-mandibulate and 2) loss of ecf in crustaceans. C The Pancrustacea concept: crustaceans and insects are allied with insects representing an in-group of the crustaceans; [three]{.underline} evolutionary events: 1) gain of ecf in stem mandibulate followed by 2) a loss in the crustacean stem and finally 3) a re-gain in the insect lineage. D Convergence. Ecf does not represent a synapomorphy of insects and myriapods but evolved two times independently. Not shown: The Myriochelata concept: chelicerates and myriapods are allied and insects and crustaceans are allied; three evolutionary events in the non-convergence scenario: 1) gain of ecf in arthropod stem followed by 2) a loss in crustaceans and 3) chelicerates; Note that in the case of convergence Myriochelata is as likely as Pancrustacea/Tetraconata (shown in D). :::  ::: The conserved expression of *col*in the tritocerebral segments in insects may thus indeed represent an evolutionary novelty, but then the presence of *col*in the homologous region in myriapods has to be considered as another independently evolved evolutionary novelty as well. Contradictory data in arthropod phylogeny: A case of homology versus convergence -------------------------------------------------------------------------------- The data presented here support the traditional Atelocerata theory, as do a number of morphological studies. Other data support the Myriochelata hypothesis joining chelicerates with myriapods. However the majority of data available today, including some morphological studies and most nucleotide sequence analysis, clearly support the close relationship of insects and crustaceans (Tetraconata) or even consider the insects as an in-group of the crustaceans (Pancrustacea). Consequently some of the data supporting contradicting evolutionary relationships must be considered to be either artificial, incorrectly interpreted or the result of convergent evolution. Convergent evolution, or convergence, is a much-discussed possibility to explain the presence of morphological data contradicting the Tetraconata/Pancrustacea hypothesis. It describes a scenario where similar morphological structures evolved independently in not (closely) related organisms as a response to similar environmental conditions. But convergent evolution is of course not restricted to morphological features only but must also be reflected by the underlying genetic levels controlling morphology. It is often argued that single genes or even genetic networks, or part of it, may be involved in the development of non-homologous structures (e.g. \[[@B56],[@B57]\]). In other words during evolution a single gene may be recruited independently because of its given function. Likewise gene networks may be recruited because of the conserved interaction of genes (e.g. \[[@B3],[@B58],[@B59]\]). For the given case described in this paper this would mean that the *collier*gene could have been recruited independently in the formation of the appendage-less tritocerebral segment in insects and myriapods (Figure [4D](#F4){ref-type="fig"}). In that case the genetic network or at least part of it (the action of the *collier*gene) would be conserved (homologous), but the resulting modification of the tritocerebral segment, the lack of an appendage on this segment, would not. Further investigation of the function of *collier*, and the genetic network within which it operates, may answer this question in the future. If, as seems most likely, the formation of the appendage-less tritocerebal segment is convergent in myriapods and insects, the hint of the same genetic mechanism behind this convergence offers a rare and important insight into the genetic basis of convergence. The degree to which the genetic patterning mechanism matches the two cases may offer important insights into how genes and their regulatory apparatus are recruited during the origin of novelties. Conclusions =========== One of the key players in the development of the limb-less tritocerebral segment in insects (intercalary segment), the COE-family HLH transcription factor *collier*, is also specifically expressed in the homologous limb-less segment in myriapods. This finding contradicts the suggestion that the role of *col*in the development of the anterior head is an insect novelty \[[@B33]\]. Historically insects and myriapods have been united in the Atelocerata (or Tracheata), and the morphology of the tritocerebral segments was used as the main synapomorphy to support this group. Modern sequence-based phylogenetic analysis, however, now rather suggests a sister- or even in-group relationship of insects to crustaceans (Tetraconata or Pancrustacea). The apparently synapomorphic limb-less tritocerebral segment has been explained as an example of convergent evolution, since it appeared likely that a structure (like one of many appendages) could easily be lost independently. Our data question this argumentation, because we show that it is not only the mere loss of an appendage, but also the involvement of a specific gene that may argue in favour of the Atelocerata. This study shows that comprehensive data (and taxon) sampling is often crucial to allow secure evolutionary statements. Although in line with the current opinion, i.e. the Pancrustacea/Tetraconata hypothesis, the data by \[[@B33]\] somewhat prematurely concluded that the involvement of *col*in the formation of the tritocerebral segment in insects would represent an evolutionary novelty. Our data strengthen a possible synapomorphy (limb-less tritocerebral segment) for the unlikely Atelocerata concept, either challenging modern phylogenies, or presenting a complex case of parallel evolution. To find out which of either is the case must be subject of future investigation including an in-depth analysis of the genetic network involved in the formation of the tritocerebral segment in arthropods. Authors\' contributions ======================= RJ designed the study, conducted the experiments and wrote the first draft manuscript. WGMD and GEB were involved in data discussion and writing the final version of the manuscript. All authors approved the final version of the manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Figure S1: Expression of *Glomeris collier*and *Lithobius collier*at the blastoderm stage (stage 0)**. **A/a**Bright field (A) and DAPI fluorescent (a) picture of the same *Glomeris*embryo showing expression at blastoderm stage. The asterisk in the DAPI stained embryo marks the cumulus. **B/C**Anterior expression in a closed ring in a blastoderm stage embryo of *Glomeris*(B) and *Lithobius*(C) respectively. **D-F**showing the same *Glomeris*embryo from different angles: ventral view (D), lateral view (E) and dorsal view (F). **G-I**showing the same *Lithobius*embryo from different angles: ventral view (G), lateral view (H) and dorsal view (I). Note that in both species dorsal expression is weaker, but in a broader domain. J-L Schematic drawing showing conserved ring-morphology of *col*expression in *Glomeris*and *Lithobius*at the blastoderm stage. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Figure S2: Additional aspects of *collier*expression in *Euperipatoides***. **A**Early stage II embryo with beginning expression in the brain (black arrow). **B**Early stage III embryo. Arrowheads mark expression in the anterior of the developing limbs. **C**Late stage III embryo showing expression in a continuous anterior to posterior stripe ventral to jaw, slime papilla and walking limbs. Dot-like expression is also visible in the limbs. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We wish to thank all reviewers involved in the publication of this paper. Their comments were helpful and most appreciated. This work has been supported by the Swedish Research Council (VR: grant to GEB), the European Union via the Marie Curie Training network \"ZOONET\" (MRTN-CT-2004-005624 (to GEB, WGMD and RJ)) and the DFG via SFB 572 of the University of Cologne (to WGMD and RJ). The authors wish to thank Jean Joss, Rolf Ericsson and especially Noel Tait for their help during onychophoran collection.

PubMed Central

2024-06-05T04:04:19.511013

2011-2-24

PMC3053237

Background ========== Biliary atresia (BA) is a progressive, inflammatory, fibrosclerotic cholangiopathy resulting in complete obliteration of the extrahepatic bile ducts \[[@B1]\]. The obstruction of bile flow leads to worsening cholestasis, hepatic fibrosis, biliary cirrhosis, end-stage liver disease, and death within a few years \[[@B2]\]. Currently, Kasai operation or hepatoportoenterostomy constitutes the initial surgical treatment of choice for infants with BA. Although Kasai procedure can successfully establish bile flow to the gastrointestinal tract, a number of BA children progress to hepatic cirrhosis, portal hypertension and ultimately require liver transplantation \[[@B3]\]. To date, the etiology and pathogenesis of BA have not been completely understood; however, several mechanisms have been proposed including genetic defects, perinatal viral infections, morphogenic abnormalities, immune mediated bile duct injuries, and autoimmune disorders involving the bile ducts \[[@B4],[@B5]\]. Bile duct inflammation, cytokine responses, and bile acid toxicity are three major contributors of liver parenchymal destruction and hepatic fibrosis in BA patients \[[@B2]\]. After hepatic stellate cells (HSC) are activated, these key effecter cells in hepatic fibrogenesis are transformed into extracellular matrix-producing myofibroblast. This process results in the production and the accumulation of collagen and other extracellular matrix in liver parenchyma, thus initiating and perpetuating the liver fibrosis \[[@B6],[@B7]\]. Recent studies showed the role of adipokines in hepatic fibrogenesis of various chronic liver diseases \[[@B8]\]. In this study, we focused on a unique adipokine, adiponectin. Adiponectin, a 244 amino acid polypeptide, is the most abundant adipokine exclusively produced and secreted by adipocytes into systemic circulation in trimeric, hexameric, and larger multimeric high-molecular-weight (HMW) forms \[[@B9],[@B10]\]. Adiponectin is structurally homologue to tumor necrosis factor-α (TNF-α); however, these two molecules antagonize each other\'s effects in the target organs \[[@B11]\]. Adiponectin exerts its anti-inflammatory effects through the reduction of pro-inflammatory cytokines release including TNF-α and interleukin-6, and inducing the expression of anti-inflammatory cytokines, such as interleukin-10 \[[@B12],[@B13]\]. Adiponectin is also renowned for its anti-diabetic, anti-atherosclerotic, and anti-obesity effects. It is believed that adiponectin plays a protective role in liver diseases. In animal studies, adiponectin-knockout mice developed more severe carbon tetrachloride-induced liver fibrosis compared with wild type mice, and adiponectin injection prior to carbon tetrachloride treatment could prevent it \[[@B14]\]. In non-alcoholic obese mice, administration of recombinant adiponectin could attenuate hepatomegaly, hepatic steatosis, and aminotransferase abnormality \[[@B15]\]. Moreover, elevated adiponectin levels correlate positively with the severity of liver cirrhosis and negatively with hepatic protein synthesis \[[@B11],[@B16]\]. In contrast, low adiponectin levels have been shown in non-alcoholic fatty liver disease \[[@B17]\]. Adiponectin levels correlate negatively with liver fat and hepatic insulin resistance in diabetic patients \[[@B18]\]. Adiponectin is currently a subject of research interest since it has the potential to be a useful marker for liver fibrosis, and a possible target for a new therapeutic approach. Recently, it has been reported that a number of cytokines and growth factors have been studied in BA patients including osteopontin \[[@B19]\], basic fibroblast growth factor \[[@B20]\], and stem cell factor \[[@B21]\]. To the best of our knowledge, there have been no published studied on serum adiponectin levels from various clinical stages of BA. This is the first study to evaluate the correlation of serum adiponectin, liver stiffness and clinical outcomes in postoperative BA. In the present study, we postulated that serum adiponectin could be associated with the severity of clinical outcomes and the liver stiffness in BA patients, and to prove this hypothesis, we analyzed serum adiponectin and liver stiffness in BA patients compared with healthy controls. Therefore, the purpose of this study was to determine serum adiponectin levels collected from BA patients and to examine the possible correlations of serum adiponectin and outcome parameters of postoperative BA patients. Methods ======= All parents of children were informed of the purpose of the study and of any interventions involved in this study. Written informed consents were obtained from participants\' parents prior to the children entering the study. This study complied with the ethical guidelines of the 1975 Declaration of Helsinki, and was approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University. Study Population ---------------- Sixty BA patients (32 girls and 28 boys with mean age of 9.6 ± 0.7 years) and 20 healthy children (10 girls and 10 boys with mean age of 10.1 ± 0.7 years) were recruited in this study. All BA patients had undergone hepatic portojejunostomy with Roux-en-Y reconstruction (original Kasai procedure), and they were generally in good health; no signs of suspected infection or bleeding abnormalities at the time of blood sampling. None of them had undergone liver transplantation. Healthy controls who attended the Well Baby Clinic at King Chulalongkorn Memorial hospital for vaccination had normal physical findings and no underlying disease. BA patients were classified into two groups according to serum total bilirubin (TB), serum alanine aminotransferase (ALT), and liver stiffness score. Based on their jaundice status, BA children were divided into a non-jaundice group (TB \< 2 mg/dl) and a persistent jaundice group (TB ≥ 2 mg/dl). Subsequently, BA patients were categorized into a non-significant fibrosis group (liver stiffness \< 7 kPa) and a significant fibrosis group (liver stiffness≥7 kPa). The cut-off point of liver stiffness score for significant fibrosis was based on the study by Castera L, et al. \[[@B22]\] with sensitivity of 67% and specificity of 89%. Laboratory methods ------------------ After overnight fast, samples of peripheral venous blood were collected in the morning from every participant, centrifuged for 15 min at 1000 × g, and stored immediately at -80°C for further analysis. Quantitative determination of adiponectin concentration in serum was performed using commercially available enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Inc., Minneapolis, MN, USA). According to the manufacturer\'s protocol, 50 μl of recombinant human adiponectin standards and serum samples were pipetted into each well, which has been pre-coated with specific antibody for adiponectin. After incubating for 2 h at room temperature, every well was washed thoroughly with wash buffer for 4 times. 200 μl of a horseradish peroxidase-conjugated monoclonal antibody specific for adiponectin was then added to each well and incubated for a further 2 h at room temperature. After 4 washes, substrate solution was pipetted into the wells and then microplate was incubated for 30 min at room temperature with protection from light. Lastly, the reaction was stopped by the stop solution and the color intensity was measured with an automated microplate reader at 450 nm. The adiponectin concentration was determined by a standard optical density-concentration curve. Twofold serial dilutions of recombinant human adiponectin with a concentration of 3.9-250 ng/ml were used as standards. The manufacturer reported precision was 2.5-4.7% (intra-assay) and 5.8-6.9% (inter-assay). The sensitivity of this assay was 0.246 ng/ml. The liver function tests including serum albumin, total bilirubin (TB), direct bilirubin (DB), aspatate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase (ALP) were measured using a Hitachi 912 automated machine at the central laboratory of our hospital. The aspartate aminotransferase to platelets ratio index (APRI) was calculated as follows: (AST/upper limit of normal) × 100/platelet count (10^9^/l) \[[@B23]\]. Liver stiffness measurement --------------------------- Transient elastography (FibroScan) measured the liver stiffness between 25 to 65 mm from the skin surface, which is approximately equivalent to the volume of a cylinder of 1 cm wide and 4 cm long. The measurements were performed by placing a transducer probe of FibroScan on the intercostal space at the area of right lobe of liver with patients lying in a dorsal decubitus position with maximal abduction of right arm. The target location for measurement was a liver portion that was at least 6 cm thick, and was free of major vascular structures. The measurements were performed until 10 validated results were achieved with a success rate of at least 80%. The median value of 10 validated scores was considered the elastic modulus of the liver, and it was expressed in kilopascals (kPa). Statistical analysis -------------------- Statistical analysis was performed using SPSS software version 16.0 for Windows. Mann-Whitney U test was used to compare the difference of serum adiponectin concentrations between groups. Correlation between serum adiponectin levels and other serological markers, and liver stiffness scores were calculated using Pearson\'s correlation coefficient (*r*). Data were expressed as mean ± SEM. *P*-values \< 0.05 were considered to be statistically significant. Results ======= Comparisons between BA patients and healthy controls ---------------------------------------------------- A total of 60 BA patients and 20 healthy controls were enrolled in this study. The characteristics of participants in both groups were demonstrated in Table [1](#T1){ref-type="table"}. Mean age and gender ratio in controls and BA patients were not different whereas serum adiponectin levels in BA patients were markedly elevated compared with those in controls (15.5 ± 1.1 vs. 11.1 ± 1.1 μg/ml, *P*= 0.03) (Figure [1](#F1){ref-type="fig"}). Furthermore, BA patients had significantly higher hyaluronic acid than controls (50.3 ± 7.1 vs. 23.9 ± 1.5 ng/ml, *P*= 0.03). Additionally, liver stiffness scores in BA patients were dramatically higher than those in controls (30.1 ± 3.0 vs. 5.1 ± 0.5 kPa, *P*\< 0.001). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Demographic data, biochemical characteristics, and liver stiffness scores of controls and biliary atresia patients. ::: ---------------------------------------------------------------------- Variables Controls\ BA Patients\ *P*-value (n = 20) (n = 60) ------------------------------ ------------ -------------- ----------- Age (years) 10.1 ± 0.7 9.6 ± 0.7 0.1 Gender (Female: Male) 10:10 32:28 0.5 BMI (kg/m^2^) 18.5 ± 0.5 18.0 ± 0.6 0.5 Albumin (g/dl) \- 4.3 ± 0.1 NA Total bilirubin (mg/dl) \- 2.7 ± 0.5 NA Direct bilirubin (mg/dl) \- 2.2 ± 0.5 NA AST (IU/l) \- 128.1 ± 11.2 NA ALT (IU/l) \- 109.5 ± 10.7 NA ALP (IU/l) \- 431.0 ± 28.0 NA Platelet count (10^3^/mm^3^) \- 162.7 ± 12.8 NA APRI \- 3.0 ± 0.3 NA Adiponectin (μg/ml) 11.1 ± 1.1 15.5 ± 1.1 0.03 Hyaluronic acid (ng/ml) 23.9 ± 1.5 50.3 ± 7.1 0.03 Liver stiffness (kPa) 5.1 ± 0.5 30.1 ± 3.0 \<0.001 ---------------------------------------------------------------------- The data was expressed as mean ± SEM. BA, biliary atresia; BMI, body mass index; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; APRI, aspartate aminotransferase to platelets ratio index; NA, not applicable ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Comparison of serum adiponectin levels in biliary atresia patients and healthy controls**. The data was expressed as mean ± SEM. :::  ::: Comparisons between BA patients with and without persistent jaundice -------------------------------------------------------------------- We further categorized BA patients into jaundice (n = 20) and non-jaundice group (n = 40). As presented in Table [2](#T2){ref-type="table"}, BA patients with jaundice had significantly higher serum bilirubin, AST, ALT, ALP, APRI and liver stiffness values compared to those without jaundice. Moreover, serum adiponectin levels in BA patients with persistent jaundice were greater than those in BA patients without jaundice (24.4 ± 1.4 vs. 11.0 ± 0.7 μg/ml, *P*\< 0.001) (Figure [2](#F2){ref-type="fig"}). Similarly, BA patients with significant fibrosis (n = 44) possessed remarkably higher serum adiponectin than those with insignificant fibrosis (n = 16) (17.7 ± 1.2 vs. 9.4 ± 1.1 μg/ml, *P*\< 0.001). We also found that BA patients with persistent jaundice had substantially higher levels of serum hyaluronic acid than those without jaundice (93.2 ± 16.7 vs. 28.9 ± 3.4 ng/ml, *P*\< 0.001). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Comparison between biliary atresia patients without and with jaundice. ::: ------------------------------------------------------------------------------------------- Variables BA Patients\ BA Patients with Jaundice\ *P*-value without jaundice\ (n = 20) (n = 40) ------------------------------ ------------------- ---------------------------- ----------- Age (years) 9.6 ± 0.9 9.5 ± 1.1 0.5 Gender (Female: Male) 20:20 12:8 0.5 BMI (kg/m^2^) 18.2 ± 0.6 17.1 ± 0.6 0.1 Albumin (g/dl) 4.5 ± 0.1 3.9 ± 0.1 \<0.001 Total bilirubin (mg/dl) 0.7 ± 0.1 6.5 ± 1.2 \<0.001 Direct bilirubin (mg/dl) 0.4 ± 0.1 5.9 ± 1.2 \<0.001 AST (IU/l) 98.3 ± 11.6 187.7 ± 18.2 \<0.001 ALT (IU/l) 94.1 ± 9.7 140.3 ± 24.5 0.04 ALP (IU/l) 369.8 ± 34.8 553.6 ± 33.9 \<0.001 Platelet count (10^3^/mm^3^) 183.5 ± 15.5 121.2 ± 19.9 0.01 APRI 1.9 ± 0.4 5.0 ± 0.5 \<0.001 Adiponectin (μg/ml) 11.0 ± 0.7 24.4 ± 1.4 \<0.001 Hyaluronic acid (ng/ml) 28.9 ± 3.4 93.2 ± 16.7 \<0.001 Liver stiffness (kPa) 18.5 ± 2.4 53.3 ± 4.3 \<0.001 ------------------------------------------------------------------------------------------- The data was expressed as mean ± SEM. BA, biliary atresia; BMI, body mass index; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; APRI, aspartate aminotransferase to platelets ratio index ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Comparison of serum adiponectin levels in controls, biliary atresia patients without jaundice, and biliary atresia patients with jaundice**. The data was expressed as mean ± SEM. :::  ::: Further analysis demonstrated that serum adiponectin levels positively correlated with TB (*r*= 0.58, *P*\< 0.001), serum hyaluronic acid (*r*= 0.46, *P*\< 0.001), AST (*r*= 0.41, *P*= 0.001), ALP (*r*= 0.30, *P*= 0.02), and liver stiffness values (*r*= 0.60, *P*\< 0.001). Conversely, serum levels of adiponectin inversely correlated with serum albumin (*r*= -0.65, *P*\< 0.001). Correlations between serum adiponectin and total bilirubin, hyaluronic acid, liver stiffness, and serum albumin were shown in Figure [3](#F3){ref-type="fig"}. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Correlation analysis in biliary atresia patients**. Serum adiponectin is correlated with (A) total bilirubin (*r*= 0.58, *P*\< 0.001), (B) hyaluronic acid (*r*= 0.46, *P*\< 0.001), (C) liver stiffness (*r*= 0.60, *P*\< 0.001) and (D) albumin (*r*= -0.65, *P*\< 0.001) in patients with biliary atresia. :::  ::: Discussion ========== Biliary atresia is an intractable liver disorder affecting infants and children. Despite early diagnosis and successful Kasai operation, the great majority of BA patients inevitably develop liver fibrosis, portal hypertension, and liver failure. Therefore, the investigation of fibrogenic progression in BA is undoubtedly important. Liver biopsy is considered a gold standard in diagnosing liver fibrosis and determining its severity. However, it is a painful and invasive procedure with infrequent but possible life-threatening complications \[[@B24]\]. Furthermore, there have been questions concerning the accuracy of liver biopsy, which is adversely affected sampling errors, intra- and inter-observer variability. These problems may result in false staging of liver fibrosis \[[@B25],[@B22]\]. In an attempt to develop noninvasive methods for the assessment of liver fibrosis, transient elastography or FibroScan has emerged as the most promising tool in BA patients \[[@B26]\]. Transient elastography or FibroScan (Echosens, Paris, France) is a novel, rapid, and non-invasive technique for measuring the degree of liver fibrosis, and it can be performed in the out-patient setting. The transducer probe creates mild amplitude and low frequency (50 Hz) vibration, which induces an elastic shear wave in the tissues underneath. A pulse-echo ultrasound is used to follow the propagation of the elastic shear wave and to measure its velocity, which is in direct proportion to tissue stiffness \[[@B25],[@B22]\]. According to the equation E = ρV^2^(E = Elastic modulus, V = Shear velocity, ρ = mass density), the stiffer the tissue is, the faster the wave can pass through it. FibroScan can measure liver stiffness in a volume of 100 times bigger than that obtained from liver biopsy and is, therefore, a better representative for the whole liver parenchyma \[[@B22]\]. Hepatic stellate cell activation followed by extracellular matrix production and accumulation is a major mechanism contributing to the development of liver fibrosis \[[@B6]\]. A number of cytokines are believed to play essential roles in this process and they become topics of research interest in an attempt to evaluate the use of serum cytokine as a biochemical marker of liver fibrosis. In the present study, we investigated the relationship of serum adiponectin with clinical outcomes and liver stiffness scores in postoperative BA patients. Adiponectin - also known as complement-related protein 30 (Acrp30), adipose most abundant gene transcript (apM1) and adipoQ - is mostly synthesized by adipose tissue \[[@B27]\]. Various animal models and clinical researches showed that adiponectin mediated anti-obesity, anti-atherosclerotic, and anti-inflammatory effects \[[@B28]\]. Direct effects on hepatocytes via a specific receptor (AdipoR2 receptor) and anti-inflammatory properties are partly mediated by its antagonism against TNF-α \[[@B29]\]. This raises the postulation on the potential hepatoprotective role of adiponectin against liver fibrosis and cirrhosis. However, the possible role of adiponectin in the pathogenesis of BA remains as yet unclear. The present study showed that serum adiponectin levels were significantly elevated in BA patients compared with healthy controls. In addition, serum adiponectin levels were substantially higher in BA patients with persistent jaundice than those without jaundice. Subsequent analysis revealed that serum adiponectin was positively correlated with serum total bilirubin, suggesting that serum adiponectin was associated with jaundice status in BA patients. Furthermore, jaundice status in BA patients seems to be an indicator for intrahepatic biliary obstruction. Thus, these findings indicate that adiponectin may play a potential role in the pathogenesis of hepatocellular damage in BA. To our knowledge, this study demonstrates for the first time increased serum adiponectin levels in BA patients. We also found that serum adiponectin positively correlated with AST, ALP, hyaluronic acid, and liver stiffness, but negatively correlated with serum albumin. These results support that serum adiponectin is associated with clinical outcomes (jaundice status, hepatic dysfunction, and liver fibrosis) in BA patients post Kasai operation. Elevated serum adiponectin has been documented in various liver diseases, including acute hepatitis, chronic hepatitis, liver cirrhosis, hepatocellular carcinoma, and primary biliary cirrhosis \[[@B11],[@B30]-[@B32]\]. In agreement with our findings, Tacke and colleagues demonstrated that adiponectin was increased and associated with inflammation and hepatic damage in chronic liver disease \[[@B11]\]. Elevated adiponectin concentrations following bile duct ligation in mice and in human bile from cholestatic patients suggest that biliary secretion is involved in adiponectin clearance. High adiponectin levels in patients with liver cirrhosis correlated positively with the severity of cirrhosis and negatively with hepatic protein synthesis, indicating that adiponectin might be used as a marker for liver cell injury \[[@B11],[@B16]\]. These findings suggest that elevated circulating adiponectin is related to hepatic damage and hence reflects liver dysfunction. Accordingly, our results also revealed that serum adiponectin was positively associated with degree of liver stiffness determined using FibroScan. Liver stiffness values are well correlated with advanced stages of hepatic fibrosis and cirrhosis in children \[[@B33]\]. Wolf and coworkers showed that mice received adiponectin before concanavalin A treatment developed less hepatic damage \[[@B12]\]. Adiponectin expression was upregulated in concanavalin A-induced liver failure. Therefore, adiponectin could play a possible role in the regulation of hepatic inflammation. Future studies on HMW form of adiponectin may help identify more pieces of the inflammatory jigsaw of BA; nevertheless, the challenge remains to piece them together to originate a rational solid hypothesis pertaining to their exact role. Several possible mechanisms may be responsible for the significant elevation of adiponectin in BA patients, particularly in those with a poor outcome. Increased serum adiponectin could be attributable to imbalance between adiponectin production and adiponectin clearance. In advance BA stages, reduced biliary clearance of adiponectin may plausibly contribute to elevated serum adiponectin levels. Moreover, extrahepatic organs can produce and secrete adiponectin in circulation. The higher adiponectin levels might be regarded as indicating hepatic injury and cholestasis in BA patients. Further clinical studies could render more valuable information on the pathophysiological roles of adiponectin in BA. It should be noted, however, that there are some limitations in this study. Firstly, the sample size of participants was not large enough to arrive at definite conclusions. Secondly, we investigated only those subjects who attended King Chulalongkorn Memorial Hospital, a tertiary care center, for assessment or treatment of BA. As this investigation was designed as a cross-sectional study, therefore we could not determine the causal relationship between adiponectin and liver fibrosis. Other limitation would be the use of FibroScan instead of liver biopsy to evaluate the stage of liver fibrosis, which was not the definitive diagnosis. Furthermore, the sensitivity for significant fibrosis of 67% is not high enough to use Castera score as a screening tool of liver fibrosis in BA patients \[[@B22]\]. Prospective studies with a longitudinal design and hepatic expression of adiponectin would provide useful information on the role of adiponectin in hepatic fibrogenesis. Conclusion ========== This is the first study to demonstrate the elevation of serum adiponectin, hyaluronic acid, and liver stiffness values in BA patients. Serum adiponectin were correlated well with clinical parameters and the degree of liver fibrosis determined by FibroScan. Accordingly, serum adiponectin, hyaluronic acid, and transient elastrography could be used as non-invasive biomarkers reflecting the severity and progression of disease in BA patients post Kasai operation. Further studies will be needed to determine the precise role of adiponectin in the process of liver fibrogenesis. Abbreviations ============= ALT: alanine aminotransferase; ALP: alkaline phosphatase; APRI: aspartate aminotransferase to platelets ratio index; AST: aspatate aminotransferase; BA: biliary atresia; DB: direct bilirubin; HSC: hepatic stellate cells; TNF-α: tumor necrosis factor-α; TB: total bilirubin; kPa: kilopascals; HMW: high-molecular-weight. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= SH and MC have conceived the study, analyzed the data, and have written the manuscript. AT, KP, and WU performed laboratory analysis. VC, PV, and YP were involved in the diagnosis and recruitment of cases. YP was responsible for the design of the study. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-230X/11/16/prepub> Acknowledgements ================ This investigation is made possible by Ratchadaphiseksomphot Fund, Faculty of Medicine, Chulalongkorn University, Thailand Research Fund, the Commission on Higher Education, the National Research University Project of CHE and the Ratchadaphiseksomphot Endowment Fund (HR 1155A). The authors are profoundly grateful to Dr. Prachya Kongtawelert, Ms. Linda Vimolket, and the entire staff of the Center of Excellence in Clinical Virology, Chulalongkorn University and King Chulalongkorn Memorial Hospital for their combined effort in this study.

PubMed Central

2024-06-05T04:04:19.513967

2011-2-28

PMC3053238

Background ========== Venous thromboembolism (VTE) has an incidence of 1-2 per 1000 individuals per year but is close to 1% per annum in those aged over 70 years\[[@B1]\]. VTE is a leading cause of death in hospital inpatients\[[@B2]\] and is a major cause of morbidity and mortality particularly in older people and among those with cancer\[[@B1],[@B3]\]. Despite the fact that 70% of patients with VTE are aged over 60\[[@B1]\], there are few studies of risk factors in the elderly. Other studies have analyzed effects of pre-existing statin medication on incident VTE \[[@B4]-[@B6]\] or have compared statin use in case control studies of VTE\[[@B7]-[@B9]\]. A *post hoc*analysis of the Heart and Estrogen/progestin Replacement Study (HERS)\[[@B4]\] and analyses of other\[[@B5]-[@B8]\], but not all \[[@B9]\], population or case control studies indicated a decreased risk of VTE with statin use. A systematic review of observational studies suggested that statins may be useful in the prevention of VTE\[[@B10]\]. However, observational studies are prone to confounding hence randomized controlled trials are required to assess reliably the effects of statins on VTE risk. Recently a randomized control trial of rosuvastatin in the prevention of VTE, in middle-aged subjects with low LDL cholesterol and raised C-reactive protein levels, (JUPITER)\[[@B11]\] indicated that rosuvastatin significantly reduced the occurrence of symptomatic VTE, (hazard ratio 0.57, 95% confidence interval 0.37 - 0.86, p = 0.007). Reviews of this trial and of recent case control studies \[[@B12],[@B13]\] have renewed the debate on the efficacy of statins in the prevention of VTE and the call for analysis of prospective data. A meta-analysis suggested that statin treatment was likely to reduce the risk of VTE, however there was significant heterogeneity of study outcome \[[@B14]\], and as the majority of studies looked at middle-aged rather than elderly populations and there was no separate analysis by age, it was not certain that elderly people benefit. The Prospective Study of Pravastatin in the Elderly at Risk (PROSPER) was a multi-centre, randomized, double-blind, placebo-controlled trial of pravastatin in the prevention of vascular disease in the elderly\[[@B15]\]. The present study is an analysis of incident VTE in this population of men and women aged 70-82 using data from the PROSPER database. The aim was to determine whether pravastatin reduces VTE incidence in older people. In addition, we assessed the impact of clinical, hematological, lipid, inflammatory and vascular risk factors for incident VTE in older age. Methods ======= Subjects -------- The design and outcome of PROSPER is described elsewhere\[[@B15]-[@B17]\]. Men and women aged 70-82 (n = 5804) with evidence of pre-existing vascular disease or at least one major risk factor for vascular disease were randomized to placebo or pravastatin (40 mg per day) treatment. Follow up was 3.2 years on average. For the present study the full medical records were retrieved for all suspected VTE in the PROSPER database and examined by three clinicians in a pre-planned *post hoc*analysis. Events were categorized as definite VTE if death from VTE was confirmed on the death certificate or if there was recorded evidence from investigations such as ultrasound venography ventilation perfusion lung scans or computed tomography pulmonary angiography. Events were categorized as probable VTE when there was a record of at least 3 months\' continuous anticoagulant treatment with warfarin or heparin or a clinician diagnosis of VTE in the trial records plus documentation of venous insufficiency or venous leg ulceration but no confirmation by clinical test or anticoagulant treatment. A consensus of two clinical opinions was required to confirm the event categorization. Where data were available for the entire cohort (Figure [1](#F1){ref-type="fig"}, Table [1](#T1){ref-type="table"}), the utility of a parameter in predicting risk of VTE was assessed using all cases (n = 48 definite VTE, n = 72 combined definite and probable VTE) and non-cases (n = 5627) not on warfarin treatment. For hemostatic variables and IGF-1 (Table [2](#T2){ref-type="table"}) a nested case control study was performed. We matched each probable and definite case (n = 76) with 2 controls (n = 152) selected at random from all non-cases on the basis of age (using two-year age categories), smoking status and country of origin (Figure [1](#F1){ref-type="fig"}). We excluded those controls that matched with a probable case, leaving 96 matching 48 definite cases, 3 of these 96 controls were on warfarin and removed from the analysis leaving 93 controls (Figure [1](#F1){ref-type="fig"}). In order to assess the impact of cancer as a precipitating factor for VTE, cases were separated into those who did and those who did not develop cancer during the study (diagnosis confirmed by endpoint committee). The institutional ethics review boards of all centres approved the protocol \[the Argyll and Clyde Local Research Ethics Committee, the Glasgow Royal Infirmary Local Research Ethics Committee, Greater Glasgow Primary Care and Mental Health Research Ethics Committee, Lanarkshire Health Board Local Research Ethics Committee, Dumfries and Galloway Health Board Local Research Ethics Committee, Forth Valley Health Board Local Research Ethics Committee, METC board of Leiden University Medical Center and the Clinical Research Ethics Committee of The Cork Teaching Hospitals\], and all participants gave written informed consent. The protocol was consistent with the Declaration of Helsinki. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Flow diagram demonstrating identification of VTE cases, non-cases and controls in PROSPER**. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Variables at baseline in the whole cohort not on warfarin and risk of definite VTE ::: Cases (n = 48) Non-cases (n = 5627) Unadjusted Hazard ratio (95% CI) P-value ---------------------------------------------------------------- ---------------- ---------------------- ---------------------------------- ---------- ***Conventional risk factors*** Age (years) 74.9 (3.1) 75.3 (3.4) 0.83 (0.54, 1.28) 0.39 Body mass index (kg/m^2^) 28.5 (5.2) 26.8 (4.2) 1.09 (1.02, 1.15) 0.0075 Total cholesterol (mmol/L) 5.5 (1.1) 5.7 (0.9) 0.80 (0.57, 1.10) 0.17 Triglyceride (mmol/L) 1.5 (0.9) 1.5 (0.7) 0.86 (0.56, 1.34) 0.51 LDL (mmol/L) 3.6 (0.9) 3.8 (0.8) 0.78 (0.54, 1.12) 0.18 HDL (mmol/L) 1.3 (0.4) 1.3 (0.3) 1.04 (0.89, 1.21) 0.65 Systolic blood pressure (mmHg) 147.7 (21.9) 154.7 (21.8) 0.74 (0.57, 0.97) 0.027 Diastolic blood pressure (mmHg) 83.2 (12.4) 83.8 (11.4) 0.96 (0.74, 1.23) 0.73 Men 22 (45.8) 2694 (47.9) 0.94 (0.53, 1.66) 0.83 Current smoker 12 (25.0) 1519 (27.0) 0.93 (0.48, 1.79) 0.82 ***History of vascular disease*** History of hypertension 32 (66.7) 3493 (62.1) 1.22 (0.67, 2.22) 0.52 History of diabetes 5 (10.4) 603 (10.7) 1.00 (0.40, 2.53) 1.00 History of vascular disease 16 (33.3) 2446 (43.5) 0.66 (0.36, 1.20) 0.18 History of MI 7 (14.6) 722 (12.8) 1.19 (0.53, 2.65) 0.67 History of angina 13 (27.1) 1519 (27.0) 1.01 (0.54, 1.91) 0.97 History of CHD 15 (31.3) 1758 (31.2) 1.01 (0.55, 1.86) 0.97 History of claudication 2 (4.1) 371 (6.6) 0.63 (0.15, 2.58) 0.52 History of arterial surgery or amputation for vascular disease 2 (4.1) 312 (5.5) 0.75 (0.18, 3.10) 0.69 History of peripheral arterial disease 3 (6.3) 603 (10.7) 0.57 (0.18, 1.82) 0.34 History of stroke/TIA 4 (8.3) 615 (10.9) 0.75 (0.27, 2.10) 0.59 ***Disability and cognition*** IADL score 13.6 (1.1) 13.6 (1.0) 0.93 (0.72, 1.19) 0.55 Mini mental state exam (MMSE) score 27.6 (1.6) 28.0 (1.6) 0.84 (0.71, 0.99) 0.034 Barthel index score 19.7 (0.7) 19.8 (0.7) 0.93 (0.66, 1.30) 0.66 Years in education 15.0 (2.0) 15.1 (2.0) 0.98 (0.85, 1.13) 0.76 ***Country of origin*** Scotland 21 (43.8) 2466 (43.8) 4.26 (1.00, 18.21) 0.034 Ireland 25 (52.1) 2104 (37.4) 6.16 (1.46, 26.00) Netherlands 2 (4.2) 1057 (18.8) Referent ***Drug treatment*** Pravastatin treatment 28 (58.3) 2795 (49.7) 1.42 (0.80, 2.52) 0.23 ***Co-morbidities*** Cancer during the study 11 (22.9) 415 (7.4) 4.58 (2.33, 9.01) \<0.0001 ***Inflammation*** IL-6\* (pg/ml) 2.8 (2.1) 2.6 (1.9) 1.21 (0.80, 1.85) 0.37 sICAM-1\* (ng/mL) 358 (1.41) 372 (1.39) 0.74 (0.31, 1.76) 0.50 CRP\* (mg/L) 3.6 (3.1) 3.1 (3.1) 1.16 (0.90, 1.50) 0.25 Leptin\* (ng/mL) 16.9 (2.4) 13.4 (2.4) 1.34 (0.97, 1.85) 0.077 ***Lipids*** Lp(a)\* (mg/dL) 11.2 (2.8) 13.6 (3.5) 0.89 (0.71, 1.11) 0.30 Apo A1 (g/L) 1.3 (0.2) 1.3 (0.2) 0.92 (0.29, 2.97) 0.89 Apo B (g/L) 1.1 (0.2) 1.1 (0.2) 0.45 (0.12, 1.66) 0.23 ***Biochemistry*** Glucose\* (mmol/L) 5.2 (1.2) 5.3 (1.2) 0.58 (0.13, 2.58) 0.47 Creatinine (umol/L) 104.2 (23.5) 101.1 (22.3) 1.01 (0.99, 1.02) 0.33 Urea\*(mmol/L) 5.7 (1.4) 6.1 (1.3) 0.45 (0.16, 1.24) 0.12 ALT\* (U/L) 22.3 (1.5) 21.3 (1.5) 1.32 (0.69, 2.54) 0.40 AST\*(U/L) 24.3 (1.3) 23.5 (1.4) 1.38 (0.59, 3.21) 0.45 Creatine Kinase\* (U/L) 84.0 (1.6) 82.6 (1.6) 1.06 (0.60, 1.86) 0.84 Free T4\* (nmol/L) 17.6 (1.2) 16.3 (1.2) 1.77 (0.61, 5.15) 0.30 TSH\* (mU/L) 1.6 (2.0) 1.8 (2.4) 0.89 (0.66, 1.20) 0.45 ***Hematology*** Hemoglobin (g/dL) 14.2 (1.2) 14.0 (1.2) 1.15 (0.92, 1.44) 0.22 Hematocrit (L/L) 0.4 (0.03) 0.4 (0.04) 1.22 (0.90, 1.65) 0.19 Mean and standard deviation (SD), \*geometric mean (SD) calculated from the log transformed distribution or n (%) for categorical variables are shown. Hazard ratios and confidence intervals are for increases of 1 unit, except for age (increase of 5 years), SBP (increase of 20 mmHg), DBP (increase of 10 mmHg), hematocrit (increase of 0.04 L/L) and log free T4 (increase of 0.2 log unit). ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Plasma risk markers in nested case control study where subjects were not on warfarin ::: Cases (n = 48) Controls (n = 93) Odds ratio (95% CI) P-value --------------------- ---------------- ------------------- --------------------- --------- Factor VIII (iu/dL) 158.4 (42.0) 144.6 (38.2) 1.52 (1.01, 2.28) 0.044 Fibrinogen (g/L) 3.60 (0.61) 3.73 (0.79) 0.81 (0.54, 1.24) 0.34 PAI-1 (ng/mL) 58.5 (35.9) 58.8 (30.9) 0.97 (0.68, 1.38) 0.86 IGF-1 (ng/mL) 74.0 (19.2) 76.0 (24.3) 0.92 (0.63, 1.34) 0.66 Factor VII (iu/dL) 135.0 (32.8) 137.9 (32.5) 0.85 (0.56, 1.29) 0.45 Factor IX (iu/dL) 138.6 (30.0) 140.4 (26.8) 0.94 (0.66, 1.33) 0.72 APC ratio 3.83 (1.19) 3.87 (1.12) 0.96 (0.66, 1.40) 0.85 Means (SD) are shown. Unadjusted odds ratios for continuous variables with approximately 1 SD change (SD in control group) are shown. ::: Baseline data ------------- Baseline demographics, clinical history, activities of daily living (20-point Barthel index), instrumental activities of daily living (14-point IADL scale) and Mini Mental State Examination (MMSE) were collected as described previously\[[@B16]\]. Baseline blood analyses ----------------------- Baseline blood glucose, serum creatinine, urea, alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatine kinase, free T4, thyroid stimulating hormone, hemoglobin, hematocrit, white blood cell count and platelet count were carried out by routine methods\[[@B17]\]. Baseline plasma lipids and lipoproteins were analyzed as described previously\[[@B18]\]. Baseline apolipoproteins AI and B were assayed by turbidimetric assays (Hitachi/Roche) as described previously\[[@B18]\]. Plasminogen Activator Inhibitor-1, interleukin-6, soluble Intracellular Adhesion Molecule-1 (sICAM-1) and Insulin-like Growth Factor-1 were assayed by commercial ELISA (Biopool, R&D). Baseline plasma leptin was measured by an \'in house\' radioimmunoassay\[[@B19]\]. Baseline Factors VII, VIII & IX, fibrinogen and activated protein C (APC) ratio were assessed using commercial coagulation assays (MDA Coagulometer, Biomerieux, Basingstoke, UK). Baseline high sensitivity C-reactive protein (CRP)\[[@B20]\] and lipoprotein (a) (Lp(a))\[[@B21]\] were analyzed by automated particle-enhanced immunoturbimetric assay and latex agglutination assay respectively. For blood analyses listed in Table [1](#T1){ref-type="table"} data was already available in the PROSPER database for all participants. For the coagulation analytes and IGF-1, the analyses were carried out specifically for the purposes of the current study and were performed in cases and matched controls only due to limited resource. Statistical analysis -------------------- Data were analyzed using SAS vs9.1 (SAS Institute Inc, Cary NC). Where necessary continuous variables were transformed logarithmically to give a near-normal distribution of data for parametric analysis. Where data were available from the entire cohort, analyses were carried out using cases and non-cases (n = 5627) not using warfarin at baseline, including 2834 patients allocated to pravastatin and 2865 to placebo. The time to VTE was quantified by univariate hazard ratios and 95% confidence intervals calculated with Cox\'s proportional hazard model for each variable of interest. In the multivariate analyses, forward stepwise regression was undertaken, where all variables significant at the 5% level on univariate analysis were allowed to enter the model, to determine the subsets of variables that were independently associated with VTE. Where data were available only for the cases and controls (Factor VIII, fibrinogen, PAI-1, IGF-1, Factor VII, Factor IX, APC ratio), statistical analyses included calculation of the conditional logistic regression univariate odds ratio (which accounts for matching). For continuous variables, the odds ratio is that associated with 1 standard deviation difference (in the control group) in the variable of interest. A *post hoc*power calculation suggests that we had 60-97% power to detect an odds ratio of similar magnitude (0.4-0.6) to previously published observational studies \[[@B4]-[@B9]\]. Results ======= Pravastatin use and VTE ----------------------- There were 28 cases of definite VTE in 2834 non-warfarinised patients in the pravastatin group and 20 cases in 2865 allocated to placebo (Additional file [1](#S1){ref-type="supplementary-material"} Table S1). The unadjusted hazard ratio for VTE in the pravastatin versus placebo group was 1.42 (95% confidence interval 0.80, 2.52; p = 0.23) (Table [1](#T1){ref-type="table"}). The results were similar after adjusting for cancer diagnosis in a multivariate analysis \[1.36 (0.77, 2.42) p = 0.29\] and when analysing only those participants without an incident cancer \[1.51 (0.78, 2.90) p = 0.22\]. An analysis was also carried out in the group combining a definite and a probable diagnosis of VTE where the unadjusted hazard ratio for VTE in the pravastatin versus placebo group was 1.20 (0.75, 1.90), p = 0.45. Clinical and plasma risk markers in entire PROSPER cohort --------------------------------------------------------- Baseline data for definite VTE cases and non-cases in the PROSPER cohort are shown in Table [1](#T1){ref-type="table"}. Higher body mass index was associated with an increase in risk for VTE \[1.09 (1.02, 1.15) p = 0.0075\]. Systolic blood pressure (increase of 20 mmHg) was associated with a reduced risk of VTE \[0.74 (0.57, 0.97) p = 0.027. Higher MMSE score was associated with a reduced risk of VTE \[0.84 (0.71, 0.99) p = 0.034\]. Subjects living in Scotland were 4.26 (1.00, 18.21) and subjects in Ireland were 6.16 (1.46, 26.00) times more likely to have a VTE compared to subjects living in the Netherlands (p = 0.034). In the stepwise model only BMI \[1.09 (1.03, 1.16) p = 0.0053\], systolic blood pressure (increase of 20 mmHg) \[0.73 (0.56, 0.95) p = 0.021\] and country \[Scotland vs Netherlands 3.98 (0.93, 17.03) and Ireland vs Netherlands 5.62 (1.33, 23.76) p = 0.049\] were independently predictive of VTE. Analysis using a combined group of definite and possible VTE gave similar results, apart from the association with systolic blood pressure which was not a significant predictor of combined VTE in both univariate and multivariate analysis. Plasma risk markers in nested case control study ------------------------------------------------ Baseline data in the nested case (n = 48) control (n = 93) study for Factors VIII, fibrinogen, PAI-1, IGF-1, Factor VII, Factor IX and APC ratio are shown in Table [2](#T2){ref-type="table"}. Plasma factor VIII levels were associated with an increased risk of definite VTE \[1.52 (1.01, 2.28) p = 0.044\]. None of these risk markers was associated with an increased risk of combined definite and probable VTE. Risk factor profile in cancer-associated VTE and non cancer-associated VTE -------------------------------------------------------------------------- Cancer is a recognised risk factor for VTE. In the current study cases were divided into cancer-associated (n = 11) and non cancer-associated (n = 37) VTE. The hazard ratio for cancer as a predictor of VTE was 4.58 (2.33, 9.01), P \< 0.001 when comparing cases with non-cases (Table [1](#T1){ref-type="table"}). Including cancer as a variable in the stepwise multivariate model had no impact on the associations we observed. There were more new cases of cancer in the pravastatin group than in the placebo group \[[@B15]\] (Additional file [1](#S1){ref-type="supplementary-material"} Table S1) and since the etiology of VTE might differ between those with a diagnosis of cancer and those without we explored the possibility that baseline risk factor profile differed between cases who had cancer and cases who did not have cancer. We classified cancer-associated VTE as those individuals who had a VTE subsequent to their cancer diagnosis (n = 5) and those who had a VTE prior to their cancer diagnosis (n = 6), total 22.9%. There were no differences in risk factor profile between those who had the VTE before and those who had a VTE after their cancer diagnosis. Discussion ========== This study showed with data from a randomized controlled trial a lack of a protective effect of pravastatin on VTE incidence in the elderly \[1.42 (0.80, 2.52) p = 0.23\]. The recently published data from the JUPITER study\[[@B11]\] indicated that rosuvastatin significantly reduced the occurrence of symptomatic VTE \[0.57 (0.37, 0.86) p = 0.007\] in middle-aged subjects with low LDL cholesterol and raised C-reactive protein levels. The PROSPER study had 48 VTE events (from 18,363 person-years of follow-up in those not on warfarin), fewer than the 94 events seen in JUPITER (17,802 participants with median follow up of 1.9 years). However incidence rates were similar: 0.26% for PROSPER and 0.28% for JUPITER. In PROSPER new cancer diagnoses were more frequent in the pravastatin treated group than in the placebo treated group\[[@B15]\]. However the lack of effect of pravastatin on VTE incidence was also observed when only considering participants with no diagnosis of cancer. Interestingly a very large, unselected population-based cohort study (n = 129,288) looking at statin use and a number of health outcomes \[[@B22]\] found a similar lack of effect of statins on VTE incidence to that observed in PROSPER. The differing results in the PROSPER and JUPITER studies may be explained by the different characteristics of the populations. PROSPER was an elderly, high risk population. JUPITER was carried out in a younger, initially healthy population although they did note a similar, but not statistically significant, benefit in a high risk subgroup of elderly participants ≥ 70 years of age. In PROSPER, LDL cholesterol levels were approximately 1.0 mmol/L higher and CRP levels approximately 1.0 mg/L lower than those observed in the JUPITER study. It is possible that more VTE events in JUPITER had an underlying etiology involving pro-inflammatory pathways \[[@B23]\] that may be susceptible to statin intervention. Notably inflammation markers, and in particular CRP, were not independently linked to VTE risk in PROSPER. Since our study involved the use of pravastatin we cannot rule out the possibility that the apparent lack of efficacy in VTE prevention is confined to pravastatin alone. However both pravastatin and rosuvastatin which was used in JUPITER are water-soluble statins and thus differences in tissue distribution are unlikely to explain the difference between the effects of these drugs on VTE events. Individual data meta-analyses of VTE risk in randomized placebo controlled trials with statins would be useful to address this issue and the low power of single trials. Nevertheless, our results go against a class effect of statins in VTE which is somewhat surprising since all statins lessen cardiovascular disease risk including in the elderly. BMI, lower systolic blood pressure and country of origin were the only significant independent predictors of VTE in PROSPER. BMI was previously observed to be a risk factor for VTE in the prospective Copenhagen City Heart Study \[[@B24]\], Physicians\' Health Study\[[@B25]\] and LITE study\[[@B26]\] where the subjects were predominately aged \<70 years. BMI appears to be more strongly associated with VTE than with either stroke or coronary heart disease\[[@B25]\]. In our elderly population the magnitude of risk associated with BMI appears less than that observed in younger populations. It is possible that the association between BMI and VTE is attenuated by age. Many mechanisms by which obesity may increase the risk of VTE have been proposed, including promotion of inflammation by adipokines, increased coagulation activity, decreased fibrinolysis, increased oxidative stress, metabolic disturbances and endothelial dysfunction\[[@B27]\]. We found that low systolic blood pressure was an independent risk factor for incident VTE. This contrasts with prospective studies of VTE in younger adults, in whom a positive association has been observed with high blood pressure \[[@B24]\]. However it is possible that there is a change in the relationship of blood pressure with risk of VTE with advancing age. In support of this blood pressure levels in very elderly people have an apparently paradoxical inverse relationship with risk of other adverse clinical events including total mortality. Low blood pressure in elderly people is likely to be associated with more severe underlying covert disease including coronary heart disease; often these underlying diseases are unrecognized because there may be no classical symptoms \[[@B28]\]. It is likely that low blood pressure in older people is a risk marker for VTE, rather than directly causative. The lower incidence of definite VTE in the Netherlands (0.19%) compared to Scotland (0.84%) and Ireland (1.17%) is likely to reflect a healthier cohort in this country; in the Netherlands there was lower incidence of most recorded adverse outcomes, including ischemic vascular events, compared to the other 2 countries. Baseline BMI was not different between randomised subjects from each country \[[@B16]\]. Our finding that coagulation Factor VIII was associated with risk of definite DVT is consistent with reports from several other cross-sectional studies, and from two prospective studies (reviewed in \[[@B29]\]). There were limitations to our study. It is a *post hoc*analysis on a subgroup from the PROSPER study. The subjects were from a selected population with either evidence of pre-existing vascular disease or at least one major risk factor for vascular disease. Thus our results can only be applied to such high risk individuals. However this is the very subgroup of older people who are likely to be considered for statin therapy. The number of VTE is small and we may not be able to detect a small effect of pravastatin treatment. The incidence of definite VTE in the current study of 0.26% per annum is less than that predicted for the very old (1% per annum\[[@B1]\]) but equivalent to that observed in JUPITER (0.28% per annum). This may have been due to a higher proportion of relatively healthy and active individuals volunteering to take part in a randomized controlled trial, the use of cancer at baseline as an exclusion criterion and insufficient recorded data to identify all cases. The analysis presented here used definite VTE cases only however analysis using a combination of definite and probable VTE gave similar results. The PROSPER database did not have information on transient or precipitating risk factors, other than cancer, that may have preceded a VTE, such as immobilization or previous general surgery. Conclusions =========== In conclusion we found that pravastatin does not prevent VTE in elderly people at risk of vascular disease. Blood markers of haemostasis and inflammation are not strongly predictive of VTE in the elderly however BMI, country and lower systolic blood pressure are independently associated with VTE risk. Abbreviations ============= ALT: alanine aminotransferase; apo: apolipoprotein; APC: activated protein C; AST: aspartate aminotransferase; BMI: body mass index; CHD: coronary heart disease; CRP: C reactive protein; DBP: diastolic blood pressure; HDL: high density lipoprotein; IADL: instrumental activities of daily living score; IGF-1: insulin-like growth factor 1; IL-6: interleukin 6; LDL: low density lipoprotein; Lp(a): lipoprotein (a); MI: myocardial infarction; MMSE: mini mental state exam; PAI-I: plasminogen activator inhibitor 1; PROSPER: Prospective Study of Pravastatin in the Elderly at Risk; SBP: systolic blood pressure; sICAM-1: soluble intercellular adhesion molecule; SD: standard deviation; T4: thyroxine; TIA: transient ischemic attack; TSH: thyroid stimulating hormone; VTE: venous thromboembolism. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= DJF, IF, AG, IAG, GDOL and DJS were involved in the initial conception and design of the study. AJMdC, IF, AG, JWJ, NS and DJS were investigators in the PROSPER study. Data collection was carried out by EAB, ER, EST, MF and PES. IF and MR were responsible for data retrieval from the PROSPER database and statistical analysis. DJF, AG, IAG, GDOL, MR and DJS were responsible for the main data interpretation and analysis. All authors contributed to the writing of this manuscript. All authors have read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2318/11/8/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Table S1**. Characteristics at baseline by treatment group in PROSPER participants not on warfarin ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Sources of funding: Chest, Heart, Stroke Scotland Project Grant R05A89, Chief Scientist Office NHS Clinical Research Grant 2006, and a GlaxoSmithKline Clinical Fellowship and Tenovus Scotland research grant (to EST).

PubMed Central

2024-06-05T04:04:19.516178

2011-2-22

PMC3053239

Background ========== Oropharyngeal dysphagia is a common clinical problem, predominantly in the elderly, and is associated with numerous pathologies including, cerebrovascular accident (CVA; stroke), neurodegenerative diseases (e. g. Alzheimer\'s disease, Parkinson\'s disease and multiple sclerosis), certain advanced cancers, particularly head and neck, and may result from a traumatic injury \[[@B1]-[@B6]\]. Dysphagia has been most widely investigated in the context of CVA where the incidence of swallowing difficulties ranges from 40 to 70% of patients \[[@B7]-[@B9]\]. A major complication of dysphagia is aspiration which results in morbidity due to the development of pneumonia - referred to as aspiration pneumonia \[[@B8],[@B10]-[@B13]\]. The facts that aspiration occurs in about 40-50% of CVA patients and the mortality from aspiration pneumonia has been reported to be as high as 6% in the first year after a CVA, highlight the extent of the problem \[[@B11]-[@B18]\]. The causal relationship between aspiration and the development of pneumonia is well-established \[[@B19]-[@B22]\]. Likewise, it is well known that thin liquid is the most likely consistency to be aspirated \[[@B23]\]. The conventional treatment options that are recommended for patients that aspirate thin liquids are limited. Although postural interventions have been shown to be beneficial, the orthodox treatment remains consumption of thickened liquids, typically to an extremely, moderately or mildly thick consistency and a diet consisting of modified solids \[[@B24]-[@B26]\]. However, there are major concerns associated with the prescription of thickened fluids and modified solid consistencies \[[@B27],[@B28]\]. These are predominantly patient dissatisfaction and in certain cases non-compliance leading to medical problems associated with hydration and malnutrition \[[@B27],[@B28]\]. The problems associated with patient disapproval of thickened liquids and modified solid consistency diets are well known. They have created an enduring debate related to the management of people with dysphagia who aspirate on thin liquids. In an attempt to overcome these problems an authoritative protocol was developed at the Frazier Rehab Center in Louisville, US about 25 years ago for the management of patients with dysphagia. This free water protocol provides detailed guidelines and strict procedures for allowing access of water to people with dysphagia who, have insurance coverage as well as the ability to tolerate three hours of physical rehabilitation daily for six days per week. Although the protocol is based on a solid rationale for the provision of water and is supported by scientific research, albeit limited, there is still much confusion in the clinic. An important study published in 1997 addressed the issues related to the effects of water on aspiration pneumonia, hydration and quality of life in patients known to aspirate thin liquids \[[@B29]\]. This one year randomised-control study compared two groups of stroke patients who were known to aspirate on thin fluids \[[@B29]\]. The control group of ten had only thickened fluids and the study group of ten had thickened fluids as well as access to free water\[[@B29]\]. The findings indicated that there were no instances of pneumonia, dehydration or complications in either group\[[@B29]\]. However patient satisfaction was much superior in the study group whereas only one person in the control group was happy with the thickened fluids \[[@B29]\]. However, the limitation of this study is that it only involved 20 very carefully selected patients. Namely, only patients in a stroke rehabilitation unit or who had suffered a CVA within three weeks of admission were included in the study \[[@B29]\]. Patients with previous CVA, neurodegenerative diseases, multiple medical diagnoses and those unable to self-feed were excluded from the study. Due to the limitations of the study, even though there were no instances of pneumonia, dehydration or other complications, the major conclusion drawn was that access to water should only be provided to patients refusing thickened liquids or when hydration becomes an issue of concern \[[@B29]\]. Given that modified diets raise serious clinical management problems with respect to hydration and quality of life and in the absence of conclusive scientific evidence, numerous hospitals, medical centres and aged care homes have adopted policies allowing free access to water. However, this is not a uniform policy with the argument being that it is predominantly based on anecdotal support, since there is minimal documented evidence to reflect the practise. Therefore, we designed a randomised-control prospective study, to further investigate the effects of oral intake of water in people with dysphagia with previously identified aspiration. Methods ======= Institution and ethics ---------------------- The study was completed at a Regional Tertiary Teaching Hospital in the Gippsland region in Victoria, Australia. The application entitled \"effects of thin water in patients who have been prescribed thickened fluids by a Speech Pathologist to assess hydration and quality of life\", was submitted and approved by the local human research ethics committee. The study was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12608000107325) and informed consent was provided by the patients or medical enduring power of attorney before inclusion. Participants ------------ A total of 100 patients from the Regional Hospital in Victoria, Australia, were recruited in this study. People with dysphagia, who had been prescribed a modified or thickened fluid diet by a Speech Pathologist, over the age of 18 and without a diagnosis of chronic respiratory conditions or prior tracheostomy, were eligible for inclusion. From these, 91 completed the trial, with five being discharged before completion, two not willing to complete and two being excluded due to discomfort with ingestion of thin liquid. All patients were confirmed to aspirate thin liquids by independent assessment by two experienced speech pathologists. Aspiration of thin liquids was also verified in selected patients (10) using conventional radioactive barium videofluoroscopic swallow evaluations. Due to ethical considerations and the inherent dangers of the barium swallow test, videofluoroscopic examination is not routine for all patients exhibiting signs and symptoms of aspiration in the Regional hospital at which this study was completed. Patients undergo the procedure only when an experienced clinician is unsatisfied with the results of the assessment at the bedside examination. As mentioned, the bedside clinical examination was completed independently by two Speech Pathologists to determine the level of dysphagia and aspiration of thin liquids. Pulse oximetry and cervical auscultation were utilised as non-invasive tools to assist with the detection of silent aspiration and make a clinical judgement. We are aware of evidence regarding limitations surrounding both the pulse oximetry and cervical auscultation and where the Speech Pathologists were uncertain regarding detection of silent aspiration by clinical bedside assessment alone, a videofluoroscopy was performed. Additionally, 12 of the participants had a videofluoroscopic procedure completed within one month of being included in the study, indicating aspiration of thin liquids. Therefore, further videofluoroscopic studies were contraindicated due to the recent exposure to radiation and due to the severity dysphagia not having been altered upon bedside clinical examination. We also considered esophageal dysmotility by evaluating patient case histories, previous medical histories and current medications. Where gastroesophageal reflux was present or achalasia suspected, a barium swallow was completed (total of two participants) and the radiology reports indicated normal esophageal peristalsis. Initially patients were recruited from either an acute unit (total 15) or subacute units (total 85) and randomly assigned (age- and sex-matched) either to the control group (thickened fluids only; eight acute and 41 subacute) or to the intervention group (thickened fluids and free access to water; seven acute and 44 subacute). Once consent was obtained participant details were entered into a freely available database, a unique identifier was obtained and each participant was allocated into one of the study groups using the clinical trial randomizer. The research nurses and clinical speech pathologists were aware of the patient allocations however, medical staff completing the chest examinations described below, were blinded to the study. Participant groups and study interventions are outlined in Figure [1](#F1){ref-type="fig"}. As described in the discussion, the findings from the acute patients were not subjected to further analysis given the limitations and obvious conclusions that could be drawn from this group. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Outline of participant groupings and study methodology**. :::  ::: The characteristics of the subacute patients that participated in the study are summarized in Table [1](#T1){ref-type="table"}. A total of 76 subacute patients completed the study and were placed in either the control (n = 34; 18 female and 16 male) or intervention (n = 42; 19 female and 23 male) groups. The mean ± standard deviation ages were 79 ± 11 and 80 ± 7 years, in the control and intervention groups, respectively. An important issue for clarification is that of underlying clinical pathology. In Table [1](#T1){ref-type="table"} the major condition diagnosed on admission by the physicians, is noted for each patient. However, patients generally suffered from multiple ailments, the most common being CVA with accompanying depression and CVA and dementia. All of the patients had been prescribed modified thickened consistencies by qualified Speech Pathologists as indicated (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Control and intervention group subacute patient characteristics ::: Control Intervention ------------------------ --------- -------------- **Patients** Total 34 42 Age (mean ± SD) 79 ± 11 80 ± 7 Female 18 19 Male 16 23 **Pathology\*** CVA\*\* 15 25 Dementia 6 5 Alzheimer\'s disease 3 4 Parkinson\'s disease 1 4 Cancer 6 4 Motor neuron disease 1 0 Huntington\'s disease 1 0 Motor vehicle accident 1 0 **Oral** Own teeth 7 8 Dentures 25 31 No teeth 2 3 **Mobility** Mobile 13 26 Low mobility 13 12 Immobile 8 4 **Modified Diet** Puree/Honey 18 17 Puree/Pudding 3 2 Puree/Nectar 1 3 Minced/Honey 6 10 Minced/Nectar 2 3 Soft/Minced/Honey 1 3 Soft/Minced/Nectar 3 4 ::: Study Design ------------ The study was completed in two distinct stages. In the first, education sessions were provided to all nursing, medical and allied health staff allocated to areas in which people with dysphagia were to be admitted. Furthermore, an organisation wide oral hygiene protocol was implemented. Nursing staff were given specific training in the implementation or promotion of vigorous oral hygiene as well as strict parameters for the provision of water to participants in the study. Water could only be provided, at least 30 minutes after a meal, following thorough oral toilet which was identical for all patients across the control and intervention groups. The oral hygiene screening tool developed for nursing staff and utilised for all patients involved a thorough brushing of teeth or cleaning of dentures, to ensure there was no food build up or residue noted on or between teeth. Chlorhexidine mouthwash was used where necessary to ensure thorough cleaning of the oral cavity prior to provision of water. In the implementation phase (stage 2), the recruited patients were monitored for 72 hours while they consumed only the prescribed thickened fluids and modified consistencies. Following this initial pre-intervention phase (referred to as Days -3 to -1 throughout this paper), participants in the intervention group consumed the prescribed thickened modified diet and under strict nursing guidance, were given access to water as requested. Participants in the control group continued to consume the prescribed thickened diets only. All participants were monitored for a further five days (referred to as the post-intervention phase, Days 1 to 5 throughout this paper). Data collection --------------- Patients were examined for chest status daily by experienced physicians (blinded to patient assignment) and any irregularities were noted in the patient medical file. After careful consultation with four medical experts, a period of five days for the determination of aspiration pneumonia was recommended. The principle being that pneumonia would most likely develop within 24-48 hours of aspiration of thin liquids. Although we are aware that aspiration pneumonia may occur beyond the five day period, the opinion of experts was considered and adopted for the study. Further, nursing staff recorded core body temperature three times per day for the entire eight day period. In addition, total daily fluid intake for each participant was noted daily. Quality of life surveys ----------------------- Quality of life surveys were administered on the final day of the pre-intervention period (Day -1) and on the final day of the post-intervention period (Day 5). A total of 18 participants, 5 from the control group and 13 from the intervention group, completed the survey. Patients were asked the following series of four simple questions related to quality of life: 1\. How have you been feeling? 2\. Are you happy with the drinks? 3\. Have you been feeling thirsty? 4\. How clean does your mouth feel? Participants were asked to point to one of six faces that best represented their feeling and were assessed on the basis of the Wong and Baker, 1997, pain scale rating chart \[[@B30]\]. This test assesses pain on a zero to ten scale with an increment of two, with the use of drawn faces ranging from a smiley to a crying sad face, and was adapted to be utilised for the purpose of defining quality of life \[[@B30]\]. Statistical analysis -------------------- Paired t-tests were performed to calculate p-values for evaluating statistical significance for the hydration and quality of life surveys. The low response rate (overall 24%) for the quality of life surveys is likely to bias the data particularly given that only 5 from the control group and 13 from intervention group completed the surveys. However, it is noteworthy that the dissatisfaction of people with dysphagia to thickened diets is well known and non-compliance is a very common clinical problem. Results ======= Lung related complications -------------------------- Our findings indicate that six patients (14.3%; two female and four male) developed lung related complications in the intervention group, with three (7.1%) being diagnosed with aspiration pneumonia and three (7.1%) had lower quadrant bibasal crepitations (indicative of aspiration pneumonia but not confirmed) \[[@B31]\]. As expected an increase in mean core body temperature, corresponding to the time of diagnosis of the first signs of respiratory symptoms by experienced physicians, was noted in these patients (Figure [2](#F2){ref-type="fig"}). The increase in core body temperature and the first signs of aspiration pneumonia or bibasal crepitations typically occurred two to three days after the ingestion of water. In contrast, to the intervention group, none of the patients in the control group developed lung related problems during the study. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Core body temperature changes in patients that were diagnosed with aspiration pneumonia or bibasal crepitations**. The core body temperature of each patient on the study was monitored three times per day for the three day (Days -3 to -1) pre-intervention period and five days (Days 1 to 5) post-intervention. The mean daily temperature of the six patients in the intervention group (each patient is indicated by a different symbol), that developed aspiration pneumonia or bibasal crepitations is shown. The findings indicate a spike in mean core body temperature two to three days after the ingestion of water, corresponding to the time of the first signs of aspiration pneumonia or bibasal crepitations. In all cases, the rise in temperature subsided with the administration of antibiotics. :::  ::: Daily fluid intake ------------------ Patients in the control group were allowed only thickened fluids for the total eight day period (Days -3 to -1 pre-intervention and Days 1 to 5 post-intervention), and patients in the intervention group were allowed access to water for the five day post-intervention period. The total daily oral liquid intake by each of the participant in the study was noted. The findings indicate that the two groups had a similar intake of thickened fluids during the three day (Days -3 to -1) pre-intervention period with a mean ± standard deviation of 1340 ± 9.5 mL and 1428 ± 7.0 mL for the control and intervention cohorts, respectively. Comparison of the mean total daily oral liquid intake after the allowance of water (Days 1 to 5) for the intervention group highlights a significant difference with a mean ± standard deviation of 1378 ± 33.7 mL and 1767 ± 10.7 mL (P = \<0.001), for the control and intervention groups, respectively (Figure [3A](#F3){ref-type="fig"}). The increase in total fluid intake in the intervention group, from a daily mean of 1428 mL in the pre-intervention period to 1767 mL (P = \<0.001) during the intervention period, represents a modest decrease in thickened fluids to a mean ± standard deviation of 1185 ± 20.7 mL. However, adequate compensation in terms of hydration was provided with the supplementation of water which accounted for a daily mean ± standard deviation of 582 ± 15.8 mL during the post-intervention period (Figure [3B](#F3){ref-type="fig"}). It should be noted that two patients from the control group and one patient from the rehabilitation group required intravenous fluid during the study. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Total daily oral liquid intake by the patients in the study**. Comparison of the mean total daily oral liquid intake by patients in the control and intervention groups (A). Patients in the control group were allowed only thickened fluids for the total eight day period (Days -3 to -1 pre-intervention and Days 1 to 5 post-intervention). Patients in the intervention group were allowed access to water for the five day post-intervention period. The findings indicate a significant difference in the total fluid intake in the intervention group compared to the control group after the allowance of water (Days 1 to 5). Comparison of the mean proportions of daily intake of thickened and thin (water) fluids in the intervention group (B). The findings indicate only a modest decrease in the volume of daily intake of thickened fluids and an overall increase in the total fluid intake, following the allowance of water in the intervention group (Days -3 to -1 compared to Days 1 to 5). :::  ::: Quality of life --------------- Quality of life surveys were administered on the final day of the pre-intervention period (Day -1) and on the final day of the post-intervention period (Day 5). The survey comprised of four questions related to the feelings of the participants and the Wong and Baker, 1986, faces rating chart was used to calculate scores. The findings indicate that the patients in both the control and intervention groups were generally feeling okay (mean scores 5.6 and 4.9, respectively) at the end of the pre-intervention period in which all patient were on a thickened modified diet for three days (Figure [4](#F4){ref-type="fig"}). However, during the same period they were largely dissatisfied when asked about how they felt about the drinks (mean 6.4 and 7.1 for the control and intervention groups, respectively), their level of thirst (mean 8.8 and 7.4) and mouth cleanliness (mean 8.4 and 6.6). Profound differences were observed between the two groups in the follow-up survey at the end of the study. The intervention group reported remarkably higher levels of satisfaction with the drinks (mean 8.8 and 1.4 for the control and intervention groups, respectively; P = \<0.001), their level of thirst (mean 10 and 0.9; P = \<0.001) and mouth cleanliness (9.2 and 2.0; P = \<0.001) than the control group. Intriguingly, this did not appear to correspond to an overall increase in positive feeling. Although the intervention group did report a more positive response than the control group (means 5.8 and 8.4, respectively; P = 0.094), there appeared to be a slightly worse general feeling amongst the group compared to the pre-intervention period (means for the intervention group were 4.9 and 5.8 at the end of the pre-intervention and post-intervention periods, respectively; P = 0.111). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Results of the quality of life surveys for patients in the control and intervention groups**. Patients were asked a series of four simple questions related to quality of life on the final day of the pre-intervention period (Day -1) and on the final day of the post-intervention period (Day 5). Scores represent the mean of each group on the basis of the Wong and Baker, 1986, faces rating chart which assesses pain on a zero to ten scale with an increment of two; therefore, the lower scores represent lower pain or greater happiness (0 is no pain, represented by a smiley face and 10 is worst pain imaginable, represented by sad weeping face). It is interesting to note the remarkable differences in the two groups when asked a specific question (questions two to four), with the intervention group showing much greater satisfaction than the control group. In contrast, the difference is much smaller between the two groups, with the control group indicating slightly more positive responses than the intervention group, for the more general question one. :::  ::: Discussion ========== Our findings indicate that six subacute patients (14.3%) in the intervention group developed lung related complications during the study. The increase in core body temperature corresponding with the diagnosis of lung crepitations typically occurred two to three days after the ingestion of water, indicating that close monitoring of patients is critical during the initial 72 hour period from the deviation from the thickened fluid or modified diet. The significant difference in patients in the intervention (14.3%) compared to the control group (no cases), suggests a causal relationship between aspiration of water and the development of lung complications including aspiration pneumonia. This is in contrast with the findings from a previously published authoritative paper, in which it was found that none of the stroke patients with identified aspiration of thin liquids developed lung complications as a result of being given access to water \[[@B29]\]. The major differences between the two studies are the number of patients (ten per group in the previous study compared to 42 in the intervention group in this study) and the selection criteria. In the previous only newly admitted CVA patients were included and patients with previous CVA, neurodegenerative diseases, multiple medical diagnoses and those unable to self-feed were excluded from the study \[[@B29]\]. In contrast the majority of our participants had multiple medical diagnoses including disease other than CVA and patients with degenerative neurologic dysfunction and immobile patients were included. In our study, of the six participants (two female and four male) that developed lung complications in the intervention group, two had Alzheimer\'s disease, two had Parkinson\'s disease, one had a congenital intellectual disability and the other suffered from cancer. One of the Parkinson\'s disease and the cancer patients also had a CVA. Another feature of the patients that developed lung complications was that three were classified as being immobile and three had low mobility; immobile patients were bedridden and required assistance for feeding and patients with low mobility were predominantly bedridden but able to self-feed and walk with limited assistance. Given that patients with similar characteristics in the intervention group (seven) did not develop lung complications, suggests that neurological dysfunction and immobility or low mobility can be considered as increased risk factors but not a certainty for the development of aspiration-induced lung complications. Importantly, had we considered only the newly admitted CVA patients with relative mobility (total of 12) in the intervention group, our findings would indicate no cases of lung complications and be in complete accordance with the previously published findings \[[@B29]\]. One of the well known issues with the management of people with dysphagia is the discontent and in numerous cases the refusal to adhere to diets composed of only thickened fluids and modified solid consistencies \[[@B27],[@B28]\]. Apart from the ethical considerations, which are important and discussed below, this may represent serious issues with hydration. This was highlighted in our findings of daily fluid intake which indicate that patients in the intervention group consumed an average of \>300 mL more fluid after water was allowed compared to their intake in the pre-intervention period and compared to the fluid intake of the control group. It was noted previously, and quite surprisingly, that the access to water did not result in a dramatic decrease in the intake of thickened fluids and the intake of water did not exceed that of thickened fluids for any of the participants \[[@B29]\]. This is in accordance with our findings which indicate the mean intake of thickened fluids for the intervention group during the post-intervention period was 1185 ± 21 mL and mean water intake was 582 ± 16 mL. Similarly, the amount of water consumption did not exceed thickened fluid intake for any of the patients. Regarding hydration, only one patient required intravenous fluid in the intervention group, and this was during the pre-intervention phase when access to water was not allowed. Two patients in the control group required intravenous fluid. Only five (15%) and 13 (31%) patients from the control and intervention groups, respectively, completed the quality of life surveys in our study. The main reasons for this low completion rate are cognitive inability which accounts for many of the participants and also we were not persistent with patients who were experiencing difficulties on any of the two assessment days. Although this low completion rate may bias the results, the overall findings of the quality of life surveys in combination with our knowledge from extensive (over 10 years) clinical experience, indicate the dissatisfaction of patients to diets composed of only thickened fluids and modified solid consistencies. Remarkable differences were observed between the two groups in the follow-up survey at the end of the study with the intervention group reporting much higher levels of satisfaction with the drinks, their level of thirst and mouth cleanliness compared to the control group. As mentioned earlier, it is intriguing, that this did not appear to correspond to an overall increase in positive feeling. Admittedly, this was most likely due to a poor choice of generic question (How are you feeling?) which is, with hindsight, inappropriate for the cohort of participants. Finally, it is important to note that we initially designed the study in such a way as to collect data for an acute-phase cohort. However, the acute group represented unique problems related to data reliability, with the patients included in the intervention cohort refusing thickened consistencies and the majority requiring intravenous fluids. Although the acute patients were initially included in the study and we attempted to monitor them according to the protocol, serious complications including the development of aspiration pneumonia in the intervention group as a result of feeding at risk (i.e. they refused thickened fluids and modified solid consistencies) prompted us to re-think the strategy. Although ethical considerations related to quality of life are important, our observations highlight the need to strongly encourage thickened fluids and modified solid consistencies to acute patients. Given the complications with this group and the obvious recommendation, the data did not require analysis, no further consideration of this cohort is appropriate. Conclusions =========== Overall, our findings indicate a significantly increased risk in the development lung complications in patients given access to water when they are known to aspirate thin liquids. However, we appear to have narrowed down patients at highest risk, namely those with degenerative neurologic dysfunction who are immobile or have low mobility. On the basis of our findings we recommend that acute patients, patients with severe neurological dysfunction and immobility should be strongly encouraged to adhere to a thickened fluid or modified solid consistency diet. In the case of rehabilitation patients, a previous authoritative study concluded that access to water should be permitted when there is a refusal to consume thickened fluids or modified solid consistencies and when hydration becomes a medical problem. On the basis of our findings, particularly, our daily fluid intake and survey data, we propose an extension to this conclusion by recognising that the ethical considerations are too strong to ignore. We recommend subacute patients with relatively good mobility should have choice after being well-informed of the relative risk. Indeed, our current findings do not obviate the need to deviate from the Frazier Rehab Center free water protocol. Research aimed at further identification of subsets of patients with a high risk of developing aspiration pneumonia is warranted. In addition, clarification of risk factors for lung complications following the ingestion thin liquids, including water, in people with dysphagia is important. Perhaps, further consideration should be given to factors such as bacterial colonisation and immunological competence. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= MJP conceived of the study, and participated in its design, coordination and data collection. LC was involved in the study design and data collection. TCK was involved in study design and preparation of the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2318/11/9/prepub> Acknowledgements ================ The support of medical staff at the Latrobe Regional Hospital and the input from the Department of Statistics, Monash University, Gippsland, VIC, are greatly appreciated. The support of the Australian Institute of Nuclear Science and Engineering (AINSE) is acknowledged. TCK was the recipient of AINSE awards. Epigenomic Medicine is supported by the National Health and Medical Research Council of Australia (566559). MJPK and LC were both Speech Pathologists at the Latrobe Regional Hospital during the course of the study.

PubMed Central

2024-06-05T04:04:19.519352

2011-3-1

PMC3053240

Background ========== The prevalence of α^+^thalassemia allele in Peninsular Arabs varies between 0.07 and 0.58 \[[@B1]-[@B5]\]. In UAE nationals, the frequency of α^+^thalassemia allele is reported to be between 0.30 and 0.37 \[[@B1],[@B2]\]. In these populations, the most common mutations are deletional, involving one of the linked pair of α-globin genes. Consequently, the genotypes are heterozygous α^+^thalassemia (-α/αα), homozygous α^+^thalassemia (-α/-α) and normal genotype (αα/αα). In contrast, deletional mutations of both paired genes (-/αα), α^o^thalassemia, are absent in Gulf Arabs \[[@B1],[@B6]\]. Most α-globin gene mutations decrease the size of red cells and alter other indices \[[@B6]-[@B9]\]. α^+^Thalassemia homozygote is characterized by lower hemoglobin (Hb), lower mean corpuscular volume (MCV), lower mean corpuscular hemoglobin (MCH) and higher red blood cell (RBC) count. α^+^Thalassemia heterozygotes have red cell indices that are between the normal genotype and α^+^thalassemia homozygotes \[[@B7],[@B8]\]. A high frequency of α^+^thalassemia allele in a population often causes microcytosis and misinterpretation of the blood counts. This frequently leads to unnecessary testing and increases health costs. Additionally, in such populations the mean values of red cell indices are expected to be decreased and the standard deviations to be increased, as is also apparent from the reference intervals of red cells indices in two such populations \[[@B10],[@B11]\]. Therefore, it is inappropriate for populations with high frequency of α^+^thalassemia allele to use the \"Western\" erythroid standards, developed for people with little α^+^thalassemia. In populations with high frequency of α^+^thalassemia allele, the use of separate reference values for normal and α^+^thalassemia homozygotes, similar to those in use for different genders and age groups, is more appropriate. The erythroid standards for populations with high frequency of α^+^thalassemia are best defined with genotyping \[[@B8]\]. However, for most of them this is expensive and technically challenging. Nonetheless, the same goal may well be achieved with a combination of red cell phenotyping and mixture analysis. The resulting reference values, if properly validated, are still more appropriate than currently used standards derived for genetically different populations. This study is conducted to establish red cell reference standards for Emirati population by phenotyping and to validate them through comparison with the results of other studies. Methods ======= Setting and study population ---------------------------- Details on study subjects have been reported previously \[[@B12]\]. Briefly, data were collected from 1,079 native UAE citizens, ethnically Arab, 538 females, 539 males and two of unknown gender. The age (mean ± SD) was 24.3 ± 6.3 years (range, 11 - 69); only 5 individuals were younger than 15 years. The UAE population is tribal (67 tribes as per the 1968 census) and endogamous, and has high frequencies of α- and β-globin gene mutations \[[@B1],[@B2],[@B13]\]. Consanguineous marriages, which increase the likelihood of homozygosis, are common \[[@B14]\]. The government mandates and fully funds a premarital screening program for UAE citizens. The main purpose of the program is to decrease the incidence of *β*-thalassemia and sickle cell disease and marriages are not officially recognized without screening. All study subjects were participants of this program between March and August 2007. Study variables --------------- Blood was collected in EDTA-tubes. Complete blood counts were performed once on each subject, using the Cell-Dyn Sapphire (Abbot Diagnostics, USA) analyzer. The hospital laboratory subscribed to external quality control conducted by United Kingdom National External Quality Assessment Scheme and met the analytic standards. Hemoglobin analysis was performed using high-pressure liquid chromatography (Variant II, Biorad Co.). Genotyping was not performed in this program. Selection criteria ------------------ Subjects (39) with abnormal hemoglobin (hemoglobin A~2~\> 3.5% or presence of hemoglobin S, D or E) were excluded from analysis. As iron deficiency was relatively common among UAE females and iron measurements were not routinely performed in this program, RDW ≥ 14.0 was used to exclude subjects (136) with iron deficiency. This cutoff was the upper limit (mean+2SD) for our male subjects (see Results). In the absence of iron deficiency, MCV differences between males and females were not statistically significantly different, corroborating the assumption of equivalency of their red cell sizes \[[@B7],[@B8],[@B15]\]. Analytics --------- Erroneous or missing data were excluded: 13 RDW, five RBC counts, two MCV, six MCH and one hemoglobin A~2~. The study subjects were grouped and compared by gender and tribe, as identified by their last name. Standard descriptive and analytic statistical methods, such as histograms, linear regression and independent samples t-test were used. As the distribution of MCV was visibly bimodal, a statistical mixture analysis was performed using PC-Normix program <http://www.alumni.caltech.edu/~wolfe/normix.htm>\[[@B16]\]. This analysis can identify two or more clusters with normal (Gaussian) distributions within a mixed population. In this study, the distribution of MCV appeared to comprise only two clusters (Figure [1](#F1){ref-type="fig"}). Further analysis of red cell parameters in the two populations (clusters) was performed using SPSS for Windows, Version 17.1. The allele frequency was derived from the frequency of low-MCV phenotype which was assumed to represent α^+^thalassemia homozygote. These calculations were performed for whole population and for each of the ten tribes. The phenotype-derived genotype frequencies were then calculated using Hardy-Weinberg formula corrected for inbreeding,. viz. for α^+^thalassemia homozygotes q^2^(1 - F) + qF, for α^+^thalassemia heterozygotes 2pq(1 - F) and normal homozygotes p^2^(1 - F) + pF \[[@B17]\]. The mean coefficient of inbreeding (*F*) in this population was previously found to be 0.022 (14), which we assumed to apply. The level of significance was set at \<0.05. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Frequency distribution of MCV**. :::  ::: Ethical approval ---------------- The study was approved by Al Ain Medical District Human Research Ethics Committee. Results ======= The erythrocyte indices of 1,040 subjects without hemoglobinopathies other than α^+^thalassemia are shown in Table [1A](#T1){ref-type="table"}. Subjects with RDW ≥14.0% were presumed to have iron deficiency (Table [2](#T2){ref-type="table"}) and were excluded from estimating the reference intervals; an additional 8 subjects with incomplete data were also excluded. Their exclusion however only changed the values in the females (Table [1B](#T1){ref-type="table"}). The frequency distribution of MCV in the remaining 896 subjects (Table [1B](#T1){ref-type="table"}) showed two distinct subpopulations, but bimodal distribution of MCH was less distinct (Figure [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}); the distributions of hemoglobin and hematocrit were apparently homogenous. Thus MCV was chosen to separate phenotypically normal from small red cells. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Erythrocyte indices in the study subjects (A^a^) and in those with RDW \<14.0% (B^b^) ::: RBC(×10^6^/μL) Hb (g/dL) Hct (%) MCV (fL) MCH (pg) MCHC(%) RDW (%) ------- ----- ------- ---------------- ----------- ----------- ----------- ----------- ----------- ----------- **A** All 1,040 5.3 14 42.9 82.1 26.8 32.6 12.6 4.0-6.5 10.5-17.5 33.2-52.7 66.6-97.6 20.6-33.0 30.4-34.8 9.2-15.9 M 520 5.6 15.3 46.5 83.6 27.5 32.8 12.1 4.5-6.7 13.0-17.6 40.5-53.0 70.1-97.1 22.0-33.0 30.7-35.0 10.3-13.9 F 520 4.9 12.8 39.4 80.6 26.1 32.4 13.1 4.0-5.9 10.2-15.4 32.5-46.2 63.8-97.3 19.5-32.7 30.2-34.6 9.0-17.2 **B** All 896 5.3 14.4 43.8 83.4 27.3 32.3 12.1 4.1-6.5 11.4-17.3 35.4-52.2 70.2-96.7 22.0-32.7 30.7-34.8 10.5-13.7 M 495 5.6 15.3 46.6 83.8 27.6 32.8 12.0 4.5-6.7 13.1-17.4 40.4-52.8 70.2-96.7 22.2-32.9 30.7-35.0 10.6-13.4 F 401 4.9 13.2 40.4 82.9 27.1 32.6 12.2 4.0-5.8 11.3-15.1 35.2-45.6 69.6-96.3 21.7-32.5 30.7-34.6 10.5-14.0 ^a^Subjects with abnormal hemoglobin electrophoresis (n = 39) were excluded. ^b^Subjects with RDW \>14.0% (n = 136, Table 2) were presumed to have iron deficiency and were excluded from estimating the reference intervals. Erroneous or incomplete data (n = 8) were also exclude from estimating the reference intervals. ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Erythrocyte indices of subjects with RDW \> 14.0%. ::: RBC(×10^6^/μL) Hb (g/dL) Hct (%) MCV (fL) MCH (pg) MCHC(%) RDW (%) ----- ----- ---------------- ----------- ----------- ----------- ----------- ----------- ----------- All 136 5.1 11.6 36.7 72.1 22.8 31.5 16.1 3.8-6.4 8.1-15.0 26.9-46.5 57.1-87.0 16.9-28.7 29.2-33.9 11.7-20.4 M 14 5.9 14.0 44.2 75.2 23.8 31.6 15.2 4.1-7.8 9.1-18.9 30.4-58.1 59.0-91.4 17.7-29.9 29.3-33.9 12.4-18.0 F 122 5.0 11.3 35.7 71.7 22.6 31.5 16.2 4.0-6.0 8.6-13.9 28.5-43.0 57.0-86.3 16.8-28.5 29.1-33.9 11.7-20.6 ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Frequency distribution of MCH**. :::  ::: Mixture analysis of MCV of 896 subjects found78.0 fl to best separate phenotypically normal from small red cells in the sense that the estimated probability of being phenotypically normal exceeded 0.50 for values ≥78 fl. Using a P-P (probability) plot, the empirical cumulative probability distribution of our MCV data plotted against that obtained by mixture analysis yielded an almost perfect straight, diagonal, line, suggesting an excellent fit. Consequently, reference intervals for subjects with normal phenotypes and α^+^thalassemia homozygous phenotype were based on the estimated means and standard deviations of the two constituent normal distributions of the mixture distribution and are shown in Table [3](#T3){ref-type="table"}. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Erythrocyte indices in subjects with normal (A) and small (B) red cells^a^ ::: RBC(×10^6^/μL) Hb (g/dL) Hct (%) MCV (fL) MCH (pg) MCHC(%) RDW (%) ------- ----- ----- ---------------- ----------- ----------- ----------- ----------- ----------- ----------- **A** All 715 5.1 14.5 44.1 86.1 28.4 33.0 11.9 4.1-6.2 11.7-17.4 35.9-52.3 77.5-94.6 24.7-32.1 31.1-34.9 10.5-13.3 M 398 5.4 15.4 46.8 86.4 28.6 33.1 11.8 4.5-6.3 13.4-17.5 40.5-53.0 78.2-95.6 25.1-32.1 31.1-35.0 10.6-13.0 F 317 4.8 13.4 40.8 85.6 28.1 32.9 12.0 4.1-5.5 11.7-15.1 35.9-45.7 76.6-94.6 24.3-31.9 31.1-34.6 10.4-13.7 **B** All 181 5.9\* 13.6\* 42.6\* 73.0\* 23.2\* 31.9\* 12.8\* 4.6-7.1 10.7-16.4 33.8-51.4 67.2-78.7 20.8-25.7 30.2-33.5 11.4-14.1 M 97 6.3 14.6 45.8 73.1 23.3 31.9 12.7 5.4-7.1 12.8-16.4 40.1-51.6 67.2-79.0 20.9-25.7 30.2-33.5 11.4-14.1 F 84 5.4 12.4 39.0 72.8 23.2 31.8 12.9 4.5-6.2 10.6-14.2 33.7-44.2 67.1-78.4 20.6-25.7 30.2-33.4 11.6-14.3 ^a^MCV = 78.0 fl used as a separating value. \* P \< 0.0001 ::: When we treated the population as homogeneous, the estimated prevalence of phenotype-derived α^+^thalassemia homozygotes (181 of 896) was 0.20 and the prevalences of phenotype-derived α^+^thalassemia heterozygotes and normal homozygotes were 0.49 and 0.31, respectively; the estimated frequency of phenotype-derived α^+^thalassemia allele was 0.44. However, the prevalence of phenotype-derived α^+^thalassemia homozygotes varied substantially among the ten largest tribes (mean = 0.15), potentially vitiating the assumption of a single homogeneous population (Figure [3](#F3){ref-type="fig"}). Using stratification by tribe, the prevalence of phenotype-derived α^+^thalassemia heterozygotes was 0.45 and that of normal phenotype 0.40. Phenotype-derived α^+^thalassemia allele frequency in each tribe was adjusted for inbreeding and found to vary from 0 to 0.55. The aggregate frequency of the allele adjusted for population structure and inbreeding was 0.34. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Prevalence of phenotype-derived α^+^thalassemia homozygotes in ten largest tribes**. :::  ::: Discussion ========== The red cell reference intervals for male and female Emiratis are noticeably broader (Table [1](#T1){ref-type="table"}) than in other Caucasian populations in which α^+^thalassemia is rare \[[@B15]\]. Similar observations were reported on young adults in Saudi Arabia, in whom the frequency of α^+^thalassemia allele varies between 0.07 and 0.5, as well as in Palestinians \[[@B10],[@B11]\]. This finding is expected in any population with considerable variations in the number and size of red cells, i.e., α^+^thalassemia heterozygotes and homozygotes and normal homozygotes. In general, the effect of α^+^thalassemia allele frequency on the three genotypes and their aggregate effects on the mean values of red cell parameters are shown in Figure [4](#F4){ref-type="fig"}. This analysis shows that the standards developed in populations with frequent α^+^thalassemia are shifted to one side and wider, less precise, and critically depend on the frequency of α^+^thalassemia allele. Therefore, such populations require separate reference intervals, one for subjects with phenotypically normal red cells and another for those with small red cells. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Mean Hb, MCV and RBC values (thick lines) at different α^+^thalassemia allele frequencies**. Genotype frequencies (thin lines) are of αα/αα (normal homozygote), -α/αα (α^+^thalassemia heterozygote) and -α/-α (α^+^thalassemia homozygote). Data were created using Hardy-Weinberg equation and published values for the three genotypes \[[@B8],[@B18]\]. :::  ::: We separated two cell populations based on their size (MCV) rather than MCH, despite earlier reports that MCH is more useful in separating α^+^thalassemia homozygotes from normal homozygotes \[[@B8]\]. The main reason for preferring MCV over MCH was that in our sample the frequency distribution of MCV was more clearly bimodal than that of MCH (Figures [1](#F1){ref-type="fig"} and [2](#F2){ref-type="fig"}). As we were using mixture analysis (which breaks down a population into its constituent subpopulations by decomposing a frequency distribution into a mix of two normal distributions; here belonging to the normal and α^+^thalassemia homozygote phenotype subpopulations, respectively), MCV rather than MCH appeared a better parameter to accomplish this task. In the study that found MCH a more useful than MCV in separating normal from α^+^thalassemia homozygotes, the subjects\' genotypes were known and α^+^thalassemia heterozygotes were excluded from analysis \[[@B8]\], which we could not do. Nonetheless, the distributions of normal and small red cell populations clearly overlap (Figure [1](#F1){ref-type="fig"}). In our study, a value of 78.0 fl seems to best separate subjects with normal red cells from those with small red cells. The validity of this finding is supported by the finding in another study in which the same value of MCV best predicted α^+^thalassemia homozygote defined by genotyping \[[@B18]\]. When tribal population stratification was taken into account (Figure [3](#F3){ref-type="fig"}) and adjusted for inbreeding, phenotype-derived α^+^thalassemia allele frequency was estimated at 0.34, and. the prevalence of phenotype-derived α^+^thalassemia homozygotes at 0.12. These results are similar to reports of homozygote frequencies obtained using genotyping (0.11) and other phenotyping (0.14) methods on the same population \[[@B1],[@B2]\]. Heterogeneity in allele frequency among tribes may well be due to founder effects, with random numbers of α^+^thalassemia alleles segregated into subpopulations at the time of the foundation of current tribes. These differences were preserved by the practice of endogamy, which limits gene exchanges between the tribes. This substructure in the Emirati population (the consequence of tribal history) is also present in other Gulf Arab societies, and may explain reported variations in α^+^thalassemia frequency among different Arab populations. Remarkably, nearly half of the studied population is deduced to be α^+^thalassemia heterozygous. Although in clinical practice these individuals are indistinguishable from normal, their erythroid indices are between the normal and α^+^thalassemia homozygotes \[[@B7],[@B8]\]. Yet, contrary to expectations, the high prevalence of these phenotype-derived heterozygotes did not \"blur\" the bimodality of the distribution of MCV (Figure [1](#F1){ref-type="fig"}), suggesting that MCV values of most phenotype-derived α^+^thalassemia heterozygotes are well within the normal range. Indeed, in another study of red cell sizes in known genotypes, 64% of the α^+^thalassemia heterozygotes had MCV \>78.0 fl \[[@B18]\]. As expected, the erythroid parameters for phenotype-derived normal and phenotype-derived α^+^thalassemia homozygotes are significantly different (Table [3](#T3){ref-type="table"}). For phenotypically normal subjects, the reference intervals closely overlap with those for Caucasians in which α^+^thalassemia homozygosis is rare (Figure [5](#F5){ref-type="fig"}) \[[@B19]-[@B24]\]. Additionally, reference intervals markedly or completely overlap with the intervals published for adults genotyped as αα/αα (Figure [6](#F6){ref-type="fig"}) \[[@B7]\]. The results show that phenotypically normal Arabs have the same erythroid parameters as people of European origin. For phenotype-derived α^+^thalassemia homozygotes, the reference intervals overlap with those for adults with -α/-α genotype (Figure [7](#F7){ref-type="fig"}) \[[@B7]\]. These comparisons validate our results obtained with phenotyping and mixture analysis of phenotypes. The observed variations of reference intervals in Figures [5](#F5){ref-type="fig"}, [6](#F6){ref-type="fig"}, [7](#F7){ref-type="fig"} and in other studies would seem attributable to sample size, subject selection, and sample handling and processing. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Erythrocyte reference intervals for phenotypically normal Emiratis (thick lines, data from Table 3A) and Caucasians (thin lines)**. The RBC and Hb values are in males and the MCV and MCH values are in both genders. The numbers indicate references. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Erythrocyte reference intervals for phenotypically normal Emiratis (thick lines, data from Table 3A) and Spanish population with αα^/^αα genotype (thin lines, data from Reference**\[[@B7]\]). :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Erythrocyte reference intervals for Emiratis with phenotype-derived α^+^thalassemia homozygosis (thick lines, data from Table 3B) and Spanish population with -α/-α genotype (thin lines, data from Reference**\[[@B7]\]). :::  ::: A possibly contentious issue is the use of RDW ≥14.0% to identify iron deficiency. This unsatisfactory test may have introduced errors in estimating the prevalence of phenotype-derived α^+^thalassemia homozygotes. This bias however is likely to be small, as nine times more women than men are excluded (Table [2](#T2){ref-type="table"}), and the prevalence of phenotype-derived α^+^thalassemia homozygotes in two genders is not significantly different (*p*= 0.34). In general, in the absence of significant iron deficiency, which is more prevalent in women than in men, there is no evidence that MCV of men and women are different. Thus, the use of the upper limit of normal RDW in males (Table [1A](#T1){ref-type="table"}) to exclude iron deficiency in the females seems reasonable. A similar value for the upper limit of normality of RDW is found in Caucasian males of comparable age \[[@B15]\]. Conclusion ========== For clinical purposes, two sets of erythroid intervals are needed for populations with a high frequency of α^+^thalassemia allele. This study is the first to propose such reference intervals for clinical use, one for a population with normal and another for a population with small red cells (Table [3](#T3){ref-type="table"}). The studied population is heterogeneous as regards the α^+^thalassemia allele, a heterogeneity that differs by tribal alliance. This study demonstrates a new approach for the development of red cell reference standards through a combination of phenotyping and mixture analysis. The reference interval it produced appear to be consistent with those obtained by genotyping. In addition, α^+^thalassemia allele frequency estimates using this method are similar to those obtained by genotyping. Thus both results support the validity of this approach. Our results are applicable to other Gulf Arabs of the same origin, i.e. old Bedouin. Also, our methods can easily be utilized in other populations with a high frequency of α^+^thalassemia homozygotes in which genotyping is not feasible or affordable. Competing interests =================== The authors report no conflict of interest. The authors alone are responsible for the content and writing of this article. Authors\' contributions ======================= SD conceived, designed and organized the study; furthermore, analyzed the results and wrote the manuscript. AKS conceived, designed the study, analyzed the results and wrote the manuscript. NN conceived and performed the statistical analysis and helped with the draft of the manuscript. SS participated in the design of the study, performed the tests and data collection. GB participated in analysis of the study. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2326/11/1/prepub> Acknowledgements ================ This work was supported by the grant from the Faculty of Medicine and Health Sciences, United Arab Emirates University, UAE. We acknowledge helpful comments of reviewers.

PubMed Central

2024-06-05T04:04:19.522339

2011-2-24

PMC3053241

Background ========== Inflammatory responses to a wide variety of stimuli are largely attributable to up-regulation of the pro-inflammatory transcription nuclear factor kappaB (NF-κB). Specifically, reactive oxygen species (ROS) up-regulate the pro-inflammatory NF-κB transcription factor. The increased transport of NF-κB to the cell nucleus enhances expression of numerous genes encoding proteins that contribute to the inflammatory process, including inducible nitric oxide synthase (iNOS), cyclooxygenase-2 (COX-2), tumor necrosis factors (TNF-α, TNF-β), interleukins (IL-1, IL-6), chemokines (IL-8, MCP1, and MIP1α), activator protein-1 (AP-1) and adhesion factors (ICAM, and VCAM). Several of the proteins encoded by genes that are up-regulated by NF-κB are also potent NF-κB activators, thereby, forming an auto-activating loop. With ageing, the capacity to maintain a proper redox balance weakens with a concomitant up-regulation of NF-κB \[[@B1],[@B2]\]. Age-associated activation of NF-κB has the expected effect of increasing serum levels of TNF-α and nitric oxide (NO), and increased NO production has been observed during senescence \[[@B1]-[@B3]\]. These changes appear to be important as there is increasing evidence to support the concept that chronic, low-grade, and systemic inflammation contributes to the development of metabolic syndrome, dementia, cancer, atherosclerosis, osteoporosis, and other age-related diseases \[[@B1],[@B2]\]. NF-κB is normally activated by degradation of inhibitory kappaB (IκB). When this occurs, NF-κB translocates to the nucleus and binds to specific promoter regions of genes encoding pro-inflammatory proteins. Over the past decade, investigators have identified a number of compounds that selectively interfere with the NF-κB pathway. Several of these compounds are plant-derived antioxidants \[[@B4]\]. Quercetin and other flavonoids \[[@B5]-[@B10]\], and tocotrienols \[[@B11]-[@B18]\] suppress protein kinase-mediated degradation of inhibitory kappaB (IκB). This prevents NF-κB activation, and the corresponding increase in production of various inflammatory proteins \[[@B7],[@B9]\]. These ubiquitous plant constituents have been reported to suppress the progression of age-related diseases \[[@B19]\]. Quercetin and related flavonoids improve mental acuity \[[@B20]\], promote bone health \[[@B21]\], and attenuate carcinogenesis \[[@B7],[@B22],[@B23]\], development of atherosclerotic plaque \[[@B22]-[@B24]\], and diabetes \[[@B25]\]. Tocols, principally the tocotrienols, attenuate diabetic neuropathy \[[@B16]\], slow diabetes-associated cognitive decline \[[@B17]\], protect neurons from glutamate toxicity \[[@B26]\], and support bone formation \[[@B27]\]. The impact of α-, β-, γ- and δ-tocotrienols, the naturally-occurring farnesylated (unsaturated side-chain) analogs of α-, β-, γ- and δ-tocopherols, on atherosclerosis \[[@B28]-[@B34]\] and cancer \[[@B35]-[@B39]\] has been reported and reviewed by several investigators (30-39). The antioxidant activity associated with all tocols suppresses protein kinase-mediated inhibitory kappaB (IκB) degradation and subsequent NF-κB activation. Tocotrienols appear to be unique among the tocols, as they suppress hepatic HMG-CoA reductase activity and the synthesis of mevalonate-derived products through a yet to be defined \[[@B40]\] post-transcriptional mechanism \[[@B41]\]. As a consequence, tocotrienols uniquely lower serum total cholesterol and LDL-cholesterol levels \[[@B31]-[@B34],[@B42]-[@B46]\]. Moreover, when evaluated *in vitro*, tocotrienols suppress the synthesis of inflammatory cytokines with greater potency than the tocopherols \[[@B11]-[@B17],[@B47]\]. As delineated above, tocopherols and tocotrienols suppress protein kinase-mediated IκB degradation and concomitant NF-κB activation initiated by reactive oxygen species (ROS). The tocotrienols additionally suppress TNF-α stimulated NF-κB activation. This suppression is reversed by treating with mevalonate and other products of HMG-CoA reductase activity \[[@B12],[@B14]\]. This reversal is consistent with findings that Rho GTPases involved in the regulation of NF-κB require a post-translational modification, specifically prenylation, in order to be active \[[@B48]\]. δ-Tocotrienol, quercetin, riboflavin, amiloride, dexamethasone and (-) Corey lactone (an intermediate in prostanoid synthesis), are all FDA approved compounds and have been used by humans for several decades without any alarming adverse affects. All these compounds inhibit lipopolysaccharide (LPS)-induced TNF-α and iNOS gene expression in macrophages prepared from RAW 264.7 cells, BALB/c and C57BL/6 mice \[unpublished results\]. These key mediators of inflammation are up-regulated during the ageing process \[[@B1]-[@B3]\]. Therefore, we now evaluate the impact of the aforementioned compounds on serum NO and TNF-α levels. We tested the hypothesis that a dietary blend consisting of δ-tocotrienol plus quercetin, riboflavin, (-) Corey lactone, amiloride, or dexamethasone (Figure [1](#F1){ref-type="fig"}) would be more effective than the individual compounds in lowering serum NO and TNF-α levels. As described earlier, δ-tocotrienol additionally suppresses cholesterol synthesis and concomitantly lowers serum total cholesterol and LDL-cholesterol levels in various experimental animal models and humans \[[@B41]-[@B45]\]. We focused on the impact of these compounds on serum levels of TNF-α, NO, total cholesterol, LDL-cholesterol, and triglyceride in young female chickens. This avian model, chickens \[[@B49]\], differing from the widely employed rodent model \[[@B50]\], closely reflects human lipid metabolism. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Chemical structures of various compounds used in this study**. :::  ::: Materials and methods ===================== Materials --------- Sources of all chemicals, substrates, and diagnostic kits have been identified previously \[[@B51]\]. Chemicals and solvents were of analytical grade. Riboflavin, (-) Corey lactone = 4-phenylbenzoate alcohol = (3aR,4S,5R,6aS)-hexahydro-4-hydroxymethyl-5-(4-phenylbenzoyloxy) cyclopenta \[b\] furon-2-one\], amiloride, and dexamethasone were purchased from Sigma-Aldrich (St. Louis, MO, USA). Quercetin was purchased from Alfa Aesar (Johnson Matthey Co. Lancastor, UK); lipid extracts of annatto seeds consisting of 50% δ-tocotrienol were purchased from American River Nutrition (Hadley, MA, USA). Isolation of δ-tocotrienol from lipid extract of annatto seeds -------------------------------------------------------------- Silica gel (Merck, 230-400 mesh, 60 Å, 500 g) suspended in 1000 mL of hexane was poured into a 2-L glass funnel with a fritted disk. The gel was washed with 2 L of hexane prior to being loaded with 100 g of the 50% δ-tocotrienol fraction of annatto seed in 200 mL of hexane. Contaminants were washed with two liters of 10% diethyl ether in hexane. δ-Tocotrienol was then eluted with 30% diethyl ether in hexane (2 L). The eluted fraction, evaporated under vacuum at 50°C, yielded 36 g of δ-tocotrienol. The purity (98%) was determined by high pressure liquid chromatography (HPLC) as described \[[@B43]\]. Diets, experimental conditions ------------------------------ White Leghorn one-day-old female chicken and diet ingredients were supplied by the Poultry Research Laboratory, University of Wisconsin, Madison, WI, USA. The diet consists of corn (8.8% protein, 615 g), soybean meal (44% protein, 335 g), tocols-stripped corn oil (10 g), calcium carbonate (10 g), dicalcium phosphate (20 g), iodized salt (5 g), and mineral and vitamin mixtures (2.5 g of each), which provided per kg feed: zinc sulfate.H~2~O, 110 mg; manganese sulfate.5H~2~O, 70 mg; ferric citrate.H~2~O, 500 mg; copper sulfate.5H~2~O, 16 mg; sodium selenite, 0.2 mg; d,l-methionine, 2.5 mg; choline chloride (50%), 1.5 g; ethoxyquin, 125 mg; thiamine.HCl, 1.8 mg; vitamin A, 1500 units; vitamin D3, 400 units; vitamin E, 10 mg; riboflavin, 3.6 mg; calcium pantothenate, 10 mg; niacin 25 mg; pyridoxine.HCl, 3 mg; folacin, 0.55 mg; biotin, 0.15 mg; vitamin B~12~0.01 mg; and vitamin K, 0.55 mg. Corn additionally provided an estimated 2 mg/kg tocols consisting of 12% α-tocopherol, 3% α-tocotrienol, 7% β-tocotrienol, 70% γ-tocopherol, 6% γ-tocotrienol, 2% δ-tocopherol, per g \[[@B52]\]. Day-old female chickens were fed a corn-soy meal diet for one week before being randomly assigned to one of 24 groups. Three control groups (*n*= 18, housed 6/cage/group) was fed the aforementioned commercial diet for four weeks. One experimental group (*n*= 6) was fed a diet supplemented with 50 ppm δ-tocotrienol (125 μM/kg); twenty experimental groups were fed a diet supplemented with 25 or 50 ppm quercetin (74, or 148 μM/kg), 25 or 50 ppm riboflavin (66.5 or 133 μM/kg), 25 or 50 ppm (-) Corey lactone (71 or 142 μM/kg), 5 or 10 ppm amiloride (16.5 or 33 μM/kg), or 0.5 or 1 ppm dexamethasone (1.27 or 2.55 μM/kg) with or without 50 ppm δ-tocotrienol (125 μM/kg) for four weeks. Dietary concentrations of the aforementioned compounds were derived from our preliminary studies of their effects on TNF-α, and nitric oxide production by RAW 264.7 cells and by macrophages prepared from several strains of mice \[unpublished results\]. All treatment compounds except dexamethasone were dissolved in 50 mL 95% ethanol; dexamethasone was dissolved in 50 mL deionized water (100°C). The compounds were mixed with the commercial diet (5 kg) in Food Mixer for 30 min to eliminate the solvent. The experimental diets were kept at room temperature throughout the feeding period. Groups were housed in a single brooder with 24 h light and free access to water and diet. Chickens were weighed at the start and end of the trial. At the end of 4-week feeding period, the birds were fasted for 12 h prior to sacrifice to facilitate chylomicron and very-low-density lipoprotein (VLDL) clearance. The chickens were sacrificed by severing their carotid arteries, rather than gas euthanasia, in order to keep the blood (serum) composition intact for determining TNF-α and NO levels. Livers were collected, weighed, and stored at -70°C; a small portion of each liver was stored in 10% formalin and stored at -70°C pending histological analyses. The blood samples were incubated at 37°C for 20 min and centrifuged at 10,000 × g for 20 min to collect sera, which were held at -70°C, pending analyses. The protocol was reviewed and approved by the University of Wisconsin-Madison College of Agriculture and Life Sciences Animal Care and Use Committee. The study was carried out under a FDA approved IND number 36906. General biochemical methods and techniques ------------------------------------------ ### Assays of serum lipid parameters The analyses of coded samples were performed at the University of Missouri, Kansas City, School of Medicine, MO, USA. Serum cholesterol and triglyceride levels were estimated using Kits \# 352 and 336, respectively, purchased from Sigma Chemical Co. St. Louis, MO, USA. LDL-cholesterol was precipitated from 200 μL of serum with 25 μL of a mixture of 9.7 mM phosphotungstic acid and 0.4 M MgCl~2~. The preparation was mixed for 10 min at room temperature and then centrifuged at 12,000 × g for 20 min. The supernatant was decanted and analyzed for HDL-cholesterol. The remaining precipitate was dissolved in 200 μL of 0.1 M sodium citrate and the level of LDL-cholesterol estimated as described for total cholesterol \[[@B51]\]. All assays for each treatment were carried out at the same time under similar conditions to minimize standard deviation. Measurement of TNF-α level in serum of 5-week-old female chickens ----------------------------------------------------------------- Levels of TNF-α in serum of chickens were determined by Quantikine M ELISA kit (R & D System, Minneapolis, MN, USA) according to manufacturer\'s instructions. The lower limit of detection for TNF-α in this method is approximately, 5.0 pg/mL \[[@B53]\]. Measurement of nitric Oxide (NO) level in serum of 5-week-old female chickens ----------------------------------------------------------------------------- Production of NO in serum of chickens was determined by measuring the amount of nitrite, a stable metabolic product of nitric oxide as described previously \[[@B54]\]. The assay mixture consisted of medium (100 μL) and Griess reagent (100 μL) placed in round-bottom 96-well tissue culture plates (incubation time 30 min) and absorption was measured at 570 nm on a \"Microplate Reader\" (MR 5000; Dynatech Labs, Inc. USA). The amount of nitrite was determined by comparison of unknowns with a NaNO~2~standard curve. The nitrite detection limit is 0.20 nM. Histological studies of liver samples of 5-week-old female chickens ------------------------------------------------------------------- Liver tissues, fixed in 10% formalin, were embedded with paraffin and cut in the sagital plane. The sections were stained with hematoxylin and eosin and examined by light microscopy, and were evaluated by two pathologists, each blinded to the treatments. A semi-quantitative evaluation of histological analyses of these liver samples was carried out according to published methods \[[@B55]\]. Mean scores were assigned to each sample, scored range 5 (presence of impact) to 40 (very severe impact). Sample with normal appearance received a score of \"0\". The means of assigned values for each group were determined, and based on these evaluations summary of each treatment were reported. Microarray data and pathway analyses of RNA of liver samples of 5-week-old female chickens ------------------------------------------------------------------------------------------ Small frozen sections, randomly collected from each liver within a treatment group, were pooled and the RNA was isolated and purified using an affinity resin column (RNeasy, Qiagen, Chatsworth, CA, USA), as described previously \[[@B56]\] and then analyzed at Mayo Clinic (Rochester, Minnesota, USA) using an Affymetrix Gene-Chip (chicken), and Expression array analysis of chicken \[[@B57]\]. Gene expression data were first imported in Genespring program (Agilento Palo Alto, CA, USA). The expression values of up-regulated genes showed positive numbers, whereas the down-regulated genes showed negative numbers. Various genes were identified by using GeneSifter software. A number of sets of cluster analyses of various ranges of genes sets (226, 465, 500, 1000, 1500 and 2000) were mapped to get quantitative data. The analysis of amiloride was not carried out due to published reports \[[@B58]\]. Statistical analysis -------------------- Stat-View software (4.01) was used for the analyses of treatment-mediated effects (1992; Abacus Concepts, Berkeley, CA, USA). Treatment-mediated differences in serum lipid (total cholesterol, LDL-cholesterol, HDL-cholesterol, triglycerides), TNF-α and NO levels and weight (gain, liver weight, relative liver weight) were identified using a two-way analysis of variance (ANOVA). When the F test indicated a significant effect, the differences between the means were analyzed by Fisher\'s Protected Least Significance Difference (LSD) test. Data are reported as mean ± SD in the text and tables. The statistical significance level was set at 5% (*P*\< 0.05). Results ======= The impact of 2-concentrations of quercetin, riboflavin, (-) Corey lactone (25 and 50 ppm), amiloride (5 and 10 ppm) and dexamethasone (0.5 and 1 ppm) with and without δ-tocotrienol (50 ppm) on inflammatory markers and lipid parameters were studied in 5-week-old female chickens. On completion of the four-week study the data revealed that varying the concentration of these compounds failed to produce significant differences. Consequently, we present results only for the data reflecting results obtained by feeding the higher concentrations. There were 24 groups in the study, including 3 control groups. The birds of control group \# 1 were sacrificed in the beginning, \# 2 in the middle (after group \# 12) and birds of third control group (\# 3) were sacrificed after group \# 23. The values presented for the control group represent the average of three control group values (*n*= 18). Effect of various compounds on weight gain and relative liver weight gain of 5-week-old female chickens ------------------------------------------------------------------------------------------------------- Each of the compounds, with the exception of (-) Corey lactone and riboflavin, significantly lowered body weight gain (Table [1](#T1){ref-type="table"}). Treatments combining δ-tocotrienol and (-) Corey lactone or amiloride significantly increased body weight gain when compared to the effect of the individual compound and the control. The opposite response, a significant lowering of body weight gain, was detected for a treatment combining δ-tocotrienol and dexamethasone; a similar trend, though not significant, was detected for the treatment combining δ-tocotrienol and riboflavin. Treatment-mediated impacts on feed efficiency (grams feed/gram weight gain) and liver weight paralleled those noted for weight gain (Table [1](#T1){ref-type="table"}). Riboflavin, amiloride, and to a much greater extent, dexamethasone, significantly increased relative liver weights when compared to that of the control group. δ-Tocotrienol attenuated the impact of amiloride on relative liver weight as shown in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Effects of various compounds on body weight gain and relative liver weight of 5-week-old female chickens^1^. ::: Nutritional state Body weight Liver weight/100 g ---- ----------------------------------------- -------------------- -------------------- **Gain (g)** **Body weight** 1 Control Diet (CD)^2^ 240.12 ± 11.37^bc^ 2.41 ± 0.29^ef^ 2 CD + δ-Tocotrienol (50 ppm) 218.83 ± 8.16^e^ 2.42 ± 0.19^ef^ 3 CD + Quercetin (50 ppm) 225.50 ± 5.96^de^ 2.35 ±0.23^ef^ 4 CD + Riboflavin (50 ppm) 231.17 ± 7.52^cd^ 2.61 ± 0.18^cd^ 5 CD + (-) Corey lactone (50 ppm) 244.00 ± 5.02^b^ 2.60 ± 0.18^cde^ 6 CD + Amiloride (10 ppm) 230.33 ± 11.91^de^ 2.80 ± 0.25^c^ 7 CD + Dexamethasone (1.0 ppm) 57.00 ± 8.41^f^ 4.44 ± 0.21^b^ **δ-Tocotrienol (50 ppm) blend**^**3**^ 8 CD + δ-T3 + Quercetin (50 ppm) 243.50 ± 10.0^b^ 2.55 ± 0.23^cdef^ 9 CD + δ-T3 + Riboflavin (50 ppm) 228.50 ± 13.50^de^ 2.75 ± 0.20^cd^ 10 CD + δ-T3 + (-) Corey lactone (50 ppm) 266.83 ± 13.86^a^ 2.48 ± 0.27^def^ 11 CD + δ-T3 + Amiloride (10 ppm) 259.50 ± 4.18^a^ 2.27 ± 0.14^f^ 12 CD + δ-T3 + Dexamethasone (1.0 ppm) 39.33 ± 7.20^g^ 5.75 ± 0.48^a^ ^1^Feeding period was 4 weeks; Data expressed as means ± SD = 6 chickens per group. ^2^The value of control group was an average of 3 control groups. ^3^The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). ^a-g^Values in columns not sharing a common superscript letter are significantly different at *P*\< 0.05. ::: The impacts of the treatments on liver slices (histological analyses) are recorded in Table [2](#T2){ref-type="table"}. Severe chronic inflammation with mild fatty infiltration was recorded for the liver samples of control group. All individual treatments reduced the extent of inflammation (Table [2](#T2){ref-type="table"}; Figure [2](#F2){ref-type="fig"}). Combining δ-tocotrienol with quercetin, riboflavin, (-) Corey lactone, or amiloride, yielded greater reductions in inflammation and fatty infiltration than the reductions achieved with individual compounds. Combining δ-tocotrienol with dexamethasone yielded a significant increase in relative liver weight compared to that of the dexamethasone group (Table [1](#T1){ref-type="table"}), and livers were severely compromised as shown by histological analyses (Figure [2](#F2){ref-type="fig"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Effects of δ-tocotrienol and various compounds on histological analyses of liver samples of 5-week-old female chickens^1^. ::: \# Nutritional state Histological analyses of chicken liver samples ---- ----------------------------------------- --------------------------------------------------------- 1 Control Diet (CD) Severe chronic inflammation & mild fatty infiltration 2 CD + δ-Tocotrienol (50 ppm)^2^ Mild chronic inflammation 3 CD + Quercetin (50 ppm) Moderate chronic inflammation & mild fatty infiltration 4 CD + Riboflavin (50 ppm) Moderate chronic inflammation & mild fatty infiltration 5 CD + (-) Corey lactone (50 ppm) Mild chronic inflammation & moderate fatty infiltration 6 CD + Amiloride (10 ppm) Mild chronic inflammation & moderate fatty infiltration 7 CD + Dexamethasone (1.0 ppm) Mild chronic inflammation & moderate fatty infiltration **δ-Tocotrienol (50 ppm) blend**^**2**^ 8 CD + δ-T3 + Quercetin (50 ppm) Moderate chronic inflammation & mild fatty infiltration 9 CD + δ-T3 + Riboflavin (50 ppm) Very mild chronic inflammation 10 CD + δ-T3 + (-) Corey lactone (50 ppm) Very mild chronic inflammation & autolysis 11 CD + δ-T3 + Amiloride (10 ppm) Very mild chronic inflammation 12 CD + δ-T3 + Dexamethasone (1.0 ppm) Autolysis ^1^Feeding period was 4 wk; Data expressed as means ± SD = 6 chickens per group. ^2^The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Effects of dietary supplements on histological analyses of livers of 5-week-old female chickens**. The scans 1-12 shows the histological evaluation of representative liver sections from chickens treated with or without δ-tocotrienol with other compounds: 1. control diet; 2. δ-tocotrienol (δ-T3); 3. quercetin; 4. riboflavin; 5. (-) Corey lactone; 6. amiloride; 7. dexamethasone; 8. δ-T3 + quercetin; 9. δ-T3 + riboflavin; 10. δ-T3 + (-) Corey lactone; 11. δ-T3 + amiloride; 12. δ-T3 + dexamethasone. :::  ::: Effects of various compounds on serum levels of TNF-α and NO of 5-week-old female chickens ------------------------------------------------------------------------------------------ The data shown in Figures [3](#F3){ref-type="fig"}, [4](#F4){ref-type="fig"}, [5](#F5){ref-type="fig"}, [6](#F6){ref-type="fig"}, [7](#F7){ref-type="fig"}, [8](#F8){ref-type="fig"} are presented in two different formats. For each Figure., \'A\' shows the raw values for each of the treatments and control, and \'B\' shows the percent change compared to control. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Effects of dietary supplements on serum levels of TNF-α of 5-week-old female chickens**. The information regarding treatment groups were described in detail in the Methods section. Chickens were fed for 4 weeks. The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). Data expressed as means ± SD, *n*= 6 chickens per group. The value for the control group was the average of 3 control groups. Values in columns not sharing a common symbol were significantly different at *P*\< 0.05. For each figure, \'A\' shows the raw values for each of the treatments and control, and \'B\' shows the percent change compared to control. The groups 1-12 correspond to: 1. control diet; 2; δ-tocotrienol (δ-T3); 3. quercetin; 4. riboflavin; 5. (-) Corey lactone; 6. amiloride; 7. dexamethasone; 8. δ-T3 + quercetin; 9. δ-T3 + riboflavin; 10. δ-T3 + (-) Corey lactone; 11. δ-T3 + amiloride; 12. δ-T3 + dexamethasone. :::  ::: ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Effects of dietary supplements on the serum levels of nitric oxide (NO) of 5-week-old female chickens**. Chickens were fed for 4 weeks. The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). Data expressed as means ± SD, *n*= 6 chickens per group. The value for the control group was the average of 3 control groups. Values in columns not sharing a common symbol were significantly different at *P*\< 0.05. For each figure, \'A\' shows the raw values for each of the treatments and control, and \'B\' shows the percent change compared to control. The groups 1 - 12 correspond to: 1. control diet; 2; δ-tocotrienol (δ-T3); 3. quercetin; 4. riboflavin; 5. (-) Corey lactone; 6. amiloride; 7. dexamethasone; 8. δ-T3 + quercetin; 9. δ-T3 + riboflavin; 10. δ-T3 + (-) Corey lactone; 11. δ-T3 + amiloride; 12. δ-T3 + dexamethasone. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Effects of dietary supplements on the serum levels of total cholesterol of 5-week-old female chickens**. Chickens were fed for 4 weeks. The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). Data expressed as means ± SD, *n*= 6 chickens per group. The results were reported in SI units (mmo/L). The value for the control group was the average of 3 control groups. Values in columns not sharing a common symbol were significantly different at *P*\< 0.05. For each figure, \'A\' shows the raw values for each of the treatments and control, and \'B\' shows the percent change compared to control. The groups 1 - 12 correspond to: 1. control diet; 2; δ-tocotrienol (δ-T3); 3. quercetin; 4. riboflavin; 5. (-) Corey lactone; 6. amiloride; 7. dexamethasone; 8. δ-T3 + quercetin; 9. δ-T3 + riboflavin; 10. δ-T3 + (-) Corey lactone; 11. δ-T3 + amiloride; 12. δ-T3 + dexamethasone. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Effects of dietary supplements on the serum levels of LDL-cholesterol of 5-week-old female chickens**. Chickens were fed for 4 weeks. The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). Data expressed as means ± SD, *n*= 6 chickens per group. The results were reported in SI units (mmo/L). The value for the control group was the average of 3 control groups. Values in columns not sharing a common symbol were significantly different at *P*\< 0.05. For each figure, \'A\' shows the raw values for each of the treatments and control, and \'B\' shows the percent change compared to control. The groups 1 - 12 correspond to: 1. control diet; 2; δ-tocotrienol (δ-T3); 3. quercetin; 4. riboflavin; 5. (-) Corey lactone; 6. amiloride; 7. dexamethasone; 8. δ-T3 + quercetin; 9. δ-T3 + riboflavin; 10. δ-T3 + (-) Corey lactone; 11. δ-T3 + amiloride; 12. δ-T3 + dexamethasone. :::  ::: ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Effects of dietary supplements on the serum levels of HDL-cholesterol of 5-week-old female chickens**. Chickens were fed for 4 weeks. The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). Data expressed as means ± SD, *n*= 6 chickens per group. The results were reported in SI units (mmo/L). The value for the control group was the average of 3 control groups. Values in columns not sharing a common symbol were significantly different at *P*\< 0.05. For each figure, \'A\' shows the raw values for each of the treatments and control, and \'B\' shows the percent change compared to control. The groups 1 - 12 correspond to: 1. control diet; 2; δ-tocotrienol (δ-T3); 3. quercetin; 4. riboflavin; 5. (-) Corey lactone; 6. amiloride; 7. dexamethasone; 8. δ-T3 + quercetin; 9. δ-T3 + riboflavin; 10. δ-T3 + (-) Corey lactone; 11. δ-T3 + amiloride; 12. δ-T3 + dexamethasone. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Effects of dietary supplements on the serum levels of triglyceride of 5-week-old female chickens**. Chickens were fed for 4 weeks. The diets of groups 8 - 12 were supplemented with δ-tocotrienol (50 ppm). Data expressed as means ± SD, *n*= 6 chickens per group. The results were reported in SI units (mmo/L). The value for the control group was the average of 3 control groups. Values in columns not sharing a common symbol were significantly different at *P*\< 0.05. For each figure, \'A\' shows the raw values for each of the treatments and control, and \'B\' shows the percent change compared to control. The groups 1 - 12 correspond to: 1. control diet; 2; δ-tocotrienol (δ-T3); 3. quercetin; 4. riboflavin; 5. (-) Corey lactone; 6. amiloride; 7. dexamethasone; 8. δ-T3 + quercetin; 9. δ-T3 + riboflavin; 10. δ-T3 + (-) Corey lactone; 11. δ-T3 + amiloride; 12. δ-T3 + dexamethasone. :::  ::: Each of the experimental compounds significantly lowered serum TNF-α levels (Figure [3A, B](#F3){ref-type="fig"}). The decreases in serum levels of TNF-α with δ-tocotrienol (83%), quercetin (84%), riboflavin (81%), (-) Corey lactone (83%), amiloride (72%), and dexamethasone (61%), were significant (*P*\< 0.01) as compared to control. Treatments combining δ-tocotrienol with the remaining compounds produced further reductions in serum levels of TNF-α with quercetin (5%), riboflavin (11%), (-) Corey lactone (5), amiloride (16%), and dexamethasone (13%) that were considered non-significant due to large standard deviation (\>18% - 45%) in all groups (Figure [3B](#F3){ref-type="fig"}). Serum NO levels were significantly lowered by treatments comprising only of δ-tocotrienol (45%), (-) Corey lactone (31%), and dexamethasone (67%) (*P*\< 0.002) compared to control (Figure [4A, B](#F4){ref-type="fig"}). However, in comparing the results of combined supplementation vs. single compound supplementation, the addition of δ-tocotrienol to all compounds tested produced further significant (*P*\< 0.001) reductions in serum NO levels with quercetin (25%), riboflavin (34%), (-) Corey lactone (18%), amiloride (32%), and dexamethasone (11%), as compared to their respective individual reduction (Figure [4B](#F4){ref-type="fig"}). It is important to note that the reduction of serum NO with δ-tocotrienol (45%) alone closely resembles the values of combined treatments of δ-tocotrienol with quercetin, riboflavin, (-) Corey lactone, amiloride (approximately 50%), except dexamethasone, which showed maximum reduction of serum NO (67%) alone versus combined treatment with δ-tocotrienol (78%; Figure [4B](#F4){ref-type="fig"}). These pronounced reductions of TNF-α and NO by δ-tocotrienol with a dose of 50 ppm may be due to maximal attenuation achieved with this dose, indicating that it is a very potent anti-inflammatory agent. The lower dose of δ-tocotrienol (10 or 20 ppm) in combination with quercetin, riboflavin or (-) Corey lactone may produce far better additive effects than observed in the present study. In summary, the pronounced reductions of serum TNF-α and NO levels by a near saturating concentration of δ-tocotrienol, likely, masked the responses to be gained with combined treatments. Effects of various compounds on serum levels of lipid parameters of 5-week-old female chickens ---------------------------------------------------------------------------------------------- Serum total and LDL-cholesterol levels in chickens fed the δ-tocotrienol-supplemented diet were 77%, and 55% of control level, respectively (Figures [5](#F5){ref-type="fig"}, [6](#F6){ref-type="fig"}). These observations are consistent with results from numerous clinical trials \[[@B18],[@B31]-[@B34],[@B51]\]. All of the remaining compounds, with the exception of dexamethasone, significantly lowered serum total cholesterol levels, (-) Corey lactone by 31%, riboflavin by 22%, quercetin by 13% and amiloride by 9% (*P*\< 0.05) compared to control (Figures [5](#F5){ref-type="fig"}, [6](#F6){ref-type="fig"}). Dexamethasone significantly increased serum level of total cholesterol to 220% (*P*\< 0.001) of control level (Figure [5](#F5){ref-type="fig"}), an impact that was attenuated, by the addition of δ-tocotrienol. Although the combined treatments produced modest reductions (6% to 12%) in serum total cholesterol level when compared to reductions achieved with the individual compounds (Figure [5B](#F5){ref-type="fig"}), only the combination of δ-tocotrienol and (-) Corey lactone produced a greater reduction in serum total cholesterol levels than that achieved with δ-tocotrienol (Figure [5B](#F5){ref-type="fig"}). Treatment-mediated effects on serum LDL-cholesterol levels were generally comparable to the effects of total cholesterol. With the exception of dexamethasone, all of the individual treatments yielded significantly lowered serum LDL-cholesterol levels (21% - 51%; *P*\< 0.001) as compared to control (Figure [6](#F6){ref-type="fig"}). Dexamethasone significantly increased serum LDL-cholesterol level to 230% of control level (Figure [6B](#F6){ref-type="fig"}). The effect of dexamethasone was significantly attenuated by the addition of δ-tocotrienol (Figure [6B](#F6){ref-type="fig"}). Although combined treatments comprised of δ-tocotrienol and quercetin, riboflavin, (-) Corey lactone and amiloride yielded serum LDL-cholesterol levels lower than those attained with the individual agents, only the combination of δ-tocotrienol and (-) Corey lactone yielded a level significantly lower than that attained with δ-tocotrienol (Figure [6B](#F6){ref-type="fig"}). Consistent with other reports \[[@B18],[@B31]-[@B34],[@B51]\], supplementation with δ-tocotrienol had no impact on HDL-cholesterol level (Figure [7](#F7){ref-type="fig"}). Quercetin, riboflavin and (-) Corey lactone lowered serum HDL-cholesterol levels modestly. Dexamethasone significantly increased serum HDL-cholesterol level to 218% of control (Figure [7B](#F7){ref-type="fig"}). The increase observed with dexamethasone alone was also attenuated by the addition of δ-tocotrienol, comparable to the findings with serum total cholesterol and LDL-cholesterol levels. Serum HDL-cholesterol levels in chickens treated with dexamethasone plus δ-tocotrienol were 182% of control level (Figure [7B](#F7){ref-type="fig"}). In comparing the results of combined dietary supplementation with single compound, the addition of δ-tocotrienol to all compounds tested, except amiloride, did not produce further reductions in serum HDL-cholesterol levels (Figure [7B](#F7){ref-type="fig"}). Individually, neither δ-tocotrienol nor amiloride reduced serum HDL-cholesterol levels; the combination of δ-tocotrienol plus amiloride, however, modestly reduced serum HDL-cholesterol level (Figure [7B](#F7){ref-type="fig"}). With the exception of dexamethasone, all of the individual treatments yielded significantly lowered serum triglyceride levels (9% - 24%; *P*\< 0.002) of control group (Figure [8](#F8){ref-type="fig"}). Dexamethasone significantly increased the serum triglyceride level by 17% (*P*\< 0.001) of control (Figure [8B](#F8){ref-type="fig"}). This increase observed with dexamethasone was completely abrogated by the addition of δ-tocotrienol; serum triglyceride level in chickens treated with dexamethasone plus δ-tocotrienol was 103% of control level (Figure [8B](#F8){ref-type="fig"}). In comparing the results of combined treatments with single compound treatments, the addition of δ-tocotrienol to all compounds tested produced additional reductions in serum triglyceride levels (Figure [8B](#F8){ref-type="fig"}), only that achieved by combining δ-tocotrienol with amiloride was significant. To summarize the results presented above, serum total cholesterol, LDL cholesterol, and triglyceride levels were generally decreased by diets supplemented with δ-tocotrienol, quercetin, riboflavin, or (-) Corey lactone, and minimally with amiloride. Treatments combining δ-tocotrienol with quercetin, riboflavin, (-) Corey lactone, or amiloride generally produced additional but mostly non-significant reductions. Dexamethasone increased serum total cholesterol, LDL-cholesterol, HDL-cholesterol, and triglyceride levels (Figures [5B](#F5){ref-type="fig"}, [8B](#F8){ref-type="fig"}). These increases in serum total and LDL-cholesterol levels with dexamethasone were attenuated by the addition of δ-tocotrienol. Microarray analyses of RNA of liver samples of 5-week-old female chickens ------------------------------------------------------------------------- Cluster microarray data analyses of mRNA from pooled liver samples of each treatment using GeneSifter program provided valuable information comprising of 465 genes. Out of 465 genes, there were at least 62 genes whose expression was either up-regulated or down-regulated by δ-tocotrienol, quercetin, riboflavin, (-) Corey lactone, and dexamethasone. These 62 genes were categorized under inflammation, ageing, cardiovascular disease and cancer (Tables [3](#T3){ref-type="table"}, [4](#T4){ref-type="table"}). Out of 62 genes, only 39 genes were up-regulated (Table [3](#T3){ref-type="table"}) and 23 down-regulated (Table [4](#T4){ref-type="table"}) by these compounds. The expression of genes up-regulated by these compounds were associated with inflammation (9 genes), ageing (7 genes), cardiovascular disease (20 genes), and cancer (3 genes) as shown in Table [3](#T3){ref-type="table"}. These compounds modulated the expression of a number of genes, such as: interferon 1 receptor, cytokine signaling, NFκB and ubiquitin protein lipase (inflammation), heat shock protein, RIKEN cDNA, ATPase, T cell receptor gamma (ageing), FAS, myosin, squalene epoxidase, NADH dehydrogenase, Prostaglandin D2 synthase, coagulation factor II (cardiovascular), and RAN, member of RAS oncogene family (cancer) as shown in Table [3](#T3){ref-type="table"}. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Microarray analyses of RNA of livers of chicken after treatment with various compounds. ::: *Genes Up-regulated* ---- -------------------- ---------------------- ---------- ---------- ---------- ------------ ---------- ---------------------------------------------------- \# **Genes** **Control** **δ-T3** **Quer** **Ribo** **Co lac** **Dexa** **Description** **Inflammation** 1 BX932265 0.15 1.30 3.69 1.57 2.78 1.83 Endothelin receptor type 2 2 BU449947 0.81 1.92 3.54 2.05 2.81 1.73 Endothelin receptor type B 3 AJ719289 1.10 2.65 3.12 2.10 2.75 2.07 Ubiquitin protein lipase E3C 4 NM\_204124 1.12 2.40 3.12 3.09 3.11 2.03 Nuclear receptor subfamily 2, group C, member 1 5 AF082666 1.22 1.89 3.24 1.92 1.86 1.01 Interleukin 10 receptor, beta 6 CF256615 1.96 4.37 4.40 3.60 4.55 3.26 Interferon-gamma receptor alpha chain precursor 7 BU313956 2.41 2.75 4.64 2.99 3.24 3.21 Suppressor of cytokine signaling 1 8 AJ720966 2.54 4.72 4.95 4.29 4.66 4.09 Nuclear factor kappaB related to binding protein 9 NM\_205485 3.35 4.40 5.91 5.06 5.28 5.10 Interferon 1 receptor, type 1 **Ageing** 10 CR523238 0.50 1.46 2.55 0.44 2.21 0.92 Palmitoyl-protein thioesterase 1 11 U22666 0.75 4.62 4.48 4.50 3.00 2.66 T cell receptor gamma 12 BX950427 0.97 3.37 4.68 2.44 4.25 2.29 RIKEN cDNA C230081A13 13 AF175433 1.96 4.00 5.05 3.80 3.50 2.20 T cell receptor delta chain (TCRD) 14 NM\_205520 3.23 4.04 5.55 3.66 4.15 3.32 ATPase, Na+/K+ transporting, beta 1 polypeptide 15 AF387865 4.67 6.17 7.02 5.55 6.49 6.44 Heat shock protein 90 Da beta (Grp94), member 1 16 BX932093 6.04 8.07 8.34 7.24 8.47 8.01 PIT 54 protein **Cardiovascular** 17 CR524241 0.46 3.47 6.00 2.39 6.61 6.57 Succinate-CoA ligase, GDP-forming, beta subunit 18 ENSGAL T4 0.48 1.73 2.55 1.81 1.91 1.06 Glycerol kinase 5 19 CR325238 0.50 1.46 2.55 0.44 2.21 0.92 Palmitoyl-protein thioesterase 1 20 ENSGAL T26 0.55 2.92 3.79 2.96 2.92 1.90 Cytosolic methionine-S-sulfoxide reductase 21 NM\_205274 0.71 2.15 3.51 2.00 3.49 2.39 Myosin, heavy chain 11, smooth muscle 22 J04598 0.85 2.79 3.69 1.89 3.15 1.32 Collagen, type VI, alpha 1 23 ENSGAL T6 1.45 3.57 4.09 2.79 3.88 2.40 Plasminogen precursor 24 ENSGAL T5 1.53 2.99 4.09 3.17 3.78 3.29 Aldehyde dehydrogenase 9 family, member A1 25 BU235638 1.54 3.51 5.21 4.32 4.21 2.15 Serine hydroxymethyltransferase 1 (soluble) 26 BX930122 1.75 3.12 3.64 3.11 3.90 2.61 Aldehyde dehdrogenase 8 family member A1 27 ENSGAL T12 2.25 4.26 4.82 3.59 5.05 3.61 Glucose 6-phosphate translocase 28 AJ720577 2.37 4.78 6.86 4.80 6.74 5.97 NADH dehydrogenase (Ubiquinone) Fe-S protein 1 29 NM\_204259 2.48 5.88 6.02 5.75 5.70 3.67 Prostaglandin D2 synthase 21 KDa 30 AJ720030 3.29 5.00 5.62 5.35 5.13 1.62 Squalene epoxidase 31 NM\_205483 3.50 5.60 5.60 4.48 3.97 3.72 Lipoprotein 32 ENSGAL T10 3.83 5.08 5.43 5.49 5.89 3.28 Enoyl-coenzyme A, hydratase/3-hydroxyacyl coenz 33 NM\_2051155 4.73 5.52 6.87 5.48 5.79 6.09 Fatty acid synthase 34 ENSGAL T20 5.09 5.86 7.01 6.32 7.11 5.70 Alcohol dehydrogenase 1C (class 1), g-polypeptide 35 NM\_204605 5.83 7.57 7.98 7.05 7.68 6.68 Coagulation factor II (thrombin) 36 BX931521 6.36 8.27 8.94 8.12 8.43 8.63 Glutathione peroxide 3 precursor (GSHPx-3) **Cancer** 37 AJ720735 0.18 1.76 2.70 0.90 2.87 1.37 Endoplasmic reticulum protein 29 38 BU117693 1.51 2.98 3.56 2.68 2.93 2.48 RAN, member RAS oncogene family 39 ENSGAL T6 2.60 5.10 5.88 4.53 5.81 4.98 Similar to cytochrome P450, family 2, subfamily W, δ-T3 = δ-Tocotrienol; Quer. = Quercetin; Ribo = Riboflavin; (-) Co lac = (-) Corey lactone; Dexa = Dexamethasone ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Microarray analyses of RNA of livers of chickens after treatment with various compounds. ::: *Genes down-regulated* -------- -------------------- ------------------------ ---------- ---------- ---------- ------------ ---------- --------------------------------------------------------- **\#** **Genes** **Control** **δ-T3** **Quer** **Ribo** **Co lac** **Dexa** **Description** **Inflammation** 1 AJ719859 3.33 2.03 2.42 1.74 1.93 1.18 Proteasome (prosome, macropain) activator subunit 4 2 BX950555 4.01 4.11 4.38 3.78 3.38 1.64 Tumor necrosis factor superfamily, \# 5-induced protein 3 BU397996 4.53 2.82 2.18 2.74 2.20 2.87 Protein Kinase **Ageing** 4 CR353144 4.03 1.66 3.57 3.12 2.92 3.18 Nuclear DNA-binding protein 5 BU426315 4.50 2.20 2.42 2.14 1.70 1.33 Carnitine palmitoyltransferase 1A (liver) 6 CR523285 4.60 2.35 3.26 1.90 3.68 8.02 Finshed cDNA clone ChEST613j16 7 CD735693 6.05 4.25 5.44 4.66 3.68 3.39 Heat shock protein 25 8 BU123548 6.47 6.16 7.31 6.30 6.14 4.20 Finshed cDNA clone ChEST495e19 9 NM\_205471 8.73 6.99 8.76 7.69 7.42 6.28 Phosphoenolpyruvate carboxykinase 1 10 CR523582 6.79 5.31 4.78 5.19 4.66 4.84 B-cell CLL/lymphoma 9 **Cardiovascular** 11 M64990 3.94 1.69 1.51 2.10 1.02 1.25 Prostaglandin-endoperoxide synthase 2 12 CR522967 4.42 3.41 2.18 3.20 2.53 3.67 KIAA1285 protein 13 ENSGAL T24 5.31 3.93 5.00 4.42 4.60 3.05 Aproteindipose differentiation-related 14 BX935098 5.79 4.45 5.41 4.35 4.52 3.09 Glutathione S-transferase theta 1 15 BU422942 5.84 4.14 3.65 4.09 3.79 4.41 Glycogen synthase kinase 3 beta 16 BU272340 6.58 5.63 4.23 5.47 5.39 5.81 Inositol hexaphosphate kinase 2 **Cancer** 17 AJ447153 2.19 0.83 2.06 1.56 1.21 0.01 Protein tyrosine phosphate, non-receptor type 2 18 BU463093 3.63 2.67 4.07 2.78 3.28 0.90 Amino acid transporter system A1 19 AL585963 3.95 2.15 1.74 1.20 2.07 1.67 RAS guanyl releasing protein 3 20 ENSGAL T0 4.24 3.18 1.62 2.42 1.85 2.97 Breast cancer-associated antigen BRCAA 1 21 BU131710 4.47 4.36 5.24 5.28 4.35 2.31 Isopentenyl-diphosphate delta isomerase 1 22 BU111042 6.54 5.46 4.39 4.81 4.98 5.95 Chrosome 6 open reading frame 111; SR rich protein 23 BU458470 7.84 5.78 7.08 6.96 5.90 5.90 Jun oncogene δ-T3 = δ-Tocotrienol; Quer. = Quercetin; Ribo = Riboflavin; Co lac = (-) Corey lactone; Dexa = Dexamethasone ::: The expression of genes down-regulated by these compounds were associated with inflammation (9 genes), ageing (7 genes), cardiovascular disease (6 genes), and cancer (7 genes) (Table [4](#T4){ref-type="table"}). Some of the important genes whose expression was modulated by these compounds included those of proteasome, protein kinase, tumor necrosis factor (inflammation), carnitine palmitoyltransferase 1A, nuclear DNA-binding protein (ageing), glycogen synthase kinase, glutathione S-transferase (cardiovascular), RAS guanyl releasing protein 3, and Jun oncogene (cancer) (Table [4](#T4){ref-type="table"}). The Jun oncogene also plays important role in ageing. The detailed analyses of genes whose expressions modulated by δ-tocotrienol, quercetin, riboflavin, (-) Corey lactone and dexamethasone, regulated differentially, were also selected (Tables [5](#T5){ref-type="table"}, [6](#T6){ref-type="table"}). Among these compounds, the first four were associated with lowering of serum lipids and latter, dexamethasone, associated as a lipid-elevating compound. Expression of 20 genes was up-regulated and 7 genes down-regulated by δ-tocotrienol, quercetin, riboflavin, (-) Corey lactone, except for a lipid elevating, dexamethasone (Table [5](#T5){ref-type="table"}). Expression of 8 genes was up-regulated by first four compounds and down-regulated by dexamethasone, and 2 genes were down-regulated by first four compounds and up-regulated by dexamethasone (Table [6](#T6){ref-type="table"}). Moreover, expression of 2 genes was up-regulated by dexamethasone, and not by the four lipid-lowering compounds and 1 gene was down-regulated by dexamethasone, and no effect was observed with the four lipid-lowering (δ-tocotrienol, quercetin, riboflavin, (-) Corey lactone) compounds (Table [6](#T6){ref-type="table"}). ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Microarray analyses of RNA of livers of chickens after treatment with various compounds. ::: *Genes up-regulated by the first four lipid-lowering compounds and lipid-raising dexamethasone* -------- ------------ ------------------------------------------------------------------------------------------------- ---------- ---------- ---------- ------------ ---------- -------------------------------------------------------- **\#** **Genes** **Control** **δ-T3** **Quer** **Ribo** **Co lac** **Dexa** **Description** 1 BX275358 3.42 6.22 6.09 6.66 5.39 3.31 Putative ISG 12-2 protein 2 BU235638 1.54 3.51 5.21 4.32 4.21 2.15 Similar to serine hydroxymethyltransferase 1 (soluble) 3 ENSGL T6 3.98 6.05 6.93 5.53 7.15 4.59 UDP glucuronosyltransferase 1 family, polypeptide A10 4 CF256116 1.85 3.58 4.64 3.76 4.57 1.91 Phosphoribosyl pyrophosphate amidotransferase 5 NM\_204858 -0.48 0.95 2.05 1.49 1.70 -0.13 Interferon (alpha, beta, and omega) receptor 2 6 AY534896 6.19 8.22 8.56 8.00 7.91 6.40 Gal 10 7 JO4598 0.85 2.79 3.69 1.89 3.15 1.32 Collagen, type IV, alpha 1 8 ENSGALT2 6.32 8.29 8.95 7.64 8.48 6.78 Similar to inter-alpha (globulin) inhibitor H3 9 ENSGALT27 0.84 3.10 3.18 2.23 2.50 0.87 Similar to KDEL (Lys-Asp-Glu-Leu) containing 1 10 M60069 6.14 8.06 6.71 7.57 7.88 6.00 Phosphoribosyl pyrophosphate amidotransferase 11 ENSGALT15 4.35 5.38 7.00 5.96 6.71 4.52 Similar to MGC 107895 protein 12 BX275222 0.23 2.43 2.64 1.24 1.59 0.19 Hypothetical protein LOC69044 13 ENSGALT27 4.80 6.48 7.13 5.83 6.84 5.06 Similar to thrombin-activatable fibrinosis inhibitor 14 BU219227 1.60 2.82 4.09 3.16 3.21 1.92 WD repeat domain 61 15 NM\_205299 2.29 3.38 4.62 3.39 3.66 2.33 Dystrophin 16 ENSGALT26 2.10 3.46 4.27 3.23 3.67 2.66 Similar to methylmalonyl coenzyme A mutase 17 NM\_205355 4.83 6.24 7.15 5.98 6.12 5.21 Ring finger protein 13 18 ENSGALT3 3.62 4.91 5.89 4.33 6.40 3.78 Similar to complement regulator factor H 19 CR290617 3.27 4.77 5.74 4.17 4.96 3.41 RER1 retention in endoplasmic reticulum 1 homolog 20 ENSGALT4 0.99 2.51 3.36 1.79 2.62 1.12 Similar to U1 snRNP-specific protein C 21 AF082666 1.22 1.89 3.24 1.92 1.86 1.01 Interleukin 10 receptor, beta ***Genes down-regulated by the first four lipid-lowering compounds except dexamethasone*** **\#** **Genes** **Control** **δ-T3** **Quer** **Ribo** **Co lac** **Dexa** **Description** 1 ENSGALT20 2.01 1.14 -0.34 1.29 0.73 2.35 Similar to Trans-Golgi p230 2 BU270035 2.54 1.06 1.03 -0.08 1.25 2.59 Similar to Expressed soquence A1314180 3 BU208119 3.84 3.14 1.33 2.13 1.93 3.51 Similar to RIKEN cDNA D130059P03 gene 4 BU229724 5.44 3.99 2.96 3.49 3.57 5.55 Finshed cDNA, clone Chest295h22 5 ENSGALT9 1.84 -0.43 -0.23 0.75 -0.57 2.50 Similar to KIAA07 protein 6 BU305188 6.08 4.75 3.02 4.19 3.99 5.71 PCF11, cleavage and polyadenylation factor subunit 7 BU426927 3.76 2.44 -0.14 1.83 0.97 3.63 Triple functional domain (PTPRF interacting) δ-T3 = δ-Tocotrienol; Quer. = Quercetin; Ribo = Riboflavin; Co lac = (-) Corey lactone; Dexa = Dexamethasone ::: ::: {#T6 .table-wrap} Table 6 ::: {.caption} ###### Microarray analyses of RNA of livers of chickens after treatment with various compounds. ::: *Genes up-regulated by the first four lipid-lowering compounds,* -------- ------------ ------------------------------------------------------------------------ ---------- ---------- ---------- ------------ ---------- --------------------------------------------------- ***and down-regulated by dexamethasone*** **\#** **Genes** **Control** **δ-T3** **Quer** **Ribo** **Co lac** **Dexa** **Description** 1 CR353609 4.47 7.23 7.13 6.67 7.12 3.95 Finished cDNA, clone CHEST110e20 2 AJ720605 3.07 5.39 5.65 5.23 5.04 2.66 Ornithine aminotransferase 3 BX265212 2.81 4.72 5.12 4.54 5.14 2.38 Similar to L-Kynurenine hydrrlase 4 CR405837 2.76 4.89 5.27 4.49 4.95 1.89 Finished cDNA, clone CHEST884a21 5 BU250153 3.57 5.25 5.97 5.13 5.77 3.27 Selenoprotein P, plasma, 1 6 AJ720030 3.29 5.00 5.62 5.35 5.13 1.62 Squalene epoxidase 7 ENSGALT10 2.97 4.76 5.19 4.00 4.83 3.86 Assembly = WASHUC1ǀchr = 20ǀstrand = forwardǀcdna 8 L07842 2.48 3.90 4.80 3.18 5.70 1.48 Antithrombin III **Genes** ***Genes down-regulated by the first four lipid-lowering compounds,*** ***and up-regulated by dexamethasone*** 1 CR389189 4.60 2.35 3.26 1.90 3.68 8.02 Finshed cDNA, clone CHEST613j 16 2 NM204114 3.82 0.66 2.35 1.33 2.22 5.97 Deiodenase, iodothyronine, type II 3 NM\_205155 4.73 5.52 6.87 5.48 5.79 9.09 Fatty acid synthase **Genes** ***Genes up-regulated only by dexamethasone*** 1 BU359098 0.84 1.08 1.04 1.25 0.88 3.69 BUB 1 uninhibited by benzimidazoles 1 homolog **Genes** ***Genes down-regulated only by dexamethasone*** 1 BU384885 2.11 3.48 4.56 2.90 3.43 1.23 Hypothetical LOC771662 δ-T3 = δ-Tocotrienol; Quer. = Quercetin; Ribo = Riboflavin; Co lac = (-) Corey lactone; Dexa = Dexamethasone ::: Discussion ========== On average, the weight of chickens fed the control diet increased by 240 g during the four-week trial. In the current study, we found chickens fed a diet supplemented with δ-tocotrienol gained significantly less weight. This finding differs from results of earlier trials demonstrating that diet supplementation with δ-tocotrienol produced either no change, or an increase in weight gain \[[@B43],[@B51]\]. Quercetin, amiloride, and dexamethasone yielded significantly lower weight gain, whereas riboflavin and (-) Corey lactone had no significant effect on weight gain. Interestingly, combined supplementation consisting of δ-tocotrienol plus either (-) Corey lactone or amiloride significantly increased weight gain; the combination of δ-tocotrienol and quercetin produced a weight gain equal to that of the control. Thus for (-) Corey lactone, amiloride, and quercetin, additional supplementation with δ-tocotrienol appeared to increase weight gains, as compared to dietary supplementation with each of these compounds alone. The addition of δ-tocotrienol to riboflavin did not improve weight gains compared to dietary supplementation with riboflavin alone. As reported elsewhere, dietary supplementation with dexamethasone markedly reduced weight gain compared to control \[[@B59],[@B60]\], and this detrimental effect was enhanced by combining dexamethasone with δ-tocotrienol. Histological examination demonstrated decreased inflammation in livers from chickens receiving individual treatments. Treatments combining δ-tocotrienol with either riboflavin, (-) Corey lactone, or amiloride yielded further decreases in hepatic inflammation and fatty infiltration. On the other hand δ-tocotrienol potentiated the toxic impact of dexamethasone. Dexamethasone toxicity, previously demonstrated in rats, was manifested by impaired growth, enlarged livers, and elevated serum total cholesterol and triglyceride levels \[[@B59]-[@B61]\]. All predictors of cardiovascular risk evaluated in this study were substantially decreased by all compounds with the exception of dexamethasone. Summarizing the overall effects of the individual compound on the five serum factors under consideration (TNF-α, NO, total cholesterol, LDL-cholesterol, and triglyceride) leads to the conclusion that the cumulative risk of atherosclerosis is reduced effectively by δ-tocotrienol, quercetin, riboflavin, and (-) Corey lactone. Serum TNF-α levels of chickens receiving each of the compounds were uniformly lower than those recorded for chickens fed the control diet. Serum levels of TNF-α of chickens treated with δ-tocotrienol, quercetin, riboflavin, and (-) Corey lactone were reduced by approximately 80%. Serum TNF-α levels of chickens treated with dexamethasone and amiloride were reduced by approximately 41% and 70%, respectively. These findings are consistent with prior reports of the effects riboflavin \[[@B62]-[@B64]\], quercetin \[[@B5],[@B9],[@B10],[@B65]\], δ-tocotrienol \[[@B15]-[@B17]\], amiloride \[[@B66]\] and dexamethasone \[[@B67]\]*in vitro*\[[@B9],[@B10],[@B15]-[@B17],[@B21]\] and *in vivo*\[[@B62]-[@B64]\] on TNF-α level. δ-Tocotrienol produced a 45% reduction in the serum NO level. Quercetin, riboflavin, (-) Corey lactone, amiloride, and dexamethasone reduced serum NO levels by 14%, 14%, 31%, 25%, and 67%, respectively. Quercetin has been reported to down-regulate inducible-NO synthase (iNOS) activity *in vitro*\[[@B5]-[@B10],[@B68]\], and riboflavin, delivered by injection or infusion, inhibits NO synthesis and the concomitant increases in serum NO level in LPS-challenged mice \[[@B62]-[@B64]\]. Perhaps the most important result of the present study is the finding that combining δ-tocotrienol with other dietary supplements enhances suppression of serum NO and TNF-α levels, as compared to single compound supplementation; these results are particularly striking for NO level. Although, reduction of serum NO level with δ-tocotrienol alone closely resembles the reductions of δ-tocotrienol combined with other compounds, which may be due to maximal attenuations achieved with a dose of δ-tocotrienol (50 ppm) used in the present study. In our earlier dose-response study of δ-tocotrienol effects on the serum levels of total cholesterol and LDL-cholesterol in chickens, the maximum effective dose was found to be 200 ppm (51). Therefore, a minimum effective dose of 50 ppm was selected for the present study. This is the first report that describes the effects of δ-tocotrienol for avian pro-inflammatory markers (TNF-α and NO), which is consistent with findings that δ-tocotrienol is very potent anti-inflammatory compound as reported recently (47), and it is possible a lower dosage (10 or 20 ppm) of δ-tocotrienol may potentiate the anti-inflammatory actions of quercetin, riboflavin and (-) Corey lactone. All treatments, except those involving dexamethasone, resulted in significantly lower serum total cholesterol, LDL-cholesterol and triglyceride levels. The effects of δ-tocotrienol and (-) Corey lactone on serum levels of total and LDL-cholesterol were significantly greater than those of the other compounds (Figures [5](#F5){ref-type="fig"}, [6](#F6){ref-type="fig"}, [8](#F8){ref-type="fig"}). Our current study appears to be the first observation of the cholesterol-lowering impact of (-) Corey lactone. δ-Tocotrienol is widely reported to effectively suppress HMG-CoA reductase activity and concomitantly lower serum total cholesterol and LDL-cholesterol levels \[[@B30]-[@B34],[@B40]-[@B43],[@B69],[@B70]\]. Quercetin suppresses HMG-CoA reductase activity *in vitro*\[[@B71]\] and *in vivo*\[[@B72]\], and dietary intake of quercetin has been inversely correlated with total cholesterol and LDL-cholesterol levels in Japanese women \[[@B73]\]. Cholesterol levels in an elderly population were inversely correlated with serum riboflavin levels \[[@B74]\]. Thus, our findings that quercetin, and more potently, riboflavin lowered serum cholesterol levels are supported by the literature \[[@B72]-[@B74]\]. We also found that diet supplementation with amiloride, an FDA approved diuretic, was least effective in reducing total cholesterol level (9%). This finding is somewhat similar to results of a human study which showed that serum total cholesterol levels were not altered in subjects receiving amiloride concomitantly with hydrochlorothiazide \[[@B75]\]. Amiloride also failed to impact serum levels of cardioprotective HDL-cholesterol in the current study. HDL-cholesterol levels in chickens receiving quercetin, riboflavin or (-) Corey lactone were modestly reduced. Combining δ-tocotrienol with these compounds failed to raise HDL-cholesterol. On the other hand, the anti-inflammatory compound, dexamethasone, dramatically increased serum HDL-cholesterol level. All compounds other than dexamethasone resulted in a significant lowering of serum triglyceride levels (Figure [8](#F8){ref-type="fig"}). The additive effect of combining δ-tocotrienol with another compound, with the exception of amiloride, on serum triglyceride was insignificant (Figure [8](#F8){ref-type="fig"}). A recent report points to the superiority of the HDL-cholesterol/total cholesterol (HDL-chol/TC) ratio for monitoring cardiovascular risk compared to serum total cholesterol and LDL-cholesterol levels \[[@B76]\]. With the exceptions of those incorporating dexamethasone, all treatments resulted in higher HDL-chol/TC ratios than that calculated for the control group (0.51). Within the individual treatment group δ-tocotrienol (0.65; 127%), riboflavin (0.61, 120%), (-) Corey lactone (0.68, 133%) appear to be most effective in improving the ratios. The ratio calculated for δ-tocotrienol + quercetin (0.65, 127%) was modestly improved relative to that calculated for the quercetin group (0.56, 110%). These modest improvements were also observed by δ-tocotrienol combination with riboflavin (129%), and (-) Corey lactone (145%), without improving amiloride or dexamethasone ratios. In addition to the potential ability of quercetin, to reduce the cumulative serum risk factors for cardiovascular disease (total cholesterol, LDL-cholesterol, triglyceride, NO and TNF-α levels) we found exceptional risk-reducing value in two vitamins, δ-tocotrienol (a member of the vitamin E group; Figure [1](#F1){ref-type="fig"}), and riboflavin when fed at levels 4- and 10-times higher, respectively, than those normally found in commercial chicken feed. Based on the data presented in this study, it is reasonable to propose that supplementing a 2500 kcal diet with 40 mg of either vitamin (i.e. δ-tocotrienol or riboflavin) could be potentially beneficial in reducing cardiovascular disease risk in humans. Microarray analyses of liver samples identified 62 genes whose expression was up-regulated (39 genes) or down-regulated (23 genes) by all compounds suggesting common impact on serum levels of TNF-α, NO, and lipid parameters. The most important up-regulated gene expression modulated by these compounds were associated with cytokine signaling, NFκB and ubiquitin protein lipase (inflammation), heat shock protein, RIKEN cDNA, T cell receptor gamma (ageing), FAS, myosin, squalene epoxidase, NADH dehydrogenase, Prostaglandin D (cardiovascular disease), and RAN, member RAS oncogene family (cancer). The down-regulated genes were associated with proteasome, tumor necrosis factor (inflammation), carnitine palmitoyltransferase1A (ageing), glycogen synthase kinase, glutathione S-transferase (cardiovascular disease), and Jun oncogene (cancer) as reported earlier \[[@B76]-[@B80]\]. The microarray array analyses further identified several other genes whose expression was differentially impacted by the compounds shown to lower serum lipid levels and dexamethasone, associated with markedly elevated serum lipids. Conclusions =========== Levels of serum markers for risk of inflammatory diseases (NO and TNF-α) are decreased by oral dietary treatments supplemented with naturally-occurring, synthetic or FDA approved compounds, δ-tocotrienol, quercetin, riboflavin, dexamethasone, (-) Corey lactone, and a diuretic, amiloride. When administered in combination with an apparently saturating dose of δ-tocotrienol, the risk-lowering impact of the remaining compounds was only modestly increased. Therefore, this finding suggests the possibility of a pronounced additive effect in the presence of a lower dose of the δ-tocotrienol. Serum NO levels increase during ageing process, as a consequence of a diminished regulation of the activation of NF-κB signaling \[[@B1],[@B2]\]. These compounds may also block the activation of NF-κB and result in lowering serum TNF-α and NO levels. Confirming numerous reports, δ-tocotrienol, a post-transcriptional suppressor of HMG-CoA reductase activity, effectively lowered serum total cholesterol, LDL-cholesterol and triglyceride levels. Moreover, treatments incorporating the anti-inflammatory compounds, quercetin, riboflavin, and (-) Corey lactone alone and in combination with δ-tocotrienol resulted in lower serum total and LDL-cholesterol levels. However, anti-inflammatory dexamethasone increased serum lipid levels, actions partially attenuated in the presence of δ-tocotrienol. These novel findings demonstrate the potential value to be gained through investigations of the impact of various nutritional supplements, specifically the flavonoids, alone and in combination with δ-tocotrienol on predictors of age-associated diseases. Abbreviations ============= AP-1: activator protein-1; COX-2: cyclooxygenase-2; HMG-CoA: β-hydroxy-β-methylglutaryl coenzyme A; ICAM: intracellular adhesion molecule-1; IκB: inhibitory kappaB; IL-1α: interleukin-1α; IL-6: interleukin-6; IL-8: interleukin-8; iNOS: inducible nitric oxide synthase; LPS: lipopolysaccharide; MCP-1: macrophage chemoattractant protein-1; MIP-1α: macrophage inflammatory protein-1α; NF-κB: nuclear factor-kappaB; NO: nitric oxide; TNF-α: tumor necrosis factor-α; ROS: reactive oxygen species; VCAM: vascular cell adhesion molecule-1. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= All the authors were involved in the designing of the study. DMS completed the paper work of \"University of Wisconsin-Madison Animal Care and Use Protocol Review Form\". He also supervised the feeding of the chickens at Poultry research Laboratory, University of Wisconsin, Madison, WI. CJP edited the manuscript. JCR has checked the statistical analyses of all the data. All the authors have read and approved the final version. Acknowledgements ================ We thank Ms. Dawn Irish (Poultry Research Laboratory) for preparation of various diets, daily care (feeding, water, weighing), and assistance with terminal procedures, and Ms. Terry Jobsis (Supervisor of campus animal operations of Department of Animal Sciences, Madison, WI, USA) for supervising the feeding regimen, sacrifice, blood and liver collections. We also thank Drs. Agostino Molteni (Professor of Pathology, UMKC, USA), Victor Flauta, Daniel Lingamfelter (Residents), and Mr. Tim L Quinn (Senior Laboratory Technician, and Central Laboratory Manager, UMKC, USA) for the histological study of liver samples, and Dr. Sreekumar Raghavakaimal (present address: National Foundation for Cancer Research, 4600 East West Highway, Suite 525, Bethesda, MD 20814, USA), and Mr. Christopher P. Kolbert, MS, RM (AAM), Supervisor-Advanced Genomic Technology Center, Microarray shared Resources (Mayo Clinic, Rochester, Minnesota, USA) for microarray analyses of serum samples. Our thanks go to Ms. Eleanor G Zuvanich (UMKC, USA) for the estimation of serum levels TNF-α and nitric oxide. We thank Mr. Keith Gilchrist (USDA, ARS, MWA, Cereals and Crops Research Laboratory, Madison, WI, 33726, USA, for carrying out statistical analyses of all the data. We also thank Dr. CE Elson (Emeritus Professor, Department of Nutritional Sciences, University of Wisconsin, Madison, WI, 53706, USA) for helpful discussion and checking the statistical analyses. This study was supported in part by Advanced Medical Research (AMR) and NIH grants GM-50870, AI-54962, AI-57168 (NQ).

PubMed Central

2024-06-05T04:04:19.524650

2011-2-28

PMC3053242

Background ========== To understand the need for adoption of standard Public Health Informatics (PHI) design methods requires a brief background of PHI, the environment in which public health information systems are used and the unique challenges encountered in designing information systems for use in public health practice. The unique field of Public Health Informatics: population-focused and diverse ----------------------------------------------------------------------------- Public Health Informatics is the \"systematic application of information and computer science and technology to public health practice, research, and learning that integrates public health and information technology\"\[[@B1]\]. And, as stated by Yasnoff et al.\[[@B2]\], PHI needs to use a \"systematic and informed approach to the application of information science and technology in order to take full advantage of its potential to enhance and facilitate public health activities.\" Such activities include: • Promotion of the health of populations as \"opposed to the health of specific individuals.\" • Prevention of disease and injury \"by altering the conditions or the environment that put populations of individuals at risk.\" • Prevention at all vulnerable points in the \"causal chains leading to disease, injury, or disability \... not restricted to particular social, behavioral, or environmental contexts.\" • Reflection of the \"governmental context in which public health is practiced \"\[[@B2]\]. In the application of information science and technology to population health, the challenges of PHI are unique. For example, unlike clinical informatics, PHI design activities cannot simply focus on an individual\'s data in an Electronic Health Record (EHR), the transmission of a specific patient visit in a data stream between a health care provider organization and public health agency or an intervention targeted to an individual patient. Instead, PHI design activities must take a more systemic approach that allows access to population-level data with functionality tailored to a variety of uses by different groups of practitioners. Instead of individual patient data, public health information systems need to include and manage entire datasets of patient encounters. Supporting prevention-centered activities moves the focus further up the causal chain of health and is less likely to manifest in an individual patient visit, but might instead be seen as a population-level prevention intervention, e.g., a city-wide anti-tobacco campaign. These prevention activities take place in diverse environments since the interventions may be social, behavioral, environmental or individual interventions. In addition, public health informaticians must always consider the governmental context in which these population-level prevention activities take place. Understanding the health services structure and how that structure affects the work done by public health practitioners is an important step in creating systems that will function within that context. Diversity in public health practice presents challenges for public health informaticians. A perceived lack of standard work practices across local health jurisdictions on the part of practitioners and administrators complicates public health information system design activities. If each state or local jurisdiction has unique information needs and business processes, then each must make individual design decisions to ensure that their needs are met. Recent efforts have been undertaken to create standards for business processes in the nation\'s local health jurisdictions. These efforts have uncovered unexpected commonalities\[[@B3]\] as well as developed a taxonomy of common public health work tasks, knowledge and resources\[[@B4]\]. Yet it has been argued that there is a great deal of variability in local health jurisdictions\[[@B5]\] and further research is necessary to document the needs and tasks of public health practitioners within specific local health jurisdictions to design technology to support these needs and tasks\[[@B6]\]. Adding to this diversity of practice and settings is the diversity of national and international settings in which these activities take place. The globalization of public health necessitates a globalization of PHI to address the challenges of disparate data, tools and services within and across nations\--both resource-rich and -constrained. Issues of conflicting data standards, the need for interoperable tools for exchanging and sharing data and the need for innovative solutions to address integrated disease surveillance, among many other issues, are driving forces to formalize design strategies in PHI. The development of public health information systems requires an understanding of the principles, practices, structures and settings in which these systems operate. Although this combination of factors presents unique challenges, PHI is at a point in its evolution as a discipline that public health informaticians can learn from the design theories and experiences of other fields to address these challenges. Evolution of public health information systems in the US and Europe: three waves -------------------------------------------------------------------------------- According to Lumpkin, federal-state public health information systems in the United States have been developed in three \"waves\"\[[@B7]\]. The first wave began with the independent development of state and federal systems which could not exchange data; the second wave, supported by federal funding to develop state level systems capable of exchanging data using standards in building independent\--and therefore costly\--state systems\[[@B7]\]; and the third wave which focused on reducing costs through the integration of \"the benefits of state-level system development with the tools of software reuse\" and a requirement that \"each system that is developed must be standards based\"\[[@B7]\]. Similarly, in Europe, public health information systems developed in three comparable waves\[[@B8]\]: a first generation of systems focused on collecting basic data on a population and/or the health of a population (for example, birth and death data); a second generation of systems that combined stratifications of time-series data and international comparative data as seen in comparisons of age-standardized breast cancer incidence rates by sex over time and country; and the third generation in the present, with the public health information system that is made possible by information technology advances and is referred to as an \'\'integrated knowledge system\'\' in its integration of data, descriptive and analytical information, and evidence-based knowledge\[[@B8]\]. In both the European and US contexts, this current third wave of public health information systems utilizes electronic data exchanges and standards to solve the information needs of public health workers at different organizational levels, across organizations and with various health care and government stakeholders\--thus addressing the challenges of population focus and diversity discussed above. While a common approach to this third wave presumes that system designers both have sufficient knowledge of the work of public health practitioners and can systematically integrate this knowledge into system design, past technology failures suggest that this is not the case and numerous studies have underscored the fact that a system will not be used by health professionals in everyday practice unless the system is usable. While a literature review on challenges to adopting or deploying health information systems is beyond the scope of this paper, these challenges are pertinent to the topic of design and PHI. The reader is referred to the work of Kushniruk et al.\[[@B9]\] and Peute et al.\[[@B10]\] for additional information. Complex problems and design challenges in public health ------------------------------------------------------- Ignoring the user experience has led to a literature replete with numerous examples of health information systems being \"turned off\" or rejected by health professionals because these systems were developed without an understanding of the information needs, workflow, or architecture needs of system users\[[@B11],[@B12]\]. As previously noted, public health processes and practices can be difficult to define, developing clear descriptions of this work requires concerted effort\[[@B13]\], and a public health information system that meets the needs of one group may simultaneously create more work for other groups of public health practitioners\[[@B14]\]. The focus of public health agencies on the development and implementation of information systems is a relatively new phenomenon. Public health informaticians face significant challenges in designing and implementing flexible, interoperable, usable systems to meet the needs of public health practice. Recommendations for a national agenda for PHI in the United States were not formally outlined until 2001\[[@B15]\] and competencies for public health informaticians that include \"expertise in both public health programs and information systems to help design, implement, and manage computer applications that support public health goals\"\[[@B16]\] were only finalized in 2009. The outcome is that many public health information systems often came into existence through an ad hoc and informal design and development process\[[@B17],[@B18]\] yet the public health domain represents a complex design setting. Complex design problems of the type encountered in the public health domain are not new. Following World War II, Weaver defined problems of disorganized complexity and organized complexity\[[@B19]\]. Weinberg discusses problems of disorganized complexity as large, random populations that are subject to statistical treatment while individual problem cases are members of small, structured populations that are subject to individual analysis\[[@B20]\]. Information design in PHI addresses a problem area of organized complexity that is subject to neither individual nor statistical analysis. Rittel and Webber frame their design discussion by enumerating a list of problems and claim that defining what systems do and planning what they should do in terms of desired outcomes can be difficult, if not impossible in large, societal systems\[[@B21]\]. Buchanan holds that \"the problem for designers is to conceive and plan what does not yet exist and this occurs in the context of the indeterminancy of wicked problems\" and this \"indeterminancy implies there are no definitive conditions of limits to design problems\"\[[@B22]\]. Nevertheless, public health practice hinges on information that is managed by information systems and those systems *must*be designed. Our proposed approach for PHI is not to specify design solutions in advance but to specify a flexible process that involves public health practitioners to design solutions for public health systems of organized complexity. This approach is supported by the writings of Cross who summarizes forty years of design and notes that the \"wicked problems\" characterized by Rittel are more appropriately satisfied by \"an \'argumentative\', participatory process in which designers are partners with the problem \'owners\'\" rather than by a rigid, step-wise process\[[@B23]\]. Problem representation in the design of complex systems requires reflection. Simon explicated the formidable task of representing new problems that do not fit with previously known patterns\[[@B24]\] while Lawson reflects on the need for a process that allows problems and solutions to emerge simultaneously and reflect each other\[[@B25]\]. Numerous design thinkers cite the contribution of Donald Schön\'s model of the \"reflective practitioner\" and a model of inquiry that relies on interactive problem framing \[[@B23],[@B25]-[@B27]\]. In formalizing his approach to scenario-based design, John M. Carroll has built on Schön\'s metaphor of design as a conversation\[[@B28]\]. We propose that PHI can benefit from design theories and methods developed in other disciplines to improve public health information systems by addressing the unique work of public health practitioners. In the next section we describe six general design problems as defined by Carroll\[[@B29]\] and provide examples of how those problems express themselves in the unique context of the public health domain. We then present a strategy for creating reusable design knowledge using established participatory and scenario-based design methodologies. Finally, we address some potential challenges to the approach we propose. Discussion ========== The overarching challenges associated with system design within the public health context can be described, in part, as traditional design problems. As mentioned previously, Carroll uses a \"design as conversation\" metaphor for interactive problem-framing and -solving. Carroll and his collaborators, long-time proponents of design methods in Human-Computer Interaction, formalized this approach in scenario-based design and described six general problems in design work\[[@B29]\]. While the philosophy and practice of modern design have a well-documented history\[[@B22]-[@B24],[@B26],[@B27],[@B30]-[@B32]\] that dates back to John Dewey in the early part of the last century\[[@B33],[@B34]\], we believe that Carroll\'s approach is a simple way to highlight the main categories of design challenges for public health information systems. To demonstrate how traditional design problems play out in the public health context, in Table [1](#T1){ref-type="table"} we have mapped Carroll\'s six design problems to public health information examples. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Design problems and solutions in the public health context ::: Design Problem Example Design Problem in the Public Health Context Reusable Design Solution ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------ --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- **Incomplete Problem Description** *Problems in system design rarely make their conditions clear early in the design process* Few formal needs assessment studies\[[@B35]\] Needs assessments conducted in coordination with participatory design to clarify problems by learning from their users, users\' goals and users\' data needs\[[@B36]\] **Unclear Design Pathway** *While there are many possible steps that a designer can take while clarifying a problem description, the best path is not obvious* Need to understand the work practices of public health practitioners\[[@B36]\] Participatory design\[[@B37]\] to create design specifications so a Public Health Informatics solution is incorporated into an optimized workflow and environment **Impact of Design Solutions are Hard to Predict** *Although the general type of problem may be understood, the solution to the problem can exceed the problem itself* Public health systems are designed and deployed without complete knowledge of the working environment system\[[@B38]\] Create a representation of an intentional future\[[@B12]\] by using Human-Computer Interaction, Interaction Design, User-Centered Design and Contextual Design to understand the public health context. **Trade-Offs Due to Competition Among Resources** *Project components compete for resources, some elements of the project constrain the design of other elements, and these competing interests challenge resolving conflicts between the design elements* Limited resources mean not all system requirements can be met Documented design knowledge that allows for comparisons between competing solutions Balancing competing characteristics such as sensitivity and false negatives\[[@B39],[@B40]\] **Integrated and Interdependent Needs** *Collaboration is needed between designers and users to pool knowledge* Diverse work and diverse information systems\[[@B14]\] Participatory design that includes a mixture of representative public health work roles and teams Create scenarios that include a mixture of representative public health work roles and teams **Unintended Consequences** *Unintended consequences can have a significant impact on users and those outside the originally conceived group of stakeholders* Data produced by public health systems is used for different purposes by a diverse set of individuals Use scenario-based design to describe the tasks and activities necessary to deliver essential services Health practitioners have complex roles and workflows\[[@B41],[@B42]\] ::: Patel and Kaufman frame informatics as a \"local science of design\"\[[@B43]\], and make the case for the incorporation of design methods from other disciplines to solve practical problems in informatics and set the stage for development of specific guidelines for design within informatics specialties. Although they caution against generalizing widely in the abstract, we believe, as they do, that principles of usability and design can be incorporated within specific domains once a domain has been defined. The strategy we propose is in concert with the development of PHI as a maturing domain in which to incorporate design methods from other disciplines. Specifically, we propose participatory design and scenario-based design as key components in a reusable design strategy. In the following section, we propose a unified strategy for public health information system design based on methods that have been developed in other disciplines. Participatory design in health care settings -------------------------------------------- Participatory design aims to actively engage all stakeholders in the design process to ensure that the system meets their needs and expectations, and ultimately is adopted by the target practitioners. This approach represents a departure from traditional approaches to information system design in that it promotes rapid prototyping and iterative approaches to implementation rather than segregated phases of design and technical production. The concept of participatory design originated in the Scandinavian countries where research projects on user participation and involvement date back to the early 1970s \[[@B44]\]. These early projects focused primarily on empowering workers in actively participating in the design and implementation of systems and workflow processes to improve working conditions. Some of the early work also focused specifically on health care. The Florence project\[[@B45]\] followed a participatory design approach to empower nurses to play an active role in the development of work processes and information technology applications in hospitals. The participatory design approach began a slow adoption process in the US in the late 80s and early 90s, however it has not been extensively employed in health care settings\[[@B46]\]. Case studies and small scale efforts demonstrate its potential; Sainfort et al.\[[@B47]\] point out how user involvement can ensure four specific qualities for information system interfaces that are appropriate for clinical providers and patients, namely multimodal, personalized, context aware and adaptive. Involving practitioners in the early design stages can maximize these qualities as a system prototype is developed and ultimately implemented. As Pilemalm and Timpka point out, participatory design methods in the disciplines of health informatics have been mainly applied to the development of small-scale systems with homogeneous user groups in local settings\[[@B48]\]. They present a participatory design framework for large-scale system design. The proposed framework was designed and validated within a PHI project aimed at developing a system for 175,000 users. Scenario-based design --------------------- Scenario-based design is a methodology that places the focus of system design on the activities of the people who use an information system rather than the system itself or the capabilities of technology\[[@B28],[@B49],[@B50]\]. It is a participatory method in that it solicits the needs and values of work practitioners by bringing them directly into the design process. Scenarios are stories that use everyday language to describe people and their work activities\[[@B28]\]. These narratives can be used to communicate with laypersons who may lack technology or software design training\[[@B51]\]. The use of everyday language to describe narratives of use makes scenario-based design a practical methodology when applied in a variety of contexts, including that of public health. Reusable design knowledge in the public health context ------------------------------------------------------ Reusable design knowledge is design advice that is \"generalized so it can be reused in a wide range of contexts\"\[[@B52]\]. Reusable design knowledge can come in the form of data standards for information storage and exchange\[[@B40],[@B53]-[@B57]\], software design patterns used by programmers when they develop a system\[[@B58]-[@B63]\] or installable software that is implemented in multiple settings\[[@B64]\]. However, this discussion pertains to design methods and the designed activities and workflows that result from those methods. Whittaker et al. propose the notion of shared tasks called \"reference tasks\" for use by designers in different application domains and state \"\[t\]he goal of reference tasks is to capture and share knowledge and focus attention on common problems\"\[[@B65]\]. Scenario-based design is one means to accomplish the goal of creating reference tasks in the public health domain as Carroll asserts that \"\[s\]cenarios can also be abstracted and categorized, helping designers to recognize, capture and reuse generalizations\"\[[@B28]\]. In addition, Wahid has proposed the reuse of positive and negative claims about particular design moves used to weigh trade-offs in scenario-based design\[[@B66]\] while Sutcliffe has done considerable work to advance the effective reuse of scenarios\[[@B52],[@B67],[@B68]\]. Within the context of public health we propose using scenario-based design to describe the tasks and activities necessary to deliver public health activities and to \"jointly identify reference tasks, collect data, analyze the tasks, and disseminate and make use of the results\"\[[@B65]\]. We believe that these processes and the resulting products will facilitate the reuse of design knowledge and resource savings in the public health domain. While there is growing work in identifying and documenting information needs for public health practitioners\[[@B69]\], we propose that the next steps in formalizing this documentation and transforming it into reusable design knowledge should include the development of scenarios of use to inform the design of information systems. Collaboration and practitioner involvement are two important aspects of developing useful scenarios. Our experiences have shown that public health practitioners are enthusiastic participants in the scenario development process\[[@B14],[@B70]\]. Scenarios can be developed in a participatory fashion using data collected from interviews, focus groups, surveys and the review of artifacts such as paper documents and electronic files. The dissemination of reusable design knowledge is important if the goal of reducing work across public health jurisdictions is to be realized. Therefore a plan for taking advantage of existing public health groups and communities of practice should be considered as part of a formal dissemination strategy. By making scenarios available through existing information dissemination channels, public health informaticians and system designers could pick and choose scenarios tailored to the work practices of particular roles and the size of a local health jurisdiction. Scenario-based design has the potential to offer great improvement in the information systems developed for public health. Aside from encouraging the development of more useful systems, scenario-based design also offers savings in time and resources. Reusing design knowledge can make systems less expensive to develop and offers a potential reduction in failed and resource intensive one-off development projects. Potential challenges to the approach ------------------------------------ Regardless of size, available resources, or geographic location, each public health organization addresses the most pressing population health concerns of its community\--which may explain the variety of public health information systems\[[@B69]\]. As previously discussed, public health organizations also maintain unique work practices that fulfill the requirements of their policy makers, and honor the organizational culture of the individual group. Before widespread Internet access provided a mechanism for public health agencies to exchange and report information electronically, efforts to standardize information systems across local health jurisdictions were limited. Now, as the argument for electronic information exchange becomes more compelling, there is a need to reevaluate the design of information systems in public health. The approach to information system design within different communities of practice has, so far, reflected the disparate organization of local and state public health agencies. With few means to acquire information systems that have been validated for interoperability, and facing an obligation to optimally address public health concerns despite this gap, public health organizations have found information management solutions through independent development, or through the purchase of commercially available systems. In either case, information management systems are designed to interact with a very limited scope of external entities, thereby reducing the value of important public health data. Without a concerted effort to integrated formal design principles into public health, this trend will continue. While this paper has focused largely on Public Health Informatics efforts in the United States, it addresses concerns of system design that are relevant to international public health efforts. Public Health Informatics efforts must be informed by the governmental context in which they operate and the environment that obtains in the United States is but one example in the set of all international examples. Working directly with each community of practice, and building a knowledgebase of activities in each, is an essential component of the reusable design approach we propose, regardless of public health environment. For example, in developing countries where resources and capacity are limited, there are several initiatives using participatory design methods to understand the information needs of users as well as the scalability and sustainability of national information systems for providing access to information and supporting the collection, handling and dissemination of health data\[[@B71],[@B72]\]. Summary ======= While the technical capacity to develop robust information systems has existed for many years, we have yet to identify the consummate model of public health information processes and develop systems to support those processes. Public Health Informatics can play a role in helping to resolve public health policy challenges, and in the process, define the necessary systems to promote population health, improve PHI as a discipline, and advance all the disciplines of biomedical and health informatics. The major public health challenges in the coming decade will be solved with the help of policy changes. Information systems are technology artifacts that implement the processes defined by policy. We believe that the policies described in the \"National Agenda for Public Health Informatics\"\[[@B15]\] should be supported with even more specific policies related to design methods. Additionally, we assert that a policy failing to describe well-tested processes will result in an information system that fails to make optimal use of the information that public health practitioners carefully collect and analyze. Policy should promote the use (and reuse) of design for the development of information systems grounded in public health practice. Policy changes in the Unites States are already transforming the health IT landscape. The information designs that result in response to these changes will contribute to the global health informatics conversation. Financial incentives for meaningful use of EHR technology reinforce the importance of health information exchange between providers and public health organizations. As policy changes place a greater emphasis on the design and use of effective health IT, the need for systematic design approaches should increase. In the case of compliance with meaningful use criteria, methods for design reuse can facilitate work that has already begun for the exchange of health information within and between provider and public health systems\[[@B73]\]. The resulting design knowledge should include specification of how, and with whom, systems interoperate. As we have suggested throughout this paper, identifying the information needs of public health practitioners is a foundational step toward developing a design strategy for PHI. There are many examples of information system failure since the advent of computing systems in the workplace when target audiences are excluded from the design process. Scenario-based design is a proven methodology for information system design but has not been applied widely within the public health domain. When coupled with methods of design-based inquiry and practitioner participation, it can be a powerful tool to move quickly and efficiently from needs and task documentation to prototypes to implemented systems. Working directly with each community of practice and building a knowledgebase of activities in each is essential. By including practitioners in a reusable design strategy for public health, we aim to increase the likelihood of future information system adoption in public health practice and, by extension, increase the likelihood of quality public health service delivery. We suggest, as part of the PHI reusable design effort, that information needs studies and technology use cases should be regularly documented and indexed for ready access by informaticians and others. Future work toward a reusable design strategy will include a survey of public health practitioners about their willingness and the ways they would like to participate in design work and a systematic review of design methods used within the specialized corpus of public health technology knowledge. We believe that the identification and consistent use of rigorous methods for system design will help move the discipline of Public Health Informatics forward. Toward this end, we believe that the design methods outlined in this paper fit into the applied, empirical and theoretical areas that should be incorporated into PHI practice and curricula for professional training at the undergraduate and graduate levels. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= BR conceived the idea for the manuscript, participated in coordination of research and writing of all manuscript revisions. RH participated in coordination of research and writing of all manuscript revisions. GD contributed material on participatory design and participated in writing of all manuscript revisions. DR contributed material on socio-technical and international issues and participated in writing of final manuscript revisions. JP participated in coordination of research and writing of manuscript revisions. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/116/prepub> Acknowledgements ================ This project was funded in part by the National Library of Medicine Medical Informatics Training Grant T15 LM007442-07. The authors would like to thank Jacob O. Wobbrock for information about disciplines related to Human-Computer Interaction, the two reviewers whose suggestions and comments made this manuscript much stronger and the generous support of the Division of Biomedical and Health Informatics in the Department of Medical Education and Biomedical Informatics at the University of Washington.

PubMed Central

2024-06-05T04:04:19.534209

2011-2-18

PMC3053243

Background ========== Binge drinking is an increasing problem among young adolescents in the Netherlands. The recent use of alcohol among pupils in secondary education (12 to 16 years of age) in the Netherlands is declining, while binge drinking among these pupils is increasing. Nowadays, 75% of the Dutch 12 to 16 year olds who drink alcohol also engage in binge drinking \[[@B1]\]; this implies consuming five or more alcoholic drinks on one occasion in the past month. The largest proportion of binge drinkers are found in the age category of 15 and 16 years old. In comparison to other European countries, Dutch adolescents are among the leaders in drinking frequency and binge drinking \[[@B2],[@B3]\]. In adolescents, heavy alcohol consumption is associated with premature and violent deaths, e.g. traffic accidents, having risky sexual intercourse \[[@B4],[@B5]\] and poor academic performance, learning difficulties and school dropout \[[@B6]-[@B8]\]. In addition, heavy alcohol use during puberty appears to be related to damage to the development of cognitive and emotional abilities \[[@B9],[@B10]\] and an elevated risk of later dependence and misuse \[[@B11],[@B12]\]. Alcohol-related risks to cognitive functions seem to be higher in adolescents than in adults \[[@B11]\]. From the point of view of public health, prevention of heavy alcohol use among adolescents is essential. There is little scientific evidence that universal prevention programmes aimed at youngsters affect drinking behaviour. Recent meta-analyses show that such programmes have small or no effects on alcohol use and binge drinking \[[@B3],[@B13],[@B14]\]. Exceptions to this are interventions aimed at both adolescents and their parents \[[@B15]\] and integrated programmes with multiple years of intervention and professional support \[[@B13],[@B16],[@B17]\]. Meta-analyses of school-based substance use prevention programmes have concluded that selective prevention programmes, targeting populations at increased risk, generally yield higher effects than universal programmes (e.g. \[[@B13],[@B18]\]). According to Cuijpers and colleagues \[[@B13]\], selective prevention programmes have proved effective, but the availability of these programmes is limited. Therefore there is a recognized need in the field of substance use prevention for selective prevention programmes. Preventure ---------- Preventure is a selective prevention programme and is one of the few school-based programmes with long-term effects on adolescents\' drinking behaviour and binge drinking \[[@B16],[@B19],[@B20]\]. In research conducted in Canadian and English samples of adolescents, effects of the programme were found on abstinence, quantity and frequency of drinking, binge drinking, and problem drinking symptoms at four months and one year after the programme \[[@B16],[@B19]\]. In addition to the effects on alcohol use, positive effects were found on emotional and behavioural problems, i.e. depression, panic attacks, truancy, and shoplifting \[[@B21]\]. The Preventure programme specifically targets young adolescents who have two well-known risk factors for heavy alcohol consumption: early-onset alcohol use \[[@B22],[@B23]\] and personality risk for alcohol abuse (e.g. \[[@B24]\]). The programme is based on the theory that personality is an important construct for understanding adolescents\' alcohol use and abuse. Two personality dimensions were previously found to be predictive of heavy alcohol use and alcohol use disorders, namely (1) an impulsive sensation seeking dimension, and (2) a behavioural inhibition dimension \[[@B16]\]. The first category involves young sensation seekers and young people with low impulse control, the second reflects a neurotic personality involving more anxious and negative thinking young people. Within these two dimensions, Conrod and colleagues \[[@B16]\] distinguished four personality profiles at higher risk of developing alcohol problems: Sensation Seeking (SS), Impulsivity (IMP), Anxiety Sensitivity (AS) and Negative Thinking (NT). The four personality profiles were subsequently found to be strongly related to adolescents\' quantity and frequency of drinking, frequency of binge drinking, and severity of alcohol problems \[[@B25],[@B26]\]. Each personality profile is associated with specific substance misuse patterns, maladaptive motives for use, and vulnerability to specific forms of co-morbid psychopathology in adolescents \[[@B27],[@B28]\]. Impulsivity is related to an increased risk of the early onset of alcohol and drug problems \[[@B29]\]. Sensation seekers drink more \[[@B30]\], tend to drink in order to enhance euphoric (intoxicating) effects \[[@B28]\], and are more at risk of adverse drinking outcomes (e.g. \[[@B30]\]). Highly anxiety sensitive persons show increased levels of drinking \[[@B31]\], are more responsive to the anxiety-reducing effect of alcohol, and are more likely to use alcohol to cope with negative feelings \[[@B28]\]. Persons with high levels of hopelessness often have depression-specific motives for alcohol use \[[@B32]\] and usually drink to cope with negative feelings \[[@B16],[@B28],[@B33],[@B34]\]. The Preventure programme screens a school population for pupils who already drink alcohol and, additionally, belong to one of the four high-risk personality profiles. The programme identifies and treats high-risk adolescents, with the aim of preventing or intervening early before the high-risk adolescents engage in risky behaviours and/or these behaviours become problematic. The selected pupils are offered a tailored intervention based on cognitive behaviour therapy (CBT) and motivational interviewing. Cognitive behavioural techniques are used to target maladaptive thinking and coping skill deficits, and motivational interviewing techniques are used to address motivation to take responsibility for one\'s problematic behaviours. Motivational interviewing has proven to be effective for alcohol- and drug-related behaviour, and CBT can lead to reduction in anxiety sensitivity, depressive cognitions, and impulsivity (e.g. \[[@B35],[@B36]\]). The manualized intervention, developed by Conrod and colleagues \[[@B35]\], provides personalized feedback and personality-specific cognitive-behavioural exercises designed to facilitate more adaptive coping. The focus is not on drinking (or drug use) per se but on risky ways of coping with personality, such as avoidance, distraction, and aggressive thinking, that may lead to substance misuse or other risky behaviour. Aims and hypotheses ------------------- In 2009, a project was started to develop and test Preventure in the Netherlands, where currently there is no selective school-based alcohol prevention available \[[@B37]\]. The main objective of this project is to study the effectiveness of Preventure on drinking behaviour of young adolescents in secondary education in the Netherlands. The effectiveness of the Dutch Preventure is being assessed by conducting a clustered randomized controlled trial (RCT), with two conditions (treatment and control arms). This is the first time that Preventure has been studied outside the setting where it was developed, England and Canada, to prove its effectiveness outside this setting. The most relevant outcomes are percentage reductions in binge drinking (≥ five drinks on one occasion in the past four weeks), weekly drinking, and drinking-related problems after 2, 6, and 12 months. The main hypothesis is that high-risk students who receive the personality targeted intervention will score lower on these outcomes relative to those in the no-treatment control group. In addition, our secondary aim is to test the effects of the programme on emotional and behavioural problems (e.g. aggression, truancy, and shoplifting). Our hypothesis is that Preventure facilitates lower depression rates, lower anxiety rates, lower delinquent behaviour rates, less problem behaviour, and lower truancy. Methods/Design ============== Study design and time frame --------------------------- The Preventure study is a 1-year RCT with two arms, an intervention and a control condition, testing the prevention programme effects, at 2, 6, and 12 months after the intervention (see Figure [1](#F1){ref-type="fig"}). Randomization is carried out at school level. The intervention condition consists of two group sessions based on cognitive behaviour therapy and motivational interviewing. The control condition receives no further intervention (business as usual). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Study design**. :::  ::: The recruitment, inclusion, and randomization of the participants (schools and students) started in Spring 2009. The data collection started in 2010. The final follow-up measurement is planned for the end of 2011. Participants ------------ ### Recruitment A total of 100 schools were selected randomly from a list of all public secondary schools (N = 405) in four regions in the Netherlands (Zuid-Holland, Utrecht, Gelderland, Overijssel). Schools were invited to participate in the study, if the following inclusion criteria were met: 1. school had at least 600 students, 2. \< 25% of students were from migrant populations, 3. school did not offer special education. A total of 15 schools were willing to participate and fulfilled the inclusion criteria. The main reasons for schools not participating were lack of time and no interest in participating in research in general. ### Eligibility Students were eligible to enter the trial if they fulfilled the following inclusion criteria: 1. life time prevalence of alcohol use (i.e. having drunk at least one glass of alcohol once in their life), 2. belonging to one of the four personality high-risk groups for (future) heavy drinking (AS, SS, NT or IMP) and 3. informed consent of the student and his or her parents. The study is aimed at students from 13 to 15 years of age. This is in contrast to Conrod et al.\'s study \[[@B16]\], in which students aged 14 to 17 were studied. The reason for this difference is the age of onset, which is lower among Dutch youngsters than among their study sample. In order to select those students fulfilling the selection criteria, a screening survey among all students attending grade 8 and grade 9 of the 15 schools was carried out. The students who scored more than one standard deviation above the sample mean on one of the four personality risk scales (AS, SS, NT, or IMP) of the Substance Use Risk Profile Scale (SURPS) \[[@B25]\], were classified as belonging to a risk group for the development of alcohol problems. If a student scored high on more than one subscale, he or she was assigned to the personality group in which he or she showed the largest statistical deviation with respect to the z-scores. ### Consent Parents were informed of the study (screening and intervention) through a letter sent home from the schools asking them to contact the researchers by phone or e-mail if they did not wish their child to participate in the study (passive informed consent). Parents were told that the intervention was coping-skill training designed to reduce adolescent risk taking, with alcohol abuse as an example. To assure participants\' confidentiality, parents were not explicitly informed about any of the selection variables of the study. On the day of the screening, students were given information on the screening, the ethical issues (confidentiality and the voluntary nature of participation), and the intervention. Parents and students provided active informed consent to participate in the intervention part of the study. The study was evaluated by the Medical Ethical Commission for Mental Health (METIGG), which considered the study did not fall within the WMO Act (Medical Research Involving Human Subject Act). As a result no ethical approval was necessary. However, for the consent procedure, we adhered to the guidelines and advices of the METIGG. ### Randomization Randomization occurred at the school level to avoid contamination between conditions. An independent statistician assigned the participating schools randomly to one of the two conditions: intervention or control. Randomization was carried out using a randomization scheme, stratified by level of education and school size, with the schools as units of randomization. Sample size ----------- ### Power The power calculation reflects the idea that we want to induce a reduction in the percentage of students engaging in binge drinking (drinking five or more glasses of alcohol on one occasion) at least once during the last four weeks, from the current estimated 50% (among life-time users grade 9/10; estimate based on the results of a national school survey, \[[@B1]\]) to 35%. For a 15% reduction after 12 months among the students in grade 9/10, a sample size of N = 183 in each condition was required to test the hypothesis in a 2-sided test at alpha = 0.05 and a power of (1-beta) = 0.80. Because of the loss of power due to randomization of schools (and not students) and the increase in error because of applying a multiple imputation procedure to fill in missing values, 183\*1.4 = 256 respondents per condition (intervention and control) needed to be included at baseline to test the effectiveness of the Dutch Preventure programme. ### Number of students According to the power analyses, a net sample of 256 respondents in each condition was needed. On the assumption of a 40% participation rate, 45% of respondents belonging to one of the risk groups (estimates based on \[[@B16],[@B19]\]), a life time prevalence at baseline of 77%, and 93% of children present in the class at the data collection time (estimates based on \[[@B1]\]), a survey sample of N = 3,972 students was needed. Study intervention ------------------ To develop the Dutch Preventure programme, the principles and guidelines of the original Canada/UK programme were followed in collaboration with the original developers of Preventure. ### Theoretical basis Preventure incorporates the principles from motivational and cognitive behavioural therapy and is adapted to different personality profiles for substance abuse: anxiety sensitivity, negative thinking, sensation seeking, and impulsivity. The intervention is brief, as the literature strongly suggests that brief interventions can be very effective in changing drinking patterns and related problems. An effective component of successful brief interventions for alcohol abuse is the persuasiveness of individualized feedback. Therefore, Preventure provides pupils with personalized feedback on their results from a personality and motivational assessment. Preventure also includes cognitive behavioural skills training specifically relevant to each personality profile. The literature has shown that successful cognitive behavioural therapy can lead to reductions in anxiety sensitivity in anxiety patients, depressive cognitions in depressed patients, and impulsivity in adolescents with externalizing disorders \[[@B38],[@B39],[@B36]\]. The intervention consists of three main components: (1) psycho-education, (2) behavioural coping skills, and (3) cognitive coping skills \[[@B16]\]. In the coping skills sections, students are engaged in activities to induce automatic thoughts. Simultaneously, they are trained to use cognitive restructuring techniques to counter such thoughts. Cognitive restructuring training has been shown to have a positive impact on the reduction of alcohol and drug abuse and symptoms of psychological disorders \[[@B35]\]. ### Intervention condition The intervention involved two group sessions, carried out at the participants\' schools. The group sessions were adapted to one of the four personality profiles. This means that there were four different groups of two sessions each. Both group sessions lasted 90 minutes and were spread across two weeks. The intervention was provided by a qualified counsellor and a co-facilitator. The three counsellors and two co-facilitators had received two days training from Dr. P.J. Conrod, who developed the original intervention. Furthermore, all the counsellors had practiced the two group sessions at a pilot school with students who met the inclusion criteria (drinkers with high-risk personality profiles). The intervention used student manuals. The original student manuals, developed in Canada, were translated and adapted to the cultural and school context of the Netherlands. The examples, the real-life stories, and the illustrations used in the programme manuals were adapted to the Dutch situation. The student manuals consist of text, exercises, and real-life experiences or scenarios. The real-life scenarios were generated by previously organized focus groups of high-risk personality adolescents. In four focus groups (one group for each personality risk factor), students were asked to share their own experiences regarding, for example, alcohol and drugs. The student manuals had been tested during the pilot sessions at the pilot school. Students were asked to give their opinion on the content, the illustrations, and real-life stories used in the manuals. In the first group session, psycho-educational strategies were used to educate students about the target personality variable (NT, AS, IMP, or SS) and the associated problematic coping behaviours, such as interpersonal dependence, aggression, risky behaviours, and substance misuse. Students were motivated to explore their personality and ways of coping with their personality through a goal-setting exercise. Thereafter, they were introduced to the cognitive behavioural model by analysing a personal experience according to the physical, cognitive, and behavioural responses. In the second session, participants were encouraged to identify and challenge personality-specific cognitive thoughts that lead to problematic behaviours. For example, the impulsivity intervention focused on not thinking things through and aggressive thinking, and the sensation-seeking intervention focused on challenging cognitive thoughts associated with reward seeking and boredom susceptibility. ### Control condition Students assigned to the control group received no further intervention. An inventory among the participating schools will reveal whether other specific substance use prevention programmes were being used, apart from the common lessons in the curriculum, e.g. biology classes. Data collection and instruments ------------------------------- The screening survey collected data by means of an online questionnaire on alcohol use, demographics, and personality risk factors. The data collection took place during a regular lesson (approximately 50 minutes), and questionnaires were administered by a research assistant from the Trimbos Institute. Those students randomly assigned to the experimental or control condition have completed, or will complete, the post-treatment survey after 2, 6, and 12 months. Data for the follow-up measurements have been, or will be, also collected online at school. The follow-up survey contains the same assessments as the screening survey. An overview of all measurements is given in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Overview of measurements ::: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Measurement Baseline (screening) Follow-up I (2 months after baseline) Follow-up II (6 months after baseline) Follow-up III (12 months after baseline) ----------------------------- ---------------------- --------------------------------------- ---------------------------------------- ------------------------------------------ Demographic characteristics \* \* \* \* Truancy\ \* \* \* \* Alcohol:  Drinking behaviour \* \* \* \*  Drinking motives \* \* \* \*  Drinking problems \* \* \* \*  Perceived parental rules \* \* \* \*  Drinking parents \* \* \* \* Tobacco: \* \* \* \*  Smoking behavior \* \* \* \*  Smoking parents \* \* \* \*  Perceived parental rules \* \* \* \* Marijuana: \* \* \* \*  Marijuana-using behaviour \* \* \* \*  Marijuana parents \* \* \* \* Other: \* \* \* \*  Personality \* \* \* \*  Anxiety \* \* \* \*  Psychological problems \* \* \* \*  Delinquency \* \* \* \*  Depression \* \* \* \*  Self control \* \* -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ::: As already mentioned, the SURPS \[[@B25]\] distinguishes four personality profiles. Each profile is assessed using five to seven items that could be answered on a 4-point scale, 1 = *strongly disagree*, 2 = *disagree*, 3 = *agree*, 4 = *strongly agree*. The SURPS scale has 23 non-overlapping items that assist in discriminating personality dimensions independent of substance use behaviour. Negative Thinking (7 items) refers to hopelessness, which might lead to depressive symptoms. A sample item on the Negative Thinking subscale is \'I feel that I\'m a failure.\' The Anxiety Sensitivity dimension (5 items) measures fear of bodily sensations, and an example item is \'It frightens me when I feel my heart beat change.\' The Sensation Seeking subscale (6 items) measures the tendency to seek out thrilling experiences, e.g. \'I would like to learn how to drive a motorcycle.\' The tendency to act without thinking is measured by the Impulsivity subscale (5 items), and an example of this subscale is \'I often don\'t think things through before I speak.\' Studies in both adolescent and adult samples in several countries, including the Netherlands, have shown that this scale has good internal reliability, good convergent and discriminant validity, and adequate test-retest reliability \[[@B34],[@B40],[@B41],[@B25],[@B27]\]. The instrument was translated into Dutch by an English speaking language consultant, has been successfully applied \[[@B34]\], and was tested at schools before use in the screening survey. Outcomes -------- When the data analysis takes place, the primary outcomes will be percentage reductions in binge drinking, weekly and weekend drinking, and drinking-related problems. To assess life-time alcohol use and binge drinking, two questions will be used that are widely used in school surveys, including the ESPAD study \[[@B2]\], Monitoring the Future \[[@B42]\], and the national school surveys in the Netherlands \[[@B1],[@B43]\]. The average standard units in the last week will be assessed with the Weekly Recall \[[@B44],[@B45]\]. Weekly and weekend alcohol use is defined by the quantity-frequency measure \[[@B46],[@B47]\]. To assess behavioural symptoms of adolescent problem drinking, the Rutgers Alcohol Problems Index (RAPI) \[[@B48]\] will be used. The RAPI has been well validated for use with both clinical and community adolescent samples \[[@B49]-[@B51],[@B48]\]. Other outcomes will include percentage reductions in depressive feelings, anxiety symptoms, problem behaviour, drinking motives, truancy, and delinquent behaviour. Depressive feelings will be measured with the widely used 20-item (Dutch version) of the Centre for Epidemiological Studies Depression Scale (CES-D) \[[@B52],[@B53]\]. The Childhood Anxiety Sensitivity Index (CASI) \[[@B54]\] is a self-report questionnaire to assess children\'s and adolescents\' fear of anxiety symptoms. The CASI has good internal consistency and acceptable 2-week test-retest reliability \[[@B55]\]. The Strengths and Difficulties Questionnaire (SDQ) \[[@B56]\] will be used as a behavioural screening instrument for early detection of psychological problems. The DMQ-R \[[@B57]\] is the most widely used instrument to assess drinking motives among young people. The DMQ-R has been well validated in several international (e.g. \[[@B58]\]) and national studies (e.g. \[[@B51]\]). Statistical Analyses -------------------- Descriptive analyses will be conducted to examine whether randomization resulted in a balanced distribution of important demographic characteristics and the outcome variables in the two conditions. To control for potential bias, possible confounders will be included in all further analyses. Analyses will be performed according to the intention-to-treat and completers-only principles, controlling for sex, age, and educational level. Intention-to-treat means that all participants will be analysed in the condition to which they were assigned by randomization. Therefore, missing data at follow up will be imputed using regression imputation. With respect to the completers-only analyses, only the participants with scores on all time points will be included, without the inclusion of imputed data. In both the intention-to-treat and the completers-only analyses, the effects of the intervention condition will be compared with those of the control condition. For continuous outcome measures, t-tests, or Man Whitney U if non-parametric distributions, will be performed. When correction for confounding variables is necessary, multivariate regression analyses will be performed. The fact that the data are clustered, because groups of respondents that are attending the same class and/or school are investigated, will be taken into account in the analyses. Discussion ========== The present study protocol presents the design of a randomized controlled trial evaluating the effectiveness of a prevention programme called Preventure. The intervention programme aims to prevent adolescents from (problematic) alcohol drinking. It is hypothesized that, after one year of follow-up, students in the intervention condition will be engaging less in binge drinking and weekly drinking, and will have fewer drinking-related problems than those students in the control condition. Strengths and limitations ------------------------- A first strength of Preventure is that it is one of the few school-based programmes with proven effects on drinking behaviour of adolescents \[[@B16],[@B20]\]. Second, the programme is a selective prevention programme. In the field of substance use prevention in The Netherlands, there is a recognized need for selective prevention programmes \[[@B13]\]. Third, the intervention incorporates elements of motivational and cognitive behavioural theory, which have been proven to be effective in reducing alcohol abuse and associated psychological problems. Fourth, Preventure is a short intervention (two sessions), which makes it less time-consuming than regular prevention programmes and therefore easier to implement in schools. A limitation of the study is that the information on the behaviour of the adolescents is based on self-reports, which might lead to measurements errors. However, studies have shown that self-report data of adolescents about their own drinking, smoking, and drug use are generally reliable (e.g. \[[@B59],[@B60]\]). A general issue with targeted interventions is the selection of participants and providing information to the participants and their parents in an accurate manner. In this study, neither the parents nor the teachers at the school were explicitly informed about the selection variables of the study, to avoid stigmatization of the students. This ethical issue should also be taken into account if the programme is implemented at other schools in the Netherlands in the future. Implications for practice ------------------------- If the Preventure prevention programme is effective, it can be implemented widely in schools in The Netherlands - for example, as part of the Dutch national school prevention programme The Healthy School and Drugs. The Healthy School and Drugs has a large network among institutions for care and treatment of drug addicts and schools. Conclusion ========== This study has described a programme, currently on trial in the Netherlands, for preventing and reducing binge drinking in adolescence. Evaluation of the programme will provide insight into the effectiveness of Preventure in the Netherlands and the precursors of alcohol use among Dutch adolescents. Competing interests =================== Patricia Conrod is one of the developers of the Preventure programme and is the principal investigator of the Preventure trials in the UK. The other authors declare that they have no competing interests. Authors\' contributions ======================= JL, FG, and SL are responsible for data collection, data analysis, and reporting the study results. PC is the principal investigator of the Preventure trials in the UK and Canada. MK, RE, and RW are supervisors. KM and LL have contributed to the grant. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/126/prepub> Acknowledgements ================ This study is funded by a grant from ZonMw, The Netherlands Organization for Health Research and Development (project no. 50-50105-96-505).

PubMed Central

2024-06-05T04:04:19.536955

2011-2-21

PMC3053244

Background ========== It has been estimated that worldwide, approximately 285.000.000 persons have diabetes mellitus (DM) and this number is expected to rise to 439.000.000 by 2030 \[[@B1]\]. This sharp increase in prevalence is partly due to the ageing of the population and the increase in the number of people having overweight and who are physical inactive \[[@B2]\]. DM is a chronic disease characterized by high levels of blood glucose resulting from a deficit in the secretion of the hormone insulin (absolute insulin deficiency: Type 1 DM), or insufficient insulin action (insulin resistance) and/or a failure of the beta-cells to produce enough insulin, i.e. beta-cell dysfunction (Type 2 DM)\[[@B3]\]. Type 2 DM is the most frequent form, prevalent in approximately 90% of the cases \[[@B4]\]. For all patients with diabetes, appropriate glycemic control by means of adequate self-management is required in order to prevent or delay the development of diabetes complications: macrovascular complications (e.g. coronary artery disease) and microvascular complications (i.e. nephropathy, retinopathy, and neuropathy). This self-management consists for example of a healthy diet, physical exercise, frequent assessment of the blood glucose levels, and (in many cases) the use of medication to control the blood glucose (an oral hypoglycemic agent and/or insulin)\[[@B3]\]. A considerable proportion of the patients with diabetes (20-40%) experience emotional problems, which vary from disease-specific worries (such as fear of hypo\'s or worries about complications) to more general symptoms of distress, anxiety and depression \[[@B5]-[@B7]\]. These problems are not only unpleasant for the persons experiencing them, but studies show also that co-morbid emotional distress in patients with DM is associated with reduced quality of life \[[@B8]\], poor self-care behaviors \[[@B9]\], more negative appraisals of insulin therapy \[[@B10]\], reduced glycemic control \[[@B9]\] and subsequent adverse cardiovascular outcomes, and even mortality \[[@B8],[@B11]\]. Despite the fact that emotional distress is so common in patients with diabetes, the attention for emotional problems in both practice and research is still limited. In practice, the recognition rates of emotional problems by clinicians are low and so are treatment rates \[[@B12]\]. Similarly, in the field of diabetes research, studies that test psychological interventions to improve the emotional well-being in patients are scarce. In addition, those studies that have been conducted often have important limitations, such as lack of controlled designs and small sample sizes \[[@B13]\]. However, it is true that the attention for emotional problems in DM patients is increasing since the last decade. In practice, screening for emotional problems is recommended in the latest versions of the standards of care of the International Diabetes Federation \[[@B14]\] and the American Diabetes Association \[[@B15]\]. In addition, in research, there are more and more studies examining the effectiveness of psychological interventions on clinical depression. Anxiety symptoms and diabetes-specific emotional distress have received, however, much less attention in diabetes research. The results of well-designed studies that have been conducted to test psychological interventions in diabetes patients with emotional problems show that cognitive behavioral therapy is an effective treatment for major depression in diabetes patients \[[@B16]\] and a potentially effective treatment in reducing symptoms of anxiety, perceived stress, and diabetes-specific emotional distress \[[@B17],[@B18]\]. Another promising psychological intervention that may help to reduce symptoms of depression, anxiety, and diabetes-specific emotional distress is based on the cultivation of mindfulness. Mindfulness can be defined as the self-regulation of one\'s attention focusing on direct experience, while adopting a curious, open, and accepting attitude towards these experiences, especially one\'s psychological processes, such as thoughts and feelings \[[@B19]\]. Mindfulness-based interventions can be considered promising as positive effects of mindfulness programs on emotional well-being and quality of life have been reported in diverse patients groups, like chronic pain patients and patients with cancer \[[@B20]-[@B22]\]. In addition, research has shown that the program is beneficial for patients with recurrent depression. In these patients, the intervention appeared to be superior to traditional cognitive behavioral therapy in the prevention of relapse of depression \[[@B23]\]. While in a recent meta-analysis, mindfulness-based interventions have been found to have medium to large effects regarding the reduction of symptoms of anxiety and depression \[[@B24]\], only two studies have focused on the effects of a mindfulness program on patients with diabetes. These studies showed that the interventions resulted in improved emotional well-being (Cohen\'s *d*= 0.86 for depressive symptoms and 0.43 for anxiety) \[[@B25]\], improved self-care behaviors (Cohen\'s *d*= 0.68)\[[@B26]\] and decreased HbA~1c~values (Cohen\'s *d*= 0.46) \[[@B25]\] (Cohen\'s *d*= 0.35)\[[@B26]\]. However, only one of the two studies was a randomized controlled trial, but this study did not include emotional well-being as an outcome measure \[[@B26]\]. Therefore, there is a clear need for randomized controlled studies in this area, examining the potential benefits of mindfulness interventions on emotional well-being in patients with DM. In addition to the potential contribution of mindfulness to emotional well-being, mindfulness may also have a beneficial influence on self-management in DM patients. First, because mindfulness interventions enhance emotional well-being, and a better mood is associated with better self-care behavior, mindfulness is expected to enhance self-care behavior as well. Second, mindfulness improves one\'s ability to behave oneself constructively in harmony with his/her values, even during the experience of difficult thoughts and emotions \[[@B27]\]. Finally, because considerable focus is placed on bodily sensations during mindfulness interventions, mindfulness leads people to become more in touch with their body and its signals and needs \[[@B28]\], which is expected to lead to better self-care. A positive effect of a mindfulness intervention may also be expected on blood pressure. It is known that episodic or chronic emotional distress can increase blood pressure \[[@B29]\]. This is unfavorable for patients with diabetes, as the prevalence of high blood pressure in people with diabetes is approximately twice as high compared to the general population \[[@B30]\] and both diabetes and high blood pressure are important risk factors for cardiovascular disease \[[@B30]\]. To date, only two studies have examined the influence of mindfulness on blood pressure, finding a reduction in mean arterial pressure in DM patients \[[@B25]\] and a reduction in resting systolic blood pressure in normotensive students \[[@B31]\]. If the mindfulness intervention appears to be effective in reducing emotional distress in DM patients, it is useful to know which patients will most likely benefit from the intervention. However, research into differences in effectiveness of the intervention between groups of patients with different characteristics (e.g. differences in personality and demographic and medical characteristics) is scarce. Objectives and hypotheses ------------------------- Because of the paucity of studies on the effectiveness of psychological interventions for emotional problems in DM patients, especially regarding mindfulness based interventions, the purpose of this study is to test the effectiveness of a mindfulness-based psychological intervention. The primary outcome is emotional distress (symptoms of depression, anxiety, diabetes-specific stress, and general perceived stress). Secondary outcomes are quality of life, (diabetes-specific) mindfulness, and self-esteem, and our tertiary outcomes are self-management, glycemic control, and blood pressure. Finally, we will examine which patient characteristics (like the extent of complications and personality) predict which patients most likely benefit from the intervention. With regard to our first outcome measure, we hypothesize that after an eight-week mindfulness-based training DM patients will experience significant greater reductions in emotional distress (diabetes-specific stress, anxiety, and depressive symptoms) compared to a wait-list control group. Concerning our secondary outcome measures, we hypothesize that the intervention will lead to greater improvements in psychological quality of life, dispositional (diabetes-specific) mindfulness, and self-esteem compared to a wait-list control group. In addition, we expect, consistent with two other studies \[[@B32],[@B33]\], that changes in dispositional mindfulness will mediate the hypothesized effects of the intervention on above mentioned outcomes. Regarding our tertiary, more clinical, outcome measures, we hypothesize that the intervention will lead to better diabetes self-care, lower HbA~1c~, and lower BP in people with elevated BP. Finally, we are interested in personality and clinical factors that could possibly act as effect modifiers. We do not have clear expectations about the direction of the possible influence of these characteristics, being complications (yes/no), diabetes type (DM1 versus DM2), co-morbidity (yes/no), and personality. Methods/Design ============== Study design ------------ The Diabetes and Mindfulness (DiaMind) study will be conducted as a randomized controlled trial (RCT). Dutch speaking patients with DM with low levels of emotional well-being recruited from outpatient diabetes clinics will be randomly allocated to the intervention group or the wait-list (usual care) control group. The intervention group will receive the mindfulness program immediately, while the control group will receive the program eight months later (six months after the training of the intervention group). Eligibility ----------- We will include adult men and women (aged 18 - 80 years) who were diagnosed with diabetes (type 1 or 2) and have poor emotional well-being as evidenced by a score of \< 13 on the WHO-5 well-being Index \[[@B34]\]. Patients will be excluded when they have a recent history of severe psychopathology (i.e., psychosis, risk of suicide attempts); or alcohol/drugs abuse; have a severe physical co-morbidity (i.e., severe forms of cancer or heart failure); when they have insufficient reading and comprehension skills of the Dutch language; when they are already in an (extensive) psychological treatment which started within a period of 6 weeks before the start of the training; and when they already have meditation experience (with Vipassana, Zen, or Dzogchen). Recruitment and screening process and enrolment ----------------------------------------------- The recruitment of patients will take place in outpatient diabetes clinics. At present, internists of the outpatient clinic of the Maxima Medical Center (Veldhoven and Eindhoven), the Catharina Hospital (Eindhoven) and the TweeSteden Hospital (Tilburg) have agreed to participate. Depending on the preferences of the diabetes teams in the participating hospitals, the screening will be conducted by the internists, diabetes nurses or by a researcher of our team. The screening tool will be the WHO-5 well-being Index \[[@B34]\]. This short (5-item) screening instrument, which assesses emotional well-being, is validated in the general population, but also in different patient samples, including patients with type 1 diabetes \[[@B35]\] and is recommended for use by the International Diabetes Federation (IDF) \[[@B14]\]. If patients have a low score on this scale (\<13) they will receive an information letter about the training and study. One week later these patients will be telephoned and asked whether they are interested to participate in the project. Eligible patients who are interested to participate will be invited for a short interview. This interview will take place in the hospital or, if logistically difficult for the patient, at their home. During this interview the in- and exclusion criteria will be checked once again, and expectations of the patient about the training will be checked and if necessary adjusted. Eligible patients will receive a written informed consent form during the interview that has to be signed and returned before the start of the training. Furthermore, at the end of the interview the blood pressure of the patients will be measured. If blood pressure is elevated (≥140/90 mmHg) the patient will be attached to an ambulatory blood pressure device which will measure blood pressure the following 24 hours. One day later the device will be brought back to the hospital by the patient or be picked up at the patient\'s home by a researcher. Randomization ------------- After the inclusion of every 16 to 20 patients the participants will be randomized into two groups: the intervention group and the control group. The randomization will be done as follows. After the interview and upon receipt of a signed informed consent form, the patient will be assigned to a participant number. Subsequently the participant will receive the baseline questionnaire by mail or email. When the first author receives the baseline assessment, she passes on the corresponding participant number to the second author who has no further involvement in the practical recruitment, enrolment, and assessment of the patients. The second author will refer to the computer generated (through PASW Statisitics 17) random list (uneditable and concealed for others) prepared by a statistician with no involvement in the trial. The second author will inform the first author about the allocation by email and will archive the allocations in a secured document on his computer. The first author will document the allocation in the general inclusion database, which will be checked by the second author. The first author will inform the patients both by telephone and letter about their allocation. The above described procedure will eliminate experimenter bias in group assignment. The computer program the statistician will use utilizes a random number generator and will be programmed to insure equal numbers of subjects in each arm of the study within a block of 4 participants. Masking ------- The nature of this psychological intervention does not allow \"masking\" or blinding of patients, and trainers. However, the statistician who will be involved in data analyses will be blinded for treatment allocation. All the questionnaires and homework forms will only be marked with a participant number, which is unknown to the trainer and researcher. Intervention ------------ ### Training The protocolised mindfulness training is based on the Mindfulness-Based Stress Reduction and Mindfulness-Based Cognitive Therapy programs as described in Kabat-Zinn \[[@B28]\] and Segal et al. \[[@B36]\], consisting of eight weekly two-hour sessions. A few modifications have been made to the original protocol and the workbook in order to make the intervention suitable for patients with diabetes. For example, the trainers explain the potential associations between emotional problems and diabetes management and diabetes outcomes (e.g., the associations between emotional stress and eating behaviors will be discussed). Instead of the silent day that is part of the original program, a two-hour booster session has been added three months after the end of the intervention. At each session the participants will receive homework assignments that take about 30 minutes 5 days per week. All the sessions will be supervised by certified psychologists who have at least four years practical experience with mindfulness, and also completed a mindfulness instructors training of eight days in The Netherlands. The standard format of the sessions is as follows: the session starts with an exercise, which is followed by a discussion about the exercise and the homework exercises of the preceding week. Subsequently, a discussion about the theme of the session takes place, followed by the administration and discussion of one or two other mindfulness exercises. Finally, the homework assignments of the following week are discussed. All sessions end with a short meditation and/or a relevant poem or story. #### Session 1 The first session of the course will consist of: 1) an invitation to all participants to introduce themselves sequentially, 2) a brief overview of the coming eight weeks, 3) a discussion about the relationship between diabetes, diabetes management, diabetes outcomes and emotional distress. After this, participants will be invited to slowly and mindfully eat a raisin with all their senses (the raisin exercise), which will be followed by a discussion about the fact that many people live their lives in a unmindful way, often do not pay much attention to what they are doing. Subsequently, the participants will be led through a body scan exercise of thirty-five minutes, during which the participants are encouraged to focus on the physical sensations in different body parts and to bring their attention back to that focus when they get distracted from it by thoughts or other momentary phenomena. At the end of the first session the participants will receive a CD containing several exercises and will be asked to practice the bodyscan on five occasions in the following week. In addition, they will be advised to perform one daily activity (e.g. to wash the dishes, to brush their teeth) in a mindful way and to eat one meal per day mindfully. #### Session 2 At the beginning of the second session the participants will be guided again through the body scan, after which they will be invited to express and share their experiences with the mindfulness exercises and the homework of the preceding week. Subsequently there will be a discussion about coping with obstacles to doing the exercises (e.g., irritation, wandering mind) and attitudes that support the cultivation of mindfulness (e.g., not judging, letting go). After that, participants will be led through an exercise \'thoughts and feelings\', which explains a basic component of cognitive behavioral therapy: our emotions are not caused by events but by our perception of events. Subsequently, the participants will practice \'sitting meditation\', with breathing as the primary object of attention. Home practice for the following week will include awareness of pleasant events, a new routine daily activity, the bodyscan, and brief sitting meditation. #### Session 3 The third session will start with a short \'Seeing and/or hearing exercise\', in which participants are instructed to exercise non-judgmental seeing or hearing for a couple of minutes. This exercise will be followed by a sitting meditation with focus on breathing and bodily sensations. After the homework discussion, there will be a talk by the trainer about attention to breathing, followed by the practice and discussion of the \'3-minute breathing space\'. This meditation has three phases: attention to the experiences in the moment, attention to the breathing, attention to the body. Subsequently, an exercise in mindful body movements will be introduced and discussed. Exercises to be practiced during the following week will include sitting meditation, bodyscan or body movement exercises, 3-minute breathing space, new mindful daily activity, and awareness of unpleasant events. #### Session 4 The fourth sessions will start with a sitting meditation with attention to the breathing, the body, sounds, and thoughts (so called \'sitting meditation with four focuses\'). Subsequently, there will be a discussion about the stress response and common reactions of individuals to difficult situations, and alternative attitudes and reactions will be discussed. Then a Dutch documentary about MBSR (entitled \'Aandachttraining\') will be shown to the participants. At the end of the session an exercise in mindful walking will be introduced and practiced. The homework assignments for this fourth session will consist of practicing sitting meditation, the bodyscan or mindful body movements, 3-minute breathing space (also during unpleasant events), to fill in the questionnaire \'automatic thoughts\' (optional), and to read an inspiring book about mindfulness (optional). #### Session 5 At the beginning of the fifth session the participants will be guided through the sitting meditation with four focuses (see Session 4). A poem will be read and the psychologist will discuss the theme of the fifth session: \'Acknowledging and accepting the reality of the present situation as it is.\' Subsequently, the second series of mindful body movements will be introduced. The recommended home work will consist of sitting meditation, 3-minute breathing space, 3-minute breathing space during unpleasant events, and a new daily activity performed mindfully. #### Session 6 The sixth session will begin with the 3-minute breathing space. The homework assignments will be discussed mindfully in pairs. An exercise called \'mood, thoughts, and different view points\' will be discussed plenary, as well as the theme of the sixth session: \'The content of our thoughts is (often) not factual\'. Subsequently, the participants will be guided through four linked up meditation exercises, with a total duration of one hour. For the following week participants will be encouraged to choose a combination of meditations that fits their needs and personal preferences. In addition, the home practice includes the 3-minute breathing space (also during unpleasant events), a new mindful daily activity, and having a mindful conversation while paying attention to one\'s automatic patterns (optional). #### Session 7 The seventh sessions will start with the meditation exercise with four focuses and open awareness (to whatever enters consciousness from moment to moment). The theme of the session, \'What is the best way to care for oneself\', will be discussed, followed by an exercise during which participants explore daily activities that are pleasant versus unpleasant and learn to plan sufficient pleasant activities. The 3-minute breathing space is performed, followed by a loving-kindness meditation during which one practices kindness towards oneself and others. The home practice includes a meditation combination of one\'s choice, 3-minute breathing space (also during unpleasant events), a new daily activity performed mindfully, and filling in a form including warning signals for emotional distress. #### Session 8 The session will begin with a body-scan. The theme of the last session will be discussed: \'Using what you have learned\'. Subsequently the psychologist will guide the participants through a 3-minute breathing space and will read a story called \'Five chapters\', which is about the steps one can take to overcome our pitfalls in the practice of mindfulness. Finally, the whole training will be evaluated with the participants. For example the following questions will be asked: did the training meet their expectations, do participants have a sense of \'personal growth\', do they feel that they have expanded their coping skills and do they want to continue practicing mindfulness. In addition, resources for continued practice will be given (e.g. titles of books and addresses of local meditation centers). Evaluation ---------- After completing the training the patients will be invited for a short interview, in which the patient\'s perceptions regarding the quality of the intervention will be evaluated and the procedure with the ambulant blood pressure measurement will be repeated. Data collection --------------- Table [1](#T1){ref-type="table"} shows the assessment instruments and data collection time points. Assessments will take place before the start of the intervention, at 4 weeks, at 8 weeks (= post intervention), at 12 weeks and at 34 weeks (6 months post intervention). The control group will receive another four measures, namely at 38 weeks (4 weeks after start of control group intervention), at 42 weeks (post control group intervention), at 46 weeks, and at 62 weeks (6 months post control group intervention). Process variables (e.g., attendance, adherence, and drop-out) will be collected continuously. The patients who have access to the internet and email will receive the questionnaire via the internet. If this is not the case, they will receive a paper version of the questionnaire along with a stamped addressed envelope. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Measurements and time points ::: Concept Questionnaire Measurement time points --------------------------------------------------------------- -------------------------- ------- ------------------------- ------- ------- ------- --- --- --- --- Perceived stress PSS (14)\* **x** **x** **x** **x** x x x Anxiety and depressive symptoms HADS (14) **x** **x** **x** **x** x x x Mood POMS (32) **x** **x** **x** **x** x x x Diabetes related problems PAID (20) **x** **x** **x** **x** x x x Quality of life SF12 (12)+3 items WHOQOL **x** **x** **x** x x Self-care behavior ADSCI (11) **x** **x** **x** x x Extraversion EPQ-E (12) **x** Type D Personality DS14 (14) **x** **x** **x** x x Mindfulness FFMQ(31) **x** **x** **x** x x Diabetes related mindfulness AADQ (11) **x** **x** **x** x x Self-esteem RSES (10) **x** **x** **x** x x Patient\'s background (demographics, clinical, psychological) \- (21) **x** **Clinical measurements** **Measurement** Glycemic control (from hospital database) hbA1c **x** **x** **x** x x Blood pressure Amb.24 h BP **x** **x** T1 - baseline; T2 - after 4 weeks of intervention; T3 - after 8 weeks (post intervention); T4 - after 12 weeks (4 weeks post intervention); T5 - after 8 months (6 months post intervention = start training for control group). T6 thru T9 only for control group: repetition of T2 thru T5. PSS - Perceived Stress Scale; HADS - Hospital Anxiety and Depression Scale; POMS - Profile of Mood States; PAID - Problem Areas in Diabetes Survey; SF-12 - Short Form Health Survey; 3 items WHOQOL-bref^9^; ADSCI - Amsterdam Diabetes Self Care Inventory; EPQ-E - Eysenck Personality Questionnaire subscale Extraversion; DS14 - Type D personality scale; FFMQ - Five Factor Mindfulness Questionnaire; AADQ - Acceptation and Action Diabetes Questionnaire; RSES - Rosenberg Self-esteem Scale. HbA1c = hemoglobin A1c (the amount of glycated hemoglobin in blood). Amb. 24 h BP = Ambulatory 24 hour Blood Pressure monitoring. \*= number of items in questionnaire. Total number of items = 205. ::: Outcome parameters ------------------ ### 1. Primary outcome measures #### Stress, anxiety and depressive symptoms We will include the Hospital Anxiety and Depression Scale (HADS) to measure symptoms of anxiety (HADS-A: e.g., \"Worrying thoughts go through my mind\") and depression (HADS-D: e.g., \"I feel as I am slowed down\") in DM patients \[[@B37]\]. The self-report scale consists of 14 items, answered on a 4-point Likert scale, with the HADS-A and HADS-D both comprising 7 items (*0-3*). The score range for the anxiety and the depressive symptoms subscales is 0 to 21. The HADS has been validated in the general population and in somatic and psychiatric patients \[[@B38],[@B39]\]. It has been shown to be a valid and reliable instrument with Cronbach\'s α ranging from 0.67 to 0.93 for the two subscales \[[@B37]-[@B39]\]. The Dutch version of the Perceived Stress Scale (PSS) will be included to measure general perceived stress (the degree to which situations in one\'s life are appraised as stressful) (e.g. \"In the last month, how often have you felt nervous and stressed?\"). It has been shown to be a reliable measure with Cronbach\'s α ranging from 0.84 to 0.86 \[[@B40]\]. The present version consists of 10 items which are answered on a five-point Likert scale, ranging from \'never\' to \'very often\' *(0-4)*\[[@B41]\]. The short Dutch version of the Profile of Mood States (POMS) \[[@B42]\] will be used to assess transient, fluctuating mood states. The scale consists of 32 adjectives about positive and negative mood states which have to be rated on a five-point Likert scale *(0 = not at all, 4 = very much)*based on how well each item describes one\'s mood during the last couple of weeks. For the interpretation of the results the items are divided into five subscales: Tension-anxiety (6 items); Depression-dejection (8 items); Anger-hostility (7 items); Vigor-activity (5 items); and Fatigue-inertia (6 items). The scale has sufficient consistency reliability and construct validity, with Cronbach\'s α\'s of the five subscales varying between 0.82 and 0.91 \[[@B42],[@B43]\]. #### Diabetes specific emotional problems We will include the Dutch version of the self-report questionnaire Problem Areas in Diabetes Survey (PAID), which consists of 20 statements about common negative feelings related to living with diabetes (e.g., \"Feeling depressed when you think about living with diabetes\", \"feeling discouraged with your diabetes regimen\") \[[@B44]\]. The items are rated on a six-point Likert scale *(1 = not a problem, 6 = a serious problem)*. PAID scores are transformed to a 0-100 scale, to facilitate interpretation \[[@B45]\]. A higher score indicates more emotional distress, with a cut-off score of 40 indicating seriously elevated emotional distress \[[@B46]\]. Factor analysis revealed that the factor structure can be represented by a one factor model (PAID total: emotional adjustment) and a four factor model (subscales: emotional problems, treatment problems, food problems, and lack of social support) \[[@B45]\]. We will examine the results with both models. The survey has proven to be a reliable measure, with a Cronbach\'s α of 0.95 for the one factor model. For the four factor model the Cronbach\'s α\'s are in range of 0.93 (emotional problems), 0.74-0.76 (treatment problems), 0.70-0.74 (food problems), and 0.69-0.72 (lack of social support) \[[@B45]\]. In addition, research in Dutch and American patients found support for convergent and discriminative validity of the PAID \[[@B45],[@B47]\]. ### 2. Secondary outcome measures #### Quality of life We will include the Dutch version of the Short-Form Health Survey (SF-12) to measure health related quality of life. The self-report scale consists of 12 items that are grouped into two component summary scores: a physical (PCS) and a mental component score (MCS). Both component scores are measured on a scale from 0 to 100, with a high score indicating good health related quality of life. The SF12 has proven to be a valid and reliable measurement \[[@B48]\]. In addition, we will use three items of the WHOQOL-BREF \[[@B49]\] to measure satisfaction with oneself and life in general: \"How much do you enjoy life?\"; \"To what extent do you feel your life to be meaningful?\"; \"How satisfied are you with yourself?\". The questions will be analyzed separately at item level if Cronbach\'s α will show to be below 0.65. #### Mindfulness To measure mindfulness we will include the Dutch version of the Five Facet Mindfulness Questionnaire (FFMQ). FFMQ is based on a factor analytic study of five independently developed mindfulness measurement scales \[[@B50]\]. The factors we will use are: *Observing*, which refers to noticing or attending to internal and external experiences, such as thoughts, sensations, emotions, sounds, sights, and smells; *Acting with awareness*, which includes attending to one\'s activities in the present moment; *Non-judging of inner experience*, which refers to taking a non-evaluative attitude toward thoughts and feelings; and *Non-reactivity to inner experience*, which includes allowing thoughts and feelings to come and go, without getting caught up in or carried away by them \[[@B51]\]. The factor *Describing*will be omitted, because describing one\'s emotions and feelings is not a primary focus of MBSR/MBCT and we do not want to burden our patients unnecessarily. Each factor consists of 7 or 8 items, which will be answered on a five-point Likert scale (*1 = never or very rarely true, 5 =very often or always true*). The four facets demonstrated adequate to good internal consistency (alphas ranging from 0.75 to 0.87) \[[@B50]\]. Changes in diabetes-specific mindfulness and acceptance processes will be assessed by the Acceptance and Action Diabetes Questionnaire (AADQ) \[[@B26]\]. This self-report questionnaire consists of 11-items that are answered on a 7-point Likert scale. It measures acceptance of diabetes-related feelings and thoughts and the extent to which they interfere with valued action (e.g., \"I avoid thinking about what diabetes can do to me\") \[[@B26]\]. For this measure, a Cronbach\'s α of 0.94 has been reported \[[@B26]\]. #### Self-esteem Self-esteem will be measured by the Dutch version of the Rosenberg Self-Esteem Scale (RSES), wherein self-esteem is defined as a person\'s overall evaluation of his or her worthiness as a human being \[[@B52]\]. The questionnaire consists of 10 items answered on a four-point Likert scale, ranging from \'I totally disagree\' to \'I totally agree\'. A sample item is \"I take a positive attitude toward myself\". A higher score reflects a higher global self-esteem. The Cronbach\'s alpha was .86, indicating a high internal consistency \[[@B52]\]. ### 3. Tertiary outcome measures #### Self-care behavior Self-care behavior (or self-management) will be assessed with the Amsterdam Diabetes Self Care Inventory (ADSCI). The questionnaire has 11 items, which will be examined separately. Each item is divided into three or four subquestions. Higher scores indicate higher levels of self-care. The questionnaire is developed by colleagues of the VU University Medical Center in Amsterdam and used in previous trials \[[@B53],[@B54]\]. The psychometric properties of the ADSCI have not yet been published. #### Glycemic control To measure glycemic control we will consult the three-monthly standard measurements of HbA1c of the hospital, which are documented in the hospital\'s patient database. The percentage HbA1c is a function of the glucose concentration to which the red blood cells are exposed and gives an indication of the average blood glucose concentration of the preceding six to eight weeks \[[@B3]\]. #### Blood pressure The patients\' blood pressure will be measured by an ambulant device, called the Mobil-O-Graph, which is based on the oscillometric method and which has been shown to provide reliable and valid assessments \[[@B55]\]. The patients, who have elevated high blood pressure (≥ 140 mmHg systolic or 90 mmHg diastolic of the mean values of 3 measurements) during the interview, will wear the meter 24 hours following the interview. The ambulatory measurements will occur two times per hour. To keep the burden for the patients as low as possible, we decided to measure the blood pressure only during the daytime. The device will be attached to the patient by the researcher who does the interviews (JvS). ### 1. Additional outcome measures #### Personality To measure personality dimensions we will use the subscale Extraversion (EPQ-E) of the shortened revised Eysenck Personality Questionnaire (EPQ-RSS) and the DS-14 to measure Type D personality. Persons with a Type D personality tend to experience high levels of emotional distress, but do generally not express these emotions. For the EPQ-E, the subscale Extraversion reflects sociability, assertiveness, and the tendency to experience positive emotions. The subscale consists of 12 dichotomous (yes/no) items \[[@B56]\]. The Committee on Test Affairs Netherlands rates the EPQ to be a reliable measure \[[@B57]\]. The 14-item Type D scale (DS14) measures both negative affectivity (NA) (e.g. \"I often feel unhappy\"; 7 items) and social inhibition (SI) (e.g. \"I am a closed kind of person\"; 7 items) \[[@B58]\]. Items are answered on a five-point Likert scale ranging from \'false\' to \'true\' *(0-4)*, with the score range for the NA and SI subscales being 0 to 28 \[[@B58]\]. We will use the standardized cut-off ≥ 10 on both subscales to identify Type D caseness \[[@B58],[@B59]\]. The DS14 has been shown to be a valid and reliable instrument, with Cronbach\'s α of 0.88 and 0.86 and a 3-month test-retest reliability of r = 0.72 and 0.82 for the NA and SI subscales, respectively \[[@B58]\]. Furthermore, a recent study showed that Type D personality is a stable construct over an 18-month period and is not confounded by disease severity and measures of anxiety and depression \[[@B60],[@B61]\]. Results from a study by Denollet et al \[[@B62]\] suggest that particularly the interaction between NA and SI is predictive of adverse health outcomes, more than the single traits \[[@B62]\]. #### Demographic and clinical variables Demographic variables will be collected by means of a questionnaire, which will be completed by the patients during the baseline assessment. Data will be collected regarding the participant\'s age, marital status, education, and job status. In addition, the baseline questionnaire will be used to gather information on history of emotional problems and use of psychotropic medication, level of meditation experience, sleep quantity and quality, participant\'s perceived importance of adequate blood glucose regulation, and amount of participant\'s effort to manage their blood glucose. Furthermore, the baseline questionnaire will be used to assess clinical data, such as length and weight of participants, last HbA1c, number of severe hypoglycemic episodes and/or hospitalizations because of a diabetic coma during the past year, existing diabetes complications, co-morbidities, and smoking and drinking behavior. The clinical variables (which are available) will also be retrieved from the hospitals\' patient information database. ### 1. Additional process measures During the intervention process data will be collected, like drop-out, patients\' attendance of the sessions, patients\' adherence to the home work assignments, and trainers\' adherence to the protocol. The adherence to the protocol will be determined each session by the trainers by means of the completion of a checklist including all components of each session. The components are divided into important and less important, facilitating decision making when time constraints are present. Statistical analyses -------------------- ### Sample size and power calculation The total number of patients that have to be included and complete the study is 160, with 80 patients in each arm, taking into account a medium effect size, a power of 0.80, an alpha of 0.05, patient attrition, and subgroup analyses involving two equally sized groups (e.g. yes vs. no complications). It is expected to reach this number of participants during the course of the study. ### Planned analyses Repeated measures analysis of variance ((M)ANCOVA) will be used to test the hypotheses concerning the differences between groups on the dependent variables over time. In these tests age, sex, and co-morbidity will be included as covariates, in addition to those other potentially confounding variables that will show an at least marginally significant (p \< .10) difference between groups. The analyses concerning the subgroup effects will be conducted on the sample as a whole, whereby possible moderating variables, like complications and personality will be included in the analyses as between-subjects factors. All analyses will be based on the intention-to-treat approach. Ethical principles ------------------ The study protocol has been approved by the medical ethical committee of the St. Elisabeth Hospital in Tilburg, The Netherlands (P0948). Discussion ========== The present article provides an outline of the background and design of the Diabetes and Mindfulness (DiaMind) study. This study\'s objective is to test an intervention based on mindfulness training supplemented with elements from cognitive therapy that aims to improve emotional well-being of distressed patients with diabetes. For this purpose, an intervention group will be compared with a wait-list control group with treatment as usual. Research on psychological interventions that aim to reduce emotional problems in DM patients is important, as emotional problems are associated with a higher risk of impaired quality of life \[[@B8]\], less optimal self-care \[[@B9]\], incident complications, and mortality \[[@B8],[@B11]\]. However, to date there are few methodologically well-designed studies on the effectiveness of psychological interventions for people with diabetes and emotional complaints. In addition, the majority of these studies have focused on improving self-management instead of emotional well-being. Moreover, the few randomized studies that have examined the effectiveness of interventions in reducing emotional problems in people with diabetes often focused on depression. For this reason, this study is clearly of added value in the research field of diabetes and emotional problems, as we have taken a broader perspective and focused on emotional distress (depression, anxiety, and general distress). The study design has a number of important strengths. First, the present study has more power to generate knowledge on the effectiveness of a mindfulness-based intervention for DM patients with emotional problems than former studies examining mindfulness in DM patients. As mentioned before, only two studies have been done on this subject, of which one lacked a control group and one did not include emotional well-being as an outcome \[[@B25],[@B26]\]. Second, an additional strength of this study is the inclusion of both psychological and clinical measures, while most randomized controlled trials examining mindfulness interventions include psychological measures only. Finally, another strong aspect of this study is the examination of factors that potentially moderate the effectiveness of the mindfulness intervention, including clinical (e.g. complications) and psychological (e.g. personality) variables. Results may facilitate more effective allocation of patients to treatments. The recruitment has already started in the Catharina Hospital in Eindhoven in spring 2010 and in the Maxima Medical Center in autumn 2010. At present, 969 individuals have been screened for emotional problems. Of those, 811 did not qualify: 658 had a score above 12 on the WHO-5 screening instrument, and 153 were ineligible (no mastery of the Dutch language, psychiatric disorder, or severe physical condition). Of the eligible patients, 120 declined participation (no perceived need or motivation for the training or an incompatible working schedule). At present, 42 patients are interested to participate, of whom 16 have been enrolled so far. The recruitment of a sufficient number of patients will be a challenge. For instance, the mindfulness intervention requires participants to commit to attending eight two-hour group sessions and to practice exercises at home for five days per week. For people who have busy lives this sometimes is a challenge to comply with, which may make them decline participation or drop out of the study. Concerning the benefits of the present study, the results will provide valuable information regarding the effectiveness of the mindfulness intervention for DM patients in improving emotional well-being. We will also gain insight in the pros and cons of a possible implementation of the intervention in usual medical care of DM patients. By doing this, we aim to find ways that can contribute to increased emotional well-being and quality of life in patients with diabetes mellitus. In conclusion, considering the prospected increase of the prevalence of diabetes, the levels of emotional distress in this patient group, and the incompleteness of the current literature on effective psychological interventions for the distressed patients, well-designed trials on psychological interventions are a welcome addition in the diabetes research. In the present randomized controlled trial the effectiveness of a psychological intervention will be examined in which the cultivation of mindfulness (the direction of attention to one\'s experiences in the present moment in an open and accepting manner) plays a central role. This intervention may be an effective alternative to traditional cognitive behavioral therapy in improving emotional well-being in DM patients. The first results of the study will become available in 2012. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= All authors contributed to the design of the study. IN is the principal investigator of the study. IN and FP coordinate and supervise JvS (a PhD-student). JvS drafted the manuscript. All authors contacted professionals (i.e. internal medicine doctors) involved in the recruitment of patients. JvS is responsible for the logistics of the study, and takes care of the recruitment of participants and data collection. IN, FP and JvS prepared the patients\' work book. JvS prepared the detailed protocol for the trainers of the intervention. All authors provided comments, read and approved the various versions including the final version of the manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/131/prepub> Acknowledgements ================ The study is supported by grants from the Dutch Diabetes Research Foundation or \'Diabetesfonds\' (project number 2008.13.005, awarded to Dr I. Nyklíček) and Tilburg University. The authors wish to acknowledge and express their special thanks to internal medicine physician Dr Marion Blonk, for her efforts in recruiting patients at the Catharina Hospital Eindhoven. In addition, we want to thank the internal medicine physicians Dr Bravenboer, Dr Erdtsieck, Dr Lieverse, and Dr Spooren, their secretaries, and the diabetes nurses of the participating hospitals, for their continuing assistance in the recruitment of patients for this study. Moreover, we want to thank the general practitioners and nurse practitioners of the Health Centre Plus Eindhoven and Vlaslant Valkenswaard, and their patients for their participation in the pilot study.

PubMed Central

2024-06-05T04:04:19.540354

2011-2-24

PMC3053245

Introduction ============ Overweight and obesity is one of the major threats to the health of Australians, as it affects a significant proportion of the population (about 60% of adults and 25% of children) and is a key risk factor in the development of chronic diseases such as type II diabetes, coronary heart disease and many cancers \[[@B1]\]\]. Obesity prevalence has risen rapidly in recent decades and also disproportionately affects people from socially and economically disadvantaged backgrounds \[[@B2]-[@B7]\]. The behavioural patterns contributing to high rates of obesity include increased consumption of high energy density foods and beverages, low consumption of fruit and vegetables and a shift to less active transport and more sedentary leisure time activities \[[@B8]-[@B10]\]. Obesity develops in a social and environmental context in which facilities, policies, economic factors and socio-cultural influences promote these behaviours \[[@B11],[@B12]\]. Thus, it is likely that education and treatment approaches alone will not be sufficient to reverse the obesity epidemic or its socio-economic gradient \[[@B13]\]. A number of countries have initiated [C]{.underline}ommunity-[B]{.underline}ased [O]{.underline}besity [P]{.underline}revention [I]{.underline}nitiatives (CBOPIs) as part of their attempts to address the obesity epidemic. The rationale for such an approach is multi-faceted and strong \[[@B14]\]. The strengths of CBOPIs to prevent obesity include the ability to influence a wide range of determinants of nutrition and physical activity behaviours and the ability to utilise and strengthen existing community assets and capacity in multiple community-settings \[[@B15],[@B16]\]. Whole-of-community interventions are consistent with a socio-ecological understanding of obesity, involving multiple spheres of influence, and are a crucial complement to individual-focused activities. In addition, interventions which focus on promoting healthier environments are likely to be more equitable than those primarily using educational approaches \[[@B17],[@B18]\]. However, CBOPIs themselves can only be part of a more comprehensive response to obesity which would include national (and even international) policies and actions - especially in creating healthier food supply and marketing systems. Internationally, Ensemble prevenons l\'obesitie des enfants (EPODE) is one of the first CBOPIs to show an effect on childhood obesity. In a pilot study, this whole of community intervention showed a significant decrease in obesity for intervention compared to comparison towns after 12 years \[[@B19]\] and has now been extended across France, Spain and Belgium \[[@B20]\]. In a number of other countries, including the USA, Australia and New Zealand, evaluation of demonstration projects in community-based obesity prevention have recently begun to show promising results, particularly in children \[[@B16],[@B17],[@B21],[@B22]\]. Although results to date have been relatively modest in most cases, small changes affecting large populations will result in significant population health gains \[[@B23]\]. There are as many as 12 large-scale (highly evaluated) CBOPI demonstration projects around Australia, many of which are ongoing. Many other similar large scale programs are planned or underway. There are also countless smaller, setting-specific interventions (e.g. schools, pre-schools, worksites). This range of projects is developing the evidence and expertise about what works and what does not work in promoting healthy eating and physical activity and sharing this information is important for national and international audiences. A central network could ensure that community projects are informed by the highest quality, best available knowledge, evidence, expertise and experience collated from existing CBOPI. A national collaboration may help to reduce the unnecessary costs of duplication of effort in activities such as extracting evidence from the literature, designing programs, developing measurement tools, creating intervention resources. In this paper we introduce the Australian [CO]{.underline}mmunity-based [O]{.underline}besity [P]{.underline}revention [S]{.underline}ites (CO-OPS)Collaboration, with the view to informing the establishment of similar collaborative networks worldwide. We introduce CBOPIs and describe the development and form of the CO-OPS network. The domains of activity for the CO-OPS Collaboration are described in detail and include; establishing the depth and breadth of CBOPI in Australia; establishing a collaborative network of these CBOPI; developing a set of best practice principles for CBOPI; promoting evidence, evaluation and analysis in practice; establishing a practice relevant evidence base; and knowledge translation and exchange (including dissemination). Collaboration of Community-based Obesity Prevention Sites - The CO-OPS Collaboration ------------------------------------------------------------------------------------ In 2007, three Australian universities, led by Deakin University and in collaboration with the University of Melbourne and the University of Sydney, received four years of funding from the Australian Federal Government Department of Health and Ageing to create the CO-OPS Collaboration. The core aims of the Collaboration are to: • identify and analyse the lessons learned from a range of community-based initiatives aimed at tackling obesity, and • identify the elements that make community-based obesity prevention initiatives successful and share the knowledge gained with other communities The structure of the CO-OPS Collaboration incorporates a steering committee, advisory committee, secretariat and the broader membership of practitioners, researchers and others with an interest in community-based obesity prevention. The secretariat is the central (staffed) structure of the Collaboration, with expertise in public health research, program evaluation, community engagement and knowledge translation and exchange. Defining the boundaries - what is community-based obesity prevention? --------------------------------------------------------------------- Initiatives were considered to be \'obesity prevention\' if they intended to promote healthy weight or prevent unhealthy weight gain and/or promote healthy eating or physical activity in a manner that could be expected to influence energy balance in communities or populations. Defining whether an initiative is \'community-based\' is more problematic. Definitions of \'community\' include common themes of geography, social interaction (mutual support, a sense of belonging, interlinked networks) and/or common ties (beliefs, activities, culture, interest, experience, political and social movements, etc). \'Community-based\' may be as broad as any program that involves community engagement or participation \[[@B24]\]. For the CO-OPS Collaboration, \'community-based\' means projects that focused on whole communities (however defined) or which intervene at a population level. This excludes programs specifically or exclusively focused on individual education or behaviour change. Community-based programs are most often delivered at or through local community settings (e.g., schools, workplaces, community centres, etc), although they may be centrally organised but locally delivered. Clinical activities in the community are excluded as they are individual level activities delivered in the community rather than population-focused promotion of health \[[@B25]\]. How many initiatives are there in Australia? -------------------------------------------- A \'top-down\' snowball sample was employed to develop the CO-OPS network and better understand current community-based obesity prevention in Australia. Top-down recruitment involved identification of CBOPIs by state government health departments and snowball sampling asked participants in the survey to identify other relevant initiatives known to them. Of the 78 initiatives identified, around a quarter of projects were focused on the entire spectrum of ages (24%), while among those that targeted specific groups, the largest proportion of projects targeted primary school children (37%) and adults (33%). The projects ranged from less than 100 participants (11% of projects) to interventions in communities of more than 50,000 (26%). Establishing a collaborative network ------------------------------------ The CO-OPS Collaboration was established to provide members access to networks of health professionals, researchers and government employees all interested in community-based obesity prevention. CO-OPS as a \'community of practice\' grew from networks of known individual practitioners working within existing initiatives to a community of more than 1300 professional members representing practitioners and researchers in CBOPI across Australia and internationally. CO-OPS is the \'community of practice\' for those working in community-based obesity prevention in Australia. CO-OPS extends beyond an information sharing network by intentionally creating new knowledge in collaboration. As a community of practice this new network has created a system with qualities and capacities beyond those present within individuals acting separately \[[@B26]\]. CO-OPS can be conceived of as a \'community of practice\' providing a forum for people to \"share a concern, a set of problems, or a passion about a topic, and who deepen their understanding and knowledge of this area by interacting on an ongoing basis\"\[[@B27]\] Communities of practice comprise three essential elements. Firstly they have a *domain*: this is the concern, issue or passion, and for CO-OPS the domain is obesity prevention (including promotion of nutrition and physical activity); secondly they have a *community*, these are the people who care about the domain (obesity prevention), they share and combine ideas, thoughts and questions, for CO-OPS the community consists of those who have become members of the website, those who attend training or workshop events and those working with CBOPI who are yet to join CO-OPS; thirdly they have practices, practices help the community (CO-OPS members) to develop new knowledge about the domain (obesity prevention). For CO-OPS, practices include the Best Practice Principles, Evidence Summaries, National Workshop presentations and resources from member (community) projects. Practices can be written or codified and they can also be beliefs or stances i.e. obesity requires interventions at the community level.\[[@B27]\] Communities of practice have varying levels of participation by members from a core group (CO-OPS Secretariat) through active (Advisory Committee members, those regularly contributing to newsletter) and occasional participants (members who assist to organise training, provide resources for the website, contribute to case studies, attend the National Workshop, submit requests for assistance through the website) to peripheral (network members who have chosen to sign up to receive the newsletter however do not contact or contribute to CO-OPS and those who know of CO-OPS however have decided not to join) and transactional participants (these are people, organisations or contractors who provide services). Levels of participation are not static and CO-OPS members easily shift between roles depending on the need of the network and the expertise available amongst its members \[[@B28]\]. Success in the establishment of the collaborative network to date has been demonstrated by the rapid growth in membership from an initial 30 founding members to now include over 1300 professionals (end 2010) who are active in research or practice relating to community-based obesity prevention. Of equal importance is the breadth of membership: members of the network include representatives from federal government departments, all state and territory governments (various departments, including health, human services, transport and planning) 52 different local government authorities, 42 community health organisations, 23 non-governmental organisations (NGOs) and 22 Australian universities. The size of the network and the interactions that network members have with the Collaboration is approximated by the number of visits to the CO-OPS website. Total numbers of visitors to the website, by month, are shown in figure [1](#F1){ref-type="fig"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Web site page visits to the CO-OPS Collaboration web site, May 2009 to November 2010**. :::  ::: Establishing Best Practice Principles for Community-based Obesity Prevention ---------------------------------------------------------------------------- A set of \"Best Practice Principles for Community-based Obesity Prevention Initiatives\" (BPP) was developed as a core document to guide the activities of the CO-OPS Collaboration and provide an important reference and guide for its members\[[@B24]\]. The BPP were developed by combining information obtained from a two-stage literature review with extensive consultation to elicit the experience and knowledge of stakeholders in well-developed and well-evaluated obesity prevention programs across Australia. Following extensive consultations and reviews, 22 Best Practice Principles were developed in five key areas. Details of the development, content and application of the principles are published elsewhere\[[@B24]\], however the principles are summarised in table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Best Practice Principles for Community-based Obesity Prevention ::: --------------------------------------------------- ------------------------------------------------------------ **Community engagement** **Evaluation** C1. Approach to community engagement E1. Evaluation framework and approach C2. Community analysis E2. Evaluation plan C3. Implementation partnerships E3. Data collection and management C4. Program sustainability and community capacity E4. Evaluation context E5. Active dissemination **Program design and planning** **Implementation and sustainability** P1. Problem analysis and program focus I1. Consumer testing of messages, resources and approaches P2. Framing of the problem I2. Quality implementation and monitoring P3. Planning context I3. Adaptations and responsiveness P4. Evidence and innovation P5. Theory of change/change process P6. Feasibility **Governance and accountability** P7. Program plan G1. Explicit funding sources P8. Target groups G2. Program management structure --------------------------------------------------- ------------------------------------------------------------ ::: As a practical resource for members of the CO-OPS network, the BPP were published as three connected resources. The first BPP resource, the *Outline and Rationale of Best Practice Principles for Community-based Obesity Prevention*is the most detailed section. It presents principles, supporting literature and commentary on specific issues related to the application of the principle to community-based obesity prevention. The second resource, *Guiding Questions for Community-based Obesity Prevention*, provides a set of practical guiding questions relating to each of the principles. These questions are designed to assist in applying the principles and are intended as a learning tool for practitioners. The third resource, *Short Guiding Questions for Community-based Obesity Prevention*, presents a simple checklist and easy introduction to the full documentation of BPP and the guiding questions. The documents are available for download from <http://www.co-ops.net.au>. To date, the response to the BPP has been very positive, and the resource has been downloaded over 800 times, in addition to over 1500 hard copies distributed through professional development sessions and networking opportunities. This clearly demonstrates the reach of the BPP and a strong demand for such a resource. Elements of the BPP have been incorporated into the quality framework for a very large federally funded community-based obesity prevention initiative (Healthy Communities) and preliminary qualitative feedback indicates that the BPP have been employed in a range of settings to guide planning, implementation and evaluation of programs. A more comprehensive review of the use and implementation of the BPP is planned for early 2011. Promoting evidence, evaluation and analysis in practice ------------------------------------------------------- A core aim of the CO-OPS Collaboration is to promote knowledge generation and translation among community-based obesity prevention initiatives and thereby reduce duplication of effort. Included in this aim is the need to promote high quality evaluation of existing programs and dissemination of these results. Key activities to promote evidence quality and analysis include identifying robust tools and methods for measurement and exploring the potential of data management processes for meta-analyses. Creating evaluation and data management support service for projects will be very important in promoting evaluation results which can be translated to other settings and inform future practice and to enhance capacity within community initiatives. Currently, very little is known about the scope and rigour of evaluations being conducted in the majority of CBOPIs and initial consultation with community based professionals including members of the CO-OPS Collaboration identified the greatest need for skill development was in evaluation. The limited academic literature indicates little agreement about evaluation design, methodology (particularly data collection) and as a result limited consistency or comparability across evaluations. High quality evaluation and clear evidence of effectiveness is needed for community-based interventions to progress. CO-OPS is beginning to address this need through: • identification and promotion of high quality evaluation tools and instruments for consistent and comparable data collection, • provision of support and advice to practitioners evaluating CBOPI, • dissemination of evaluation findings to the network, to promote successes and reduce repetition of failures, and • meta-analyses of CBOPI evaluation findings. Establishing a practice relevant evidence base ---------------------------------------------- A consistent finding of national consultations was the need for accessible and practice relevant summaries of knowledge and evidence summaries. In response, CO-OPS supported the development of three practitioner-led evidence summaries. The process involved workshops to develop an \'answerable\' research question with practitioners, discussion of the best ways in which to present evidence summaries, a review of systematic reviews related to the question as defined (guided by the results of the workshops) and publication and dissemination of the evidence summary both as hard copies and via the project web site. The topics of the three evidence summaries developed to date are: • Remote and rural issues in the prevention of obesity for pre-adolescents and adolescents • Achieving equity in community-based obesity prevention interventions for children and adolescents • Considerations regarding harm minimisation for obesity prevention policies and programs for pre-adolescents and adolescents Local government is an important platform for the implementation of CBOPIs and this is an increasing level of activity in this area. The role of local government in obesity prevention was emphasised by the recent commitment of Australian governments to the Healthy Communities Initiative, which charged local government with delivering nutrition and physical activity promotion through community-based programs. A second round of evidence summaries were developed to clearly understand the role of local government in obesity prevention and to provide evidence relevant to this setting. Consultation with sites in the CO-OPS network was conducted to understand the capacity of local government in using research evidence by developing and evaluating an \'evidence tool kit\'. The purpose of the toolkit was to strengthen capacity of local governments to use research evidence when formulating obesity prevention policies and programs at the local level. Australia does not have a national, uniform system for monitoring overweight and obesity among children. Such a system would allow examination of trends in obesity over time and also provide information for evaluation of program effectiveness. CO-OPS has commissioned a series of reports to inform the development of a monitoring system, addressing; *monitoring system design; ethics assessment; and, feedback systems of monitoring information*. In addition to the six newly developed evidence summaries and the three reports described above, the CO-OPS resource library currently includes a wide variety of resources relevant to evidence-based practice, including systematic reviews, validated evaluation tools and project reports from large and well-evaluated programs (84 resources at the end of 2010). The evidence summaries have been among the most popular resources downloaded form the CO-OPS website, surpassed only by the BPP. The growth in usage of the CO-OPS website (figure [1](#F1){ref-type="fig"}) demonstrates that it is becoming established as a key resource for practitioners. Knowledge translation and dissemination --------------------------------------- A key function of the CO-OPS Collaboration is to facilitate knowledge translation and exchange among members of the network. Specific activities include identifying and addressing knowledge gaps, translation of research evidence into practice-relevant knowledge and evidence-based recommendations, sharing of lessons learned in practice between initiatives, and collating and disseminating both research- and practice-based evidence. Available evidence suggests that the most effective knowledge translation occurs when recipients feel they are competent and confident in the use of evidence and *a priori*learnings, accompanied with easily accessible information, key messages, a facilitated network and access to a knowledge broker role. This is the approach that the CO-OPS Collaboration takes by having a knowledge broker as a core part of the Secretariat. Additional strategies used by CO-OPS to encourage successful translation of knowledge include: ▪ working with practitioners working in CBOPIs to identify knowledge needs, ▪ sourcing and developing resources, evidence summaries and other knowledge resources for the sites, and ▪ \'translating\' the knowledge for CO-OPS sites and a wider audience through a variety of approaches including websites, emails, newsletters, training programs and site visits, As a national collaboration in a geographically very large country, internet-based resources and services are a key component of CO-OPS\' communication with the network. The CO-OPS website <http://www.co-ops.net.au> includes electronic access to all new resources developed for CO-OPS (including those described above), a number of discussion forums, a searchable database of existing CBOPI in Australia and a searchable resource library providing access to a wealth of relevant information and resources for CBOPI, including evaluation tools and protocols, project reports and systematic reviews of the evidence. The website is the practical repository for knowledge received and created by CO-OPS members and project staff. Contributions to these resources and knowledge come from members in the form of evaluation tools, program designs, academic journals, training presentations and external sources. As a network it is imperative that CO-OPS codify and disseminate knowledge of relevance to its members to ensure it retains its position of importance in members work and professional estimations. CO-OPS achieves this by using data such as web-statistics to monitor what internet resources are accessed; to assess which parts of its electronic newsletter are most visited; through discussion with members on their training needs; through its governance structure of an Advisory and Steering Committee; and tracking requests for assistance. Evidence from these data sources indicates that resources are well used and highly accessed, particularly those that relate to case studies of best practice (typically \>120 downloads per individual case study and over 400 downloads and 700 hard copies for the collection of case studies collated from the 2009 National Workshop) and evidence syntheses (\>570 downloads in addition to \>600 hard copies distributed). Since its establishment, members of the CO-OPS secretariat have conducted 26 professional development sessions across Australia, for more than 650 practitioners, in addition to over 70 other presentations to a variety of national and international audiences. There has been high demand for these activities, with sessions regularly being over-subscribed well in advance. In addition, it appears from the available data that rates of access and use of the online (website) resources are heavily dependent upon active knowledge translation activities, including the dissemination of newsletters highlighting available evidence and resources, but particularly the networking and professional development events run by CO-OPS. These events, which often introduce available resources and guide members in how to access and use them, appear to facilitate members to engage in a more in-depth manner with resources and evidence. The levels of website usage shown in figure [1](#F1){ref-type="fig"} reflect key events, particularly the national workshops held in October 2009 and 2010 and a series of state-based workshops around February to March and May to June 2010 (the spike in August 2010 resulted from a variety of online promotion activities and presence of links to CO-OPS in the communications of other professional networks). As CO-OPS grows in member numbers, available resources and stature it is becoming and needs to continue to become increasingly sophisticated in how it reaches and works with members to steward and disseminate knowledge. CO-OPS is enhancing the meaning and quality of its interactions with members through increasing its modes of assistance, depth of discussion and member case studies. This is achieved by working with members and non-members to identify the specific training required to better tailor the information, case studies and type of training it delivers. CO-OPS is also moving to expand the availability of its training and the discussions within these sessions to a broader audience and to continue the discussions beyond traditional face-to-face training towards electronic options including web forums. These expanding, more inclusive and increasingly tailored opportunities to steward and disseminate knowledge will see CO-OPS develop as a network and progress through the various stages of development of a community of practice. Discussion ========== The establishment and development of the CO-OPS Collaboration provides an important model for support and knowledge exchange for community-based obesity prevention. Emerging evidence from successful community-based efforts to prevent obesity suggests that a significant level of centralised support is necessary, especially in relation to evaluation and evidence. The widely adopted EPODE model of intervention, which originated in France, began with a funding structure allocating 50% of funds to interventions and 50% to central support and evaluation. Currently EPODE costs around €2 per capita from the local government and 20c is used centrally (10%) for evaluation (personal communication). Australian demonstration programs in community-based obesity prevention have also shown the importance of strong support and central infrastructure. The CO-OPS Collaboration further expands this notion to bring together obesity prevention initiatives intervening at the community level from across the country, facilitating knowledge exchange and translation between initiatives of varying size, scope and focus. This ensures that lessons learned are shared and translated to other relevant contexts as well as providing support to practitioners who may be intervening in novel ways, with less common target groups or who may be geographically isolated. A number of research-based networks have been established internationally to bring together expertise on obesity prevention, especially for children. There are fewer examples for models of collaboration to build a network of professionals in practice, policy and research. Notable recent exceptions include the Canadian Obesity Network <http://www.obesitynetwork.ca/> and the Canadian Partnership Against Cancer Coalitions Linking Action and Science for Prevention <http://www.partnershipagainstcancer.ca/priorities/primary-prevention/strategic-initiatives/coalitions>. The expertise in implementing community-based initiatives which is held in the non-academic sector is vast and valuable. A strength of the CO-OPS Collaboration is this breadth of stakeholder representation, the model of learning and exchange between academic and non academic sectors, and the recognition of a variety of experiences and types of evidence and information. Our experience in establishing the network and consulting widely with a range of stakeholders has provided a number of important lessons about the needs of practitioners working in CBOPI. Chief among these is the need for professional networks and for support in planning and conducting program evaluation and using and contributing to current research. Key success factors in the process of setting up the collaboration have been the political, multi-level government policy, and commonwealth funding support and the overarching governance arrangements, each of which have helped ensure a systems integrated approach. Operationally, this resulted in the establishment of a secretariat, a steering committee, an advisory committee and defining the form and function of the network through consultation with stakeholders. The input of key stakeholders from government, research and practice in the steering and advisory committees has been crucial to guiding the direction of CO-OPS secretariat activities and developing the network. Another key factor in successful developments to date has been a secretariat and partner university collaborators with the skills and capacity to provide advice and support to network members in knowledge translation, program evaluation and training and supporting networking and training. Accommodating the diversity of CBOPI across in Australia, from very large, well funded programs with academic support through to small community-health led activities in single communities has proved a major challenge. The geographical size (and attendant remoteness of some areas) of Australia and the cultural diversity of communities provide a second challenge. The CO-OPS secretariat have endeavoured to overcome this in part by regularly travelling to each state and territory, including outside of capital cities, for consultation, networking activities and professional development, and in part by ensuring that as many of the resources and support services as possible are available online. The CO-OPS Collaboration is now at the forefront of community-based obesity prevention in Australia, bringing together research, practice and policy. As the policy and practice environment continues to evolve, it will be critical to ensure that CO-OPS retains its relevance and adapts to new challenges, opportunities and initiatives. The greater the critical mass of action in community-level obesity prevention grows, the greater the need for a central support and knowledge translation network will become. Significant gaps remain in knowledge and evidence related to CBOPI. An important future function of CO-OPS will be to facilitate further evidence generation and translation, including combining findings from a range of initiatives to draw broader conclusions about process, impacts and outcomes of CBOPI. Given the global significance of the obesity epidemic, the natural next stage to share experiences and develop the evidence will be to build international networks along the CO-OPS model, bringing together stakeholders in research, practice and policy. Conclusion ========== Community-based interventions are a promising approach and an important component of a comprehensive response to obesity. The establishment of the CO-OPS Collaboration is a significant step toward strengthening action in this area, by bringing together research, practice and policy to promote best practice, high quality evaluation and knowledge translation and exchange. The lessons learned from the CO-OPS Collaboration provide valuable insight into the development of national collaborative efforts and should be taken into consideration when establishing similar national partnerships and in combined international community based efforts to prevent obesity. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= SA, BS, MN, CF were involved in all activities reported within the paper. EW, RA led the evidence summary work presented in this paper. LK and TG led the work described under best practice principles. SA, MN, CF, RR led the writing of initial drafts of this paper. All authors have been involved in drafting the manuscript or revising it critically for important intellectual content. All authors have given final approval of the version to be published Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/132/prepub>

PubMed Central

2024-06-05T04:04:19.544815

2011-2-24

PMC3053246

Background ========== China is the largest producer and consumer of tobacco in the world. The prevalence of cigarette smoking is above 60% for men aged 15 and above, and 50% of women in this age group and adolescent are passive smokers \[[@B1]\]. Tobacco smoking was estimated as being responsible for approximately 67,300 premature deaths in Chinese adults aged 40 years and over in 2005 \[[@B2]\]. The premature deaths, productivity losses, and a substantial number of preventable diseases and health care costs related to smoking have resulted in a significant economic burden in China \[[@B3]-[@B7]\]. However, few people are aware of the harm caused by smoking and passive smoking \[[@B8]\]. Instead, cigarettes are considered a good vehicle for communication, and a popular gift to relatives or friends in China, especially for holidays. The World Health Organization Framework Convention on Tobacco Control (WHO FCTC), the world\'s first public health treaty, calls for warning labels to be displayed as large and clear health warnings covering 30% to 50% of the package in the form of pictures, pictograms or text. Every person should be informed of the health consequences, addictive nature, and mortal threat posed by tobacco use and exposure to tobacco smoke. This can be achieved by the warning labels on cigarette packages \[[@B9]\]. In 1991, the Chinese Congress enacted legislation requiring cigarette warnings to state \'smoking is harmful to your health\' in Chinese. This warning appeared on one of the side panels of every cigarette package \[[@B10]\]. On 9^th^of January 2006, WHO FCTC was ratified in China. In 2008, China implemented new regulations according to the FCTC and the legislation. Cigarette warnings were moved from the side panels and covered at least 30% of the front and back of the pack, with Chinese on the front, and English on the back. The warnings include two rotating sets of text in Chinese and English. One set states \'smoking is harmful to your health\', and \'quitting smoking reduces health risk\'. Another states \'smoking is harmful to your health\', and \'quitting smoking early is good for your health\' \[[@B10]\]. The purpose of the current study was primarily to compare the difference in reactions to different types of warning labels on cigarette packages with a specific focus on whether the new warning label is better than the previous one and labels used abroad. Methods ======= Study sites ----------- The study was conducted in 2008 in Nantong and Zhangjiagang cities, Jiangsu Province, one of the economically booming areas in Eastern part of China. Nantong is an urban city, and is a moderate developed city in the Province. Zhangjiagang is a rural area, and belongs to Suzhou City, a highly developed city in the Province. Participants ------------ Eligible study participants included in this survey were those aged 18 years and over and those working in hospitals, schools, bus/train stations, government offices, restaurants and bars. Altogether 1000 adults were approached, and 876 participants agreed to participate and finished the survey. All participants were asked to complete a face-to-face interview using a standard questionnaire; informed consent was sought prior to the interview being undertaken. The study was approved by the ethical board of Jiangsu Provincial Centre for Disease Control and Prevention. Verbal consent was obtained from each participant. Smoking status and demographic variables ---------------------------------------- Information was obtained from all participants about their smoking status. Smoking status was measured by asking if participants had ever smoked. Participants were grouped into two categories smokers and former smokers. Smokers were defined as those having smoked at least 100 cigarettes in their lifetime, and those having smoked at least one cigarette per day at the time of the survey. Former smokers were defined as individuals who had quit smoking at least one month prior to the survey and those who smoked at least one cigarette per day prior to quitting. Participants also reported their gender, age, and education level. Warning labels of cigarette packages ------------------------------------ Six warning labels were included in the interview questionnaire (Additional file [1](#S1){ref-type="supplementary-material"}). They were coded A-F. Label A was the old Chinese warning label, with \'smoking is harmful to your health\' written in Chinese on one of the side panels of the pack (Figure [1](#F1){ref-type="fig"}). Label B was the new Chinese warning label, with \'smoking is harmful to your health\', and \'quitting smoking reduces health risk\' written on the front and back faces of the package, in Chinese and English respectively. Label C was a famous Chinese brand which is exported abroad and includes text, pictorials, and quitline information on the whole front face and 1/3 back face, respectively. The text on the pack reads \'smoking damages your blood vessels, which can prevent blood circulation, particularly to your legs or feet. This can result in blood clots, infection, gangrene, even amputation.\' The other labels were foreign brands. Label D warned that cigarette smoking can result in mouth and oropharynx cancers by using text and pictures on 50% of the cigarette package. Label E, also used text and pictures to show that smoking when pregnant harms your baby. Label F indicated that smoking can lead to laryngeal cancer using both text and pictorials on 50% of the cigarette package. All English health warnings were translated into Chinese during the interview. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Six cigarette warning labels**. Label A was the old Chinese warning label, with \'smoking is harmful to your health\' in Chinese on one of the side panels of the pack. Label B was the new Chinese warning label, with \'smoking is harmful to your health\', and \'quit smoking reduces health risk\' on the front and back faces of the package, in Chinese and English respectively. Label C was a famous Chinese brand exported abroad, with text, pictorials, and quitline on the whole front face and 1/3 back face, respectively. The text said \'smoking damages your blood vessels, which can prevent blood circulation, particularly to your legs or feet. This can result in blood clots, infection, gangrene, even amputation.\' The other labels were foreign brands. Label D warned that cigarette smoking can result in mouth and oropharynx cancers in text and pictures on 50% of cigarette package. Label E showed smoking when pregnant harms your baby in large area with text and pictures. Label F indicated that smoking can lead to laryngeal cancer in text and pictorial on 50% of cigarette package. :::  ::: The harm warning provided by the label -------------------------------------- Making reference to Labels A to F, participants were asked if the each label gave them clear information on the harm which cigarette smoking can have on health and the specific diseases that occur related to cigarette smoking. Participants were also asked if Labels C, D, E, and F gave them clear information on specific diseases smoking can cause (as described above). The perceived impact of giving cigarettes as a gift --------------------------------------------------- Three questions on the perceived impact of giving cigarettes as a gift were presented. These included: 1) If you want to use cigarettes as a gift, do the following cigarette labels (A-F) make you change your mind and not do so? 2) If you want to give cigarettes as a gift, which warning label is least likely to stop you using cigarettes as a gift? 3) If you want to give cigarettes as a gift, which warning label is most likely to stop you using cigarettes as a gift? The perceived impact on the decision to quit smoking ---------------------------------------------------- Participants were asked three questions about the perceived impact of quitting smoking. These include: 1) If you were a cigarette smoker, would the following labels (A-F) make you want to quit smoking? 2) If you were a cigarette smoker, which warning label is most likely to cause you to quit? 3) If you were a cigarette smoker, which warning label is least likely to cause you to quit? Knowledge of the FCTC and its provision for cigarette packaging --------------------------------------------------------------- Participants were asked if they knew that China had ratified the WHO FCTC. If they answered yes, participants were then asked if they were aware of the FCTC provision that health warnings on cigarette packaging should be large, clear, visible and legible. Statistic analysis ------------------ Univariate and bivariate analyses were conducted to examine how much impact each of the different cigarette warning labels had and the knowledge of the FCTC by age groups, gender, education levels and smoking status. To compare the new Chinese label with international labels, Label C, D, E, and F were aggregated into one group. Chi-square tests were used to assess differences among groups where appropriate. All analyses were conducted using SPSS 13.0 (SPSS Inc., Chicago, IL, USA). Results ======= General information ------------------- Table [1](#T1){ref-type="table"} demonstrates the demographic characteristics of the sample. A total of 876 participants (374 male and 502 female) were involved in the study. The average age was 34.0 ± 11.0 years, a higher proportion of males reported that they were current smokers compared to females and 82.7% of participants had graduated from technical secondary school or higher. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Characteristics of the study sample ::: Male Female Total ------------------------ ------------------------------------- ----------- ----------- ----------- Number 374 502 876 Age groups (%) 20-29 127(34.0) 211(42.0) 338(38.6) 30-39 114(30.4) 179(35.7) 293(33.4) 40 and above 133(35.6) 112(22.3) 245(28.0) Educational levels (%) Low (High school and below) 75(20.1) 76(15.2) 151(17.3) Medium (technical secondary school) 28(7.5) 105(21.0) 133(15.2) High (Junior college and above) 270(72.4) 320(63.9) 590(67.5) Smoking status (%) Current smokers 152(40.6) 10(2.0) 162(18.5) Former smokers 42(11.2) 4(0.8) 46(5.3) Non-smokers 180(48.1) 488(97.2) 668(76.3) ::: The harm warning provided by the label -------------------------------------- Of the participants, 18.3% said Label A provided adequate information on the harm of cigarette smoking. Among them, 16.5% (14/85) of participants said Label A gave adequate information on the relationship between cigarette smoking and respiratory diseases, including lung cancer, and 16.5% and 3.5% respectively mentioned cancer and cardiovascular diseases. Overall, 31.2% said Label B gave adequate information on the harm of cigarette smoking. Among them, 36.6% said Label B provided adequate information on the relationship between cigarette smoking and respiratory diseases, 5.3% and 3.1% could identify the relationship of smoking with cancer and cardiovascular diseases respectively based on the information on label B. Similar percentage of participants said Label C-F gave adequate information on the harm of cigarette smoking, 90.5% for Label C, 92.7% for Label D, 92.4% for Label E and 92.7% for Label F. Compared to Label A, a higher proportion of participants said that Label B gave them clearer information on the harm of smoking across all subcategories, except those with low education level and current smokers. Labels C-F performed better than Label B in providing harm information for all sub-groups (Table [2](#T2){ref-type="table"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### The proportion of positive responses to the harm information provided by different cigarette labels by gender, age groups, educational levels and smoking status ::: n Label A Label B Label C Label D Label E Label F **P for Label B vs A**^†^ **P for Label B vs CDEF**^†‡^ -------------------- ---------------- ----- --------- --------- --------- --------- --------- --------- --------------------------- ------------------------------- Gender Male 374 18.7 31.0 88.2 89.8 88.8 88.8 \< 0.001 \< 0.001 Female 502 17.7 31.3 92.2 94.8 95.0 95.6 \< 0.001 \< 0.001 Age (yrs) 20- 338 16.3 32.0 90.8 94.1 93.2 93.8 \< 0.001 \< 0.001 30- 293 14.0 25.3 92.2 93.2 92.5 92.8 0.001 \< 0.001 40 and above 245 25.7 37.1 88.2 90.2 91.0 91.0 0.006 \< 0.001 Education levels\* Low 151 23.8 30.5 89.4 92.1 91.4 94.0 0.196 \< 0.001 Medium 133 15.8 27.1 88.0 90.2 94.0 92.5 0.025 \< 0.001 High 590 17.1 32.4 91.4 93.4 92.2 92.4 \< 0.001 \< 0.001 Smoking status Current smoker 162 18.5 23.9 80.9 82.7 83.3 81.5 0.222 \< 0.001 Former smoker 46 17.4 24.1 84.8 89.1 82.6 89.1 0.021 \< 0.001 Non-smoker 668 18.1 33.4 93.3 95.4 95.2 95.7 \< 0.001 \< 0.001 Total 876 18.3 31.2 90.5 92.7 92.4 92.7 \< 0.001 \< 0.001 A positive answer means participants can understand the harm information provided by the label. \*low, medium and high education refers to high school and below, technical secondary school, college and above, respectively. ^†^By chi-square test. ^‡^Combination of labels C-F (positive answer to all labels were regarded as positive). ::: The perceived impact of giving cigarettes as a gift --------------------------------------------------- Of the participants, 20.8% and 25.2% reported that they would not give cigarettes as a gift to somebody with Labels A and B (respectively) on the package. Over 80% of participants refused to give cigarettes as a gift if the package displayed warning Labels C-F. The proportion of those who would not give cigarettes as gift was higher among female, those who had never smoked and those having a higher educational level. Generally, there was no difference between the sub-groups, in terms of those who would not give cigarettes as a gift, for Label A and Label B, except the proportions were marginally higher among non-smokers and those aged between 30-40 for Label B. When comparing Label B to the combined labels, the proportion of respondents who would not give cigarettes as a gift was higher if any of Labels C-F were on the package (Table [3](#T3){ref-type="table"}). The majority of participants (70.4%) considered that Label A was least likely to stop them using cigarettes as a gift, and the proportion was 20.2% for Label B. Almost half of participants (46.8%) considered that Label C was most likely to stop them using cigarettes as a gift, and the proportion was 5.7% and 3.4% for Label A and Label B respectively. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### The perceived impact of not giving cigarette as a gift by gender, age groups, education levels and smoking status ::: n Label A Label B Label C Label D Label E Label F **P for Label B vs A**^†^ **P for Label B vs CDEF**^†‡^ ----------------- ---------------- ----- --------- --------- --------- --------- --------- --------- --------------------------- ------------------------------- Gender Male 374 18.4 21.7 80.5 80.5 79.1 81.0 0.273 \< 0.001 Female 502 22.5 27.9 90.2 90.2 89.0 89.6 0.05 \< 0.001 Age (yrs) 20- 338 22.5 26.3 87.9 87.9 85.5 87.3 0.244 \< 0.001 30- 293 13.7 19.8 90.1 90.1 89.4 89.8 0.046 \< 0.001 40 and above 245 26.9 30.2 78.8 78.8 78.4 79.6 0.424 \< 0.001 Education level Low 151 23.2 29.8 76.8 76.8 75.5 75.5 0.192 \< 0.001 Medium 133 22.6 24.8 89.5 89.5 85.0 86.5 0.665 \< 0.001 High 590 19.7 24.1 87.6 87.6 87.1 88.5 0.067 \< 0.001 Smoking status Current smoker 162 17.3 21.0 79.6 79.6 75.3 77.8 0.397 \< 0.001 Former smoker 46 22.0 19.6 71.7 71.7 71.7 73.9 0.582 \< 0.001 Non-smoker 668 15.2 26.6 88.6 88.6 88.0 88.8 0.048 \< 0.001 Total 876 20.8 25.2 86.1 86.1 84.8 86.0 0.027 \< 0.001 \*low, medium and high education refers to high school and below, technical secondary school, college and above, respectively. ^†^By chi-square test. ^‡^Combination of labels C-F (positive answer to all labels were regarded as positive). ::: The perceived impact on the decision to quit smoking ---------------------------------------------------- There were 26.8% and 31.5% of the participants who reported thinking about quitting due to warning Label A and Label B, respectively. In addition, the proportions were all above 80% for Labels C-F. We asked non-smokers if they were smokers, if the labels would impact on a decision to quit smoking. Non-smokers were more likely to quit smoking due to Label C-F, compared to those who were smokers. It was shown that due to the warning on Label B, those more likely to quit were females, those with higher educational level and non-smokers when compared to Label A. Label B was less likely to make the participants quit smoking compared to Labels C-F combined (Table [4](#T4){ref-type="table"}). Almost half of participants (43.3%) considered that Label C was most likely to cause them to quit. The proportion was only 4.5% and 3.7% for Label A and B, respectively. The majority of participants (69.9%) considered that Label A was least likely to cause them to quit, and the proportion was 20.2% for Label B. There was no significant difference between smoking status groups in terms of the impact Label A and Label B had on a decision on quit smoking. ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### The perceived impact on the decision to quit smoking by gender, age groups education levels and smoking status ::: n Label A Label B Label C Label D Label E Label F **P for Label B vs A**^†^ **P for Label B vs CDEF**^†‡^ ------------------- ---------------- ----- --------- --------- --------- --------- --------- --------- --------------------------- ------------------------------- Gender Male 374 27.0 28.6 81.3 80.7 81.0 82.6 0.624 \< 0.001 Female 502 26.7 33.7 91.6 91.6 93.0 93.2 0.016 \< 0.001 Age group 20- 338 33.1 33.7 88.8 87.6 89.1 89.1 0.870 \< 0.001 30- 293 20.5 26.6 88.4 87.7 88.4 90.4 0.080 \< 0.001 40 and above 245 25.7 34.3 83.7 85.3 85.7 86.1 0.038 \< 0.001 Education level\* Low 151 28.5 34.4 86.1 84.8 84.8 85.4 0.265 \< 0.001 Medium 133 24.8 30.1 88.0 84.2 90.2 91.0 0.336 \< 0.001 High 590 26.8 31.0 87.3 88.1 88.1 89.0 0.108 \< 0.001 Smoking status Current smoker 162 24.7 25.9 76.5 75.9 74.7 79.0 0.798 \< 0.001 Former smoker 46 28.3 26.1 84.8 80.4 80.4 78.3 0.115 \< 0.001 Never smoker 668 13.0 33.2 90.0 90.1 91.6 91.8 0.050 \< 0.001 Total 876 26.8 31.5 87.2 87.0 87.9 88.7 0.031 \< 0.001 \*low, medium and high education refers to high school and below, technical secondary school, college and above, respectively. ^†^By chi-square test. ^‡^Combination of labels C-F (positive answer to all labels were regarded as positive). ::: Knowledge of the WHO FCTC and its provision for cigarette packaging ------------------------------------------------------------------- Overall 32.4% of participants knew of the FCTC. Among them, 77.1% and 72.4% were non-smokers and those with the highest educational level, respectively. Furthermore, 75.4% (214/284) knew that China have ratified the FCTC, and 77.5% knew the provision of the FCTC that health warnings on cigarette packaging should be large, clear, visible and legible. Discussion and Conclusions ========================== Our study has shown that both the old and new Chinese warning labels have a low effect on the participants\' knowledge of the harmful effects of smoking, on giving cigarettes as a gift, and quitting smoking. Labels used abroad were far more effective than the labels used in the Chinese market. Over 90% of the participants knew \'smoking was harmful to their health\', while the knowledge of smoking-related disease, such as cardiovascular diseases, stroke etc. was relatively low. The result is consistent with another report from six cities in China \[[@B11]\]. From our survey, neither the old Chinese label nor the new one is able to provide details of smoking-related disease to smokers or nonsmokers, although there was a difference in the level of information related to the harm of smoking provided by Labels A and B, which may be due to their distinct location on the new pack. The result of no difference in low educational groups between Label A and B suggested that only text warnings cannot provide useful information to poor literacy population. In addition, the text-only labels cannot provide health warnings to current smokers, and smokers were failed to take notice of the difference between the old and new labels, even they take out cigarettes from packages every day. Our survey showed that text-plus-graphic warning labels were more effective than text-only labels, which is also consistent with other reports \[[@B12]\]. Graphic warnings can clearly express the consequences of smoking, and they are especially useful for populations with poor literacy and difficulty understanding text-based warnings. Moreover, graphic warning labels appear to be an important source of information regarding health risks for non-smokers, which may lead to increased pressure to quit from members of a smokers\' social network \[[@B13]\]. More and more countries have mandated the inclusion of graphic imagery on cigarette warning labels (e.g., Australia, Brazil, Canada, Chile, Singapore, Thailand, Uruguay, and Venezuela), with other countries soon to follow (e.g., Belgium and New Zealand) \[[@B14]\]. Warning labels can not only increase awareness of the health hazards, but also provide information on assistance for quitting and can promote interest in quitting. In our study, labels with detailed risk information and graphics had a more effective on the decision to quit. While both the old and the new Chinese labels had less effective with no information on specific smoking-related diseases, and no useful information on cessation. Canadian warning labels on cigarette packs are considered one of the most effective in the world, and are very useful for tobacco cessation. The requirements of the warning label with big, clear and direct health messages provides a strong incentive for smoking cessation \[[@B15]-[@B17]\]. Approximately one third of the smokers reported a likelihood of quitting and 20% of smokers reported smoking less, as a result of warning labels with graphic and detailed health risk and cessation information. Smokers were more likely to quit, make an attempt to quit, or reduce their smoking because of increased level of fear and disgust for the labels with text and large graphics \[[@B18]\]. Thus, graphic messages on warning labels appear more effective than text-only messages in promoting quitting \[[@B12],[@B14],[@B16],[@B19]\]. Recent surveys have also shown increased cessation activities due to newly introduced text-and-graphic warnings in countries such as Australia, the United Kingdom, and the United States \[[@B19]-[@B21]\]. As a traditional culture, cigarettes are usually considered a valued gift to give, especially on special days, such as Chinese Spring Festival and other holidays. Chinese cigarette packages are always designed with beautiful brand names and graphics, and with one sentence of text warning about the harm but without information related to specific smoking-related diseases. Beautiful designed packs and high prevalence of cigarette smoking in male make cigarettes popular for giving in social communication. Giving cigarette is giving harm. While, the current Chinese warning labels have limited effect on not giving cigarettes as a gift. Compared with foreign warning label requirements, both the old Chinese warning labels and the new ones are relatively weak. The impact will increase if a country changes from smaller to larger and more contrasting warnings \[[@B19]\]. A key limitation of this study concerns the use of a convenience sample which may not be representative of the Jiangsu population. However, purposive selection of groups in six types of work or public places in two cities enabled data collection from a wide range of population segments across a relatively small number of groups \[[@B14]\]. Another limitation was that the educational level of the participants was higher than the general population, thus this is not representative of the average education level of local residents. But, even in the population with higher educational levels, the proportion knowing the harm of smoking and WHO FCTC was not high. We estimate that the proportion is likely to be much lower in the general population. Dissemination of smoking-related knowledge needs be spread widely, especially in smokers and those with lower educational levels. As the first report in Jiangsu Province, our findings suggest that the new Chinese warning labels are still not effective for this target population. People do not receive sufficient information on the harm of smoking and smoking-related diseases from these labels. In addition, the new warning labels do not effectively increase the desire to quit, or prevent individuals from giving cigarettes to others. The findings from this study indicate that cigarette packaging may benefit from more noticeable, readable, believable and memorable warnings in line with the WHO FCTC and this may be an important policy element in reducing the attractiveness of smoking especially among young adults and teenagers. Warning labels should be part of a larger public health education effort. Abbreviations ============= (\'WHO FCTC\'): World Health Organization Framework Convention on Tobacco Control; Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= YQ contributed to the field work, data collection, quality control, analysis, and manuscript writing. MW, QX, and MZ contributed to the implementation in the field and gave advice on the manuscript writing. XP, JH, and ZG contributed to the field work, data collection and quality control. ZS contributed to the statistical advice and critical English review. All authors have read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/133/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Questionnaire for warning labels of cigarette**. As requested by the editor, the questionnaire used in the study was translated into English and presented for the readers. The questionnaire was designed by the China CDC-PUMC-JHSPH Project Group. Any use of it should be noticed to the Group and must be properly cited in any related research products. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We thank all the participants involved in the survey. The help of the health workers from Nantong, Zhangjiagang in recruiting subjects is very much acknowledged. We are grateful to the staff from Jiangsu Provincial, Nangtong and Zhangjiagang Cities Centre for Disease Control and Prevention. We thank Constance Kourbelis, Tiffany Gill, and Virginia Routley for help in proof reading. There was no conflict of interest among all authors. The study is part of the China CDC-PUMC-JHSPH Project \'towards to smoke-free China\', and financed by Bloomberg Foundation.

PubMed Central

2024-06-05T04:04:19.548473

2011-2-25

PMC3053247

Background ========== With the ageing population and the shortage of acute hospital beds, the cost of unplanned readmission (UR) among elderly patients to health service is high. Hospital readmission in 2004 accounted for about 17% of the total Medicare\'s hospital bill in the United States \[[@B1]\]. UR is also associated with high levels of psychological distress and frustration in patients and their relatives. In addition, repeated hospital admissions have been suggested to lead to dependency and result in a self-perpetuating cycle \[[@B2]\]. Readmission to hospital has often been considered as an indicator of quality of hospital care. It can be defined as admission to the same hospital within a specified period time since discharge from a hospital. Various time period such as 7, 14, 15, 28, 30 days after discharge have been reported in the literature \[[@B3]-[@B6]\]. Previous studies have suggested that increased UR was associated with frailty, severe disability \[[@B3]\], age, lack of a family physician, reduced social support, heart disease, diabetes \[[@B7]\], tumor, dementia \[[@B8]\], anemia \[[@B9],[@B10]\], falls, urinary incontinence, less supportive living arrangement \[[@B11]\], length of prior admission \[[@B8],[@B12]\], numbers of medical problems, numbers of previous admission \[[@B13]\], activities of daily living \[[@B14],[@B15]\], Bathel Index, Mini-Mental State Examination, and the triceps skin fold thickness \[[@B8]\]. Although some the above factors are caused by patients\' frailty and progression of their chronic diseases, it has been estimated that up to 48% of all readmissions have been judged to be preventable by measures such as better predischarge assessment, patient education, and postdischarge care \[[@B16]\]. Therefore, by identifying factors associated with UR, particularly those that can be obtained non-invasively on admission, geriatric medical units can allocate available resources to allow a more effective management of any high-risk UR patients. Frequency domain heart rate variability (HRV) analysis has been used as a non-invasive tool to assess the status of cardiac autonomic functions. The standards of measurement, physiologic interpretations, and clinical applications of HRV analyses have well been documented since 1996 \[[@B17]\]. Recent studies indicated that HRV reflects the dynamic operations of neural regulation of the heart. It can serve as signal markers for physiological or pathological events in healthy vegetarian women \[[@B18]\], in patients with brain death \[[@B19]\] or with terminal hepatic cancers \[[@B20]\], and those in intensive care unit \[[@B21]\]. Therefore, we conducted a prospective study to explore whether cardiac autonomic functions could be used as a predictor for UR in geriatric patients. The objective of this study is to investigate the association between frequency domain indices of heart rate variability on admission and 14-day UR in geriatric patients. Methods ======= The study was approved by the Human Research Ethics Committee of the Buddhist Dalin Tzu Chi General Hospital and all subjects gave informed consent. Patients admitted to the geriatric ward at the Buddhist Dalin Tzu Chi General Hospital, Chiayi, Taiwan during the period from July 2006 to June 2007 were eligible for inclusion the study. In this study, UR was defined as readmission to the hospital within 14 days after being discharged from an earlier hospital stay. A readmission period of 14 days was chosen in this study over other durations because it is the standard duration used by the Taiwan\'s Bureau of National Health Insurance for monitoring the quality of hospitalization cases \[[@B22]\]. Patients with invasive tubes, such as tracheostomy tube, Foley\'s tube, and naso-gastric tube; heavy smokers; heavy alcohol drinkers; or taking medications that might influence HRV, such as alpha-blockers or beta-blockers antihypertensives were excluded from the study. Patients who were discharged from the study hospital between 15 and 30 days were excluded from the study to act as a wash-out period in order to minimize the potential influences from their previous hospital stay. In addition, patients who had multiple 14-day readmissions during the study period were excluded. Therefore, each patient can only be considered as either an index admission or a readmission, but not both, in the data analysis. Demographic characteristics, biochemical test results, and clinical symptoms and signs of the patients were collected by a team comprised of physicians and nurses within 24 hours of admission. Biochemical test variables included white blood cell (WBC) counts, blood urea nitrogen (BUN), blood glucose, and serum glutamic oxaloacetic transaminase (GOT). Clinical symptoms and signs included blood pressure, heart rhythm, Glasgow Coma Scale scores, and activity levels of the patients. Five-minute electrocardiography (ECG) recordings were performed on admission between 14:00 and 16:00 hours with patients in supine position. ECG signals were taken by precordial leads and were recorded using a 12 bit analog-digital converter (PCL-818, Advantech, Taiwan) with a sampling rate of 1024 Hz. Each QRS complex was identified. The R point of each valid QRS complex was defined as the time point of each heart beat, and the interval between two R points (R-R interval) was estimated as the interval between the current and latter R points (PPI). Frequency domain analysis of PPI was performed using the nonparametric method of fast Fourier transform (FFT). For each time segment (288 seconds, 2048 data points), the algorithm estimated the power spectral density on the basis of FFT. The resulting spectrum was corrected for attenuation resulting from the sampling and Hamming window. The power spectrum was subsequently quantified by integration into frequency domain indices including total spectrum power (TP), high frequency power (HF, 0.15-0.4 Hz), low frequency power (LF, 0.04-0.15 Hz), and the ratio of lower frequency power to high frequency power (LF/HF ratio). The power content of the HF component corresponds to respiratory sinus arrhythmia and is modulated solely by the parasympathetic nervous system. The power content in the LF component is modulated jointly by the sympathetic and parasympathetic nervous systems. The LF/HF ratio is used to reflect sympathovagal balance. The power of the very low frequency component (0-0.04 Hz) was not calculated because it was reported to be unreliable over short recording periods \[[@B17]\]. Data entry and analysis were performed using commercially available software (SAS, version 9.1.3, SAS Institute Inc., Cary, NC, USA). All statistical assessments were two-sided and statistical significance was set at the 0.05 level. The basic characteristics between UR and non-UR patients were analyzed using nonparametric Mann-Whitney test for continuous variables or Fisher\'s exact test for categorical variables. The frequency domain HRV indices were natural-logarithm transformed to reduce skewness of the data distribution \[[@B18]\]. Multiple logistic regression analyses were conducted to analyze the association between the frequency domain HRV indices and UR adjusted for age and length of hospitalization. Hosmer and Lemeshow Goodness-of-Fit Test was conducted to determine the adequacy of the fitted logistic models. Results ======= A total of 131 patients were admitted to the geriatric ward of the study hospital during the period from July 3, 2006 to June 30, 2007. Fifty-three patients were excluded from the study and they included 30 patients who were discharged from the hospital within last 15 to 30 days, 6 patients who died during admission course, 14 patients who were lost to follow up, and 3 patients who were on alpha or beta blocker\'s antihypertensives. Therefore, 78 patients were included in the analysis. Within the study period, 19.2% of the patients (15 of 78) were readmitted within 14 days after discharge from their index admission. The basic characteristics of the patients are shown in Table [1](#T1){ref-type="table"}. There were no significant differences in sex, age, lengths of hospitalization, proportions with hypertension and diabetes, distributions of diagnosis on admission, activity levels, Glasgow Coma Scale scores, blood urea nitrogen levels, glucose levels, and serum GOT levels between UR and non-UR patients. Counts in WBC before discharge was significantly higher (p = 0.027) in UR patients than those in non-UR patients. The two most frequently diagnoses at index admission were pneumonia and urinary tract infection. Regarding their activity levels in the hospital, most patients (82.1%) were bedridden. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### The basic characteristics of study subjects ::: ---------------------------------------------------------------------------------------------------------------------------------- Variable Unplanned\ Non-unplanned readmission\ *p-*value Readmission\ *n*= 63\ *n*= 15\ n (%) n (%) ---------------------------------- ---------------------------------- ------------------- ---------------------------- ----------- Sex male 5 (33.3) 28 (44.4) 0.434 female 10 (66.7) 35 (55.6) Age (years) 80 (72, 90)\* 81 (73, 86) 0.929 Length of hospitalization (days) 10 (6, 16) 9 (7, 13) 0.541 Hypertension 9 (60) 41 (65) 0.713 Diabetes 7 (47) 19 (30) 0.223 Diagnosis at index admission pneumonia 9 (60) 25 (40) 0.154 urinary tract infection 8 (53) 44 (70) 0.223 electrolyte imbalance 3 (20) 9 (14) 0.582 gastro-intestinal tract bleeding 1 (7) 2 (3) 0.527 cerebral vascular accident 2 (13) 8 (13) 0.947 Activity level 0.288 can walk in hospital room 1 (7) 3 (5) on bed and can change position 0 9 (14) bedridden 14 (93) 50 (79) Glasgow Coma Scale 12 (7, 14) 10 (8, 14) 0.880 WBC before discharge (×10^3^/μl) 8.8 (6.7, 10.0) 6.8 (5.6, 8.5) 0.027 Blood urea nitrogen (mg/dL) 25 (13, 32) 16.5 (13.0, 29.0) 0.584 Blood glucose (mg/dL) 142 (110, 239) 128 (105, 164) 0.302 Serum GOT (mg/dL) 20 (14, 23) 22 (16, 32) 0.212 ---------------------------------------------------------------------------------------------------------------------------------- WBC = white blood cell; GOT = glutamic oxaloacetic transaminase \* continuous variables were expressed as median (first quartile, third quartile) ::: Comparison of frequency domain HRV indices between UR and non-UR patients is showed in Table [2](#T2){ref-type="table"}. Levels of total power, LF, HF, and LF/HF were significantly higher in non-UR patients compared to UR patients. Table [3](#T3){ref-type="table"} showed the odds ratios of the associations between HRV indices and non-UR. Odd ratios greater than 1 mean that lower level of HRV indices were associated with increased risk of UR. The risk of UR was significantly higher in patients with lower levels of total power (OR = 1.39; 95% CI = 1.04-2.00), low frequency power (LF) (OR = 1.22; 95% CI = 1.03-1.49), high frequency power (HF) (OR = 1.27; 95% CI = 1.02-1.64), and lower ratios of low frequency power to high frequency power (LF/HF ratio) (OR = 1.96; 95% CI = 1.07-3.84). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Comparison of frequency domain heart rate variability indices between unplanned readmission and non-unplanned readmission in geriatric patients ::: --------------------------------------------------------------------------------------- Variable Unplanned\ Non-unplanned readmission\ *p*-value Readmission\ *n*= 63 *n*= 15 ------------------------ --------------------- ---------------------------- ----------- Total power, ln(ms^2^) 5.57 (4.07, 6.34) 6.80 (4.68, 8.58) 0.021 LF power, ln(ms^2^) 3.85 (0.45, 5.07) 5.54 (2.48, 7.04) 0.013 HF power, ln(ms^2^) 3.51 (1.17, 4.29) 4.72 (3.17, 6.03) 0.035 LF/HF ratio -0.43 (-0.71, 1.19) 0.76 (-0.23, 1.48) 0.025 --------------------------------------------------------------------------------------- LF = low frequency; HF = high frequency \* variables were expressed as median (first quartile, third quartile) ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Results of multiple logistic regression of frequency domain heart rate variability indices associated with non-unplanned readmission in geriatric patients ::: HRV indices Odds ratio 95% confident interval *p*-value ------------------------ ------------ ------------------------ ----------- Total power, ln(ms^2^) 1.39 1.04-2.00 0.046 LF power, ln(ms^2^) 1.22 1.03-1.49 0.032 HF power, ln(ms^2^) 1.27 1.02-1.64 0.046 LF/HF ratio 1.96 1.07-3.84 0.036 LF = low frequency; HF = high frequency \*All models were adjusted with age and length of hospitalization. \*P-values of Hosmer and Lemeshow Goodness-of-Fit Test: Total power = 0.622, LF = 0.078., HF = 0.727, and LF/HF ratio = 0.723. ::: Discussion ========== The present prospective study is the first study in evaluating the association between cardiac autonomic functions as measured by frequency domain HRV and the risk of UR in geriatric patients. Lower levels of TP, LF, HF, and LF/HF ratio on admission were significantly associated with increased risk of UR in geriatric patients. Using HRV indices for UR prediction has the advantages of accessibility, ease of standardization, and noninvasiveness. There were no previous studies available in the literature that explored the use of HRV indices in UR prediction in geriatric patients. At present, UR prediction depends mostly on screening of clinical factors, which subjects to variations associated with the experience and judgment of clinicians. On the other hand, the use of objective HRV indices can minimize the influences of individual variations. Frequency domain HRV analysis is a noninvasive tool for measuring cardiac neural regulation. It has been well established that the HF component of HRV is equivalent to the vagal regulation of the heart whereas the LF/HF ratio can be considered to reflect sympathetic modulations. The levels of LF and total power can reflect the strength of autonomic regulation on the heart \[[@B17]\]. Since the autonomic neural control may play an important role in regulating cerebral circulation \[[@B23]\], greater strength of cardiac autonomic functions imply a better cerebral circulation. This may explain the association between greater strength of cardiac autonomic functions, as reflected by the increased levels of TP and LF, and a lower risk of UR in the present study. Previous studies have shown that HRV could reflect the body\'s adaptability to a stressed physiologic state \[[@B24]\]. Therefore, it is plausible that HRV can also reflect the general body reserve of geriatric patients and in turn, their risk of UR. The exact mechanisms linking HRV and UR in geriatric patients warrant additional exploration. Regarding the laboratory data, the levels of WBC before discharge from the hospital was significantly higher in UR patients compared with those in non-UR patients. The increased in WBC counts might reflect a subclinical inflammatory status in the UR patients. Other biochemical data were not significantly different between the two groups of patients. This is in contrast to a previous study designed to examine the rate of UR within the most recent postoperative year for heart transplant patients. The authors reported that BUN and creatinine levels were significantly higher in readmitted patients than in patients who were not \[[@B25]\]. A few limitations in this study should be noted. First, we excluded patients who had multiple readmissions from the study and therefore, our results are applicable only to the case of single unplanned readmissions. Whether cardiac autonomic functions are associated with UR in patients with multiple readmissions will require further investigations. Second, our findings from a single regional hospital may not be representative for other geriatric wards with different sources of patients. Most of our patients came from nursing homes rather than from their own homes. Third, the loss of 14 patients to follow up raised the concern that their readmission might have occurred in other hospitals. However, these patients were residents of nursing homes which had contracts with the study hospital to provide health services to their residents. Therefore, readmission to other hospitals was unlikely to have happened. Conclusions =========== This is the first study to evaluate the association between cardiac autonomic functions as measured by frequency domain HRV and the risk of UR in geriatric patients. Lower TP, LF, HF, and LF/HF ratios on admission were significantly associated with increased risk of UR in geriatric patients. Additional studies are required to confirm the value and feasibility of using HRV indices on admission as a non-invasive tool to assist the prediction of UR in geriatric patients. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= JKC conceived the research questions, designed the study and drafted the initial manuscript. CHF was involved in data analysis and revisions. TBJK was involved in preparatory field works and data collection. MK was involved in the interpretation of data and revisions of the manuscript for publication. All authors revised, read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/137/prepub> Acknowledgements ================ This study was supported by grants from Buddhist Dalin Tzu Chi General Hospital (DTCRD: 96-18, DTCRD99(2)-I-04)

PubMed Central

2024-06-05T04:04:19.551590

2011-2-27

PMC3053248

Background ========== Socio-economic inequalities in mortality are well reported in Western European countries \[[@B1]-[@B5]\]. Excess coronary heart disease mortality \[[@B6],[@B7]\] as well as unnatural mortality, namely suicide \[[@B8]-[@B10]\], alcohol-related deaths \[[@B9],[@B11],[@B12]\], and accidental and violent causes of death \[[@B13]\], have been reported in lower socio-economic groups. Socio-economic variation is also significant in Finland in these specific causes of death \[[@B10],[@B11],[@B14],[@B15]\]. Over the past 40 years, the socio-economic inequalities in mortality have widened in several countries \[[@B16],[@B17]\]. In their search for new explanations for socio-economic disparities, scholars exploring the links between socio-economic position and health are moving beyond the material and behavioural factors, which do not fully account for these disparities. Psychological indicators, such as negative emotions (including depression and anxiety) \[[@B18]\], stress \[[@B19],[@B20]\] and insomnia \[[@B21]\], have been proposed as a plausible explanation for the socio-economic gradient in health. It is suggested that socio-economic differences in health are at least partly mediated by psychological distress stemming from socio-economic deprivation. Supposedly, not only absolute deprivation, but also relative deprivation, that is, one\'s position in the hierarchy vis-à-vis others, is important and associated with health \[[@B22]\]. In Williams\' \[[@B23]\] conceptual framework, psychosocial factors (consisting of risky health practices, social ties, perception of control, stress and affective states) are seen as critical mediators between social structure and health status. Marmot and Brunner \[[@B22]\] proposed a model in which social structure is linked to health and disease via material, psychosocial and behavioural pathways. These approaches view psychological distress not as the property of an individual but as the response of the individual to the external environment acting upon him or her. According to Schnittker \[[@B24]\], the resources provided by socio-economic position are related to the inferences individuals draw about the self, and these psychological states might affect physical and mental health. Wilkinson states that the psychological pain resulting from low social status affects patterns of violence, disrespect, shame, poor social relations and depression \[[@B25]\]. An association between lower socio-economic position and poor mental health \[[@B26]-[@B30]\] has been reported, as well as an excess mortality associated with psychological distress, especially with death from unnatural causes and cardiovascular disease \[[@B31]-[@B37]\]. The evidence regarding the potential mediating role of psychological factors on the relationship between socio-economic position and health is not unambiguous. Schnittker \[[@B24]\] examined whether four psychological factors (self-esteem, mastery, neuroticism and depressive symptoms) mediated the relationship between socio-economic position and three indicators of health (self-rated health, functional limitations and chronic conditions) and found only weak mediating effects. The results provided the strongest evidence for mediation in cases of neuroticism or depressive symptoms. Marmot et al. \[[@B38]\] found evidence for the mediation of psychological well-being measures (control/self-efficacy) on the association between education and health. Likewise, a Hungarian population study found that depressive symptom severity mediated between relative socio-economic deprivation and higher self-rated morbidity rates \[[@B36]\]. To our knowledge, few studies have examined the question of whether psychological factors contribute to socio-economic differences in cause-specific mortality. It has been proposed that psychological factors play a mediating role in the socio-economic differences associated with cardiovascular mortality \[[@B7],[@B39]\]. In a U.S. study \[[@B40]\], psychological distress as measured by hopelessness, depression and life dissatisfaction was not a significant contributor to socio-economic disparities in all-cause mortality. Van Oort et al. \[[@B41]\] found that, when independent of material factors, psychosocial factors contribute little to the explanation of educational inequalities in all-cause mortality. Aim of the study ---------------- The aim of this study was to explore whether self-reported psychological distress, measured by depression, stress and insomnia, mediates socio-economic differences (indicated by educational level, employment status and household income) in unnatural (suicide, accidents and violence, and alcohol-related mortality) and coronary heart disease (CHD) mortality in the 28-year follow-up (see Additional file [1](#S1){ref-type="supplementary-material"}). In other words, the aim of the study is to examine the contribution of psychological distress to relative differences in cause-specific mortality by socio-economic position. Methods ======= Data ---- The basic data source is the nationwide, repeated cross-sectional survey, \"Health Behaviour and Health among the Finnish Adult Population\", conducted annually since 1978 by the National Public Health Institute of Finland \[[@B42]\]. The survey questionnaire is mailed to a random sample of 5,000 Finns aged 15-64 years. The simple random sample was conducted by The Finnish Population Information System which is a computerized national register that contains basic on-line information about all Finnish citizens residing permanently in Finland. The survey years covered in this study are 1979-2002. The year 1985 has been excluded from the survey due to missing personal identification codes for that year. Respondents under 25 years of age have been excluded from this study because their socio-economic status is not established. We have supplemented the survey data with education and household income variables from Statistics Finland Register Data from the years 1979-2002 and the Finnish National Causes of Death Register follow-up data from the years 1979-2006. The mortality data include immediate, contributing and underlying causes of death, as well as the exact date of death. The data linkages were derived by using the personal identification codes assigned to all persons living permanently in Finland. After excluding the missing data on psychological distress variables (N = 1129, 1,6%), the total number of cases was 67871 (average annual response rate 73.5%), out of which 32451 were men (average response rate 69%) and 35420 were women (average response rate 78%). Our study is reviewed and supported by The Institutional Review Board of National Institute for Health and Welfare, (THL) (IRB 00007085, FWA 00014588). Psychological distress variables -------------------------------- We questioned the respondents about 14 health problems or symptoms, among them depression and insomnia, by the following single question: \"Have you had any of the following symptoms or health problems during past 30 days?\" (Yes, if so). Stress was addressed in a separate question on a four-point scale (1 = unbearable situation, 4 = no stress at all); respondents were asked if they had symptoms of tension or had been under great stress or considerable strain during the past 30 days. We considered unbearable stress as having the most negative effect on health and mortality and being associated with social disadvantage. Therefore, we classified those reporting an unbearable situation as having stress. We investigated the correlation of the psychological distress measures in another paper using this same data and showed that single-question depression (males r = .58; females r = .55) and insomnia (males and females r = .38) correlated with the general mental health inventory (MHI-5) \[[@B30]\]. In this study, self-reported psychological distress is thought to reflect the subjective experience of psychological well-being, and it is used to explore the role of psychological distress in generating socio-economic differences in mortality at an extensive population level \[[@B43]\]. Socio-economic variables ------------------------ Socio-economic variables included education and household income from the register data and employment status from the survey questionnaire. We collected the register data for education and income from 1980 statistics for the survey years 1979-1983, from 1985 statistics for the survey years 1984-1986 and annually from 1987 until 2000. For the survey years 2001-2002, we collected the socio-economic data from the year 2000. The educational level was derived from the Register of Educational Qualifications and Degrees, which follows, as far as possible, the principles and categories of the revised UNESCO International Standard Classification of Education 1997 (ISCED 1997). We divided the respondent\'s educational qualification into three categories: the lowest level included respondents with no education, an unknown education or with lower secondary education; the intermediate level included respondents with upper secondary or post-secondary non-tertiary education; and the highest level included respondents with tertiary education. Household income has been found to be more strongly and consistently associated with health than individual income. \[[@B44]\] We calculated household income as taxable total gross income for a household per year without transfer payment, divided by the consumption unit of the OECD equivalence scale. The first adult in the household was weighted as 1.0, other adults as 0.7 and children under 18 as 0.5 \[[@B45]\]. We further divided household income per consumption unit into tertiles by every study year, in order to keep the comparability of the variable over time. Employment status during most of the year consisted of the categories of employed and unemployed. Additional categories, that is housewives/husbands, students and retired people were excluded from the analyses concerning mortality differences by employment status. Mortality --------- In this study we analysed unnatural causes of death like suicide, accidents and violence, and alcohol-related mortality, and, for purposes of comparison, a general cause of death, coronary heart disease (CHD) mortality. We identified the causes of death using the International Classification of Diseases (ICD, WHO; 1974, 1978, 1992). The 8th revision was used for the years 1979-1986, the 9th revision for the years 1987-1995 and the 10th revision for the years subsequent to 1996. The classifications for suicide were E950-E959 (ICD-8 and ICD-9) and X60-X84 (ICD-10). Accidents and violence were ICD codes E800-E859, E861-E949, E960-E999 (ICD-8), E800-E849, E852-E949, E960-E999 (ICD-9) and V00-V99, W00-W99, X00-X44, X46-X59, X85-X99, Y00-Y89 (ICD-10). The definition of alcohol-related deaths included injuries, diseases and poisonings where alcohol was the main cause of the death (ICD-8 codes 291, 303, 571.0, 577, and E860; ICD-9 codes 291, 303, 357.5, 425.5, 535.3, 571.0-571.3, 577.0D-577.0F, 577.1C-577.1D and E851; ICD-10 codes F10, G31.2, G62.1, G72.1, I42.6, K29.2, K70, K86.0, O35.4 and X45). We grouped suicides, accidents and violence and alcohol-related deaths together and called them \'unnatural\' causes of death. The classification for coronary heart disease (CHD) mortality was 410-414 for ICD-8 and 9 codes and I20-I25 for ICD-10 codes. Statistical methods ------------------- For preliminary analyses, we examined associations between psychological distress and socio-economic position with a logistic regression model reporting odds ratios (OR) with 95% confidence intervals (CI) (see Additional file [2](#S2){ref-type="supplementary-material"}) and between psychological distress and mortality with a Cox proportional hazard model reporting hazard ratios (HR) with 95% confidence intervals (CI) (see Additional file [3](#S3){ref-type="supplementary-material"}) \[[@B46]\]. We conducted the main analyses using the Cox proportional hazard model (Tables [1](#T1){ref-type="table"}, [2](#T2){ref-type="table"}, [3](#T3){ref-type="table"}). All the analyses were performed with age and study year as covariates. Variation over time was taken into account by adjusting for the study year. To take into account the non-linear association of age with unnatural mortality, we adjusted mortality analyses for age squared. We carried out all the statistical analyses separately for men and women, using the statistical package SPSS 17.0 for Windows (SPSS Corporation 2008). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### The effect of adjusting for self-reported psychological distress on educational level differences in excess mortality. ::: Unnatural mortality CHD mortality ---------------------------------------- --------------------- ------------------ ----- ------------------ ------ -------------- ------------------ ----- ------------------ ----- Educational level High Intermediate Low High Intermediate Low Men HR (95% CI) (%) HR (95% CI) (%) HR (95% CI) (%) HR (95% CI) (%) 1: confounders\* 1 1.43 (1.15-1.77) 1.58 (1.28-1.94) 1 1.19 (0.98-1.43) 1.36 (1.16-1.59) 1+depression 1 1.40 (1.13-1.74) -7 1.53 (1.25-1.89) -9 1 1.18 (0.98-1.43) -5 1.36 (1.16-1.59) 0 1+stress 1 1.41 (1.14-1.75) -5 1.55 (1.26-1.91) -5 1 1.18 (0.98-1.43) -5 1.36 (1.16-1.59) 0 1+insomnia 1 1.41 (1.13-1.74) -5 1.53 (1.24-1.88) -9 1 1.19 (0.98-1.43) 0 1.36 (1.16-1.59) 0 1+all psychological distress variables 1 1.38 (1.11-1.72) -11 1.51 (1.23-1.85) -12 1 1.18 (0.98-1.43) -5 1.35 (1.15-1.58) -3 Women 1: confounders\* 1 1.29 (0.88-1.89) 1.16 (0.79-1.69) 1 1.54 (1.02-2.31) 2.32 (1.62-3.31) 1+depression 1 1.28 (0.87-1.87) -3 1.13 (0.77-1.65) -19 1 1.52 (1.01-2.28) -4 2.27 (1.59-3.24) -4 1+stress 1 1.28 (0.88-1.88) 0 1.13 (0.77-1.65) -19 1 1.53 (1.02-2.29) -2 2.29 (1.60-3.27) -3 1+insomnia 1 1.29 (0.88-1.90) 0 1.14 (0.78-1.67) -19 1 1.53 (1.02-2.30) -2 2.30 (1.61-3.28) -2 1+all psychological distress variables 1 1.27 (0.87-1.86) -7 1.11 (0.76-1.62) -31 1 1.51 (1.01-2.27) -6 2.25 (1.57-3.21) -5 Hazard ratios (95% CIs) and percent reduction (%) in mortality among those with an intermediate or low education compared to those with a high education after adjusting for psychological distress. \*The confounders: age, age squared, study year. ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### The effect of adjusting for self-reported psychological distress on employment status differences in excess mortality. ::: Unnatural mortality CHD mortality ---------------------------------------- --------------------- ------------------ --------------- ---------- ------------------ ----- Employment status Employed Unemployed Employed Unemployed Men HR (95% CI) (%) HR (95% CI) (%) 1: confounders\* 1 4.12 (3.26-5.21) 1 2.06 (1.58-2.68) 1+depression 1 3.50 (2.75-4.44) -20 1 2.00 (1.53-2.59) -6 1+stress 1 3.78 (2.98-4.79) -11 1 2.03 (1.56-2.64) -3 1+insomnia 1 3.45 (2.72-4.36) -21 1 2.01 (1.55-2.62) -5 1+all psychological distress variables 1 3.15 (2.47-4.00) -31 1 1.96 (1.50-2.56) -9 Women 1: confounders\* 1 3.50 (2.13-5.75) 1 1.91 (1.18-3.11) 1+depression 1 3.09 (1.87-5.11) -16 1 1.83 (1.12-3.00) -9 1+stress 1 3.22 (1.95-5.32) -11 1 1.91 (1.18-3.11) 0 1+insomnia 1 3.15 (1.91-5.19) -14 1 1.88 (1.16-3.07) -3 1+all psychological distress variables 1 2.86 (1.73-4.73) -26 1 1.83 (1.12-2.99) -9 Hazard ratios (95% CIs) and percent reduction (%) in mortality among the unemployed compared to the employed after adjusting for self-reported psychological distress. \*The confounders: age, age squared, study year ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### The effect of adjusting for self-reported psychological distress on household income differences in excess mortality. ::: Unnatural mortality CHD mortality ---------------------------------------- --------------------- ------------------ ----- ------------------ --------------- ------ ------------------ ----- ------------------ ----- Income High Intermediate (%) Low (%) High Intermediate (%) Low (%) Men HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) 1: confounders\* 1 1.19 (0.98-1.44) 1.70 (1.42-2.04) 1 1.37 (1.19-1.58) 1.58 (1.38-1.81) 1+depression 1 1.16 (0.95-1.40) -16 1.62 (1.35-1.94) -11 1 1.36 (1.18-1.57) -3 1.56 (1.36-1.79) -3 1+stress 1 1.18 (0.97-1.43) -5 1.65 (1.38-1.98) -7 1 1.37 (1.19-1.58) 0 1.57 (1.38-1.80) -2 1+insomnia 1 1.19 (0.98-1.45) 0 1.65 (1.38-1.98) -7 1 1.37 (1.19-1.58) 0 1.57 (1.38-1.80) -2 1+all psychological distress variables 1 1.18 (0.97-1.43) -5 1.59 (1.33-1.92) -16 1 1.37 (1.19-1.57) 0 1.56 (1.36-1.78) -3 Women 1: confounders\* 1 1.08 (0.75-1.55) 1.80 (1.29-2.50) 1 1.37 (1.07-1.75) 2.13 (1.71-2.65) 1+depression 1 1.07 (0.75-1.54) -13 1.73 (1.24-2.40) -9 1 1.36 (1.06-1.74) -3 2.09 (1.68-2.60) -4 1+stress 1 1.08 (0.75-1.54) 0 1.73 (1.24-2.40) -9 1 1.37 (1.07-1.75) 0 2.10 (1.69-2.62) -3 1+insomnia 1 1.09 (0.76-1.56) +13 1.77 (1.27-2.46) -4 1 1.37 (1.07-1.75) 0 2.11 (1.69-2.63) -2 1+all psychological distress variables 1 1.08 (0.75-1.55) 0 1.69 (1.21-2.35) -14 1 1.36 (1.06-1.74) -3 2.07 (1.66-2.58) -5 Hazard ratios (95% CIs) and percent of reduction (%) in mortality among intermediate or low household income levels compared to high income level after adjustments of self-reported psychological distress.\*The confounders: age, age squared, study year ::: In the main Cox proportional hazard analyses, we first carried out the base model to explore the relative differences in mortality outcomes by socio-economic variables adjusted for age, age squared and study year. In the following models, we adjusted for each of the psychological distress variables separately and, finally, for all of them simultaneously to see whether those variables contributed to the socio-economic disparities in mortality. To assess the impact of the adjustment of different variables on the base model hazard ratio, we calculated the percentage reduction of the HR as follows: \[(base model HR-base model plus other factors HR)/(base model HR-1)\] × 100 \[[@B7],[@B47]\]. We interpreted the reduction in the hazard ratio to tell how much of the association between the individual socio-economic variables and mortality was accounted for by the measures of psychological distress. Results ======= Table [4](#T4){ref-type="table"} describes the follow-up data by socio-economic position. The total number of deaths for unnatural causes was 716 for men and 222 for women, while the numbers for CHD mortality were 1,389 for men and 635 for women. Fourteen per cent of men and 18% of women reported depression, 18% of men and 19% of women reported insomnia, and 2.6% of men and 2.4% of women reported unbearable stress (not shown in the table). ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Description of the mortality follow-up data by socio-economic position among men and women ::: Men Women -------------- ------- -------------------------------- ------------------------------------ ------------------------------ --------------------------------- ------- -------------------------------- ------------------------------------ ------------------------------ --------------------------------- **N** **Person years at follow- up** **N deaths (unnatural mortality)** **N deaths (CHD mortality)** **Mean follow-up time (years)** **N** **Person years at follow- up** **N deaths (unnatural mortality)** **N deaths (CHD mortality)** **Mean follow-up time (years)** Education Highest 8296 125895 129 194 15.2 9623 139138 42 34 14.5 Intermediate 11036 169469 238 238 15.4 11875 182635 73 79 15.4 Lowest 13119 220049 349 957 16.8 13922 245434 107 522 17.6 Total 32451 515413 716 1389 15.9 35420 567207 222 635 16.0 Empl. status Employed 25749 428906 483 804 16.7 25075 408378 124 196 16.3 Unemployed 1718 21244 90 62 12.4 1619 20153 19 18 12.4 Total 27467 450150 573 866 16.4 26694 428531 143 214 16.1 Income Highest 11121 180305 197 368 16.2 11439 185284 57 110 16.2 Intermediate 10656 171646 215 422 16.1 11853 190711 61 151 16.1 Lowest 10321 158719 281 533 15.4 11915 187955 100 355 15.8 Total 32098 510670 693 1323 15.9 35207 563950 218 616 16.0 ::: The preliminary adjusted logistic regression analysis confirmed the associations between low socio-economic position and psychological distress for all indicators (see Additional file [2](#S2){ref-type="supplementary-material"}). The second preliminary analysis (see Additional file [3](#S3){ref-type="supplementary-material"}), based on the Cox proportional hazard model, demonstrated statistically significant hazard ratios for both unnatural and coronary heart disease mortality by psychological distress. Hazard ratios for psychological distress were higher for unnatural causes of death than for CHD mortality Contribution of psychological distress to educational differences in mortality ------------------------------------------------------------------------------ In the main Cox proportional hazard model analyses, we examined the contribution of the psychological distress variables to excess mortality by socio-economic position for each of the socio-economic variables separately (Tables [1](#T1){ref-type="table"}-3). The hazard ratios for educational level in Table [1](#T1){ref-type="table"} present the effect of adjusting for psychological distress variables on the relative differences by educational level in unnatural and CHD mortality among men and women. In the base model for men, we found excess unnatural mortality in the intermediate and lowest educational levels, and excess CHD mortality in the lowest level of education. However, adjusting for measures of psychological distress, when considered both separately and simultaneously, resulted in a very modest reduction in the relative mortality difference by educational level (5-12%) in unnatural causes of death, and no change in CHD mortality (0-5%). In women, the level of education was statistically significantly associated only with CHD mortality, where the contribution of psychological distress variables was equivalent to men. Contribution of psychological distress to employment status differences in mortality ------------------------------------------------------------------------------------ In the base model presented in Table [2](#T2){ref-type="table"}, unemployment was associated with increased mortality in both genders. For unnatural cause of death, adjusting for psychological distress variables separately and simultaneously accounted for 11-31% of the excess mortality in unemployed men and 11-26% in women. Adjusting for all of the measures of psychological distress combined resulted in further reductions to the excess risk of unnatural mortality among the unemployed. Adjusting for psychological distress attenuated the association between employment status and CHD mortality at the most 9% among both men and women. Contribution of psychological distress to household income level differences in mortality ----------------------------------------------------------------------------------------- In the base model for mortality by household income level, we found a higher risk of mortality in the lowest income group compared to the highest income group among men and women in both unnatural and CHD mortality (Table [3](#T3){ref-type="table"}). After controlling for the psychological distress variables separately or combined, psychological distress accounted for 4-16% of the differences in unnatural mortality among those at the lowest income level in both men and women. Again, the effect of adjustment for psychological distress measures in CHD mortality by income level appeared weak (0-5%). Discussion ========== Based on our results, we can conclude that psychological distress partly accounted for employment status and household income level differences in unnatural mortality (suicide, accidents and violence, and alcohol-related mortality) in both genders, and for educational level differences in unnatural mortality among men; among women no significant educational differences were found in unnatural mortality in the first place. The contribution of psychological distress variables to socio-economic differences in CHD mortality, on the other hand, was negligible. The strength of our study is the nationally representative data from repeated population surveys, which was supplemented with extensive socio-economic register data and national causes of death register data, providing for a prospective study design with a 28-year follow-up. However, the cross-sectional measure of socio-economic factors and psychological distress variables allows for no conclusions about the direction of the association, that is, health selection versus causation, which may both contribute to the associations between socio-economic position and psychological factors \[[@B18]\]. The response rate of the survey is similar to that of other population surveys \[[@B48]\]. However, in our non-respondent analysis of this data \[[@B49]\] we found lower response rates for the lower educated. Total and cause specific (for example, alcohol, external causes, suicide) excess mortality rates were higher among survey non-respondents and this is partly explained by educational and income differences between respondents and non-respondents \[[@B50]\]. These results indicate that non-respondents have more severe illnesses, mental health problems and depression as well as unhealthy lifestyles, such as smoking and alcohol use. They also indicate that the comparability of the results of the different socio-economic groups may be biased and, therefore, the socio-economic differences may actually be stronger than those observed in this data. Additional analyses for respondents with missing data on psychological distress variables (N = 1129, 1,6%), although containing relatively small number, showed that those with missing data on psychological distress measures were also more likely to be in the lower SES groups. One principal limitation of the study is that the measures of psychological distress are very simple self-reported single-item questions. These measures may cover a variety of transient or chronic psychological symptoms, a wide range of meanings from the temporary decrease of psychological well-being to deeply impaired, even life-threatening disorders. Therefore, the main focus of these indicators is not to detect clinical disorders but to reflect the subjective experience of mental health, and to study mental well-being at an extensive population level \[[@B43]\]. Nevertheless, single-item psychological distress variables demonstrated significant associations with cause-specific mortality, indicating that self-reported psychological distress have an implication for health. Another limitation concerning measures used in this study is the unemployed versus employed classification, which is a crude measure of employment status. In the previous studies psychological factors only weakly or moderately mediate the relationship between SES and all-cause mortality \[[@B40],[@B41]\]. In this study, we analysed three different measures of psychological distress and found some mediation for unnatural mortality and SES, and weak mediation for CHD mortality by employment status. It has been proposed that the excess CHD mortality among those in a lower socio-economic position is dependent on socio-economic differences in behavioural and biological risk factors, such as smoking, blood pressure and serum cholesterol levels \[[@B51]\]. A previous study based on the same data examined health behaviours as explanations for educational level differences in CHD mortality \[[@B47]\]. Health behaviours, most importantly smoking, physical activity and vegetable intake, explained about 50% of the educational differences in CHD mortality among men, but did not explain much of the differences among women. Compared to these results, psychological factors examined in the present study did not add to the contribution made by behavioural factors in explaining socio-economic differences in CHD mortality. However, psychological distress explaining some of the inequalities in suicide, accidents and violence, and alcohol-related mortality indicates that in these specific causes of death, poor mental health is related to more severe consequences in the lower socio-economic status groups than in the higher SES groups. It is possibly due to poor coping strategies of psychological distress in the lower SES. Obviously, that includes risky behaviour and, above all, heavy alcohol consumption which may be aimed at relieving psychological symptoms. Theories and models which propose psychosocial factors as mediators in the SES-health relationship also emphasize that health status is the result of complex causes. Health behaviour, socio-demographic factors and early environmental, genetic, biomedical and medical factors are all seen as related to this phenomenon. Our results suggest that psychological distress may explain some of the cause-specific mortality disparities between socio-economic groups. Conclusions =========== Psychological distress partly accounted for socio-economic disparities in unnatural mortality, but notably less for CHD mortality. Improvement of psychological well-being in lower socio-economic groups may reduce some of the socio-economic disparities in cause-specific mortality. Especially, the possible mental health problems of the unemployed should be taken into account when searching for a means to decrease these inequalities. Further studies are needed to explicate the role and mechanisms of psychological distress in generating socio-economic differences, particularly in cause-specific mortality. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= KTM processed the data, carried out the statistical analyses and drafted the manuscript. TKML, TMH, AIO were involved in interpreting the data and drafting the manuscript. TPM supervised the first author and was involved in interpreting the data and drafting the manuscript. RSP was involved in data management, coordinated the study, supervised the first author and was involved in drafting the manuscript. All authors revised the text critically for important intellectual content and read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/138/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Appendix figure S1**. Conceptual framework of the study. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Appendix table S1**. Logistic regression model (Odds Ratios, 95% Confidence Intervals) for psychological distress by socio-economic position. Males and females. Adjusted for age and study year. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Appendix table S2**. Adjusted Cox proportional hazard model for the unnatural and CHD mortality in self-reported psychological distress. Adjusted for age, age squared and study year. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Funding: This work was supported by the Academy of Finland \[grant number 214126\]. PhD Ari Haukkala\'s comments in the final stage of the manuscript are gratefully acknowledged.

PubMed Central

2024-06-05T04:04:19.554491

2011-2-28

PMC3053249

Background ========== Education and other measures of socioeconomic position, such as occupation and income, are consistently inversely associated with incidence of cardiovascular disease in developed countries \[[@B1],[@B2]\]. Elevated blood pressure, a major risk factor for cardiovascular disease, has been demonstrated in cross-sectional studies to be associated with low education and lower levels of other SEP measures \[[@B3]\]. Because of the limitations of cross-sectional studies, further investigation of whether educational attainment may be causally related to blood pressure can be achieved through prospective designs that measure longitudinal trajectories of blood pressure. Few studies have investigated longitudinal blood pressure trajectories, especially over a substantial proportion of the life course \[[@B4]-[@B7]\]. Furthermore, little is known about the effects of adjusting for potential explanatory/mediating mechanisms such as smoking, alcohol consumption, obesity, or use of antihypertensive medications \[[@B4]-[@B7]\]. The objectives of this study were to determine whether low educational attainment was associated with increased blood pressure from multiple longitudinal assessments over 30 years. Furthermore, we aimed to separate \'antecedant\' effects of education, and other related factors, that might have caused baseline differences in blood pressure, from potential long-term effects of education on post-baseline changes in blood pressure. Analyses prioritized measures of systolic blood pressure (SBP) over diastolic blood pressure (DBP), as systolic hypertension is substantially more common than diastolic hypertension, and SBP contributes more to the global disease burden attributable to hypertension than DBP \[[@B8]\]. Methods ======= Study sample ------------ The Framingham Heart Study is a community-based, longitudinal, observational cohort study that was initiated in 1948 to prospectively investigate risk factors for coronary heart disease. The Framingham Offspring Study began in 1971 with recruitment of 5124 men and women who were offspring (or offspring\'s spouses) of the Original Cohort of the Framingham Heart Study. The design and selection criteria of the Framingham Offspring Study have been described elsewhere.\[[@B9]\] Participants were prospectively assessed during 7 examinations between 1971 and 2001. The consecutive examination dates were as follows: 1971-1975; 1979-1982; 1984-1987; 1987-1990; 1991-1995; 1996-1998, and 1998-2001. At each examination visit, participants underwent medical history, physical examination, anthropometry, and laboratory assessment of coronary heart disease risk factors, as previously described.\[[@B9]\] Framingham participants signed informed consent and the study is reviewed annually by the Boston University Medical Center Institutional Review Board. There were 5124 participants who completed Offspring Examination 1 (in 1971-1975), of which 4989 (97%) agreed for their data to be in the open-access dataset. Of these, 1099 subjects were excluded from the present analyses because of missing education data (primarily among the participants who did not attend exams 2 or 3 when education was assessed) or being \<28 years of age (n = 60) when education was assessed. Participants were restricted to those aged ≥28 years at the time education was assessed in order to allow at least 10 years from expected completion of high school (at age 18 years, on average), during which the participants could obtain higher levels of education. Consequently, 3890 subjects were included in the data analyses. Education --------- The participants\' own education was measured directly from Framingham Offspring Study participants at Examinations 2 (1979-1982) and 3 (1984-1987). Examination 3 education was used whenever available, otherwise the Examination 2 measure was used. In the original data, education was recorded in 6 categories of completed years of education: 0-4, 5-8, 9-11, 12, 13-16, ≥17 years. For current analyses, the participants\' own education was collapsed into 3 groups: ≤12 years (reflecting high school or less), 13-16 years (indicative of some post-secondary education including technical school and college degree) and ≥17 years education (approximating those with more than an undergraduate college degree). This grouping was motivated by both (i) statistical power considerations (to ensure adequate number of participants in each category) and (ii) substantive reasons, whereby the education categories represent educational milestones recognized to influence earnings, occupation type, and socioeconomic position in society. Blood Pressure -------------- Each participant rested for at least five minutes before blood pressure measurement. While the participant remained seated, a physician measured SBP and DBP each twice in the left arm with a mercury-column sphygmomanometer, according to a standardized protocol \[[@B10]\]. The average of the two readings was used for analyses. Covariates ---------- Covariates were measured at each exam. A binary indicator of current cigarette smoking was determined by self-report, defined as smoking regularly in the year prior to the examination (yes/no). Alcohol consumption was evaluated by self-reported average number of alcoholic drinks (e.g. beer, wine, cocktails) per week. Body mass index was calculated as the weight in kilograms divided by the square of the height in meters (kg/m^2^). Current antihypertensive medication use was self-reported and modeled as a binary variable (yes/no). \"Baseline age\" represented age at Examination 1. \"Time from baseline age\" was calculated as the difference between age at a given examination and the baseline age. Statistical Analyses -------------------- Primary analyses focused on associations of education (categorized as ≤12, 13-16, and ≥17 years, as described above) with longitudinal trajectories of SBP and DBP. Analyses relied on multivariable mixed linear models, which extend multiple linear regression to longitudinal analyses of repeated measures \[[@B11]\]. Accordingly, all effects reported in this study are likelihood-based estimates from mixed models, as are all 95% confidence intervals and test statistics used for inference about these estimates. The blood pressure measures from consecutive examinations represented repeated values of the continuous dependent variable. To model the dependence between repeated outcome measures, we used the autoregressive order 1 AR(1) covariance structure of the residuals, which assumed that blood pressure values measured at consecutive visits are correlated more strongly than those separated by longer time intervals \[[@B12]\]. All models adjusted for baseline age and time since baseline assessment (the former allowed us to adjust for potential cohort effects, such as increasing education over time in the United States). In further analyses, we additionally adjusted for several time-varying conventional risk factors for hypertension expected to be involved in potential mechanisms by which educational attainment may influence blood pressure (representing visit-specific binary indicators of current use of anti-hypertensive medication and current smoking, as well as time-varying continuous measures of alcohol consumption and body mass index). AR(1) is a standard choice for the covariance matrix in mixed model analyses of longitudinal data. Because there are no well established tests to compare fit of models\' based on alternative covariance structures, the AR(1) structure is usually selected *a priori*. Assessment of the consistency of the AR(1) assumption is shown in the Results section. We implemented this longitudinal analysis by using PROC MIXED, with AR(1) covariance structure specified in REPEATED statement in SAS \[[@B13]\]. The second class of models additionally adjusted all the education effects for baseline blood pressure. Accordingly, in these models, baseline blood pressure values were *not*used as the outcome measure, so that the number of dependent value measures for each subject was reduced by one. Adjustment for baseline blood pressure effectively implied that we compared post-baseline trajectories of blood pressure as if participants with different education had the *same*baseline blood pressure. This approach allowed us to separate the antecedent effects of education, and other related factors, that might have resulted in the baseline differences in blood pressure, from the potential long-term effects of education on post-baseline changes in blood pressure. Preliminary analyses were carried out to determine the most accurate representation of the effects of baseline age, and time since baseline. In particular, we *a priori*expected that both between- and within-subject effects of aging on blood pressure may be non-linear. Accordingly, we gradually expanded the basic model, with linear effects of age and time, by adding and testing first quadratic and then cubic effects of each of the two variables, while adjusting for the use of anti-hypertensive medication and for education. Furthermore, because we expected that the impact of within-subject aging (time) on blood pressure may vary depending on the baseline age, we also tested linear and quadratic interactions between age and time. All the multivariable mixed models employed in the final analyses, described below, adjusted for only those non-linear effects of age or time, and those interactions between these effects, that were statistically significant, based on Wald test with 2-tailed α = 0.05. Once the optimal representations of the effects of age and time, as well as of their interactions, were determined, these representations were used in the final analyses of the adjusted association of education with SBP and DBP. The final representation of the effects of baseline age and time from baseline for analyses on SBP was as follows: age+age^2^+time+time^2^+age\*time+age\*time^2^+age^2^\*time. For DBP, it was as follows: age+age^2^+time+ time^2^+age\*time+age\*time^2^+age^2^\*time+ age^2^\*time^2^. All analyses were sex-specific, as a formal test for sex-by-education interaction suggested that the effects of education may differ between males and females (p = 0.046 for SBP; p = 0.063 for DBP). Results ======= Participants included in the current analyses had higher mean age (36.7 vs. 34.5 years, P \< 0.001), and lower smoking rates (43.1% vs. 49.8%, P \< 0.001) than excluded participants. Included and excluded participants had similar distributions of sex and baseline values of SBP, DBP, body mass index, alcohol consumption and antihypertensive medication use. In females, unadjusted analyses demonstrated that education was inversely associated with baseline values of age, SBP, DBP, anti-hypertensive medication use, body mass index and smoking, and directly associated with alcohol consumption (Table [1](#T1){ref-type="table"}). In males, education was inversely associated with age, SBP, DBP, body mass index, alcohol consumption and smoking. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Baseline characteristics (means and proportion) of the Framingham Heart Study Offspring Cohort, according to educational attainment. Statistical variance shown in parentheses represents 95% confidence intervals. ::: Educational Attainment (years) ------------------------------------ ------------------- -------------------------------- ------------------- ------------------- ---------- **Female**  N 2024 970 838 216  Age (years) 36.2 (35.8, 36.6) 38.3 (37.7, 38.9) 34.8 (34.1, 35.4) 32.3 (31.2, 33.5) \<0.0001  Year of Birth 1937 (1936, 1937) 1935 (1934, 1935) 1938 (1938, 1939) 1941 (1940, 1942) \<0.0001  Systolic Blood Pressure, mmHg 118 (117, 119) 120 (119, 121) 117 (116, 118) 114 (112, 116) \<0.0001  Diastolic Blood Pressure, mmHg 75.9 (75.5, 76.3) 76.8 (76.1, 77.5) 75.7 (75.0, 76.3) 72.9 (71.7, 74.1) \<0.0001  Anti-Hypertensive Medication, % 3.0^†^ 3.8 2.5 1.4 0.035  Alcohol Consumption (drinks/week) 4.4 (4.2, 4.7) 4.2 (3.8, 4.6) 4.5 (4.1, 4.9) 5.2 (4.2, 6.1) 0.037  Body Mass Index, kg/m^2^ 24.0 (23.8, 24.1) 24.5 (24.3, 24.8) 23.5 (23.2, 23.7) 23.2 (22.7, 23.7) \<0.0001  Current Smoker, % 42.6^†^ 47.3 39.3 33.8 \<0.0001 **Male**  N 1866 757 697 412  Age (years) 37.2 (36.8, 37.7) 40.0 (39.3, 40.7) 35.6 (34.9, 36.3) 34.8 (33.9, 35.7) \<0.0001  Year of Birth 1936 (1935, 1936) 1933 (1932, 1934) 1937 (193, 1938) 1938 (1937, 1939) \<0.0001  Systolic Blood Pressure, mmHg 126 (126, 127) 128 (127, 129) 126 (125, 127) 124 (123, 125) \<0.0001  Diastolic Blood Pressure, mmHg 81.9 (81.5, 82.4) 82.9 (82.2, 83.7) 81.5 (80.7, 82.2) 80.8 (79.9, 81.7) 0.0008  Anti-Hypertensive Medication, % 3.5^†^ 4.0 2.9 3.4 0.482  Alcohol Consumption (drinks/week) 11.0 (10.5, 11.6) 12.6 (11.6, 13.7) 10.5 (9.6, 11.3) 9.1 (8.1, 10.1) \<0.0001  Body Mass Index, kg/m^2^ 26.5 (26.4, 26.7) 26.9 (26.6, 27.1) 26.4 (26.2, 26.7) 26.1 (25.7, 26.4) 0.0002  Current Smoker, % 43.7^†^ 49.4 47.7 26.3 \<0.0001 To test the trend across education level, we used t tests for continuous variables, and Cochran-Armitage tests for categorical variables. ^†^Pearson chi-square tests comparing males and females for proportion taking anti-hypertensive medication, and proportion current smokers demonstrated P-values of 0.41 and 0.49, respectively. ::: Using multivariable mixed linear models, mean SBP across the assessment times was higher for participants with low education compared with high education (Table [2](#T2){ref-type="table"}, Figure [1](#F1){ref-type="fig"}), after adjusting for baseline age and time from baseline age, including their selected quadratic effects and two-way interactions, as shown in footnotes of Table [2](#T2){ref-type="table"}. Specifically, in these analyses, the mean difference across all 7 visits in SBP for ≤12 *vs*. ≥17 years education was 3.26 (95% CI: 1.46, 5.05) mmHg in females, and 2.26 (95% CI: 0.87, 3.66) mmHg in males (Table [2](#T2){ref-type="table"}). Further adjustment for conventional time-dependent covariates representing current (up-dated) values of antihypertensive medication use, smoking, body mass index and alcohol consumption reduced the difference in females to 2.86 (95% CI: 1.13, 4.59) mmHg, and in males to 1.25 (95% CI: -0.16, 2.66) mmHg. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Multivariable-adjusted mixed linear models, demonstrating associations between educational attainment and longitudinal trajectories of mean systolic blood pressure, Framingham Offspring Study, 1971-2001. ::: ------------------- ----------------------- ----------------------- ------------------------------------ ---------------------------------- ------------------------------------------------------------- **Model Adjustment** **Sex (n)** **Education (Years)** **Age** **Age, Conventional Risk Factors** **Age, Baseline Blood Pressure** **Age, Baseline Blood Pressure, Conventional Risk Factors** Female (n = 2024) ≤12 **3.26 (1.46, 5.05)** **2.86 (1.13, 4.59)** **2.69 (1.09, 4.30)** **2.53 (0.93, 4.14)** 13-16 **2.00 (0.20, 3.79)** **2.14 (0.42, 3.87)** 1.30 (-0.31, 2.91) 1.47 (-0.12, 3.07) ≥17 0 0 0 0 Male (n = 1866) ≤12 **2.26 (0.87, 3.66)** 1.25 (-0.16, 2.66) 1.20 (-0.07, 2.46) 0.34 (-0.97, 1.64) 13-16 **1.55 (0.16, 2.94)** 0.88 (-0.51, 2.27) 0.98 (-0.28, 2.24) 0.39 (-0.90, 1.68) ≥17 0 0 0 0 ------------------- ----------------------- ----------------------- ------------------------------------ ---------------------------------- ------------------------------------------------------------- Point estimates (and 95% confidence intervals shown in parentheses) represent mean differences in systolic blood pressure (mmHg) between comparison and referent groups. Age adjustment refers to adjustment for baseline age and time from baseline age. Modeling for baseline age and time from baseline was as follows: age+age^2^+time+ time^2^+age\*time+age\*time^2^+age^2^\*time. Conventional risk factors include antihypertensive medication, smoking, body mass index and alcohol consumption. ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Mixed linear models adjusted for age, demonstrating associations of educational attainment with longitudinal trajectories of mean systolic blood pressure (SBP) in (A) females and (B) males**. Age adjustment refers to adjustment for baseline age and time from baseline age. Modeling for baseline age and time from baseline was as follows: age+age^2^+time+time^2^+ age\*time+age\*time^2^+age^2^\*time. Error bars represent 95% confidence intervals. Framingham Offspring Study, 1971-2001. :::  ::: A second set of analyses adjusted for baseline SBP, in an effort to evaluate if there were educational differences in the post-baseline values of blood pressure, independent of the baseline differences. In analyses adjusted for baseline age, time from baseline age, and baseline SBP, females with ≤12 years education had 2.69 (95% CI: 1.09, 4.30) mmHg higher SBP over follow-up compared with females with ≥17 years education (Table [2](#T2){ref-type="table"}). Further adjustment for conventional risk factors had minimal impact on the effect strength in females (effect reduced to 2.53 (95% CI: 0.93, 4.14) mmHg for ≤12 years vs. ≥17 years education). Associations were weaker in males, where those with ≤12 years education had 1.20 (95% CI: -0.07, 2.46) mmHg higher SBP over follow-up compared to males with ≥17 years of education, after adjustment for baseline age, time from baseline age, and blood pressure; effects were substantially reduced after adjusting for conventional risk factors (Table [2](#T2){ref-type="table"}). As a formal test for sex-by-education interaction suggested that the effects of education may differ between males and females (p = 0.046 for SBP), and of the covariates only alcohol consumption showed differential associations with education between males and females (Table [1](#T1){ref-type="table"}), analyses were repeated without adjusting for alcohol consumption. This approach evaluated if gender differences in associations between education and SBP persisted with and without adjusting for alcohol consumption. Analyses adjusted for all aforementioned covariates with the exception of alcohol (i.e. adjusted for baseline age, time from baseline age, baseline SBP, antihypertensive medication, smoking, and body mass index) demonstrated persistent gender differences in associations where mean difference across all 7 visits in SBP for ≤12 *vs*. ≥17 years education was 2.20 (95% CI: 0.59, 3.81) mmHg in females, and 0.60 (95% CI: -0.72, 1.92) mmHg in males, suggesting gender differences in the association consumption between alcohol and education were not a substantial explanation for gender differences in observed associations between education and SBP. DBP was higher in female, but less so in male, study participants of low compared to high educational attainment after adjusting for baseline age and time since baseline assessment including the selected quadratic effects and two-way interactions described in Table [3](#T3){ref-type="table"}. Specifically, the mean difference in DBP across all assessment times, for ≤12 *vs*. ≥17 years education was 1.47 (95% CI: 0.43, 2.50) mmHg in females, and 0.66 (95% CI: -0.17, 1.50) mmHg in males (Table [3](#T3){ref-type="table"}). Further adjustment for conventional risk factors including antihypertensive medication, smoking, body mass index and alcohol consumption reduced the association strength of low education with DBP somewhat in females, resulting in a smaller difference of 1.26 (95% CI: 0.25, 2.26) mmHg, and eliminated any association in males, with adjusted difference of 0.05 (95% CI: -0.78, 0.86) mmHg. In analyses adjusted for baseline age, time from baseline age, and baseline SBP, there was no association between education and post-baseline values of DBP, among participants with the same baseline DBP, in either sex (Table [3](#T3){ref-type="table"}). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Multivariable-adjusted mixed linear models, demonstrating associations between educational attainment and longitudinal trajectories of mean diastolic blood pressure, Framingham Offspring Study, 1971-2001. ::: Model Adjustment ------------------- ------- ----------------------- ----------------------- -------------------- --------------------- Female (n = 2024) ≤12 **1.47 (0.43, 2.50)** **1.26 (0.25, 2.26)** 0.62 (-0.31, 1.55) 0.51 (-0.42, 1.43) 13-16 **1.29 (0.26, 2.33)** **1.40 (0.40, 2.40)** 0.33 (-0.60, 1.26) 0.44 (-0.48, 1.36) ≥17 0 0 0 0 Male (n = 1866) ≤12 0.66 (-0.17, 1.50) 0.05 (-0.78, 0.86) 0.42 (-0.33, 1.18) -0.09 (-0.85, 0.68) 13-16 0.60 (-0.23, 1.43) 0.14 (-0.68, 0.96) 0.49 (-0.27, 1.24) 0.09 (-0.67, 0.85) ≥17 0 0 0 0 Point estimates (and 95% confidence intervals shown in parentheses) represent mean differences in diastolic blood pressure (mmHg) between comparison and referent groups. Age adjustment refers to adjustment for baseline age and time from baseline age. Modeling for baseline age and time from baseline was as follows: age+age^2^+time+ time^2^+age\*time+age\*time^2^+age^2^\*time+age^2^\*time^2^. Conventional risk factors include antihypertensive medication, smoking, body mass index and alcohol consumption. ::: In order to assess the consistency of the AR(1) assumption with our data, we estimated the Pearson correlation coefficients between measurements at different time points, for both SBP and DBP, separately for female and male. The results of Pearson correlation coefficients (shown in Tables [4](#T4){ref-type="table"}, [5](#T5){ref-type="table"}, [6](#T6){ref-type="table"}, [7](#T7){ref-type="table"}) indeed show an autoregressive correlation structure, that is, the correlation decreases systematically as the distance in time between two measurements increases. In addition, we compared the AIC of the AR(1) model with that based on another popular structure: the exchangeable structure. As expected, based on Tables [4](#T4){ref-type="table"}, [5](#T5){ref-type="table"}, [6](#T6){ref-type="table"}, [7](#T7){ref-type="table"} and *a priori*considerations, the AR(1) (e.g. for male in the model specified in the last column of Table [2](#T2){ref-type="table"}, AIC = 73281) yielded better (i.e. lower AIC) than the exchangeable structure (AIC = 73434). ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Pearson correlation coefficients of systolic blood pressure among females for examinations 1-7. ::: Exam 1 Exam 2 Exam 3 Exam 4 Exam 5 Exam 6 Exam 7 -------- -------- -------- -------- -------- -------- -------- -------- Exam 1 1.000 0.620 0.590 0.531 0.492 0.441 0.370 Exam 2 0.620 1.000 0.717 0.660 0.600 0.556 0.465 Exam 3 0.590 0.717 1.000 0.744 0.671 0.629 0.526 Exam 4 0.531 0.660 0.744 1.000 0.707 0.637 0.546 Exam 5 0.492 0.600 0.671 0.707 1.000 0.646 0.572 Exam 6 0.441 0.556 0.629 0.637 0.646 1.000 0.669 Exam 7 0.370 0.465 0.526 0.546 0.572 0.669 1.000 ::: ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### Pearson correlation coefficients of systolic blood pressure among males for examinations 1-7. ::: Exam 1 Exam 2 Exam 3 Exam 4 Exam 5 Exam 6 Exam 7 -------- -------- -------- -------- -------- -------- -------- -------- Exam 1 1.000 0.554 0.513 0.461 0.406 0.330 0.303 Exam 2 0.554 1.000 0.681 0.592 0.495 0.462 0.418 Exam 3 0.513 0.681 1.000 0.679 0.588 0.538 0.460 Exam 4 0.461 0.592 0.679 1.000 0.661 0.567 0.452 Exam 5 0.406 0.495 0.588 0.661 1.000 0.592 0.516 Exam 6 0.330 0.462 0.538 0.567 0.592 1.000 0.584 Exam 7 0.303 0.418 0.460 0.452 0.516 0.584 1.000 ::: ::: {#T6 .table-wrap} Table 6 ::: {.caption} ###### Pearson correlation coefficients of diastolic blood pressure among females for examinations 1-7. ::: Exam 1 Exam 2 Exam 3 Exam 4 Exam 5 Exam 6 Exam 7 -------- -------- -------- -------- -------- -------- -------- -------- Exam 1 1.000 0.572 0.511 0.443 0.350 0.250 0.121 Exam 2 0.572 1.000 0.686 0.597 0.498 0.369 0.240 Exam 3 0.511 0.686 1.000 0.679 0.548 0.465 0.291 Exam 4 0.443 0.597 0.679 1.000 0.589 0.483 0.329 Exam 5 0.350 0.498 0.548 0.589 1.000 0.557 0.455 Exam 6 0.250 0.369 0.465 0.483 0.557 1.000 0.576 Exam 7 0.121 0.240 0.291 0.329 0.455 0.576 1.000 ::: ::: {#T7 .table-wrap} Table 7 ::: {.caption} ###### Pearson correlation coefficients of diastolic blood pressure among males for examinations 1-7. ::: Exam 1 Exam 2 Exam 3 Exam 4 Exam 5 Exam 6 Exam 7 -------- -------- -------- -------- -------- -------- -------- -------- Exam 1 1.000 0.533 0.453 0.373 0.273 0.158 0.061 Exam 2 0.533 1.000 0.625 0.498 0.406 0.292 0.202 Exam 3 0.453 0.625 1.000 0.593 0.489 0.375 0.322 Exam 4 0.373 0.498 0.593 1.000 0.555 0.447 0.337 Exam 5 0.273 0.406 0.489 0.555 1.000 0.520 0.447 Exam 6 0.158 0.292 0.375 0.447 0.520 1.000 0.566 Exam 7 0.061 0.202 0.322 0.337 0.447 0.566 1.000 ::: Discussion ========== Findings in this paper demonstrated that education was inversely associated with longitudinal trajectories of mean SBP in females and males. Furthermore, especially in females, lower education was associated with higher post-baseline SBP even among the participants with the same baseline SBP. This suggests that low education may have a long-term impact on changes over time in blood pressure in females. Adjusting for the time-varying values of conventional risk factors, measured at the same time as the blood pressure, typically reduced the strength of these associations. Associations of education with DBP were generally weaker than with SBP, for both females and males. Prior Literature ---------------- Few studies have investigated sex-specific longitudinal trajectories of blood pressure, particularly over a substantial proportion of the life course. Diez Roux *et al*. in the ARIC cohort (n = 8555) aged 45 to 64 years at baseline and followed using 4 examinations over a period of 9 years, found in white participants, that education was marginally inversely associated with increases in blood pressure after adjusting for age, sex, center, medication use, and reported interactions between time and sex, and interactions between time and baseline age \[[@B4]\]. The 5-year change in mean SBP was 6.0 mmHg for those with \<high school degree and 5.3 mmHg for those with a college degree. Further adjustment for baseline SBP somewhat reduced the association strength, to 5.9 mmHg for \<high school and 5.5 mmHg for participants with college degree. Associations were weaker in black participants. Strand et al. demonstrated, in a large prospective study of 48,422 males and females aged 35-49 followed for 14 years using three examinations, that education was inversely associated with increases over time in SBP in males and females, after adjusting for year of birth \[[@B6]\]; socioeconomic disparities widened over time in females but not males. In a study on the Framingham Offspring cohort that included only participants aged 20-29 years at baseline (many of whom may not have completed education yet), education was not significantly associated with mean 8-year change in SBP or DBP in males or females, after adjusting for age \[[@B5]\]. In the CARDIA study of 2913 participants aged 18-30 years at baseline education was significantly inversely associated with mean 15-year change in both SBP and DBP \[[@B7]\]. Specifically for SBP, those with \<high school degree had a 15-year mean increase of 8.2 mmHg versus only 0.7 mmHg for participants with \>college graduate degree. However the observed associations were not adjusted for covariates \[[@B7]\]. Although prior cross-sectional studies suggested that associations may be stronger in females than males \[[@B3]\], little is known about sex-specific associations between education and blood pressure trajectories, particularly over long periods of the life course (\>20 years follow-up). Finally, little is known about the effects of adjusting for use of antihypertensive medications, body mass index, alcohol consumption, smoking or other potential mechanisms that may, at least partly, mediate the impact of lower education on longitudinal trajectories of blood pressure. This study added to the literature sex-specific information demonstrating that education is inversely associated with longitudinal trajectories of mean SBP in females and males over a substantial proportion of the life course (approximately 30 years) and that association may be stronger in females than males. Furthermore, in females, lower education was associated with a higher mean post-baseline SBP even among participants with the same baseline SBP, suggesting a possible long-term impact of lower education. Adjusting for up-dated values of conventional risk factors typically reduced strengths of association, but in females the impact of lower education remained statistically significant. For DBP, association strengths were generally weaker for both females and males. Mechanisms ---------- The primary candidate mechanisms by which education may influence longitudinal trajectories of blood pressure involve conventional risk factors for hypertension, including smoking, obesity, blood pressure medication use, and alcohol consumption. In this study, in females, education was inversely associated with anti-hypertensive medication use, body mass index and smoking, and directly associated with alcohol consumption. In males, education was inversely associated with body mass index, alcohol consumption and smoking, and not associated with antihypertensive use. Furthermore, the estimated effects of education tended to somewhat decrease after adjusting for these potential mechanisms (particularly in males), suggesting that they may be at least partial explanatory pathways for the observed association between educational attainment and longitudinal trajectories of blood pressure. It is important to note that biases can be induced by adjusting for variables that may partly mediate the effect of exposure; therefore, these mechanistic findings should be interpreted with caution \[[@B14]\]. Furthermore, there remain plausible confounders unadjusted for, such as childhood socioeconomic circumstances (which are associated with adulthood education and blood pressure \[[@B15]\]), parental blood pressure (which may be associated with offspring education and has been related to offspring blood pressure \[[@B16]\]), intelligence (which is associated with educational attainment and CHD risk \[[@B17]\]), and early life obesity (that could affect upward social mobility via obesity discrimination particularly in women \[[@B18],[@B19]\], and is related to blood pressure in adulthood \[[@B20]\]. Consequently, residual confounding remains a possibility. Low educational attainment has been demonstrated to predispose individuals to high strain jobs, characterized by high levels of demand and low levels of control, which have been associated with elevated blood pressure \[[@B21],[@B22]\]. Other related mechanisms involve stress-induced sympathetic nervous system activation due to stressful conditions outside of work, that are also associated with low educational attainment. These may be particularly important for women. It has been shown that women with low education have higher risk of co-occurring psychosocial determinants of poor health, including single-parenting, depression, income below the poverty threshold, and unemployment, compared to men with low education \[[@B23]\]. Consequently, low socioeconomic position may be a stronger determinant of hypertension risk in women compared with men. This may be one of the explanations for why we found a significant interaction between sex and education, and somewhat stronger associations between education and blood pressure in women than men. The extent of health care available for people of low socioeconomic position is typically less than what is available for those with high socioeconomic position, hence limiting access to treatments of hypertension \[[@B24]\]. Furthermore, there is evidence that people of low socioeconomic position have less healthful diets, such as lower rates of fruit and vegetable consumption, and higher salt intake, which may be additional mechanisms contributing to disparities in blood pressure \[[@B25],[@B26]\]. It has been demonstrated that although both SBP and DBP are positively associated with incidence of coronary heart disease, there are differences in the way SBP and DBP evolve over the life course. SBP tends to increase steadily with age, while DBP tends to increase until age 50 years, and to decrease steadily after that age \[[@B4],[@B8],[@B27]\]. The mechanisms responsible for the age-related increase in DBP among younger people likely involve an atherosclerotic increase in peripheral resistance, caused by narrowing of the smaller arteries and arterioles \[[@B8],[@B28]\]. In contrast, for older individuals, structural damage and calcification due to arteriosclerosis in the larger conduit arteries can result in loss of arterial compliance, which can cause a rise in SBP, but a reduction in DBP \[[@B8],[@B28]\]. As the burden of hypertension is greatest after the age of 50 years, and it is exceedingly uncommon to have diastolic hypertension without concurrent systolic hypertension in adults over the age of 50 years, it has been argued that SBP is by far the more important measure of the two in terms of predictive importance for population health \[[@B8]\]. Studies generally show consistent inverse associations between educational attainment and longitudinal changes in SBP \[[@B4],[@B6],[@B7]\], with the exception of young participants aged 20-29 years at baseline, followed over 8 years in the study by Hubert *et al*. \[[@B5]\] However, findings are less consistent for DBP, where studies have shown inverse \[[@B7]\], null \[[@B5]\], or even positive \[[@B4]\] associations between educational attainment and longitudinal changes in DBP. Our findings demonstrated fairly robust inverse associations of education with SBP, and weaker inconsistent associations with DBP. The pathophysiological mechanisms (e.g. smoking, obesity, alcohol consumption) that cause steady increases over the life course for SBP but not DBP, and also tend to be inversely associated with socioeconomic position, may explain the more consistent findings for the inverse association between education and changes in SBP rather than DBP over the life course. However, adjustment for these variables in our study appeared to account for only a small amount of the association in females, and a larger amount of the (weaker) association in males, suggesting there may be other explanatory factors, particularly in females. Strengths and Weaknesses ------------------------ Strengths of this study include having access to data on approximately 30 years of longitudinal blood pressure measurements. Furthermore, follow-up rates of the Framingham Heart Study are considered to be high for observational studies, decreasing risk of bias due to loss-to-follow-up. Finally, measurements of blood pressure were performed using methods and equipment providing good accuracy and precision, and analyses relied on statistical methods appropriate for longitudinal repeated-measures studies. With regard to weaknesses, because the historical design of the Framingham Offspring Study reflected the population of Framingham, Massachusetts at study onset in 1948, the Original and Offspring cohorts are largely composed of white participants. Consequently, the generalizability of our findings to other races/ethnicities is uncertain. Furthermore, although we had up to 7 measurements for each covariate, we expect there to be reasonable residual confounding due to imperfect measurement of obesity (body mass index), and self-reported alcohol consumption, smoking and antihypertensive medication use. Conclusion ========== This study provides evidence that education is inversely associated with systolic blood pressure throughout a 30 year life course span, and associations may be stronger in females than males. These findings provide evidence that education may be a potential risk factor for elevated blood pressure across the life course. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= EL initially conceived of the study and further developed the study objectives in collaboration with all co-authors. EL wrote much of the initial draft of the manuscript. MA was the senior biostatistics advisor and drafted the analytic approach in the Methods section. YX performed the analyses and advised on the analytic approach. JL made substantial contributions to the analytic approach, and advised on the subject matter related to socioeconomic disparities in blood pressure. All authors were involved with drafting the final manuscript, and revising it as needed for important intellectual content. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/139/prepub> Acknowledgements ================ We are grateful to the National Heart, Lung and Blood Institute for providing us with the limited access dataset, in conjunction with Framingham Heart Study investigators who provided other important variables for analyses. Research funding was provided by Canadian Institutes of Health operating grant MOP81239, and a Canadian Institutes of Health New Investigator Award (EL).

PubMed Central

2024-06-05T04:04:19.559030

2011-2-28

PMC3053250

Background ========== Tuberculosis (TB) continues to be a global public health problem, with an estimated 9.4 million incident cases of TB and 1.8 million deaths in 2008\[[@B1]\]. Drug resistance and obstacles to successful directly observed therapy short-course (DOTS) impede disease control. Among patients being retreated for TB because of initial treatment failure, default from initial treatment, or relapse following initial treatment, drug resistance is common and retreatment outcomes inferior\[[@B2],[@B3]\]. Patients who fail, default from, or relapse after completion of standard first-line TB treatment and present for retreatment were previously grouped together by the World Health Organization (WHO) as Category II cases, and, in settings where individual drug susceptibility testing (DST) was not universally accessible, these patients were often treated with a standard retreatment regimen of first-line agents (a regimen that adds a single drug to the standard initial TB treatment regimen)\[[@B4]\]. Retreatment outcomes, however, are often poor, especially in patients with treatment failure or default\[[@B5]\]. DST may help identify those patients with multidrug-resistant (MDR) TB so that the appropriate antibiotics can be administered. Identifying local patient characteristics that confer higher risk of relapse, failure, or default from primary TB treatment may help inform country-specific prevention strategies aiming to reduce the need for retreatment, resulting in cost savings and diminished morbidity and transmission. In Morocco, the incidence of TB in 2008 was 81 per 100,000 overall but was significantly higher in several urban centers, or \"hot spots\". Of the roughly 28,000 new TB cases nationally each year, 12% are retreatment cases\[[@B5]\]. National TB treatment guidelines in 2007 and 2008 recommended a Category I treatment regimen \-- 2 months of isoniazid, rifampin, pyrazinamide, and streptomycin followed by 4 months of rifampin and isoniazid (2SHRZ/4RH) \-- for new smear-positive cases and a Category II regimen \-- 2HRZES/1RHEZ/5RHE (E = ethambutol) \-- for retreatment cases. Beginning in 2009, ethambutol replaced streptomycin in Category I regimens. Among retreatment cases in Morocco, 12.2% are infected with *Mycobacterium tuberculosis*strains that are resistant to rifampin and isoniazid, or MDR TB\[[@B6]\]. National guidelines suggest drug susceptibility testing (DST) for all retreatment patients. This retrospective cohort study examines the efficacy of standard TB retreatment in urban centers in Morocco, evaluates the uptake of DST testing for retreatment patients, and, using a nested case-control design, explores the risk factors that led to the need for retreatment in the first place. Methods ======= Sites ----- Morocco\'s National Tuberculosis Program (NTP) is well-established. TB care is provided free of charge by the Ministry of Health. Regional TB programs are divided into sectors, and each sector has a public health center (CDTMR) staffed by a specialist in TB care. TB clinical care is provided at CDTMR, while TB medications are delivered via DOTS at local clinics or dispensaries. *M. tuberculosis*culture and DST are performed at the National Reference Laboratory at the National Institute of Hygiene (INH) in Rabat or at the regional reference laboratory at the Institute Pasteur in Casablanca (IPM). National guidelines recommend DST for all retreatment patients. We chose nine urban CDTMR in high-incidence settings in Rabat, Casablanca, Fes, Tangier, and El Jedida as sites. Design ------ A retrospective cohort study of retreatment cases focusing on patient characteristics, treatment outcomes, and drug susceptibility was conducted. To determine the risk factors for failure, relapse, or default after initial treatment that led to the need for retreatment, a nested case-control study was performed. Study population ---------------- Moroccan patients with smear- or culture-confirmed pulmonary TB presenting for retreatment between June 2007 and August 2008 were identified using clinic registries. Those who had failed initial treatment, had relapse after completing initial treatment, or defaulted after at least two months of initial treatment were included in the cohort study. Among patients in the cohort study, those whose initial treatment had ended within two years of starting their retreatment regimen were eligible to be included in the nested case-control study as cases. The rationale for the two-year limit was to make it more likely that initial treatment charts would be available, to ensure that standards of practice for initial TB treatment would be similar across cases, and to increase the likelihood that recurrent disease represented relapse rather than reinfection. Controls were chosen from among patients with successful initial treatment without failure, default, or early relapse selected from the same center and treatment week. Definitions ----------- As per national guidelines, a patient with positive sputum smears for acid-fast bacilli after 5 months of continuous Category I treatment had treatment failure. A patient with initial treatment success after TB therapy of sufficient length (9 months for severe disease, 6 months for all others) that developed recurrent TB had treatment relapse. Treatment default was defined as interruption of treatment for ≥2 consecutive months. Treatment success was defined as treatment completion or cure. Retreatment patients were those receiving their first retreatment regimen after relapse, failure, or default. Data collection --------------- DST results were reviewed at INH and IPM to identify sites that used DST testing services. At clinics chosen as study sites, the TB registry and medical records of included patients were reviewed. For retreatment patients, information from initial TB treatment was collected if initial treatment was within two years of retreatment, and this information was used for risk factor analyses. This study was approved by the Ministry of Health of Morocco and by the Institutional Review Board of the Johns Hopkins University School of Medicine. Drug Susceptibility Testing --------------------------- Smear microscopy and culture were performed using standard methods. Specimens demonstrating growth of *M. tuberculosis*on Lowenstein-Jensen medium were tested for susceptibility to isoniazid, rifampin, ethambutol, and streptomycin using the proportion method. Critical concentrations were as follows: RIF 40 mcg/mL, INH 0.2 mcg/mL, streptomycin 4 mcg/mL, and ethambutol 2 mcg/mL. DST quality control was provided by the Supranational Laboratory at the Institute Pasteur of Algiers, as per standard laboratory operating procedures. Statistical analysis -------------------- Data analyses were performed in EpiInfo™ (Version 3.3.2, Centers for Disease Control, Atlanta, GA) except for multivariate logistic regressions, which were performed using STATA software, version 10.0 (StataCorp LP, College Station, Texas). Demographic and clinical characteristics of cases and controls during [initial]{.underline}TB treatment were compared using Pearson\'s χ^2^or Fisher\'s exact tests for categorical variables and student\'s *t*tests for continuous variables. Variables known to be risk factors for relapse, failure, or default as well as factors found to be associated with these outcomes in univariate analyses were included in multivariable logistic regression models. Significance tests were two-sided with p-values of ≤ 0.05 considered statistically significant. Results ======= Tuberculosis retreatment patients: population description, treatment outcomes, and DST results ---------------------------------------------------------------------------------------------- 291 patients with smear- or culture-positive pulmonary TB presenting for retreatment were identified and included in the study. Of retreatment cases, 232 (80%) had relapsed after completing an initial treatment regimen, 21 (7%) had failed an initial treatment regimen, and 38 (13%) had defaulted from initial treatment. The mean age was 37 years (IQR 27-46), and 78% were men. HIV and diabetes mellitus were rare (recorded for 1 and 3 patients, respectively). A standard Category II retreatment regimen was used in 272 (93%) of 291 patients. Retreatment outcomes were as follows: 172 (59%) cured, 26 (9%) completed treatment, 7 (2%) died, 73 (25%) defaulted, and 11 (4%) failed. Retreatment was successful in 173 (74%) of relapse patients, 10 (48%) of failure patients, and 15 (41%) of default patients.(p \< 0.01) Not surprisingly, retreatment failure was more common among patients who had failed initial treatment (24%) than among relapse (3%) or default (0%) patients (p \< 0.01). Default was very common among retreatment patients at 25% overall and was particularly high among patients with initial treatment default (57%) (vs. 20% among relapse patients and 24% among failure patients (p \< 0.01)). Among relapse patients, the median time from the end of initial treatment to diagnosis of relapse was 7.0 years (range 9 months to 44 years). Only 30 (10%) of 291 retreatment patients had DST testing performed with results available in the chart. Of three patients being retreated because of initial treatment failure who were tested, all had resistance to HRS, none of five being retreated because of default from initial treatment who were tested had pan-sensitive *M. tuberculosis*, and 3 of 22 (14%) patients who were being retreated because of relapse within two years after initial TB treatment had MDR-TB. Risk factors for relapse, failure, or default from initial TB treatment: results of the nested case-control study ----------------------------------------------------------------------------------------------------------------- Of 291 retreatment patients, 104 (36%) had started a retreatment regimen within two years of completing or stopping initial TB treatment and were, thus, eligible for the risk factor analysis. Of these, *initial*treatment charts were available for 83 (80%), and 80 were suitable for data extraction (cases, n = 80). 266 patients with initial treatment success were included as controls (controls, n = 266). (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Patient and disease characteristics of individuals receiving standard initial tuberculosis treatment, comparing patients with treatment relapse, failure, or default (cases) to patients with treatment success without relapse (controls) ::: ---------------------------------------------------------------------------------------------------------- Characteristic Cases\ Controls\ p-value (N = 80) (N = 266) ---------------------------------------------------------------------- ------------ ------------ --------- **At treatment initiation**  Male gender (n, %) 65 (81) 175 (66) \< 0.01  Marital status (n, %)§ 0.70   Single 32 (59) 101 (53)   Married 21(39) 85(45)   Other (widowed, divorced) 1(2) 4(2)  Medical comorbidities (n,%)\* 13(16) 30(11) 0.23  Cavity on chest x-ray (n,%)‡ 32(46) 101(43) 0.65  Age (mean, sd) 33.2(13.6) 34.5(14.3) 0.47  Weight at treatment initiation (mean, sd) 55.1 (9.5) 56.0 (9.4) 0.46  Habits (n, %)√   Tobacco use 18 (23) 42(16) 0.16   Alcohol use 2(3) 6(2) 0.89   Illicit drug use 6(8) 4(2) \< 0.01 **During treatment**  Sputum smear conversion to negative by 3 months (n,%)† 28(45) 220 (87) \< 0.01  Weight gain after two months of treatment (kg, sd)\*\* 2.6 (3.8) 3.4 (3.6) 0.09  Weight gain over initial treatment period (5-6 months) (kg, sd)\*\* 3.8 (5.8) 6.6 (5.3) \< 0.01  Missed doses during intensive phase 14(18) 10(4) \< 0.01  Hospitalization 21(26) 38(14) 0.01 ---------------------------------------------------------------------------------------------------------- §N = 54 for cases, 190 for controls, as marital status was not consistently recorded in the charts \*Patients with no evidence of comorbid conditions in chart review were combined with those for whom the absence of cormorbidities was expressly noted. ‡N = 70 for cases, 237 for controls, as x-rays were not performed and recorded for all patients. √Patients with no evidence of tobacco, alcohol, or illicit drug use in chart review were combined with those for whom the absence of tobacco, alcohol, or illicit use was expressly noted. †N = 62 for cases (15 of 27 patients with default did not have sputum smear conversion data, 3 patients with relapse were culture-positive, smear-negative cases); N = 251 for controls. \*\*N = 65 for cases (12 default patients, 1 failure, and 2 relapse patients did not have 2-month weight data), 263 for controls at 2 months; n = 50 for cases, 257 for controls after 5-6 months of treatment ::: In a multivariable logistic regression analysis, patients undergoing initial treatment for TB were at higher risk of a composite endpoint of failure, default, or relapse within two years if they were male (OR = 2.29, 95% CI 1.10-4.77), failed to have sputum smear conversion to negative by 3 months of treatment (OR 7.14, 95% CI 4.04-13.2), or required hospitalization during treatment (OR 2.09, 95% CI 1.01-4.34). There was a trend towards increased risk of this composite endpoint among those with poor weight gain (less than 10% by two months of treatment) or missed doses during the intensive phase of treatment, but these differences did not reach statistical significance. Odds of the composite outcome were 4% lower for each 1 kilogram increase in weight at treatment initiation (OR 0.96, 95% CI 0.93-0.99). Alcohol use and HIV were uncommon (2% and \<1%, respectively). Risk factors appeared to differ by subgroup, though analyses were limited by sample size (Table [2](#T2){ref-type="table"}). Risk factors for treatment default included male sex, substance use, missed doses during the intensive phase, and hospitalization. Risk factors for failure or relapse were harder to identify. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Risk factors for relapse, failure, or default from initial TB treatment, a subgroup analysis. ::: ---------------------------------------------------------------------------------------------------------- Subgroup\ Univariate analysis\ Multivariate analysis\ Risk factor (OR, 95% CI) (OR, 95% CI) ---------------------------------------------------------- ---------------------- ------------------------ **Relapse (N = 35)**  Lack of sputum smear conversion to negative by 3 months 4.58 (2.22-9.43) 4.14 (1.92-9.01)  Missed doses during the intensive phase 3.23 (0.96-10.9) NS **Treatment failure (N = 18)**  Lack of sputum smear conversion to negative by 3 months \*\* \*\*  Comorbid conditions 3.29 (1.08-9.99) NS **Treatment default (N = 27)**  Male gender 6.50 (1.51-28.1) 4.56 (0.96-20.7)  Substance use† 4.19 (1.83-9.56) 2.73 (1.04-7.15)  Missed doses during the intensive phase 12.9 (4.65-35.7) 10.9 (3.47-34.0)  Hospitalization during treatment 2.56 (1.04-6.23) 3.84 (1.41-10.5) ---------------------------------------------------------------------------------------------------------- \*\* Three-month sputum smear conversion to negative was negative in all failure patients so could not be included in logistic regression models. †Tobacco, alcohol, or illicit drug use. ::: Discussion ========== In this study of 291 patients undergoing retreatment for TB, outcomes differed considerably by group \-- 74% of patients with relapse, 48% of patients with failure, and 41% of patients with default had treatment success \-- similar to previous studies\[[@B5],[@B7]\]. Default from retreatment was extremely common at 25%, higher than previous country-wide estimates\[[@B5]\]. This may reflect temporal changes in treatment completion but more likely represents differences in study populations, as we focused on TB \"hot spots\", or urban centers with comparatively high TB incidence. Recent studies have demonstrated that, in urban settings, adherence is linked to patient knowledge about TB and provision of disease-specific education by the health care provider to the patient\[[@B8]\]. In busy urban clinics, time for education may be limited. Default from retreatment was most frequent among those who had defaulted from initial treatment, while failure was most common among those with previous failure. Although retreatment guidelines are often the same for patients with failure, default from, or relapse after initial treatment,\[[@B4]\] these results suggest that groups may benefit from different management strategies\[[@B9],[@B10]\]. For example, treatment failure is commonly due to drug resistance, while recurrence may be due to poor adherence, high mycobacterial burden (such as in cavitary disease), or exogenous reinfection. Default patients may require intensified case management and education, rather than more intensive treatment. The present study shows that, even when available, drug susceptibility testing is underutilized. It was performed in only 10% of retreatment patients. All 3 failure patients who underwent DST testing had MDR-TB, while 3 of 22 of relapse patients and 0 of 5 default patients tested did. While these DST results were only available for three failure patients and, therefore, not representative, these data and those from other studies suggest that MDR risk is not uniform among retreatment subgroups, with increased prevalence of MDR among patients with initial treatment failure\[[@B2],[@B11]-[@B13]\]. According to a population-based study conducted among retreatment cases in Morocco, 12.2% had MDR-TB, but the study did not divide retreatment patients into failure, relapse, or default subgroups\[[@B6]\]. Taken together, these findings support use of DST in all retreatment patients, earlier DST testing in those with clinical and microbiological indications of impending treatment failure, and use of second-line drugs for retreatment of patients with initial treatment failure until DST results are known. In Morocco, DOTS coverage is 100%, and concerted efforts to dramatically enhance DST use are underway. Published medical risk factors for failure or relapse include HIV infection, diabetes mellitus, low body weight, cavitation on chest x-ray, high bacterial burden, short treatment duration, drug resistance, and positive culture after two months of treatment\[[@B14]-[@B16]\]. Sociodemographic factors include unemployment, drug abuse, alcoholism, smoking, and poor treatment adherence. Treatment default is known to be associated with substance abuse, foreign birth, male gender, previous default, low socioeconomic status, psychiatric illness, unemployment, migration, side effects, \[[@B17],[@B18]\] long distance to the clinic, social stigma, and poorly-implemented DOTS but, of course, differ by setting \[[@B16],[@B19]-[@B21]\]. In our study population, HIV infection is rare; among TB patients, less than 1% are HIV-infected (unpublished data, Morocco NTP). Further, alcohol use in Morocco is uncommon, and smoking is extremely uncommon among women. Moreover, in the urban clinics studied, the majority of patients are non-immigrants, the clinics are geographically accessible, and DOTS coverage is 100%. Thus, many traditional risk factors for poor TB treatment outcomes are less prominent in Morocco, making it harder to prospectively identify patients at risk. However, continued sputum smear positivity after 3 months of treatment is a strong predictor of subsequent poor outcomes, and should prompt DST testing in all patients. As missed treatment doses may herald impending default, enhanced communication between the local clinics that dispense TB treatment and physicians at the regional health centers that prescribe it may be one country-specific strategy to help pinpoint those individuals who are missing doses and are at high risk of defaulting altogether. Small sample sizes limited our ability to evaluate subgroups, but even so, we were able to identify male sex, substance use (tobacco, alcohol, or illicit drug use), and missed doses during the intensive phase as likely risk factors for treatment default. Higher odds of hospitalization probably reflected the need for hospitalization to ensure adherence rather than increased disease severity. Further exploring risk factors for treatment default may help control programs identify those likely to benefit from targeted interventions such as health education, substance abuse counseling, enhanced tracking, or reinforcement of DOTS supervision\[[@B22],[@B23]\]. As a retrospective chart review, our risk factor evaluation was limited by the availability of data present in clinical charts. Information about tobacco, alcohol, and illicit drug use was not routinely recorded, for example. Classifying those with missing substance use data as nonusers likely biased analyses of this risk factor toward the null; however, estimates of smoking, tobacco, and illicit drug use were similar to national substance use statistics\[[@B24]\]. Also, in cases of recurrence, it was not possible to distinguish between relapse and reinfection, so we limited our risk factor analysis to those who had received initial TB treatment within two years of recurrence and were, thus, more likely to have true relapse. Our ability to identify independent risk factors in subgroup analyses was limited by small sample sizes; questions regarding risk factors in these subgroups would best be answered in larger, prospective studies. Finally, DST testing was not universally performed in retreatment patients, so selection bias is possible, as clinicians are more likely to send those at high risk of resistance for testing. In our study, 20% of retreatment patients with DST had MDR-TB, compared with 12% in a national prevalence survey\[[@B6]\]. Conclusions =========== Patients presenting for TB retreatment - those with relapse, failure of initial treatment, or default \-- are often grouped together and treated with a standard Category II retreatment regimen. However, these groups have distinct demographic and clinical characteristics, important differences in retreatment outcomes, and likely have different rates of drug resistant *M. tuberculosis*. Default from retreatment is common in high-incidence urban centers in Morocco, pointing to the need for strategies to address adherence. DST is essential for identifying retreatment patients with drug resistance, but even when available, it is underutilized, likely due to practical constraints. Preventing the need for retreatment in the first place is the best strategy given the individual and public health consequences of poor initial TB treatment outcomes, so strategies to identify and address country-specific risk factors are warranted to maximize treatment success. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= KED designed the study and helped with data analysis, interpretation of results, and drafting of the manuscript. OL assisted with access to and interpretation of laboratory drug susceptibility testing results. IG conceived of the study and assisted with study design, interpretation of results, and editing of the manuscript and provided invaluable clinical insight. JK helped with data collection and study design and implementation. MDE assisted with access to laboratory drug susceptibility testing results. IC performed the data analysis and assisted with interpretation of results. REA assisted with study design and study implementation and provided oversight of key study personnel at INH. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/140/prepub> Acknowledgements ================ The authors wish to thank Professor Mohammed Hassar for his assistance with obtaining DST results from the IPM; Dr. Afrah Akdi for her patience in teaching KED and JK about TB treatment in the field in Morocco; Amir Mohareb for his help with data collection in multiple Moroccan cities; Mrs. Radia Sabounie for her help retrieving DST results at the National TB Reference Laboratory; Dr. Baya Benhassine for her French translations; Dr. Jonathan Golub for his input and assistance; Dr. Khalil Ghanem for his thoughtful reading of the manuscript; and the CDTMR doctors and nurses who so kindly allowed us access to their patient charts and helped us in so many ways. KED is supported by NIH K23AI080842.

PubMed Central

2024-06-05T04:04:19.563339

2011-2-28

PMC3053251

Background ========== The increasing complexity of cancer care makes organisation of clinical decision-making one of the key elements in high-quality cancer care \[[@B1]\]. This raises the question of how good professional co-operation is to be achieved in day-to-day clinical practice. Pre-eminent among the different answers to this question is multidisciplinary teamwork, an approach that has emerged in parallel with the accelerated process of specialisation in cancer among health professionals. Recent reviews of the literature have associated a multidisciplinary approach to cancer care with better adherence to clinical practice guidelines \[[@B2]\], increased patient access to clinical trials \[[@B3]\] and enhanced co-ordination of hospital services \[[@B4],[@B5]\]. These outcomes, together with the role assigned to multidisciplinary care in various cancer plans, indicate its strategic role for health systems in general and for quality of care in particular in the organisation of cancer services, something that was highlighted at the Lisbon round-table held under the Portuguese EU Presidency (2007) \[[@B6]\]. The development of multidisciplinary cancer care involves a redistribution of health professionals\' tasks when it comes to clinical decision-making and patient care processes. The development of specific organisational frameworks for dealing with different types of cancer care must be seen in the context of a services-based hospital structure. This process modifies a highly sensitive aspect, i.e., the way in which professionals interact and are co-ordinated. The following are the main aspects defining a multidisciplinary approach in the organisation of cancer care: \- professional specialisation by disease (including diagnostic disciplines); \- standardisation of the process and clinical criteria in guidelines and pathways; \- redistribution of tasks at a multidisciplinary-team level; and, \- trend towards identifying and allocating specific resources according to disease/organ The Spanish National Health System (SNHS) recognised multidisciplinary care as being a health priority as far back as 2006, when a national strategy for promoting quality in cancer care was first published \[[@B7]\]. As one of its basic principles, this *Cancer Strategy*document stipulates that cancer patients should be diagnosed and treated in the context of a multidisciplinary team (MDT), and goes on to identify tumour boards as the main mechanism for deciding and planning therapy. This institutional effort is being implemented on a co-operative basis within the context of Spain\'s decentralised health care system. The priority assigned by the respective regional health services to multidisciplinary care since health service management became operationally decentralised (2002) and the specific mode of organisation introduced at each hospital, together determine the starting point of the implementation of an MDT approach. A high degree of variability is thus to be expected. This study addresses the question of how multidisciplinary cancer care has been implemented and the critical factors linked to this process, with special stress laid on the knowledge of policy required to ensure effective team work. The study was undertaken against a common backdrop of a growing cancer care burden and a rapidly expanding range of potentially effective treatments, which involves \"therapeutic dilemmas about treatment options\" \[[@B8]\]. A qualitative approach was chosen in order to better understand the perceptions and beliefs of all the professional partners involved at the different public hospitals spread across the nation\'s various health care regions. Methods ======= Study design and setting ------------------------ A qualitative research method was used in order to describe health professionals\' points of view and experiences of multidisciplinary cancer care, and explore the barriers to be considered in future policy development. As an MDT approach in cancer had not been previously studied in the context of the Spanish health system, a pilot test was deemed appropriate. The aim was to identify a set of analytical categories which, along with a review of the literature, would define a theoretical basis for the interviews. The pilot scheme was undertaken by three teams at different hospitals in Catalonia, focusing on different tumours (breast, lung and colorectal). In order to ensure the relevance and appropriateness of the categories yielded by the test, health professionals from different disciplines were then asked to give their considered opinion. The methodology of analysis consisted of semi-structured interviews conducted *in situ*from October 2008 to January 2009 with professionals involved in the diagnosis and treatment of cancer patients at public hospitals of differing levels of importance, situated in the most populated regions of Spain, namely, Andalusia, Catalonia, Madrid, Galicia and Valencia. Participants were interviewed by an experienced, qualitative researcher. Recruitment ----------- A total of thirty-nine health professionals were recruited. They were deemed eligible in any case where they performed their professional task, wholly or in part, in an MDT. For the selection of informants and composition of the theoretical sample, three inclusion criteria were established. To ensure that the views of different health professionals could be explored, the first criterion laid down that five medical specialisations had to be covered in each region: these were medical oncology, radiation oncology, surgery, pathology or radiology, and oncology nursing (Table [1](#T1){ref-type="table"}). The second criterion reinforced a systematic approach to the phenomenon under study, by insisting on the presence of three opinions from each region (table [2](#T2){ref-type="table"}). The third criterion took the form of a restriction on interviewing professionals belonging to the same team, with the aim of preventing biased versions on the subject of study and so contribute to the internal validity of our research. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Detailed breakdown of the 39 professionals interviewed ::: Medical specialisations: No. of professionals ----------------------------------------- ----------------------  Medical oncologist 10  Radiation oncologist 8  Surgeon 7  Radiologist/Pathologist 6  Nurse 5  Other (palliative care, gynaecologist) 3 Cancer site:  Breast 19  Colorectal 8  Lung 4  Other 8 ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Profiles required for selection of key informants ::: From a process standpoint Professionals usually working with this type of organisational approach ---------------------------------- ----------------------------------------------------------------------------------------------------------------- From a technical standpoint Professionals who have led organisational change towards more integrated forms of multidisciplinary cancer care From an institutional standpoint Professionals who frame the situation of multidisciplinary care within a hospital management model ::: Interviews ---------- A semi-structured list of questions ensured that critical points were covered in every interview. To elicit beliefs and experiences, participants were given the necessary flexibility to enable them to volunteer information on topics that were relevant to them. The selected health professionals were interviewed on a one-to-one basis for 60-100 minutes at their hospital offices. Interviews started with a general question on the cancer team\'s approach and ended with a question on how multidisciplinary care could be promoted by health policy measures. No notes were taken by the researcher during the interview; instead, all interviews were audio-taped and transcribed in full by the researcher. These data were then compiled into a documentary record and rendered anonymous to protect confidentiality. Every transcription was checked against its corresponding audio record and accuracy was found to be good. A preliminary analysis was conducted after each interview. Analysis -------- Interview data were examined inductively, using content analysis to generate categories and an explanatory framework. Grounded theory methodology was considered appropriate for describing the organisation and culture of health professionals belonging to multidisciplinary cancer teams. As our study was theoretical and aimed at incorporating the organisational context in which cancer teams practised, we used an axial coding, as described by Strauss and Corbin \[[@B9]\]. Data were electronically coded with ATLAS.ti \[[@B10]\]. Whereas the thematic analysis enabled language use to be understood and professionals\' beliefs to be communicated, the method of constant comparison ensured that recurring views and experiences were obtained. The coding process and emerging themes were derived, on the one hand, from *a priori*issues drawn from the pilot test and previous research, and on the other, from issues raised by participants. Examples of codes were \"nature of agreements\", \"access of a patient to the team during his/her journey\" and \"impact management on care performance\". The consistency of coding/interpretation was checked during analysis by reviewing the transcripts at different moments in time. This process allowed for labelling and developing a reference of the data for subsequent exploration and identification. Accordingly, a thematic framework based on models of and barriers to effective multidisciplinary teamwork was identified. A specific effort was made to capture this stage of interpretation, i.e., by mapping, creating typologies \[[@B11]\] and finding associations among themes. Moreover, the preliminary findings were discussed at a workshop (held by the scientific societies), to which some of the health professionals interviewed and a number of social science researchers were invited. These discussions were useful for reinforcing team types for the Spanish health system and clarifying some barrier-related aspects. Ethical considerations ---------------------- The data for this study was based on professionally conducted interviews which, other than the consent of the professionals themselves, require no \"ethical approval\" from any research committee. However, as the aim of our study is sensitive to hospital organisation and relations between specialisations and health professionals, confidentiality is of the essence. Accordingly, the strategy pursued was to prioritise a selection of participants whose opinion on these issues was deemed crucial, and so, contact with regional cancer control policymakers was made. The implementation of the study began through contact with the heads of the regional cancer plans, who proposed a short list of health professionals selected in accordance with our criteria (see table [2](#T2){ref-type="table"}). Likewise, the resulting list was endorsed by the scientific societies of medical and radiation oncology. The health professionals concerned were then sent a letter of invitation explaining the research goals and a confidentiality agreement. On being advised by telephone and receiving an assurance as to the confidentiality of any information provided, all the professionals selected for study consented to the interviews being recorded. The consent form was formally signed at the meeting, with any doubts as to the designated purpose and method of research being discussed with the professionals. No-one refused the invitation to participate. Results ======= We identified three models of development of multidisciplinary cancer care into which all teams could be classified (table [3](#T3){ref-type="table"}). Their internal consistency means that they can be seen as *models of co-operation*, and despite being general categories, any given team model may well contain elements of the other two. Moreover, the qualitative classification described here may even occur within the same hospital for different teams, with each of these assuming responsibility for a different tumour. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Models of co-operation in multidisciplinary cancer care ::: 1\. Advisory committee 2\. Formal co-adaptation 3\. Integrated care process --------------------------------------------------------- ---------------------------------------------- -------------------------------------------- ---------------------------------------------------- Cases submitted (approx. %) \"Complex\" cases or off-protocol: 10% - 50% All \"possible\" cases: 50% - 80% Initial source of clinical assessment: 90%-100% Patient access to team Treatment (whether or not initiated) Diagnosis or treatment Suspect or diagnosis *(early access)* Nature of agreements Recommendations Consensus decisions not always implemented Binding decisions defended by the team Professional team roles Negative perception Chair, tumour board co-ordinator Chair, co-ordinator, nurse case manager Impact on clinical process management Minor changes Some segments of care Whole process (*cross-boundary*frequent) Specialist participation No diagnostic specialisations Absences due only to timetable problems Professionals associated with a clinical committee Junior doctors and nursing role, in terms of attendance Considered inappropriate Open meeting, participation encouraged Mandatory presence Hospital executive board role Lack of interest Acknowledgement without express support Express support (room, clerk, etc.) Presence in health system 40% 50% 10% ::: Rather than targeting specific forms of organisation (tumour board, cancer unit), this approach focuses instead on team capabilities, based on each team\'s work method and the overall scope (breadth and depth) of the tasks performed, elements of analysis that emerged during the study and indicate the nature of interaction among professionals. While all cancer teams fulfilled the role of assessing patients and complying with and updating clinical protocols, obvious differences in their respective abilities to achieve quality of care were nevertheless observed. 1.- Advisory committee ---------------------- This is a group of professionals, largely made up of specialists in different therapeutic fields, which meets regularly on a informal basis to discuss cases considered clinically complex. Since the patients may already have received some of the treatments (usually surgery), the multidisciplinary meeting is aimed at referring them to other professionals for further treatment. This approach implies rigorous respect for the autonomy of the clinician and for overlapping boundaries between the team and the multidisciplinary meeting: patient assessment is made without other health care performance considerations. Judging by our results, this model continues to enjoy an important presence in the system (40%). 2.- Formal co-adaptation ------------------------ Owing to the high degree of interaction (mutual adaptation) among the professionals involved, consensus plays a key role in this model. The team acts as the reference framework for professionals, who share their views on the diagnosis, treatment and monitoring of a specific cancer type. The meeting is open to all professionals involved in patient management and it is here that the roles of team tumour board co-ordinator and chair appear. The key factor for fostering such an approach is agreement on the need for: joint decision-making to precede application of any treatment; and all cases to be dealt with at the multidisciplinary meeting. Both aspects are hampered, however, by hospital service inertia when it comes to disease management. This formula accounts for half (50%) of all MDT meetings in the health care system. 3.- Integrated care process --------------------------- The teams that work under this model share wide aims on patient management, including co-ordination of clinical research and economic evaluation of treatment. As this model provides teams with early access to patients, the latter\'s preferences along with knowledge of their co-morbidities and psychosocial context are incorporated into the multidisciplinary discussions. This occurs through systematic follow-up of patients throughout their journey, from suspicion of cancer, to diagnosis, therapeutic decision-making and follow-up. The presence of professional team roles has an impact on the entire care process and on progress towards achieving seamless care. The hospital executive board plays a pivotal role by consolidating meeting times, and ensuring health professionals\' attendance as well as their commitment to the MDT. This model is not frequent in the system (10%). Critical factors in multidisciplinary cancer team development ------------------------------------------------------------- For many clinicians, development of multidisciplinary care has involved a cultural change, as can be seen from the pathway which resulted from *recommendations*that were proposed by cancer team members for clinical management of patients and have gradually become *binding decisions*. The key critical factors identified for this change are as follows: ### Existence of different gateways for the same patient profile Many clinicians acknowledged significant variability in clinical practice as a result of diagnosing and treating patients who, despite having similar symptoms and diagnoses, might receive different initial therapy because access to hospital took place through different departments. A typical example is provided by the different therapeutic approaches proposed to a prostate cancer patient depending on the initial hospital department responsible for his diagnosis. Although a shared clinical protocol on such patients for the whole hospital plus an agreement to submit the patient to the MDT meeting would limit this variability, it is the unification of hospital-access gateways into a single hospital department that would have the necessary transforming quality in terms of standardisation of clinical criteria and pathways. Indeed, a recurring example of this type of organisational change is afforded by unification of admission to radiology for breast cancer or to gastroenterology for colorectal cancer in the case of patients displaying symptoms with a high risk of cancer. The experience of agreeing upon a common gateway for suspects has three effects: it provides teams with early access to patients; it reduces the feeling of patients being the \"property\" of any given clinician or department; and it sets up a primary care reference catchment area, as it becomes easier and clearer to determine where and how subjects with high risk of cancer should be referred and who the specialists of reference are. Moreover, where the department taking on the gateway unification process of the clinical pathway is a diagnostic unit, this implies that it should have a more relevant role within the team. \"At the hospital, there is a breast cancer unit that is beset by a root problem, i.e., two treatment options depending on whether the patient has been admitted via gynaecology or surgery. These are internal battles waged by the respective competencies.\" (Breast surgeon) \"In this hospital there are two chairs of surgery: one comes to the meetings, the other doesn\'t. We know that they administer different forms of treatment. The percentage of cases in which this occurs is by no means inconsiderable.\" (Medical oncologist) ### Variability in the development and use of clinical protocols and guidelines Evidence-based decisions are a source of concern to professionals, and the updating of clinical protocols by the team reflects this concern. Many clinicians felt, however, that this goal was conditioned by the implementation and dissemination of cancer clinical guidelines in the Spanish Health System. They argued that the absence of common guidelines for the whole country and the lack of co-ordination strategies for implementing the few that did exist resulted in reduced use and a lack of systematic assessment of existing levels of adherence. Owing to this perceived situation, clinical protocols at a hospital level are very often based on foreign guidelines, and efforts made to produce Spanish ones are of little use. This in turn leads to three common situations which impact at a team level. Firstly, hospitals that refer cases (e.g., because of their clinical complexity) have protocols based on different guidelines that are not standardised across the health care system. Secondly, multidisciplinary cancer care displays different levels of development, so that patients in one hospital may be referred to a specific department, but not to the tumour board, in another, the point here being the absence of pre-specified criteria for referral among levels of care. Thus, some decisions are made without the scientific consensus of an MDT. Finally, a cancer team might change the original treatment plan for a patient referred from a lower level hospital. This was a concern voiced by several clinicians, since decisions sometimes tend to differ widely, causing confusion and lack of trust in the patient. This perception was not shared by health professionals who work for cancer networks. \"When a surgeon comes along saying that he has operated on a given patient without the consensus of the committee but -according to him- this course of action was \'in line with the evidence\'\... this is unacceptable. It\'s an issue to be addressed by the respective cancer plans. When guidelines are reviewed, this should be the starting point, i.e., the tumour committee report should be seen before the surgical report.\" (Radiation oncologist) ### The role of the hospital executive board Most health professionals believed that, while they had not been hampered by the hospital executive board, neither had they been specifically supported to better organise clinical pathways and MDT activity. In their view, the main problem was that MDT work time was not recognised as a health care activity (or \"real work\" to quote them). Half of those interviewed felt that hospital managers knew little about their tasks, goals, level of involvement and management problems. These professionals identified two clear priorities for hospital executive boards, namely: to protect multidisciplinary meetings and work time; and to promote new professional roles, such as nurse case-managers or administrative support. Those with management responsibilities stated that cancer teams were not reflected in the organisational chart but were very important in terms of quality of care, and more innovative and responsive insofar as health care organisation achievement was concerned. \"If you tell management that you have to attend a committee meeting, they view it as something that is all well and good but nevertheless ancillary, and so not meriting consideration as part of the daily work load. Yet such attendance should be accorded health care and scientific value, i.e., so many hours correspond to committee work, which is equal to time spent seeing patients in a medical practice.\" (Colorectal [s]{.underline}urgeon) \"Your personal efforts are not appreciated, regardless of whether you\'ve taken part in drawing up a protocol or whether you\'ve devoted one day or three weeks to the job\... And, as no stress is laid on the importance of teamwork, there are pockets of resistance that don\'t change\". (Radiologist) ### Outcome assessment The main goal of any multidisciplinary cancer team is to enhance the effectiveness of diagnosis and treatment of a specific disease. Assessment of the MDTs that had been put in place revealed relevant differences among the views held by the professionals themselves. To most of them, the ultimate consequence of the efforts of some hospital units or specialised professionals that regularly collect clinical data was evaluation or a study aimed at assessing MDT outcomes. Others, in contrast, described process evaluation involving initial inter-departmental consensus on indicators, development of a specific data-collection methodology, and periodic analysis of results using a shared database. Above all, this situation defines different approaches to the possibility of taking clinical outcomes and process indicators, and linking these to actions aimed at improving cancer care. There were two recurring arguments associated with possible ways of achieving organisational change: the first centred on the key role to be played by the health care service in reaching a technical definition of and agreement on a minimum set of indicators for the entire hospital system and a proposed level of transparency vis-à-vis outcomes; the second addressed the pervasive \"culture of efficiency\" currently prevailing in hospital departments, insisting on the need to limit its influence and instead give increased relevance to clinical and process indicators. An experience that has had remarkable success in various health care regions and has served as the basis for the evaluation of each MDT, is the implementation of a fast-track, colorectal, breast and lung cancer diagnosis and treatment programme, a *driving force*in promoting integration among services and MDTs. Its implementation has shown the key role that health care policy could play in enhancing the organisation of cancer care. \"The problem is that each specialisation has developed its own indicators of toxicity, clinical results, etc. There should be at least one database in which the team\'s outcomes are reflected. This is something that the hospital ought to demand. We could then say in real time, \'this, or that, is what\'s happening in prostate cancer\'.\" (Medical oncologist) \"Yesterday I saw 37 patients: I can\'t devote myself to recording that much information in the database without any support\... It\'s difficult for everyone\'s survival to be ascertained under such conditions. We tend to move within a \'dead\' database context, that\'s to say, we get together at the end of the year to see how things have gone\...\" (Colorectal surgeon) ### Recording and documenting of clinical decisions in an MDT setting The more formalised MDTs become, the more important easy access to and transparency of decisions and the rationale behind them are. The reason for this is that recording decisions reflects the outcome of consensus building and the value that professionals attribute to their work. Half of all clinicians interviewed stated that they noted their decisions on the electronic clinical record. Not only does such action clearly define the end-point of the decision-making process, it also renders it more transparent, something that, in turn, generates a positive perception of the entire hospital environment. In contrast, there are many cases where team decisions do not extend beyond the strict limits of the tumour board, as shown by the first comment in Excerpt 5. The major weaknesses in recording clinical decisions stem from the lack of standardisation achieved in tumour-board Minute-taking, due to absence of common forms, failure to identify clear recording responsibilities and, very often, lack of administrative support. What this tends to mean is that only the decisions affecting of-protocol patients are recorded, thus hindering the possibility of establishing a reference database for a specific cancer. One last very important aspect for any team is the recording of decisions made in those cases where there is no consensus. Though infrequent, this situation is thought to play a relevant role in terms of medico-legal implications. \"In one of the hospital teams, there are professionals who find it difficult to accept consensus-based decisions. Accordingly, we consider it appropriate that, in addition to the decision being recorded in the digital clinical history, a file should be circulated to all team members so that decisions taken with respect to all patients are \'known\' to them\...\" (Medical oncologist) \"We keep a number of formal records, I mean to say that there are several specialists who record details of patients in their files\... but there is no single overall record.\" (Nurse case manager) \"There is an element of administration (which should be the task of a secretary) entailed in the drafting and signing of Minutes. This is generally performed by a physician, but if he\'s absent for any reason, then no-one does it. It\'s always the same old story: it\'s all a matter of personal involvement.\" (Pathologist) Discussion ========== The reference to Jean-Baptiste Lamarck (1744-1829) in the title of this article (\"the function creates the organ\") is a description that is both accurate and useful for understanding the evolution of multidisciplinary cancer care in the Spanish health system. In a manner similar to the apocryphal example of Lamarck\'s giraffe, which craned its neck to match the height of the trees, the development of cancer teams results from adaptation to the immediate hospital environment accompanied by a lack of policy orientation. While the law lays down that all hospitals are to have cancer boards for the most prevalent diseases, no specific aims, organisational requirements or performance assessment standards have been prescribed. There is a valuable lesson to be learnt in the path taken by the UK National Health Service. The publication of the Calman-Hine \[[@B12]\] report in 1995 highlighted the importance of a successful institutional framework for cancer services. As Haward \[[@B13],[@B14]\] pointed out, however, the effort to define their performance \-- including multidisciplinary care\-- in detail \[[@B15]\], without addressing the factors that were to facilitate the transition, resulted in slow, uneven change. The Spanish experience failed to develop this type of learning process. Our study therefore sought to identify the cultural and organisational dimensions that influence the incorporation of planned actions. This approach is reinforced by the EUROCARE-4 study, which identifies the organisational elements in the care process by the latter\'s ability to improve the survival and quality of life of cancer patients, as evidenced by the differences among European countries \[[@B16],[@B17]\]. Impact on decision making ------------------------- Implementation of multidisciplinary care involves a redistribution of the responsibilities assumed by the respective professionals, with the aim of developing greater potential for enhancing their joint clinical effectiveness. It is a specific organisational answer to the complexity of cancer care, and enables new approaches to be taken and known problems, such as variability in clinical practice, to be tackled. In this connection, note should be taken of the overall strategy adopted by the National Breast and Ovarian Cancer Centre of Australia, which, along with several other authors \[[@B18]-[@B20]\], identifies multidisciplinary care with the standardisation of clinical practice in the health system. Most professionals interviewed by us regard MDT as the main tool for ensuring that the expertise of each discipline is involved in the clinical decision making process affecting any given patient. Furthermore, high levels of adherence to clinical protocols improve the efficiency of MD meetings by better discerning the transition from simple to complex case discussions. Our study confirms previous research in underscoring the high degree to which the effectiveness of multidisciplinary interventions is dependent on the organisational context in which cancer care is delivered. Some technical aspects stressed are the need: for administrative support for team activity and organisation \[[@B8]\]; and for all decisions taken to be entered into the electronic clinical record, since failure to keep a record hinders application of such decisions to the patient, as shown by a study that targeted breast cancer teams in 2006 \[[@B21]\]. Moreover, a treatment-planning register can be helpful when it comes to assessing similar cases or auditing an MDT\'s performance \[[@B22]\]. Nevertheless, the key factor is communication among team members as a sign of professional team trust. This is the most relevant dimension to be discerned in the above-described models of integration of clinical care. The fact that decisions should be binding upon team members, that there is continued participation by specialists in the meetings, that the impact on the entire patient pathway is perceived as positive, that there is a role for clinical co-ordinators and nurse case managers, and that both residents and nurses participate in training, are indicators of the ability of the clinicians involved to abandon a sequential and relatively unco-ordinated model of cancer care and progress instead towards achieving a model of integrated care based on consensus decision-making. Specialisation in a given area of cancer diagnosis and treatment facilitates communication among different specialisations and professionals, through using specialist knowledge and expertise as a departure point for addressing specific patients rather than for the performance of specific tasks. Experiences, such as reaching an agreement on a common gateway for patients with high-risk symptoms for cancer or protecting the time for multidisciplinary teamwork, can also be key groundwork for promoting effective team communication. Other researchers emphasise aspects of the importance of professional integration within cancer networks \[[@B23]\], or the improvement in mental wellbeing and professional satisfaction that comes with MDT development, as a result of lower anxiety and better feelings about personal performance \[[@B24],[@B25]\]. Another approach is also the need to achieve consistent care from the standpoint of the cancer patient \[[@B26]\]. This was well illustrated by affording patients access to the MDT in the early stages of the diagnostic process, a way of preventing the possibility of initial treatment being administered without team discussion, and communicational fragmentation with the patient being increased. Strengths and limitations ------------------------- This study has some strengths and limitations that must be taken into account when assessing its results. Insofar as its strengths are concerned, it should be noted that, rather than approaching MDTs from the standpoint of the specific structures which frame teamwork, we sought instead to understand MDTs from the standpoint of the capabilities of the professionals and teams themselves. This enabled us to obtain a better insight into the ways in which professionals interacted and the nature of the agreements and commitments reached within an MDT. The synthesis of our results in the form of three models of multidisciplinary cancer care to be found in Spain facilitates the transfer of such findings to SNHS hospitals. Indeed, as our study shows, multidisciplinary care displays significant variability in its methodology and degree of implementation among hospitals and regions, but not in the critical factors that have influenced its development. These models have been checked with the health professionals involved in the study. A clear limitation of the study resides in the selection process, based on proposals put forward by the chairmen of scientific societies of medical and radiotherapy oncology and the heads of regional cancer plans, which could have biased our selection of professionals towards those with sensitivity to multidisciplinary care and organisational change *per se*. The selection criteria vis-à-vis the different profiles, plus the fact that major university teaching hospitals were involved in the study, were intended to minimise this limitation. Moreover, our interpretation of the findings and the model proposed here were discussed with different specialists, hospitals and regions. As with all qualitative studies, there was not a large number of participants. Our research focused on the views of key informants, thereby implicitly ruling out the possibility of capturing all the experiences and best practices that might exist in the health system. Lastly, mention should be made of the fact that one third of all interviewees belonged to the field of breast cancer, a disease that frequently becomes a model for others. Conclusions =========== This is the first qualitative study of multidisciplinary cancer care in southern Europe. The delay in MDT implementation poses the need for health policy not only to acknowledge and promote it, but also to provide quality standards. In addition, there is a clear need to respect and promote good practices existing in the health care system. In this regard, this study may help understand how professionals conceptualise this approach, which is relevant when interest lies in developing more comprehensive care by placing multidisciplinary care at the core of cancer department, as stated in Spain\'s official cancer strategy. Moreover, metaphors play a key role in the way professionals imagine and explain teamwork in cancer care (table [4](#T4){ref-type="table"}). ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### Research metaphors ::: The \"black box\" This metaphor is often used by health professionals outside MD meetings because of little knowledge of their internal functioning. --------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- \"The Lone Ranger\" The \"Lone Ranger doing \'clinical justice\' is outdated but we still have many \'Lone Rangers\' riding in our health system\", says one interviewee. Lone Rangers, in this context, are clinicians who unilaterally assume the management of cancer processes. \"Orchestra\" vs. \"Big Band Jazz\" In the case of the orchestra, a multidisciplinary team requires a \"baton to lead it\", a few \"first violins to give the health care symphony order and structure\" and several \"instruments\" which may stand out to a greater or lesser degree but must nevertheless all play in harmony so that the ensemble sounds good as a whole. To this end, developing an internal organisation based on commonly shared rules and roles is a crucial factor. Other professionals view \"Big Band Jazz\" as a more appropriate metaphor. They understand the functioning of the multidisciplinary team in a much less rational and formalised way, a human group in which improvisation and voluntary actions play a key role, with individual creativity as an essential component for ensuring that the process has a good outcome. \"Partitions and walls\" Professionals refer to the different metaphorical thickness of the partitions and walls to explain the mental distances that often separate them. \"Main actors, supporting (secondary) actors, and guests artists\" The feeling of playing specific roles in teams varies among professionals. Some of them express their involvement in terms of being main actors, and others as supporting actors or guests artists who attend the meeting only because they are invited. The \"snowball\" The large volume of visits entailed in long-term follow-up of cancer survivors, equivalent to one third of the time of activity for some professionals, leads them to refer to this process as a \"snowball\". In fact, one physician interviewed stated that, \"you almost marry a patient with cancer\". In \"no man\'s land\" or \"trapped between the two health systems\" These expressions are used in cases where good practices for taking care of cancer survivors are seen to be lacking, and the current intervention model is ineffective. Primary and specialised care are organised and conceived without identifying specific needs and consistent responsibilities vis-à-vis cancer survivors. ::: MDT development often entails a process of decentralisation inside hospitals, which may involve some redistribution of power. This is an adaptive challenge for hospital managers in terms of clinical governance, i.e., making structures more permeable to organisation of expertise without losing efficiency in the management of shared resources. A culture of evaluation of clinical and process outcomes should emerge, aimed at directing and justifying organisational innovation so as to achieve the best possible performance in terms of care of cancer patients. Multidisciplinary care occurs simultaneously with rapid changes in treatment and use of clinical practice guidelines, all of which makes it more difficult to identify its specific advantages \[[@B27]\]. This is why health policy plays an important role in promoting an organisational approach that changes the way in which professionals develop their clinical practice, a key issue in a disease such as cancer, characterised by its clinical complexity, involvement of different clinical specialists and need to face the new challenge of managing patient preferences. List of abreviations ==================== MDT: Multidisciplinary Team; SNHS: Spanish National Health System Competing interests =================== The authors declare that they have no competing interests. Ethical approval ================ Not required. Authors\' contributions ======================= JMB had the initial idea for this study. JMB and JP designed the study and drafted the research proposal. JP conducted pilot interviews, while JMB provided guidance and critical review of this information and helped with the review of the literature. JP undertook the main fieldwork for the study, interviewed, coded, charted and analysed the data for this paper, which was scrutinised and discussed by JMB. JP and JMB interpreted the results and wrote the first draft and final version of this article. JP and JMB read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/141/prepub> Acknowledgements ================ We should like to thank all the health professionals and decision makers who so unstintingly gave of their time, and shared their experiences and thoughts with us. We are also grateful to Tàrsila Ferro and Paula Manchon for their valuable contributions to the discussions held during the course of this study. This study was partially supported by an MSD grant from the Spanish Society of Medical Oncology (*Sociedad Española de Oncología Médica*- *SEOM*). Neither the MSD nor the SEOM had any involvement in the research process or the drafting of this paper. Support from the Carlos III Institute of Health (*Instituto de Salud Carlos III*- *ISCIII*) to the Cancer Research Network (RD 06/0020/0089) is also acknowledged.

PubMed Central

2024-06-05T04:04:19.565563

2011-2-28

PMC3053252

Background ========== The growth of the older population in the UK is substantial and the social care and health needs of older people represent a considerable cost to health and social care services \[[@B1]\]. There is clear scientific evidence of the importance for \'successful aging\' of maintaining an active lifestyle-physically, mentally and socially \[[@B2]\] and considerable attention has been given to the need to promote the health of older people, to help them maintain active and independent lives in the community and delay the need for residential care \[[@B3]\]. There is also a need to address mental health problems among the elderly and wider issues associated with social isolation and exclusion. It is widely understood that the NHS cannot alone address the diverse needs of older people, but has to work in partnership with other statutory bodies and voluntary agencies in the development of services and models of intervention to assist people in living independently and maintaining health and wellbeing \[[@B4]-[@B7]\]. In recent years there has been growing interest in the potential value of the arts in addressing significant social issues \[[@B8],[@B9]\]. The social utility of the arts has been promoted by the national Arts Councils, and both DCMS and DH have supported research to explore the contribution of the arts in the field of mental health\[[@B10]\]. The Department of Health established a Review of Arts and Health Working Group which reported in 2007. This led to the Department of Health and Arts Council England publishing A Prospectus for Arts and Health (2007)\[[@B11]\], which provides an overview of the current Arts and Health field in the UK, and offers recommendations for future developments. Coinciding with publication of the Prospectus, Arts Council England published its own strategic framework on Arts, Health and Wellbeing \[[@B12]\].Under the Labour Government the Home Office recognised the value of the arts in work with offenders, and the Treasury funded several projects on arts and community health under the Invest to Save Budget framework (see for example Manchester Metropolitan \'Invest to Save Arts in Health Project\' <http://www.miriad.mmu.ac.uk/investtosave/>. There is growing recognition of the value of arts activities in improving the lives of older people \[[@B13]-[@B16]\], and particularly the value of live music and musical participation for older people \[[@B17]\], including those affected by Alzheimer\'s and other forms of dementia \[[@B18],[@B19]\]. These studies are useful in highlighting some of the standardized measures which may be employed in studies of arts and health. They do, however, address very disparate interventions within a single study, making cause-effect conclusions problematical. Further, many studies use subjects as their own controls, rather than incorporating a control or comparison group, which would further enhance scientific rigour. We have recently completed a systematic review of all available non-clinical research studies concerned with the possible benefits of active engagement in singing. A systematic search of the published literature identified 56 reports giving attention to singing and its benefits. Of these reports, 18 were excluded from consideration as being of limited interest. The remaining 36 sources are extremely diverse with respect to the research problem addressed, the participants involved, research design and methods of data gathering and analysis. These papers have been carefully scrutinised but limited synthesis is possible. Only two studies out of these 36 attempted to assess the wellbeing/health benefits of participation in group singing using standardised measures of health \[[@B20],[@B21]\]. Both studies were concerned with older people. In each study, a positivistic research model was adopted in which pre and post assessments were made of participants in intervention and control groups, with statistical comparison of means between the groups. Houston et al \[[@B20]\] (UK) assessed the impact of four hour-long sessions of singing over four weeks with 31 residents in three care homes using GHQ-28 and HADS. A non-intervention group of 30 residents in three further care homes were also assessed (mean age 84 years). The authors claim that the intervention group showed significant improvements in measured anxiety and depression over the four weeks compared with the non-intervention group. Cohen et al. \[[@B21]\] (USA) assessed the impact of weekly participation in a community choir on 90 people aged 65 and older. The study ran over a period of two years, and a range of measures of physical health, health service utilisation, mental health and social activity were employed at pre-test and follow up after one and then two years. A matched non-intervention comparison group (N = 76) were also assessed. The authors claim to find a range of positive effects from participation in the singing group, including higher rating of physical health, fewer doctor visits, fewer falls and better mental health. Both studies have limitations and both methodological and analytical weaknesses, which raise doubts about the validity of their conclusions. Neither study justified sample sizes in terms of study power. In neither study is there evidence of randomization to intervention or control group, or any formal assessment of cost effectiveness of the intervention. A central focus of the current proposal is the evaluation of an innovative initiative - the Silver Song Club Project - which provides opportunities for older people to come together on a regular basis to make music and sing, with the support of professional musicians and volunteers drawn from established choral societies and singing groups. Currently, over 40 Silver Song Clubs are in operation across the South East of England, managed by a third sector organization, Sing For Your Life Ltd. (SFYL). A qualitative and process-oriented formative evaluation has been completed \[[@B22]\], providing information on how the project is run, the experiences of facilitators, volunteers, participants and carers and perceived benefits gained. Data were collected on six of the clubs through observation of sessions, focus groups with volunteers and interviews with club facilitators and participants, venue managers and SFYL directors and facilitators. Older participants have reported positive health benefits across psychological, cognitive, social and physiological domains, supporting previous research findings. There is now a need to develop a more controlled and objective assessment of the benefits for older people of participation in Silver Song Club activities. Aims of the study ----------------- 1\. To assess the effectiveness for older people of active engagement in community music activities on measures of physical and mental health. 2\. To evaluate the cost-effectiveness for older people of active engagement in community singing. Methods ======= The study is a pragmatic randomised controlled trial with two parallel arms. The study has been granted ethical approval by Surrey Research Ethics Committee REC ref: 10/H1109/5 and complies with the Helsinki Declaration. Hypotheses ---------- ### Primary hypothesis Stated as null hypothesis Singing groups for older people are no better than usual activities in increasing health-related quality of life in older people measured six months after randomisation using the York SF-12. ### Secondary hypotheses Stated as null hypotheses 1\. Singing groups for older people are no more cost-effective than usual activities 2\. Singing groups for older people are no better at reducing anxiety and depression when compared with usual activities at six months after randomisation measured using the Hospital Anxiety and Depression Scale (HADS). Participants ------------ ### Inclusion criteria The study will, as far as possible, reflect the inclusive nature of existing Silver Song Clubs. Participants will be adults over 60 years of age, able to speak English and able both to provide informed consent and to complete questionnaires. ### Exclusion criteria Individuals unable to provide informed consent or complete questionnaires in English. ### Recruitment Researchers, in conjunction with the third sector partner organization providing the intervention, will seek suitable venues across East Kent for song clubs to take place. Adverts will be placed in the venues and local newspapers and leaflets will be delivered to homes within the vicinity of the venues. These will specify the locality of the five venues and request that interested individuals should specify a locality to which they will be attached should they decide to take part. Adverts will also state that participants will be randomly allocated into either the singing group or the \'usual activities\' group within that area. Local radio publicity will also be sought. Two \'taster sessions\' will be held in each venue which will include the provision of information, an invitation for questions and an opportunity for individuals to sample the nature of the proposed intervention programme. A dedicated phone line and email address will be set up to deal with responses and queries and a proforma will be devised to collate details of those volunteering to participate. Personal details will be held in a locked filing cabinet within the research centre. ### Randomisation and consent All individuals indicating an interest in taking part will be sent a baseline questionnaire. They will also be sent an information sheet outlining the purpose of the study and what to expect if taking part, plus a consent form to be returned with the questionnaire. All returned, consented questionnaires will be randomly assigned to either intervention or control arm within one of five geographical areas. Randomisation will be conducted using the secure remote randomisation service based at the University of Kent. Because of the known preponderance of women in singing groups, stratification by gender will be used. Individuals who would attend together if allocated to a singing group (e.g. couples or friends providing transport) will be randomised in pairs on request. Intervention ------------ The Silver Song Club model is a well established format for participative singing for older people and this will be replicated in a standardised form. Trained and experienced facilitators under the guidance of SFYL will meet to compile a twelve week 90 minute programme comprising songs from different eras and a variety of genres. This will be followed by a series of \'unification\' meetings, to ensure that all facilitators are aware of how to access the material and deliver it in the same way (e.g. accompaniment, key, acquiring copyright etc). The programme will be developmental, progressing from singing melody lines to harmonising, layering and singing in rounds. Chime bars will also be introduced where appropriate and there will be an opportunity for participants to request items. All clubs will deliver the same programme concurrently, and consistency will be monitored by a programme manager who will make unannounced visits to each club 5-6 times during the intervention. A song book will be especially produced for the trial and a register of attendees will be maintained. Control ------- Individuals in the control group will continue with their normal activities. Should they decide to join a singing group outside the research they will not be excluded but data will be collected on these activities. Sample size ----------- A power calculation was carried out for the primary outcome (health related quality of life measured on the York version of the SF-12). There are no previous randomised controlled trials evaluating the impact of singing groups on quality of life in this population. Using data from quasi-experimental and observational studies we hypothesise that a minimally important difference in health related quality of life between the intervention group and control groups is in the order of 5 points on the York SF-12, equivalent to a medium effect size difference of 0.5. To detect this difference using a two-tailed test, alpha of 0.05 and power at 80% requires 63 participants in each of the two arms, a total of 126. We anticipate 5 singing groups and 5 controls and need to take account of any clustering effect in calculating sample size. We have used a conservative estimate of intra-class correlation coefficient of 0.02, similar to general practice populations and a cluster size of 12. This inflates the required sample size by a factor of 1.2, 77 in each group, a total of 154. Previous research would suggest the loss to follow up at 6 months for this population would be in the order of 20% and this further inflates the sample size to 92 in each group, a total of 184. We do not anticipate that all those enquiring would be willing to take part in the evaluation but we do have evidence that the consent rate would be in the order of 70%. This would mean we would require a total of 240 potential participants, 24 in each singing group and control group over the course of the study. Outcomes -------- The study will use standardised instruments with demonstrated validity, reliability and responsiveness. In addition, data will be collected on health and social care service utilisation and activities (including musical and singing activities) engaged in over a period six months prior, and six months after randomisation. Outcomes will be measured using a postal questionnaire. ### Primary outcome The primary outcome measure will be self-assessed health related quality of life assessed using the York SF-12 \[[@B23]\]. This is a frequently used measure that has been validated for use with older people and for which population norms exist. The twelve multiple choice questions cover both physical and mental domains of health. Most questions ask about health over the past week. Measures will be taken at baseline (prior to randomisation), 3 and 6 months after randomisation. The instrument allows the generation of an overall health-related quality of life score and additional two component scores addressing physical and mental health aspects of health-related quality of life. ### Secondary outcomes Health utility will be measured using the EQ-5D \[[@B24]\]. This is a short, 3-level, 5-dimensional instrument allows the generation of Quality Adjusted Life Years. It is routinely used in the economic evaluation of health care and recommended for cost-effectiveness analyses. Anxiety and depression will be measured using the Hospital Anxiety and Depression Scale (HADS) \[[@B25]\]. This has validity for use in community as well as hospital settings and has been used in previous studies evaluating arts and music interventions. Health and social care service utilisation will be measured using a specially designed questionnaire used in a number of evaluations involving older populations. All secondary outcomes will be measured at baseline, and then 3 and 6 months after randomisation. In addition we will collect process data consisting of individual attendance at singing groups for participants and information on who delivered singing group sessions. Data analysis ------------- ### Effectiveness analysis Data analysis will be conducted blind to the allocation of participants. The primary analysis will be intention to treat, where participants are analysed as part of their randomised group irrespective of their attendance. This provides the most pragmatic estimate of effectiveness. The primary outcome measure, the SF-12, will be analysed at 6 months using analysis of covariance to adjust for baseline differences between the groups. Due to the clustered nature of the study in which participants are nested within singing groups multi-level modelling will be undertaken to adjust for any cluster effects. Results will be presented as means and 95% confidence intervals. Secondary outcomes will be analysed in a similar manner. We will also explore the potential efficacy of the intervention by conducting a per protocol analysis in which only those participants who attended at least 50% of sessions are included. Other secondary analysis will include a regression analysis to explore both potential predictors of outcome and any baseline x treatment interactions. ### Cost-effectiveness analysis The incremental cost-effectiveness of singing groups compared with usual group activities will be assessed from both a health and personal social services perspective using NICE guidelines \[[@B26]\] and a wider public sector resource perspective \[[@B27]\]. The costs of setting up singing groups will be assessed using the actual resource costs for these activities based upon local service costs including costs associated with premises and managerial overheads. The costs of providing singing groups will be based on information gathered regarding participant contacts (all intervention contacts the participants receive, including the time and resources used to deliver the song groups which the participants attended) throughout the study. The participants\' use of health, social care or welfare services at baseline, 3 and 6 months will be calculated from the service use questionnaire. Units of resources used will be combined with national sources of unit costs \[[@B28],[@B29]\]. Because the resource use occurs within a 12 month period no discounting will be applied. The EQ-5D will be used with population values and the quality adjusted life year (QALY) change calculated using the area under the curve method. Bootstrapping will be conducted to explore the sensitivity and stability of derived estimates and incremental cost-effectiveness ratios and cost-acceptability curves presented. Discussion ========== The aim of this study is to evaluate the effect of group singing on the health related quality of life in older people. The existing body of evidence suggests that singing may be of benefit to this population group, but leaves many open questions. The two most relevant published controlled studies have significant methodological limitations, which call into question the credibility of findings. This study proposes to advance knowledge through a number of mechanisms, including: the use of a scientifically rigorous methodological design, use of a power calculation to estimate sample size, a standardized intervention which can be replicated, the assessment of cost-effectiveness of the intervention and single-blind analysis. The proposed research should be regarded as a step in a progressive programme of studies investigating the value of group singing for older people. Taking the MRC framework for the evaluation of complex interventions \[[@B30]\] as a model, this protocol represents the second (evaluation) stage, building on the feasibility study (our formative evaluation - \[[@B22]\]). Further refinements may be undertaken in the future, building on this study and dependent upon greater resources. This may include: offering an alternative activity for the control group, a longer term follow-up and a greater geographical spread. If group singing proves to be both effective and cost-effective, it may offer an alternative to more expensive, labour-intensive and less enjoyable means to maintaining the health of people in later life. This would be beneficial to the NHS and social care organizations as well as to individuals, groups and communities. Conclusions =========== This study is, we believe, the first to investigate the health benefits of singing for older people based on a stringent randomised controlled trial design which includes a cost-effectiveness evaluation. Its intervention, based on an existing model of participatory singing for older people, maximizes its potential for acceptability as a health promotion initiative for this age group should the outcome prove positive. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= AS drafted the manuscript and is coordinating the study; SMC conceived of the study, participated in the design and is chief investigator; SPC revised the manuscript and advised on the design and data analysis; JR is principal investigator and NHS lead. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2458/11/142/prepub> Acknowledgements ================ This paper presents independent research commissioned by the National Institute for Health Research (NIHR), Research for Patient Benefit programme ref: PB-PG-0408-16038. The views expressed herein are those of the authors and do not necessarily reflect those of the NHS, NIHR or the Department of Health.

PubMed Central

2024-06-05T04:04:19.570033

2011-2-28

PMC3053255

Background ========== The canonical genetic code is not universal although it is present in most complex genomes. Its establishment is still under discussion once the discovery of non-standard genetic codes altered the \"frozen accident\" \[[@B1]\]. Woese \[[@B2]\] was one of the first to consider the adaptability of the genetic code. He suggested that the patterns within the standard genetic code reflect the physicochemical properties of amino acids. An argument in favor is the fact that in the canonical genetic code the amino acids with similar chemical properties are coded by similar codons. There are three basic theories on the origin of the organization of the genetic code \[[@B3]\]. The stereochemical theory claims that the origin of the genetic code must lie in the stereochemical interactions between anticodons or codons and amino acids. The second one is the physicochemical theory, which claims that the force defining the origin of the genetic code structure was the one that tended to reduce the deleterious effects of physicochemical distances between amino acids codified by codons differing in one base. The third one is the coevolution hypothesis \[[@B4],[@B5]\], which suggests that the structure of the genetic code reflects the biosynthetic pathways of amino acids through time and the error minimization at the protein level is just a consequence of this process. This coevolution theory suggests that codons, originally assigned to prebiotic precursor amino acids, were progressively assigned to new amino acids derived from the precursors as biosynthetic pathways evolved. For other authors as Higgs \[[@B6]\], the driving force during the build-up of the standard code is not the minimization of the effects of translational error, and the main factor that influenced the positions in which new amino acids were added is that there should be minimal disruption of the protein sequences that were already encoded. Nevertheless, the code that results is one in which the translational error is minimized. Several previous works have studied the genetic code optimality. Most authors have quantified the efficiency of a possible code taking into account the possible errors in the codon bases. Generally, a measurement of changes in a basic property of the codified amino acids was used considering all the possible mutations in a generated code. The most efficient code is one that minimizes the effects of mutations, as this minimization implies a smaller phenotypic change in the codified proteins. Once the efficiency of a code has been measured, different criteria are used to assess whether the genetic code is in some sense optimal. These analyses fall into two main classes: statistical \[[@B7]\] and engineering \[[@B8]\]. The first one considers the probability of random codes more efficient than the standard genetic code. With this alternative for measuring code optimality, the standard genetic code is compared with many randomly generated alternative codes. These considerations define the so-called \"statistical approach\" \[[@B7]\]. Comparing the error values of the standard genetic code and alternative codes indicates, according to the authors using this approach \[[@B9]-[@B13]\], the role of selection. The main conclusion of these authors is that the genetic code conserves amino acid properties far better than expected from a random code. In a first computational experiment with this alternative, Haig and Hurst \[[@B12]\] corroborated that the canonical code is optimized to a certain extent. They found that of 10,000 randomly generated codes, only two performed better at minimizing the effects of errors, when polar requirement \[[@B2]\] was taken as the amino acid property, concluding that the canonical code was a product of natural selection for load minimization. Freeland and Hurst \[[@B9]\] investigated the effect of weighting transition errors differently from transversion errors and the effect of weighting each base differently, depending on reported mistranslation biases. When they used weightings to allow for biases in translation, they found that only one in every million randomly generated alternative codes was more efficient than the standard genetic code. With a similar methodology, Gilis et al. \[[@B14]\] took into account the frequency at which different amino acids occur in proteins and found that the fraction of random codes that beat the canonical code decreases. Torabi et al. \[[@B15]\] considered both relative frequencies of amino acids and relative gene copy frequencies of tRNAs in genomic sequences which were used as estimates of the tRNA content \[[@B16]\]. Zhu et al. \[[@B17]\] included codon usage differences between species and Marquez et al. \[[@B18]\] tested the idea that organisms optimize their codon usage as well as their genetic code: codons with lower error values might be used in preference to those with higher error values, to reduce the overall probability of errors, although their conclusions were the opposite. Sammet et al. \[[@B19]\], using a genotype-to-phenotype mapping based on a quantitative model of protein folding, compared the standard genetic code to seven of its naturally occurring variants with respect to the fitness loss associated to mistranslation and mutation. According to the authors\' methodology, most of the alternative genetic codes were found to be disadvantageous to the standard code, that is, the standard code is generally better able to reduce the translation load than the naturally occurring variants. The second alternative for measuring code optimality is the so-called \"engineering approach\", followed, for example, by Di Giulio \[[@B8],[@B20]\]. The approach uses a \"percentage distance minimization\" (p.d.m.) which compares the standard genetic code with the best possible alternative. The p.d.m. determines code optimality on a linear scale, as it is calculated as the percentage in which the canonical genetic code is in relation to the randomized mean code and the most optimized code. Therefore, it is defined as *(∆~mean~- ∆~code~)/(∆~mean~- ∆~low~)*, where *∆~mean~*is the average error value, obtained by averaging over many random codes, and *∆~low~*is the best (or approximated) *∆*value. This approach tends to indicate that the genetic code is still far from optimal. With this methodology, Di Giulio \[[@B21]\] estimated that the standard genetic code has achieved 68% minimization of polarity distance, by comparing the standard code with random codes that reflect the structure of the canonical code and with the best code that the author obtained by a simulated annealing technique. The author indicates that the minimization percentage can be interpreted as the optimization level reached during genetic code evolution. With this methodology, the authors in \[[@B22]\] also considered the evolution of the code under the coevolution theory. We previously analyzed the evolution of codes within the coevolution theory \[[@B23]\]. We used the mean square (MS) measurement \[[@B9],[@B12]\] (Methods Section) to quantify the relative efficiency of any given code. We considered two possibilities to generate alternative codes: the first one is the model of hypothetical codes that reflects the current genetic code translation table (model 1), which is most used in the literature. Two restrictions were considered \[[@B9],[@B12]\]: 1\. The codon space (the 64 codons) was divided into the 21 nonoverlapping sets of codons observed in the standard genetic code, each set comprising all codons specifying a particular amino acid in the standard code. 2\. Each alternative code is formed by randomly assigning each of the 20 amino acids to one of these sets. The three stop codons remain invariant in position, these being the same stop codons of the standard code. This choice of a small part of the vast space of possible codes, with these conservative restrictions, as Novozhilov et al. \[[@B24]\] indicate, \"is based on the notion that the block structure of the standard code is a consequence of the structure of the complex between the cognate tRNA and the codon in mRNA where the third base of the codon plays a minimum role as a specificity determinant\". As the codon set structure of the standard genetic code is unchanged, only considering permutations of the amino acids coded in the 20 sets, there are 20! (2.43·10^18^) possible hypothetical codes. Without restrictions in the mapping of the 64 codons to the 21 labels there would be more than 1.51·10^84^general codes \[[@B25]\]. In this work we considered the commented restrictive codes. Nevertheless, as Higgs \[[@B6]\] indicates, none of the known examples of codon reassignment occurs by swapping the amino acids assigned to two codon blocks. Instead, one or more codons assigned to one amino acid are reassigned to another, so one block of codons decreases in size while the other increases. Furthermore, the amino acid that acquires the codon is almost always a neighbor of the one that loses it. As Higgs \[[@B6]\] states, \"The reason for this is that reassignments of codons to neighbouring amino acids can be done by changing only a single base in the tRNA anticodon\". Hence, we also studied a second alternative with these possible restricted hypothetical codes which consider these codon reassignments (model 2), model not considered in the previous literature. Methods ======= The optimality of a code is related to its relative efficiency when different errors are considered in the DNA sequence or in the transcription and translation machinery of the protein synthesis. The efficiency generally considers these possible errors to take into account the possible changes in codified amino acids and their properties \[[@B7]-[@B18],[@B20]-[@B27]\]. A code which, on average, generates fewer changes is more efficient, as the effects of errors are minimized. Encoding and genetic operators ------------------------------ An adapted genetic algorithm (GA) \[[@B28],[@B29]\] was used to search for alternative codes that were more optimized than the standard genetic code. Each individual of the genetic population must encode a hypothetical code. Model 1 of alternative codes considered permutations of the amino acids coded in the 20 codon sets observed in the canonical code, so each individual has 20 positions, and each position encodes the particular amino acid associated with the codon set (Figure [1](#F1){ref-type="fig"}). The use of a simple algorithm ensures that the individuals of the initial population encode the 20 amino acids. Three codons are used for the stop label, which are the same as those of the canonical code. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Evolutionary computing methodology applied to the search for better adapted codes**. (a) Each individual of a genetic population encodes a possible hypothetical code, which is evaluated according to its efficiency to changes or mutations in the codon bases. (b) This efficiency defines the fitness of each individual, which is used by the genetic algorithm to select the individuals (c) in which a genetic operator (swap) is applied to define the individuals of the next generation (model 1 of hypothetical codes). (d) This evolutionary process is iterated through generations. :::  ::: In model 1, the GA used a swap operator. The operator interchanges the contents of two codon sets, that is, once two positions are randomly selected, the amino acids codified by the two respective codon sets are swapped. Figure [1](#F1){ref-type="fig"} shows how this operator works. In model 2 of hypothetical codes each individual has 64 positions, corresponding to the 64 codons. In each hypothetical code, 3 codons are reserved for the stop signal. In this case, the genetic operator models the known codon reassignments \[[@B6]\]. This operator can be summarized as follows: 1\. Choose a random codon from the 61 codons that encode an amino acid. 2\. The encoded amino acid is copied (duplicated) in another codon (randomly chosen) which differs only in one letter with respect to the first codon. If the amino acid to replace is the only instance in the hypothetical code, then the operator is not applied. In both models, tournament was used as selection operator. It chooses the best in a window of randomly selected individuals from the population. Hence, the size of the window determines the required selective pressure. Moreover, elitism of the best individual was used, that is, this individual is kept in the next generation without changes. Fitness function in the Genetic Algorithm ----------------------------------------- The fitness function was the measurement that calculates the mean square (MS) change in an amino acid property resulting from all possible changes to each base of all the codons within a given code \[[@B9],[@B12]\]. Any one change is calculated as the squared difference between the property value of the amino acid coded for by the original codon and the value of the amino acid coded for by the new (mutated) codon. As most authors \[[@B9],[@B12],[@B20]-[@B22]\] we used the polar requirement as the amino acid property. This property can be considered as a measurement of hydrophobicity and it was introduced by Woese as a measurement for the polarity of an amino acid, which is defined as a partitioning coefficient of an amino acid in a water/pyrimidine system \[[@B2]\]. The final error is an average of the effects of all the substitutions over the whole code. Hence, the error ∆ is defined as: $$\Delta^{2} = \frac{\sum_{ij}{(X_{i} - X_{j})}^{2}}{\sum_{ij}N_{ij}}$$ where *N~ij~*is the number of times the *i-th*amino acid changes into the *j-th*amino acid, and *X~i~*is the value of the amino acid property of the *i-th*amino acid. The changes from and to \"stop\" codons are ignored, while synonymous changes (the mutated codon encoding the same amino acid) are included in the calculation. The MS value defines the fitness value of a given code and the evolutionary algorithm will try to minimize it. Results and discussion ====================== We tested the implemented GA, searching for alternative codes, with the two definitions of models of hypothetical codes previously explained. Figure [2](#F2){ref-type="fig"} shows the evolution of the MS across the generations of the genetic algorithm. The quality (fitness) of the best individual and the average quality of the population were the result of an average of 50 evolutions with different initial populations. The population size was 1,000 individuals for the different tests and we used tournament selection with a size of 3% of the population. The selected individuals pass to the next generation, applying the suitable genetic operators for each model (Methods Section). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Evolutions of alternative codes with MS and tMS**. Evolution of the best individual and the average quality of the population using MS and tMS as fitness with codes that reflect the canonical code translation table (model 1) and with the model of natural codon reassignments (model 2). MS is the average mean square error of the effects of possible changes in the three codon bases. tMS incorporates the quantification of mistranslation in each particular base. These curves of fitness evolution were the result of an average of 50 evolutions with different initial populations. The horizontal lines indicate the corresponding MS or tMS values of the canonical code. :::  ::: The mean value of the best final codes was 3.506 using model 1, with a low standard deviation of 0.031. The best value found by Freeland and Hurst \[[@B9]\] was 4.7 and the MS value of the standard genetic code is 5.19. The p.d.m., using the best value obtained by the GA, was 71% with these restrictive codes. Figure [2](#F2){ref-type="fig"} shows that evolution easily finds better adapted codes, although the p.d.m. value shows good adaptability of the standard genetic code. The p.d.m. with the codes of model 2 was 68%, this value being lower since the freer evolution of codes can obtain better optimal codes. We repeated the analysis taking into account the errors as a function of the base position in the codon. Table [1](#T1){ref-type="table"} shows the quantification of mistranslation used in \[[@B9]\] as well as in this work to weight the relative efficiency of the three bases. It presents a summary of the empirical data on the frequency of transition and transversion mutations at the three codon positions. The MS is changed to tMS, which weights the errors according to the values shown in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Quantification of mistranslation used to weight the relative efficiency of the three bases in the tMS calculation. ::: Combined weighting First base Second base Third base -------------------- ------------ ------------- ------------ For transitions 1 0.5 1 For transversions 0.5 0.1 1 ::: Using model 1, there was an increase from a p.d.m. value of 71% in the MS case to a p.d.m. value of 84% when the mistranslation biases were considered in the fitness calculation. Using model 2, the increase was larger, from a p.d.m. value of 68% in the MS case to a value of 89% using tMS. This implies that the standard code is better adapted when we consider the quantification of mistranslations. This agrees with the results obtained in the statistical study of Freeland and Hurst \[[@B9]\] (these authors used only model 1). Note that using the two fitness functions, MS and tMS, model 2 obtains better values, although using tMS the GA needs more generations to overcome the corresponding values found with model 1, so the evolutions with model 2 are shown with more generations. The reason of the better values with model 2 is that, with the movements of this model, there is the possibility to reach the codes obtained with model 1, so the GA has a larger landscape where to find better codes. The evolution of the quality curves leads to the same conclusion: Evolution requires more generations to obtain a better individual with a better value than that of the canonical code when using tMS. This is clearer with the known codon reassignments model. With the average quality we have the same effect, as the GA has greater difficulty in obtaining better individuals than the canonical code. Figure [3](#F3){ref-type="fig"}, with the usual representation of the genetic code, corresponds with the assignments of best evolved codes using MS and tMS in the restrictive codes as well as with the model of codon reassignments. The position of each codified amino acid is shaded by a gray scale representing its polar requirement value. Although there are very different assignments of amino acids with respect to the canonical code, the two alternative restrictive codes present two patterns that are correlated with systematic errors in the processes of replication and translation, which are also present in the standard genetic code \[[@B30]\]. Pattern I: Amino acids are more similar to each other along the first codon position than they are along the second. This \"column-like\" pattern corresponds to the high rate of translational misreading in the first codon position; Pattern II: Along the second position, amino acids associated with pyrimidine bases (U,C) or purine bases (A,G) are more similar within these sets than between them. This is associated with mutational bias in replication, in which transitions (mutations within these base sets) occur more frequently than transversions (mutations of a base in one set to a base in the other set). Pattern I is present in all the evolved codes except for the evolved code using model 2 and MS, where is more difficult to recognize such pattern. Pattern II is clearer in the best codes with tMS, especially with model 2 of hypothetical codes. This is logical because the tMS variant models the different frequency of transition and transversion mutations. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Tables of best evolved codes with MS and tMS in restrictive codes (model 1), and best evolved codes in the codon reassignments model (model 2)**. The position of each codified amino acid is shaded by a gray scale, representing its polar requirement value. :::  ::: The MS or tMS values of each sample of codes in each generation form a probability distribution against which the standard genetic code MS or tMS values may be compared. Figure [4](#F4){ref-type="fig"} shows the histograms at four stages of the evolutionary processes: initial population, generations 5, 10 and 50. The histograms of the initial populations present a similar distribution as the ones of Freeland and Hurst \[[@B9]\], as the populations are random. A better code (better than the canonical code) was not found by chance in those initial populations. At the end of the evolutionary processes, the situation changed radically, where most of the individuals showed a better MS/tMS than that of the standard genetic code. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Histograms of different generations of evolutionary runs with both models of hypothetical codes**. The histograms correspond to the initial generation, generations 5 and 10, and generation 50 when there is not significant improvement in the evolutionary processes. In the histograms, the *x*-axis gives a particular range of categories of MS/tMS values whereas the *y*-axis gives the number of individuals with an MS/tMS in that category. The vertical lines in the histograms indicate the category into which the MS/tMS value of the canonical code falls. :::  ::: Conclusions =========== We used a genetic algorithm to search for better adapted hypothetical codes and as a method to guess the difficulty in finding such alternative codes, allowing to clearly situate the canonical code in the fitness landscape. We are emphasizing what simulated evolution search can provide about such difficulty of discovering possible better codes than the canonical one, and we must take into account that our methodology does not provide possible evolutionary pathways by which the canonical code reached its current state, as done by other authors \[[@B6]\]. From our GA simulations we can infer several conclusions. First, our results are not in disagreement with the main result of the statistical approach, as it is shown in the histograms of the initial populations, because such distributions of codes demonstrate, using the MS and tMS cost functions, that the canonical code is much better than random codes. Moreover, we agree with Knight et al. \[[@B26]\] when they state that the code could be trapped in a local, rather than global, optimum, and when they indicate that the average effect of amino acid changes in proteins is unlikely to be perfectly recaptured by a single linear scale of physical properties \[[@B26]\]. Nevertheless, with the information provided by the evolution of the histograms (Figure [4](#F4){ref-type="fig"}), now we do not agree with the authors who focus their analyses on the statistical approach \[[@B7],[@B9]-[@B11],[@B27]\] when they favor it because, as they emphasize, the approach takes into consideration that the possible random codes form a Gaussian distribution of error values \[[@B13]\]. According to the authors, the canonical genetic code is \"extremely efficient\" \[[@B9]\]. When they used an amino acid similarity based on the PAM 74-100 matrix, Freeland et al. \[[@B27]\] stated \"if theoretically possible code structures are limited to reflect plausible biological constraints, and amino acid similarity is quantified using empirical data of substitution frequencies, the canonical code is at or very close to a global optimum for error minimization\" \[[@B27]\]. Nevertheless, Di Giulio has questioned this work, as the title of his article \"the origins of the genetic code cannot be studied using measurements based on the PAM matrix because this matrix reflects the code itself, making any such analysis tautologous\" clearly explains \[[@B31]\]. However, regarding the comments of the authors focused on the statistical approach, even beginning with the Gaussian distributions of random codes in the initial genetic populations, the GA simulations indicate that it is very easy to improve the adaptability level of the standard genetic code. The better codes were obtained with low selective pressure and in few generations. Hence, the canonical code is clearly far from optimal, as also revealed by the position of the optimality values of the canonical code in the curves of quality evolution (Figure [2](#F2){ref-type="fig"}) for the two models considered. In this sense, we agree with the engineering approach as this alternative tends to indicate that the canonical code is still far from optimal. Nevertheless, the more realistic model of the known codon reassignments shows a slightly better efficiency of the canonical code with respect to the first model, as revealed by the greater difficulty of the GA to overcome the optimality value of the canonical code, as Figures [2](#F2){ref-type="fig"} and [4](#F4){ref-type="fig"} indicate. Authors\' contributions ======================= JS planned the studies and wrote the manuscript. AM performed the different experiments. Both authors discussed the results and implications and commented on the manuscript at all stages. Both authors read and approved the final manuscript. Acknowledgements ================ This work was funded by the Ministry of Science and Innovation of Spain through project TIN2007-64330.

PubMed Central

2024-06-05T04:04:19.572717

2011-2-21

PMC3053256

Background ========== Biological databases, such as pathway databases, are very useful in helping biologists discover new and existing biological mechanisms. Each entry in a database is based on experimental results. Yet as the amount of published research increases, constructing databases becomes more difficult. Biomedical text mining is a key technology for automatically extracting important information from biomedical literature. One key task in biomedical text mining is gene normalization (GN), mapping genes mentioned in the text to their unique database IDs. This task is difficult because one gene mention may map to different genes in various species. For example, papers containing *in vivo*experiments may describe mouse and human genes with the same name (e.g. IL4) in the same paper. Closely related to GN is the important and fundamental task of identifying proteins for inclusion in protein-protein interaction (PPI) databases. PPIs are of great interest to biomedical researchers because of their crucial role in elucidating signal pathways, controlling central biological processes such as transcription factors involved in cell division and DNA transcription \[[@B1]\], and their implications in a range of human diseases including cancer and neurodegeneration \[[@B2],[@B3]\]. To provide efficient widespread access to PPIs information, some organizations have begun compiling structured PPI annotation in public databases. MINT \[[@B4]\], IntAct \[[@B5]\], and BioGRID \[[@B6]\] are examples of PPI databases containing large numbers of verified PPIs. PPI databases are also very useful in building databases of signalling pathways, like BioCarta \[[@B3]\], or protein networks. Since most PPI information exists in published papers, text mining is an ideal way to speed up construction of these databases. Several research institutes and groups have been active in the effort to curate resources for PPI text mining and lay our roadmaps for the development of PPI mining tools. In 2005, BioCreAtIvE I/II \[[@B7],[@B8]\], held by the Centro Nacional de Investigaciones Oncológicas (CNIO), established the GN task, in which participant systems must map gene names to database identifiers. Then in 2009, CNIO launched the interactor normalization task (INT) \[[@B9],[@B10]\] at the BioCreAtIvE II.5 challenge. Another more important and difficult task in the BioCreAtIvE II.5 challenge is the interaction pair task (IPT). In the IPT task, systems are further asked to return a list of interaction identifier pairs for each article. The goal of IPT is to map interaction pairs in well-formed full-text articles to UniProt identifiers (see \[[@B9],[@B10]\] for details on IPT) and to rank these identifier pairs according to their probability of being interactors (refers to the Interactor Ranking for detail). Ranking PPI pairs is very important because only a small percentage of the identifiers are suitable for database curation (23.10% in abstracts, and 7.02% in full-text articles). In a quality ranked list, the curatable identifier pairs should be placed at the top. Such a ranked list would be more useful for human curators. Unlike other biomedical text mining tasks, which use only abstracts as research data, the BioCreAtIvE II.5 challenge compiled a dataset of full-length journal articles, which were formatted in a well-formed block structure. Figure [1](#F1){ref-type="fig"} shows an example article with three blocks. The first line of each block defines the section type and its corresponding text content. The section types include TITLE, ABSTRACT, and BODY for the main article content, as well as FIGURE, KEYWORD, and TABLE. The second line defines the section and the subsection headings. For example, the article has a keyword \"SOCS3\", and a subsection heading \"SOCS3 interacts with MAP1 S in vivo and in vitro\". ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Snippet of a full text article in the BioCreAtIvE II.5 dataset**. :::  ::: Our three main contributions in this paper are a section-optimized ranking model, an efficient dynamic-programming-based relational re-ranking algorithm, and the interaction pair ranking function. Firstly, our ranking model is optimized to take advantage of section-specific information. Several studies \[[@B11]-[@B13]\] have shown that scientific authors do, in the majority of cases, follow the basic principles of the research article structure and assign information accurately to each section. Each section of the paper has different characteristics which we can use to guide GN and the ranking algorithm. For example, the Introduction often contains information that repeatedly appears throughout the article (key genes), while the Results section presents new scientific findings, such as PPIs. In addition to section-specific information, our ranking model also makes use of the metadata that is included in full-text articles, such as the keywords fields. Secondly, we have developed a re-ranking algorithm that considers co-occurrence among interactors in an article. Co-mentioned genes influence each other\'s rank \[[@B14],[@B15]\]. If two gene names frequently occur alongside each other in the same sentences in an article, they are likely to have an influence on each other\'s rank. Take a low-ranked interactor mentioned only twice in an article for example. If both mentions happen to be alongside the highest-ranked interactor in the article, then the low-ranked interactor\'s rank should be significantly boosted. Our re-ranking algorithm is designed to use this information to improve accuracy of interactor ranking. Using a greedy computational approach, the re-ranking procedure requires large amounts of computer resources and time to calculate each gene identifier\'s rank simultaneously and find the best ranked list. Therefore, to maximize computational efficiency, we implemented our re-ranking algorithm using dynamic programming. Finally, we describe an approach to generating a ranked list of binary interaction pairs which combines the INT re-ranking results with a labor-saving unsupervised approach that still achieves competitive results. Results and discussion ====================== Implementation -------------- ### Overview of the proposed INT and IPT system Figure [2](#F2){ref-type="fig"} shows a flowchart of our system for INT and IPT. The top block depicts the fundamental steps for both tasks. The well-formed full-text article is preprocessed to resolve the conjunction problems presented by \[[@B16]\]. We use several rules proposed in our previous work \[[@B17]\] to expand collapsed ranges, such as SOCS1-SOCS7 in the Introduction section of Figure [1](#F1){ref-type="fig"}, into their individual components. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **System workflow**. :::  ::: After preprocessing, the multi-stage GN procedure \[[@B18]\] is executed (see \"Multi-stage gene normalization\" sub-section). For each identifier extracted by the multi-stage processing, the corresponding context information is used to extract features and the identifier is ranked by a support vector machine (SVM) classifier (see \"SVM-based ranking model\" sub-section). In order to further refine the ranking results, we create a re-ranking algorithm that takes into consideration the rank of an identifier and the genes that co-occur alongside. The re-ranking results are treated as an interaction candidate list for the article. To further extract interaction pairs, an unsupervised association analysis method is employed to determine their association scores. We then combine the scores with the results of INT re-ranking to generate the list of interaction pairs (see \"Using re-ranked INT results for IPT\" section). Algorithm --------- ### SVM-based ranking model Our model is based on the SVM algorithm, a well-known Machine Learning (ML) algorithm that has proved useful for text classification \[[@B19]\]. We have extracted the following features related to rank identifiers found in the GN procedure: #### Frequency features The frequency with which the identifier appears in the entire article is used as a feature. In addition, based on the work of \[[@B20]\], who found that molecular interaction descriptions usually appear in the Results section, we added the percentage of an identifier found in the Results section against in the other sections as a feature. #### Location features The features indicate where the identifier appears in the full text. The following location types are taken into consideration: (1) title; (2) abstract; (3) the first section (usually the Introduction section), the Results section, the last section (usually the Conclusion section), the Appendix section, and the other sections; (4) section and sub-section titles, such as \"3 Results 3.1 SOCS3 interacts with MAP1 S in vivo and in vitro\" in the Figure [1](#F1){ref-type="fig"}; and (5) Figure and Table captions. #### Field features These features indicate if the target identifier\'s name occurs in the full-text article\'s Keyword field, full name, or abbreviation field. #### Co-occurrence features These features describe co-occurrence of the target interactor with three types of the paper\'s key information: keywords, full names/abbreviations, and references to figures in the article text. Co-occurring with the paper\'s keywords or full names/abbreviations implies that the identifier is very likely to be key identifier in this article \[[@B21]\] and \[[@B22]\] have shown that figures often concisely summarize the most important results or methods used and described in an article. The paragraphs containing figure references often summarize the content of a figure. Therefore it follows that identifiers that co-occur beside figure references have a higher possibility of being interactors. For instance, the interactor \"SOCS3\" appears beside \"Figure [1C](#F1){ref-type="fig"} and [1D](#F1){ref-type="fig"}\" in the Figure [1](#F1){ref-type="fig"}. #### Indicator phrase features \[[@B23]\] and \[[@B24]\] have shown that there are commonly occurring structures which explicitly state the article\'s main knowledge claim or the assertion for which the authors hope to be cited and credited in future articles. Examples are \"the aim/purpose of this paper/article/study\" and \"we conclude/propose\". Those phrases that signal important sentences in a text are referred to as indicator phrases \[[@B25],[@B26]\]. Such sentences can be used to create an extract-type summary of the text. We add a feature to indicate whether or not the identifier occurs in the same sentences that contain indicator phrases. Table [1](#T1){ref-type="table"} shows an example of the features used for SVM-based ranking and a sample ranking result, *r*= \[P40337, P40338, Q05513, Q9NPB6\]. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Re-ranking example ::: ------------------------------------------------------------------------------------------------------------------------------------------------------------- Gene ID P40337 P40338 Q05513 Q9NPB6 ------------------------- ----------------------------------------------------- --------------------- -------------------------------- ---------------------- GN procedure Exact Match Exact Match Exact Match Exact Match frequency 45/0.76 14/0.00 10/0.20 7/0.57 location Title, first, Results, last, sub-section title, fig Abstract, fig Abstract, first, Results, last first, Results, last co-occurrence keywords, abbr-full, fig keywords, abbr-full keywords indicator phrase ✓ ✓ ✓ SVM ranking result 1 2 3 4 Concurrence information P40338:1\ P40337:1 P40337:2\ P40337:2\ Q05513:2\ Q9NPB6:1 Q05513:1 Q9NPB6:2 ------------------------------------------------------------------------------------------------------------------------------------------------------------- ::: ### Re-ranking algorithm based on co-mentioned genes The ranking procedure tends to rank the focus genes higher. However, the main concern of INT and IPT is to extract interactors. Our re-ranking algorithm is designed to boost the rank of interactors by considering co-occurrence among gene mentions in a single article. The example shown in Table [1](#T1){ref-type="table"} gives a general illustration of how the algorithm works. Table [1](#T1){ref-type="table"} shows the initial ranking results, *r*= \[P40337, P40338, Q05513, Q9NPB6\], determined by SVM. The algorithm starts by using a function, **newRankedList**(see \"Step1: Generate possible ranked lists\" in Methods). Given the rank-*i*identifier *x*and all other identifiers *y*in *r*that co-occur with *x*, **newRankedList**generates a ranked list *l~x~*, where the 1^st^to *i*-1^th^position are empty slots, *x*is put in the *i*th slot. Each *y*is put in a slot starting at *i*+1^th^from highest to lowest **association**(*x*,*y*) score. The remaining slots are empty slots. The generated lists for Table [1](#T1){ref-type="table"} are shown in Table [2](#T2){ref-type="table"}. We use an unsupervised approach based on a sentence-level mutual information (MI) \[[@B27]\] to measure the association (see Methods section) because the dataset provided by the BioCreAtIvE II.5 challenge did not contain sentence-level PPI pair annotation. In this example, the new ranked list generated by P40337 is \[P40337, Q9NPB6, Q05513, P40338\]. For P40338, the ranked list is \[\_, P40338, P40337, \_\]. The others are excluded because P40338 only co-occurs with P40337. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Possible ranked lists determined by all identifiers ::: Gene ID\\Rank 1 2 3 4 ------------------ -------- -------- -------- -------- [P40337](P40337) P40337 Q9NPB6 Q05513 P40338 [P40338](P40338) P40338 P40337 [Q05513](Q05513) Q05113 Q9NPB6 [Q9NPB6](Q9NPB6) Q9NPB6 ::: After generating all possible ranked lists, there are new candidates in each rank. As shown in Table [2](#T2){ref-type="table"}, Q9NPB6 and P40338 are candidates in rank 2 and 4 respectively. Therefore, there are several possible combinations and each one represents a possible re-ranked list. We then define a **score**function that estimates the likelihood of an identifier *x*being re-ranked in *i*as follows: $$\begin{array}{l} {\textbf{score}(x,i,\textbf{L}) =} \\ {\textbf{rankN}\_\textbf{Ratio}(x,i,\textbf{L}) \times \textbf{svmAccuracy}(i) \times} \\ {\textbf{svmAccuracy}(\textbf{deciderRank}(x,i,\textbf{L}))} \\ \end{array}$$ where the **rankN\_Ratio**function is calculated based on all possible ranked lists, **L**(see \"Step 2: Assign scores to identifiers in all possible ranked lists\" for the details); **svmAccuracy**is the function that returns the INT accuracy of rank *i*in our SVM ranking (Details and rank accuracies are listed in \"Preliminary experiments on the INT training set\" section); the **deciderRank**function returns the rank of the highest ranked identifier that supports an identifier *x*in rank *i*. For example, assuming that the identifier *w*\'s corresponding new list is \[*w*, *x*, *y*, *z*\] and *x*\'s is \[\_, *x*, *y*, *w*\], *y*will be ranked third, and **L**= {\[*w*, *x*, *y*, *z*\], \[\_, *x*, *y*, *w*\]}. In this case, **deciderRank(***y*,3,**L**), will return 1 since *w*and *x*support *y*in rank 3, but *w*\'s rank is higher than *x*\'s. For a possible re-ranked list *r*, its overall score can be calculated as follow: $$\textbf{overallscore}(r',\textbf{L}) = {\prod_{i = 1}^{|r'|}{\textbf{score}(r\text{'[}i\text{],}i\text{,}\textbf{L})}}$$ where *r*\'\[*i*\] is the *i*th element of *r*\'. The re-ranked list with the highest **overallscore**among all possible combinations is chosen as the final output re-ranked list. If the duplication of identifiers in a re-ranked list is permitted, the optimal ranked list can be directly found by choosing the identifier with the highest ***score***value for each rank. However, a legal ranked list cannot have any duplicates. To avoid duplication, we add a duplication constraint on the **score**function: when estimating **score**(*x*,*i*,**L**), if the identifier *x*has been determined in the previous rank, *k*, the **score**(*x*, *k*,**L**) function must return 0 (i.e. **overallscore**(*x*,*k*,**L**) is 0). Unfortunately, the duplication constraint increases the computational complexity of finding the optimal ranked list. In order to find the optimal ranked list and avoid computational overhead, we propose a dynamic-programming-based algorithm. Details on the algorithm are described in the \"Implementation of the re-ranking algorithm\" sub-section. ### Using re-ranked INT results for IPT In order to extract interaction pairs, we follow the workflow shown in Figure [2](#F2){ref-type="fig"} to process the article and generate a ranked list of identifiers. If two identifiers are described in one sentence, they are extracted as an interaction pair candidate. The mutual information (MI) analysis described in the Methods section is used to determine their association score. We then combine the association analysis with the results of INT ranking. Based on the assumption that the **IPTScore**(which estimates if *x*and *y*is an interactor pair given *x*in rank-*i*and *y*in rank-*j*) is positively relevant to **association**(*x*,*y*) and *x*and *y*\'s individual interactor scores, we propose the following formula to combine the results of association analysis and re-ranking: $$\begin{matrix} {\textbf{IPTScore}(x,i,y,j)} \\ {\text{=}\textbf{association}(x,y)\text{×}\textbf{interactor}\text{\_}\textbf{score}(x,i)} \\ {\times \textbf{interactor}\text{\_}\textbf{score}(y,j)} \\ \end{matrix}$$ When using the **score**function introduced in Equation 1 as the **interactor\_score**function, the formula can be rewritten as: $$\textbf{IPTScore}(x,i,{}y,j)\text{=}\textbf{association}(x,y)\text{×}\textbf{score}(x,i,\textbf{L})\text{×}\textbf{score}(y,j,\textbf{L})$$ Because we use MI to determine the association, the formula can be rewrite as follows: $$\textbf{IPTScore}(x,i,{}y,j)\text{=}\textbf{MI}(x,y)\text{×}\textbf{score}(x,i,\textbf{L})\text{×}\textbf{score}(y,j,\textbf{L})$$ Testing ------- ### Dataset and evaluation metrics The BioCreAtIvE II.5 \[[@B28]\] provides 124 journal articles selected mainly from FEBS Letters for evaluating INT and IPT systems. Following the format of the BioCreAtIvE II.5 challenge, we use 61 articles published in 2008 (50%) as our training set and 63 articles published in 2007 or earlier (50%) as our test set. The dataset is preprocessed to convert Greek alphabet characters (e.g. α, β) to corresponding diacritics (e.g. alpha, beta). The most common information extraction (IE) evaluation metric is centered on F-measure, an evaluation score generated from combining precision and recall. This evaluation has an obvious shortcoming: it does not take ranking of results into account. E.g., two systems reporting the same two correct and eight wrong hits for a document would produce the same F-scores, including precision and recall values, no matter the ranking of the results. Therefore, the area under curve (AUC) \[[@B29]\] of the interpolated precision/recall (iP/R) curve used in the BioCreAtIvE II.5 challenge is used to evaluate the proposed approach. The AUC of the iP/R function *f~pr~*is defined as follows: $$\begin{matrix} {\textbf{AUC~iP}/\textbf{R}(f_{pr}) = {\sum_{j = 1}^{n}\left( {p_{i_{j}} \times {\text{(}r}_{\text{i}}-r_{i-1})} \right)}} \\ {p_{i}(r) = \max_{r'}{}_{\geq r}p(r)} \\ \end{matrix}$$ where *n*is the total number of correct hits and *p~i~*is the highest interpolated precision for the correct hit *j*at *r~j~*, the recall at that hit. Interpolated precision *p~i~*is calculated for each recall *r*by taking the highest precision at *r*or any *r*\' ≥ *r*. ### Preliminary experiments on the training set To examine calculate the accuracies used in the **score**function (Equation 1), we apply the thirty-fold cross validation on the training set. We use the SVM-based ranking procedure to determine ranks, and then calculate each rank\'s accuracy using the training set. The accuracy of rank *i*is calculated by the following formula: $$\textbf{svmAccuracy}(i) = \frac{\text{the~number~of~correctly~nomalized~identifiers~in~rank~}r}{\text{the~number~of~normalized~identifiers~in~rank~}r}$$ Figure [3](#F3){ref-type="fig"} shows the accuracies of ranks 1 to 15. We can see that as the rank increases, the accuracy drops. This implies that higher ranks predicted by our SVM-based ranking method are more reliable than lower ranks. As mentioned in Equation 1, we utilize this phenomenon for scoring each identifier. After all ranks\' accuracies are calculated, the re-ranking algorithm is employed. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **SVM-based ranking accuracy across different ranks in the training set**. :::  ::: ### INT test set performance Figure [4](#F4){ref-type="fig"} shows the AUC iP/R scores of three configurations of our system. In the configuration (SVM ranking/Rank1), multi-stage GN and SVM-based ranking are employed. In the configuration (SVM ranking+Re-ranking), the proposed re-ranking algorithm is added. We also implement a baseline method (Freq) which ranks all identifiers according to their frequency. If two or more identifiers have the same frequency, two criteria are employed sequentially to rank them: (1) highest frequently in the Results sections (2) mentioned earliest in the article. Lastly, Figure [4](#F4){ref-type="fig"} also shows the AUC iP/R scores of the top three teams and the average AUC iP/R score of all BioCreAtIvE II.5 INT participants (Average). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **AUC iP/R scores of different INT approaches**. :::  ::: As shown in Figure [4](#F4){ref-type="fig"}, after employing our re-ranking algorithm, AUC iP/R performance increases by 1.84% over the previous top score in BioCreAtIvE II.5. According to our analysis, before re-ranking, gene identifiers whose feature values rarely appear in the training set are often incorrectly ranked because their feature values are underweighted in the ranking model. However, if these identifiers co-occur with higher-ranked identifiers whose feature values frequently appeared, our re-ranking algorithm is very likely to increase their ranks. This results in the improved AUC iP/R score. ### IPT test set performance Figure [5](#F5){ref-type="fig"} compares the results of three configurations of our IPT system. In the first configuration (MI), we rank all possible interaction pairs according to their MI scores. In the second configurations (MI+svmAccuracy), we simplify the original Equation 2 to rank all pairs. The **score**function is replaced by **svmAccuracy**. In the third configuration (MI+Re-rank), we use Equation 2 to rank all pairs. Figure [5](#F5){ref-type="fig"} also shows the AUC iP/R scores of the top three BioCreAtIvE II.5 IPT teams and the average AUC iP/R score of all BioCreAtIvE II.5 IPT teams (Average). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **AUC iP/R scores of different IPT approaches**. :::  ::: As shown in Figure [5](#F5){ref-type="fig"}, MI achieves a very low AUC iP/R (2.07%), while MI+SVMAccuracy achieves a much better AUC iP/R (16.02%) than Average (12.80%). MI+Re-ranking (our proposed method) further achieves a competitive AUC iP/R of 23.86% (outperforms the rank-1 system in BioCreAtIvE II.5 IPT). The results show that by using the results of INT re-ranking, the AUC iP/R score can be improved by 7.84% (compared to our system without re-ranking MI+SVMAccuracy) Comparing with the top team \[[@B10],[@B30],[@B31]\], which constructed a syntactic filter by semi-automatically annotating a PPI syntactic path corpus based on the GENIA corpus \[[@B32]\], and the third team \[[@B33]\], which manually annotated interaction proteins according to SDA information to construct a corpus for supervised learning, our approach requires less annotated data to construct and has the potential to be improved by using advanced **association**(*x*,*y*) calculation techniques. ### Comparison of SVM-based ranking and re-ranking We apply a two-sample *t*-test to examine whether the proposed re-ranking method is better than the original SVM-based ranking method by a statistically significant difference in INT and IPT. The null hypothesis, which states that there is no difference between the two configurations A and B, is given as $$H_{0}:\mu_{A} = \mu_{B}.$$ where *μ~A~*is the true mean AUC iP/R of configuration A, *μ~B~*is the mean of the configuration B, and the alternative hypothesis is $$H_{1}:\mu_{A} > \mu_{B}.$$ A two-sample *t*-test is applied since we assume the samples are independent. As the number of samples is large and the samples\' standard deviations are known, the following two-sample *t*-test is suitable: $$t = \frac{({\overline{X}}_{A} - {\overline{X}}_{B})}{\sqrt{\frac{s_{A}^{2}}{n_{A}} + \frac{s_{B}^{2}}{n_{B}}}}$$ If the resulting *t*-score is equal to or less than 1.67 with a degree of freedom of 89 and a statistical significance level of 95%, the null hypothesis is accepted; otherwise it is rejected. To get the average AUC scores and their deviations required for the *t*-test, we randomly sampled ninety training sets (*g*~1~,\..., *g*~90~) and ninety test sets (*d*~1~,\..., *d*~90~) from the 124 BioCreAtIvE II.5 journal articles. We trained the model with baseline ranking on *g*~i~and tested it on *d*~i~. We then performed the re-ranking procedure on the test results for *d*~i~. Following that, we summed the scores for all ninety test sets, and calculated the averages for performance comparison. Table [3](#T3){ref-type="table"} shows the results. We can see that after re-ranking, INT and IPT performances are improved by 3.12% and 3.88% respectively on the AUC iP/R scores with a statistically significant difference. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Comparison of INT and IPT performance on SVM-based ranking and SVM-based ranking + re-ranking ::: Avg AUC iP/R(%) Stdev AUC iP/R(%) *t* *t*\> 1.67? ----- ----------------- ------------------- ------ ------------- ------ --- INT 47.29 50.41 6.30 6.34 3.31 Y IPT 15.87 19.75 6.39 6.85 3.93 Y ::: Conclusions =========== In this paper, we have proposed a SVM-based ranking procedure with a relational re-ranking algorithm that considers the associations among gene identifiers to further improve performance on the BioCreAtIvE II.5 INT and IPT task. We formulated the re-ranking problem as an optimization problem and solved it by using dynamic programming to reduce computational complexity. We evaluated our approach on the BioCreAtIvE II.5 full-text dataset. In INT, the highest AUC iP/R achieved by our re-ranking system is 45.34%, 1.84% higher than that of our SVM-based system (Rank 1 in the BioCreAtIvE II.5 INT challenge). In IPT, our unsupervised method incorporating re-ranking not only achieves a promising AUC iP/R of 23.86%, which exceeds the best score in the BioCreAtIvE II.5 IPT challenge by 1.64%, but also saves significant annotation effort in comparison to other top teams\' supervised methods. A statistical significance *t*-test using ninety randomly selected training/test sets confirms that our additional re-ranking procedure significantly improves performance over the baseline ranking method. The proposed re-ranking algorithm relies heavily on association information, which it uses to generate possible ranked lists as described in the Algorithm section. We believe that our proposed approach could be combined with other advanced association scoring methods to further improve results. In the future, we plan to integrate our relation extraction method, which was developed for extracting hypertension-related genes in \[[@B34]\], into our INT and IPT system, and study the performance gains. We will also continue to improve our online web service to allow users to upload full text articles for PPI pair extraction. Methods ======= Multi-stage gene normalization ------------------------------ We have updated our previous one-stage GN system \[[@B17]\] with keyword-based species determination processing as well as multi-stage processing. In the following paragraphs, we briefly describe the improvement. Please refer to \[[@B18]\] for details. Keyword-based species determination finds the species for a given gene by checking the surrounding text for species keywords. If keywords are found, the corresponding identifier is assigned, otherwise the most frequently described species in the article is chosen. In order to exploit the characteristic of difference sections, we propose a three-stage GN procedure. In the first stage, GN is carried out on the Title, Abstract and Introduction sections. Successfully normalized gene mentions and corresponding identifiers are collected into a dictionary. In the second stage, we search the whole article for mentions recorded in this dictionary. The Title, Abstract, and Introduction sections are rechecked in case our machine learning (ML)-based gene mention tagger \[[@B35],[@B36]\] missed any instances. In the third stage, the remaining paper sections (except Title, Abstract, and Introduction) including figure/table captions and appendix descriptions are processed by our GN system. When combined with the dictionary-based approach used in stage two, disagreement of boundaries or identifiers may occur. In such cases, we select the candidate identifier with the longest gene mention string. ### Implementation of the re-ranking algorithm Figure [6](#F6){ref-type="fig"} shows the main steps of the proposed re-ranking algorithm. It accepts the output of the ranking procedure mentioned in \"SVM-based ranking model\" section as its input, and generates the re-ranked list as its output. In the following paragraphs, we describe the algorithm in detail. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **The re-ranking algorithm**. :::  ::: Notation -------- Let *r*be the ranked list generated by the SVM ranking procedure and *x*be any identifier in *r*. The possible ranked list generated by *x*is denoted by *l~x~*. The set of all possible ranked lists is denoted as **L**= {*l~x~*\|*x*in *r*}. For a possible re-ranked list, *r*\', the rank-*i*identifier is denoted as *r*\'\[*i*\]. ### Step 1: Generate possible ranked lists In the following paragraphs, we describe the algorithm, **newRankedList**, and functions that are used for generating possible ranked lists. A pseudo code implementation of the algorithm using python syntax is shown in Figure [7](#F7){ref-type="fig"}. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **A pseudo code implementation for *newRankedList*function**. :::  ::: **association**(*x*,*y*): This function measures the association between an identifier (interactor) *x*and another identifier *y*within an article and returns an association score. Several approaches can be used to measure this score, such as rule learning \[[@B37],[@B38]\], co-citation analysis \[[@B39]\], maximum entropy model \[[@B40]\], and conditional random fields \[[@B41]\]. Because the dataset provided by the BioCreAtIvE II.5 challenge did not contain sentence-level PPI pair annotation, we use an unsupervised approach based on mutual information (MI) \[[@B27]\] to measure the association: $$\textbf{MI}(x,y) = P(x,y)/\left( {P(x) \times P(y)} \right)$$ In the above formula, *P*(*x*) and *P*(*y*) are estimated by dividing *x*\'s and *y*\'s frequencies by *N*, the number of sentences containing gene identifiers. The joint probability, *P*(*x*,*y*), is estimated by dividing the frequency that *x*co-occurs with *y*in the same sentence by *N*. **newRankedList**(*x*,*i*)**:** Give the rank-*i*identifier *x*and all other identifiers *y*in *r*that co-occurs with *x*, **newRankedList**generates a ranked list *l~x~*, where the 1^st^to *i*-1^th^position are empty slots, *x*is put in the *i*th slot. Each *y*is put in a slot starting at *i*+1^th^from highest to lowest **association**(*x*,*y*) score (see Figure [7](#F7){ref-type="fig"}, lines 5-7). The remaining are empty slots. Take a ranked list of four identifiers *w*, *x*, *y*and *z*for example. For the rank 2 identifier *x*, the output of **newRankedList**(*x*,2) will be \[\_, *x*, *z*, *y*\] if **association**(*x*,*z*) \>**association**(*x*,*y*). ### Step 2: Assign scores to identifiers in all possible ranked lists Given an identifier *x*, a rank *i*and the set of possible ranked lists **L**generated in step 1, the score of *x*in *i*th rank is calculated as follows: $$\begin{matrix} {\textbf{score}(x,i,\textbf{L}) = \textbf{rankN}\_\textbf{Ratio}(x,i,\textbf{L}) \times \textbf{svmAccuracy}(r)} \\ {\times \textbf{svmAccuracy}(\textbf{deciderRank}(x,r,\textbf{L}))} \\ \end{matrix}$$ The following are definitions of functions used for defining **score**: **svmAccuracy**(*i*): Given a rank *i*, the function returns the INT accuracy of rank *i*in our SVM ranking. The accuracy is calculated based on a three-fold cross validation carried out on the training set. **deciderRabk**(*x*,*i*,**L**): For an identifier *x*whose new rank *i*is determined by more than one identifier, the function returns the rank of the highest ranked identifier. If only one identifier, *y*, determines *x*in *i*, the function returns the rank of *y*. **rankN\_Ratio**(*x*,*i*,**L**): Given an identifier *x*, rank *i*and the set of all possible ranked lists **L**, the **rankN\_Ratio**score is calculated based on the following equation: $$\textbf{rankN}\_\text{Ratio}(x,i,\textbf{L}) = \frac{\sum_{y}{\textbf{isRankN}(x,i,l_{y})}}{\sum_{y}{\sum_{j = 1}^{|r|}{\textbf{isRankN}(x,j,l_{y})}}}$$ where *y*is any identifier except *x*in *r* The intuition behind **rankN\_Ratio**is that the more ranked lists that agree with *x*in rank *i*, the more likely it is that *x*is in rank *i*. **isRankN**(*x*,*i*,*l~y~*) indicates whether the identifier *x*is in rank *i*of *l~y~*(return 1) or not (returns 0). Therefore, $\sum_{y}{\sum_{j = 1}^{|r|}{\textbf{isRankN}(x,j,l_{y})}}$ is the total number of times *x*appeared in all ranks of all lists in **L**. Figure [8](#F8){ref-type="fig"} shows the algorithm illustrated in step 2. It first generates all possible ranked lists for each identifier in *r*(lines 1-2). For each rank, the corresponding identifiers and their scores are calculated and stored in a dictionary-like data structure, *scoreInfo*(lines 3-6). Table [4](#T4){ref-type="table"} shows the data structure and its attributes. ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **Pseudo code implementation of *calculateScores*function**. :::  ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### The *scoreInfo*data structure ::: *scoreInfo* ------------------------------------------------------------------------------------------------------------------------------------------------------------------- A dictionary maps the rank (an integer) to a list of tuples: (*id*, *score*, *overallscore*, *from*). The dictionary\'s keys are the ranks in the re-ranked list. **Attributes** tuple.id: the identifier tuple.score: the score of tuple.id tuple.overallscore: the overall score of the ranked list after considering tuple.id tuple.from: the identifier in the previous rank, which leads to the optimal tuple.overallscore **Methods** *scoreInfo*\[*key*\]: Return the list of tuples in *si*with key *key*. *scoreInfo*\[*key*\]\[*i*\]: Return the *i*th tuple in the list in *si*with key *key*. ::: ### Step 3: Find the optimal overall score Given a possible re-ranked list, *r*\', its score is defined as follows: $$\textbf{overallscore}(r',\textbf{L}) = {\prod_{i = 1}^{|r'|}{\textbf{score}(r'\lbrack i\rbrack,i,\textbf{L})}}$$ where *r*\'\[*i*\] is the *i*th element of *r\'*. We can now formulate the re-ranking problem as an optimization problem that maximizes the overall scores over all possible rank orders: $$\underset{r'}{\arg\max}\,\,\textbf{overallscore}(r',\textbf{L})$$ The duplication constraint on the **score**function can be defined as follows: when estimating **score**(*x*,*i*,**L**), if the identifier *x*has been determined in the previous rank, *k*, the **score**(*x*,*k*,**L**) function must return 0 (i.e. **overallscore**(*x*,*k*,**L**) equals 0). For example, consider two possible re-ranked lists: *r*~1~= \[*x*, *w*, *y*,\...\] and *r*~2~= \[*x*, *y*, *z*,\...\]. Assuming that **overallscore**(*r*~1~,**L)\> overallscore**(*r*~2~,**L)**, and we now want to determine the value of **overallscore**function when *w*is in rank 4, then *r*~1~\'s **overallscore**becomes 0 because **score**(*w*,2,**L**) = 0 and *r*~2~\'s **overallscore**is greater than zero. Therefore, even though in rank 3, *r*~1~\'s **overallscore**is higher than *r*~2~\'s, the algorithm will not choose *r*~1~as the optimal sub-ranking when considering *w*in rank 4. Figure [9](#F9){ref-type="fig"} shows the dynamic-programming-based re-ranking algorithm. The algorithm starts by using *calculateScores*defined in Step2 to generate all possible ranked lists and their scores, which are stored in the data structure *scoreInfo*(lines 1). Lines 2-18 of Figure [9](#F9){ref-type="fig"} encompass the dynamic programming approach employed to find the optimal ranked list. ::: {#F9 .fig} Figure 9 ::: {.caption} ###### **Pseudo code implementation for *findOptimalScore*function**. :::  ::: In the forward phase (lines 3-15), the algorithm computes the optimal overall score for each identifier in each rank. Lines 8-11 find the maximum overall score in *scoreInfo*\[*i*-1\] in which the identifier, *scoreInfo*\[*i*\]\[*j*\] = *candidatesIDs*\[*j*\], does not appear among rank 1 to rank *i*-1. In the following formula, *scoreInfo*\[*i*\]\[*j*\].overallscore is shorted to **overallscore**(*i*,*j*), which is the optimal overall score from rank 1 to rank *i*when rank *i*\'s *j*th candidate is placed at rank *i*. *scoreInfo*\[i\]\[j\].identifier is shorted to, which stands for rank *i*\'s *j*th candidate. The score can be recursively calculated as follows: $$- \left\{ \begin{matrix} {score(ID(i,j),i,\textbf{L}) \times \max\begin{Bmatrix} {overallscore(i - 1,0)} \\ \vdots \\ {overallscore(i - 1,k - 1)} \\ \end{Bmatrix}\text{~if~i} > 2} \\ {score(ID(1,0),1,\textbf{L}\text{)~if~}i = 1,j = 0} \\ \end{matrix} \right.$$ where *k*is the number of tuples in the *scoreInfo*\[*i*-1\]. Lines 12-13 calculate the score. In the backward phase (lines 16-18), the optimal ranking is reconstructed by tracing the \"from\" attribute of the tuple in the last rank with maximal overall score (the *optimal\_end*) until the value of \"from\" is None. ### Step 4: Interaction pair ranking function After Step3, the algorithm has generated the re-ranked list for INT. The final step of the algorithm is to employ the IPTScore function defined in Equation 2 to generate the ranked list for IPT. Availability and requirement ============================ We have developed a demo website to demonstrate the proposed re-ranking algorithm. The service is available at <http://biosmile.cse.yzu.edu.tw/DPRerankAlgorithmForAbstractDemoWebsite/>. The web-based service has been tested and run on the Firefox 3.5+, Chrome 7+, and IE7+. Authors\' contributions ======================= RTHT designed all the experiments and wrote most of this paper. PTL wrote the programs and conducted all experiments. RTHT guided the whole project. All authors approved the manuscript. Acknowledgements ================ This research was supported in part by the National Science Council under grant NSC 98-2221-E-155-060-MY3. We especially thank the BMC reviewers for their valuable comments, which helped us improve the quality of the paper.

PubMed Central

2024-06-05T04:04:19.575742

2011-2-23

PMC3053257

Background ========== Problem Description ------------------- Mendelian syndromes are monogenic hereditary diseases, each manifesting as a combination of clinical signs and symptoms (phenotypes). Syndromes can differ from each other by as few as one or two phenotypes. Not all patients display all of the known phenotypes, even if they all have the same syndrome; for example, for the Cystic Fibrosis syndrome, only 10% of the patients display the \"*meconium ileus*\" phenotype \[OMIM 219700\]. Genetic Mendelian syndromes are specifically associated to a gene or group of genes. The most important curated, comprehensive, reliable, and updated source of information in human genetics is the Online Mendelian Inheritance in Man (OMIM) databank \[[@B1]\], which consists of some 17,000 detailed entries on human genes and genetic disorders including 4,802 syndromes with a Clinical Synopsis section (described below). This database is routinely used by clinicians to diagnose patients and is also used widely for data mining purposes \[[@B2]-[@B4]\]. Each OMIM entry contains free text descriptions of: genetic loci, inheritance patterns, allelic variants, biochemical and clinical features, and molecular and population genetics. Most OMIM syndrome entries also include a Clinical Synopsis (CS) section in structured text that outlines signs and symptoms (phenotypic features) accompanying the disease. The phenotypes are presented within one or two levels of contextual information: the first describing the heading or body system (such as \"Head\", \"Heme\" or \"Labs\") and an optional subheading (for example \"Eyes\" may come under \"Head and Neck\" and \"Airways\" comes under \"Respiratory\" in \[OMIM 272800\]). Two problems with searching and mining OMIM are that the same phenotype can be described using very different phrases and that negations are used widely to indicate lack of a phenotype. The similarity between two syndromes can be calculated using the number of shared phenotype phrases. Regrettably, information in the CS sections is not represented in a uniform manner. No controlled vocabulary is used for phenotype and location names. Typing errors and synonyms for the same entity, as well as different nomenclature for overlapping concepts, are observed in large numbers: \"*trunk ataxia*\" vs. \"*truncal ataxia*\", \"*hypoplastic radius*\" vs. \"*hypoplastic radii*\", \"*brachymetacarpalia*\" vs. \"*brachymetacarpals*\", \"*cystic renal dysplasia*\" vs. \"*renal dysplasia cystic*\" vs. \"*renal cystic dysplasia*\" and many more. Negation phrases are phenotype phrases describing symptoms the patients do not have, therefore adding mistakes to search results. For example, when searching for \"*loss of vision*\", we are not interested in retrieving syndromes containing the description \"*no significant loss of vision*\". This further confounds calculating the similarity between two syndromes since sharing a negation phrase does not mean the syndromes are similar. Previous Work ------------- OMIM has already been used as a data source for data mining in previous research. Freudenberg *et al*. \[[@B5]\] manually processed syndrome data in OMIM and collected 5 fields of information for each syndrome (episodic, etiology, tissue, onset and inheritance). The syndrome information was used to find similarity between syndromes for prediction of gene involvement in diseases. Their work was performed manually and the results capture only limited aspects of the syndrome descriptions. In GFINDer \[[@B4]\], OMIM syndromes are used for enrichment of gene lists obtained with microarray experiments. Cantor and Lussier \[[@B2]\] used phenotype headings for clustering syndromes. Only the headings and sub-headings were used due to the problematic nature of the free-form clinical synopsis information in OMIM. In this work, we address this limitation by analyzing the free-form text of the phenotypes. In 2006, Driel *et al*. \[[@B6]\] have created an algorithm for automatically mapping the OMIM descriptions to MESH (Medical Subject Headings) \[[@B7]\] terms. This mapping also aims at normalizing the free-form OMIM phenotype descriptions. The objective of this processing was similar to ours: to discover connections between phenotypes and syndromes. The MESH controlled vocabulary does not contain all the concepts appearing in OMIM. As a result, this approach leads to significant loss of data in OMIM, which has no MESH terms associated with it, as well as some of the phenotype context. Unified Medical Language System (UMLS) \[[@B8]\] is a larger medical language resource provided by the NLM. It consists of a collection of medical vocabularies. UMLS MetaMap \[[@B9]\] maps texts into UMLS concepts. Lage *et al*. \[[@B3]\] used UMLS MetaMap \[[@B9]\] for mapping syndromes to UMLS terms (instead of MESH). Their aim was to use syndrome similarity in order to predict disease gene association. This approach produces limited success, as many of the associated syndromes found were simply variants of the same syndrome (\"*Waardenburg syndrome type IIA*\" and \"*Waardenburg syndrome type IIB*\") or poor quality association. For example, \"*Tuberous sclerosis (TS)*\" and \"*Chordoma*\" were found as similar in their analysis. This association is based on reports of both diseases found in 3 patients. This association was not described in the Clinical Synopsis, as it is likely a random association due to the relative large number of TS patients. Further comparison to this work is not possible as the full results of the OMIM analysis by Lage *et al*. \[[@B3]\] were not published. The methods of Driel *et al*. (2006) \[[@B6]\] and Lage *et al*. (2007) \[[@B3]\] both ignore the context in which terms occur in the OMIM entries and do not detect negative contexts around term entities (negation detection such as NegEx \[[@B10]\]). Neither method provides a practical software system for searching OMIM. UMLS MetaMap performance in extracting biomedical terms was examined in a few articles. Pratt and Yetisgen-Yildiz \[[@B11]\] compared MetaMap to three human experts and found precision of 55% and recall of 93%. Chapman *et al*. \[[@B12]\] examined MetaMap in the domain of respiratory findings in emergency department reports and reported a precision of 56% and recall of 72% for clinical terms. Meystre and Haug \[[@B13]\] examined results using MetaMap with all of UMLS or when defining a subset of UMLS relevant to 80 medical conditions they were interested in. The default data set produced precision of 76% percent and recall of 74% while the customized subset produced similar precision (69%) with an improved recall of 90%. Another system for identifying UMLS terms in a document is BioMedLEE by Lussier and Friedman \[[@B14]\], with reported precision of 89% and recall of 77%. Unfortunately BioMedLEE was not available for comparison with our work. Robinson *et al*. (2008) \[[@B15]\] manually created a database of phenotypes and syndrome relations from OMIM, reported in 2010 \[[@B16]\] to contain more than 9,500 phenotype phrases built into ontology. Our method is automatic and captures more phenotype phrases (almost 3 times as many phrases) in a more superficial manner (as manual clustering is more precise). S2G (Syndrome to Gene) \[[@B17]\], is a tool for finding candidate genes for hereditary syndromes in suspect loci. S2G is comprised of two parts: one for ranking genes in a locus in comparison to a known gene causing the syndrome, the other, for syndromes with no known genes uses our search application for choosing a known gene causing the most phenotypically similar syndrome. Our Approach ------------ Recognizing which phenotype phrases denote similar phenotypes and recognizing negation phrases helps searching (we can match the query term to all the variations under which it appears in the database) and provides critical input to further data-mining of the rich information stored in OMIM. Our objective in this work is to pre-process the natural language descriptions that appear in OMIM to identify synonymous phrases denoting the same phenotype. For example, we want to identify that \"*unilateral kidney agenesis*\" and \"*unilateral renal agenesis*\" both denote the same phenotype. Once we have identified such clusters of synonymous phrases, we can normalize the natural language descriptions in OMIM: we can recognize that two syndromes include the same phenotype even though their natural language descriptions are different. In our approach, we also attempt to use MetaMap and the alignment it can generate between the raw OMIM phrases and UMLS terms (as done by \[[@B3]\]). But this data provides only partial categorization and we found it not to be robust. We do not assume a priori that OMIM terms will be found in an existing ontology or controlled category (such as MESH or UMLS). In contrast, we identify clusters of similar phrases (and combine several clues to determine similarity). Our approach is close, in terms of methodology, to the WordNet (a lexical database for the English language) approach to thesaurus construction: WordNet does not assume a priori that a sense hierarchy exists. Instead, it identifies classes of words that share a similar sense by grouping them in the same class. These classes are called synsets. Similarly, in our method, the clusters of similar phenotype phrases emerge through the computation of pair-wise similarity. We find in our experiment that this method is more robust to the noisy data observed in a large corpus such as OMIM, which evolves over a long period of time and is maintained by a variety of experts. Precision found in predicted synonymous phrase pairs was 93.5% (500 pairs of phrases were sampled randomly from the resulting phrase clusters and evaluated by a geneticist as similar or non-similar). 10% of the phrases were recognized as negations with a precision of 89% with most false positive being ambiguous. Using the discovered phenotype clusters, we have created a Web application for searching OMIM, called CSI-OMIM, which provides much more efficient search than the original OMIM site. CSI-OMIM supports incremental search in the following manner: the clinician enters free text description for a phenotype. The application displays a list of matching phenotype clusters found by partial string matching to the query. The user can select the best matching clusters and continue searching for more phenotypes. At any time, the user can also search for the most similar syndromes matching the list of selected phenotype clusters. CSI-OMIM obtains high recall in the search for phenotypes, because it captures the wide variability found in the OMIM free-text descriptions. It also has improved precision (with respect to the original OMIM search engine) because it avoids retrieving phrases in negative contexts. Finally, the matching from phenotype clusters to syndrome is highly effective, because it uses the reliable syndrome-similarity measure (cosine similarity over the all the phenotypes) computed over the phenotype clusters. In the rest of the paper, we present our method to compute clusters of similar phenotype phrases from the various OMIM free-text descriptions, using natural language processing methods. We then present an evaluation of the quality of the acquired phenotype clusters and of the syndrome similarity measure induced from the clusters of similar phenotype phrases. Construction and content ======================== Data Collection --------------- To prepare our analysis, we extracted and formatted data out of the OMIM database according to the following method. All Clinical Synopsis (CS) entries of the syndromes were used to build a database of syndrome phenotypes; the OMIM text data is obtained from the NCBI site (see details in Additional File [1](#S1){ref-type="supplementary-material"}) and the system is updated every three months using an automatic process. The updated data, OMIM data used for the analysis in this paper was obtain on October 13, 2010, is automatically processed using the process described below (parsing using a Context Free Grammar and annotating semantic types using MetaMap. We constructed manually the CFG grammar used to parse the text, but the parsing is performed automatically on each update). Defining Phenotype Areas ------------------------ The contextual information of a phrase (its heading and sub-heading) is crucial for phenotype phrase comparison. Only phrases with similar context should be compared to avoid false positive results. The list of headings and sub- headings is not strictly maintained in OMIM. The same heading appears in different syndromes under slightly different names. Phenotypes can also appear directly below headings, without a sub-heading refinement. This variability makes it difficult to systematically exploit the contextual information provided by the heading/sub-heading classification. To address this lack of consistency, we manually defined 26 areas. We selected the areas to best describe the large and overlapping number of domains and sub domains. For example: the \"Head and Neck\" heading includes both the \"Head\" and \"Neck\" domains. \"Head and Neck\" and \"Head\" are both included in area 17 - \"Head\", area 16 - \"Neck\" includes \"Neck\" and \"Head and Neck\".Table [1](#T1){ref-type="table"} in the Results section shows the list of areas we constructed. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Areas of phenotypes identified ::: Area Name Area\# \#Distinct phrases \#Clusters identified Avg Similar Phrases Cluster Size \% Phrases clustered in area ---------------- -------- -------------------- ----------------------- ---------------------------------- ------------------------------ Syndrome names 1 4,801 278 3.45 19.9 Abdomen/gi 2 707 33 2.3 10.1 Respiratory 3 546 39 2.17 15.6 Gu/renal 4 985 44 2.25 10.0 Gu/genitalia 5 975 44 2.25 10.0 Cardiovascular 6 808 38 2.13 10.0 Muscle 7 1143 68 2.98 17.8 Endo 8 344 16 2.25 10.4 Neuro 9 4.721 247 2.68 14.0 Oncology 10 891 21 2.62 6.17 Heme 11 576 16 2.75 7.6 Immune 12 542 19 2.05 7.2 Eyes 13 2,265 152 2.5 16.8 Face 14 3,927 283 2.56 18.4 Teeth 15 3,542 258 2.48 18.1 Neck 16 3,490 252 2.49 17.9 Head 17 4,196 281 2.51 16.8 Limb 18 4,522 283 3.15 19.7 Skel 19 3,936 247 3.02 18.9 Chest 20 4,307 273 3.03 19.2 Growth 21 497 37 4.24 31.6 Nails 22 1,617 106 2.41 15.8 Skin 23 2,098 140 2.42 16.2 Hair 24 1,701 116 2.41 16.5 Lab 25 3,553 106 2.9 8.6 Misc 26 4,021 Was not clustered ::: Similarity Computation ---------------------- ### Combining Cues to Compare Phrases The task we now approach is the following: given a set of phrases within a single area, how can we determine the level of similarity between any two phrases. For example, how do we recognize that \"*flaring of the iliac wings*\" and \"*flared iliac wing*\" are similar, but \"*flared metaphyses*\" and \"*flared iliac wing*\" are not. The phrases observed in the OMIM dataset are all noun phrases. On average, they include 3.9 words; with a maximum of 64 words for the most complex phrase (found in syndrome 255125 in OMIM). Syntactically, the phrases exhibit structures such as Noun, Adjective Noun, and a few Prepositional Phrases. Figure [1](#F1){ref-type="fig"} illustrates the distribution of the length of the Noun Phrases observed in the OMIM Clinical Synopses. Most of the phrases (80% out of 67,470 phrases) include between 2 and 6 words (see Figure [1](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Distribution of phrase instances by number of words**. Distribution of length of Noun Phrases in OMIM Clinical Synopsis. :::  ::: Our intuition when comparing noun phrases is to combine several sources of knowledge about the phrases to decide on their similarity. First, because the vocabulary used in OMIM is complex and not fully consistent (spelling variations and usage of synonyms are both common), we attempt to align each word with a known controlled vocabulary (we use UMLS and MetaMap). Second, because phrases exhibit complex syntactic structures (2 to 6 words), we attempt to parse the phrases into trees with explicit modification relations. Third, because there can be slight spelling variations in the phrases, we also use a simple edit-distance metrics to compare words. In the rest of this section, we first describe each of the components of our analysis: alignment of words with UMLS concepts; parsing; and edit-distance computation as well as detection of negations. We then describe the method used to combine these elements into a similarity computation for any pair of phrases. ### Mapping Phrases to UMLS UMLS terms can be used to describe a large proportion of the words or entities in the phenotype phrases. A term may be a pathology name like \"*obesity*\", a modifier such as \"*mild*\", an anatomy location such as \"*wrist*\" or a multiple word entity like \"talipes *equinovarus*\". When we find them, UMLS terms help us recognize multiple word entities and classify terms and words into categories such as pathology, modifier, anatomy or named entity. We used MetaMap Transfer \[[@B9]\] (version 2.4 with 2006 AA UMLS version using the default parameters) to annotate all the phrases found in the OMIM Clinical Synopsis with UMLS information (we use default settings, this stage is automatic and takes about 12 hours per update). MetaMap produces UMLS annotations and performs two major tasks: 1\. It identifies multi-word expressions as single terms. For example \"*congenital abnormality*\" is tagged as a single token. 2\. It classifies each term according to the UMLS controlled vocabulary. For example, \"*strokes due to coagulopathy*\" is mapped to \"Strokes \[CUI 0038454/Disease or Syndrome\] and Coagulopathy \[CUI 0005779/Disease or Syndrome\]\". Not all term classifications computed by MetaMap are correct. Previous work indicates precision as low as 0.69 \[[@B12]\] or 0.56 \[[@B13]\] if no precaution is taken. For example, \"*adams stokes attacks*\" is mapped to \"*adams*\" in category: \"*Health Care Related Organization*\"; in \"*functional defects in the cortical and subcortical motor related areas of the frontal lobe*\", \"*cortical*\" is mapped to \"*BARK (ADRBK1 gene)*\" in category \"*Gene or Genome*\". Such associations would introduce too much noise in our phrase similarity computation. Low precision stems from the very large number of different terms in UMLS and the variety of vocabularies combined in UMLS (over 30 distinct vocabularies are merged in the free version of UMLS). In many applications, the precision of MetaMap is improved by focusing on a subset of the UMLS terms, and defining smaller collections of terms in specific domains within UMLS. As OMIM phenotypes are extremely varied and cover all domains, such an approach would miss too many of the candidate UMLS concepts. Instead, we rely on UMLS Semantic Categories (such as \"Congenital Abnormalities\") to filter UMLS concepts returned by MetaMap. We manually defined a set of non-noisy UMLS Semantic Types and keep only concepts that belong to those types in the result of MetaMap. For example, terms that belong to the \"Fish\",\"Bird\",\"Organization\" semantic types in UMLS are ignored. Those that belong to the categories shown in Table [1](#T1){ref-type="table"} at Additional File [1](#S1){ref-type="supplementary-material"} are kept. In our application, we exploit the MetaMap UMLS annotations in terms of \"rough semantic\" categories. We manually mapped UMLS Semantic Types into groups according to the functional role of the terms in the phenotype descriptions. This reclassification of UMLS concepts into new semantic groups has been suggested by Fan and Friedman \[[@B18]\]. We divided the concepts into 4 semantic roles relevant for OMIM phenotype phrases using the existing semantic classifications of UMLS.. The semantic roles we created attempt to describe the phenotype structure and include the following 4 categories: pathology, anatomy, named entities and modifiers (See Table [2](#T2){ref-type="table"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Rough Semantic Categories. ::: Rough Semantic Category Description Examples ------------------------- ------------------------------------------------------------- ----------------------------------------------------------------------- Pathology or Finding Names and symptoms of diseases \"Perthes\", \"Hexadactyly\", \"Diffuse atrophy\" or \"Short finger\" Named entities Names of chemicals, functions, microorganisms or proteins \"Actin\", \"Tyrosine\", \"Insulin\" or \"Agglutination\". Anatomy The body part or organ the phenotypes occurs in. \"Cranium bifidum\", \"Thumb\", or \"Distal femur\" Modifiers Concepts describing the phenotype and changing its meaning. \"Absent\", \"Hypoplastic\", \"Mild\", \"Enlarged\" ::: We also use UMLS tagging to reduce the syntactic complexity by recognizing multi-word expressions. We present quantitative evaluation of this simplification in the Results section. Both the chunking of multiple words into single tokens and their semantic categorization help build a relevant grammar for parsing the phrases in a robust manner. ### Parsing The syntactic structure of the phrases is critical to decide on their similarity. Consider the task of comparing these two phrases: \"*fifth finger single interphalangeal crease*\" and \"*single flexion crease of fifth fingers*\". The sub-phrase \"*fifth finger*\" appears at the beginning of the first and at the end of the second - the edit distance between the two strings would therefore be quite high. However, knowledge of the syntactic structure (NN NN vs. NN of NN) and knowledge that the phrase \"*fifth finger*\" is identified as a single token of type \"Anatomy\" can help us compute a more precise similarity between the two phrases. Because the syntactic structure of phrases in OMIM can be complex (on average 3.9 words with many phrases containing more than 6 words), our approach is to try to parse these expressions, and then compare the parse trees, while taking into account the modification relations that exist between the elements of the phrases. As is well known in natural language processing, parsing noun phrases can be complex because the structure of the noun phrase can be ambiguous. For example, \"*altered melatonin secretion*\" may be parsed as (\"*altered melatonin*\" + \"*secretion*\") or (\"*altered*\" + \"*melatonin secretion*\"). Consider the following group of phrases found in OMIM: • *ossification defect of skull* • *absent ossification of skull vault* • *decreased skull ossification* • *deficient skull ossification* Our algorithm concluded they all belong to the same cluster of similar phenotype phrases. Wrong parsing of the phrases could have prevented this clustering. For example: (deficient skull ossification) parsed as (deficient skull) (ossification) would not have been similar to (ossification defect) (of skull). In order to parse Noun Phrases, we must decide on a grammar and a set of categories that annotate words and groups of words within phrases. We use regular parts of speech (verb, adjective, noun, etc) and, in addition to traditional grammar, we introduce as well the UMLS classification of terms into the 4 categories described above: pathology, anatomy, named entity and modifier. If a word or a term is classified by UMLS, we keep its UMLS tag; else we use the parts of speech tagged by a standard parts of speech tagger (we used the LingPipe \[[@B19]\] tagger trained on the Genia corpus \[[@B20]\]). We then manually constructed a Context Free Grammar (CFG) that captures the structure of the phenotype description in the OMIM Clinical Synopsis. The leaves of the parse trees are tokens annotated either by UMLS categories (pathology, anatomy, named entity or modifier) or by parts of speech tags (conjunction, adverb, etc). Note that as discussed above, the tokens annotated by UMLS can cover multiple words. We create the tree using the method of chart parsing, gradually building the parse tree by combining simpler parts of the tree together based on the CFG starting from the labels over the phrase in a bottom-up manner (for more information see \[[@B21],[@B22]\]). We used the Chart parser code from AIMA \[[@B22]\]. Phrases longer than 11 words (1,438 such phrases were found out of 68 K) were not parsed as the parser may not finish parsing in a satisfactory time; these phrases are still compared to other phrases using edit distance (computed with dynamic programming) and still affect the similarity measure of the syndromes containing them. Negation Detection ------------------ Phrases that include negation are not incorporated in any cluster and are not used for computing similarity among syndromes. For example, \"*no significant loss of vision\"*is not clustered together with \"*loss of vision\"*, even though they are syntactically similar. Negation is discovered using regular expressions in the same spirit as NegEx \[[@B10]\]. We extend this approach by relying on the structure of the parse tree using the following rules: 1\. Phrases with the word \"*normal*\" appearing anywhere (not as part of another word such as \"*abnormal*\") are marked as negations. (This is implemented as a regular expression test.) 2\. Phrases where the word \"*no*\" appears are marked as negations only if they contain a \"Pathology\" node in their parse tree (in order to remove sentences describing absent functions or body parts). For example, \"*no dysarthria\"*is marked as a negation, because *\"dysarthria\"*is a pathology but *\"no natal teeth*\" is not marked as a negation. ### Clustering of Similar Phenotype Phrases We compared all pairs of phrases within each area. Altogether, we performed about 77 Million pairwise phrases comparisons (this stage takes about 24 hours using one 3.0 GHZ CPU and is executed each time we update the data from OMIM). We define a set of rules to decide on the similarity of 2 parsed phrases: Successfully parsed phrases (based on our context free grammar described above) are similar if they share synonymous Pathology/Named Entities and Anatomy node terms and are modified by synonymous modifiers. Nodes annotated by MetaMap are compared by comparing their UMLS tag (CUI). Nodes that are not annotated by MetaMap are compared as case-insensitive string comparison. This comparison relies on the parse tree to determine which parts should be compared (for example: nodes annotated in both trees as \"Modifiers\" would be compared to each other when comparing the \"Pathology\" they modify. Both phrases must have a similar \"Pathology\" node for the rest of the nodes to be compared). Phrases without a parse tree were compared using string distance (allowing maximum distance of 2). Results ======= Our objectives are: create cluster of similar phenotype phrase and identify negation phrases. In this section we evaluate each of the steps in our method: (1) classification of phenotype phrases into areas; (2) MetaMap tagging of the phenotype phrases; (3) parts of speech tagging of the phrases; (4) parsing of the phrases according to our CFG grammar; (5)negation detection; and (6) clustering of the phrases. Distribution of phenotype phrases into areas -------------------------------------------- Our method identifies clusters of similar phenotype phrases (*i.e*., phrases have similar meaning) within each area. We originally found 159 distinct headings and sub-headings in the Clinical Synopsis sections in OMIM. We mapped every phenotype phrase to one or more of the 26 areas on the basis of its heading/sub-heading location in OMIM, syndrome names were gathered in a dedicated area (see Table [2](#T2){ref-type="table"} of the Additional File [1](#S1){ref-type="supplementary-material"}). Area 26, \"Misc\", is not used further in clustering experiments. Phrases in this area include: \"*variable phenotype\"*or \"*reduced penetrance*\". This manual classification improved overall the consistency of the rough classification of phenotypes. The granularity we obtain with 25 areas allows us to significantly reduce the complexity of the task of finding synonym phrases: we avoid false positive matches by only comparing phrases within the same area. In addition, the consistent classification in areas improves the precision of UMLS mapping, as has been discussed in the past by Chapman *et al*. \[[@B12]\]. MetaMap Performance ------------------- In the dataset of all Clinical Synopsis phrases in the clustered areas (31,778 distinct phrases out of 67,470 phrases extracted from OMIM.txt), Meta Map tagged 78% of the word tokens and identified on average 1.9 concepts (CUIs) per phrase. MetaMap tagged 41,566 concept instances in the dataset covering 104,673 words. Many of the identified concepts include more than one word (33,591 concept instances of more than one word were identified). The grouping of words into multi-word tokens by MetaMap significantly reduces the syntactic complexity of the phrases: before the MetaMap treatment, phrases have an average length of 3.86 tokens (standard deviation of 2.02); after the MetaMap tagging, phrases have an average length of 2.48 tokens (with standard deviation of 1.4). For example: \"*bowed radius*\", \"*bowing of the radius*\", \"*bowing of radius*\" and \"*bowed radii*\" are all changed to \"*bowing radius*\". We did not measure explicitly the precision of the MetaMap annotations after filtering according to UMLS Semantic Types. Parsing Results --------------- The parse trees we obtained have an average height of 3.75 and size of 5.3 nodes on average. The parser found an acceptable parse tree (*i.e*., one matching our grammar) in about 90% of the phrases. The percentage of phrases parsed successfully varies between the different areas (84% - 93%). In one area (25 - lab) parsing encountered problems due to the different structure of the phrases which includes many prepositions in the same sentence and a large number of named entities. Figure [2](#F2){ref-type="fig"} shows example parse trees that demonstrate the multi-word tokens tagged by MetaMap. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Parsing result**. a) The phrase \"brainstem hypoplasia\" parsed. The token \"brainstem\" was recognized by MetaMap as UMLS term (CUI C0006121) of semantic type \"Body Part, Organ, or Organ Component\" and the token \"hypoplasia\" was identified as CUI C0243069 of semantic type \"Pathologic Function\". b) Parse tree of the phrase \"mri shows brainstem hypoplasia\", same concepts were recognized as in (a), \"mri\" is marked as a noun and \"shows\" is marked as a noun as well since we view the phrases as noun phrases without verbs. c) Parse tree of the phrase \"hypoplasia of the brainstem\", the entire phrase is reduced by MetaMap to the two concepts identified in (a), only in reversed order. :::  ::: Negation Detection Results -------------------------- 860 distinct phrases were identified as negation phrases (3.2% of the phrases in the applied areas: 2 - 25). Precision (the fraction of correctly identified negations out of all identified negations), measured by manual inspection of 10% of the negation phrases (chosen randomly and tagged manually as negated or not-negated), was 89% with most false positive phrases being ambiguous for human readers as well (\"*low to normal IQ*\", \"*plasma testosterone is normal or increased*\"). Measuring the recall (number of negation phrases identified/number of negation phrases in the database), in the absence of a well annotated examples set, is too difficult. When searching OMIM using our website, negation phrases are marked by italics. These phrases are also not included in the similar phrase clusters (\"*loss of vision*\" is not grouped with \"*no significant loss of vision*\" even though both contain the phenotype \"*loss of vision*\") see Figure [3](#F3){ref-type="fig"}. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **CSI-OMIM: Negation detection**. Negations are marked with italics and are ignored in the clustering process. :::  ::: Clustering of Similar Phenotype Phrases Results ----------------------------------------------- Our algorithm groups similar phrases into clusters that are recognized as synonyms (see Figure [4](#F4){ref-type="fig"}). We obtained altogether 1,680 clusters of similar phenotype phrases covering 4,551 distinct phrases with on average 2.7 phrases per clusters of similar phenotype phrases. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **CSI-OMIM: Clusters of similar phrases**. Clusters of similar phrases results of \"white matter abnormalities\" search. :::  ::: Clusters of similar phenotype phrases cover 13,009 phrase instances (about 20% of the overall phrase instances in the clustered areas). The distribution of clusters per area that we eventually obtained is shown in Table [1](#T1){ref-type="table"}. For example: \"*hyperreflexia in the lower limbs*\" is clustered with \"*hyperreflexiaespecially of the lower limbs*\", for two reasons: a) in the parse tree of both the \"Pathology\" node is similar b) both nodes are connected to a similar \"Anatomy\" node. In Figure [2](#F2){ref-type="fig"} we see 3 phrases which are clustered together due to an identical PathlogyNounPhrase sub-tree in the parse tree of phrases (a) and (b), and a PathlogyNounPhrase sub-tree with identical Pathology and Anatomy between (c) and (a). Precision found in predicted synonymous phrase pairs was 93.5% (one thousand pairs of phrases were sampled randomly from the resulting phrase clusters and evaluated by a geneticist as similar or non-similar). Recall was not measured due to the complexity of the problem and lack of an expert annotated corpus. The clusters obtained through our method are used in two ways: when a user enters a term that is recognized as a member of a cluster, the system searches for all variants of the term as well. For example, if the user searched for \"*thickened cranium*\", results that include \"*thickened skull vault*\" and \"*thickened cranial vault*\" will also be retrieved. We measure the quality of the acquired clusters by measuring the improvement they bring to a data mining application. We want to compute a similarity measure between syndromes. To this end, we compare the Clinical Synopsis section of the syndromes, after they have been parsed and normalized - that is, all occurrences of phrases that belongs to the same cluster, are replaced by the same term. For example, the syndrome \"CUTIS MARMORATA TELANGIECTATICA CONGENITA\" (OMIM 219250) is normalized \"*bowed legs*\" replaced by the cluster: \"*bowing of the legs*; *bowed legs*; *bowing of legs*\". We then compute the cosine distance of the normalized syndromes and compare the results when computing the cosine distance without normalization. Cluster normalization increased the number of similar phenotypes among syndromes: 79,770 new connections between syndromes were discovered among 4,802 syndromes, adding 16 new connections (shared pehnotypes) per syndrome on average. For example: the aforementioned syndrome 219250 is now connected to \"BOWING OF LEGS, ANTERIOR, WITH DWARFISM\" (OMIM 112350) through the phenotype \"*bowing of the legs*\" due to the aforementioned cluster. Conclusions =========== Our work focused on pre-processing the textual component of the OMIM dataset using Natural Language Processing methods, and produced clusters of similar phenotype descriptions. These similar clusters allow us to propose an enhanced search system of the OMIM database and to perform effective data-mining on the OMIM textual descriptions and discover similarities among syndromes. Our work unifies parts of the methods used by Lage *et al*. \[[@B3]\] and Van Driel *et al*. \[[@B6]\] and improves the precision of the results by dividing Clinical Synopsis phrases into areas and using MetaMap to get both UMLS and MESH information. We provide a robust application capable of being updated regularly and offering a wealth of organized phenotypic data of OMIM syndromes. Precision found in predicted synonymous phrase pairs was 93.5%. 3.2% of the phrases were recognized as negations with a precision of 89%. We identified 79,770 new connections between syndromes - on average 16 new connections per syndrome. Using the new curated database, we provide an online search application for clinicians with improved search accuracy. The improved data is used to find similarity between syndromes in order to find candidate genes for hereditary syndromes as part of the S2G application \[[@B17]\]. The enhanced OMIM database we produced can be further used for bioinformatics purposes as a basis for identifying connections between syndromes. Another product of our work is an accurate connection of syndromes\' phenotypes to the UMLS and MeSH, this can be applied further in the development of literature mining tools and search applications. The database is available in our website (in the FAQ section). Our system contains over to 31 thousand phenotype phrases in comparison to 9,500 in the Human Phenotype Ontology (HPO) \[[@B15],[@B16]\]. Our method may be used as a way to augment HPO by suggesting synonyms or improve syndrome comparison by including terms which are not yet in the vocabulary but exist in OMIM. Another aspect of our work is the principle of detecting similarity between phenotype descriptions that can be applied to vocabularies using the semantic roles we defined as anchors for comparison. These rough semantic categories (Pathology or finding, Anatomy or body area, Modifiers and named entities, see Table [2](#T2){ref-type="table"}.) can assist in detecting similarity of concepts by identifying important cues in the phrases: pathologies and anatomy region and named entities (in this order), and then comparing the modifiers. This distinction can enhance both statistical driven methods for similarity computation by measuring co-occurrence in documents and construction and improvement of manually built ontologies such as HPO. Our pre-processing allows finding similarity between syndromes based on different measures of the descriptions using the rough semantic categories. For example similarity can be calculated based only on anatomical location terms or only pathology terms. Syndrome similarity can also be calculated using only a subset of the anatomical areas such as: similarity based only on the \"Hair\", \"Nails\" and \"Skin\" areas. Authors\' contributions ======================= RC and AG contributed to the design of the study equally. RC carried out the design of the comparison algorithm the implementation and drafting the article. ME contributed in evaluating the results and drafting the article. OB coordinated the study and helped draft the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Supplementary Material**. Details of data acquisition, UMLS division into rough semantic categories and division of heading/sub-headings into areas. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The authors deeply thank the Morris Kahn Family Foundation for making this study possible.

PubMed Central

2024-06-05T04:04:19.579374

2011-3-1

PMC3053258

Background ========== MicroRNAs (miRNAs) are \~21 nucleotides long sequences, which are endogenously generated both in plants and animals. They are one of the key players in gene regulation, typically inhibitory in nature, and act either at the post-transcriptional level (by triggering target messenger RNA degradation) or at the translational level (by inhibition of translation) (see review by \[[@B1]\]). In plants, miRNA genes are initially transcribed as a primary miRNA sequence (i.e., pri-miRNA), which folds into a hairpin loop with small overhanging regions at both ends. The pri-miRNAs are then processed into hairpin sequences (i.e., precursor miRNA) by a riboendonuclease, named DICER. The DICER protein removes the loop region from the hairpin, and the remaining duplex is transported out of the nucleus, where the complementary sequence (also called star) is removed to allow mature miRNA to be ready for action \[[@B2]\]. Various tools have been developed to discover conserved and/or novel miRNAs. Initially, the miRNAs have been discovered by direct cloning and sequencing \[[@B3],[@B4]\], which suggested a high degree of sequence conservation of miRNA across species \[[@B5]\]. With the availability of a complete genome sequence, it is possible to use computational approaches to discover the miRNA homologs. Recent progress in high-throughput sequencing further enabled genome-wide discovery of miRNAs, including novel miRNAs, and their expression profiling. Over the past few years, several web-servers and standalone applications, analyzing deep sequencing data for miRNA discovery and/or expression profiling, have been reported; the web-servers include miRCat \[[@B6]\], miRAnalyzer \[[@B7]\], miRTools \[[@B8]\], whereas the standalone tools include miRDeep \[[@B9]\], miRExpress \[[@B10]\], MiroPipeline \[[@B11]\], etc. In addition to the above categories, there have been few miRNA discovery studies, reported in species, like *Arabidopsis*\[[@B12]\] rice \[[@B13]\], maize \[[@B14]\], and human \[[@B15]\], largely carried out using in-house scripts. There are two major limitations with these tools or protocols: 1) several of them use known miRNAs to identify only the homologs present in deep-sequencing data, thus limiting the scope of discovery of *novel*miRNAs, despite the availability of complete genome sequence; 2) several others are available as web-servers, which constrains the analysis in terms of upload/analysis time for larger datasets and also by limitations in the choice of reference genomes. In contrast, miRDeep does not have these constraints. In addition, miRDeep scores the predictions, which makes it easier to pick the better candidates from the rest \[[@B9]\]. The scoring used in miRDeep is based on computation of posterior probabilities of statistical features, like minimum free energy (MFE), conservation of the core region of mature miRNA, etc.. Some of these features are commonly used by other tools/analyses (see Methods for details). The posterior probabilities are usually computed based on parameters derived from a known set of real and background precursors; the current parameter set used in miRDeep is from a nematode, *C. elegans*, and reported to be effective for other animals as well. However, there are major differences between the properties of plant and animal miRNAs. For example, the maximum length of plant miRNA precursors can be about \~900 nt long; the extent of pairing and bulge size of duplex (of mature and star) in plants is very different from that of animals \[[@B2]\]. Therefore, the differences in properties between plant and animal miRNAs can affect the parameters required for statistical scoring used in miRDeep. In the current study we examined the key differences between plant and animal miRNAs, and their effects on general and miRDeep-specific statistical features, and further estimated the plant specific parameters of these features. With these new miRDeep parameters, we validated the prediction by applying it to a set of newly discovered miRNA in maize seedlings. The results were compared to that obtained using default miRDeep parameters. Results ======= Frequency distribution of Minimum Free Energies of plant miRNA precursors ------------------------------------------------------------------------- The frequency distribution of minimum free energies (MFE or *abs*) of real plant and animal miRNA precursors showed large differences: plant miRNA precursors showed a broader distribution and a lower mean MFE compared to that of animals (Figure [1](#F1){ref-type="fig"}). An underlying factor likely to be accountable for this difference is the length of precursors. While the length of animal miRNA precursors mostly lie in the range of 45-215 nt (mean = \~87 nt), those from plants show large heterogeneity, lying in the range of 55-930 nt (mean = \~146 nt). An examination of a relation between precursor length and MFE showed a linear relation between them, with the standard deviation of MFE, interestingly, increasing with length (Additional file [1](#S1){ref-type="supplementary-material"}: Figure 1). The effect of length on MFE was also evident when the cumulative frequency distribution (instead of mean) of MFE was plotted (Figure [2A](#F2){ref-type="fig"}). The distributions were distinct for precursors of different length. When these distributions were compared with that of background precursors of corresponding length (maize genome has been the source of background precursors unless specified), they were largely overlapping, except in the left tail region (Figure [2A](#F2){ref-type="fig"}). This indicated that the minimum free energy may not be a strong discriminant between real and background precursors of plants. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Frequency distribution of Minimum Free Energy of all known miRNA precursors of plants and animals**. :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Behavior of Minimum Free Energy distribution in plants**. (A) Cumulative distribution of MFE of real and background precursors of three different lengths: 95-105 nt, 180-190 nt, and 260-270 nt. The Bgr indicates background precursors (from maize genome). (B) Cumulative distribution of MFE of background precursor of size 260-90 nt from five plant species. The source species does affect the nature of distribution, and they tend to cluster into monocots and dicots. While the differences in distributions within the group were insignificant (at confidence level \<0.05), those from two groups were highly significant. (C) Same as in A, but the MFE has been normalized by its length. For clarity the distribution of only two values of length have been shown. (D) Empirical distribution of log-odds score of MFE and its best fit theoretical approximation by a modified sigmoid function. The probabilities for background precursors, used in computing log-odds, were average of that of three monocots, namely *Z. mays*, *S. bicolor*, and *O. sativa*. :::  ::: In order to examine if the source (species) of precursor sequences has any bias on the above observation, the MFE distribution of real and background precursors from four additional plant species (with complete genome sequence), namely sorghum (*S. bicolor*), rice (*O. sativa*), *Arabidopsisthaliana*, and *Medicagotrancatula*, were compared. While the MFE distributions of background precursors of sorghum and rice converged to that of maize, the distributions of *Arabidopsis*and *Medicago*were very distinct (Figure [2B](#F2){ref-type="fig"}). This led us to speculate that differences possibly exists at the level of monocots and dicots. For real precursors, although differences did exist between distributions of monocot and dicot species, however, these were statistically insignificant (at confidence level ≤0.05) (data not shown). These observations imply that the miRDeep training, in a plant species, is largely independent of the choice for the real precursors, but sensitive for the background precursors. This also affects the log-odds of MFE in dicots, as the MFE distributions show large differences between real and background (Additional file [2](#S2){ref-type="supplementary-material"}: Figure 1), as against the monocots (Figure [2A](#F2){ref-type="fig"}). The multiple distributions, characterizing MFE of plant miRNA precursors (Figure [2A](#F2){ref-type="fig"}), posed practical difficulties in employing them in a miRNA prediction algorithm. Given the linear dependency between the MFE and the length of precursors, normalization of MFE by the length proved a better solution to the problem of multiple distributions. As shown in Figure [2C](#F2){ref-type="fig"}, the cumulative frequency distributions, of precursors of varying length, overlapped almost completely. For theoretical approximation of these distributions, \'Gumbel\' or \'Extreme-value\' distribution functions were the ideal choice, as the MFE values are \'minimum\' over all possible free energies. On fitting these functions, the sum of errors, for the best fit curve, was within the allowable limit of 10% (Additional file [3](#S3){ref-type="supplementary-material"}: Figure 1). However, the quality of fit with empirical distributions in the left tail region was poor (in Additional file [3](#S3){ref-type="supplementary-material"}: Figure 1, note the region of the plot for normalized MFE values ranging between -0.4 to -0.1). A better approximation of left tail regions is desirable, as it has the potential to contribute larger scores. As an alternative the log-odds scores were directly computed from the two distributions, and were approximated by a modified sigmoid function, which fitted the left tail of curve very well (Figure [2D](#F2){ref-type="fig"}). The new parameters are listed in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### List of miRDeep parameters estimated for plant specific miRNAs. ::: ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Feature Original parameters of miRDeep Plant specific (monocot) Dicot specific (if any) -------------------------------------------- ---------------------- ---------------------------------------------------------------------------- -------------------------------------------------- ------------------------------------------------- **MFE** Known Cumulative Distribution Function: $F(x) = e^{- e^{(x - location)/scale}}$\ Log-odds score: $f(x) = \frac{a}{(b + e^{x*c})}$ Log-odds score:$f(x) = \frac{a}{(b + e^{x*c})}$ Location = 32; Scale = 5.5 Background Cumulative Distribution Function: $F(x) = e^{- e^{(x - location)/scale}}$\ a = 1.339e-12\ a = 4.46e-4\ Location = 23; Scale = 4.8 b = 2.778e-13\ b = 9.125e-5\ c = 45.843 c = 26.929 **Stability (log-odds)** Stable 1.6 1.37 0.63 Unstable -2.2 -3.624 -3.17 **Nucleus conservation (log-odds)** Conserved 3 7.63 Non-conserved -0.6 -1.17 **Excision length** 140 nt 300 nt **Paired** Total ≥14 ≥15 nt **Unpaired** Total NA ≤5 nt Consecutive unpaired NA ≤3 nt **Bulge** Total could be as high as 5 nt ^\#^ ≤2 nt **Maximum multiple hits of deep-seq read** 5 20 ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- The cumulative distribution of MFE of known and background precursors, in miRDeep algorithm, are represented by Gumbel distribution functions, where *F(x)*is the cumulative frequency of MFE less than or equal to variable *x*, and *location*is used for mean centering the distribution of MFE. On the other hand, in the parameterized algorithm, log-odds of MFE were directly obtained from the distributions of real and background precursors, and then represented by a modified sigmoid function where *f(x)*denotes log-odds of a given normalized MFE. ^\#^The size of bulge was estimated based on allowed difference between length of mature and star sequences in miRDeep. The bulges often appear when one of the duplex sequences has additional bases than its counterpart. ::: Stability of secondary structure of plant miRNA precursors ---------------------------------------------------------- Another feature which distinguishes real precursors from background precursors is the marked difference in secondary structure when the sequence of a candidate precursor is shuffled (referred henceforth as stability). Bonnet *et al*. \[[@B16]\] reported that MFE of real miRNA precursors significantly differ from their shuffled counterparts. The miRDeep implements this feature through the use of a tool RANDfold \[[@B16]\] to distinguish real and background precursors by examining the stability of secondary structure on shuffling. RANDfold first generates an ensemble of shuffled sequences, either by mononucleotide or dinucleotide shuffling, and computes the frequency (*p-value*) of shuffled sequences exceeding MFE of the candidates in question. So a *p-value*of 0.01, for instance, signifies that only 1% of the shuffled sequences have similar or better folding than the candidates. Typically, those candidates with a *p-*value of up to 0.05 are categorized as *stable*. For the real precursors, the RANDfold was run on a sample of 500 precursors, sampled across all lengths, which yielded \~0.98 of them as stable (Figure [3](#F3){ref-type="fig"}). Like the MFE distribution, the length of miRNA precursor may also affect the stability of structures. In order to know how longer precursors behave, this program was executed on precursors with length ≥ 400 nt. There was a slight, but statistically insignificant, decline in the frequency of stable precursors (data not shown). On the contrary, the frequency of stable background precursors differed substantially for varying length. RANDfold was run on background precursors of three distinct lengths- 100, 200, and 300 nt. As Figure [3](#F3){ref-type="fig"} demonstrates, the frequency of stable precursors varied considerably: the frequency for 100 nt long background precursors was slightly lower (0.8), the same declined drastically for longer background precursors (ranging from 0.18-0.25). Considering the excision length (details in later section), the frequency for the \~300 nt long background precursors was chosen as the parameter for statistical scoring (Table [1](#T1){ref-type="table"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Cumulative frequency of *p-value*of real and background miRNA precursors**. Three different lengths were examined: 100, 200, and 300 nt. At a threshold of ≤ 0.05, there was small difference in frequencies for real and background of length 100 nt. However, with increase in length of background precursors, the p-value declined drastically. :::  ::: We further examined whether the choice of source of background precursors affected the stability. As in the case of MFE, the stability of the precursors differs substantially between monocots and dicots (Additional file [4](#S4){ref-type="supplementary-material"}: Figure 1). The dicot background precursors were found to have a higher fraction of stable structures (for example, 50-55% for 300 nt size range), as opposed to their monocot counterparts (only 20-25% for the corresponding length). Therefore, in dicots, the higher frequency of stable background precursors will make the distinction between real and background precursors more blurred. Core conservation in mature miRNAs of plants -------------------------------------------- The mature miRNAs generally show high conservation within a family. Moreover, for a particular miRNA family, conservation is also observed across species within (plant or animal) kingdoms. Apart from nucleotides, the conservation also shows positional pattern, that is, certain positions within mature miRNAs are consistently conserved even across miRNA families. This may have some distant implications on the nature of binding of miRNA to its target, as it is also position specific. In animal miRNA families, positions 2-8 are often regarded as the nucleus or core region, and the same has been implemented in miRDeep algorithm. However, given the better complementarity between mature and star sequence in plants \[[@B2]\], we examined the occurrence of any deviation in the positional conservation profile. A dataset comprising 18 plant miRNA families were randomly sampled, such that each family had representatives from at least four species. Results showed that the nucleus conservation pattern in plants is quite different from that in animals (Figure [4](#F4){ref-type="fig"}). Considering 75% as the threshold for positional conservation across species, two distinct conservation blocks were identified: 2-13 and 16-19. For parameter estimation, the frequency of conservation at position 2-12 was found adequate; this choice was based on simulations (data not shown) for finding the minimum length of nucleus at which specificity of match of any miRNA within its family is maximized. To compute the frequency of conservation among real miRNAs, the complete set of real miRNAs were split into two sets, 10% was used as a test set and its conservation at specified positions was checked in the remaining 90%. This was repeated 10 times with disjointed test-sets. For background miRNAs, conservation was examined in a complete set of real miRNAs. While the frequency of conservation in real miRNA was 0.69, those for the background was based on a pseudo-count of 1/3000 (= 0.0003) (see Table [1](#T1){ref-type="table"} for estimated parameters). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Positional conservation in 18 plant miRNA families**. The height of a bar at any position indicates the number of families in which said position was conserved. :::  ::: Precursor excision length ------------------------- Several miRNA discovery tools, whether based on purely computational approaches or deep sequencing data analysis, involve excision of genomic regions based on the occurrence of an inverted repeat or mapping of sequencing reads, respectively. These excised precursors are subjected to secondary structure prediction to check for imperfect fold-back or hairpin structures. Since the length of plant miRNA precursors varies widely, from 50 to 900 nt (Figure [5](#F5){ref-type="fig"}), an appropriate choice of excision length becomes necessary. It will be inappropriate to use the maximal length as excision length, because too few real precursors of that length would be available to estimate the parameters. Moreover, longer precursors also tend to have large variations in MFE, which is likely to decrease the prediction accuracy (Additional file [1](#S1){ref-type="supplementary-material"}: Figure 1). Therefore, an optimal choice would ensure 1) the majority of plant precursors get covered, and 2) adequate sample size of real precursors (greater than 30) is available for parameter estimation. Two of the optimal lengths, 277 and 336, which cover 96 and 98 percent of sequences, respectively, were found to agree with the above stated criteria (Table [1](#T1){ref-type="table"}). ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Frequency distribution (density and cumulative) of precursor length**. Two of the possible thresholds which cover 96% and 98% of the real precursors have been indicated by down-arrows (↓). At these lengths, a reasonable sample size of miRNAs were available for estimation of parameters. :::  ::: Other miscellaneous properties ------------------------------ Unlike animal miRNAs, plant miRNAs show better complementarity with miRNA star. Based on the hairpins available in miRBase release-14, the miRNA-miRNA\* duplexes were found to have a maximum of 5 unpaired bases (however the majority have up to 3 mismatches), two consecutive unpaired bases, and bulge size up to 1 (Table [1](#T1){ref-type="table"} and [2](#T2){ref-type="table"}). For an accurate prediction of plant specific miRNAs by any tool including miRDeep, these properties must be taken into account. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Duplex (miRNA-miRNA\*) associated parameters in miRDeep compared to plants miRNAs. ::: miRDeep Plants ---------------------------- ------------------------------- --------- Total unpaired Up to 8 (for 22nt long miRNA) Up to 5 Consecutive unpaired Up to 8 Only 3 Total bulge Up to 5-6 1 Consecutive bases in bulge Up to 5-6 Nil ::: While computational structure prediction yields minimum free energy structures according to the Turner model \[[@B17]\], the way miRNA precursors fold in their cellular context may differ from the one predicted by this model. This could be more likely for longer miRNA precursors as they may have higher degrees of freedom in which to be folded. Therefore, we examined how faithfully the software RNAfold predicted the secondary structure of real plant precursors longer than the mean length. We used RNAfold \[[@B18]\] to generate structures of 11 real plant miRNA precursors, of size in the range of 375-425 nt. The predicted structures were consistent with the secondary structures reported in miRBase \[[@B19]\]. The probability distribution of mapping of reads to the real and background precursors modeled by geometric distribution in miRDeep, was assumed to be same in plants, as the nature of mapping of reads to mature-loop and star-region are likely to be the same in both plants and animals. However, a minor difference (in the distribution) cannot be ruled out due to the longer loop region in plants. Validation of predictions by miRDeep parameterized for plant miRNAs ------------------------------------------------------------------- In order to validate if the above parameterization improves miRDeep\'s prediction accuracy for plant miRNAs, we selected recently discovered miRNAs in maize as a reference set \[[@B14]\]. These miRNAs were not covered by the training set (miRBase release 14) used in our study. The reference set included about 150 miRNAs at the whole genome scale, and a large fraction of them were experimentally validated. In this study, we report the results for maize chromosome 5, which was reported to have 23 miRNAs transcriptionally active in maize seedlings \[[@B14]\]. The parameterized miRDeep predicted 17 candidates with log-odds score well above 0, and 16 of them were found to overlap with the reference dataset (Table [3](#T3){ref-type="table"}). Out of the seven known miRNAs which were missed out, discovery of six of them failed as these miRNAs were either located within coding sequences (CDS) or showed high homology to CDS, thereby, reads corresponding to such cases have been excluded from our prediction. Although, this can be considered as a minor methodological limitation that reduces sensitivity, keeping this filter ON, helps minimize the false positives originating from CDS. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Validation of miRNAs identified by parameterized miRDeep. ::: miRNA Family miRNA name Confirmation by RACE Seedling RPM If identified by parameterized miRDeep Reads aligned Star (count) -------------- ------------- ---------------------- -------------- ---------------------------------------- --------------- -------------- miR156 zma-MIR156d 14111.2 ✔ 5932 1 miR156 zma-MIR156l 12645.91 ✔ 1659 1 miR160 zma-MIR160b ✔ 292.65 miRNA overlapping with CDS miR162 zma-MIR162 ✔ 2.74 poor abundance of reads miR166 zma-MIR166d 314.6 ✔ 328 1 miR166 zma-MIR166k ✔ 6.52 miRNA homologous to CDS miR166 zma-MIR166m ✔ 49.75 miRNA homologous to CDS miR167 zma-MIR167b ✔ 2860.25 ✔ 17402 1 miR167 zma-MIR167c ✔ 2872.95 ✔ 72684 0 miR168 zma-MIR168a 10364.43 ✔ 32179 1 miR169 zma-MIR169f 1234.4 ✔ 3371 0 miR171 zma-MIR171f ✔ 15483.87 ✔ 78 1 miR171 zma-MIR171m 4.12 ✔ 508 1 miR172 zma-MIR172b 134.49 ✔ 81 1 miR390 zma-MIR390b 2220.41 ✔ 6504 1 miR394 zma-MIR394a ✔ 4.46 MiRNA overlapping with CDS miR396 zma-MIR396f 849.46 ✔ 2503 1 miR396 zma-MIR396g 43.23 miRNA homologous to CDS miR397 zma-MIR397b 556.13 ✔ 1628 1 miR399 zma-MIR399h 43.23 ✔ 129 1 miR399 zma-MIR399i 90.57 ✔ 272 0 miR529 zma-MIR529 359.2 ✔ 1050 1 miR827 zma-MIR827 711.89 miRNA overlapping with CDS Unclassified ? \- 334 0 The candidates were compared against a recently discovered set of miRNAs, from chromosome 5 of maize, expressed in seedlings \[[@B14]\]. The data in third and fourth columns are from \[[@B14]\], while those in last two columns are from current study. The cases missed out by parameterized miRDeep appeared mainly due to a filter which discarded reads mapped to coding sequences (CDS). Even chance homology of a miRNA, either complete or at least in mature/star region, to any CDS can result into the miRNA being unidentified. RPM: Reads Per Million. ::: The miRDeep algorithm with default parameters, on the contrary, predicted only two candidates from chromosome 5 at a log-odds score of \>0, and both failed to match known miRNAs. On a whole genome scale, it predicted a total of sixteen candidates, out of which only two overlapped with known miRNAs. When checked for the presence of basic features typical to plant miRNAs, we found the majority of them failed to meet the desired metrics. Most of the discrepancies are related to the number of mismatches in the duplex region, the bulge size, non-specific homology, or precursor length, etc. Therefore, the newly parameterized miRDeep consistently identified known miRNAs (from chromosome 5, expressed in maize seedlings) with higher specificity and sensitivity than the original parameterization. Discussion ========== The plant miRNAs differ from animal miRNAs in several aspects, mainly in the hairpin length and in the nature of complementarity with the star sequence \[[@B2]\]. Although the length of mature sequences largely remains the same, the length of the loop region differs substantially in plants, owing to their recent evolution \[[@B19]\]. These differences strongly influence the statistical features used for their prediction. The minimum free energy (MFE), a commonly used measure for characterizing secondary structure of different types of RNA \[[@B20],[@B21]\], is also being used for characterization and/or prediction of miRNAs \[[@B3],[@B4],[@B22]-[@B24]\]. For screening miRNA candidates, the majority of previous studies have either used a fixed MFE threshold (for example, -18 kcal/mol) \[[@B13]\], or a variable threshold \[[@B25]\]. The miRDeep, however, involves comparison of (posterior probabilities of) MFE of real and background hairpins, enabling a more robust discrimination between them. This comparison in plants, however, did not prove to be so straightforward, as diversity in miRNA hairpin length results in multiple distributions of MFE. Although, the effect of hairpin length on MFE in plants has been reported earlier \[[@B25]\], these reports did not give a systematic evaluation of potential impacts of this relationship on prediction of plant miRNAs. In the present study, we observed that the MFE distributions become length-free by normalizing the MFE of precursor with its length, which renders the MFE of hairpins, of different length, comparable. We further observed only minor differences between the MFE distributions of real and background precursors of comparable length. This suggested that MFE alone may not be a good discriminator between real and background miRNAs, and more weight should be placed on other measures besides MFE. These findings however do not hold for dicot species, as they do show substantial differences between real and background, rendering MFE a more important discriminating feature. Differences between the secondary structures of candidate precursors and their shuffled counterparts is another important feature exploited for miRNA prediction \[[@B9]\]. The real precursors generally display substantial difference in the nature of folding (as well as in MFE) from their shuffled counterparts \[[@B16]\]. This difference is quantified by the p-value, which is the fraction of shuffled sequences with MFE lower than original precursors; a candidate with *p-value*≤0.05 can be statistically considered as stable. Although, plant and animal miRNA precursors, of the same length, are expected to have a similar frequency of stable precursors, due to the diversity in length of plant miRNA precursors, the parameter estimated for animals becomes inapplicable to plants. In the latter case, while p-values of real precursors remains almost the same even with increased length, that of background precursors declined substantially with length. These criteria become more effective when it comes to predicting longer candidate precursors, which is often the case with plants. Furthermore, the conservation of mature miRNAs is yet another important feature for miRNA discovery, exploited by miRDeep and several other tools, where the former takes into account the conservation in the nucleus region of mature miRNAs \[[@B9]\]. The positions which constitute the nucleus in animals are 7-8 nt in length, starting from position 2 \[[@B27]-[@B29]\], whereas in plants, there is near-perfect complementarity all along its length. We examined the nature of the positional conservation pattern in plants, and found a relatively longer conserved motif, wherein two conservation blocks were apparent: positions 2-13, and 16-19, with position 4 completely conserved. Implementation of positional conservation pattern in plants has improved the specificity of miRNA homolog prediction. Since plant miRNA precursors show a relatively broader distribution of length compared to animals, this in turn, necessitates a different choice of excision length(s) for candidate prediction. While Sunkar et al. \[[@B13]\] considered 200 nt as a threshold at which 90% of the real precursors in rice were covered, Jones-Rhoades et al. \[[@B30]\] took a higher precursor length (500 nt) for prediction in *A. thalina/O. sativa*. In another study by Adai *et al.*\[[@B25]\], in *A. thaliana*again, the maximum precursor size was set to 400 nt. Based on the distribution of length of plant miRNA precursors from miRBase database (release 14), length(s) which covered the maximum number of miRNAs, and at the same time, had an adequate number of precursors (30, for instance, which have properties of a normally distributed population) for parameter estimation, were chosen. We observed two thresholds satisfying the above constraints, 276 and 336, covering 96% and 98% of the population, respectively. To show how much the new parameterization improves the prediction accuracy, we used the miRDeep with the default parameters to predict miRNA candidates. Results suggested that a major fraction of miRNAs, predicted using the default parameters, did not match with experimentally identified miRNAs. The observed values of key features, such as number of total mismatches, bulges, nucleus conservation, excision length, etc., of the predicted candidates were atypical for plant miRNAs. Moreover, shorter nucleus size in default miRDeep led to identification of several false miRNA homologs. However, prediction using new parameters on the same dataset showed very high prediction accuracy, with good sensitivity and even better specificity. Notably, any proposed improvement in plant miRNA discovery must meet the criteria laid out for miRNA annotation in plants \[[@B12],[@B31]\]. Despite the parameter adjustments in the miRDeep algorithm, the primary criterion for miRNA annotation, namely precise excision of mature miRNA from the stem of a stem-loop precursor, is implemented faithfully. The parameterization doesn\'t interfere in miRDeep\'s core method. Further, two of the miRDeep\'s statistical features, namely characterization of stem-loop and mapping of reads onto precursors, are enough to prevent a siRNA being misclassified as miRNA. Besides, there have been recent reports of few plant miRNAs being processed by riboendonucleases other than DCL1 \[[@B32]\], therefore, the predictive methods should also be capable of their identification. This however does not pose much problem to the tools based on deep-sequencing reads, as their methods are guided primarily by the sequences. So, a DCL3 processed miRNA, for instance, will be analyzed just like the DCL1 generated miRNAs, despite the longer product size of the former. Furthermore, there have been rare reports of multi-functional stem-loops \[[@B12]\], which poses challenges to the tools available for miRNA discovery. We are skeptical about the ability of the current form of miRDeep algorithm to handle such complexity. This study also brings forward some issues that can be studied in the future. Increasing the number of plant genomes can allow researchers to further test whether MFE distributions of monocots and dicots truly differ and if so, study the underlying mechanisms. Furthermore, improved genome annotation will also improve the discovery of miRNAs missed out due to overlap with an otherwise incorrectly annotated CDS. Other desired advancements include modules for identification of other kinds of sRNA and the ability to characterize multi-functional stem-loops. Methods ======= Known miRNA sequences --------------------- The sequences of precursor and mature miRNA of plant and animal species were downloaded from miRBase (release 14) \[[@B19]\] and pooled into respective sets. Additionally, plant specific miRNA sequences were also extracted (as on April 15^th^2010) from the \'Plant miRNA Database\' \[[@B33]\], which contains additional numbers of precursors, which are largely computationally predicted. This dataset was further processed to remove redundancy, as in several instances, the precursor sequences are almost identical except for a few changes in nucleotides. Such sequences are likely to create bias towards the over-represented members of a miRNA family. This bias may affect some of the important analysis such as cumulative distribution of MFE and the effect of length on MFE. For obtaining non-redundant sets, the precursors were first divided into individual families and were then subjected to multiple alignment. Based on the alignment, the highly similar sequences were manually discarded, resulting in a set of 1904 precursors out of 2034. Generation of background sequences ---------------------------------- A background miRNA precursor is one that exhibits similar physical properties as that of real precursors, however, they are never transcribed into a miRNA. Since it is known that plant miRNAs are transcribed largely from the intergenic region and coding sequences are unlikely to contain any miRNA. Therefore, we used protein-coding sequences of five different species namely maize \[[@B34]\], sorghum, rice, *Medicago*, and *Arabidopsis*(sequences of remaining four species were obtained from \[[@B35]\]), for generation of background precursors (in the present study, maize was the default choice for background sequences unless otherwise specified). We used *de novo*miRNA discovery program, miRCheck \[[@B36]\], for identification of background sequences, over a broad length range, with default parameter settings, except the excision length was increased to 500 nt. Statistical scoring in miRDeep ------------------------------ The miRDeep score is given by, $$score = log\left( \frac{P(pre|data)}{P(bgr|data)} \right)$$ where *P (pre\|data)*is posterior probability of a test precursor being a \'real precursor\' (*pre*) given the values of multiple statistical features (*data*), and *P (bgr\|data)*is posterior probability of a test precursor being \'background precursor\' (*bgr*) given the *data*. When Eq. (1) is expanded using Bayes theorem, $$score = log\left( \frac{P(data|pre)P(pre)}{p(data|bgr)P(bgr)} \right)$$ where *P(data\|pre)*is conditional probability of observing *data*in real precursors, *P(data\|bgr)*is conditional probability of observing *data*in background precursors, and *P(pre)*and *P(bgr)*are prior probabilities of real and background precursors, respectively. The *data*, for any precursor, is described by five statistical features: 1) absolute value of minimum free energy (*abs*), 2) stability of secondary structure against randomized counterparts (*rel*), 3) signature (or pattern) of mapping of reads to the precursor (*sig*), 4) sequence conservation in the nucleus (or core) region of mature miRNA (*nuc*), and 5) presence of at least one read mapping star region (*star*) of the precursor. Therefore, the term *P(data\|pre)*, for instance, can be expanded into, $$score(data|pre) = P(abs|pre)P(rel|pre)P(sig|pre)P(nuc|pre)P(star|pre).$$ Out of these five probability distributions, *P(abs\|pre)*and *P(sig\|pre)*are continuous distributions (takes any value within specified range), while the remaining three are discrete distributions (takes one of the values from a set). The above also applies to the probability distributions for background (bgr) data. The cumulative frequency distribution of MFE, *P(abs\|pre)*or *P(abs\|bgr)*, shows gumbel (or extreme value) distribution, while that of signature, *P(sig\|pre)*or *P(sig\|bgr)*, shows geometric distribution \[[@B9]\]. The stability (*rel)*takes either of the two values: 0 if frequency of unstable is more than 5%, otherwise 1. Similarly, if nucleus regions is non-conserved at the specified positions, the value of *nuc*will be 0, otherwise 1. The value of *star*will be 0, if any read fails to map with star region, otherwise 1. The miRDeep essentially involves determination of these probability distributions, mentioned in Eq. (3), for known datasets of real (*pre*) and background *(bgr)*precursor/mature sequences. Here in current study, these distributions will be estimated for plants miRNAs. Frequency distribution of Minimum Free Energies ----------------------------------------------- The frequency distribution of MFE, plotted in Figure [1](#F1){ref-type="fig"}, was obtained from secondary structure prediction (by RNAfold \[[@B18]\]) of all known animal and plant miRNA precursors (available in miRBase release-14). To obtain a relationship between precursor length and MFE (as in Additional file [1](#S1){ref-type="supplementary-material"}: Figure 1), samples of plant precursors of different lengths, with sample size 60 (however for highest length, the sample size dropped to 21), each sample being homogeneous in length with a variation of 5 nucleotides (however 10 for the samples having members less than said sample size) were obtained. The MFE was computed by RNAfold and mean MFE values were plotted against the average precursor length of each sample. For comparison of distributions of MFE of background precursors from multiple plant species (shown in Figure [2B](#F2){ref-type="fig"}), two-sample Kolmogorov-Smirnov tests were conducted using \'ks.test()\' function of R statistical package \[[@B37]\]. For theoretical approximation of distributions shown in Additional file [3](#S3){ref-type="supplementary-material"}: Figure 1, the empirical distributions of MFE of 47 real and 554 background precursors, of length 260-290 nt, were fit with Gumbel distribution function (minimum), $$F(x) = 1 - e^{- e^{\frac{x}{b}}}$$ where *b*is a scaling factor. Since the fitting errors were high so the log-odds were directly computed. The MFE log-odds score, displayed in Figure [2D](#F2){ref-type="fig"}, were obtained by $$score(MFE) = log\left( \frac{P(abs|pre)}{P(abs|bgr)} \right)$$ The log of ratio of conditional probabilities of MFE of real and background precursors was computed for each of the bins, and were plotted. The frequency distribution of MFE was obtained with bin-size = 0.01. The empirical distribution was modeled by a modified sigmoid function \[[@B38]\]: $$f(normalized\_ MFE) = \frac{a}{(b + e^{c*normalized\_ MFE})}$$ where *a*, *b*, and *c*are fitting parameters. Core conservation in mature miRNAs ---------------------------------- To study the positional conservation in mature miRNA families of plants, we obtained all members of 18 miRNA families, sampled randomly. It was however insured that each family must have representatives from at least four species. Members of each miRNA family were aligned using CLUSTALX \[[@B39]\] and a position with ≥ 0.9 conservation was marked conserved. Then positional conservation profile of all 18 families were summed. That is, count of families conserved at each of the positions, starting from 1 to 23, was obtained and plotted. Since higher evolutionary divergence is likely *across*families (than *within*family), therefore threshold for defining a position as *conserved*was further relaxed, that is, ≥ 0.75 (Figure [4](#F4){ref-type="fig"}). In order to obtain the frequency of nucleus conservation in real miRNAs, a strategy of 10-fold cross-validation was applied. That is, the complete set of known mature miRNAs were first shuffled, and divided into two fractions with ratio 9:1. The smaller fraction was used as a \'test set\' and its conservation (at positions 2-12) was examined in sequences of larger fraction, and the frequency of sequences from test-set hitting \'training-set\' was recorded. This was repeated 10 times, and an average of all ten frequencies was obtained. For background, the conservation of a large set of putative mature sequences, against the training-set described above, were examined in a similar way. The log-odds of observing conserved (*nuc*= 1) and non-conserved (*nuc*= 0) nucleus were $score(conserved\, nucleus) = log\left( \frac{P(nuc = 1|pre)}{P(nuc = 1|bgr)} \right) = 7.63$ and $score(non - conserved\, nucleus) = log\left( \frac{P(nuc = 0|pre)}{P(nuc = 0|bgr)} \right) = 1.17$, respectively (Table [1](#T1){ref-type="table"}). Stability of secondary structures of miRNA precursors ----------------------------------------------------- For computing frequency of stable and unstable real precursors, analysis was done on a sample of 500 plant miRNA precursors of varying lengths. Each of them were subjected to mononucleotide shuffling for 999 times, followed by computing *p-value*, $$p - value = \frac{R}{(N + 1)}$$ where R is the number of shuffled sequences with MFE greater than that of original sequence, N is the number of iterations. For background precursors, the analysis was carried on three samples of lengths: 100, 200, and 300 nt, each sample having 100 sequences. They were subjected to 499 iterations. The estimated log-odd scores for stable (*rel = 1*) and unstable (*rel = 0*) were given by, $score(stable) = log\left( \frac{P(rel = 1|pre)}{P(rel = 1|bgr)} \right) = 1.37$ and $score(unstable) = log\left( \frac{P(rel = 0|pre)}{P(rel = 0|bgr)} \right) = 3.624$, respectively (Table [1](#T1){ref-type="table"}). Running default and parameterized miRDeep on a sample deep-sequencing data -------------------------------------------------------------------------- The Illumina reads, generated from Maize seedlings transcriptome, were downloaded from NCBI Gene Expression Omnibus (ID: GSM448856) \[[@B40]\]. The reads, already trimmed for adaptors, were 7.922 millions in number. The pre-processing of Illumina reads was done using MiroPipeline \[[@B11]\], which involved low complexity filtering, inclusion of reads without adaptor, and replacing identical sequences with single representatives. For mapping reads in miRDeep, the default choice of mapper, namely mega-blast \[[@B41]\], was replaced with one of faster mapper, namely SOAP-v2.2 \[[@B42]\]. Mapping was done onto unmasked Maize genome \[[@B43]\], with repeated hits allowed, and the maximum number of mismatches was set to 1. The hits were later filtered for number of multiple best hits, a maximum of 20 was set as a threshold (so that reads repetitive in nature get excluded, at the same time those most likely mapping to multiple members of a miRNA gene family are considered), and then converted into miRDeep compatible format. The hits with protein coding sequences and with various types of non-coding RNA (eg., rRNA, tRNA) were also filtered out \[[@B44],[@B45]\]. The excision length was set to a maximum of 300. In order to speed up the mapping of filtered reads onto precursors (default program: auto\_blast.pl of miRDeep package), we used MiroPipeline (configured for using seqmap as an alignment tool) \[[@B11]\]. The core program (*miRDeep.pl*) was run with score threshold 0 and stability check on. One of the miRDeep\'s filtering criteria, to exclude candidates with bifurcation in secondary structure, was set off to improve sensitivity. The key syntactic changes for incorporating plant specific parameters in PERL scripts of miRDeep are listed in Additional file [5](#S5){ref-type="supplementary-material"}. Conclusions =========== The difference in properties of plant and animal miRNAs has large impact on the statistical features used for miRNA prediction. The parameters used for animal miRNA prediction cannot be used to predict plant miRNAs. Among the statistical features, the minimum free energy was found to have marginal difference between real and background in monocots. However dicots showed a different behavior wherein MFE scoring is potentially a key discriminator. The stability pattern in plant miRNAs was different to animals, in particular among the background sequences. The positional conservation profile was relatively longer in plants, so does the associated frequencies. The new set of parameters identified in this study will substantially improve our capacity to predict plant miRNAs. Conflict of interest ==================== The authors declare that they do not have competing interests. Authors\' contributions ======================= VT, SW, MX, RB, WPQ, AM, and XZ conceived the study. VT and SW carried out the experiments. VT, SW, AM, XZ analyzed the results and wrote the paper. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Figure 1**. Mean MFE as a function of (precursor) length. The vertical bars display the standard deviation. The best fit linear curve has a slope of 0.48. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Figure 1**. Comparison of cumulative frequency distributions of (length normalized) MFE of real and background precursors from dicot species. These precursors are of length 260-290 nt. Bgr: background. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Figure 1**. Best fit curve of Gumbel distribution (minimum) for the cumulative distributions of MFE of real and background precursors (length: 260-290 nt). Except for small range of normalized MFE values, largely in middle, the corresponding curves do not fit well. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **Figure 1**. Cumulative frequency distribution of *p-value*of background precursors of five species (size = 300 nt). Those from dicot species have higher stability, this, rendering them less distinguishable from the real precursors. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 5 **Text**. Key changes made in the syntax of miRDeep to incorporate plant-specific parameters. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ The current study was supported by Bill and Melinda Gates Foundation.

PubMed Central

2024-06-05T04:04:19.584384

2011-2-16

PMC3053259

Background ========== Now that rinderpest has been eradicated, *M. mycoides*subsp. *mycoides*Small Colony (MmmSC), the aetiologic agent of contagious bovine pleuropneumonia (CBPP), is considered to be the most important threat to cattle farming in affected countries. Although CBPP has been eradicated from most continents, it persists in Africa. The disease is notifiable to the world organisation for animal health (OIE) and, following notification, export of live cattle to countries free of CBPP is forbidden. Programmes for the control of CBPP involve vaccination, but eradication is based solely on slaughter of affected herds and strict control of animal movements. Because of its economic importance, CBPP has received much attention especially when it affected industrialized countries. This was the case in the 1960\'s when Australia started an eradication programme and in the 80\'s and 90\'s when Europe suffered a re-emergence of the disease \[[@B1],[@B2]\]. Identification and diagnostic methods have been improved. This task was difficult as MmmSC belongs to a complex of species, the so-called \"*M. mycoides*cluster\". This cluster consists of five closely related mycoplasmas that are referred to as: *M. mycoides*subsp. *mycoides*Small Colony (MmmSC), *M. mycoides*subsp. *capri*(Mmc), *M. capricolum*subsp. *capricolum*(Mcc), *M. capricolum*subsp. *capripneumoniae*(Mccp), and *Mycoplasmaleachii* (Ml), the last being a group of strains that had remained unassigned until recently when a modification of the cluster taxonomy was proposed based on both phenotypic and recent phylogenetic studies \[[@B3],[@B4]\]. The designation *Mycoplasma mycoides*subsp. *mycoides*Large Colony (MmmLC) was discarded and the corresponding \"LC\" isolates are now considered to be an additional serovar of Mmc \[[@B4]\]. All members of the *M. mycoides*cluster share phenotypic and genetic traits. One very close relative of MmmSC is Mmc, and indeed, the reference growth inhibition test using rabbit hyperimmune serum does not allow differentiation between these two subspecies. However, it is important to be able to identify MmmSC and Mmc without ambiguity because the two organisms differ greatly in terms of pathogenicity, geographical distribution and quarantine regulations. Unlike MmmSC, Mmc affects mostly small ruminants where it can induce a syndrome called \"MAKePS\" with lesions including mastitis, arthritis, kerato-conjonctivitis, pneumonia and septicaemia \[[@B5]\]. It can also be found in the ear canal of asymptomatic animals \[[@B6]\]. Mmc strains are found world-wide, especially where goats are raised. Before 1994, the distinction of the two subspecies *in vitro*was difficult and findings were sometimes ambiguous as few tests were sufficiently discriminatory (Table [1](#T1){ref-type="table"}). Analysis of cellular proteins by one-dimensional SDS PAGE showed that this approach could be used to distinguish MmmSC from Mmc \[[@B7]\]. Specific detection methods were then developed based on PCR technology and the empirical search for specific restriction sites in the amplified fragment or with PCR primers allowing a specific amplification \[[@B8],[@B9]\]. Finally, MmmSC specific monoclonal antibodies were obtained and, more recently, specific real-time PCR methods have also been validated \[[@B10]-[@B13]\]. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Phenototypic characteristics differentiating *M. mycoides*subsp. *mycoide*s SC (MmmSC) and *capri*(Mmc) ::: *Characteristics* *MmmSC* *Mmc* ---------------------------------------- ----------- ----------- Mean colony size after 137 h at 37°C 0.99 mm 2.26 mm Liquefaction of inspissated serum weak \+ Casein digestion \- \+ Sorbitol fermentation \- \+ Presence of ornithine transcarbamylase \- \+ Presence of alpha-glucosidase \- \+ Thermal stability at 45°C sensitive resistant Digestion of DNA by MboI \+ \- DpnI \- \+ *Substrate oxidation (12-50 μM)*  Maltose \- \+  Trehalose \- V  Mannose \- \+  Glucosamine \- \+ List of phenotypic differences that allow the differentiation of *M. mycoides*subsp. *mycoides*small colony (SC) and *M. mycoides*subsp. *capri in vitro*. Adapted from \[[@B39],[@B68]-[@B70]\]. ::: The two subspecies have been considered to be very close relatives that could be distinguished only by minute differences *in vitro*in spite of their marked differences of physiology *in vivo*. Recent advances in sequencing and bioinformatics allow the comparison of whole bacterial genomes. The complete genome sequence of the MmmSC reference strain PG1 was made public in 2004 \[[@B14]\]. Here, we report the availability of the complete genomic sequence of an Mmc strain and compare it with that of the MmmSC PG1 strain. We describe differences in chromosome organization, gene repertoire, sequence polymorphism, and consider possible links between these differences and the physiology of the two bacteria. In addition, the recent publication of the genome sequence of strain GM12 (Mmc) allowed a genome-wide comparison for these two strains and an evaluation of intra-species polymorphism. Results ======= General genome features ----------------------- The Mmc 95010 genome consists of a circular chromosome of 1,153,998 bp (EMBL/GenBank accession number Q377874) and of a plasmid of 1,840 bp (pMmc-95010, GenBank accession number [FQ790215](FQ790215)). The chromosome has a G+C content of only 23.8 mole%. Genome annotation revealed 924 putative CDS, two Integrative Conjugative Elements (ICE), each being on average 30 kbp long and including 18 CDS, and 24 recognized copies of Insertion Sequences (IS) belonging to five different IS types. Two sets of rRNA genes and 30 tRNAs can be predicted. Putative genes corresponding to the tmRNA and the RNA component of the RNase P were also found, as in other mycoplasmas. Overview of genome structure and plasticity ------------------------------------------- The MAUVE alignment of Mmc 95010, MmmSC PG1, and the slightly more distant California kid (Mcc) genome sequences allowed the identification of 62 Locally Collinear Blocks (LCB) that were interspaced by Mmc-specific DNA stretches of various lengths (Figure [1](#F1){ref-type="fig"}). Six of these Mmc-specific stretches merit special attention. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Comparative genome structure of strains 95010 (Mmc), PG1 (MmmSC) and California kid (Mcc)**. The graph represents an alignment of the colinear blocks, identified by MAUVE, that are conserved in three closely related genomes: Mmc 95010; MmmSC PG1; Mcc California kid. Numbers 1 to 6 refer to large DNA fragments of the Mmc 95010 genome with apparently no homology in the other two genomes. Numbers 7 to 11 indicate DNA inversions (blocks below the central line) that distinguish Mmc from MmmSC. Comparison with the strain California kid genome allows identification of the genome in which these inversions have taken place. For blocks 7, 9, 10 and 11, the inversion took place in the MmmSC PG1 genome, whereas inversion 8 took place in the Mmc 95010 genome. The higher number of block rearrangements in the Mcc genome is in agreement with its taxonomic position and greater evolutionary distance from Mmc, as compared with MmmSC. :::  ::: The first contains ten CDS (MLC\_0740 to MLC\_0840), six of which constitute a predicted lipoprotein (lpp) family for which the best similarity scores were found with *M. agalactiae*or Mcc, whereas no homologues were found in MmmSC. This gene family was previously described as a candidate for an Horizontal Gene Transfer (HGT) between *M. agalactiae*and the M. mycoides cluster \[[@B15],[@B16]\]. The absence of homologues in MmmSC may be the consequence of a deletion from this pathogen after the HGT event. The second is a stretch of 12 CDS (MLC-1730-1830) forming a maltodextrin/maltose gene cluster. Most of these CDS exhibit substantial similarity with genes in distantly related genera, such as *Listeria*or *Bacillus*, again suggestive of acquisition by HGT. Interestingly, the last four genes (MLC\_1810 to 1840) had homologues in Mcc, an indication that Mcc may have previous contained, and then lost, the first eight CDS of the cluster. The third (MLC\_2080 to MLC\_2280) and fourth (MLC\_2890 to MLC\_3080) specific DNA stretches correspond to two copies of the Integrative Conjugative Element (ICE) (see below) that are 30 and 28 kbp long, respectively. The fifth specific DNA stretch contains a pseudogene and six CDS (MLC\_3580 to MLC\_3640). The second and third CDS were similar to two ICE CDS (MLC\_3590 and MLC\_3600) with the last four CDS being identified as IS*1296*copies. Therefore, this DNA stretch appears to be the remains of an ICE copy after a partial deletion. Finally, the sixth specific DNA stretch contains five CDS (MLC\_7610 to MLC\_7650) with four of them showing either no similarity to known sequences in any organism or showing similarity with sequences in non-mollicute organisms (*Treponema denticola, Trichomonas vaginalis, Finegoldia magna*), suggesting again a possible indication of HGT. MLC\_7610 and MLC\_7620 show similarities with viral A-type inclusion protein, consistent with this possibility. The MAUVE genome alignment clearly indicated inversions of large DNA fragments including one or more collinear blocks. Only five of such events were identified between Mmc and MmmSC (Figure [1](#F1){ref-type="fig"} LCB 7 to 11) and nine with Mcc, consistent with Mmc and MmmSC being more closely related to each other than to Mcc. The initial orientation of these LCB could be deduced for those that were colinear and with a similar orientation in two of the three genomes. For four of the five (LCB 7, 9, 10, 11), the DNA inversion occurred in the MmmSC genome with the inverted LCBs being flanked by IS copies (IS*1634*or IS*1296*). The exception is LCB 8, which is flanked by two ICE copies in opposite direction in the Mmc-95010 genome, an indication that this inversion occurred during the duplication and insertion of the first ICE copy (see below) Mobile elements: Integrative Conjugative Elements ------------------------------------------------- The two ICE copies identified in the 95010 genome were not found in the other available genomes of the *M. mycoides*cluster, including Mmc GM12. This genome contains an ICE but it is located elsewhere on the chromosome. This suggests that these mobile elements were recently inserted into the 95010 genome. The two ICE copies differ slightly in size (ICEMmc95010-1a is 30,136 bp-long, with 21 CDS, and ICEMmc95010-1b is 28,965 bp-long, with 19 CDS) but are otherwise almost identical (Figure [2](#F2){ref-type="fig"}). The CDS are very similar to those of the ICE found in Mcc (six of the 18 Mcc CDS exhibit more than 70% similarity to ICEMmc95010-1a CDS). Four CDS seemed to be more conserved among mycoplasmas: CDS 1 and 22, at the extremities of the ICE but contiguous in the circular form, and CDS 5 and 17, which encode TraG and TraE, respectively. For CDS 5 and 17, the highest similarity was found with homologues in Mmc strain GM12 (98%), Mcc California kid (80%), and partial sequences of *M. agalactiae*PG2 (80%). They were more distantly related to an *M*. *fermentans*CDS (50%). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Comparison of the two ICEs found in (95010) with those in Mcc (California kid) and *M. fermentans***. Schematic representation of ICEs found in Mcc (top) and *M*. *fermentans*(bottom) with the two ICE copies found in the Mmc 95010 genome and the single copy identified in strain Mmc GM12. The numbering above refers to that of the original publications describing these ICEs \[[@B16],[@B26]\], and numbering below the ICEs refers to CDS numbers in the respective genomes. The CDS are coloured according to the similarity results from BLASTP with the ICEMmc95010-1a CDS (dark blue, \#90%; light blue, 89-70%; green, 69-50%; no colour, \< 50%). The two ICE copies in 95010 differ by two CDS (in red) which may have been deleted from the \"1b\" copy (or inserted into the \"1a\" copy). The ICE copy in GM12 is very similar as 13 CDS exhibit more than 90% similarity with ICEMmc95010 -1a CDS. However, some CDS are only present in GM12 (MMCAP2\_0566), another is duplicated (MMCAP2\_0556 and 0557) and another (MMCAP2\_0554), in yellow, is similar to MCAP\_0556. The greatest similarities were found within the TraE and TraG proteins and for the 2 CDS at the extremities. These terminal CDS may be joined and form a single CDS in the circular form of the ICE. :::  ::: The two Mmc 95010 ICE copies are in opposite orientation and flank an 80.9 kbp-long DNA fragment which is inverted, in relation to the most closely related genomes (Mmc strain GM12, MmmSC and Mcc). Examination of the direct repeats at the insertion sites of the ICE clearly showed that a recombination had occurred between them, resulting in the observed chromosomal inversion. It is therefore likely that the organization of this region is the result of successive events including the initial integration of an ICE, the duplication and integration of the ICE at a second site in the opposite orientation and finally recombination between the two copies (Figure [3](#F3){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Putative model of the insertion and duplication of an ICE copy in Mmc 95010 genome**. This figure describes the events that may have led to the presence of two ICE copies in the Mmc 95010 genome. A first ICE copy may have been acquired from a donor cell by conjugation, with the subsequent insertion in CDS A (MLC\_2290). This insertion led to the disruption of CDS A resulting in fragments (A1 and A2) and the creation of a small Direct Repeat at the extremities of the ICE (DR). The ICE duplicated and inserted elsewhere in the genome, disrupting another CDS (CDS B, MLC\_2070) and creating two additional direct repeats. Recombination between the two ICE copies then lead to the inversion of the DNA fragment located between them. The presence and the site of insertion of ICE copies in Mmc is strain specific. This was shown by PCR amplification with two primers chosen on either side of the ICE Mmc95010-1a integration site (primers VIP-F and VIP-R). In four Mmc strains, this PCR amplified the expected fragment, an indication that there was no ICE copy inserted at this site (see table 2). :::  ::: We used a PCR primer pair designed to amplify a fragment of the well conserved *traG*gene (Additional file [1](#S1){ref-type="supplementary-material"} table S1) to screen for the presence of this type of ICE in a representative set of 31 strains of the *M. mycoides*cluster and other ruminant mycoplasmas (Table [2](#T2){ref-type="table"}). Amplification products were obtained for four of the ten Mmc strains, indicating that these strains carried closely related ICE. Two additional primers (VIP-F and VIP-R), designed to flank the insertion site of one of the 95010 ICE copies, were used for PCR analysis of DNA from the same set of strains. Of the four *traG*-positive Mmc strains, three yielded the expected 1,079 bp-long amplicon, indicating that the detected ICE was integrated elsewhere on the chromosome of these strains. As expected, strain 95010 DNA did not yield any amplified product because the primers were separated by about 30 kbp and in the same orientation. The absence of amplification with \"TraG\" and \"VIP\" for the five other strains DNA was not expected; it may have been due to sequence variations that hampered the correct annealing of the primers or by other, unidentified, genetic rearrangements. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Mycoplasma strains used in this study to evaluate the presence of genetic mobile elements^a^ ::: Strain Year Origin PCR ---------------------------------------- --------- ------------- -------------- -------------- ------------ --------- --------- **IS*Mmy2*** **IS*Mmy3*** ***traG*** **VIP** **16S** *M. mycoides*subsp. capri 55507-1^b^ 1998 Germany \+ \- \- \- \+ 95010-C1^b^ 1995 France \+ \+ \+ \- \+ Y-goat^R\ b^ 1956 Australia \+ \- \+ \+ \+ 2002-054 (VP9L)^b^ \< 2002 India \- \- \- \- \+ 8756-C13^b^ \< 1987 USA \- \- \- \- \+ Kombolcha^b^ 1975 Ethiopia \- \- \+ \+ \+ WK354 1980 Switzerland \+ \+ \- \+ \+ N108 1977 Nigeria \+ \- \- \- \+ L 1975 France \+ \+ \+ \+ \+ PG3^T^ 1950 Turkey \- \- \- \- \+ *M. capricolum*subsp. *capricolum* 8086-1 1980 France \+ \- \- \- \+ 2002-053 (VP28L) \< 2002 India \- \- \- \- \+ California kid^T^ 1955 USA \- \- \- \- \+ 96038 \< 1996 Greece \- \- \- \- \+ 90122 (C1547) 1990 Ivory Coast \- \- \- \+ *M*. *mycoides*subsp. *mycoides*SC T1/44 1954 Tanzania \- \- \- \- \+ 94111 1994 Rwanda \- \- \- \- \+ 8740-Rita 1987 Cameroon \- \- \- \- \+ PG1^T^ \< 1931 Unknown \- \- \- \- \+ *M. capricolum*subsp *capripneumoniae* F38^T^ 1976 Kenya \- \- \- \- \+ GL102 1981 Tunisia \- \- \- \- \+ Gabes 1981 Tunisia \- \- \- \- \+ 95043 1995 Niger \- \- \- \- \+ *M*. bovine group 7 of Leach PG50^R^ Australia \- \- \- \+ \+ D424 \< 1990 Germany \- \- \- \+ \+ 9733 1993 India \- \- \- \+ \+ *M*. *putrefaciens* KS1^T^ 1954 USA \- \- \- \- \+ Tours 2 1972 France \- \- \- \- \+ *M. yeatsii* GIH^T^ 1981 Australia \+ \- \- \- \+ *M. cottewii* VIS^T^ 1981 Australia \+ \- \- \- \+ *M. auris* UIA^T^ 1981 Australia \- \- \- \- \+ ^a^Whenever possible, strains originating from different continents were selected to improve the representivity of the sample. PCR results are recorded as positive (+) or negative (-) with primer pairs designed to detect the presence of Insertion Sequences (ISMmy2 and ISMmy3), Integrative Conjugative Elements (*traG*) or the insertion of an integrative conjugative element at a specific location (VIP). Non specific PCR primers corresponding to 16SrDNA genes were used as positive amplification controls. ^b^These strains were formerly identified as *M*. *mycoides*subsp. *mycoides*LC ::: Mobile elements: Plasmid ------------------------ Plasmid pMmc-95010 is 1,840 bp long and has a G+C content of 29.0 mole%. It is predicted to encode two proteins probably involved in plasmid replication. The protein encoded by orfA is similar to several replication (Rep) proteins of plasmids replicating by a rolling-circle mechanism. The putative pMmc-95010 Rep protein has a conserved architecture domain corresponding to that of Rep\_2 plasmid replication proteins (pfam01719). The sequence included five motifs typical of pMV158 plasmid family \[[@B17]\]. The protein encoded by orfB has a sequence structure similar to that of the RHH\_1 ribbon-helix-helix family of CopG repressor proteins (pfam1402 domain). CopG, also known as RepA, is a transcriptional regulator that controls the plasmid copy number in the pMV158 derivative pLS1 \[[@B18]\]. Several plasmids with sizes between 1,717 and 3,432 bp have been isolated from mycoplasmas related to the *M*. *mycoides*cluster. Plasmid pKMK1 (M81470.1) was isolated from a Mmc strain (GM12); it has a size similar to that of pMmc-95010 and the predicted proteins of the two plasmids share more than 97% similarity \[[@B19]\]. Complete nucleotide sequence alignment indicated nearly identical sequences. Plasmids pADB201 (NC\_001382), pBG7AU (NC\_002569) and pMyBK1 (NC\_011102) were isolated from *M. mycoides*, *M. leachii*and *M. yeatsii*, respectively \[[@B20],[@B21]\]. pADB201 and pBG7AU encode two proteins similar to OrfA and OrfB from pMmc-95010, but pMyBK1 does not. Interestingly, nucleotide alignments of the plasmid sequences suggested a composite origin of plasmid pADB201: the first thousand nucleotides are very similar to pBG7AU sequences whereas downstream sequences were closely related to sequences found in pMmc-95010 and pKMK1. This suggests that plasmids found in various mycoplasma species of the *M. mycoides*cluster may circulate among strains and recombine to generate new hybrid plasmids. More surprisingly, blast search using pMmc-95010 nucleotide sequence identified a non-coding, GC-rich sequence of 286 nt showing 96% identity with two Mmc chromosome regions (Additional file [2](#S2){ref-type="supplementary-material"} Figure S1). These regions are located within ICEs, between CDS22 and the terminal inverted repeat. Three palindromic sequences of 7, 7 and 16 nucleotides were repeated two to five times with the potential to form several stem-loop structures. A related sequence was also found in the ICE in Mmc strain GM12 and in plasmids pKMK1 and pADB201. The distribution of this trait suggests that ICEs and plasmids have probably exchanged sequences. This highly structured region covers a 130 bp sequence identified as the single-strand origin (*sso*) of pKMK1 replication (Additional file [3](#S3){ref-type="supplementary-material"} Figure S2). In rolling-circle replication plasmids, the *sso*is the initiation site of lagging strand synthesis and is essential for the conversion of a single-strand intermediate into a double-strand molecule. The identification of an sso-type region in ICEs raises the question of whether this plasmid-derived sequence is involved in ICEMmc95010 dissemination. Mobile elements: Insertion Sequences ------------------------------------ Five different types of insertion sequences (IS) were identified in the Mmc 95010 genome. Nine copies of IS*1296*were identified. They were very similar to IS*1296*copies present in the MmmSC PG1 genome both in terms of length and deduced amino-acid sequence (98% similarity). Six copies of IS*Mmy1*were identified. The similarity with those found in MmmSC PG1 was high (87%) but lower than for IS*1296*. Six copies of a novel IS named IS*Mmy2*were found; one of the copies was considered as the prototype for a new IS and has been submitted to GenBank (EMBL/GenBank accession N° DQ887910). This IS is 1,374 bp long, contains 24-bp inverted repeats and appears to have generated 3-bp-long direct repeats at its insertion sites. It contains a putative ribosome binding site (position 66-71) 5 bp upstream from the ATG start codon of a 1,278 bp-long orf (426 aa); the orf possesses a classical DDE catalytic site motif. Accordingly, this IS was identified as belonging to the IS*3*family and the IS*150*group. The presence of IS*Mmy2*in related strains was evaluated by two techniques: Southern blotting with a DIG-labelled probe (results not shown) and PCR using a single primer corresponding to the conserved part of the inverted repeat found at the extremities of the IS element (Table [2](#T2){ref-type="table"}). The results with the two techniques were entirely consistent: elements related to IS*Mmy2*were detected in six of ten strains tested belonging to the subspecies Mmc and Mmc. However, this IS was not detected by either of the two techniques in any of four MmmSC strains. The absence of detectable IS*Mmy2*in the MmmSC subspecies was not expected, given that this IS shares substantial similarity with two contiguous MmmSC CDS. A single DNA sequence homologous to IS*Mmy2*was however found between positions 801579 and 802903 of the MmmSC PG1 sequence (NC\_005364.2). This sequence seems to represent the remnant of an IS*Mmy2*without transposase activity because it carries a mutation disrupting the transposase gene. The sequence found in MmmSC PG1 may thus be the result of ensuing genetic drift by accumulation of insertions, deletions (including the 49 terminal bases of the original IS) and point mutations. The failure to detect this sequence by Southern hybridization may be explained by the presence of *Hin*dIII sites in the MmmSC PG1 IS*Mmy2*-remnant sequence (the copies in Mmc IS*Mmy2*contain no such sites) resulting in much smaller fragments. The failure to detect these sequences by PCR was clearly due to the absence of the Right Inverted Repeat essential for PCR amplification with a single primer. Two copies of another novel insertion sequence, named IS*Mmy3*were found. IS*Mmy3*is 1,442 bp long with 24 nt long terminal inverted repeats that are only partially identical (66%) and has generated 3bp direct repeats. It encodes two putative polypeptides that may be fused to form a transposase as there is a potential frameshift position between nucleotide positions 503 and 510. This new IS therefore belongs to the IS*3*family and it closely resembles IS*1296*(60% similarity). IS*Mmy3*was detected in three of ten Mmc strains tested but was not detected in MmmSC or Mcc strains (Table [2](#T2){ref-type="table"}). A fragment of another insertion sequence was detected (MLC\_7590). This partial IS copy codes for a single protein that shares 60% similarity with transposase protein A of the IS*1296*found in MmmSC. A very similar partial copy is also found in Mcc California kid strain (96% similarity). The diversity of IS types and copy numbers in the Mmc strains can be assessed by comparison of the GM12 and 95010 genomes (Figure [4](#F4){ref-type="fig"}). Both genomes contain eight copies of IS*1296*but only two of them are inserted at the same locations. This indicates that transposition events are relatively frequent for this IS. In addition the presence of copies of IS on either side of an inverted linear block in GM12 shows that IS duplication and transposition may lead to major genome rearrangements. Note also that three other IS types (14 copies) present in 95010 do not appear to be associated with any rearrangement. The only other major rearrangement in the 95010 genome was triggered by an ICE. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Comparison of Mmc 95010 and Mmc GM12 genome organisation in relation to mobile elements**. The colinear blocks identified by Mauve software in the closely related Mmc 95010 and Mmc GM12 genomes are shown. The various mobile elements where then positioned on the graph. IS*1296*copies are represented by dark blue lozenges, IS*Mmy1*by red triangles, IS*Mmy2*by green circles, IS*Mmy3*by vertical blue lines and ICEs by red horizontal rectangles. Eight IS*1296*copies are found in each genome but only two map at the same site (copies 6 and 7 in Mmc 95010). In addition, two copies of IS*1296*are positioned on either side of an inverted DNA fragment: duplication and transposition of this IS may have been responsible for this inversion. Similarly, the duplication of the ICE in the Mmc 95010 genome may have caused the inversion of a large DNA fragment. :::  ::: Mmc-specific clusters: the maltodextrin/maltose gene cluster ------------------------------------------------------------ A specific stretch of 12 CDS was identified downstream from an IS*1296*copy (MLC\_1730 to 1840). None of these CDS have any significant similarity with the PG1 genome. All have similarities with genes involved in carbohydrate metabolism and more specifically starch/glycogen and maltose utilization (hereafter named the maltodextrin/maltose gene cluster) (Figure [5](#F5){ref-type="fig"}). This cluster is composed of genes coding for a phosphoglucomutase (EC5.4.2.6), amylases that may be involved in starch degradation, two lipoproteins that may be involved in substrate binding, a hypothetical protein with transmembrane segments, followed by *malC, malG*, and *malK*genes that may encode a translocation complex, *mapA*(EC2.4.1.8) a maltose phosphorylase, *dexA*(EC3.2.1.10) a gene encoding oligo-1,6 glycosidase and, at the end, a transcription regulator. Most of these genes are similar to genes found in other bacteria such as *Listeria*, *Mycoplasma*(*M. mobile*and *M. pulmonis*), and *Lactococcus*, but the most similar homologues were mostly in *Bacillus*in which maltose and maltodextrin utilization has been documented \[[@B22]\]. A tentative model for starch and maltodextrin utilization can be established for Mmc 95010 with the two lipoproteins encoded by the cluster possibly involved in substrate binding (Figure [6](#F6){ref-type="fig"}) and MLC\_1780 in transmembrane transport. Only four genes of this operon were found in the MccCalifornia kid genome and none were found in MmmSC PG1 although the flanking genes were present (IS*1296*and *pep*F). This suggests that this cluster was present in the ancestor of the *M*. *mycoides*cluster and that the whole cluster has been deleted from MmmSC PG1. For Mcc California kid, part of the cluster has been deleted and replaced by 20 CDS found inserted at this position. The maltodextrin/maltose gene cluster is fully conserved in the Mmc GM12 genome but there is no IS*1296*copy found upstream. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Organisation of the \"maltodextrin/maltose gene cluster\" in the Mmc 95010, MmmSC PG1 and Mcc California kid genomes**. In the Mmc 95010 genome, the \"maltodextrin/maltose gene cluster\" is composed of 12 CDS shown in green. The cluster is flanked by an IS*1296*copy and *pep*F. The cluster comprises: a phosphoglucomutase (*pgm*B), two amylases (CDS 1740 and 1750), two lipoproteins that may be involved in binding (CDS 1760 and 1770), a translocation complex (CDS 1780, *mal*C, *mal*G, *mal*K), two CDS involved in maltose degradation (*map*A, *dex*A), and a CDS probably involved in the regulation of transcription. This cluster is absent in the PG1 genome although the flanking genes (the IS*1296*copy and *pep*F) are still present. In the Mcc California kid genome, a large portion of the cluster has been deleted, (8 CDS out of 12) and been replaced by 20 CDS followed by an IS*3*copy. :::  ::: ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Putative model for starch, maltodextrin and maltose utilization in *M. mycoides*subsp. *capri***. The proteins involved are shown as circles with their relevant gene names or CDS numbers. Starch is hydrolyzed by amylases extracellularly to give maltodextrin and maltose. The resulting oligosaccharides are transported into the cytoplasm by a translocation complex. Further degradation of the sugar leads to glucose-6P which can enter glycolysis (adapted from \[[@B22]\]). :::  ::: Lipoproteins ------------ The number of predicted lipoproteins (Lpp) was strikingly higher for Mmc 95010 (N = 86) than for PG1 (N = 56). Sixty of these Lpp in Mmc 95010 can be grouped into 16 families, and seven are organised in clusters (Additional files [4](#S4){ref-type="supplementary-material"}, [5](#S5){ref-type="supplementary-material"} and [6](#S6){ref-type="supplementary-material"}, figures S3, S4 and S5). Many of these clustered Lpp are absent from MmmSC PG1 or are interrupted by IS elements. In contrast, all these clusters were found in strain GM12 with a perfect synteny with the 95010 genome. However, the similarity (between GM12 and 95010) differed greatly from one Lpp to another (from 52 to 98%), whereas the flanking housekeeping genes were well conserved (mean of 98% similarity). For example, the five Lpp illustrated in Additional file [5](#S5){ref-type="supplementary-material"} Figure S4 (MLC\_9030 9040; 9050; 9070; 9080; 9090) form a family with a highly conserved signal peptide and lipoprotein cleavage site (AVIAC) and some conserved stretches in their C terminal parts. The presence of IS at this locus may be an indication that these elements played a role in the duplication of these genes. Similar Lpp are present at the same location in the GM12 genome (MMCAP2\_0900 to 0904) and were identified as putative variable lipoproteins. The similarity of these 95010 Lpp with the GM12 Lpp was between 65 and 93%. These Lpp also shared similarities with a cluster of Lpp found in Mcc (MCAP\_0593; 0594; 0595 and 0596), identified as \"Vmc\". They vary in length and display repeated AA motifs which are typical of variable lipoproteins in mycoplasmas. Such repeated motifs were not identified in the corresponding 95010 and GM12 Lpp. Differences in lipoprotein expression may be linked to the length variation of poly TA tracts within the promoter region of the Lpp. This phase variation has been described for Vmm lipoproteins in MmmSC strains \[[@B23]\]. Poly-TA tracts with more than ten repeats were found at five positions in the PG1 chromosome. Three of these positions are immediately upstream from genes encoding Lpp (MSC\_0117; 0390 and 1005) and the two others are near to transposase genes. Such poly-TA tracts are also found in the 95010 genome, at six positions. All precede Lpp genes and the number of TA repeats differs: MLC\_1150 (N = 10); MLC\_5200 (N = 10); MLC\_9030 (N = 12); MLC\_9070 (N = 16); MLC\_9120 (N = 15); MLC\_9140 (N = 15). A LC-MS/MS proteomic approach allowed the detection of 30 lipoproteins in 95010 and only 18 in MmmSC PG1, and thus a similar proportion (about one third) of the predicted genes in each strain. Of the six Lpp at the locus shown in Additional file [5](#S5){ref-type="supplementary-material"} Figure S4, only two were detected. One Mmc 95010 Lpp family, with genes downstream from an IS*1296*element and that shares some similarity with the spmA family in *M. agalactiae*, is not present in MmmSC \[[@B15],[@B16]\] (Additional file [6](#S6){ref-type="supplementary-material"} Figure S5). In *M. agalactiae*, the genes of the homologous family are also clustered and most of them are preceded by G-rich stretch that may be involved in the regulation of expression. Interestingly, in the Mmc 95010 genome, intergenic regions upstream from these genes contain a conserved GC(T)~17-20~motif 88 to 113 nt upstream from the ATG start codon. A similar motif (GC(T)~16~) is present 97 nt upstream from the ATG of the unique member of the homologous family in Mcc. The function of these motifs that are unique in these highly AT-rich genomes, like the function of *M. agalactiae*G-rich repeats, is unknown. Three genes involved in lipoprotein processing were identified in Mmc 95010: two encoding diacylglyceryl transferases (*lgt*: MLC\_8480 and MLC\_8500) and one a lipoprotein signal peptidase (*lsp*: MLC\_5520). By contrast, only one *lgt*was found in the MmmSC PG1 genome; the gene orthologous to the second *lgt*is interrupted by a frameshift mutation resulting in two \"hypothetical proteins\" (MSC\_0936 and 0937) that are not identified as *lgt*or pseudogenes in the current annotation of the MmmSC PG1 genome. Phylogenetic analysis of diacylglyceryl transferase across mollicutes reveals that most species possess only one *lgt*gene but that duplication has occurred in the ancestor of the Spiroplasma phylogenetic group. Whether the inactivation of one copy in MmmSC PG1 has affected the lipoprotein processing in this strain is not known. Pseudogenisation in MmmSC PG1 and genes interrupted by mobile elements ---------------------------------------------------------------------- The annotation process of the Mmc 95010 genome identified 45 CDS that were highly similar to two or three contiguous CDS in MmmSC PG1. These MmmSC PG1 CDS correspond to pseudogenes (N = 96) that were not identified as such at the time of the annotation because no sequence from a related mycoplasma was available (Table [3](#T3){ref-type="table"}). The proteomic approach allowed the detection of peptides for 11 of these 45 Mmc 95010 genes. Surprisingly, five products from the corresponding interrupted MmmSC genes were also detected. These products may result from the expression of a partial gene. Alternatively, there may be errors in the MmmSC PG1 sequence. In addition to disruption of genes by frameshift mutations, five PG1 genes were truncated by insertion sequences. Only three of these five genes had orthologues in Mmc (two CHP and one putative haemolysin). Conversely, four CDS in Mmc 95010 are interrupted by IS elements but only one of them was similar to a gene in MmmSC PG1 (a lipoprotein). Three Mmc 95010 genes (MLC\_1870, 3270, 4830, coding for a leucylpeptidase, a protease and an endopeptidase, respectively) had orthologues in MmmSC PG1 that are altered: the MmmSC leucylpeptidase gene is truncated, giving two CDS (MSC\_0234, MSC\_0235) by a frameshift mutation; the protease gene is interrupted by an insertion sequence (MSC\_0343); and the endopeptidase (MSC\_0504) gene is shorter by 100 codons than the Mmc 95010 ortholog. The inactivation of these three genes may explain why MmmSC PG1 is less able to hydrolyse proteins as evidenced by the inspissated serum digestion test (Table [1](#T1){ref-type="table"}). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Frameshift mutations in MmmSC PG1 ::: Mmc CDS N° Detection of Protein ^a^ Product (Mmc) MmmSC CDS Detection of Protein ^a^ ------------ -------------------------- --------------------------- --------------------- -------------------------- 0210 \- recG 0022, 0023 and 0024 \- 0350 \- CHP 0036 and 0037 \- 0370 \- CHP 0040 and 0041 \- 0630 \- TMB protein 0075 and 0076 \- 1000 \+ lipoprotein 0103 and 0104 \+ 1500 \- lipoprotein 0169, 0170 and 0171 \- 1910 \- TMB protein 0227 and 0228 \- 1870 \+ pepA 0234 and 0235 \- 1650 \- CHP 0242 and 0243 \- 2550 \- lipoprotein 0285 and 0286 \- 2540 \- lipoprotein 0285 and 0286 \- 2530 \+ TMB protein 0287 and 0288 \+ 2330 \- protein phosphatase 0307 and 0308 \- 2320 \- TMB protein 0309 and 0310 \- 2290 \- TMB interrupted by ICE 0314 and 0315 \- 3100 \- lipoprotein 0317 and 0318 \- 3270 \- TMB and protease 0343 and 0344 \- 3280 \- lipoprotein 0345 and 0346 \- 3560 \- CHP 0378 and 0379 \- 3790 \- TMB protein 0399 and 0400 \- 4410 \- putative abc transporteur 0459 and 0460 \- 4510 \- TMB protein 0470 and 0471 \+ 4520 \+ TMB and efflux 0472 and 0473 \+ 4840 \- endopetidase O 0504 \- 4930 \- TMB substrate transport 0513, 0514 and 0515 \- 4980 \- TMB protein 0521, 0522 and 0523 \- 5170 \- TMB protein 0536 and 0537 \- 5250 \+ CHP 0545 and 0546 \- 5570 \- CHP 0589 and 0590 \- 5940 \- lipoprotein 0631 and 0632 \- 5970 \- TMB protein 0636 and 0637 \- 6080 \+ CHP 0652 and 0654 \- 6100 \+ CHP 0656, 0657 and 0658 \- 6510 \- TMB protein 0694 and 0695 \- 6580 \- TMB protein 0705 and 0706 \- 6620 \+ lipoprotein 0710 and 0711 \+ 7000 \- transcriptional regulator 0750 and 0751 \- 7010 \- TMB protein 0753 and 0754 \- 7200 \+ lipoprotein 0772 and 0773 \- 7340 \+ HAD hydrolase 0792 and 0793 \- 7420 \- CHP 0819 and 0820 \- 8740 \- TMB protein 0919 and 0920 \- 8500 \- lgt 0936 and 0937 \- 8860 \+ CHP 1012, 1013 and 1014 \- 8970 \- TMB protein 1023, 1024 and 1025 \- a \"+\" indicates that the protein was detected by LC-MS/MS with at least 2 disctinct peptides. MmmSC CDS should be considered as pseudogenes unless there was a sequencing error leading to an artefactual frameshift mutation. **List of CDS with frameshift mutations in the MmmSC PG1 genome as evidenced by comparison with Mmc CDS**. ::: The insertion of these 2 ICE copies in the Mmc 95010 genome has disrupted two CDS. The first encodes a putative lipoprotein (MLC\_2070; 142 AA) whose orthologue in Mmc GM12 (MMCAP2\_0208) is 263 AA long. The second encodes a putative permease, 1624 AA long, annotated as a pseudogene in 95010 (MLC\_2290), whose orthologue in GM12 (MMCAP2\_0272) is 1773 AA long and possesses two FtsX domains that may be associated with lipid transport. A gene orthologous to this permease was also found in the MmmSC PG1 sequence (MSC\_0033, 1796 AA long). Global comparison of gene repertoires ------------------------------------- A search for orthologous genes in the Mmc 95010, MmmSC PG1 and Mcc California kid genomes identified various clusters of orthologs that were specific to each genome or shared by two or three genomes (Figure [7](#F7){ref-type="fig"} Additional file [7](#S7){ref-type="supplementary-material"} table S2). ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Number of orthologous genes in the Mmc 95010, MmmSC PG1 and Mcc California kid genomes**. Numbers of orthologous genes in Mmc 95010, MmmSC PG1 and Mcc California kid genomes were evaluated using the MBGD database <http://mbgd.genome.ad.jp/>. The core genome of these three strains consists of 658 genes. Letters refer to supplementary table S2 giving the list of orthologues. :::  ::: The majority of the 113 PG1-specific clusters (Additional file [7](#S7){ref-type="supplementary-material"} table S2, sheet E), correspond to hypothetical proteins and predicted lipoproteins, most (63%) of which, however, are considered to be pseudogenes. In addition, two clusters identified as an Aspartate-tRNA ligase (*asp*S, MSC\_0432) and a CTP synthetase (pyrG, MSC\_0902) are in fact partial sequences that were erroneously annotated and should also be considered as pseudogenes. MSC\_0902 maps near two other CDS that are duplicated, probably when an IS element transposed (IS*1296*copy). The same event may have occurred for three clusters identified as *cps*, *glf*and *gal*E (MSC\_0970; 0974 and 0985) which are partial duplicated copies of genes with orthologues in 95010. These sequences should also have been annotated as \"putative pseudogenes\". In addition, PG1-specific clusters included restriction modification genes, notably an adenine or cytosine methylase (MSC\_0950 and 0951). As expected, some transposase genes and genes belonging to the maltodextrin/maltose gene cluster were identified as Mmc 95010-specific CDS (Additional file [7](#S7){ref-type="supplementary-material"} Table S2, sheet F). One CDS coding for a predicted MatE efflux family protein (Multi Antimicrobial Extrusion) was identified (MLC\_6090). These proteins mediate resistance to a wide range of cationic dyes, fluroquinolones, aminoglycosides and other structurally diverse antibodies and drugs, and most possess 12 alpha-helical transmembrane regions. This is the case for MLC\_6090 in strain 95010 but apparently not in strain GM12 in which the CDS has been disrupted resulting in two CDS which should be considered to be pseudogenes. Some ribosomal proteins, such as *rpm*J, *rps*S, *rps*N and a preprotein translocase *sec*G, were identified as specific for a subset of strains (Additional file [7](#S7){ref-type="supplementary-material"} Table S2, sheet C). This may be a consequence of the very small size of these proteins which resulted in very low blast e values or, more frequently, of incorrect annotation of the PG1 genome. This could be evidenced by a BLASTN search that identified *rpm*J and *rps*N in the PG1 sequence at positions 830855 and 836897, respectively, and the identification of *sec*G at position 973088. The 658 CDS clusters that were found as orthologous in all of MmmSC PG1, Mmc 95010 and Mcc California kid (Additional file [7](#S7){ref-type="supplementary-material"} Table S2, sheet A) may be considered as the \"*M*. *mycoides*core genome\". A blast analysis of whole CDS in Mmc strains 95010 and GM12 showed that 80% of these CDS shared more than 90% identity. This illustrates the very close relatedness of the two genomes and confirms their classification as belonging to the same subspecies. The divergent CDS had identity values that varied from 20 to 90% (Additional file [8](#S8){ref-type="supplementary-material"} Table S3). Many corresponded to partial sequences that may be considered as pseudogenes and the lower identity values may be due to these sequences being short or to the fact that mutations are likely to accumulate more rapidly in pseudogenes than in functional genes. Other divergent genes were identified as lipoproteins and the percentage of identity for these genes varied from 30% to 90%. Similar polymorphism was also evidenced in membrane surface expressed proteins highlighting the highly dynamic and diverse surface architecture of otherwise closely related strains. This is undoubtedly a consequence of environmental pressures that shape the variability of these surface exposed proteins. This variability also explains why reference identification tests such as the growth inhibition test may give such variable results as hyperimmune sera contain large amounts of antibodies directed towards these immunodominant antigens. Divergent CDS also included restriction and modification genes such as \"C5 Cytosine DNA methylase\" (MMCAP2\_0554), \"GCATC recognizing Type II restriction modification system\" (MMCAP2\_2020; 0900 and 0920), \"restriction endonuclease\" (MLC\_\_1620), \"modification methylase\" (MLC\_1630), \"cytosine-specific DNA methyltransferase *Sau*96I\" (MLC\_2020) and \"Type II site specific deoxyribonuclease (*Sau*96I like)\" (MLC\_2030). Discussion ========== Our goal was to compare the genomes of two very closely related Mycoplasma subspecies and accurately determine their degree of relatedness. This can be done at various levels by comparing the genome organization, the gene repertoire and the polymorphisms within the genes. Genome plasticity in mycoplasmas of the mycoides cluster is greatly influenced by the \"mobilome\" and more specifically by ICE and IS. Such elements were evidenced in the 95010 genome and have been shown to drive overall genome plasticity. Integrative conjugative elements have a modular structure and contain blocks of genes dedicated to integration into and excision from the chromosome, as well as conjugal transfer \[[@B24]\]. Until recently, ICEs had been evidenced in a limited range of hosts belonging to the four major divisions of bacteria. However, whole genome sequencing projects suggest that ICEs are widespread in bacteria and could be one of the main types of shuttle for horizontal gene transfer \[[@B25]\]. ICEs have now been identified in various *Mycoplasma*species, including *M. fermentans*, *M. agalactiae*and Mcc \[[@B16],[@B26],[@B27]\]. The ICE copy number in mycoplasmas seems to be small, for example only four copies of two ICE types in *M. fermentans*. ICE copy numbers are much higher in other bacterial species: in *Orienta tsutsugamushi*duplicated elements, including ICEs, account for more than 37% of the genome \[[@B28],[@B29]\]. Our analysis of the Mmc genome shows that these elements have a direct impact on genome rearrangements, although the exact mechanisms leading to excision, integration and/or conjugation to another cell remain to be elucidated. In the genus *Bacillus*, transfer of ICE copies seems to be favoured by high densities of cells not carrying these elements and integration into a cell apparently leads to blocking the entry of additional copies \[[@B30],[@B31]\]. The possibility that these elements are involved in transferring virulence factors to and between mycoplasmas needs to be investigated. Comparison of plasmids from the mycoides cluster suggested that various recombination events may have occurred during the spread of these plasmids among strains. More surprisingly, alignments of plasmid and ICE sequences in Mmc 95010 indicated that these mobile elements may have exchanged sequences. This new finding suggests that these two types of mobile elements could interact within mycoplasma cells and maybe even cooperate in transmission from cell to cell. Insertion sequences are another driving force for genome plasticity in the mycoplasma mycoides cluster. Comparisons of genomes revealed substantial diversity of IS type and copy number even between closely related strains (95010 and GM12) and that duplication of IS copies may lead to large DNA fragment inversions. This contrasts with the findings for two *M. agalactiae*genomes in which the presence of 15 IS copies and three ICE copies was not associated with any large-scale genetic rearrangement \[[@B15]\]. In the case of the *mycoides*cluster, the major contribution of IS to genome plasticity is well illustrated by the comparison between Mmc and MmmSC genomes. The MmmSC PG1 genome has large numbers of IS copies, IS*1634*having the highest copy number (N = 60). These IS-elements have not only led to large DNA fragment inversions but also large DNA fragment duplications and deletions. This is not unprecedented in the bacterial world and IS expansions may result from an evolutionary bottleneck due to bacterial population isolation \[[@B32]\]. In the case of MmmSC PG1, this bottleneck may have been associated with the strict adaptation of this subspecies to the bovine lung. Indeed, IS*1634*shares 97% identity with IS*Mbov3*which is found in *M*. *bovis*, a very common pathogen isolated from cow lungs \[[@B33]\]. This close relatedness certainly indicates recent HGT and the absence of IS*1634*from Mmc suggests that this IS was acquired by MmmSC from *M. bovis*. Such exchanges of IS in between the *mycoides*cluster and the *M. bovis/M. agalactiae*cluster have already been proposed \[[@B33]\]. A striking feature of the MmmSC PG1 genome, as compared to those of Mmc 95010 and GM12, is the large number of pseudogenes in the vicinity of IS elements. Altogether, more than 98 MmmSC of the originally described putative genes are certainly pseudogenes as a result of frameshift mutations or inserted insertion sequences. This represents more than 9% of the total number of MmmSC genes that were annotated in 2004. High percentages of pseudogenes are often associated with a recent adaptation to a host and to virulence, as suggested for *Yersinia pestis*\[[@B34]\]. Adaptation to a new host allows a massive clonal population growth in which all mutations affecting genes that are not essential for bacterial survival in the new environment are maintained. Such clonal expansion also explains the limited polymorphism of the housekeeping genes. Reductive evolution of this type has been described for various pathogens in addition to *Yersinia pesti*s, including *Orienta tsutsugamushi*, the agent of scrub typhus, *Ricketsia prowazeckii*, the agent of epidemic typhus and *Aliivibrio salmonicida*, the agent of cold-water vibriosis \[[@B29],[@B35],[@B36]\]. In the case of MmmSC, adaptation to a new host may also have favoured the acquisition and multiplication of new IS types, such as IS*1634*. Similarly, pseudogenisation was also observed in *M. bovis*, where an adhesin of *M. agalactiae*was inactivated upon infection of a different host \[[@B37]\]. Longer evolution times would possibly allow a streamlining of the genome with a reduction of the number of pseudogenes by a deletion process. Mmc is an ubiquitous pathogen that is present in numerous species (sheep, goats) all around the world; it is an opportunist pathogen that can infect diverse organs and can even be found in the ear canal of healthy goats (or in parasites found in the ear canal). By contrast, MmmSC is strictly pathogenic and limited to a single host, cows, and to a single organ, the lung. This is consistent with an Mmc ancestor, adapted to various ecological niches in small ruminants, adapting to a bovine host where it colonizes only the lungs, and evolving into what is now known as MmmSC. Genomic studies, and particularly the observation that intraspecies polymorphism in housekeeping genes is much more limited in MmmSC than in Mmc, support the hypothesis that MmmSC emerged only recently \[[@B38]\]. The availability of whole genome sequences may help unravel the genetic events underlying phenotypic diversity among closely related strains. As an example, the utilization of maltose in the *M. mycoides*cluster species has been studied by AbuGroun \[[@B39]\]: no maltose utilization was found in MmmSC. Maltose is utilized by Mcc and more rapidly by Mmc. However, some Mmc strains failed to metabolize maltose at all. The presence of an alpha glucosidase was also detected by a rapid colorimetric test using pNPG flooded on mycoplasma colonies \[[@B40]\]. None of the MmmSC strains tested possessed any glucosidase activity although most Mcc and Mmc strains did. However MmmSC strains express beta-glucosidases with variations which may be related to cytotoxicity \[[@B41]\]. Our findings are in accordance with these observations. What remains to be verified is the integrity of the maltose gene cluster in the Mmc strains that fail to utilize this substrate, and the ability of Mmc strains to utilize starch. Mcc California kid is expected to be unable to use starch, although other Mcc strains should be able metabolize this carbon source. Surface proteins and more specifically lipoproteins that play key roles in interactions with the environment are determinant for the lifestyle of mycoplasmas. They contribute to the uptake of nutrients and can mediate essential functions during the infection cycle. Some play a role in cytadhesion, and other bind IgAs to allow the cells to escape cellular recognition \[[@B42]\]. Surface proteins can also display mechanisms of phase variation as a means to escape the host immune responses \[[@B43]\]. At the same time they are excellent immunogens, their lipid moiety acting with adjuvant-like proinflammatory activity and their protein part evoking an immune response \[[@B44]\]. However, the type of immune response they trigger may vary according to the Lpp involved. In the case of MmmSC PG1, LppA seems to trigger a cellular response, involving CD4 cells producing interferon gamma, whereas LppQ, LppB and LppC do not \[[@B45]\]. The presence of 86 genes coding for Lpp in the Mmc 95010 genome, as compared to 56 in MmmSC PG1, is in agreement with findings in other mycoplasma species. The number of Lpp genes in two strains of *M. agalactiae*is 100 and 67, the latter number being that of reference strain PG2 \[[@B15]\]. In *M. agalactiae*, poly-G tracts are suspected to be involved in genomic rearrangements and possibly in the control of expression of genes in the region encoding the so-called spma lipoprotein family \[[@B15],[@B16]\]. A locus encoding homologous predicted lipoproteins was found in Mmc 95010 with intergenic regions containing GC(T)~7-20~motifs. This suggests that there may have been exchange of genes belonging to this family between these ruminant pathogens followed by divergent evolution of intergenic motifs and subsequent expansion of the gene families. In accordance with this hypothesis, only one single member of this family was found in Mcc genome whereas variable expansions were observed in the two strains of *M. agalactiae*. In *M. mycoides*, intergenic nucleotide tracts are found at other loci. Poly-TA tracts, with more than 10 repeats, were found at six locations both in the 95010 and the GM12 genomes. The size of these tracts differed between strains. However, these size variations should be interpreted with great care as most sequencing projects use cloned bacterial stocks: such variants may differ from the main population. In 95010, the percentage of Lpp that were detected by the proteomic analysis was 30%. This is slightly lower than reported by other studies in which amphiphilic proteins were first concentrated by Triton X-114 extraction \[[@B15]\]. As a consequence, the differences may simply be due to differences in the sensitivity detection of the techniques used. In addition, Lpp expression may be driven by environmental conditions and our results apply only to mycoplasmas grown in rich media, *in vitro*. Co-incubation and adhesion to cells may well trigger the expression of a different set of Lpp as has been demonstrated for *M. pneumoniae*in contact with lung epithelial cells \[[@B46]\]. This type of modulation of expression may play an important role in virulence. In fact, the evolution of MmmSC genome may be shaped by unconstrained population growth in infected animals, followed by extreme transmission bottlenecks from host to host. Furthermore, current MmmSC strain populations may also be shaped by CBPP control strategies based on slaughter and vaccination. Existing MmmSC strains may well have adapted to this artificial selection that has been implemented for more than 100 years. The MmmSC genome is larger than that of Mmc, mostly due to gene duplications, and the insertion of multiple copies of Insertion Sequences. IS elements seem to play a prominent role in this gene rearrangement process, as demonstrated during growth *in vitro*under conditions of stress induced by high temperature (41.5°C) \[[@B47]\]. Fever in CBPP-infected animals may induce a similar stress and favour gene rearrangements in MmmSC. However, the MmmSC genome is also characterized by a high degree of gene decay with more than 9% of the originally described genes likely to be pseudogenes. Many of these genes are not associated with any known function (\"ORFans\"), consistent with the notion that the number of genes in whole genomes is often overestimated \[[@B48]\]. This fits also well with a non-adaptative genomic complexity theory allowing duplications or pseudogenes to be maintained in the absence of an adaptive selection that would lead to purifying selection and genome streamlining \[[@B49],[@B50]\]. Genome structure in both of the subspecies seems to have been affected by mobile genetic elements despite these elements differing in kind and in numbers. Integrative conjugative elements have been identified in Mmc where they were shown to induce chromosomal rearrangements, but not in MmmSC. They may also have played a role in gene acquisition although this has not yet been demonstrated. Insertion sequences were identified in both subspecies but, here again, there are differences: Mmc and MmmSC have only two IS types in common and MmmSC possess only three IS types present in large copy numbers (95 copies) whereas Mmc possess five IS types but only in lower copy numbers (24 copies). Again the larger copy number in MmmSC may be associated with an evolutionary bottleneck such that they provide transitory selective advantages to their host such as HGT and genomic rearrangements \[[@B32]\]. Homologous recombination has been demonstrated experimentally in Mcc and Mmc strains \[[@B51]\]. This does not seem to be the case in MmmSC where multiple attempts to obtain homologous recombinations *in vitro*have failed \[[@B52]\]. These failures could be linked to the functional absence of two genes, *recG*and *recR*, which are disrupted by frameshift mutations. Conclusions =========== Comparing the genomes of two subspecies allows a micro-evolutionary analysis. Genome evolution is expected to be directly linked to the ecological niche of the two organisms and the driving forces shaping chromosome organization, gene content and sequence evolution. Preliminary data, using multilocus sequence typing, had already shown that Mmc strains were much more diverse than MmmSC strains \[[@B38]\]. This is consistent with the different ecology of the two organisms and the associated population dynamics. The comparison of the Mmc and MmmSC genomes has revealed their very close relatedness, especially evident from the sequence similarities in their housekeeping genes. However, in view of the earlier belief that the two subspecies were nearly indistinguishable, there is a surprisingly large number of differences. Many of the differences are associated with genes of unknown function, and many of these genes may have been acquired by HGT. Note also that comparison of single strain genomes is limiting because a single genome cannot reflect the gene repertoire of a species. In addition, the comparison with MmmSC PG1 could lead to some bias as this strain is a \"laboratory\" strain which has lost its pathogenicity and its genome has been subject to genetic drift since its, unknown, time of isolation. Understanding the real driving forces of genetic fluidity will certainly need further sequencing, whole genome assemblies and re-annotation of many well-characterised field strains, not only of Mmc and MmmSC but also of closely related species or and other bacterial species that share the same ecological niche. Methods ======= Bacterial strains ----------------- Mmc strain 95010 was isolated from a young female goat (alpine breed) with polyarthritis in February 1995 near Bourges (France). It was identified, at the time, by biochemical tests (Digitonine+, film and spots -, glucose+, tetrazolium reduction +, arginine - and phosphatase +) and by growth inhibition tests (positive with antiserum against strain YG and negative or partial with antisera against strains PG3, PG1, PG50, F38, California kid). It was subsequently characterized by multilocus sequence typing with *fus*A, *glp*Q, *gyr*B, *lep*A and *rpo*B partial gene sequences \[[@B38]\]. The strain was cloned to ensure its purity and \"clone C1\" was used thereafter. Other strains were included in the study (see Table [2](#T2){ref-type="table"} for the full list), notably to investigate the presence of repeated elements in various strains of each species or subspecies of the \"mycoides cluster\" or species that are found in ruminants. Sequencing and annotation ------------------------- The complete sequence of Mmc 95010 was obtained by a shotgun strategy. To prevent cloning bias, three plasmid libraries were obtained after mechanical shearing of genomic DNA and ligation of 3 kb (A) and 4 kb (D)/10 kb (B) fragments into pNAV (derived from pcdna2.1) and pCNS (derived from pSU18) vectors, respectively. In addition, large inserts (25 kb (C)), generated by *Hin*dIII partial digest, were introduced into pBeloBAC11. Vector DNAs were purified and end-sequenced (n = 15744 (A), n = 11424 (B), n = 1536 (C), n = 5184 (D)) using dye-terminator chemistry on ABI3730 sequencers. To reduce repeated sequence assembly problems, a pre-assembly was performed using the Phred/Phrap/Consed software package <http://www.phrap.com>\[[@B53]\]. The finishing step was achieved by primer walks and PCR and transposon bomb libraries. A total of 990 sequences were needed for gap closure and quality assessment. Genome annotation was performed using the CAAT-Box platform which was customized to facilitate the annotation process \[[@B54]\]. CDS were first detected using the Genemark software implemented in the CAAT-Box environment \[[@B55]\]. Putative CDS of more than 300 amino acids were used to train the Markov model (order 5). The three codons AUG, UUG, and GUG were used as potential start codons, and UAG and UAA were defined as stop codons. Once trained, the Markov model was applied to the complete genome using 80 bp as a cut-off value for the smallest CDS. Prediction of CDS with CAATBox also integrates results of BLAST searches to discriminate highly probable CDS from false ORFs \[[@B56]\]. The databases used for this purpose were SwissProt [http://www.ebi.ac.uk/swissprot/index. html](http://www.ebi.ac.uk/swissprot/index.html), trembl <http://www.ebi.ac.uk/embl/index.html>, and MolliGen <http://molligen.org>, a database dedicated to the comparative genomics of mollicutes \[[@B57]\]. To determine the extent of sequence similarity, alignments between predicted proteins and best BLAST-hit sequences were performed using NEEDLE software implementing the Needleman-Wunsch global alignment algorithm and using the BLOSUM62 matrix \[[@B58]\]. During the annotation process, proteins were considered to be homologues if the similarity in these alignments exceeded 40%. Predicted proteins with lower or only local similarities with previously characterized proteins were annotated as hypothetical proteins. Most start codons were identified according to CAAT-Box recommendations that resulted from both Genemark coding state prediction and BLAST results analysis. For CDS showing neither obvious homology relationships nor clear coding curves, the most upstream start was chosen, with a preference for the most frequently used AUG codon. Other tools incorporated into CAAT-Box were also used to improve annotation and function predictions: they included Inter-ProScan \[[@B59]\] and PrositeScan for domain detection and TMHMM for trans-membrane segments prediction \[[@B60],[@B61]\]. To recover small CDS or gene fragments that may have been discarded during the CDS prediction process, intergenic sequences of more than 80 bp were systematically compared to reference databases using BLASTX. The annotation of each CDS was manually verified by at least two annotators. The tRNAs were mapped on the chromosome using tRNAscan-SE software and the rRNA genes were identified by BLASTN searches for homology with rRNA genes in MmmSC \[[@B62]\]. Precise boundaries were established after comparisons with the sequences stored in the European Ribosomal RNA Database <http://www.psb.ugent.be/rRNA> and the 5 S Ribosomal RNA Database <http://www.man.poznan.pl/5SData>\[[@B63],[@B64]\]. The *rnp*B gene of the RNase P system and the tmRNA were sought by BLASTN searches for sequence similarity with homologues from Mcc and MmmSC, respectively. Whole genome structure comparison --------------------------------- Mmc 95010 genome was aligned with those of MmmSC PG1, Mmc GM12 and Mcc California kid with Multiple Genome Alignment software MAUVE v2.2.0 <http://gel.ahabs.wisc.edu/mauve>\[[@B65]\]. This software allows the identification of locally collinear blocks (LCB) that are conserved segments that appear to be internally free from genome rearrangements. Genomes were displayed on a map along lines representing the whole sequences with LCBs of various colours. Each of these block outlines surrounds a region of the genome sequence that aligned to part of another genome. When a block lay above the centre line, the aligned region was in the forward orientation relative to the first genome sequence. Blocks below the centre line indicate regions that aligned in the reverse complement (inverse) orientation. Regions outside blocks lacked detectable homology among the input genomes. Mobile elements --------------- Integrative conjugative elements (ICE) were identified by homology search and by similarity with other ICEs previously identified in mycoplasma species. Insertion Sequences (IS) in the Mmc genome were identified during annotation. The total number of IS copies was identified by \"auto-blast\" in the CAAT-box platform and by using 40 bp-long motifs taken at the extremities and in the middle of each IS type. Search for these motifs with Vector NTI (v:10.3.1) (Invitrogen) allowed the identification of additional truncated copies that may have been missed during annotation. The positions of repeat elements in the PG1 sequence were also retrieved from the MolliGen database. The IS copies were then positioned on a ruler along the genome and placed on the MAUVE alignment. New IS types have been submitted to the dedicated web server \"IS Finder\" <http://www-is.biotoul.fr/>. Gene repertoire, homologies and orthologous search -------------------------------------------------- Orthologous genes in Mmc 95010, MmmSC PG1 and Mcc California kid were investigated using the MBGD database <http://mbgd.genome.ad.jp/> with standard parameters \[[@B66],[@B67]\]. A BLASTP comparison was performed with all CDS from Mcc 95010 and Mcc GM12 to evaluate the similarities at each CDS level. The distribution of similarity results allowed the identification of CDS that were specific for each of the two genomes. Proteomics ---------- Strains MmmSC 8740-Rita and Mmc 95010 were grown in Hayflick medium at 37°C and harvested by centrifugation at the late exponential phase of growth. Whole mycoplasma pellets were washed twice in PBS and then solubilized in Laemmli buffer. Samples were subjected to electrophoresis in a 10% SDS acrylamide gel. Gels were sliced, treated with acetonytril, dried, and digested with trypsin. The resulting peptide mixtures were concentrated and analyzed on a Dionex U-3000 Ultimate nano LC system coupled to a nanospray LTQ XL IT mass spectrometer (Thermo-Finnigan, San Jose, CA). Data were acquired in a data dependent mode alternating an MS scan survey over the range m/z 300-1700 and three MS/MS scans in an exclusion dynamic mode. Data were searched by SEQUEST through Bioworks 3.3.1 interface (Thermo-Finnigan) against the CDS of MmmSC PG1 and the CDS of Mmc 95010. DTA files were generated for MS/MS spectra that both reached a minimal intensity, 1.10e3, and a sufficient number of ions, 10. Authors\' contributions ======================= FT and AB designed the initial genome sequencing project. VB and BV carried out the DNA libraries, sequencing, finishing and assembly of the 95010 sequence. FT, LMS and PSP carried out the genetic analysis and manual expert annotation of the genome. WS and VD analyzed the distribution of Insertion Sequences and ICE copies. AML carried out the LC-MS/MS whole proteomic sequencing. MB analyzed the plasmid and ICE sequences. DJ adapted the bioinformatic server used for the annotation. FT and PSP drafted the manuscript. FT, PSP and AB revised it. All authors read and approved the final manuscript Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **\"PCR primers used during this study\"**. This additional table lists the names and sequences of primers used in the study. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **\"Comparison of related plasmid and ICE sequences in Mmc 95010\"**. This figure describes the positions and sequences shared by the integrative conjugative elements copies and the plasmid that was characterized in the same strain. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **\"Comparison of plasmids from the mycoides cluster\"**. This figure depicts the results of a three-way BLASTN comparison of three mycoplasma plasmid sequences using the Artemis Comparison Tool \[[@B71]\] ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **\"Organization of lipoprotein gene clusters in the Mmc 95010 genome\"**. This figure is a schematic representation of 6 lipoprotein families (A to F) found in the genome of strain 95010. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 5 **\"organization of a Mmc-specific lipoprotein gene clusters in the Mmc 95010 genome\"**. This figure is a schematic representation of a lipoprotein family found in strain 95010 genome but not in the published MmmSC genomes. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 6 **\"Peculiar Mmc genomic locus encoding predicted surface proteins\"**. This figure is a schematic representation of the locus and homologies with *M. agalactiae*or *M. capricolum*subsp. *capricolum*genes. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 7 **\"List of gene clusters identified as orthologous\"**. This additional table lists the names of genes found orthologous by the **Microbial Genome Database for Comparative Analysis**<http://mbgd.genome.ad.jp/> ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 8 **\"List of CDS in Mmc strains GM12 and 95010 showing less than 90% similarity with each other\"**. This additional table lists the names of genes that are not highly similar in two strains belonging to the same subspecies, *M. mycoides*subsp. *capri*. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This research was supported by a grant from Genoscope (Evry France). We are greatly indebted to the French Embassy in Addis Abeba for a grant supporting Woubit Salah. We would like to thank Armelle Peyraud who cultivated the *Mycoplasma*strains and to all colleagues that supplied these strains. We also thank Laure Béven for advices on maltodextrine and maltose metabolism.

PubMed Central

2024-06-05T04:04:19.589597

2011-2-16

PMC3053260

Background ========== With two cattle genome assemblies available (Btau\_4 and UMD3)\[[@B1],[@B2]\], the cattle research community has been focusing on single nucleotide polymorphisms (SNPs) as the main source of genetic variation in cattle. This effort led to the development of the cattle SNP map \[[@B3]\] and the Illumina Bovine SNP50 (\> 50,000 SNP probes) genotyping array \[[@B4],[@B5]\]. Evaluations of genetic merit based on SNPs became a reality in early 2009 leading to an acceleration of improvements to dairy and beef breed stocks \[[@B6]-[@B8]\]. Widespread use of the BovineSNP50 array has resulted in the availability of tens of thousands of SNP genotyping results. Based on SNP genotyping assays, QTL distributions and artificial selection signatures in dairy cattle have been reported \[[@B9],[@B10]\]. Copy Number Variation (CNV) represents another important source of genetic variation that provides genomic structural information complementary to SNP data. Genomic structural variations ranging from 1 kb to 5 Mb comprise mainly of CNVs in the form of large-scale insertions and deletions, as well as inversions and translocations \[[@B11]\]. In humans, \~29,000 CNVs that correspond to over 8,400 CNV regions have been identified, and over 9,000 genes have been mapped within or near regions of human structural variation \[[@B12],[@B13]\]. Some of these CNVs have been shown to be important in both normal phenotypic variability and disease susceptibility. Several recent publications have reviewed the effects of CNVs on gene expression and human diseases \[[@B14]-[@B17]\]. Due to their low cost and high-density, SNP arrays have been routinely used for human CNV detection and analysis \[[@B13]\]. Compared to CGH arrays which only report relative signal intensities, SNP arrays collect normalized total intensities (Log R ratio - LRR) and allelic intensity ratios (B allele frequency - BAF) which represent overall copy numbers and allelic contrasts \[[@B18]\]. Multiple algorithms have been developed to exploit SNP data to identify CNVs, including QuantiSNP \[[@B19]\], PennCNV \[[@B20]\], Birdseye \[[@B21]\] and Cokgen \[[@B22]\]. Comparisons of the strengths and weaknesses of these algorithms have been published \[[@B23],[@B24]\]. As one of the leading methods, PennCNV incorporates multiple sources of information, including total signal intensity and allelic intensity ratio at each SNP marker, the distance between neighboring SNPs, and the allele frequency of SNPs. PennCNV also integrates a computational approach by fitting regression models with GC content to overcome \"genomic waves\" \[[@B25],[@B26]\]. Furthermore, PennCNV is capable of considering pedigree information (a parents-offspring trio) to improve call rates and accuracy of breakpoint prediction as well as to infer chromosome-specific SNP genotypes in CNVs \[[@B27]\]. Previous cattle studies have produced a number of CNV datasets. For example, our earlier array CGH survey using 3 Holstein bulls identified 25 germline CNVs \[[@B28]\]. Recently, we reported a broader, systematic CNV survey in 90 cattle using array CGH \[[@B29]\]. We identified over 200 candidate CNV regions (CNVRs); some of which are likely to underlie cattle domestication and breed formation. Fadista et al. recently reported 304 CNV regions in 20 animals of 4 cattle breeds using high-density array CGH \[[@B30]\]. Besides array CGH experiments, other evidences for cattle CNV came from SNP genotyping results, where a screen of Bovine HapMap Consortium samples (over 500 animals from multiple cattle breeds) identified 79 candidate deletions using an earlier version of cnvPartition \[[@B5]\]. However, these results only included homozygous deletions which were validated by multiple observations. A recent paper reported 368 unique CNV regions from 265 Korean Hanwoo cattle based on BovineSNP50 genotyping data; however, during the PennCNV calling, the \"genomic waves\" pattern was not discussed and pedigree information was not considered \[[@B31]\]. In this study, we reprocessed the published Bovine HapMap Consortium SNP genotyping results using optimal settings for PennCNV by adjusting for \"genomic waves\" and utilizing trio/pedigree information whenever possible. We identified 682 candidate CNV regions in a diverse panel of 521 animals from 21 different breeds. We also included 18 animals from 6 outgroups to derive the ancestral states of CNVs. We then compared this CNV call set with the existing cattle CNV call sets, validated several novel CNVR calls and discussed the evolutionary impact of cattle CNVs. Results and Discussion ====================== Optimization of cattle CNV detection ------------------------------------ A total of 58,336 markers were selected for the BovineSNP50 assay \[[@B4],[@B5]\]. Except for 1,389 markers which failed to pass manufacturer assay production pipeline, we intentionally kept all remaining 56,947 markers without any other filtering. These included 1,465 markers (2.57%) which had a call rate of 0. The markers with a call rate of 0 are resistant to the default biallelic SNP clustering and often fall in CNV regions. Compared to the standard BovineSNP50 Genotyping Beadchip v1 featuring 54,001 SNP probes, 2,946 more SNPs were included in our analysis, of which, \~17% located in cattle segmental duplication (SD) regions \[[@B32]\], \~9% overlapped with the CNVRs detected by array CGH method \[[@B29]\], and \~27% contributed to the CNVRs reported here. We tested the cattle CNV calls made with or without the -gcmodel option on Batu\_4.0 to identify the impact of genomic waves on CNV calling. Agreeing with previous results \[[@B26]\], we found the total CNVR counts were higher without -gcmodel (719) than those with -gcmodel enabled (682). However, only 86.80% (592/682) of the gcmodel calls directly overlapped with 79.28% (570/719) calls made without gcmodel, revealing a \~20% CNV discordancy rate. These discordant calls were likely due to false positives called from the differentiating signal intensities caused by \"genomic waves\" rather than by real CNV events. This further demonstrated that genomic waves have a significant effect on this type of analysis. We also compared results of PennCNV using -test, -trio and -joint options sequentially. In other words, we compared data resultant from not considering trio information (-test), considering trio information only after calling (-trio) and finally by considering trio information in a simultaneous fashion during CNV calling (-joint) (Additional file [1](#S1){ref-type="supplementary-material"}: Table S5). Consistent with the earlier comparisons using simulated and real SNP data \[[@B27],[@B33]\], trio information significantly increased our CNV call rates. The result of the -joint option (1276 calls) was significantly higher than those of the other options: -test (684 calls) and -trio (1019 calls). After merging overlapping CNVs, \~87% of the 682 CNVRs deduced from the -joint option overlapped with those deduced from the -test and -trio options (both with a total of 621 CNVRs). Due to its improved call sensitivity and breakpoint inference, the -joint option reported about 13% more CNVRs which were not detected by the -test or -trio options. Cattle CNV discovery and distribution ------------------------------------- Due to issues regarding CNVR calls, we excluded chrX and chrUn from our analysis. In our initial analysis of chrX, it was found that almost half of the potential CNVRs were unreasonably large (\> 1 Mb) and several events were present in high frequencies (\> 25%). This is likely due to the fact that PennCNV assumes two copies of each SNP as the normal copy number state, which was likely not the case within the pseudoautosomal region \[[@B34]\] and segmental duplications \[[@B32]\] on chrX. Additionally, since chrX sequence and annotation also differ dramatically between Btau\_4.0 and UMD3 builds, we considered the CNV calls on chrX as unreliable and excluded them from further analysis. Since chrUn only contains unassigned sequence contigs, it was not included due to the lack of sequence and SNPs as well as the SNP mapping uncertainty. Within the placed autosomes, a total of 3,666 CNVs in 521 samples were detected and an average of 7.09 gain or loss events were evident in each sample (Table [1](#T1){ref-type="table"}). CNVRs were determined by aggregating overlapping CNVs identified across all samples, following previously published protocols \[[@B13]\]. A total of 682 high-confidence autosomal CNVRs were identified, covering 139.8 Mb of polymorphic sequence and corresponding to 5.49% of the autosomal genome sequence (139.8/2,545.9 Mb) and 4.60% of the whole cattle genome (139.8/3,036.6 Mb, Figure [1](#F1){ref-type="fig"} and Additional file [1](#S1){ref-type="supplementary-material"}: Table S2). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### CNV events by species and breeds. ::: *Btau\_4.0* *Sample* *Count* *Unique* *Gain* *Loss* *Gene* *Total Length* ---------------- ---------- -------------- ------------ ------------- ------------- --------------- ---------------------- Taurine^a^ 366 2,256(6.23) 239(0.66) 1,454(4.02) 802(2.22) 4,744(13.10) 373,001,599(165,337) Composite 46 330(7.17) 23(0.50) 224(4.87) 106(2.30) 651(14.15) 113,483,966(142,032) Indicine 70 799(11.41) 62(0.89) 401(5.73) 398(5.69) 1,464(20.91) 57,402,891(173,948) African Breeds 39 281(7.21) 38(0.97) 213(5.46) 68(1.74) 775(19.87) 54,728,022(194,761) CNV^a^ 521 3,666(7.09) 362(0.70) 2,292(4.43) 1,374(2.66) 7,634(14.77) 598,616,478(163,288) CNVR^b^ 521 682 278^c^ 216^d^ 370^d^ 1,263 139,786,166(204,965) Outgroup^e^ CNV 18 1,003(55.72) 284(15.78) 48(2.67) 955(53.06) 2,603(144.61) 442,235,607(440,912) CNVR 18 483 187 21 458 1,593 276,846,573(573,181) The numbers in parentheses are normalized by sample counts except that the lengths in parentheses are average lengths normalized by CNV counts. ^a^At sample level, each sample has 7.09 (3666/517) CNVRs averagely and 6.23 (2256/362) for Taurine, since there are 4 taurine individuals without identified CNVs; ^b^These numbers are nonredundent CNVR counts. ^c^278 CNVRs are unique to one sample while 404 CNVRs are shared by at least 2 individuals or breeds and 18 of 404 multiple events have frequency \>5%; ^d^Besides 370 loss and 216 gain CNVRs, there are 96 CNVRs containing both loss and gain events; ^e^Outgroup animals are not included in the total counts for CNV and CNVR. ::: ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Genomic landscape of cattle copy number variations and segmental duplications**. CNV regions (682 events, 139 Mb, \~4.60% of the bovine genome) reported by 521 SNP genotyped individuals are shown above the chromosomes in green (gain), red (loss) and dark blue (both), while below are the CNV regions (177 events, 28 Mb, \~1% of the bovine genome) reported by 90 array CGH experiments by Liu et al. The bar height represents their frequencies: short (appeared in 1 sample), median (≥2 samples) and tall (≥5 samples). Segmental duplications (94.4 Mb, 3.1% of the bovine genome) predicted by two independent computational approaches are illustrated on the chromosomes in red (WSSD), blue (WGAC) or purple (both). The patterns are depicted for all duplications for ≥5 kb in length and ≥90% sequence identity. The gaps in the assembly are represented on the chromosomes as white ticks. :::  ::: To test the stability of CNV calls with respect to different genome builds and SNP mapping, we also migrated 56,408 out of 56,947 SNP markers from Btau\_4.0 to UMD3 using the UCSC liftOver tool \[[@B35]\], and repeated the entire calling analyses to ensure consistency in calls (Additional file [1](#S1){ref-type="supplementary-material"}: Tables S3 and S4). Only 61 more CNVRs were identified on the UMD3 assembly (making a total of 743 CNVRs). A simple comparison indicated that the total coverage of variable regions were 13.05% larger on UMD3 (158.0 Mb, Additional file [1](#S1){ref-type="supplementary-material"}: Table S3) than on Btau\_4.0. For all three CNVR types (gain, loss and both), counts increased slightly. This was expected as both assemblies were based on the same raw WGS reads. The most obvious difference between the two assemblies is that Btau\_4.0 unplaced contigs are placed on UMD3. This resulted in more markers that were on Btau\_4.0 ChrUn contigs to be placed on UMD3 autosomes, which could partially explain the increase in the CNVR counts. Since the majority of cattle genome annotations were performed on the Btau\_4.0 assembly, we focused on further characterization of the 682 high-confidence CNV regions from Btau\_4.0 autosomes. These 682 CNVRs include 370 loss, 216 gain and 96 both (loss and gain within the same region) events, ranging from 32,566 to 5,569,091 bp with a mean or median of 204,965 or 131,179 bp, respectively (Additional file [1](#S1){ref-type="supplementary-material"}: Table S2). Loss events are approximately 1.7-fold more common than gain events, but have slightly smaller sizes than gain regions on average. Furthermore, 278 CNVRs were found in only one sample (Unique), 404 CNVRs were present in two or more animals or breeds and 18 of 404 multiple events had a frequency \>5% (Table [1](#T1){ref-type="table"} and Additional file [1](#S1){ref-type="supplementary-material"}: Table S2). These datasets confirm that segregating CNVs exist among these 21 cattle breeds and/or groups, which is consistent with our earlier results based on array CGH \[[@B29]\]. In general, the number of CNVs identified in each sample is consistent with SNP estimates of breed-specific founding and effective population sizes and levels of polymorphism based on ≥50,000 SNPs \[[@B5]\]. As shown in Table [1](#T1){ref-type="table"}, more CNV events were detected in indicine (11.41 per sample) than in African groups (7.21 per sample) and composite (7.17 per sample). The taurine breeds (6.23 per sample) had the fewest detected CNVs. While some of these differences could be related to the fact that the SNP markers were designed based on the Btau\_4.0 reference genome (which was derived from the sequence of a Hereford cow of European origin; Dominette 01449), this observation is consistent with the concept of subspecies divergence and supports the hypothesis of multiple independent domestications of cattle in the Fertile Crescent, Southwest Asia and likely Africa \[[@B36],[@B37]\]. Cattle CNVs are distributed in a nonrandom fashion at two different levels. First, CNV content varies significantly among different chromosomes. The proportion of any given known chromosome susceptible to CNV regions varies from 1.32-8.80% (Additional file [1](#S1){ref-type="supplementary-material"}: Table S2). Chromosomes 1 and 6 show the greatest enrichment for CNV (Figure [1](#F1){ref-type="fig"} and Additional file [1](#S1){ref-type="supplementary-material"}: Table S2) with almost two-fold of the variable content of the autosomal average. It is interesting to note that these chromosomes do not have the highest SD content \[[@B29],[@B32]\]. Furthermore, similar to the human, mouse, rat and dog genomes, there are a greater proportion of CNVs near pericentromeric and subtelomeric regions. Excluding chrX and chrUn, pericentromeric and subtelomeric regions each represent 3.42% of genomic sequence but show an enrichment of 1.5-2.4-fold more CNVRs (both P values \<0.001) and contain 7.78-12.54% of all polymorphic sequence. Quality assessment of selected CNV Regions ------------------------------------------ The quality of our 682 CNV calls was assessed in multiple ways, though our first assessment was a comparison against existing cattle CNV datasets (Table [2](#T2){ref-type="table"} and Figure [2](#F2){ref-type="fig"}). One of the earlier datasets included 79 filtered deletion variants (representing 42 unique genomic loci and 9 single SNPs) reported earlier using the Illumina genotyping software module cnvPartition v1.0.2 \[[@B5]\]. Nineteen of our CNVRs overlapped with 11 of the deletion variants (21.57%) in that dataset (Table [2](#T2){ref-type="table"}). We also identified 129 CNVRs (18.91%) in our dataset that overlapped with 128 CNVRs from a SNP-based CNV study on 265 Korean Hanwoo cattle \[[@B31]\] (Figure [2B](#F2){ref-type="fig"}). The Hanwoo CNV study identified 368 CNVRs in total, so our dataset overlapped with 34.83% of their calls \[[@B31]\]. We then compared our calls against an array CGH-based study of 20 cattle from four breeds \[[@B30]\]. Since our dataset excluded CNV calls in the chrX, chrUn and mitochondrial sequences, we compared our autosomal CNVR calls (682 CNVRs) to the autosomal CNV calls of that study (254 CNVRs) \[[@B30]\]. Only 51 of our CNVRs (7.48%) directly overlapped with 55 of their calls (21.65%, Figure [2C](#F2){ref-type="fig"}). Our final comparison was against our previous array CGH-based study of 90 animals from 14 breeds which resulted in 163 autosomal CNVR calls \[[@B29]\]. In this comparison, 57 of our SNP-based CNVR calls (8.36%) overlapped with 59 CNVRs derived from array CGH (36.20%, Figure [2C](#F2){ref-type="fig"}). If we only focused on the 16 HapMap samples which were assessed by both platforms (60 CNVRs derived from array CGH and 106 CNVRs reported by SNP array), there were 21 overlapping CNVRs: 19 for array CGH (31.67%, 19/60), and 20 for SNP array (18.87%, 20/106). When we merged existing CNV datasets, a total of \~ 200 out of 682 (about 30%) newly identified CNVRs overlapped with them (Figure [2A](#F2){ref-type="fig"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Summary of genome-wide studies of cattle copy number variations ::: ---------------------------------------------------------------------------------------------------------------------------------------------- Study Assay Count CNVR Size ------------------------- ------------- ------------- ------ ------ --------------------- -------- ----------------- -------- -------- ------- Matukumalli et al. 2009 BovineSNP50 54,001 556 21 Deletion only 51^a^ 22.92-11,050.69 394.87 960.67 49.0 Liu et al. 2010 Array CGH \~385,000 90 17 Deletion,\ 163^b^ 18.00-1,261.90 86.19 153.75 25.1 insertion Bae et al. 2010 BovineSNP50 54,001 265 1 Deletion,\ 368 25.35-967.18 128.33 171.49 63.1 insertion Fadista et al. 2010 Array CGH \~6,300,000 20 4 Deletion,\ 254^c^ 1.72-2,031.34 15.51 62.26 15.8 insertion This study BovineSNP50 56,947 521 21 Deletion, insertion 682 32.57-5,569.09 131.18 204.97 139.8 ---------------------------------------------------------------------------------------------------------------------------------------------- ^a^This includes 9 independent SNPs and 42 CNVRs. The statistics are calculated for 42 CNVR excluding the 9 SNPs; ^b^This is the number excluding chrX and chrUn; ^c^This is the number excluding chrX, chrUn and mitochondrial sequence. ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Comparisons between identified 682 CNVRs in this study and the other existing cattle CNVR datasets in terms of count and length**. A, compared to the total nonredundant CNVR merged from existing published datasets; B, compared to CNVR derived from SNP array (Bae et al, 2010); C, compared to two CNVR datasets derived from array CGH studies (Liu et al,2010; Fadista et al, 2010); D, the summaries and legends of existing cattle CNVR datasets. :::  ::: It is expected that the variants identified in these studies do not overlap, suggesting a vast amount of CNVs exist in cattle population and saturation for this type of variation has not yet been approached. It is likely that many thousands of more common structural variants may still remain undiscovered in the cattle genome. A similar situation was encountered in human CNV studies using the early version of SNP, CGH arrays and detection methods \[[@B38],[@B39]\]. For example, although cnvPartition detects CNVs by processing the similar raw data as PennCNV (i.e. LRR and BAF), it is based on a different proprietary sliding window algorithm. Only those homozygous deletion events segregating in different animals were reported due to concerns with the quality of calls \[[@B5]\]. In the future, high-density SNP arrays combined with improved CNV calling algorithms could remedy these differences. Besides the technology and detection method differences, the following could also contribute to the observed differences: (1) sampling differences: 521 individuals from 21 diverse breeds and/or groups were included in our study; (2) genome coverage biases: 56,947 markers were included in our study rather than a subset of \"well-behaved\" SNPs (54,001 markers) which exclude those SNPs in CNV-rich regions; (3) correction of genomic waves in order to minimize false positive calls; and (4) trio/pedigree information was fully explored in our study to improve the accuracy and call rate of CNVs. When filtering criteria varying the CNVR length and frequency were applied, we observed significant overlap within our 2 datasets derived from SNP arrays and array CGH (Additional file [1](#S1){ref-type="supplementary-material"}: Table S8). For example, when the large CNVRs (SNP count = 10, a median \~574kb) were considered, the overlap reached 21.74%. When the CNVR frequency was increased to 1, 2, or 5% (animal count = 5, 10, or 25, respectively), the overlap increased to 89.47%. When we filtered the CNVR frequency to greater than 10% of our population (50 animals), the overlap included 100% of our filtered dataset. This further demonstrated that large, common CNVRs can be reliably detected through using different detection technologies even when the majority of samples were different. For example, our current SNP array study identified most of the large, common CNVRs which were confirmed in our published results \[[@B29]\]. After comparison with other existing datasets, we found that \~70% of our CNVR calls were not reported in the literature. In order to confirm these novel CNVRs, we performed 24 quantitative PCR (qPCR) assays for 15 low frequency, novel CNVR calls spread among seven individuals (Additional file [1](#S1){ref-type="supplementary-material"}: Table S9). Nine of the CNV regions had two target amplicons placed near two different SNP loci. Out of 24 total locations, 11 loci (\~48%) were in agreement with CNV estimates by PennCNV (Additional file [1](#S1){ref-type="supplementary-material"}: Table S9 and Figure S1). When counting the CNV regions, 9 out of 15 (60%) CNV regions had positive qPCR confirmations in at least one location. If CNVRs previously validated in the literature \[[@B29]\] were also included, approximately two third (30% + 70% × 48%) of our detected CNVRs had positive confirmations. As expected, the Bovine SNP50 platform has a large resolution limit under the current PennCNV calling criteria. The size of the CNVRs detected ranged from 32.6 kb to 5.6 Mb, with a median size of 131.2 kb (Additional file [1](#S1){ref-type="supplementary-material"}: Table S2). This is partially due to the fact that the Bovine SNP50 assay was originally developed for high-throughput SNP genotyping in association studies. Although CNV detection is feasible with SNP arrays, it is impaired by low density and non-uniform distribution of SNPs, especially in CNV and SD regions. Compared to a CGH array, a SNP array lacks non-polymorphic probes designed specifically for CNV identification. Thus, only the large CNVRs are expected to be identified with the Bovine SNP50 assay. This explains the difference in CNV length between our study and the earlier results. The discrepancies between the qPCR and PennCNV results may represent small CNV events that were missed in the PennCNV calls, or instances where SNPs caused the qPCR reaction to fail or be suboptimal but did not affect the SNP assay. Despite the fact that a two-copy state was assumed for test PCR loci in Dominette, smaller CNV events in Dominette may have evaded detection by PennCNV. If our test primers amplified a small CNV event in Dominette, that would skew the relative copy number estimates of our qPCR reactions. Although qPCR primers were designed within 250 bp around the target SNP positions, additional SNPs and small indels may have influenced the hybridization of the qPCR primers in some animals, thereby reducing primer efficiency. Other causes may also contribute to the discrepancy in CNVR validation by qPCR. The draft status of the cattle genome assembly and the low SNP density of the Bovine SNP assay make it difficult to determine the real breakpoints of CNVRs. For example, multiple, neighboring, discrete CNV events could result in a larger call by PennCNV; therefore, giving an over estimation of the CNV size. Therefore, it cannot be ruled out that the qPCR primers used to confirm the CNVRs may have been designed outside the breakpoints. CNVs overlap with segmental duplications and other genomic features ------------------------------------------------------------------- Following previous studies of other genomes, we detected the association between CNVRs and SDs. Agreeing with previous predictions regarding cattle SDs \[[@B32]\], a local tandem distribution pattern is predominant in our cattle CNVR dataset (Figure [1](#F1){ref-type="fig"}). It should be noted that about 25.66% (175/682) of CNV regions directly overlap with cattle SDs with an overlapping span of 16,283,071 bp (11.65% of the total 139,786,166 bp). Approximately 12.06% (356/2952) of the SDs (excluding chrX and chrUn) identified by WGAC and WSSD \[[@B32]\] exhibit CNVs. In comparison, 58.90% of the CNVRs (96/163) detected by using array CGH \[[@B29]\] excluding X chromosome overlap with cattle SDs, corresponding to 15,176,612 bp (60.56% of the total 25,061,646 bp). The proportion of our new CNVR calls (identified in this study) that overlap with SDs reaches approximately 40% compared to 61% in our previous study. This lower overlap fraction probably reflects the fact that the BovineSNP50 array used in this study is biased against cattle SD regions. SNP density on the array drops by one-third (from 21 probes/Mb in unique regions down to 14 probes/Mb) in SD regions. We also failed to detect any correlation between 682 CNV regions and evolutionary breakpoint regions (EBRs). Compared to the genomic averages, cattle-specific EBRs and artiodactyl-specific EBRs do not show enrichments of CNV sequences \[[@B40]\]. This negative result is consent with the fact that EBRs have fewer overlaps with SD regions. Gene Content of Cattle CNV regions ---------------------------------- Within autosomes, the 682 CNV regions overlap with 1,679 Ensembl peptides, corresponding to 1,263 unique Ensembl genes (Table [1](#T1){ref-type="table"} and Additional file [2](#S2){ref-type="supplementary-material"}: S6). About 55.57% (379/682) of high-confidence CNVRs completely or partially span cattle Ensembl genes. We assigned PANTHER accessions to a total of 1,263 overlapping genes. Statistically significant over or under representations were observed for multiple categories (Additional file [1](#S1){ref-type="supplementary-material"}: Table S10). This set of copy number variable genes possess a wide spectrum of molecular functions, and provides a rich resource for testing hypotheses on the genetic basis of phenotypic variation within and among breeds. Consistent with similar CNV analyses in other mammals (human, mouse and dog), several of these CNVs, which are important in drug detoxification, defense/innate and adaptive immunity and receptor and signal recognition, are also present in cattle. These gene families include olfactory receptors, ATP-binding cassette (ABC) transporters, Cytochrome P450, β-defensins, interleukins, the bovine MHC (BoLA) and multiple solute carrier family proteins which support the shared GO terms among mammals as shown in Additional file [1](#S1){ref-type="supplementary-material"}: Table S10. For gene families that went through cattle-specific gene duplication \[[@B32]\], such as interferon tau, pregnancy-associated glycoproteins, *SCP2*and *ULBP*and *WC1.1*subfamilies, we also detected marked variation in copy number between individuals and across diverse cattle breeds and/or groups (Additional file [2](#S2){ref-type="supplementary-material"}: Table S6). It is intriguing to note that we also detected variations of *TLR3*(toll-like receptor 3) and *PPARA*(peroxisome proliferator-activated receptor alpha). This current CNV survey further supports a hypothesis that the generation of new CNV insertions and deletions may be a constant phenomenon in multiple cattle breeds/individuals \[[@B41]\]. We also overlapped our CNVRs with two sets of genomic regions under positive or balance selection detected by iHS and F~ST~using SNP data \[[@B3],[@B10]\] (Additional file [3](#S3){ref-type="supplementary-material"}: Table S11). By doing so, we have identified CNV regions that may span potential cattle QTLs and human orthologous OMIM genes influencing disease susceptibility (Additional file [3](#S3){ref-type="supplementary-material"}: Table S11). For instance, multiple CNV regions directly overlap with QTLs for significant and typical economic traits and 87 out of 682 CNVRs correspond to loci known to cause disease in humans. However, since the cattle genome and cattle QTLs are less well defined, future study is warranted. Cattle CNV frequency differences among breeds --------------------------------------------- To highlight the potential evolutionary contributions of these CNVs to cattle breed formation and adaptation, we queried 91 CNVRs that have breed-specific CNV frequency differences (Additional file [1](#S1){ref-type="supplementary-material"}: Table S12). We only considered breeds that had at least 12 samples and any detectable variations must be identified in at least 3 individuals or 10% of samples (for Holstein, Angus and Limousin where n \> 30). Fifty-eight of these CNVRs correspond to annotated genes or gene families; many of which were identified in other mammals as influencing adaptation to the environment. Some of the annotated genes are known to be important in cattle adaptation including CNVR266(*IFNT)*in Brown Swiss, CNVR122 (*SCP2)*in Hereford \[[@B32]\] and CNVR178 (Olfactory Receptors) in most of breeds \[See also \[[@B42]\]\]. The differences of CNV frequency among cattle breeds supported our earlier hypothesis that some cattle CNVs are likely to arise independently in breeds, are likely to contribute to breed differences and are therefore related to the breed formation and adaptation. However, the observed differences between breed variations could be caused by both selection and genetic drift due to genetic bottlenecks for some breeds. Our findings, therefore, must be confirmed with an even larger sample size. CNVs in outgroup animals ------------------------ For the 18 individuals in outgroups, which were analyzed similarly together with 521 modern breeds individuals, 1003 CNVs and 483 CNVRs were detected, covering 276.8 Mb base pairs, with 21 gain and significantly more (458) loss events (Table [1](#T1){ref-type="table"} and Additional file [1](#S1){ref-type="supplementary-material"}: S7). About 34.60% of our current CNVRs (236/682) directly overlapped with 37.47% of these ancient CNVRs (181/483) derived from ancient outgroups, which indicates over one third of CNVRs were likely ancestral. We suspected this observation of more loss events than gain events was at least partially related to the high genetic divergences between these outgroup animals and the cattle reference genome. With additional cattle, sheep, goat and pig CNV papers published recently \[[@B43]-[@B46]\], it will be interesting to look into the evolutionary mechanism of CNVs within livestock animals. Conclusions =========== We have performed a comprehensive genomic analysis of cattle CNVs based on whole genome SNP genotyping data, therefore providing a valuable genomic variation resource. A total of 682 CNVRs were identified, covering 139.8 megabases (\~4.60%) of the genome. A subset of these CNVRs showed Mendelian inheritance and were also confirmed in other cattle CNV studies and other mammalian species. As high density cattle SNP genotyping data are becoming available, CNVs combined with SNPs, may help identify genes undergoing artificial selection in domesticated animals. Methods ======= Selection of cattle breeds and animals -------------------------------------- It has been demonstrated that the BovineSNP50 genotyping array provides a robust resource for genome-wide, high-density SNP genotyping of cattle and for population genetic analyses of closely related artiodactyl species \[[@B4],[@B47]\]. In which, less than 3% of markers had call rates below 99.94%, the average call rate for individual samples was greater than 97.5% and 85% of samples had call rates above 98.8% \[[@B5]\]. Cattle CNVs in this study were detected by using the same SNP genotyping results, including those collected for the Bovine HapMap project \[[@B3]\] (Additional file [1](#S1){ref-type="supplementary-material"}: Table S1). PennCNV quality filters were applied after the CNV detection, resulting in 521 distinct high quality genotyping results from the original 556 animals. This panel included 366 animals from 14 taurine dairy and beef breeds, 70 animals from three breeds of predominantly indicine background, 46 animals from two breeds that are Taurine × Indicine composites, and 39 animals from two African groups, one of which (Sheko) is an ancient hybrid. It is worth to note that for many of the breeds, individuals were sampled from more than one continent to represent the global cattle population. This panel contained 39 trios where both parents and an offspring were genotyped. Additionally, we included 18 animals from 6 outgroups (*Bos gaurus*- Gaur, *Bos bison*- North American Bison,*Bubalus depressicornis*- Lowland Anoa,*Bos javanicus*- Banteng,*Bos grunniens*- Yak, and *Syncerus caffer*-Cape Buffalo) with 1 trio information to derive the ancestral states of CNVs. Among these outgroups, the average sample call rate was 89.91%, reflecting their divergent relationship from *Bos Taurus*. Identification of cattle CNVs ----------------------------- PennCNV algorithm \[[@B20]\] was only applied to autosomes (-lastchr 29) to detect cattle CNV in this study. In our initial analysis, chrX (-chrx) was also considered separately from automosomes. PennCNV incorporates multiple sources of information together, including LRR and BAF at each SNP marker, more realistic models for state transition between different copy number states based on the distance between neighboring SNPs, population frequency of B allele (PFB), the allele frequency of SNPs, and the pedigree information where available, into a hidden Markov model (HMM). Both LRR and BAF were exported from Illumina GenomeStudio Genotyping Module v1.0 software given the default clustering file for each SNP. The PFB file was calculated based on the BAF of each marker in this population. Because there were 153 out of 556 animals (\~27.5%) with absolute values of waviness factor larger than 0.04 in our original analysis, the genomic waves were adjusted using the -gcmodel option. The cattle gcmodel file was generated by calculating the GC content of the 1Mb genomic region surrounding each marker (500kb each side). For comparison, the analysis without considering gcmodel was also conducted. Three different PennCNV options were performed wherever possible: 1) -test: the individual-calling algorithm that treats family members as if they were unrelated; 2) -trio: the posterior-calling algorithm which accommodates family information to improve the accuracy of individual-based CNV calling and boundary prediction; 3) -joint: the joint-calling algorithm that identifies CNVs using family data simultaneously. After CNV detection, filtering of low-quality samples was carried out with the default cutoffs: standard deviation (STD) of LRR as 0.30, BAF drift as 0.01 and waviness factor as 0.05. The filtered results from the three algorithms were compared in terms of CNV numbers, lengths and number of SNP in CNVs (Additional file [1](#S1){ref-type="supplementary-material"}: Table S5). The final CNVs set was the nonredundant combination of CNVs from the -joint results for family trio members and the -test results for unrelated individuals. For the outgroup animals, quality filtering was not performed due to their divergent relationship from *Bos Taurus*. CNVRs are determined by aggregating overlapping CNVs identified across all samples \[[@B13]\]. CNV validation -------------- array CGH experiments were performed as previously described \[[@B11]\]. Primers were designed for qPCR validation using the Primer3 webtool <http://frodo.wi.mit.edu/primer3/> by limiting amplicon length to 150bp to 250bp and by incorporating a GC clamp of 2. All other settings were left at the default. Primer information is shown in Additional file [1](#S1){ref-type="supplementary-material"}: Table S9. Quantitative PCR experiments were conducted using SYBR green chemistry in triplicate reactions, each with a reaction volume of 25 μl. All reactions were amplified on a BioRad MyIQ thermocycler. An intron-exon junction of the BTF3 gene was chosen as a reference location for all qPCR experiments. Analysis of resultant crossing thresholds (Ct) was performed using the ΔΔCt method \[[@B48],[@B49]\]. Calibration ΔCt values were derived from amplification of reference and test primers on a genomic DNA template derived from the European Hereford, Dominette 01449. Since all reference and test primers did not overlap with any of Dominette\'s CNV regions, two-copy states were assumed for both amplicons in Dominette. Cattle CNV distribution and association with Segmental Duplications and other features -------------------------------------------------------------------------------------- We investigated the genomic distribution of 682 CNVRs by testing the hypothesis that pericentromeric and subtelomeric regions were enriched for CNVs as described previously \[[@B32]\]. Briefly, all predicted variable bases that overlapped with these regions were totaled and chi-square tests were used to test the null hypothesis of no enrichment. CNVRs were also overlapped with SD and the other genomic features such as evolutionary breakpoint regions (ERBs), which were obtained from literature and public databases listed in web site references. Gene content ------------ Gene content of cattle CNV regions was assessed using Ensembl genes <ftp://ftp.ensembl.org/pub/current_fasta/bos_taurus/pep/>, the Glean consensus gene set, cattle RefSeq and *in silico*mapped human RefSeq (the UCSC Genome Browser website at <http://genome.ucsc.edu/>). We obtained a catalog of all bovine peptides from Ensembl. This yielded 26,271 peptides, 1,679 of which overlap with predicted 682 high-confidence CNV regions, and correspond to 1,263 unique Ensembl genes. Using the PANTHER classification system, we tested the hypothesis that the PANTHER molecular function, biological process and pathway terms were under- or overrepresented in CNV regions after Bonferroni corrections \[[@B32]\]. It is worth noting that a portion of the genes in the bovine genome has not been annotated or has been annotated with unknown function, which may influence the outcome of this analysis. Web Site References =================== The Database of Genomic Variants: <http://projects.tcag.ca/variation/> Ensembl genes <ftp://ftp.ensembl.org/pub/current_fasta/bos_taurus/pep/> PANTHER <http://www.pantherdb.org/> OMIM <http://www.ncbi.nlm.nih.gov/omim/> OMIA <http://omia.angis.org.au/> QTL <http://www.animalgenome.org/> Authors\' contributions ======================= YH, KW, EK, LKM, JS carried out computational analysis. DMB, MFC, MV carried out the experimental validations. PMV, TSS and CPVT participated in data collection. GEL conceived of the study and led in its design and coordination. All authors contributed to writing the manuscript, read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Supplemental Material file**. Table S1. Numbers of species, breeds, animals and trios used to call CNVs genotyped by BovineSNP50 assay. Table S2. Btau\_4.0 cattle CNV regions and their frequencies. Table S3. Comparison of CNV regions identified on two cattle genome assemblies. Table S4. UMD3 cattle CNV regions and their frequencies. Table S5. The comparison of CNVs from 39 trios using three CNV calling algorithms: individual-calling, posterior-calling and joint-calling. Table S7. Outgroup CNV regions and their frequencies. Table S8. The effects of CNV length and frequency on calling consistances between CNV callings based on SNP array and aCGH. Table S9. qPCR Summary. Table S10. Over/Underrepresentation of PANTHER molecular function, biological process and pathway terms. Table S12. CNVR frequency differences among breeds. Figure S1. Illustration of a typical CNV call with qPCR validation. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Supplemental Material file**. Table S6. Gene contents of cattle CNV regions. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Supplemental Material file**. Table S11. Cattle CNV regions overlap with genomic regions under positive selection, human orthologous OMIM genes and cattle QTLs. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We thank members of the Bovine HapMap Consortium for sharing their samples. We thank R. Anderson, J. Shaffer, D. Hebert, S. Schroeder and L. Shade for technical assistance. This work was supported by in part by NRI grant No. 2007-35205-17869 from USDA CSREES (now NIFA), Project 1265-31000-098-00 from USDA-ARS and Research Grant Award No. IS-4201-09 from BARD, United States - Israel Binational Agricultural Research and Development Fund.

PubMed Central

2024-06-05T04:04:19.597351

2011-2-23

PMC3053261

Background ========== Hepatocyte nuclear factor 4α (HNF4α) is a highly conserved member of the nuclear receptor superfamily. It is highly expressed in liver, kidney, intestine, and pancreas in mammals. The active form of HNF4α is a homodimer which recognizes a direct repeat (DR) of the AGGTCA motif separated by 1 nucleotide (DR1) as a binding site. HNF4α exerts direct transcriptional effects on target genes, and it also works indirectly *via*the positive regulation and negative regulation of other liver-enriched transcription factors, each of which regulates numerous downstream targets \[[@B1],[@B2]\]. In contrast to the liver-enriched transcription factors HNF1α, HNF3α, HNF6, and CAAT/enhancer-binding protein (C/EBP)α, which when disrupted in the mouse genome the mice are viable but show specific effects on hepatocyte differentiation, metabolic function, and gene expression \[[@B3]-[@B6]\], disruption of the mouse *HNF4a*gene is embryonic lethal \[[@B7]\]. Studies with HNF4α deficient mice \[[@B8]\] have established the critical role of this factor in regulating diverse liver functions, including glucose, fatty acid and cholesterol homeostasis, bile acid and urea biosynthesis \[[@B9]-[@B11]\]. Defects in HNF4α function have been linked to the human disease maturity onset diabetes of the young 1(MODY1) that results from haploinsufficiency of *HNF4a*gene \[[@B12]\]. The pivotal role of HNF4α in the maintenance of the differentiated hepatic phenotype is highlighted by the exceptionally high number of potential target genes revealed by genome-scale target search studies. Binding sites for HNF4α in genes expressed in the liver occur more frequently than those of other liver-enriched transcription factors \[[@B13]\], supporting the idea that HNF4α is a global regulator of liver gene expression. Diverse signaling pathways regulate the transcription of hepatocyte-specific genes. For example, trauma or infection results in the release of proinflammatory cytokines, *e.g.*, interleukin (IL)-6, IL-1, and tumor necrosis factor-α (TNF-α). The release of these cytokines has long been known to stimulate the acute phase response (APR) and rapidly alters rates of synthesis of a group of plasma proteins known as acute phase proteins (APPs) \[[@B14]\]. APPs are an established diagnostic tool as early indicators of inflammation and disease. Many APPs play beneficial roles in mediating the complex inflammatory response and seeking to restore homeostasis, but prolonged exposure to acute phase conditions has been correlated with inflammatory syndromes such as sepsis and multiple organ failure \[[@B15],[@B16]\]. An understanding of the molecular events that are involved in mediating the response to external stresses can lead to the development of therapeutic strategies for preventing the progression of the APR to the chronic inflammatory states, while preserving its protective effects. APP gene expression is regulated at the level of transcription. Transcriptional activation is mediated by a number of transcription factors. Beyond the well-known nuclear factor-κB, (NF-κB) and signal transducer and activator of transcription (STAT) family members, HNF4α has been shown to be involved in the regulation of liver-specific genes, including acute phase genes \[[@B17],[@B18]\]. It has been reported that in several injury models, injury leads to significant changes in binding activities of several liver-enriched transcription factors, including HNF4α \[[@B18]-[@B20]\]. However, the functional analysis of HNF4α regulated-APR genes so far has mainly relied on the description of the expression level of a few selected genes \[[@B21]\]. To achieve a global view of HNF4α during the APR, we used microarray analysis to evaluate the expression profile in HepG2 cell, a human hepatoma cell line. This cell line is similar to hepatocytes in terms of biologic responsiveness \[[@B22]-[@B24]\] and is widely used as a model system for studying the regulation of acute phase protein synthesis in human liver \[[@B25]-[@B27]\]. In this study, HepG2 cells were treated with either HNF4α short hairpin RNA (shRNA) or cytokines (IL-6, IL-1β, TNF-α) alone, or in combination of the two treatments. Our results demonstrate that HNF4α is an important regulator in liver gene expression. The highly significant overlap of genes sensitive to HNF4α knock-down and cytokine treatment suggests that HNF4α may be involved in the regulation of the liver\'s inflammatory response. Our data also show that HNF4α may mediate a certain amount of genes indirectly *via*the ETS-domain transcription factor ELK1, a mitogen-activated protein kinase (MAPK)-responsive transcription factor \[[@B28],[@B29]\], which is a previous undefined mechanism for HNF4α regulation. Results and discussion ====================== Knock-down endogenous HNF4α in HepG2 cells ------------------------------------------ To study the role of HNF4α in liver-specific gene expression and the inflammatory response, the endogenous HNF4α in HepG2 cells was knocked down by the technique of RNA interference. As shown in Figure [1](#F1){ref-type="fig"}, HNF4α shRNA caused a reduction in mRNA and protein levels of HNF4α by more than 70% relative to control levels assayed by real-time PCR and Western blot. Our results indicate that HNF4α shRNA can efficiently and specifically knock-down HNF4α in HepG2 cells \[[@B21]\]. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Knock-down of HNF4α in HepG2 cells**. HepG2 cells were transfected with non-specific shRNA control or HNF4α shRNA plasmid. mRNA and whole cell lysates were prepared for real-time PCR (A) and Western blots (B), respectively. The results shown in (A) represent the relative mRNA expression level normalized to GAPDH mRNA level. The abundance of mRNA in the controls was set at 1. Data represent mean ± SD of 4 replicates. An HNF4α antibody (sc-6556, Santa Cruz Biotechnology) and β-actin (Sigma) antibody, used as an internal loading control, were utilized for Western blot (B). :::  ::: HNF4α acts as a global regulator of liver-specific gene expression ------------------------------------------------------------------ Given that HNF4α is a central mediator in hepatocyte-specific gene expression and liver function, it is important to identify the full spectrum of genes impacted by the loss of HNF4α. To reach this end, microarray analysis was performed to probe differences in gene expression between the control and HNF4α knock-down HepG2 cells. Because we are particularly interested in how HNF4α alters global gene expression patterns during the inflammatory response, we also examined gene expression in response to the inflammatory response induced by cytokine treatment. A two-way ANOVA was performed to identify differentially expressed genes and classify the observed expression patterns. This procedure identifies genes with altered expression in response to HNF4α shRNA or cytokines alone, or in combination HNF4α knock-down with cytokine treatments. The entire dataset is available at the NCBI Gene Expression Omnibus <http://www.ncbi.nlm.nih.gov/geo/> with the accession number GSE15991. From this analysis, a total of 14,220 unique probesets were found to be present in the samples. Two-way ANOVA analysis was then used to determine which probesets were differentially expressed between the untreated control and treated groups. The ANOVA analysis identified four categories of interest (Figure [2](#F2){ref-type="fig"}): Category A, \"HNF4α shRNA only\" contains genes that are significantly regulated by HNF4α shRNA, but not by cytokines; Category B, \"Cytokine only\" contains genes that respond to cytokine treatment, but not to HNF4α knock-down; Category C, \"Additive\" contains genes whose expression is dependent on both treatments with cytokines and HNF4α shRNA, and the effect of both treatments is additive, suggesting that the two treatments likely influence expression independent of each other. Finally, Category D, \"Interactive\" contains genes that exhibit an expression pattern dependent on both treatments in which the effect of the two treatments is not additive, rather interactive between each other. The genes with significantly altered expression and fold changes in each category are listed in Additional file [1](#S1){ref-type="supplementary-material"}. Utilizing K-means clustering, the genes in the category A and category B (Figure [2A](#F2){ref-type="fig"} and [2B](#F2){ref-type="fig"}) were clustered into 2 clusters for each category; Category C and category D (Figure [2C](#F2){ref-type="fig"} and [2D](#F2){ref-type="fig"}) were clustered into 4 and 8 clusters, respectively. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **The clustering expression profiles of up- and down-regulated genes**. A global transcriptional view of HepG2 cells in response to the treatment with cytokines and HNF4α knock-down alone (2 and 3) or in combination (4) is shown. Each group has 4 replicates. Relative expression values are expressed as a color code (bar color chart on the bottom, red up- and blue down-regulation). **(A)**, Category A contains genes that are significantly regulated by HNF4α knock-down, but not by cytokines; **(B)**, Category B contains genes that respond to cytokine treatment, but not to HNF4α knock-down; **(C)**, Category C contains genes whose expression is dependent on both treatments with cytokines and HNF4α knock-down in an independent manner; **(D)**, Category D has genes that exhibit an expression pattern dependent on both treatments in an interactive manner. :::  ::: As shown in Table [1](#T1){ref-type="table"}, we identified a total of 5,173 probesets (36% of the 14,220 probesets on the Affymetrix HG U133A 2.0 GeneChip with sequence-specific signal in our experiment) that exhibited differential expression in response to HNF4α knock-down (false discovery rate, FDR\< 0.01). Of these, 3,606 probesets were only affected by HNF4α knock-down, and were not differentially expressed in response to cytokine stimulation (Category A). The remaining 1,567 probesets exhibited a differential expression pattern that was either specific to the combined effect of HNF4α knock-down and cytokine treatments (502 probesets; Category D) or the genes that were affected by HNF4α knock-down and cytokine treatments independently (1,065 probesets; Category C). Among the probesets affected by HNF4α knock-down independently (Category A and Category C), many (3,088 of 4671 probesets) were up-regulated when HNF4α expression was reduced. These observations may suggest that HNF4α directly or indirectly regulates a large number of liver-specific gene expressions. We found that in our experimental conditions more genes appeared to be down-regulated or repressed under a normal level of HNF4α in HepG2 cells. This finding is not in agreement with the description that HNF4α functions primarily as a transcription activator \[[@B8],[@B30]\], but is more in line with the observation that HNF4α can act as a suppressor of transcription \[[@B31],[@B32]\]. To exclude an off-target effect of knock-down in our study, siRNA rescue experiments were performed. A rescue effect was observed in the knock-down HNF4α HepG2 cells transfected with an HNF4α siRNA-resistant construct generated by introducing silent substitutions in the HNF4α siRNA-target region (Additional files [2](#S2){ref-type="supplementary-material"} and [3](#S3){ref-type="supplementary-material"}). These siRNA rescue experiments suggest that the HNF4α shRNA used in this study can specifically knock-down HNF4α in HepG2 cells. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Genes exhibiting altered expression by the treatment with HNF4α shRNA alone and in combination with cytokines. ::: Cut-off ^1^ HNF4 shRNA only^2^ HNF4 shRNA+Cytokines^3^ Total ------------- -------------------- ------------------------- ------- ------ ---------- FDR \< 0.01 3606 1065 502 1567 **5173** 2-fold 878 223 296 519 **1397** Two methods of cut-off ^1^, FDR (false discovery rate) and 2-fold changes compared to untreated controls, are shown. HNF4 shRNA only^2^(Category A) denotes that gene expressions are significantly changed by HNF4α shRNA, but not by cytokines. HNF4 shRNA+Cytokines^3^indicate that gene expressions are significantly changed by both HNF4α shRNA and cytokines, including the additive (Category C) and interactive (Category D) effects. ::: The observation that a large number of liver specific genes are affected by HNF4α is consistent with the recent work of Odom *et al.*\[[@B13]\], in which the authors demonstrated that HNF4α binds the promoters of 12% of genes in human liver cells, whereas HNF1α binds the promoters of only 1.6% of genes using chromatin immunoprecipitation-chip (ChIP-chip). In order to determine what fraction of genes that showed differential expression in response to HNF4α knock-down is likely to be bound by HNF4α, we compared our microarray expression data to the ChIP-chip data reported by Odom *et al.*Only those genes in the Odom\'s data for which there were probesets on our microarray were considered. We found that 54% (659 of 1,219) of the genes bound by HNF4α detected by ChIP-chip also showed differential expression in response to HNF4α knock-down on our microarray. A Fisher-Exact test was used to test whether the fraction of genes in the Odom *et al.*data that are differentially expressed in response to HNF4α knock-down is significantly greater than the fraction of genes probed on our microarray that exhibit HNF4α-dependent expression, resulting in a highly significant *p*-value of *p*= 2.9 × 10^-33^. A similar analysis determined that 13% (659/5,173, *p*= 2.8 × 10^-17^) of those genes that exhibit differential expression in response to HNF4α knock-down were reported by Odom *et al.*to be bound by HNF4α. These results suggest that a substantial fraction of HNF4α-sensitive genes may be indirect targets of HNF4α. However, the differences seen might also be caused by the significant differences in the experimental methods used such as the data from ChIP-chip were derived from human hepatocytes, while our experiment was performed on HepG2 cells. Majority of genes affected by cytokines are also affected by HNF4α knock-down ----------------------------------------------------------------------------- As shown in Table [2](#T2){ref-type="table"}, we found that expression of 15% of the probesets (2,202 of 14,220) was affected by treatment with cytokine alone (635 probesets; Category B), and treatment with both HNF4α shRNA and cytokines (1567 probesets; Category C and D). Of the probesets that are differentially expressed in response to cytokine treatment but non-interactively with HNF4α knock-down (Category B and C) almost twice as many are down-regulated (1,144 probesets) as up-regulated (556 probesets). This may reflect that during the APR, there is an important role for the down-regulation of specific genes in response to an inflammatory stimulation, although the up-regulation of APR genes has been more extensively studied than those that are down-regulated. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Genes exhibiting altered expression by the treatment with cytokines alone and in combination with HNF4α shRNA. ::: Cut-off Cytokine only^1^ HNF4 shRNA+Cytokines^2^ Total ------------- ------------------ ------------------------- ------- ------ ---------- FDR \< 0.01 635 1065 502 1567 **2202** 2-fold 78 223 296 519 **597** Cytokine only^1^(Category B) indicates that gene expressions are significantly changed by cytokines, but not by HNF4α shRNA. HNF4 shRNA+Cytokines^2^indicate that gene expressions are significantly changed by both HNF4α shRNA and cytokines, including the additive (Category C) and interactive (Category D) effects. ::: More interestingly, we found that majority of those genes that exhibit differential expressions in response to cytokine treatment also reveal differential expression in response to HNF4α knock-down. Of the 2,202 probesets found to be responsive to cytokine treatment, 1,567 (71%) of them also show altered expression in response to knock-down of HNF4α, while only 635 probesets are not influenced by HNF4α knock-down (Table [2](#T2){ref-type="table"}). The *p*-value for this level of over-representation, calculated using a Fisher Exact test, is 1.8 × 10^-207^. Some of these probesets are altered independently by cytokine treatment and reduction in HNF4α levels (1,065 probesets; Category C), while others show a pattern of expression where the effect of cytokine treatment and reduction in HNF4α levels is interdependent (502 probesets; Category D). When more stringent criteria were used for filtering our data, and only those probesets in response to cytokine treatment and/or HNF4α knock-down that change expression by more than two fold were chosen, we found that an even greater percentage (519/597, 87%) of probesets regulated by cytokines are also regulated by HNF4α (Table [2](#T2){ref-type="table"}). This over-representation suggests that HNF4α may play a significant role in orchestrating the inflammatory response in hepatic gene expression. Although there is significant overlap between those probesets regulated by cytokines and those regulated by HNF4α knock-down (Category C and D), the expression patterns in response to these two treatments are varied. For some probesets the response to HNF4α knock-down are in the same direction as cytokine treatment (99 probesets, clusters 4 and 7 in the interactive category D, Figure [2D](#F2){ref-type="fig"}). This group of genes may represent a direct linkage of HNF4α with the injury response and will serve as interesting targets for further study of the complex role of HNF4α in the response to injury. For other genes the effect of each treatment is in opposition to each other (144 probesets, clusters 3 and 5 in Category D, Figure [2D](#F2){ref-type="fig"}). These diverse regulatory patterns observed suggest that the effects of HNF4α knock-down as that of cytokine treatment are pleiotropic in nature affecting transcription events at basic levels allowing individual gene responses to be highly variable but none the less altered. Previous work by our lab has demonstrated that HNF4α binding activities are significantly reduced in a burn injury mouse model and a cytokine-induced APR cell culture model \[[@B18],[@B21]\]. We have shown, utilizing the cell culture model, that the decrease in HNF4α binding activity also affects HNF4α\'s ability to transactivate target genes \[[@B21]\]. The injury induced decrease in HNF4α binding may affect cellular transcription by simply decreasing the amount of effective HNF4α available for binding. Our ability to efficiently decrease HNF4α concentration utilizing RNA interference technique may mimic this decrease in HNF4α binding ability isolating this aspect of HNF4α\'s role in the injury response. Genes annotated as participating in inflammatory response exhibit distinct expression patterns ---------------------------------------------------------------------------------------------- In order to further explore the function of HNF4α in the inflammatory response, a set of 170 genes annotated as playing a key role in inflammatory response was obtained from Gene ontology (GO) <http://geneontology.org>. These inflammatory response genes are highly enriched in the set of probesets up-regulated by cytokines (*p*= 3.5 × 10^-3^, 334% above background), but not in those probesets down-regulated after cytokine treatment. Slightly more inflammatory response genes are up-regulated in response to HNF4α knock-down than those that are down-regulated, but the difference is not statistically significant (*p*\> 0.05). To determine whether certain expression patterns are associated with functional annotations, we further tested each cluster in each category using ANOVA analysis. While HNF4α-regulated probesets as a whole are not significantly enriched for the annotated inflammatory response genes, two expression clusters in the interactive category D are highly enriched for these genes. Of the 5,173 probesets regulated by HNF4α knock-down, 25 of them are annotated as participating in inflammatory response, and 14 of these genes fall into one of two clusters. One of these clusters (cluster 8, Category D) contains 101 probesets that are dramatically up-regulated by the combination of HNF4α shRNA and cytokine treatment, but exhibit relatively low levels of expression in the untreated controls, and the cells treated by either HNF4α knock-down or cytokines alone. Nine of 101 probesets in cluster 8 are annotated in GO as participating in the inflammatory response, which is 8.2 times more than expected by chance (*p*= 2.2 × 10^-6^) (Figure [3A](#F3){ref-type="fig"}). Another cluster (cluster 1, Category D) contains 41 probesets expressed in moderate levels under all conditions except HNF4α shRNA treatment in the absence of cytokines, in which they are markedly down-regulated. In this cluster, five genes are annotated as inflammatory response genes which is over 11.3 times more than expected by chance (*p*= 8.8 × 10^-5^) (Figure [3B](#F3){ref-type="fig"}). While it is not immediately clear why genes that play a role in inflammatory response are enriched in these particular expression clusters, it is intriguing that such genes exhibit similar HNF4α-dependent expression. These inflammatory response genes in cluster 8 and cluster 1 present good candidates for further study. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Inflammatory response genes are enriched in two clusters within the interactive category (Category D)**. Expression profiles of up- and down-regulated genes in the different groups treated by cytokines and HNF4α knock-down alone (2 and 3) or in combination (4) are shown. Each group has 4 replicates. Relative expression values are expressed as a color code (bar color chart on the bottom, red up- and blue down-regulation). Inflammatory response genes (listed at the right side of the graph) extracted from GO are highly enriched in two different clusters (A and B). Gene abbreviations: VNN1, vanin 1; SAA1/SAA2, serum amyloid A1/serum amyloid A2; C5; complement component 5; CX3CL1, chemokine (C-X3-C motif) ligand 1; LBP, lipopolysaccharide binding protein; PTX3, pentraxin-related gene, rapidly induced by IL-1 beta; PTAFR, platelet-activating factor receptor; NMI, N-myc (and STAT) interactor; ORM1/ORM2, orosomucoid 1/orosomucoid 2; C4A/C4B, complement component 4A/complement component 4B; CEBPB, CCAAT/enhancer binding protein (C/EBP), beta; SERPINA3, serpin peptidase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin), member 3; CCL20, chemokine (C-C motif) ligand 20; HIF1A, hypoxia-inducible factor 1, alpha subunit (basic helix-loop-helix transcription factor). :::  ::: To confirm the microarray results, we chose several transcripts from category D, cluster 8 (C5 and LBP) and cluster 1 (SERPINA 3 and C4A), and measured their expressions with real-time PCR using the same RNA samples used for microarray studies. All of them showed a high concordance between microarray and real-time PCR data. Figure [4](#F4){ref-type="fig"} illustrates the comparison of the expression levels between microarray (Figure [4A](#F4){ref-type="fig"}) and real-time PCR (Figure [4B](#F4){ref-type="fig"}) for the selected transcripts from cluster 8 and cluster 1 (Category D). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Confirmation of microarray results using quantitative real-time PCR**. (**A**), Raw intensity values were measured using microarray. (**B**), Real-time PCR results for the same genes are expressed as the relative mRNA expression level normalized by GAPDH mRNA level. The abundance of mRNA in the controls was set at 1. Data represent mean ± SD of 4 replicates. Ctr, Control; Cyto, Cytokines; shRNA, HNF4α shRNA; shRNA+Cyto, HNF4α shRNA plus Cytokines. \**p*\< 0.05 and \*\**p*\< 0.01 indicate a significant difference compared to control. Gene abbreviations are the same as described in Figure 3 legend. :::  ::: In addition to testing for enrichment of inflammatory response genes, enrichment for broad categories of functionality was calculated for each cluster. GO terms of all 26 top level biological processes were analyzed. We found that GO term of metabolism is significantly enriched in genes exhibiting HNF4α-dependent expression (Figure [5](#F5){ref-type="fig"}). As previous work \[[@B9],[@B30],[@B33]\] has noted that HNF4α plays a significant role in regulating metabolism in the liver, it is not surprising that the GO term for metabolism is HNF4α-dependent. Thirty percent of genes (1,530 of 5,173) that are differentially expressed in response to HNF4α knock-down are annotated as participating in a metabolic process, resulting in a Fisher Exact *p*-value of 2.6 × 10^-18^. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Gene ontology analysis**. Twenty six broad categories of biological processes were analyzed for genes that exhibit HNF4α-dependent expression. Significance levels are plotted as-log (*p*value). One-tail *p*-values were calculated using the Fisher Exact test. Threshold (line) denotes the *p*= 0.002 level, which is the threshold for significance after Bonferroni correction. :::  ::: HNF4α may regulate a large number of targets *via*the ETS family of transcription factors ----------------------------------------------------------------------------------------- To gain insight into the complex transcriptional networks that regulate hepatic gene expression, the promoter region (-1 kb to +0.5 kb relative to the transcription start site, TSS) of genes was extracted, and potential transcription factor binding sites were identified using Clover \[[@B34]\]. By extensive search of the genes found to be differentially expressed as a function of HNF4α knock-down, we found that the HNF4α motif is not statistically enriched in the promoters of genes whose expression changes in response to HNF4α knock-down relative to promoters chosen randomly from throughout the genome (*p*\> 0.05). There are several possible reasons for this observation. Firstly, HNF4α binding sites may be located far up-stream or down-stream of the TSS, although the results still remain the same when we extended our search up to 5 kb up-stream and 2.5 kb down-stream of the TSS. Secondly, there might be a fraction of sites that bind HNF4α but differ from the established consensus. Thirdly, HNF4α may interact with proximal promoters through formation of enhancer/promoter loops with HNF4α binding to distal regulatory element \[[@B35]\]. Yet another possibility is that HNF4α might act as a cofactor, interacting with other transcription factors and not directly bind to DNA. Indeed, we found that the binding sites for the ETS family of transcription factors including ELK1, ELK4 and GA binding protein transcription factor A (GABPA) are highly enriched in the HNF4α knock-down group. The binding motifs for ELK1 (*p*= 4.0 × 10^-7^), ELK4 (*p*= 1.14 × 10^-10^) and GABPA (*p*= 3.4 × 10^-5^) are very similar (Figure [6](#F6){ref-type="fig"}), and all significantly enriched in the promoter regions of genes affected by HNF4α knock-down. This mirrors a finding by Smith *et al.*\[[@B36]\], where they showed that ELK1 binding motifs are enriched in the promoters of genes also bound by HNF4α. Moreover, in this study we found that ELK1 and ELK4 genes exhibit significant differential expression in response to HNF4α knock-down. ELK1 is up-regulated when HNF4α is knocked down, while ELK4 is down-regulated. These results suggest that HNF4α may regulate a substantial number of genes *via*ELK transcription factors. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Sequence logos of ETS transcription factor binding sites**. Sequence logos of consensus DNA binding sites for the three ETS transcription factors enriched in genes regulated by HNF4α. Y axis indicates amount of information at each position in the motif. These logos were generated from information obtained from the JASPAR core database \[[@B42]\]. **(A)**, ELK1 binding motif; **(B)**, ELK4 binding motif; and **(C)**, GABPA binding motif. :::  ::: The transcription factor ELK1 is of particular interest as this protein is a nuclear target of extracellular-regulated kinases (ERKs) and plays a pivotal role in immediate early gene induction by external stimuli \[[@B28],[@B29]\]. It has been known that HNF4α expression is modulated by MAP kinase signaling \[[@B37],[@B38]\]. Li *et al.*\[[@B39]\] reported that ELK1 is involved in the inflammatory response *via*stimulation of chemokine by thrombin. These observations suggest that the relationship between ELK1 and HNF4α may be especially relevant to understanding the role of HNF4α in regulating the inflammatory response. By searching for ELK1 binding site in the HNF4α regulated genes (5173 genes), we identified 373 genes that have one or more potential ELK1 binding sites in their promoter regions (-1 kb to +0.5 kb) (Additional file [4](#S4){ref-type="supplementary-material"}). To define a functional link between HNF4α and ELK1, we selected 4 genes (COL4A1, ZNF175, MMP15 and SEC24A). These genes all were found having an ELK1 binding motif (CCGGAAG/A, Figure [6](#F6){ref-type="fig"}), but no HNF4α binding motif in their proximal promoter regions, and from our microarray analysis their expressions were shown to be either up-regulated (COL4A1, ZNF175) or down-regulated (MMP15 and SEC24A) by HNF4α knock-down. The expression of these potential ELK1 target genes was examined by real-time PCR in HepG2 cells treated with siRNA to knock-down HNF4α and/or ELK1 alone or both together. As shown in Figure [7](#F7){ref-type="fig"}, the expression of ELK1 can be efficiently knocked-down by siRNA. The knock-down of HNF4α resulted in a significantly greater ELK1 expression compared to control (*p*\< 0.05), and the ELK1 expression was not significantly affected by cytokine treatment, which are consistent with our microarray data. We propose that HNF4α may indirectly regulate a number of genes through ELK1 transcription factor based on the observation that the ELK1 binding site is highly enriched in genes affected by HNF4α knock-down. It was predicted that if a higher ELK1 expression induced by HNF4α knock-down affects the transcriptional outcomes of ELK1-target genes, we should see an opposite regulatory effect on these gene transcriptions when the ELK1 level is reduced. The results (Figure [8](#F8){ref-type="fig"}) show that when the cells were treated with siELK1 either alone or with siHNF4α, the two genes up-regulated by HNF4α knock-down (Figures [8A](#F8){ref-type="fig"} and [8B](#F8){ref-type="fig"}) are expressed at significantly lower levels compared to HNF4α knocked-down (*p*\< 0.01). In contrast, the genes down-regulated by HNF4α knock-down (Figures [8C](#F8){ref-type="fig"} and [8D](#F8){ref-type="fig"}) show significantly increased expression compared to siHNF4α (*p*\< 0.01). This result is consistent with a direct effect of ELK1 level induced by the decrease of HNF4α concentration on a subset of ELK1-target genes. However, these findings could be explained by another mechanism by which an HNF4α/ELK1 complex is formed, leading to alterations in the regulation of a group of genes. To test this hypothesis, chromatin immunoprecipitation (ChIP) assays were performed. An HNF4α antibody was used to immunoprecipitate chromatin from HepG2 cells. Specific PCR primers were utilized to amplify a DNA fragment with an ELK1 binding motif, and without an HNF4α binding motif. As shown in Figure [9](#F9){ref-type="fig"}, HNF4α was able to interact with these genes containing ELK1 binding sites, and the interactive ability of HNF4α was significantly reduced after the knock-down of ELK1. Interestingly, when the HNF4α binding motif was further searched using a web-based search tool, HNF4 Motif Finder \[[@B40]\], a potential HNF4 binding site was identified in the promoter region of MMP15, which could lead to a direct binding of HNF4α when HNF4α antibody was used to immunoprecipitate chromatin. However, if HNF4α binding to MMP15 is independent of ELK1, when ELK1 is knocked down, one would expect to see no change in the binding of MMP15 in the ChIP assay, this is not the case as shown in Figure [9](#F9){ref-type="fig"}. Furthermore, we found that ELK1 can directly bind to its potential ELK1 binding site in MMP15 utilizing electrophoretic mobility shift analysis (data not shown). Given these findings, we hypothesize that HNF4α may indirectly mediate gene expression, in part, through a co-operative interaction with ELK1, and possibly also with other ETS transcription factors. Others have reported that a number of ETS family proteins interact and crosstalk with several transcription factors including AP-1, NF-κB and Stat-5 to co-regulate the expression of cell-type specific genes. Such interactions coordinate cellular processes in response to diverse signals including cytokines, growth factors, antigen and cellular stresses \[[@B41]\], here we believe that we have uncovered a novel interplay between the transcription factors HNF4α and ELK1 for controlling gene expression in the liver. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **The ELK1 expression in HepG2 cells with HNF4α or ELK1 knock-down**. HepG2 cells were transfected with non-specific siRNA control (siControl), HNF4α shRNA plasmid (siHNF4) or siELK1, and then treated with or without cytokines. The expression of ELK1 was measured by real-time PCR. The results represent the relative mRNA expression level normalized to GAPDH mRNA level. The abundance of mRNA in the controls was set at 1. Data represent mean ± SD of three different experiments. \**p*\< 0.05 and \*\**p*\< 0.01 indicate a significant difference compared to siControls. :::  ::: ::: {#F8 .fig} Figure 8 ::: {.caption} ###### **A decrease in ELK1 expression leads to elimination of the regulatory effect of HNF4α knock-down on a subset of genes**. HepG2 cells were treated with non-specific siRNA control (siControl), HNF4α shRNA plasmid (siHNF4), siELK1 or both siHNF4 and siELK1. The expressions of COL4A1 (collagen, type IV, alpha 1), ZNF175 (zinc finger protein 175), MMP15 (matrix metallopeptidase 15) and SEC24A (SEC24 family, member A) genes were determined by real-time PCR. The abundance of mRNA in the siControl was set at 1. Data are presented as mean ± SD of three different experiments. \**p*\< 0.05 and \*\**p*\< 0.01 indicate a significant difference compared to siControl. \#*p*\< 0.01 indicates a significant difference compared to siHNF4α. :::  ::: ::: {#F9 .fig} Figure 9 ::: {.caption} ###### **The reduction of ELK1 level decreases the ability of HNF4α to interact with the promoters of a subset of genes**. **(A)**, HepG2 cells were transfected with siControl (siCtr) or siELK1. Protein interaction of HNF4α and ELK1 was determined by ChIP assay with either antibody against HNF4α or goat normal IgG (IgG). Chromatin-immunoprecipitated DNA was analyzed by PCR with primers specific for the ELK1 binding sites in the promoter regions of COL4A1, ZNF175, MMP15 and SEC24A genes. The result shown in **(A)**is a representative experiment, replicated three times with similar results. **(B)**, Histograms show densitometric analyses of relative binding abilities. Values represent mean ± SD of three separate experiments. The relative quantitative analysis was carried out by comparison of siELK1 with siControl, and the siControl was set at 1. \**p*\< 0.05 and \*\**p*\< 0.01 indicate a significant difference compared to siCtr. IP, immunoprecipitation. :::  ::: Conclusions =========== HNF4α is a major regulator of hepatic gene expression. The complex physiological effects of HNF4α on the regulation and maintenance of hepatic phenotype are likely involved both directly and indirectly in a systemic response such as that seen in a response to injury or inflammation. Our microarray data are consistent with a broad effect of HNF4α on liver functions, and more importantly, the microarray analysis provides a genomic view for the role of HNF4α in the inflammation response, which greatly extends the observations from both animal- and cell culture-injury models. However, more experimentation and a focus on individual pathways will need to be done before a full picture of HNF-4α\'s role in the modulation of hepatic phenotype by injury can be obtained. Methods ======= Cell culture and treatments --------------------------- HepG2 cells (ATCC \# HB-8065), human hepatoma cells, were grown in Dulbecco Modified Eagle Medium (DMEM) supplemented with penicillin (100 units/ml), streptomycin (100 μg/ml), and 10% heat-inactivated fetal bovine serum (Mediatech, Herndon, VA) at 37°C in a humidified atmosphere with 5% CO~2~. The inflammatory response in HepG2 cells was stimulated with a cytokine mixture consisting of 1 ng/ml of recombinant human IL-1β, 10 ng/ml of IL-6, and 10 ng/ml of TNF-α (PeproTech, Rocky Hill, NJ) in serum-free medium for 18 h. Knock-down of HNF4α in HepG2 cells was carried out as described previously \[[@B21]\]. Briefly, shRNA plasmids (HNF4α shRNA and non-specific control shRNA) were transfected into HepG2 cells with the Nucleofector (Amaxa Biosystem, Cologne, Germany) T-28 protocol following the manufacturer\'s instructions. Knock-down of ELK1 in HepG2 cells was performed by transfection with small interfering RNA (siRNA) (Ambion, Austin, TX) using lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA) according to the manufacturer\'s instructions. Twenty-four hours after transfection, the cells were placed in serum-free medium for 5 h and then either treated or untreated with cytokines for 18 h. The different treated HepG2 cells were divided into four groups: Group 1, control: the cells were transfected with non-specific control shRNA; Group 2, HNF4α shRNA: the cells were transfected with HNF4α shRNA; Group 3, cytokines: the control cells were transfected with non-specific control shRNA and then treated with cytokines; Group 4, HNF4α shRNA with cytokine treatment: the cells were transfected with HNF4α shRNA prior to the treatment with cytokines. Four biological replicates from separate experiments were performed for each of the four study groups. Total RNA isolation and real-time PCR ------------------------------------- Total RNA was extracted using the RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer\'s instructions. Quantitative real-time PCR analysis was conducted on the ABI 7000 Sequence Detection System and StepOnePlus™ Real-Time PCR System. Relative mRNA expression was quantified using the comparative Ct (ΔΔCt) method according to the ABI manual (Applied Biosystems, Foster City, CA). Amplification of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used in each reaction as an internal reference gene. Assays were performed in triplicate. TaqMan probes were used for the human HNF4α (Hs00230853\_m1), complement component 5 (C5, Hs00156197\_m1), lipopolysaccharide binding protein (LBP, Hs00188074\_m1), serpin peptidase inhibitor member 3 (SERPINA3, Hs00153674\_m1), complement component 4A (C4A, Hs00167147\_m1), ELK1 (Hs00901847\_m1), collagen, type IV, alpha 1 (COL4A1, Hs01007469\_m1), zinc finger protein 175 (ZNF175, Hs00232535\_m1), matrix metallopeptidase 15 (MMP15, Hs00233997\_m1), SEC24 family, member A (SEC24A, Hs00405228\_m1) and GAPDH (Hs99999905\_m1) from the TaqMan^®^Gene Expression Assays (Applied Biosystems). Microarray analysis ------------------- Affymetrix HG-U133A 2.0 GeneChips were used to study the expression levels for HepG2 cell RNA. Sample labeling, hybridization to microarrays, scanning and calculation of normalized expression levels were carried out at the Microarray Resource, Boston University School of Medicine, Boston, MA. RNA samples (four biological replicates for each of the four study groups: control, HNF4α shRNA, cytokines, and HNF4α shRNA with cytokine treatment) were analyzed. Microarray data were quantified and normalized using Affymetrix MicroArray Suite (MAS) 5.0. Two-way ANOVA analysis was then used to determine which probesets were differentially expressed between the untreated control and treated groups. Only probesets exhibiting differential expression with FDR \< 0.01 were included in this study. After the ANOVA analysis was performed, K-means clustering was then used to cluster probesets that exhibit similar expression patterns within each category of differential expression. In order to eliminate multiple probesets, the raw data were clustered based on Unigene identifiers (ids), which represent unique transcription loci. Unigene ids associated with multiple probesets that exhibit different patterns of expression were removed from the dataset, ensuring unambiguous data that could be easily used in subsequent enrichment analyses. This approach ensures that, if anything, the results of subsequent analyses are conservative. Gene ontology (GO) enrichment analysis -------------------------------------- Before GO enrichment could be calculated, each GO annotation was mapped to a corresponding Unigene id. SwissProt ids associated with each Unigene id were obtained and were used for this purpose. Where SwissProt ids were not available, gene symbols were used to associate GO annotations with Unigene ids. Each cluster in all 4 categories was then tested for enrichment for genes associated with inflammatory response GO database (relative to the entire set of genes on the microarray) using Fisher\'s Exact test. In addition, to identify other functionality, each cluster was tested for enrichment of each of the top-level \"biological process\" annotations in the GO database. As there are 26 top-level \"biological process\", a Bonferroni correction was applied for this top-level analysis, resulting in a *p*-value cut-off of 0.002. Promoter analysis ----------------- Weight matrices obtained from the JASPAR core database \[[@B42]\] were used to identify potential transcription factor binding sites in the up-stream of promoter regions. One promoter with each Unigene id was included in the analysis, so that genes with multiple annotated promoters/TSS did not bias the analysis. The region, extending from -1 kb to +0.5 kb relative to the TSS, was extracted. Clover \[[@B34]\] was used with a minimum log-ratio score cut-off of 8.0 to identify potential binding sites in this region. Promoters were classified based on whether they contain at least one binding site, and enrichment in different expression clusters relative to a background set of promoters derived from all probesets present on the microarray. The randomization feature of Clover was not used, instead the exact number of promoters containing binding sites in both the HNF4α-regulated set and the background set were computed, and a *p*-value was calculated using Fisher\'s Exact test. Because there are almost 140 transcription factors in the JASPAR database, a Bonferroni correction was applied, resulting in a *p*-value cut-off of 3.5 × 10^-4^. Chromatin immunoprecipitation (ChIP) assay ------------------------------------------ ChIP assay was performed as described previously \[[@B21]\]. The purified chromatin was immunoprecipitated using 10 μg of anti-HNF-4α (sc-6556, Santa Cruz Biotechnology, Santa Cruz, CA) or normal goat IgG (Santa Cruz). After DNA purification, the presence of the ELK1 putative binding motif (CCGGAAG/A) DNA sequence was assessed by PCR. The primers used were as follows: (1) COL4A1 gene: 5\'-GAGTTTAGCGCAGGATGAGG-3\' and 5\'-GCTCTCCTGCTTGGGAGTAG-3\', and the PCR product is 230 bp in length; (2) ZNF175: 5\'-TAAAAGCCCTTTGACGATGG-3\' and 5\'-CTCTAGGCCACTTCCGGTTT-3\', and the PCR product is 239 bp in length; (3) MMP15: 5\'-ATCCAGCTCGTTAAGCTTCG-3\' and 5\'-TTAATCTCTCCGAGCCTCCA-3\' for amplifying a 207 bp DNA fragment. (4) SEC24A: 5\'-GCACCAGGAGCTGTCAGG-3\' and 5\'-GGCAGCCAAACCTAGAGAGA-3\', and the PCR product is 186 bp in length. The PCR conditions were as following: 95°C for 10 min, followed by 94°C for 45 s, 58°C for 60 s, and 72°C for 60 s for a total of 31 cycles. In the ELK1 knock-down experiments, the relative quantitative analysis in binding activity was performed, utilizing densitometry and statistical analysis, by comparison of siELK1 with non-specific siRNA control, and the control was set at 1. Authors\' contributions ======================= ZW and PAB designed the experiment. ZW performed the experiments. EPB analyzed the data. ZW and EPB wrote the manuscript. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Genes that respond to the treatment of HNF4α knock-down and/or cytokines**. This file includes four sub-tables showing the genes with significantly altered expression and fold changes in four categories (Category A to D). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Construction of HNF4α siRNA-resistant mutant**. This file illustrates the generation and DNA sequencing of the HNF4α siRNA-resistant construct. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **siRNA rescue assay**. This file shows the characterization of the HNF4α siRNA-resistant construct by Western blot, and the rescue effect of this construct on HNF4α knock-down responsive genes. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **ELK1 and HNF4α binding motifs in HNF4α-regulated genes**. This file shows the sequence and position of ELK1 and HNF4α potential binding sites in the group of HNF4α-regulated genes. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This work was supported by NIH grant (R01DK064945). The authors wish to thank the Microarray Resource, Boston University School of Medicine, Boston, MA for performing the microarray experiments and for help with data analysis. We also thank Dr. Marc Lenburg and Dr. Avrum Spira for critical reading of the manuscript.

PubMed Central

2024-06-05T04:04:19.601458

2011-2-25

PMC3053262

Background ========== Inflammation of the udder tissue, known as mastitis, has become one of the most costly and prevalent diseases affecting dairy cows worldwide \[[@B1]\]. Mastitis onset has been associated with many different bacterial infections, but the most common are staphylococci, streptococci and coliform bacteria \[[@B2]\]. *Escherichia coli*(*E. coli*) is a prevalent environmental pathogen that routinely colonizes dairy cattle, and represents a significant risk of mastitis \[[@B2]\]. *E. coli*-induced mastitis is characterized as a relatively short-term disease process \[[@B3]\] and induces a distinct acute phase response (APR). The precise response kinetics, however, depend on various biological and environmental factors, such as host diet, *E. coli*strain and dose, and time of infection \[[@B4]\]. The innate immune response is triggered at the earliest stages of infection, as its function is to recognize pathogens that have not been encountered before \[[@B5]\]. Therefore, the innate immune system is effectively the first line of defence against intramammary *E. coli*infection. Many of the cytokines and other inflammatory mediators involved in the innate immune response are characterized by a specific pathogen-dependent expression profile \[[@B6]\]. Gram-negative bacteria, such as *E. coli*, are characterized by an early and high expression of cytokines and other inflammatory mediators in the milk, while Gram-positive bacteria, such as *Staphylococcus aureus*(*S. aureus*), show a lower and/or delayed expression of cytokines and other inflammatory mediators in the milk \[[@B6]\]. Moreover, the inflammatory responsiveness to *E. coli*endotoxin is highly influenced by the cows\' lactation stage \[[@B7],[@B8]\]. Recently, the Toll-Like Receptor (TLR)-4 transduction pathway was suggested as a crucial contributor to robust mammary gland immune defence against *E. coli*mastitis \[[@B9],[@B10]\]. This pathway involves induction of cellular inflammatory and apoptotic responses, and eventually leads to the activation of NF-κB factors in resident macrophages, monocytes and epithelial cells \[[@B9],[@B10]\]. Diapedesis of polymorphonuclear neutrophil leukocytes (PMN) is impaired in early lactation, and a reduced neutrophil influx into the mammary gland during the APR is believed to promote the incidence of severe *E. coli*mastitis during this period \[[@B11]\]. Hence, both macrophages and neutrophils are considered as important players in the local APR during *E. coli*mastitis. To gain a detailed understanding of the defence mechanisms employed by the mammary gland, the genes and gene pathways that are altered in response to the presence of pathogenic bacteria should be identified and characterized. The development of microarray technology has enabled comprehensive screening of defined gene expression profiles in specific tissues, such as the mammary gland. Recently, this type of analysis was carried out on udder tissue collected *post mortem*in dairy cows experimentally infected with medium to high doses of *S. aureus*\[[@B12]\], *S. uberis*\[[@B13]\] and *E. coli*\[[@B14],[@B15]\] intramammary. The principal focus of our study was on the gene-expressions in udder tissue biopsies collected *ante mortem*from dairy cows during the APR (24 h p.i) and the chronic stage (192 h p.i) of the *E. coli*infection when using a low inoculation dose of *E. coli*in early lactation. The aim being to identify the global mammary gland gene expressions and gene pathways associated with bovine *E. coli*mastitis during the acute and chronic stage of the infection in early lactating dairy cows. Furthermore, the aim was to identify common genes involved in the local mammary gland APR between our study and two other gene-expression studies on *E. coli*mastitis \[[@B14],[@B15]\] independent of *E. coli*strain and dose, cow\'s lactation stage and number, tissue collection method and gene analysis used in the studies. Results ======= Clinical examinations and para-clinical measurements ---------------------------------------------------- All challenged cows (n = 16) developed *E. coli*mastitis and were free of infections with other major udder pathogens during sampling. *E. coli*was identified in the infected quarters of all 16 cows at 24 h after inoculation. At T = 120 h p.i., *E. coli*was present in variable concentrations in eight of the cows, and by 192 h all cows had cleared the infection. Body temperature, colony forming units (CFU/ml) of *E. coli*in the milk, milk somatic cell count (SCC), and the concentrations of Milk Amyloid A (MAA) and of Serum Amyloid A (SAA) in blood in response to *E. coli*infection at time points T = 0 h, T = 24 h, and T = 192 h are presented in Figure [1](#F1){ref-type="fig"}. In the acute phase (T = 24 h), the body temperature, *E. coli*CFU in the milk, SCC, concentrations of MAA in milk and SAA in blood were significantly increased, as compared to the data from T = 0 h and T = 192 h. These findings were consistent with symptomology of *E. coli*-induced mastitis, and confirmed the model system. At T = 192 h, even though all cows had successfully cleared the *E. coli*, SCC and MAA concentrations remained higher than at T = 0 but to a lesser extent than observed at T = 24 h. These findings confirmed the establishment of the chronic stage. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Overview of the clinical examination and para-clinical measurement of the cows infected with *E. coli*at T = 0 h (black), T = 24 h (grey) and T = 192 h (white)**. The bars correspond to the mean values of the trait, while the whiskers on the bars correspond to the standard deviation. A: body temperature of the cows (°C), B: number of *E. coli*bacteria in the milk (CFU/ml), C: somatic cell count in the milk (cells/ml), D: concentration of Milk Amyloid A in milk (μg/ml), and E: concentration of Serum Amyloid A in blood (μg/ml). :::  ::: Acute phase response -------------------- In total, 982 DE transcripts were detected (*P*\< 0.05) at this stage, of which 684 corresponded to 609 annotated genes \[Additional file [1](#S1){ref-type="supplementary-material"}: Supplemental Table S1\]. Hierarchical clustering of the DE transcripts revealed two main clusters (up-regulated and down-regulated). Up-regulated genes ------------------ Cluster 1 (up-regulated) contained 699 of the DE transcripts, representing 551 distinct genes. Hypergeometric gene set analysis based on the GO information revealed overrepresentation of 92 gene sets \[Additional file [2](#S2){ref-type="supplementary-material"}: Supplemental Table S2\]. Of these gene sets, 43 were directly related to immune response. GO identifiers were related to inflammatory response (GO:0006954), chemotaxis (GO:0006935), immune response (GO:0006955), leukocyte migration (GO:0050900), response to lipopolysaccharide (GO:0032496), and TLR signalling pathway (GO:0002224). In contrast, the hypergeometric gene set analysis based on the KEGG information revealed 18 gene sets (Table [1](#T1){ref-type="table"}). These were related to chemokine signalling pathway (KEGG04062), TLR signalling pathway (KEGG04620), leukocyte transendothelial migration (KEGG04670), cytokine-cytokine receptor interaction (KEGG04060), natural killer (NK) cell-mediated cytotoxicity (KEGG04650), and RIG-I-like signalling pathway (KEGG04622). Among the DE transcripts in this cluster the genes encoding chemokine (C-C motif) ligands (CCL3, CCL4, CCL5, CCL16, CCL19), chemokine (C-X-C motif) ligands (CXCL2, CXCL5, CXCL16), interleukin 1β (IL1B), interleukin 8 (IL-8), interleukin receptors (IL1R1, IL2RA, IL2RG, IL4R, IL6R, IL8R), interferon regulatory factor 9 (IRF9), neutrophil cytosolic factors (NCF1, NCF2), and Toll-like receptors (TLR2, TLR4) were represented. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Significant KEGG identifiers detected based on the differentially expressed transcripts in cluster 1 for the acute phase response to *E. coli*infection\* ::: KEGG ID *P*value Odds Ratio Exp Count Count Size Term --------- ---------- ------------ ----------- ------- ------ ------------------------------------------- 4062 2.98E-09 5.571349 5.027606 22 106 Chemokine signaling pathway 4620 4.87E-06 6.066348 2.513803 12 53 Toll-like receptor signaling pathway 4670 2.71E-05 4.568835 3.414978 13 72 Leukocyte transendothelial migration 4060 9.62E-05 3.722762 4.363583 14 92 Cytokine-cytokine receptor interaction 4650 0.000254 4.473538 2.656094 10 56 Natural killer cell mediated cytotoxicity 4514 0.000518 3.714061 3.414978 11 72 Cell adhesion molecules (CAMs) 4622 0.000743 5.504747 1.565198 7 33 RIG-I-like receptor signaling pathway 4666 0.000873 4.103185 2.561233 9 54 Fc gamma R-mediated phagocytosis 4662 0.001521 4.767827 1.754919 7 37 B cell receptor signaling pathway 4612 0.002095 4.468453 1.84978 7 39 Antigen processing and presentation 4920 0.002438 4.332374 1.89721 7 40 Adipocytokine signaling pathway 4210 0.002843 3.718903 2.466373 8 52 Apoptosis 4610 0.00426 3.861615 2.086931 7 44 Complement and coagulation cascades 3050 0.00727 3.941793 1.754919 6 37 Proteasome 4630 0.00791 2.831847 3.509838 9 74 Jak-STAT signaling pathway 5340 0.008959 5.797582 0.853744 4 18 Primary immunodeficiency 564 0.046285 3.241129 1.375477 4 29 Glycerophospholipid metabolism 5200 0.04726 1.759624 7.778561 13 164 Pathways in cancer \*A hypergeometric gene set enrichment test was performed based on cluster 1 of the contrast T24 *vs*. C24. Overrepresentation of gene sets defined by the KEGG database was tested using the Fisher\'s exact test. A gene set was considered significant if *P*\< 0.05. ::: Down-regulated genes -------------------- Cluster 2 contained 229 DE transcripts, representing 125 known genes that were down-regulated in response to *E. coli*infection. The hypergeometric gene set analysis based on the GO information revealed over representation of 44 gene sets mainly related to fat metabolism \[Additional file [3](#S3){ref-type="supplementary-material"}: Supplemental Table S3\], including fatty acid biosynthetic process (GO:0006633), lipid metabolic process (GO:0006629), and fatty acid metabolic process (GO:00031). The hypergeometric gene set analysis based on the KEGG information revealed over representation of 7 gene sets, of which three were related to fatty acid synthesis: fatty acid metabolism (KEGG00071), fatty acid biosynthesis (KEGG00061), and glycerolipid metabolism (KEGG00561) (Table [2](#T2){ref-type="table"}). Among the DE transcripts in this cluster, genes encoding dehydrogenase/reductase (SDR family) member 3 (DHRS3), fatty acid synthase (FASN), fat storage-inducing transmembrane protein 2 (FITM2), lipase maturation factor 1 (LMF1), and lipoprotein lipase (LPL) were represented. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Significant KEGG identifiers detected based on the differentially expressed transcripts in cluster 2 for the acute phase response to *E. coli*infection\* ::: KEGG ID *P*value Odds Ratio Exp Count Count Size Term --------- ---------- ------------ ----------- ------- ------ ---------------------------------------------- 830 0.00011 19.46957 0.257269 4 24 Retinol metabolism 71 0.002574 12.51056 0.278708 3 26 Fatty acid metabolism 1100 0.016481 2.126324 6.817621 13 636 Metabolic pathways 61 0.021326 93.55556 0.021439 1 2 Fatty acid biosynthesis 980 0.029004 8.222903 0.267988 2 25 Metabolism of xenobiotics by cytochrome P450 561 0.031206 7.879108 0.278708 2 26 Glycerolipid metabolism 982 0.035802 7.270856 0.300147 2 28 Drug metabolism - cytochrome P450 \*A hypergeometric gene set enrichment test was performed based on cluster 2 of the contrast T24 *vs*. C24. Overrepresentation of gene sets defined by the KEGG database was tested using the Fisher\'s exact test. A gene set was considered significant if *P*\< 0.05. ::: Chronic stage response ---------------------- In total, only two DE transcripts (Bt.12553.1.S1\_at and Bt.7165.1.S1\_at) were detected (*P*\< 0.05). These encode haptoglobin (HP) and chemokine (C-X-C motif) ligand 5 (CXCL5), respectively. Overlap of DE genes between live tissue and related analysis in necrotic tissue ------------------------------------------------------------------------------- The DE genes were compared to two other related studies on gene expression in response to *E. coli*infection in the udder \[[@B14],[@B15]\]. In \[Additional file [4](#S4){ref-type="supplementary-material"}: Supplemental Figure S1\], the overlap of statistically significant DE genes at T = 24 h p.i. is presented. A group of 248 genes was represented in all three studies \[Additional file [5](#S5){ref-type="supplementary-material"}: Supplemental Table S4\]. Among this group are immune-related genes, namely those encoding chemokine (C-C motif) ligands (CCL4, CCL19), chemokine (C-X-C motif) ligands (CXCL2, CXCL5, CXCL16), intercellular adhesion molecules (ICAM1, ICAM3), immediate early response (IER3, IER5), interferon-induced transmembrane proteins (IFITM1, IFITM3), interferon regulatory factor 9 (IRF9), interleukins (IL1B, IL8), interleukin receptors (IL2RG, IL8RB, IL10RB), haptoglobin (HP), neutrophil cytosolic factors (NCF1, NCF2, NCF4), S100 calcium binding proteins (S100A2, S100A8, S100A9, S100A12), serum amyloid A 3 (SAA3), Toll-like receptors (TLR2, TLR4), and tumour necrosis factor receptor superfamily members (TNFRSF1A, TNFRSF6B). The hypergeometric gene set analysis based on the KEGG information revealed overrepresentation of 15 gene sets, of which one was related to fatty acid metabolism (glycerolipid metabolism (KEGG00561)) while the 14 other KEGG identifiers were related to the immune response: chemokine signalling pathway (KEGG04062), leukocyte transendothelial migration (KEGG04670), natural killer cell-mediated cytotoxicity (KEGG04650), Toll-like receptor signalling pathway (KEGG04620), and RIG-I-like receptor signalling pathway (KEGG04622) (Table [3](#T3){ref-type="table"}). ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Significant KEGG identifiers detected based on the differentially expressed genes between three studies of *E. coli*infection in the bovine udder\* ::: KEGG ID *P*value Odds Ratio Exp Count Count Size Term --------- ---------- ------------ ----------- ------- ------ ------------------------------------------- 4062 6.05E-08 6.785196 2.70837 15 106 Chemokine signaling pathway 4670 1.73E-06 7.273962 1.839648 11 72 Leukocyte transendothelial migration 4650 1.02E-05 7.646809 1.430837 9 56 Natural killer cell mediated cytotoxicity 4620 5.14E-05 7.058635 1.354185 8 53 Toll-like receptor signaling pathway 4060 0.00053 4.306462 2.350661 9 92 Cytokine-cytokine receptor interaction 4622 0.001365 6.982249 0.843172 5 33 RIG-I-like receptor signaling pathway 4514 0.002293 4.237402 1.839648 7 72 Cell adhesion molecules (CAMs) 4662 0.00231 6.105769 0.945374 5 37 B cell receptor signaling pathway 4920 0.003278 5.579882 1.022026 5 40 Adipocytokine signaling pathway 4210 0.01013 4.147677 1.328634 5 52 Apoptosis 4666 0.011834 3.977177 1.379736 5 54 Fc gamma R-mediated phagocytosis 4610 0.025259 3.88 1.124229 4 44 Complement and coagulation cascades 561 0.027695 5.044241 0.664317 3 26 Glycerolipid metabolism 4142 0.038281 2.85764 1.865198 5 73 Lysosome 4630 0.040239 2.815796 1.890749 5 74 Jak-STAT signaling pathway \*A hypergeometric gene set enrichment test was performed based on overlapping genes between the study described in this paper, Mitterhuemer *et al*. \[[@B14]\], and Rinaldi *et al*. \[[@B15]\]. Overrepresentation of gene sets defined by the KEGG database was tested using the Fisher\'s exact test. A gene set was considered significant if *P*\< 0.05. ::: Discussion ========== In this study, we analyzed the global gene expression changes in the bovine mammary gland that occurred during the acute phase and chronic stage of experimental *E. coli*mastitis in dairy cows in early lactation. Tissue samples were collected with a biopsy pistol, which made it possible to sample the same *E. coli*-infected cows *in vivo*at two different time periods. The acute phase was verified by the presence of high *E. coli*CFU/ml in milk, fever, increased SCC and the presence of acute phase proteins MAA and SAA in milk and blood, respectively. Marked changes in gene expression were detected in the acute stage at 24 h p.i. In agreement with results from other related studies the up-regulated genes were related to the immune response, with the majority of genes involved in the induction and regulation of the local inflammatory response and the APR \[[@B9],[@B16]\]. Among the top 10 KEGG pathways associated with these genes were pathways involved in chemokine signalling, Toll-like receptor signalling, leukocyte transendothelial migration, cytokine-cytokine receptor interaction, cell adhesion and Fc gamma R- mediated phagocytosis. This finding indicates the importance of recruitment and activation of macrophages and neutrophils to sites of infection during acute *E. coli*mastitis \[[@B17],[@B18]\]. In the acute phase, the group of down-regulated genes in the mammary gland was mainly associated with the lipid metabolism and fatty acid pathways. It is well known that cows with *E. coli*mastitis experience decreased milk production \[[@B19],[@B20]\]; however, the affected molecular mechanisms related to fat synthesis in those cows has been undefined \[[@B21]\]. The results presented in this paper, in conjunction with the results of Mitterhuemer *et al*. \[[@B14]\], clearly demonstrate that *E. coli*infection of the mammary gland reduces the expression of genes involved in fat metabolism. However, further research is needed to verify whether these genes actually influence the total amount of fat secretion in the milk or instead influence the fat percentage in the milk. In the chronic, subclinical stage examined at 192 h p.i., all cows were found to have cleared the *E. coli*bacteria and re-established normal body temperatures. Yet, the increased SCC and increased concentrations of MAA in milk were maintained. At this stage, gene expression had returned to control levels, with the notable exception of the genes encoding for the acute phase protein haptoglobin (HP) and the chemokine CXCL5. Both genes were found to be up-regulated at the APR and chronic stage response. CXCL5 expression on the surface of eosinophils facilitates their ability to recruit and activate CXC receptor 2 (CXCR2)-bearing cells, such as neutrophils, to the site of inflammation. In this manner, CXCL5 can influence tissue remodelling \[[@B22]\], such as would be expected in recovering udder tissue. It is likely that more up-regulated genes associated with tissue modelling could have been identified if the samples had been collected earlier. However, as the biopsy procedure itself causes tissue damage and bleeding, there is a limit to how often a biopsy can be collected without influencing the gene expression results. By comparing our study\'s results with those from two recent array studies using *post-mortem*tissues from *E. coli*infected cows \[[@B14],[@B15]\], we identified a particularly strong dataset of the statistically significant genes and genetic pathways present in the mammary tissue during the acute phase at 24 h. All three studies were conducted in the Holstein Friesian breed, but differed with regard to *E. coli*strain and dose, cows\' lactation number and stage, and the sampling procedure. In our study, a low dose (20-40 CFU/quarter) model was established in primiparous dairy cows in early lactation, and tissue samples were collected *in vivo*using biopsies at time points 24 h and 192 h p.i. in the lobuli-alveoli area. Rinaldi *et al*. \[[@B15]\] used a medium dose (400 CFU/quarter) in a two-step inoculation model in two quarters, with a 12 h time-delay in multiparous cows in late lactation; these tissue samples were collected *post-mortem*from euthanized cows at 12 h and 24 h p.i. from four different portions of the udder. Mitterhuemer *et al*. \[[@B14]\] also used a medium dose model (500 CFU/quarter), but relied on mid-lactation cows and collected tissue samples *post-mortem*at 6 h and 24 h p.i. from a less well-defined area of the udder. Furthermore, the gene analysis platforms used were different between each of the studies. All of the factors mentioned above can influence disease kinetics and gene expression levels in the mammary gland at 24 h p.i. This fact may explain why each study found a unique set of differentially expressed genes, as shown in the non-overlapping areas of the Venn-diagram \[Additional file [4](#S4){ref-type="supplementary-material"}: Supplemental Figure S1\]. A large group of DE genes were common between the three studies and were determined to be unaffected by *E. coli*strain, dose, cow factors, sampling method and gene expression analysis (\[Additional file [4](#S4){ref-type="supplementary-material"}: Supplemental Figure S1\], \[Additional file [5](#S5){ref-type="supplementary-material"}: Supplemental Table S4\]). The gene sets obtained from the analysis of the common 248 DE genes were mainly related to the immune response and are associated with the response to *E. coli*mastitis \[[@B14],[@B15]\] (Table [3](#T3){ref-type="table"}). However, in all three studies two of the 15 pathways were of interest because they are principally related to intracellular infections (e.g. viruses, protozoa and bacteria). The first pathway, the natural killer cell-mediated cytotoxicity pathway (KEGG04650), ranked fifth on our list (Table [1](#T1){ref-type="table"}) and third in the comparison between the *E. coli*gene expression studies (Table [3](#T3){ref-type="table"}). The NK cells are innate lymphocytes, which function to eliminate all aberrant cells which have completely lost, or express insufficient amounts of, MHC class I molecules; virally-infected cells or tumorigenic cells are examples of such targets \[[@B23],[@B24]\]. Moreover, NK cells are major cytokine producers and participate in the onset and shaping of the adaptive immune response. So far, bovine NK cells have been investigated in relation to intracellular infections with protozoa and mycobacteria \[[@B24]\]. The mechanisms underlying NK cells role(s) in the defence against extracellular bacteria, such as *E. coli*, remains a question of interest. NK cell lyses of damaged apoptotic and necrotic cells is one possibility. Furthermore, NK cells might be involved in a response to *E. coli*via the dendritic cells (DC) \[[@B25]\]. The main role of the DC is processing antigens and presenting antigens to other cells, namely the antigen presenting cells (APC). In this manner, the NK cells may play a role in maintaining homeostasis of the immune response during bacterial infections by regulating the APC amount in the secondary lymphoid organs \[[@B25]\]. In \[Additional file [5](#S5){ref-type="supplementary-material"}: Supplemental Table S4\], CD69, a DC-induced activation marker of NK, as well as IL2RA and IL2RG, which are part of the IL2 receptor complex on lymphocytes and NK cells, are found in the natural killer cell-mediated cytotoxicity pathway. This suggests a possible function of NK cells in the immune defence process directed against *E. coli*in the mammary gland. To our knowledge, NK and DC have not been identified in healthy or inflamed mammary gland tissues from dairy cows; and, it remains to be shown if and how they are recruited from the blood compartment or local lymph nodes to the mammary gland. The second interesting pathway, the Rig-I-like receptor signalling pathway (KEGG04622), ranked sixth on our list. Cells of the innate immune system, such as macrophages and DC, express a number of pattern-recognition receptors (PRR) that specifically recognize unique pathogen-associated molecular patterns (PAMP). In this study, as well as in others, TLR2 and TLR4 are characterized as being significantly influenced by *E. coli*mastitis. Comparison of the three related studies also provided evidence for involvement of molecules functioning in the RIG-I-like receptor pathway \[Additional file [5](#S5){ref-type="supplementary-material"}: Supplemental Table S4\]. The gene-encoded products include receptors such as RIG-I, Mda5, and LGP2, which are cytoplasmatic proteins that recognize viral DNA and RNA \[[@B26]\]. Following recognition of viral-associated molecular patterns, the RIG-I-like receptors stimulate the synthesis of multiple cytokines including the group of interferon type 1 cytokines and pro-inflammatory cytokines that influence protein synthesis, growth arrest and apoptosis. Furthermore, the RIG-I-like receptors enhance DC maturation, NK cell activation, antibody production and differentiation of cytotoxic T-lymphocytes, which in turn facilitate the adaptive immune response \[[@B26]\]. Whether the RIG-I-like receptors are up-regulated directly from contact with *E. coli*components in the mammary gland or indirectly via cross-induction through other TLRs is yet unknown. The exact role of the natural killer cell-mediated cytotoxicity pathway (KEGG04650) and the Rig-I-like receptor signalling pathway (KEGG04622) in the response to *E. coli*infection of the bovine udder is still unresolved. The fact that identical pathway-associated genes were similarly regulated during the APR in three independent studies indicates that further investigation of these pathways and their related cells will provide significant insights into the pathogenesis of *E. coli*mastitis. Furthermore, this work sets stage for additional work to determine whether patterns of gene expression associated with udder infections with Gram positive bacteria e.g. *S. aureus*and *S. uberis*differs from that of Gram negative infection and whether the timing of gene changes are similar. Conclusions =========== In conclusion, we have identified multiple DE genes involved in the APR and two genes in the chronic stage response to *E. coli*-induced mastitis in the mammary gland of dairy cows in early lactation. The up-regulated genes from the local APR were involved in the immune response, while the down-regulated genes were primarily involved in fat metabolism. Comparison of the DE genes in the APR of our study with those detected in two other related gene expression studies on *E. coli*mastitis revealed a common set of 248 DE genes existed despite differences in *E. coli*strain and dose, cow lactation stage and number, tissue collection method and gene analysis method. Based on the annotation and gene pathway analysis it is suggested that these genes play a central role in general host defence against *E. coli*infections in the mammary gland. Methods ======= Animals and Treatment --------------------- Sixteen, healthy primiparous Danish Holstein-Friesian cows were challenged intramammarily with *E. coli*at four to six weeks after parturition. Prior to disease challenge, the general health and udder health of a total of 24 dairy cows were evaluated based on body temperature, white blood cell count (WBC), glutaraldehyde test, Californian Mastitis Test (CMT; (Kruuse, Marslev, DK)) and bacteriological examinations of foremilk samples. Only those cows with normal body temperature and WBC, negative glutaraldehyde test, low CMT test (range 1-5) and found to be free from major mastitis pathogens were used in the study. Tissue quarters for *E. coli*inoculation and biopsy were selected based on CMT scores (≤2) and SCC in foremilk by using the portable DeLaval Cell Counter (DCC; DeLaval, Tumba, Sweden) (range 1-6000 × 10^3^cells/ml). The front quarter with the lowest SCC (\< 27000 cells/ml) was used for the *E. coli*inoculations. Control quarters were chosen based upon bacteriological examinations conducted prior to *E. coli*inoculation and on the quarter foremilk SCC at 24 h (\<181000 cells/ml) and 192 h (\<215000 cells/ml). The day prior to the *E. coli*inoculation, sterile Micro-Renathane polyvinyl catheters were inserted into the jugular vein and flushed with a sterile 0.9% NaCl solution containing 50 IU Na-heparin (Loevens Kemiske Fabrik, Ballerup, Denmark). The cows were housed in a traditional straw-bedded tie stall barn, where they were individually fed and given free access to water. A total mixed ration (TMR) diet including vitamins and minerals was fed *ad libitum*twice a day in equal portions at 0800 and 1530 (24 hr clock). The cows were milked at 0600 and again at 1700. All procedures involving animals were approved by the Danish Animal Experiments Inspectorate and complied with the Danish Ministry of Justice Laws concerning animal experimentation and care of experimental animals. Inspection was carried out by members of the Danish Animal Experiment committee during the acute stage of the disease. Preparation and inoculation of *E. coli* ---------------------------------------- The *E. coli*used was a Danish field isolate (k2bh2) isolated from a cow with severe, acute mastitis (kindly donated by Dr. Helle Daugaard Larsen, formerly of the Danish Institute for Food and Veterinary Research, Copenhagen, Denmark). All procedures involving handling of the *E. coli*inoculums were conducted in a laminar air flow (LAF) bench under sterile conditions. Prior to the inoculation, *E. coli*was grown overnight at 37°C in brain heart infusion medium (BHI; Merck, VWR - Bie & Berntsen, Denmark) in a 200 rpm-shaking incubator (Stuart Orbital incubator s150; VWR - Bie & Berntsen). The following day, the bacteria suspension was tested for purity on two non-selective enriched agars: mastitis blood agar plates (BA; Steins Eurofins A/S, Holstebro, Denmark) and tryptic soya agar (TSA; BD Difco™ - BD Denmark A/S) and for *E. coli*on a selective coliform MacConkey agar (MCA; Merck). The *E. coli*suspension was transferred to fresh BHI medium (1:10 dilution) and incubated for and additional 3 h at 37°C. Optimal bacteria growth was tested by measuring the OD (600 nm; value \>1.8) on a spectrophotometer (Gene-Quant Pro; Amersham Pharmacia Biotech, Uppsala, Sweden). From then on, the *E. coli*suspension was kept on ice. Centrifugation was performed at 14000 rcf, 4°C for 20 min (Sorvall, RC6+ centrifuge; Axeb Danmark A/S, Albertslund). The bacteria pellet was washed twice and resuspended in 15 ml of cold endotoxin-free 0.9% NaCl (Løven Apotek, Hammershøj, Denmark), and 10-fold dilutions were made from this suspension. The number of washed bacteria in the dilutions was counted using the plate count agar (PCA) method on small TSA and MCA plates. The inoculum was kept overnight at 4°C. The following morning, the number of colony-forming units (CFU)/ml was counted using a manual colony counter. Then, *E. coli*was diluted in sterile 0.9% endotoxin-free NaCl to obtain 2-4 CFU/ml. Each dairy cow was inoculated with 10 ml of 0.9% pyrogen-free NaCl solution containing \~20-40 CFU live *E. coli*in one front quarter immediately after evening milking (T = 0 h). The control quarter was not inoculated. Each teat was disinfected twice with cotton wool pre-wetted with 70% ethanol. The *E. coli*-NaCl solution was infused into the gland with a sterile teat cannula and the quarter was thoroughly massaged. After the inoculation, the remaining bacteria suspension was retested in the laboratory to determine *E. coli*number by using large agar plates with TSA and MCA to accommodate 1 ml volume. Clinical examinations, milk and blood sampling, and laboratory analysis ----------------------------------------------------------------------- Clinical and para-clinical measurements were recorded throughout the experiment. The body temperature, *E. coli*CFU/ml milk, SCC, Milk Amyloid A (MAA) concentration in the milk and the concentration of Serum Amyloid A (SAA) in the blood measured at 0, 24, and 192 h relative to the time of *E. coli*inoculation are reported here. ### Bacteriological analysis of milk Ten ml of foremilk samples were aseptically collected from the *E. coli*infected quarter and negative control quarter prior to infection (0 h) and immediately before the udder biopsies were collected at 24 h and 192 h. Quantification of *E. coli*(CFU/ml) was conducted on MCA plates using serial dilutions (1:10) of the milk samples in 9 ml of 0.9% NaCl. From each dilution, 0.1 ml was spread on an MCA plate and incubated for 24 h at 37°C, after which the number of pink colonies were counted using a manual colony counter. In addition, 10 μl aliquots of foremilk were cultured on BA and TSA for 48 h at 37°C in order to rule out the presence of other mastitis pathogens. The minor mastitis pathogen, coagulase negative *staphylococcus spp*., was found in three control quarters at 24 h after inoculation; specifically, two cows presented with increased SCC and these two quarters were excluded from the array analysis. ### SCC Foremilk samples were analyzed for SCC as previously described. ### MAA and SAA MAA measurements were carried out as previously described \[[@B27]\]. Blood sampling and SAA measurements were carried out as previously described \[[@B28]\]. Udder biopsies -------------- Udder biopsies were collected from the infected quarter and a healthy control quarter of each cow at 24 h (acute stage: T24; C24) and 192 h (chronic: T192; C192) after the *E. coli*challenge. Sampling was conducted after the evening milking using the following procedure: The cows were immobilized in their tie stall using a halter and rope. The skin surface of each quarter was thoroughly washed and dried. A mild sedation injection of 0.1 ml Domosedan^®^Vet per 100 kg body weight (10 mg/ml Detomidin; Orion Pharma, Espoo, Finland) was administered intravenously. A skin area of 10 × 10 cm in the middle of each quarter (lobuli-alveolar gland region) was shaved and disinfected twice with 70% ethanol. Local anesthetic consisted of xylocaine (1% lidocaine) (Astra Zeneca A/S, Albertslund, Denmark) that was sprayed onto the shaved skin area. After 10 min, a 0.3-0.6 cm long incision was made in the udder skin with a scalpel, trying to avoid blood vessels visible under the skin. Five udder biopsies were collected at the incision point using a biopsy pistol developed for obtaining biopsies in humans (Manan Automatic Biopsy System; Marmon/MDTech, Gainesville, FL). The needle of the biopsy pistol was 14 gauge with a 17 mm notch, which collected 10-15 mg tissue per biopsy. Bleeding from the incision wound was allowed briefly, after which pressure with sterile cloth tampons were placed on the incision areas to stop bleeding and avoid hematoma formation. After the udder biopsies had been collected, cows were administered a prophylactic antibiotic treatment against infection with Gram-positive bacteria by intramuscular injection of 30 ml of Penovet^®^vet (300000 IE benzylpenicillinprocain/ml; Boehringer Ingelheim Danmark A/S, Copenhagen, Denmark). The udder biopsies were snap frozen in liquid nitrogen and transported to the laboratory where the tissue was stored at -80°C until RNA extraction. Expression profiling using microarrays -------------------------------------- The number of samples used in the microarray analysis was: T24, 12; C24, 9; T192, 14; and C192, 14. Isolation and labelling of RNA and microarray processing was performed as described previously \[[@B29]\]. RNA from the udder biopsies was isolated using Trizol Reagent (Invitrogen, Taastrup, Denmark). Five μg total RNA was labelled using the SuperScript Choice System (Life Technologies), according to the manufacturer\'s instructions. Biotin-labelled cRNA was prepared using the BioArray High Yield RNA Transcript Labelling Kit (Enzo, Farmingdale, NY, USA), and 15 μg was loaded onto the probe array cartridge (Bovine Genome Array; Affymetrix, Santa Clara, CA, USA). The array contained 24128 probe sets which represented 15264 UniGene (annotation from May, 2006) to measure the global transcripts. The Bovine Genome Array annotation available from NetAffx™ Analysis Centre (Bovine.na29.annot.csv) was used. Additional and updated annotation was obtained from the Ensemble database, using the biomaRt package (version 2.0.0) in R \[[@B30]\]. In total 10401 probes on the array had a corresponding Ensemble ID, and 9369 probes were able to be assigned to the group Biological Processes (BP) of the Gene Ontology (GO) database <http://www.geneontology.org/>. Statistical analysis -------------------- The data was analyzed using R (version 2.10.0) <http://www.r-project.org/>. Normalization of the expression values for the udder samples was performed using the Robust Multi-array Average (RMA) algorithm as implemented in the Affy package (version 1.24.2). More detailed descriptions of the microarray experiment and data are available at the NCBI\'s Gene Expression Omnibus (GEO) database \[[@B31],[@B32]\] through the accession number GSE24217. Differential expression of each gene was assessed using linear modelling and empirical Bayes methods, which were implemented using the R package Limma (version 3.2.1) \[[@B33]\]. The linear models allowed for changes according to time points. Comparisons of the APR and chronic stage response were 24 h p.i. *vs*. 24 h control (T24 - C24) and 192 h p.i. *vs*. 192 h control (T192 - C192), respectively. Each transcript targeted by a probe was tested for its expression change using a modified Student\'s *t-*test. In the modified *t-*test, the residual standard deviations were moderated across the probe sets to ensure that there was a more stable inference for each transcript. The moderated standard deviations were considered as a compromise between the individual transcript standard deviations and the pooled transcripts standard deviation. Multiple testing was accounted for using the Bonferroni correction method. Probes were considered to be DE if the corrected *P*-value was below 0.05 and had a minimum log2-fold change of 1.0 for both the APR and chronic stage response. Two-way clustering of the DE genes and samples (animals) was performed using the heatmap.2 option (gplots package version 2.7.4). A hyper geometric gene set enrichment test (GOstats package version 2.12.0) was performed based on the clusters identified in the APR or chronic stage response comparison. Overrepresentation of gene sets defined by the group of BP in the GO database or the Kyoto Encyclopedia of Genes and Genomes database (KEGG) <http://www.genome.jp/kegg/> was tested using Fisher\'s exact test. For this test, only the significant genes which were annotated with an Entrez gene ID were included. When a gene had a duplicate on the array, only a single gene ID was used. A gene-set was considered significant if *P*\< 0.05. The differentially expressed genes from our study at T = 24 h p.i. were compared to the *E. coli*expression studies (also at T = 24 h p.i.) previously reported in the literature \[[@B14],[@B15]\]. The Affymetrix probe identifiers for the DE genes from the study by Mitterhuemer *et al*. \[[@B14]\] were taken from the GEO database (GSE15025), while the gene identifiers for and Rinaldi *et al*. \[[@B15]\] were taken from GSE15441). Expression levels were calculated at T = 24 h p.i. for all four regions in the udder mentioned in the study and a probe was included if it was differentially expressed in any of the regions. Since the arrays in Rinaldi *et al*. \[[@B15]\] were not Affymetrix, the probes were translated to Affymetrix probe IDs. Then, the annotated probes were matched with the list of annotated probes from our study and a Venn diagram was generated to show the overlap between the studies. A hypergeometric gene set enrichment test was performed based on the KEGG pathways, as described earlier based on the DE genes that were common between our study, Mitterhuemer *et al*. \[[@B14]\], and Rinaldi *et al*. \[[@B15]\] (GOstats package version 2.12.0). A gene set was considered significant if *P*\< 0.05. Authors\' contributions ======================= BB analyzed the microarray data and wrote the manuscript. CR carried out the animal challenges. KLI and CR collected the mammary biopsy samples. PS, CR and KLI designed the experimental plan. SME made the comparison between the gene lists of the different *E. coli*gene expression experiments available in the literature. BB, PS, CR, SME and KLI contributed to the interpretation of the results, discussion and refinement of the manuscript. The authors agreed on the contents of the paper. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Table S1: Differentially expressed transcripts for the acute phase response to *E. coli*infection in the bovine udder**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Table S2: Significant GO identifiers detected based on the differentially expressed transcripts in cluster 1 for the acute phase response to *E. coli*infection**. A hypergeometric gene set enrichment test was performed based on cluster 1 of the contrast T24 *vs*. C24. Overrepresentation of gene sets defined by the GO database <http://www.geneontology.org/> was tested using the Fisher\'s exact test. A gene set was considered significant if *P*\< 0.05. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Table S3: Significant GO identifiers detected based on the differentially expressed transcripts in cluster 2 for the acute phase response to *E. coli*infection**. A hypergeometric gene set enrichment test was performed based on cluster 2 of the contrast T24 *vs*. C24. A gene set was considered significant if *P*\< 0.05. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **Figure S1: Venn diagram showing the overlap of differentially expressed genes between three studies on *E. coli*infection in the bovine udder at T = 24 h post-infection**. 1) \"This study\": the study described in this manuscript, 2) \"Mitterhuemer\": the study described by Mitterhuemer *et al*. \[[@B14]\], and 3) \"Rinaldi\": the study described by Rinaldi *et al*. \[[@B15]\]. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 5 **Table S4: Overlap of differentially expressed genes between three studies of *E. coli*infection in the bovine udder**. Data from T = 24 h post-infection in three different studies: 1) \"This study\": the study described in this paper; 2) \"Mitterhuemer\": the study described by Mitterhuemer *et al*. \[[@B14]\]; and 3) \"Rinaldi\": the study described by Rinaldi *et al*. \[[@B15]\]. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We would like to thank and acknowledge the staff at the Department of Animal Health and Bioscience\'s Dairy Cattle facility and Martin Bjerring, Jens Clausen, Dorte Agnholt, Hanne Møller Purup, Elisabeth Mark and Lene Niklassen for their excellent technical assistance and analytical skills. This study was partly funded by the European Commission, within the 6^th^Framework Program (contract No. FOOD-CT-2006-016250) and the BIOSENS project granted by the Danish Ministry of Food, Agriculture and Fisheries (Innovations Law), Lattec I/S, the Danish Cattle Association, and the Faculty of Agricultural Sciences, Aarhus University.

PubMed Central

2024-06-05T04:04:19.605076

2011-2-28

PMC3053263

Background ========== Chromatin immunoprecipitation (ChIP) followed by genomic tiling microarray hybridization (ChIP-chip) or massively parallel sequencing (ChIP-seq) are two of the most widely used approaches for genome-wide identification and characterization of *in vivo*protein-DNA interactions. They can be used to analyze many important DNA-interacting proteins including RNA polymerases, transcription factors, transcriptional co-factors, and histone proteins \[[@B1]\]. Indeed these genome-wide ChIP analysis approaches have led to many important discoveries related to transcriptional regulation \[[@B2]-[@B4]\], epigenetic regulation through histone modification \[[@B5]\], nucleosome organization \[[@B6],[@B7]\], and interindividual variation in protein-DNA interactions \[[@B8],[@B9]\]. ChIP-chip first appeared in the literature about 10 years ago and was one of the earliest approaches to performing genome-wide mapping of protein-DNA interactions in organisms with small genomes, such as yeast \[[@B2],[@B10]\]. Currently, various tiling microarray platforms of common model organisms are well supported by commercial vendors, and many bioinformatics tools have been developed for ChIP-chip analysis \[[@B11]-[@B14]\]. Fueled by rapid development of the second generation high-throughput sequencing technologies in the past few years, ChIP-seq has emerged as an attractive alternative to ChIP-chip \[[@B1]\]. For instance, ChIP-seq generally produces profiles with higher spatial resolution, dynamic range, and genomic coverage, allowing it to have higher sensitivity and specificity over ChIP-chip in terms of protein binding site identification. Further, ChIP-seq can be used to analyze virtually any species with a sequenced genome since it is not constrained by the availability of an organism-specific microarray. Many current ChIP-seq protocols can work with a smaller amount of initial material compared to ChIP-chip \[[@B15],[@B16]\]. Moreover, ChIP-seq is already a more cost-effective way of analyzing mammalian genomes, and the cost effectiveness will likely become more apparent as the cost of high-throughput sequencing technology continues to drop. These factors have led to the rapid adoption of ChIP-seq technology. However, despite the widespread use of both ChIP-chip and ChIP-seq, only a few small-scale studies have attempted to quantitatively compare these technologies using real data. Euskirchen et al. \[[@B17]\] compared the STAT1 binding sites identified by ChIP-chip and ChIP-PET (paired-end ditag sequencing by Sanger sequencing technology) and found that there was a good overall agreement between the two technologies, particularly at identifying highly ranked enrichment regions. They nonetheless noted specific discrepancies in regions associated with repetitive elements, which can be attributed to lack of microarray probe coverage or misalignment of ChIP-PET reads. More recently, a number of studies compared genome-wide transcription factor binding datasets generated from ChIP-chip with those generated from ChIP-seq \[[@B18]-[@B22]\] (see Additional file [1](#S1){ref-type="supplementary-material"}: Table S1). The general conclusions from these studies were that binding profiles generated from ChIP-chip and ChIP-seq were largely correlated at the genome-wide level, and that ChIP-seq had superior sensitivity and specificity over ChIP-chip in terms of binding site identification as determined by motif enrichment or quantitative PCR validation. It was also found that the strongest peaks were more likely to be detected by both technologies. However, only a few pairs of ChIP-chip/ChIP-seq profiles were analyzed in these studies, and their focus was mainly on the ability to identify narrow enrichment regions using specific peak calling algorithms. As shown previously \[[@B23]\] and in this study, peak identification can be strongly dependent on the analysis algorithm, so other more general comparison metrics should be used. In addition, little is known about the technology-specific variation for analyzing histone modification data. ChIP-based histone modification data is commonly used to reconstruct average signal profiles, or \"epigenetic signatures,\" of key genomic regions such as the transcription start and end sites, but the impact of using ChIP-chip versus ChIP-seq data for constructing epigenetic signatures is largely unknown. Furthermore, it is also important to understand technology-specific biases associated with high-throughput sequencing. Recent studies indicated that the distribution of cross-linked and sonicated DNA fragments (input DNA) was affected by chromatin structure, copy number variation, occurrence of genomic repeats, mappability, genomic location, gene expression activity, and genomic GC content variation \[[@B24]-[@B26]\]. Since input DNA is commonly used as a background control for a ChIP-seq experiment, it is important to assess how such variation affects the analysis of ChIP-seq data. Therefore a thorough understanding of the technological variation between ChIP-chip and ChIP-seq is important in experimental design and data analysis. In this study, we compiled and analyzed 31 pairs of ChIP-chip/ChIP-seq profiles of technical replicates across eight immunoprecipitation (IP) factors (CBP, RNA PolII, and six histone modifications) at four developmental stages of the common fruit fly *Drosophila melanogaster*(Table [1](#T1){ref-type="table"}) as part of the model organism Encyclopedia of DNA Elements (modENCODE) project \[[@B27]\]. In addition, our compiled dataset comprises another 62 ChIP-chip profiles (biological replicates) in the same set of biological conditions (i.e., three ChIP-chip biological replicates at each developmental stage/IP combination), nine sequencing profiles of input DNA, and four pairs of ChIP-seq/ChIP-seq replicates (Table [2](#T2){ref-type="table"}). Agilent\'s tiling microarray (Agilent custom 3X244K Dmel Whole Genome Tiling Microarray) and Illumina\'s Genome Analyzer II platforms were used to generate the ChIP-chip and ChIP-seq data, respectively. All data used in this study were generated as part of the modENCODE project, and are accessible from NCBI GEO (accession numbers: GSE15292, GSE16013, and GSE20000). The goal of this study was to quantify reproducibility within and between profiles generated using ChIP-chip and ChIP-seq approaches, and to pinpoint the source of variation between the technologies, which ultimately should provide useful information for experimental design and data analysis. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Summary of the ChIP-chip and ChIP-seq profiles analyzed in this study ::: ChIP-chip and ChIP-seq replicates ----------------------------------- ------------- --------------- --------------- --------------- ---------- --------- ---------- --------- **IP factor** **E-0-4 h** **E-12-16 h** **E-16-20 h** **E-20-24 h** **chip** **seq** **chip** **seq** **chip** **seq** **chip** **seq** **CBP** 3 1 \-\-- \-\-- 3 1 3 1 **H3K27Ac** 3 1 3 1 3 1 3 1 **H3K27Me3** 3 1 3 1 3 1 3 1 **H3K4Me1** 3 1 3 1 3 1 3 1 **H3K4Me3** 3 1 3 1 3 1 3 1 **H3K9Ac** 3 1 3 1 3 1 3 1 **H3K9Me3** 3 1 3 1 3 1 3 1 **PolII** 3 1 3 1 3 1 3 1 **INPUT** **\*** 1 \* 1 \* 1 \* 1 For each combination of developmental stage and IP factor, three biological samples were profiled by ChIP-chip, and one of these samples was independently profiled by ChIP-seq. Therefore we have three biological ChIP-chip/ChIP-chip replicates, and one technical ChIP-chip/ChIP-seq replicate pair for every biological condition. \*Although no independent microarray profiles of input DNA were available, we extracted input DNA profiles from the input channel (of the two-channel microarray) for each ChIP-chip profile. We therefore have 31 × 3 = 93 INPUT-chip profiles in this collection, where all INPUT-chip profiles within each time point are technical replicates. ::: ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Summary of the additional ChIP-seq profiles analyzed in this study ::: ChIP-seq replicates only -------------------------- ------------- --------------- ----------------- ------------- **IP factor** **E-4-8 h** **AdultMale** **AdultFemale** **S2 cell** **seq** **seq** **seq** **seq** **CBP** 2 2 2 \-\-- **CTCF** \-\-- \-\-- \-\-- 2 **INPUT** 1 1 1 2 Details of the four additional pairs of ChIP-seq/ChIP-seq (biological) replicates analyzed in this study. The two CTCF ChIP-seq profiles were technical replicates, whereas all other ChIP-seq pairs were biological replicates. ::: Results ======= Analysis of input DNA profiles ------------------------------ To understand the technological differences between sequencing-based and microarray-based ChIP data, we first analyzed the profiles of cross-linked and sonicated DNA fragments (input DNA) generated by microarray (INPUT-chip) and high-throughput sequencing (INPUT-seq). Since the input DNA profile should be independent of the antibody used for ChIP, this comparison can give insight into the specific differences between these two profiling technologies. We obtained INPUT-chip data from the background channel of our two-channel microarray data. While this microarray platform uses competitive hybridization, the two channels in our Agilent microarray have been shown to be relatively independent as saturation in either channel is very rare \[[@B28]\]. Out of all the INPUT-chip profiles that we extracted, we only present the analysis of eight representative profiles here (two from each of the four developmental time points) since most of the INPUT-chip profiles are very similar (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S1). The eight INPUT-chip profiles were then compared with the nine INPUT-seq profiles collected in this study (Additional file [1](#S1){ref-type="supplementary-material"}: Table S3). One of the most striking observations is that INPUT-chip and INPUT-seq profiles appear to be substantially different, even though the same input DNA material was used for microarray hybridization and sequencing (Figure [1](#F1){ref-type="fig"}). The relative magnitude and location of the peaks seem to be consistent across the INPUT-chip profiles from multiple experiments. However, the patterns in the nine INPUT-seq profiles appear to be more variable. We can visually identify many regions that have inconsistent signal enrichment across multiple INPUT-seq profiles (highlighted in Figure [1a](#F1){ref-type="fig"}). A clustering analysis was performed to quantify this observation. We found that all eight INPUT-chip profiles clustered closely to one another (Figure [1b](#F1){ref-type="fig"}). This result shows that the background DNA distribution measured from microarray and high-throughput sequencing is different. All INPUT-chip and seven out of nine INPUT-seq profiles correlated positively with genomic GC content at the genome-wide level (Figure [1b](#F1){ref-type="fig"}), as well as around the transcription start sites (TSS) and transcription end sites (TES) (Figure [1c](#F1){ref-type="fig"}). The strength of the correlation with GC is highly consistent among INPUT-chip profiles, but highly variable among the INPUT-seq profiles (Figure [1b-c](#F1){ref-type="fig"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S2). Notably, the INPUT-seq profiles obtained at E-16-20 h (E16) and E-20-24 h (E20) do not correlate with GC content. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Comparison of input DNA profiles obtained by microarray and sequencing technologies**. (a) A genome browser view of input DNA profiles of chromosome 2R of *D. melanogaster*at various developmental stages measured by microarray (INPUT-chip; blue) and sequencing (INPUT-seq; red). (b) A heat map that summarizes the Spearman correlation coefficient between every pair of the nine INPUT-seq and eight INPUT-chip profiles along with genome-wide GC content. The number of mappable reads (in million) is written next to the name of each INPUT-seq profile. (c) The average signal profiles of INPUT-seq and INPUT-chip around transcription start sites (TSSs) and transcription end sites (TESs) are largely consistent, and their variation along these genomic regions generally coincide with GC content variation. (d) We generated 11 additional profiles from one of the INPUT-seq samples (AM) by subsampling the reads at different proportions (90%,80%,\...,10%,5%,1%). A heat map summary representation of the Spearman correlation coefficient between every pair of sub-sampled INPUT-seq profiles and GC content is shown here. (e) The relationship between sequencing depth and genomic coverage. The curve shows how sequence read subsampling (i.e., reducing sequencing depth) affects genomic coverage. The genomic coverage of the nine INPUT-seq datasets and our Agilent microarray are also shown in the plot. :::  ::: We also note that INPUT-seq with higher sequencing depth (\>4 million mapped reads) tend to cluster together more tightly than those with lower sequencing depth, suggesting that there may be a relationship between sequencing depth and input DNA variability. To test this hypothesis, we generated 11 additional INPUT-seq profiles by subsampling sequencing reads from the most deeply sequenced input DNA sample (AdultMale; AM) at different sampling proportion (Figure [1d](#F1){ref-type="fig"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S3). As expected, profiles with higher sequencing depth tend to cluster more strongly together, and their correlation with GC content variation is more consistent. However, the GC content correlation only becomes much weaker only at a very low sequencing depth (\<2 million reads; Figure [1d](#F1){ref-type="fig"}). This indicates that low sequencing depth is not the only factor affecting INPUT-seq quality. Moreover, some INPUT-seq with relatively low sequencing depth (E0 and AF, \<4 million reads) can give consistent input DNA profiles. This implies that INPUT-seq variability may also be attributed to other experimental factors. Although further studies are required to dissect the full range of experimental factors affecting variability of input DNA libraries, it could be affected by variations in the sample preparation (e.g., different chromatin preparation and sonication), run-to-run variation of the sequencer, sequencer-to-sequencer variation for the same model, and a host of other variables in experiments. The high variability among INPUT-seq profiles is indeed a critical issue, since large variability contributes to instability of density estimation in a ChIP-seq profile, which will affect downstream data analysis. As will be shown in subsequent sections of this paper, an INPUT-seq with unusually weak correlation with GC content can impact construction of average profiles at important genomic locations. It is thus imperative to sequence the input DNA to sufficient depth and to ascertain that the obtained profile is consistent with those from similar experiments. Genomic coverage is another key consideration when choosing between ChIP-chip and ChIP-seq. The genomic coverage of ChIP-chip is limited by the microarray probe design, and the coverage of ChIP-seq is dependent on sequencing depth. The genomic coverage achieved by our Agilent microarray is about 70%. Using the sub-sampled INPUT-seq data, we show that INPUT-seq generally provides higher genomic coverage at sequencing depth as low as one million reads. This trend constructed from the randomly subsampled data corroborates the observed genomic coverage of the other eight real INPUT-seq datasets (Figure [1e](#F1){ref-type="fig"}). Comparison of profile characteristics ------------------------------------- We then compared the characteristics of ChIP-chip and ChIP-seq profiles. To compare the profiles generated by the two technologies, we divided the genome into 1 kb non-overlapping bins and defined the enrichment level at each bin as the average of log ratio of the IP channel over the input channel (see the Methods section for details). We refer to a signal distribution of a ChIP profile as its distribution of enrichment values of all the bins. First, we aimed to characterize the average signal-to-noise ratio for profiles generated by the two technologies. We used the (truncated) skewness of the signal density profile after removing signals from the highest and lowest 5% of the distribution as a measure of signal-to-noise ratio of a profile. Skewness is a measure of asymmetry of a distribution and a positive skewness indicates that the tail on the right side is longer, implying a good signal-to-noise ratio. In almost all cases, a ChIP-seq profile has a higher skewness than its corresponding ChIP-chip profile of the same biological condition (Figure [2](#F2){ref-type="fig"} and Additional file [1](#S1){ref-type="supplementary-material"}: Table S4). We note that the difference of skewness is dependent on the IP factor which could be due to different antibody quality and prevalence of histone modification or binding events. The same conclusion can be drawn even if a different bin size was used (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S4). Our results confirmed the general observation that ChIP-seq usually produces a more distinctive signal profile than ChIP-chip. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Comparison of the characteristics of ChIP-chip and ChIP-seq profiles**. Figures (a) and (b) summarize the skewness of the signal distributions of all the ChIP-chip and ChIP-seq profiles. A ChIP profile with a good signal-to-noise ratio should have a signal distribution that is positively skewed (i.e., skewness \>0). Higher skewness implies a better signal-to-noise ratio. In almost all cases, a ChIP-seq profile has higher signal skewness than its corresponding ChIP-chip profile. Figures (c) and (d) show the ratios of the number and mean width of the enrichment regions identified by ChIP-chip and ChIP-seq using our heuristic approach (see the Methods section of this paper). In almost all cases, we can identify larger number and narrower peaks in a ChIP-seq profile than its corresponding ChIP-chip profile. :::  ::: Next, we characterized the enrichment regions within each ChIP profile. To perform a fair comparison, we would like to use an algorithm that performs peak calling on ChIP-seq and ChIP-chip data using the same criteria. Currently, many commonly used peak calling algorithms are specifically designed for analyzing ChIP-chip or ChIP-seq data, but not both. To overcome this limitation, we identified peaks from both ChIP-chip and ChIP-seq profiles using the same genome-scanning heuristic (see the Methods section). Our results indicate that we can almost always discover a larger number of peaks and narrower peaks using data generated from ChIP-seq compared to ChIP-chip when analyzing the same biological sample, and this conclusion is consistent regardless of the stringency of the identification criteria used (Figure [2](#F2){ref-type="fig"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S5). In practice, we can probably identify an even larger number of narrow peaks in ChIP-seq data if we explicitly make use of strand-specific information within the peak calling procedure (beside only shifting each read towards its 5\' end by a constant number of base pair), so the current analysis provides a lower bound on the effectiveness of ChIP-seq compared to ChIP-chip. Taken together, our results demonstrate that ChIP-seq provides higher spatial resolution and signal-to-noise ratio. Genome-wide signal reproducibility within and between technologies ------------------------------------------------------------------ Further, we estimated the reproducibility between ChIP-chip and/or ChIP-seq profiles at the genome-wide level (1 kb bins). To avoid biases due to differences in genomic coverage and sequence mapping (Figure [1e](#F1){ref-type="fig"}), we exclude genomic regions that do not contain any microarray probes and regions with unusually high variability across multiple INPUT-seq profiles. The Pearson correlation coefficient, *r*, was used as a measure of correlation, since it is more sensitive than the Spearman correlation coefficient for comparing the tail of two signal distributions, which is particularly important in analyzing ChIP enrichment signal profiles. The correlation between ChIP-chip replicate pairs and between ChIP-seq replicate pairs is generally high (median *r*= 0.85 and 0.82, respectively), indicating that both technologies can produce reproducible results. As expected, the cross-platform correlation between replicate pairs of ChIP-chip and ChIP-seq profiles are more modest (median *r*= 0.41; Additional file [1](#S1){ref-type="supplementary-material"}: Table S5). Similar conclusions can be reached even if we use different bin sizes for calculating inter-profile correlation (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S6). A representative scatter plot comparing each pair of technologies is shown in Figure [3b-d](#F3){ref-type="fig"}. We also observe a positive correlation between the skewness and inter-profile reproducibility (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S7), suggesting more sensitive antibodies may produce more consistent profiles between the two technologies. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Genome-wide reproducibility within and between ChIP-chip and ChIP-seq replicate profiles**. (a) Genome-wide reproducibility between two profiles was measured by Pearson\'s correlation coefficient, *r*, between their signal intensities (in 1 kb bins). Both ChIP-chip and ChIP-seq have high reproducibility (median *r*≈0.83), whereas the reproducibility between replicate ChIP-chip and ChIP-seq profiles is moderate (median *r*≈0.41). This figure also shows one typical example of genome-wide correlation between biological replicates of (b) ChIP-chip and ChIP-chip, (c) ChIP-seq and ChIP-seq, and (d) ChIP-chip and ChIP-seq. A locally weighted scatterplot smoothing (LOESS) regression line is also shown in each of these plots. :::  ::: Construction of average signal profile at TSS and TES ----------------------------------------------------- Construction of average ChIP signal profiles around important genomic features such as TSS and TES is a common way to visualize signal enrichment around these features. Therefore, we investigated the reproducibility of average TSS and TES profiles (2 kb up and 2 kb downstream) for every pair of replicate ChIP profiles (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S8). The average profiles of most replicate pairs are highly consistent. However, there are a few pairs that are significantly different, especially the profiles of H3K27Me3 and H3K9Me3 at both stage E-16-20 h and E-20-24 h (Additional file [2](#S2){ref-type="supplementary-material"}: Figures S8c and S8g). Without external validation, it is impossible to determine whether the average signal profiles generated by ChIP-chip or ChIP-seq are more accurate. Nonetheless, two lines of evidence led us to believe that the average signal profiles from ChIP-chip were more likely to be accurate. First, all three ChIP-chip replicates at these time points had very consistent average profiles. Second, the ChIP-seq average signal profiles at these biological conditions resembled the trend of GC content variation at TSS and TES (Figure [1c](#F1){ref-type="fig"}). The unusually low correlations between GC contents and the INPUT-seq profiles of E-16-20 h and E-20-24 h (Figure [1b](#F1){ref-type="fig"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S2b) prompted us to hypothesize that the observed discrepancy was due to a misrepresentation of GC content variation by the respective INPUT-seq profiles. Both H3K27Me3 and H3K9Me3 are repressive marks that are usually depleted at TSSs and TESs so any variation in background subtraction is likely much more pronounced than other histone marks that have strong signal enrichment at these genomic features. To test our hypothesis, we replaced the corresponding INPUT-seq background with the INPUT-seq from the AdultFemale sample, since it has the highest correlation with GC content variation. After the replacement, the average signal profiles generated by ChIP-seq and ChIP-chip at these two developmental stages agree (Figure [4](#F4){ref-type="fig"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S9). This result is striking since it shows that using different INPUT-seq as negative control of the same ChIP-seq profile can lead to substantially different interpretation of the data. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Illustration of how variability in an INPUT-seq profile can affect reconstruction of average signal profile at TSS and TES**. The top panel shows the average signal profiles at the TSS and TES for the ChIP-chip and ChIP-seq profiles of H3K27Me3 at E-16-20 h. These ChIP-chip and ChIP-seq profiles differ quite substantially, and the ChIP-seq profiles resemble that of the GC content variation (Figure 1c). We subsequently reprocessed the ChIP-seq sample by using the INPUT-seq at AdultFemale as background for normalization since this profile has a strong correlation with GC content variation, which more likely reflect the actual technology-specific biases of our sequencing platform. After this procedure, the average signal profiles of ChIP-chip and ChIP-seq look much more alike, indicating that the original INPUT-seq at E-16-20 h does not appropriately capture the technology-specific variation at these sites. :::  ::: Effect of using different input profiles in ChIP-seq data normalization ----------------------------------------------------------------------- Having observed the impact of INPUT-seq in constructing average TSS and TES profiles, we asked whether using different INPUT-seq profiles for background normalization significantly affects ChIP-seq peak calling results. We used SPP \[[@B29]\] to call peaks for 10 of our ChIP-seq samples (CBP, H3K9Ac, H3K9Me3, H3K27Ac, H3K27Me3 at E16-20 h and E20-24 h) where each ChIP profiles was normalized against four different INPUT-seq as background (the input from the matching time point, AdultFemale, AdultMale, and E-4-8 h). These INPUT-seq profiles were chosen because they have different sequencing depth and correlation with GC content (Figure [1b](#F1){ref-type="fig"}). A comparison of the number of peaks and median peak width is shown in Figure [5](#F5){ref-type="fig"}. We observed a large difference in number of peaks being called for any ChIP-seq sample when different INPUT-seq was used as background. In the extreme case (E-16-24 h, H3K9Me3 ChIP), the number of peaks can changes from zero to nearly 40,000 at a FDR of 5% (Figure [5a](#F5){ref-type="fig"}). In general, more statistically significant peaks (FDR \< 0.05) were detected when normalizing against a deeply sequenced input DNA sample (AdultMale and E-4-8 h in this experiment), although the absolute magnitude of the difference varies among ChIP datasets. The difference in peak number likely indicates a difference in detection power. For each ChIP sample, we calculated the proportion of overlap between each pair of peak sets generated by four different input DNA background (ie, six comparisons per ChIP sample). We found that the mean proportion of overlap with respect to the smaller peak set is about 95%, indicating that the differences in number of detected peak is likely due to different power to call weaker peaks. We observed that the strong peaks (ie, those with low detection FDR) were more likely detected in different peak sets (see Additional file [2](#S2){ref-type="supplementary-material"}: Figure S10 for an example). The median width of the detected peaks is also affected by using different INPUT-seq as background (Figure [5b](#F5){ref-type="fig"}). This analysis showed that the normalization using different INPUT-seq may have a significant, and underappreciated, impact on peak calling. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Effect of normalization with different INPUT-seq on ChIP-seq peak calling**. We compared the number of peaks (a) and median peak width (b) of 10 ChIP-seq samples (CBP, H3K9Ac, H3K9Me3, H3K27Ac, H3K27Me3 at E16-20 h and E20-24 h) where each of them was normalization against four different input DNA samples (the input for from the matching time point, AdultFemale, AdultMale, and E-4-8 h). Peak calling was performed with SPP using the same parameters. Clearly peak detection is significantly affected by using different input DNA library as background control. In general, more peaks are identified as statistically significant (FDR \< 0.05) when normalized with an INPUT-seq library with higher sequencing depth, although the magnitude of the differences vary across different ChIP datasets. :::  ::: Assessing variation due to the use of different peak callers ------------------------------------------------------------ Another important source of variation in analysis of ChIP-chip and ChIP-seq profiles originates from the use of different analysis algorithms. A large number of publicly available ChIP-chip and ChIP-seq analysis tools have been developed to date \[[@B23],[@B30]\], and all of them utilize different methods for tag shifting, profile normalization, smoothing, peak identification, and calculation of false discovery rate. It is therefore not too surprising to find that different peak callers can generate quite different results in terms of binding site identification, particularly when dealing with peaks with weak signals \[[@B23],[@B31]\]. Using our compendium of ChIP-chip and ChIP-seq datasets, we could assess how much variation in peak identification can be attributed to the use of different profiling technology and use of different peak callers. In this study, we analyzed our ChIP-chip profiles using two peak callers: MA2C \[[@B13]\] and Splitter \[[@B32]\] and analyzed our ChIP-seq profiles using another two peak callers: MACS \[[@B20]\] and SPP \[[@B29]\] (see Additional file [1](#S1){ref-type="supplementary-material"}: Table S8). These peak callers were chosen because they are widely used, publicly available, and generally show good performance in previous comparative studies \[[@B30],[@B31]\]. We calculated the overlap of the top 1,000 peaks of four of the factors (CBP, H3K4Me1, H3K4Me3, and H3K27Me3) across multiple developmental stages. The four IP factors were chosen as they were representative profiles containing broad peaks (CBP and H3K27Me3) and narrow peaks (H3K4Me1 and H3K4Me3). Here, we only present the results of comparing the top 1,000 peaks, since this is a biologically reasonable number of high-confidence enrichment sites in these profiles. The general conclusion of this analysis is robust against a variety of peak calling thresholds (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S11). Concordance between two peak sets was measured by the average proportion of overlapping peaks. As shown in Figure [6](#F6){ref-type="fig"}, the comparisons based on profiles of H3K4Me1 and H3K4Me3 yielded expected results, in which the intra-platform concordance is higher than cross-platform concordance (i.e., peak sets generated by two peak callers on the same profile are more concordant than peak sets generated by two peak callers on two profiles). However, the intra-platform concordance can be as low as the inter-platform concordance when analyzing the profiles of H3K27Me3 and CBP, implying that variation in peak calling algorithms can be as large as the use of different profiling technologies for some IP factors. The observation that current peak calling algorithms produce less concordant results for ChIP profiles with broad domains (CBP and H3K27Me3) than those with sharp peaks (H3K4Me1 and H3K4Me3) may suggest that they are less consistent at identifying broad enrichment regions, which may be an interesting subject for further investigation. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Variability due to peak calling algorithms**. We compared the average proportion of overlapping peaks identified by two ChIP-seq peak callers (red) and two ChIP-chip peak callers (blue) for the ChIP-seq and ChIP-chip profiles, respectively. Interestingly, the variation in peak identification concordance due to the use of different algorithms can be as large as technological differences, which is especially clear in the comparison of CBP and H3K27Me3 profiles. :::  ::: Discussion ========== ChIP-seq is an attractive alternative to ChIP-chip due to its many practical advantages. However, to date there is a lack of systematic comparison between ChIP-chip and ChIP-seq based on a large dataset from multiple IP factors. Using a compilation of replicate ChIP-chip and ChIP-seq datasets generated as part of the modENCODE project, we had an unprecedented opportunity to conduct such a systematic comparison. Through comparing the characteristics of the profiles generated by ChIP-chip and ChIP-seq, we showed that ChIP-seq generates profiles with higher signal-to-noise ratios and a larger number of more localized peaks. This is consistent with many previous observations that ChIP-seq generates profiles with higher spatial resolution and dynamic range. Not surprisingly, we found that inter-technology (i.e., ChIP-chip vs. ChIP-seq) reproducibility was lower than intra-technology reproducibility (ChIP-chip vs. ChIP-chip or ChIP-seq vs. ChIP-seq). We only had access to four ChIP-seq/ChIP-seq replicate profiles across two IP factors (CBP and CTCF) in this study, so the estimate of intra-platform reproducibility of ChIP-seq may be less accurate than that of ChIP-chip. Nonetheless, the overall magnitude of intra-platform reproducibility should still be instructive. Another important lesson was that variability due to the use of different peak callers can be as substantial as different profiling technologies. Only a small number of representative peak callers were selected in this study, since it was not our primary goal to compare performance of different algorithms. Instead, our goal was to estimate to what extent peak detection variability might be attributed to the use of different profiling technologies (microarray or sequencing) and use of different peak calling algorithms. The observation that a higher variability in peak detection is associated with broad enrichment domains is interesting, and it requires further assessment using larger numbers of peak callers. Perhaps one of the most striking findings is that there is high variability among different INPUT-seq compared to INPUT-chip. The differences between INPUT-seq profiles do not seem to correlate with the developmental stages, but rather to sequencing depth (Figure [1b](#F1){ref-type="fig"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S3). Nonetheless, sequencing depth alone does not account for all the observed variability. Samples E0, E16, E20 and AF all have low sequencing depth (\< 4 million mapped reads), yet AF and E0 have strong correlation with GC content, whereas E20 and E16 have little correlation with GC content (Figure [1b](#F1){ref-type="fig"} and Additional file [2](#S2){ref-type="supplementary-material"}: Figure S3). This suggests that variation in experimental conditions may also contribute to such variability. Our results demonstrate that it is important for the input DNA background to capture technology-specific biases, such as GC content variation, as this may have impact on the construction of average signal profiles at important genomic sites. Such average signal profiles have been used as features for building computational models of transcription factor binding or regulatory elements \[[@B33],[@B34]\], so it is important to ensure that the average signal profiles are accurate. In addition, we also show that the quality of input DNA profile used for background normalization when calling peaks for a ChIP-seq dataset is critical (Figure [5](#F5){ref-type="fig"}). The observation that more peaks were being called when a ChIP-seq dataset was normalized against a more deeply sequenced INPUT-seq suggests that increasing the sequencing depth in the INPUT-seq data may lead to higher statistical power in ChIP-seq peak detection. Currently the importance of input DNA in ChIP-seq analysis is largely underappreciated and most researchers do not even bother to check for the consistency of the input profiles. This work provides quantitative evidence that the success of a ChIP-seq analysis may depend on the quality of input library as much as the quality of ChIP library. Since using an appropriate input DNA profile as background is critical, we believe that obtaining high quality and deeply sequenced input DNA profile is necessary, and that INPUT-seq quality should be assessed more carefully during data analysis. We recognize there are several limitations in our study. First, we only compared ChIP-chip profiles generated from one commercial platform (Agilent tiling microarray) with ChIP-seq profiles generated from one sequencing platform (Illumina GAII). Although strictly speaking we can only make conclusion about these platforms, we believe that the key lessons learned in this study are instructive for analyzing data generated from other platforms. We note that the Agilent platform uses long oligonucleotides as probes (50-60 mers) and has relatively low noise level compared to other platforms \[[@B17],[@B31]\]; Illumina\'s GAII sequencing platform is currently the most widely used platform for ChIP-seq, so our analysis should be of interest to many users. Second, unlike previous comparative studies using transcription factor binding data (Additional file [1](#S1){ref-type="supplementary-material"}: Table S1) or \"spike in\" experiments \[[@B31]\], we do not have external validation of the true enrichment regions, which prohibited us from assessing detection sensitivity and specificity. Nonetheless, we believe that our conclusions, drawn from analyzing many pairs of replicate ChIP-chip/seq profiles, should be reasonably robust. Conclusions =========== Our findings highlight the differences between ChIP-chip and ChIP-seq, and show the variability that can arise from both technology and analysis methods. We demonstrate the importance of obtaining high quality and deeply sequenced input DNA libraries for ChIP-seq analysis, which has fundamental implication to experimental design and data analysis. Methods ======= ChIP-chip and ChIP-seq protocols -------------------------------- Chromatin immunoprecipitations have been performed as described previously \[[@B35]\]. Briefly, the biological material is homogenized in the presence of 1.8% formaldehyde. The cross-linked chromatin is sonicated using a Bioruptor (Diagenode) to an average size of 500 bp. Pre-cleared chromatin extract is incubated overnight at 4°C with the specific antibody and immunoprecipitated with protein-A Sepharose beads. After purification of the DNA and amplification of the libraries by linker-mediated PCR, the samples are labeled by incorporating Cy3 or Cy5 conjugated dUTPs. Labeled IP and Input samples are hybridized onto the Agilent 1X244K arrays using a TecanHS4800Pro at 67°C for 24 h. Details related to this custom microarray platform can be found in Additional file [1](#S1){ref-type="supplementary-material"}: Table S9. The non amplified ChIP samples have been used directly for sequencing by Solexa Genome Analyzer following Illumina protocols for libraries generation, cluster generation and sequencing. Data preprocessing ------------------ For each set of Agilent microarray data, the processed intensity value of the IP and input channels were extracted from the raw data file. For each lane of Illumina sequencing data, the raw sequence reads (36 bp single end) were extracted. Bowtie \[[@B36]\] was used to map both the microarray probe sequences and the Illumina\'s short reads onto the reference *D. melanogaster*genome assembly (dm3, FlyBase built 5.22). This procedure ensured that there was no systematic bias due to mapping to different version of the genome assembly. All microarray probe sequences can be uniquely mapped to the genome, and the proportion of mappable reads in our ChIP-seq data is available in Additional file [1](#S1){ref-type="supplementary-material"}: Table S2. For preprocessing of ChIP-seq data, the mapped reads of an IP library and its corresponding input DNA libraries from the same developmental stage were processed by an R package called SPP \[[@B29]\]. In particular, we used SPP to filter out uninformative reads, remove read anomaly and estimate the average fragment length by the cross-correlation profile (Additional file [1](#S1){ref-type="supplementary-material"}: Table S7). We then divided the genome into many 50 bp non-overlapping bins, and each bin *i*is associated with a genomic coordinate *x~i~*(corresponding to the center of the bin) and an intensity value *y~i~*An enrichment value in a bin is the log ratio of the smoothed signal intensity (or smoothed read count for ChIP-seq) of the IP sample over the INPUT sample. We used a Gaussian smoother with a bandwidth of 50 for signal smoothing. Specifically, the Gaussian smoother takes the form $$y_{i} = {\sum\limits_{j}{\exp\left\lbrack \frac{{(x'_{j} - x_{i})}^{2}}{50} \right\rbrack}}y'_{j}$$ where *x*\'*~j~*and *y*\'*~j~*are the center of the genomic coordinate and intensity value of probe *j*in the same chromosome. In practice due to computation constraints, we only consider the 400 closest probes from bin *i*(200 probes upstream and 200 probes downstream) for smoothing which already give a close approximation to the full Gaussian smoother. Profiles at this 50 bp resolution were used for construction of average profiles of TSS and TES. For other analyses presented in this study, we combined the enrichment value of every 20 adjacent 50 bp bins (by averaging) to obtain a ChIP profile at the 1 kb resolution. To avoid biases in estimation of genome-wide correlation between two profiles, we first excluded genomic regions that did not contain microarray probes and had high variability in INPUT-seq variability. INPUT-chip and INPUT-seq profiles were obtained in a similar manner, except we used the log~2~(intensity from input profile) as a measure of enrichment value. Unless specified otherwise, all data analysis in this study was performed using the R statistical programming environment \[[@B37]\]. All the signal density profiles were visualized using the Integrated Genome Browser \[[@B38]\]. Construction of average signal profile at TSS and TES ----------------------------------------------------- We used the gene model annotation from FlyBase \[[@B39]\] to define transcription start and end sites. We only included genes with a minimum length of 2 kb (7,231 of 15,186 genes) to exclude short genes in our analysis. The 80 bins (50 bp each) surrounding every relevant genomic feature (corresponding to 2 kb upstream and 2 kb downstream of the feature) were taken and averaged. The resulting average profiles were scaled such that the mean and the variance of signal in each profile were zero and one, respectively. Characterization of enrichment signals in ChIP-chip and ChIP-seq profiles ------------------------------------------------------------------------- We calculated skewness of a signal density profile using the following formula: $$skewness(x_{1},x_{2},...,x_{n}) = \frac{\frac{1}{n}{\sum_{i = 1}^{n}{(x_{i} - \overline{x})}^{3}}}{\left\lbrack {\frac{1}{n}{\sum_{i = 1}^{n}{(x_{i} - \overline{x})}}^{2}} \right\rbrack^{3/2}}$$ where *x~i~*is the enrichment value of bin *i*of a genome. We specifically removed bins with enrichment value at the lowest or the highest 5% of the distribution to remove potential outliers. We devised a simple heuristic method to detect \"peaks\" that can be identified by ChIP-chip and ChIP-seq profiles in a consistent manner. This method consists of two steps: (1) identify candidate enrichment regions using *ad hoc*criteria, and (2) assign a *p*-value to each candidate enrichment region. For step 1, we identify all bins with an enrichment value above $\overline{x} + ks$, where $\overline{x}$ and *s*are, respectively, the sample mean and standard deviation of a signal density per chromosome, and *k*is an arbitrary parameter. We present the results for *k*= 0 in this paper, but we have also performed analysis with *k*= -1 and *k*= 4, and the conclusion of the analyses is largely similar (Additional file [2](#S2){ref-type="supplementary-material"}: Figure S5). Adjacent bins are merged to form candidate enrichment regions. For step 2, we assign a statistical significance, *p*, to each enrichment region to be *p*= *P*(*l*\>*L*), where *l*is the sum of enrichment values of all the bins in the candidate enrichment region, $L \sim N(m\overline{x},ms^{2})$, and *m*is the number of bins in this enrichment region. To account for multiple comparison, we calculated a false discovery rate (FDR) value for each enrichment region using the method of Benjamini and Hochberg \[[@B40]\], and all the enrichment regions with an FDR less than 0.05 were considered statistically significant. We then characterize each set of enrichment regions from a profile by its number and median width of the enrichment regions. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= JWKH designed and performed the study, and wrote the manuscript. EB and PVK performed initial data analysis. NN and KPW generated the ChIP-chip and ChIP-seq datasets. PJP conceived, designed, and supervised the study. All authors read and approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Supplemental tables**. This file contains supplementary tables. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Supplemental figures**. This file contains supplementary figures. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We thank members of the Park laboratory for constructive comments and suggestions. This work is supported by grants RL1DE019021 and U01HG004258 (PJP), U01HG004264 (KPW), and a SysCODE interdisciplinary postdoctoral training fellowship through Common Fund grant 5RL9EB008539 (JWKH) from the National Institutes of Health.

PubMed Central

2024-06-05T04:04:19.609048

2011-2-28

PMC3053264

Background ========== Galactosylceramide (GalCer) is the most abundant sphingolipid of mammalian myelin \[[@B1]\]. It is synthesized in the endoplasmic reticulum by UDP-galactose:ceramide galactosyltransferase (encoded by the *Cgt*gene). In the Golgi apparatus, part of the GalCer is sulfated by cerebroside sulfotransferase (encoded by the *Gal3st1*gene), forming sulfatide \[[@B2]\] (see Figure [1](#F1){ref-type="fig"}). In the absence of a functional *Cgt*gene, compact myelin can be formed, which is, however, unstable and *Cgt*^-/-^mice develop tremors and ataxia at 3 to 4 weeks of age \[[@B3],[@B4]\]. This phenotype could be explained by disturbed axon-glial contacts at the paranodes in the CNS caused by mistargeting of essential adhesion molecules, NF-155 and Caspr \[[@B5],[@B6]\]. At least in part, these structural alterations are caused by the loss of sulfatide rather than GalCer, as demonstrated by a similar alteration of the paranodal region in *Gal3st1*-deficient mice, which lack sulfatide but have normal GalCer levels \[[@B7]\]. However, myelin appear to be more stable in *Gal3st1*-deficient mice, suggesting additional, yet less defined roles of GalCer in myelin \[[@B8]\]. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Pathway of sphingolipid biosynthesis in myelinating glia cells**. Changes in sphingolipids observed in *Cgt*^-/-^mice (blue arrows) and expected possible changes in *Fa2h*^-/-^/*Cgt*^-/-^mice (red arrows). Shown are the genetic symbols of genes encoding the respective enzymes: *Cgt*, UDP-galactose:ceramide galactosyltransferase; *CerS*, ceramide synthase; *Des1*, dihydroceramide desaturase; *Fa2h*, fatty acid 2-hydroxylase; *Gal3st1*, galactose-3-O-sulfotransferase 1; *Sgms*, sphingomyelin synthase; *Ugcg*, UDP-glucose ceramide glucosyltransferase. :::  ::: In *Cgt*^-/-^mice, the loss of GalCer and sulfatide is accompanied by a significant upregulation of 2-hydroxylated fatty acid-containing (HFA) glucosylceramide (HFA-GlcCer) and HFA-sphingomyelin \[[@B3],[@B4]\]. Interestingly, there are no indications for presence of HFA-gangliosides or other higher glycosylated HFA-sphingolipids in the brain of *Cgt*^-/-^mice \[[@B9]\]. The presence of HFA-GlcCer and HFA-sphingomyelin was interpreted as a compensatory upregulation, which may also (in part) functionally replace HFA-GalCer, enabling *Cgt*^-/-^mice to form compact myelin \[[@B3],[@B10]\]. HFA-sphingolipids in CNS and PNS myelin are synthesized from 2-hydroxylated fatty acids, formed by the fatty acid 2-hydroxylase (encoded by the *Fa2h*gene) \[[@B11]-[@B14]\]. Although loss of 2-hydroxylated sphingolipids in myelin does not affect initial myelin formation, it causes late onset (in mice older than 6 months) axon and myelin sheath degeneration \[[@B15]\]. *Fa2h*-deficiency was identified as the cause of a new leukodystrophy with spastic paraparesis \[[@B16]\], and hereditary spastic paraplegia SPG35 \[[@B17]\], respectively. In order to test the hypothesis that the presence of HFA-GlcCer and HFA-sphingomyelin in *Cgt*^-/-^mice is a functional important compensatory upregulation, preventing a more severe phenotype, we generated *Cgt*^-/-^mice with an additional deficiency in the *Fa2h*gene. Our analysis shows that the additional deletion of *Fa2h*does not obviously affect the phenotype of *Cgt^-/-^*mice. This suggests that HFA-GlcCer and HFA-sphingomyelin do not functionally compensate the loss of HFA-GalCer in *Cgt^-/-^*mice. Results ======= Generation of *Fa2h*^-/-^*/Cgt*^-/-^double deficient mice --------------------------------------------------------- In order to test the hypothesis that HFA-GlcCer partially compensates the loss of HFA-GalCer in *Cgt*^-/-^mice and thereby prevents a more severe phenotype, we generated *Cgt*^-/-^mice with an additional deficiency in *Fa2h*. As shown previously, young *Fa2h*^-/-^mice form structural and functional normal myelin \[[@B15]\] and did not show behavioral abnormalities that would indicate myelin deficiency. Older *Fa2h*^-/-^, however, developed a progressive axonal degeneration in peripheral nerves and brainstem, accompanied by myelin sheath degeneration \[[@B15]\]. As shown before \[[@B3],[@B4]\], *Cgt*^-/-^mice had a strongly reduced life span, whereas *Fa2h*^-/-^mice did not show increased mortality (data not shown). Survival of *Fa2h*^-/-^/*Cgt*^-/-^double deficient mice was not significant different from *Cgt*^-/-^mice (around 50% survival at four weeks of age). There were no obvious behavioral differences between the two genotypes. However, because of the low amount of age-/weight- and gender-matched 4-week-old *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice available, extensive behavioral testing could not be performed, and thus minor behavioral differences between *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice cannot be ruled out. The following biochemical and morphological analyses were done with mice at 4 weeks of age. GlcCer levels are reduced in CNS and PNS of *Fa2h*^-/-^/*Cgt*^-/-^mice when compared to *Cgt*^-/-^mice ------------------------------------------------------------------------------------------------------ TLC analysis of total brain lipids from wild-type, *Fa2h*^-/-^, *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice showed significant levels of HFA-GlcCer in *Cgt*^-/-^mice, in line with earlier reports \[[@B3],[@B4]\]. Unexpectedly, NFA-GlcCer levels in total brain of *Fa2h*^-/-^/*Cgt*^-/-^mice were strongly reduced compared to HFA-GlcCer levels in *Cgt*^-/-^mice (Figure [2A](#F2){ref-type="fig"}). Presence of NFA-GlcCer in *Fa2h*^-/-^/*Cgt*^-/-^mice was better visible when the amount of lipids from these mice loaded was increased 4-fold compared to controls (Figure [2B](#F2){ref-type="fig"}). A similar reduction in GlcCer levels was seen when sphingolipids of purified CNS myelin were examined (Figure [3A](#F3){ref-type="fig"}). Densitometry revealed a reduction of NFA-GlcCer by more than 80% compared to HFA-GlcCer levels in *Cgt*^-/-^mice (Figure [3C](#F3){ref-type="fig"}). Analysis of peripheral nerves (sciatic nerve) revealed a reduction of GlcCer in *Fa2h*^-/-^/*Cgt*^-/-^mice by about 60% compared to *Cgt*^-/-^mice (Figure [3B, D](#F3){ref-type="fig"}). MALDI-TOF mass spectrometry of hexosylceramides from *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice (CNS myelin) demonstrated that cerebrosides in both genotypes mainly contained very long chain fatty acyl (VLCFA; C22:0, C24:0, and C24:1) residues (data not shown). The TLC analysis also revealed a strong increase in the very-long chain fatty acid (VLCFA)-containing sphingomyelin (SM; upper band) in *Fa2h*^-/-^/*Cgt*^-/-^, but also in *Fa2h*^-/-^/*Cgt*^+/-^mice, when compared to *Cgt*^-/-^and wild-type mice (note that *Fa2h*^-/-^mice heterozygous for Cgt were used as a control in these experiments, because the number of *Fa2h*^-/-^/*Cgt*^+/+^mice were limited). Densitometry confirmed a 2-fold increase of sphingomyelin in CNS myelin of *Fa2h*^-/-^/*Cgt*^-/-^mice compared to wild-type controls (and 60% increase in *Fa2h*^-/-^/*Cgt*^+/-^mice; *Fa2h*^-/-^were not analyzed) (Figure [3E](#F3){ref-type="fig"}). The increase of sphingomyelin in the PNS was about 50% in *Fa2h*^-/-^/*Cgt*^-/-^mice (Figure [3F](#F3){ref-type="fig"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Brain lipid analysis of *Fa2h*^-/-^/*Cgt*^-/-^mice**. (A) TLC analysis of sphingolipids (after mild alkaline hydrolysis) from total brain of 4-week-old animals of the indicated genotypes showed a strong reduction of GlcCer in *Fa2h*^-/-^/*Cgt*^-/-^(lane 3) compared to *Cgt*^-/-^mice (lane 2). The amount of lipids isolated from 0.05 mg of tissue (wet weight) was loaded per lane. LPE, lyso-phosphatidylethanolamine; SM, sphingomyelin. (B) TLC analysis of total brain lipids (not subjected to alkaline hydrolysis). In order to demonstrate the presence of NFA-GlcCer in *Fa2h*^-/-^/*Cgt*^-/-^mice, the amounts of lipids was increased 4-fold in case of *Fa2h*^-/-^/*Cgt*^-/-^mice. :::  ::: ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Myelin lipid analysis of *Fa2h*^-/-^/*Cgt*^-/-^mice**. (A) TLC analysis of lipids extracted from purified myelin confirmed the low concentration of NFA-GlcCer in *Fa2h*^-/-^/*Cgt*^-/-^mice and showed that myelin from *Fa2h*^-/-^/*Cgt*^-/-^mice contained larger amounts of NFA-sphingomyelin (lane 3) compared to wild-type (lane 1) or *Cgt*^-/-^mice (lane 2). Lipid samples were subjected to alkaline hydrolysis before TLC. LPE, lyso-phosphatidylethanolamine; SM, sphingomyelin. (B) Total lipids were isolated from sciatic nerves of 4-week-old mice of the indicated genotypes and separated by HPTLC. Shown are 2 representative chromatograms (Note that lanes 5 to 8 are from the same TLC plate, however, 2 lanes containing lipids from mice with other genotypes have been removed from the Figure). PE, phosphatidylethanolamine. Hexosylceramide (HexCer) (C, D) and sphingomyelin levels (E, F) in CNS myelin (C, E) and sciatic nerves (D, F) were determined by densitometric scanning of HPTLC plates. HexCer levels were normalized to cholesterol. Data were combined from three independent experiments, and lipids from 3-6 animals per genotype were analyzed. Shown are the mean ± SD (CNS myelin: n = 3-5; sciatic nerves: n = 4-6); asterisks indicate statistically significant differences (ANOVA with post hoc Fisher\'s LSD test, \*p \< 0.05). :::  ::: In line with data published by Bosio et al. \[[@B18]\], HFA-ceramide was detectable in *Cgt*^-/-^mice but not in mice of other genotypes (Figure [4A](#F4){ref-type="fig"}). However, in contrast to the 18-day-old *Cgt*^-/-^mice analyzed by Saadat et al. \[[@B19]\], we did not observe a strong increase of ceramide in 4-week-old *Cgt*^-/-^mice compared to wild-type controls. This might be due to the different ages examined. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **TLC analysis of ceramides and gangliosides**. (A) TLC analysis of free ceramides in myelin confirmed the presence of HFA-ceramide in *Cgt*^-/-^(lane 2) and its absence in *Fa2h*^-/-^/*Cgt*^-/-^myelin (lane 3). NFA-ceramide levels were slightly increased in *Fa2h*^-/-^/*Cgt*^-/-^mice (lane 3). Cer, ceramide. (B) The ganglioside pattern of purified myelin was not altered in *Fa2h*^-/-^/*Cgt*^-/-^(lanes 1 and 2) compared to *Cgt*^-/-^mice (lanes 3 and 4). The same amount of the major myelin ganglioside GM1 was found in both mouse lines, as well as in wild-type controls (lanes 5 and 6). The larger amount of the gangliosides GD1a, GD1b, and GT1b, which are major gangliosides of neuronal membranes, in the two mouse lines with less compact myelin are most likely caused by a higher proportion of axonal membranes in the myelin preparation. :::  ::: To examine the possibility that GlcCer was replaced by gangliosides in myelin of *Fa2h*^-/-^/*Cgt*^-/-^mice, gangliosides were isolated and analyzed by TLC. Although the total amount of gangliosides in the myelin fraction of *Fa2h*^-/-^/*Cgt*^-/-^and *Cgt*^-/-^mice was increased compared to wild-type, the major myelin ganglioside GM1 was unchanged and only gangliosides normally found in neuronal membranes (GD1a, GD1b, GT1b) were increased (Figure [4B](#F4){ref-type="fig"}). Most likely these gangliosides are derived from neuronal membranes and possibly reflect a higher proportion of neuronal membrane contaminations in the *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice, with their significant reduced amounts of compact myelin. In line with this, Saadat et al. \[[@B19]\] showed a similar relative composition of major gangliosides in myelin preparation from mice deficient in *Ugcg*in myelinating cells. We therefore believe that myelin ganglioside levels in *Fa2h*^-/-^/*Cgt*^-/-^mice were not increased compared to *Cgt*^-/-^mice or wild-type controls. In summary, we conclude that HFA-GlcCer present in *Cgt*^-/-^CNS myelin was replaced by only low amounts of NFA-GlcCer but larger amounts of NFA-sphingomyelin in *Fa2h*^-/-^/*Cgt*^-/-^mice. Taken together, these results indicate that myelin hexosylceramides were mainly replaced by sphingomyelin in the absence of *Fa2h*expression. It should be noted that the increase of sphingomyelin and the decrease of hexosylceramides were more pronounced in CNS myelin than in the PNS. This might indicate differences between oligodendrocytes and Schwann cells in the trafficking or metabolism of sphingolipids. Myelination in *Fa2h*^-/-^/*Cgt*^-/-^mice as compared to *Cgt*^-/-^mice ----------------------------------------------------------------------- Myelin content in the brain of *Cgt*^-/-^, *Fa2h*^-/-^/*Cgt*^-/-^, and wild-type mice was examined by Western blot analysis of myelin basic protein (MBP) and by gravimetry of purified compact myelin. At the examined time point (4 weeks of age), MBP levels were reduced in *Cgt*^-/-^mice compared to wild-type controls (Figure [5A](#F5){ref-type="fig"}), in agreement with our previous observations \[[@B20]\]. A comparable reduction of MBP was found in *Fa2h*^-/-^/*Cgt*^-/-^mice. Accordingly, the amount of compact myelin isolated by sucrose gradient centrifugation was significant reduced in both, *Fa2h*^-/-^/*Cgt*^-/-^and *Cgt*^-/-^mice, compared to wild-type controls (Figure [5B](#F5){ref-type="fig"}), but there was no significant difference between the first two genotypes. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Reduced levels of MBP and compact myelin in *Fa2h*^-/-^/*Cgt*^-/-^and *Cgt*^-/-^mice**. (A) Western blot analysis of 4-week-old wild-type (lanes 1 and 2), *Cgt*^-/-^(lanes 3 and 4), and *Fa2h*^-/-^/*Cgt*^-/-^mice (lanes 5 and 6) brain homogenates showed reduced MBP levels in *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice. Shown is one representative out of three independent experiments, all of which gave comparable results. (B) The dry weight of purified myelin per mg brain wet weight isolated from *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice (n = 3 per genotype) was significantly reduced compared to wild-type controls (\*p \< 0.05, t-test). :::  ::: Absence of GalCer and HFA-sphingolipids does not affect stability of CHAPS-insoluble membrane fractions ------------------------------------------------------------------------------------------------------- Compact myelin of wild-type mice is stable against extraction with the detergent CHAPS at low \[[@B21]\] and high temperature \[[@B15]\]. Data by Simons et al. \[[@B21]\] suggest that loss of HFA-GalCer in *Cgt*^-/-^mice affects the association of the myelin proteolipid protein (PLP) with CHAPS-insoluble membrane fractions (CIMF), suggesting a role for (HFA-)GalCer in the formation or stabilization of CIMF. We have recently shown that the absence of HFA-sphingolipids does not affect stability of myelin CIMF \[[@B15]\]. However, in order to examine a possible synergistic effect of *Fa2h*and *Cgt*deficiency, purified myelin of *Fa2h*^-/-^/*Cgt*^-/-^mice was subjected to CHAPS extraction at 37°C. Optiprep gradient centrifugations were performed and the fractions were examined for sphingolipid content, as described \[[@B15]\]. These experiments showed that myelin was resistant to CHAPS extraction irrespective of the genotype (Figure [6](#F6){ref-type="fig"}). Therefore, we conclude that neither GalCer nor HFA-sphingolipids are essential for stabilization of CIMF. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **CHAPS-insoluble membrane fractions in *Fa2h*^-/-^/*Cgt*^-/-^myelin**. Myelin from wild-type (*Fa2h*^+/+^/*Cgt*^+/+^) (A), *Fa2h*^+/-^/*Cgt*^-/-^(B), and *Fa2h*^-/-^/*Cgt*^-/-^(C) mice isolated by sucrose density gradient centrifugation was treated with 20 mM CHAPS at 37°C for 30 min, followed by Optiprep density gradient centrifugation. Gradient fractions were collected from the top, and total lipids were extracted from all gradient fractions and separated by TLC. Most membrane lipids were present in the CHAPS-insoluble membrane fractions (CIMF), independent of the mouse genotype. Positions of lipid standards are indicated. These experiments have been done two times using two independent myelin preparations with similar results. PE, phosphatidylethanolamine; SM, sphingomyelin. :::  ::: Comparable myelination in *Fa2h*^-/-^/*Cgt*^-/-^and *Cgt*^-/-^mice ------------------------------------------------------------------ To evaluate the extent of myelin sheaths thickness, cross sections were obtained from the cervical spinal cord of 4-week-old mice. On each side of the midline, starting at the deep medial boundary of the ventral funiculus and extending ventrally and laterally, fibers of all calibers were included. In this unbiased sampling approach, the number of myelinated axons was unchanged in *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice (179 ± 19 versus 194 ± 30). Only few axons in both mouse mutants displayed no myelin at this postnatal stage (Figure [7A](#F7){ref-type="fig"}). The mean axon diameter did not differ significantly between *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice (3.88 μm versus 3.77 μm). The axons of *Fa2h*^-/-^/*Cgt*^-/-^mice recruit inappropriately thick myelin relative to their absolute calibers resulting in an average myelin thickness of 0.60 ± 0.02 μm versus 0.53 ± 0.01 μm in *Cgt*^-/-^(p \< 0.01, Chi-Square test, n = 3) (Figure [7B](#F7){ref-type="fig"}). However, due to the relative small difference in myelin thickness, the g-ratios did not differ significantly between the two genotypes (Figure [7C](#F7){ref-type="fig"}). Furthermore, electron microscopy of myelinated axons revealed normal compact myelin in *Fa2h*^-/-^/*Cgt*^-/-^mice (Figure [7D](#F7){ref-type="fig"}). These results demonstrate that a compact myelin sheath can be generated in the absence of GalCer and any 2-hydroxylated sphingolipids. ::: {#F7 .fig} Figure 7 ::: {.caption} ###### **Normal lamellar spacing in *Fa2h*^-/-^/*Cgt*^-/-^myelin**. (A) Typical cross sections of spinal cords from *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice (two different magnifications are shown). Arrows indicate axons with thin or without myelin. Asterisks indicate demyelinated axons. Scale bar, 40 μm (upper panel) 10 μm (lower panel). Myelin thickness (B) and g-ratios (C) were determined in the spinal cord of three mice per genotype. Though there was a shift towards increased myelin thickness in *Fa2h*^-/-^/*Cgt*^-/-^compared to *Cgt*^-/-^mice (Data shown are the mean SD \[n = 3 animals per genotype\]; \*\*p \< 0.01, Chi-Square test), the g-ratios of myelinated axons were similar in *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice. (D) Electron micrographs showing normal compact myelin in *Fa2h*^-/-^/*Cgt*^-/-^mice (Magnification: 16,000×). Scale bar, 50 nm. :::  ::: No obvious structural differences between oligodendrocytes of *Fa2h*^-/-^/*Cgt*^-/-^and *Cgt*^-/-^were observed by the histological or electron microscopic analyses, however, a detailed structural analysis was not performed, and therefore we cannot exclude subtle structural changes in *Fa2h*^-/-^/*Cgt*^-/-^when compared to *Cgt*^-/-^mice (e.g. in the paranodal region). Discussion ========== Although HFA-sphingolipids appear to be dispensable for the formation of compact myelin \[[@B15]\], they are essential for long-term myelin maintenance and may also play a role in glia-dependent axonal support. In the absence of GalCer in *Cgt*^-/-^mice, compact myelin can be formed \[[@B3],[@B4]\], which is, however, unstable, also suggesting a role of GalCer in myelin maintenance. We hypothesized that HFA-sphingolipids may also play a more subtle role in early postnatal development, which could be detectable on a *Cgt*^-/-^background, where compact but less stable myelin is formed. However, we did not observe signs of a more severe phenotype in *Fa2h*^-/-^/*Cgt*^-/-^mice compared to *Cgt*^-/-^mice. At the behavioral level, *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice were indistinguishable. Though there was a shift towards thicker myelin in *Fa2h*^-/-^/*Cgt*^-/-^compared to *Cgt*^-/-^mice, the g-ratios did not differ significantly between *Cgt*^-/-^and *Fa2h*^-/-^/*Cgt*^-/-^mice. Furthermore, myelin from *Fa2h*^-/-^/*Cgt*^-/-^and *Cgt*^-/-^mice was resistant towards extraction with CHAPS, as shown before for *Fa2h*^-/-^and wild-type mice \[[@B15]\]. Unexpectedly, HFA-GlcCer present in *Cgt*^-/-^mice was replaced by only low amounts of NFA-GlcCer in *Fa2h*^-/-^/*Cgt*^-/-^mice but higher levels of NFA-sphingomyelin. This suggests that the GlcCer concentration in myelin is not critical and furthermore that elevated GlcCer levels in *Cgt*^-/-^mice do not functionally compensate loss of GalCer. Accordingly, Saadat et al. \[[@B19]\] showed that deleting oligodendroglial glucosylceramide synthase (*Ugcg*) in *Cgt*^-/-^mice did not reinforce the myelin phenotype. This demonstrates that upregulation of (HFA)-GlcCer in *Cgt*^-/-^mice does not functionally compensate loss of GalCer in these mice. In *Ugcg*-deficient *Cgt*^-/-^mice, HFA-GlcCer was partially replaced by HFA-sphingomyelin \[[@B19]\]. Our results demonstrate that HFA-sphingomyelin can be replaced by NFA-sphingomyelin without any obvious effect on the phenotype of *Cgt*^-/-^mice. Thus, the significant up-regulation of HFA-sphingomyelin in *Cgt*^-/-^mice is not a functional compensation. The reason for the much lower NFA-GlcCer level in CNS myelin of *Fa2h*^-/-^/*Cgt*^-/-^mice compared to the HFA-GlcCer level in *Cgt*^-/-^mice is currently unknown. One possible explanation for the high HFA-GlcCer level and absence of HFA-gangliosides \[[@B9]\] in *Cgt*^-/-^mice could be the inability of the responsible glycosyltransferases to use HFA-GlcCer as a substrate. However, presence of a high amount of HFA-gangliosides in other tissues \[[@B22]\] and in tumor cells \[[@B23]\] argues against this. An alternative explanation could be differential sorting of HFA- and NFA-GlcCer, as shown for polarized epithelial cell lines \[[@B24]\], or reduced half-life of HFA-sphingolipids. Although structural changes at the paranodes of *Cgt*^-/-^mice are clearly caused by sulfatide rather than GalCer deficiency \[[@B7],[@B25]\], myelin appear to be much more stable in young adult *gal3st1*-deficient mice, whereas myelin maintenance is already affected in young *Cgt*^-/-^mice \[[@B8]\], indicating additional function roles of HFA- and/or NFA-GalCer. Taken together, our results and those of Saadat et al. \[[@B19]\] strongly suggest that the upregulation of HFA-GlcCer and HFA-sphingomyelin in *Cgt*^-/-^mice does not functionally compensate the loss of HFA-GalCer in these mice. This, however, does not exclude the possibility that GalCer could be functionally replaced by GlcCer, if the latter would be present at higher concentrations than in *Cgt*^-/-^mice. The HFA-GlcCer concentration in *Cgt*^-/-^mice is clearly below the concentration of HFA-GalCer in wild-type but also in heterozygous *Cgt*^+/-^mice (which show a wild-type phenotype). There might be a relatively high threshold of hexosylceramide concentration in myelin to efficiently fulfill its role in myelin maintenance. This possibility could be tested using transgenic mice overexpressing *Ugcg*under control of an oligodendrocyte specific promoter, which to our knowledge are not available yet. Alternatively, the functional role of HFA- and/or NFA-GalCer may not be taken over by GlcCer or other glycolipids. Conclusions =========== Our data indicate that compact myelin can be formed with non-hydroxylated sphingomyelin as the predominant sphingolipid, though myelin maintenance is impaired. While the specific role of GalCer in myelin maintenance remains mysterious, our results suggest that the presence of HFA-GlcCer and HFA-sphingomyelin in *Cgt*^-/-^mice does not functionally compensate the loss of HFA-GalCer. Methods ======= Mice ---- *Fa2h*^-/-^mice have been described previously \[[@B15]\]. Heterozygous *Cgt*^+/-^mice (kindly provided by Dr. Brian Popko, University of Chicago) were crossed with *Fa2h*^-/-^mice and double heterozygous *Fa2h*^+/-^/*Cgt*^+/-^mice were interbred or bred with *Fa2h*^-/-^/*Cgt*^+/-^mice to obtain mice of all possible nine genotypes. Genotyping was done using tail genomic DNA and the following oligonucleotides. *Fa2h*genotyping: 5\'-GCTCTTCTTCAAGAGCCATCC-3\', 5\'-GTGCTGTACCTCAGCTGGTC-3\'. 5\'-ATTCGCAGCGCATCGCCTTCTATC-3\', PCR products: 1045 bp for wild-type and 685 bp for the targeted allele; *Cgt*genotyping: 5\'-TTACCAAGGAGTTCAGCAAACC-3\', 5\'-CCTCTCAGAAGGCAGAGACATTG-3\', 5\'-TCTGCACGAGACTAGTGAGACG-3\', PCR products: 684 bp for wild-type and 820 bp for the targeted allele. All animal experiments followed internationally recognized guidelines and were approved by the Landesamt für Natur, Umwelt und Verbraucherschutz, Nordrhein-Westfalen, Germany. Lipid extraction and thin layer chromatography ---------------------------------------------- Total lipid extracts were prepared from brains, sciatic nerves, or purified myelin (from 4-week-old animals) according to Bligh and Dyer \[[@B26]\]. In some experiments, phosphoglycerolipids were removed by mild alkaline hydrolysis as described \[[@B27]\]. In order to isolate gangliosides from purified myelin, lipids were extracted as described by Folch et al \[[@B28]\] and the ganglioside containing upper phase was desalted by reversed-phase chromatography using RP-18 columns (Merck, Darmstadt, Germany). Lipids were separated by thin layer chromatography (TLC) in one of the following solvent systems: (1) chloroform/methanol/water (65/25/4) for hexosylceramides and sphingomyelin, (2) chloroform/methanol/acetic acid (190/9/1) for ceramides, and (3) chloroform/methanol/0.22% CaCl~2~(60/35/4) to separate gangliosides. HPTLC silica gel 60 plates (Merck) were used for all experiments. To visualize lipids, HPTLC plates were sprayed with cupric sulfate in aqueous phosphoric acid \[[@B27]\] followed by charring at 180°C for 5 min. Lipids were quantified by densitometry and the GalCer and GlcCer levels were normalized to cholesterol. Data are shown as the mean ± SD and were tested for statistically significant differences by ANOVA with post hoc Fisher\'s least significant difference (LSD) test using the program STATISTICA. Isolation of myelin ------------------- Compact myelin was isolated by sucrose gradient centrifugation as described by Norton and Poduslo \[[@B29]\], with minor modifications. Brains were homogenized in water using an Ultra-Turrax tissue homogenizer (IKA-Werke, Staufen, Germany). Aliquots of the homogenates were used for lipid extraction. The residual homogenates were adjusted to 10.5% (w/v) sucrose in 5 mM Tris-HCl (pH 7.4) and overlaid onto 10 ml of 30% (w/v) sucrose in 5 mM Tris-HCl (pH 7.4). After centrifugation (68,000×g, 50 min, 4°C) the enriched myelin fraction was recovered from the 10.5%/30% interphase, resuspended in 5 mM Tris-HCl (pH 7.4) and centrifuged at 68,000×g for 10 min. The resulting pellet was resuspended in 5 mM Tris-HCl (pH 7.4), centrifuged again, and the myelin was further purified by a second sucrose gradient centrifugation, followed by two washing steps in 5 mM Tris-HCl (pH 7.4). The final myelin pellet was resuspended in a small volume of water, lyophilized and stored at -80°C. Analysis of CHAPS insoluble membrane fractions ---------------------------------------------- CHAPS insoluble membrane fractions (CIMF) of purified myelin were prepared by optiprep density gradient centrifugation of myelin samples treated with 20 mM CHAPS at 37°C, as described previously \[[@B15],[@B21]\]. Six fractions of 350 μl each were removed from the top of the gradient and lipids were isolated from each fraction according to Bligh and Dyer \[[@B26]\], and analyzed by TLC as described above. MALDI-TOF mass spectrometry --------------------------- MALDI-TOF MS of hexosylceramides was done as described \[[@B13]\]. Western blotting ---------------- Brain samples were homogenized in 20 mM Tris-HCl (pH 8.0), 150 mM NaCl (TBS), containing 5 mM EDTA and 1 mM PMSF, using an Ultra-Turrax tissue homogenizer (IKA-Werke, Staufen, Germany). Homogenates were centrifuged at 1,000×g for 5 min and the supernatant was mixed with SDS-PAGE sample buffer containing 2-mercaptoethanol. Proteins were separated in 12.5%-polyacrylamide gels and transferred to nitrocellulose membranes by semi-dry blotting. Membranes were stained with rabbit anti-myelin basic protein (MBP; dilution 1:10,000; Millipore, Schwalbach, Germany) and mouse anti-alpha-tubulin (Developmental Studies Hybridoma Bank, University of Iowa). Bound secondary antibodies were detected by enhanced chemiluminescence as described \[[@B30]\]. Protein concentrations were determined with the Bio-Rad DC protein assay (Bio-Rad Laboratories, München, Germany) using bovine serum albumin as standard. Morphometrical analysis ----------------------- Axon caliber and myelin thickness were measured on toluidine-stained semithin sections of the cervical spinal cord of 4-week-old mice. Axonal caliber was determined by the diameter of a circle of area equivalent to each axon. The g-ratio was determined by dividing the diameter of the axon by the diameter of the fiber (axon with myelin). Quantification of myelinated axons was performed with a semi-automatic program on the basis of AnalySIS using a light microscope (100× objective, BX60, Olympus). All morphometric measurements were conducted in a blinded manner using coded sections. Ultrathin sections (50 nm) were photomicrographed with an EM10 electron microscope (Zeiss, Germany) at a magnification of 16.000×. Authors\' contributions ======================= MM carried out the animal work, performed lipid and Western blot analyses and CIMF experiments. JJ carried out the morphometrical analyses and participated in writing the draft manuscript. CG participated in the design and coordination of the study. VG participated in the design and coordination of the study. ME conceived and designed the study, wrote the draft manuscript, and participated in the lipid and Western blot analyses. All authors read and approved the final manuscript. Acknowledgements ================ This work was supported by the Deutsche Forschungsgemeinschaft through SFB645 of the University of Bonn. The authors thank Dr. Brain Popko for providing the *Cgt*^-/-^mouse line and Dr. Andreas Ratzka for help with preparing the final version of the manuscript.

PubMed Central

2024-06-05T04:04:19.612444

2011-3-2

PMC3053265

Background ========== Allergic rhinitis (AR) and chronic urticaria (CU) are common diseases that disturb sleep and reduce work/school productivity \[[@B1],[@B2]\]. Treatment guidelines recommend second-generation antihistamines, including cetirizine, desloratadine, ebastine, fexofenadine, levocetirizine, and loratadine, as a first-line treatment for AR \[[@B1]\] and CU \[[@B2]\]. Although data on comparative efficacy of second-generation antihistamines are limited, clinical studies demonstrate that patients with AR or CU who fail to respond to one antihistamine may benefit from a switch to another \[[@B3]-[@B6]\]. In 2006, based on a report \[[@B7]\] that drew no conclusions regarding efficacy or safety differences among the 6 antihistamines mentioned above, the Norwegian Medicines Agency mandated, as a cost-cutting measure and requirement for continued governmental reimbursement, that health care providers substitute generic cetirizine or loratadine for desloratadine, ebastine, or fexofenadine in their AR and CU patients, irrespective of treatment success or patient satisfaction with their current regimen \[[@B8]\]. Only those patients who failed treatment with both cetirizine and loratadine could switch back to the newer agents \[[@B8]\]. An online survey conducted in 2007 in Norwegian general practice settings evaluated patient experiences regarding efficacy, fatigue, and impact on work productivity and attendance after a mandatory switch from desloratadine, ebastine, or fexofenadine to cetirizine or loratadine. Methods ======= Between January and April 2007, general practice centers across Norway each invited up to 15 patients (minimum 1500 in total) to participate in an online survey. Patients were identified through a data software program (eZearch; Emetra AS, Bergen, Norway) run on the physicians\' computers. If the practice had more than 15 eligible patients, the total number was randomly reduced to a maximum of 15. These patients were sent a letter, signed by their physician, inviting them to answer an Internet-based questionnaire. The letter contained a unique username and password that permitted the patients to answer the questionnaire only once. Internet service was provided by Questback AS (Oslo, Norway), an independent IT services group. Participants aged 18 to 65 years treated with desloratadine, ebastine, or fexofenadine, and switched to cetirizine or loratadine after May 1, 2006, completed the 25-question survey on demographics, antihistamine therapy, physician visits, and treatment cost. Anonymity was maintained by Questback AS using a method approved by the Norwegian Data Inspectorate. Participants rated post-switch efficacy, fatigue, and effect on work productivity and attendance compared with their pre-switch medication using the following descriptors: \"much less,\" \"slightly less,\" \"similar,\" \"slightly more,\" or \"much more.\" Instances of no response were included in \"similar.\" Patients also reported post-switch change in number of doctor visits required (fewer, same, more), total treatment cost (lower, no change, higher), and whether they had switched back or wanted to switch back to their previous medications. Descriptive analyses were performed on main findings, and responses by patients are given as proportion of patients (%) in Table [1](#T1){ref-type="table"}. Subgroup analyses were conducted on diagnosis, pre-switch antihistamines, and post-switch therapy. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Baseline demographics and clinical characteristics (N = 409) ::: Sex, n (%) ------------------------ -------------  Female 271 (66.3%) Age (y), n  18-25 35  26-35 67  36-45 101  46-55 122  56-65 84 Diagnosis, %  Pollen allergy 63.7  Other nasal allergy 15.2  Skin allergy/eczema 6.9  Other 14.3 Pre-switch therapy, %  Desloratadine 78.4  Ebastine 16.0  Fexofenadine 5.6 Post-switch therapy, %  Cetirizine 64.7  Loratadine 35.3 ::: Results ======= A total of 343 physicians invited 1920 patients to participate. Of 493 patients (25.7%) responding to the invitation, 421 were eligible (ie, patients stating a switch of medication as a result of the new rules for reimbursement). Twelve patients who failed to specify pre-switch or post-switch antihistamine were excluded. The number of participants included in the final assessment totaled 409 (83.0% of responding patients; Table [1](#T1){ref-type="table"}). Pollen allergy was present in 63.7% of patients. Respondents also reported other nasal allergy (15.2%) and skin allergy (6.9%); other allergies comprised the remaining 14.3%. Most patients (78.4%) were taking desloratadine prior to the mandatory switch, followed by ebastine (16.0%) and fexofenadine (5.6%); post-switch, 64.7% of patients received cetirizine and 35.3% loratadine. Respondents (46.3%) rated cetirizine or loratadine as less efficacious than desloratadine, ebastine, or fexofenadine; 21.1% of respondents reported much less efficacy than with their previous treatment; and 25.2% reported slightly less efficacy (Figure [1](#F1){ref-type="fig"}). In total, 6.9% of respondents rated cetirizine/loratadine more effective than previous treatment (5.0%, slightly more effective; 1.9%, much more effective). Further, 28.7% of respondents reported increased fatigue, while 6.4% were less tired. Another 63.9% reported fatigue levels that were similar between previous and post-switch treatment. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Post-switch efficacy of cetirizine/loratadine therapy compared with efficacy of desloratadine, ebastine, or fexofenadine pre-switch**. :::  ::: In reporting impact on work productivity, 31.6% of respondents reported decreased work capacity; 63.1% said post-switch work productivity was similar to that of previous treatment, while 3.3% reported greater work productivity after switch. Post-switch, 6.0% reported reduced work attendance, and 2.0% stated that work attendance had increased. Respondents reported increased contact with their physician post-switch compared with pre-switch: 13% visited their physician more often, compared with 4% who reported fewer visits. Total costs post-switch increased for 16% of respondents and decreased for 10%. Due to limited data, the actual number of consultations and total costs were not possible to measure. Subgroup analyses stratified by diagnosis, pre-switch antihistamine, or post-switch antihistamine found no significant difference in ratings for efficacy, fatigue, or work productivity and attendance irrespective of type of allergy. This confirmed that the above-mentioned factors had no impact on the outcomes assessed. At the time of the survey, 26% of all participants had already switched back to their pre-switch antihistamine, and 25% reported dissatisfaction with cetirizine or loratadine and wanted to switch back to their original agent; 38% remained on their post-switch agent (Figure [2](#F2){ref-type="fig"}). Of 191 participants who rated cetirizine/loratadine less effective than desloratadine, ebastine, or fexofenadine, 37% had switched back to their pre-switch antihistamine; an additional 42% reported they were dissatisfied with the mandatory switch in agents and were interested in switching back to their original prescription; 11% remained on their post-switch agent. Further, 63% of respondents reporting less efficacy, more somnolence, or less productivity at work (n = 267) with cetirizine or loratadine were already receiving their previous treatment or wanted to switch back. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### Responses from patients reporting less efficacy post-switch (n = 191) :::  ::: Discussion ========== This online survey provided the first patient assessment of the impact of a mandatory switch from one antihistamine to another. Although the Norwegian Health Agency claimed that all second-generation antihistamines were equally efficacious and safe \[[@B8]\], 46.3% of respondents perceived reduced efficacy with cetirizine/loratadine compared with their previous drug, 28.7% reported increased fatigue, and 31.6% experienced decreased work productivity. Patients with AR and CU report varying degrees of symptom relief with different second-generation antihistamines, raising the issue of whether a mandated switch in treatment is ethically acceptable in patients well controlled by or satisfied with their treatment. Clinical management guidelines for urticaria recommend changing to another second-generation antihistamine in patients whose symptoms do not respond to first-choice treatment at standard or higher-than-indicated doses after a defined period \[[@B2]\]. Moreover, clinical studies and patient surveys indicate that switching patients dissatisfied with their current treatment to another second-generation antihistamine may provide symptom relief. In a post hoc analysis of data from 4 post-marketing studies in patients with AR or CU who had failed previous therapy with cetirizine, fexofenadine, or loratadine, 90.3% of patients rated efficacy with desloratadine as excellent or good \[[@B3]\]. In another study, patient satisfaction with CU treatment rose from 74.3% to 83.7% after their antihistamine therapy was modified \[[@B4]\]. A cross-sectional survey of AR patients dissatisfied with loratadine reported equal or better satisfaction after a switch to desloratadine or fexofenadine \[[@B5]\]. Those patients with severe symptoms inadequately controlled by loratadine also reported greater satisfaction with desloratadine compared with fexofenadine. Finally, a randomized, multicenter study found that more patients with AR reported moderate, marked, or complete relief of symptoms after switching to loratadine subsequent to fexofenadine failure than after switching to fexofenadine following loratadine failure \[[@B6]\]. The Norwegian Health Agency conducted no patient-impact assessment or health-economic analysis prior to mandating use of generic antihistamines to maintain governmental reimbursement. Instead, it based its decision solely on the cost of the 6 antihistamines. Post-switch, respondents reported a trend toward more consultations with their physicians, increased total costs for medication, and decreased work productivity and attendance, suggesting that the mandate may not be as cost-effective as assumed. Some limitations of this survey should be noted. There was no control group; responses were sought only from patients for whom a switch from desloratadine, ebastine, or fexofenadine to loratadine or cetirizine was mandated. In addition, differences in efficacy and tolerability among agents can only be determined through head-to-head, controlled clinical trials. The results of this survey cannot be extrapolated to other groups of allergy patients, and any drug comparisons should be made with great caution. Conclusions =========== A mandatory switch to the generic second-generation antihistamines cetirizine or loratadine to continue reimbursement in patients well controlled on desloratadine, ebastine, or fexofenadine decreased symptom control and work productivity and attendance. Taken together, the results from this survey and from clinical trials and other surveys indicate that treatment decisions should be made only after a thorough patient evaluation and then individualized for each patient. Further, periodic follow-up should be made to assess patient response post-switch. Competing interests =================== Medical writing and editorial assistance was provided by Carol Sibley and Patricia C. Abramo of AdelphiEden Health Communications, New York, New York. This assistance was funded by Merck, Sharpe and Dohme & Co. and MSD Norway AS. Authors\' contributions ======================= HC, TO and EH conceived the survey and participated in its design and coordination; they also participated in the preparation of the manuscript. FT participated in the design and coordination and had patients who answered the survey. RM performed the statistical analysis and participated in the preparation of the manuscript. All authors read and approved the final manuscript.

PubMed Central

2024-06-05T04:04:19.614623

2011-2-28

PMC3053266

Background ========== Knowledge translation, considered as both strategies and processes, has received worldwide attention in the application of health research for decision making in many universities in recent years \[[@B1],[@B2]\]. Some universities and organizations see the process as a priority \[[@B3]\]. The necessity for the optimal utilization of research has become even more prominent in low- and middle-income countries which suffer resource shortage \[[@B4]\]. Knowledge translation, on the other hand, is a complex, non-linear, ongoing and dynamic process which cannot be achieved easily \[[@B4]\]. Knowledge translation theories have developed considerably in recent decades, and various models have been put forward to explain them \[[@B5],[@B6]\]. Even though some think there is basically no need for such theories \[[@B5],[@B7]\] many researchers believe a model-based framework, adopted at different levels ranging from individual to group to organizational behaviour, is required to design and implement effective and assessable interventions for turning research results into decision making \[[@B8],[@B9]\].\[[@B5]\]. Various knowledge translation models have been designed in different fields of study (i.e., social sciences, nursing, and health services) and for a variety of organizations, each one of which emphasizes one aspect of knowledge translation. However, there is no extensive model which can be used in all situations. Basically, such a model cannot exist, because Knowledge translation strategies must be attuned to the groups\' and organizations\' needs and structures \[[@B7]\]. Only a few models have been designed for examining the performance of research organizations (e.g., universities) \[[@B5],[@B6],[@B10]\]. Therefore, the knowledge translation model was designed at Tehran University of Medical Sciences (TUMS) as a framework to identify knowledge translation capacities and weaknesses in the university and to suggest appropriate and necessary interventions, a detailed report of which has already been published elsewhere \[[@B11]\]. Since one way of applying models is to design and develop appropriate tools on the basis of an available model, a tool was designed to assess knowledge translation in research institutes using the \'Knowledge Translation Model of Tehran University of Medical Sciences\' \[[@B12]\]. Such a move can not only transform a theoretic model into applicable knowledge but also helps identify strengths and weaknesses of knowledge translation in universities, as knowledge-producing organizations, and subsequently promote the utilization of knowledge in the country. In Iran Knowledge translation has two aspects. Firstly, the Iranian healthcare system, is an integrated structure formed following the integration of the Ministry of Health into medical education in 1985 \[[@B13]\]. The system has potentially prepared the ground for health knowledge translation \[[@B14]\]. Secondly, the production of science has followed a rapid pace in recent years, pointing out the fact that the translation of this rapidly-growing knowledge needs special attention \[[@B14]\]\[[@B15]\]. Apart from designing a self-assessment tool for knowledge translation activities in research-producing institutes, TUMS research centres and faculties, this pilot study uses the tool to assess the weaknesses and strengths of knowledge translation in the university. Methods ======= Tool Development ---------------- The TUMS model provided a framework for the development of a self-assessment tool for knowledge translation \[[@B11]\]. All the determinant factors of knowledge translation in the model were summarized in homogenous domains, and a preliminary draft was prepared. To ensure internal validity, five experts were asked to give their impression on each statement by \'thinking aloud\', and necessary changes were subsequently made if the items did not convey the meanings we had in mind. Then, the tool was discussed for the sequence of questions and content validity among 23 researchers from five universities around the country and final modifications were implemented on the basis of their feedback. The finalized version of the tool consists of 50 statements in four main domains (Additional file [1](#S1){ref-type="supplementary-material"} and [2](#S2){ref-type="supplementary-material"}). Each item covers at least one of the determinant factors of knowledge translation, needing to be evaluated. The four main domains and their sub-domains are as follows: 1- The question of research: Do we identify decision makers\' research needs and convert them into research questions? This covers two sub-domains of resources (four statements) and strategies (eight statements) 2- Knowledge production: Do we produce useful evidence for decision making? (nine statements) 3- Knowledge transfer: Do we have appropriate means for disseminating the organization\'s research results to their target audiences? This covers two sub-domains of resources (nine statements) and strategies (16 statements) 4- Promoting the use of evidence: Do we help decision makers utilize research results better? (four statements) By \'resources\' we mean all the financial, equipment, legal and human resources that have been provided by the organization for knowledge translation activities. And by \'strategies\' we mean all the steps taken in the concerned domains. The tool permits the self-assessment of the organization through group discussions and consensus by its members, rather than by survey method. To assess an institute, the research authorities and its selected researchers should complete the tool. It would be preferred if other stakeholders of the researches carried out in the organization also participate in the meeting. First, the participants list the organization\'s research target audiences. Then, they review each statement and rate each item on a Likert scale upon discussion, exchange of ideas and overall consensus. Each statement will assess at least one of the aspects influencing knowledge translation. And each item will secure a score that would cover a range of five options, ranging from \'the situation is good and needs no intervention\' to \'the situation is quite unfavourable and/or there is a dire need for intervention\'. The results of the assessment will not be considered as the overall *score*obtained by the organization regarding knowledge translation activities. Instead, they will be used to identify the strengths and weaknesses of the organization for future interventions. This is why the tool is considered as a guide and not a questionnaire for cross-sectional study purposes. Reliability Assessment ---------------------- To study the reliability of the tool using intra class correlation, 21 researchers in three research centres answered the self-assessment tool twice, at two-week intervals. Internal consistency with Cronbach\'s alpha was estimated by having 45 researchers from five research centres (24 persons in addition to the first 21 researchers) complete the tool. Items having lower reliability were then revised. TUMS Pilot Testing ------------------ TUMS is the oldest and largest centre for health sciences in Iran, consisting of seven faculties, over 1,250 faculty members, 47 research centres and 16 teaching hospitals. Also, it holds the greatest share of medical publications in Iran \[[@B15]\]. Among the 47 TUMS research centres and departments, 12 research centres and eight departments were chosen to participate. One department was chosen from each faculty; in the medical faculty, a clinical and a basic science department were selected. Selection criteria included (a) Holding regular research council meetings (b) Willingness to participate in the study. The selected research groups and centres were invited to participate in the study, and were asked to introduce one of their research council members as a focal point as a contact person. The three-hour briefing session was held to familiarize the focal points with the tool, and as a drill they completed it on their own. The questionnaire was thereafter completed by the members of the research council and the researchers from different centres and departments. An average of six individuals participated in these sessions. In order to assure the quality of the study, the focal points were asked to arrange the research council meeting in each centre and inform the research team of its time so they would be available to answer possible questions regarding the tool. The research team answered the questions on phone in order to avoid any possible information bias caused by their presence in the meeting. Data analysis ------------- To assess each statement and domain, the mean score of each statement was calculated. The option \'the situation is quite unfavourable and/or there is a dire need for intervention\' scored 1 and \'the situation is good and needs no intervention\' scored 5 points. The statements which obtained the highest and lowest score were identified. To determine the lowest and highest scores, the quartiles were used. To observe ethical considerations, the participants were told that writing down their names or the name of their centre was not necessary. The research was approved by the Ethical Board Committee of TUMS. Results ======= Tool reliability ---------------- In addition to the domains, the intra-class correlation coefficient (ICC) was calculated for each of the statements. The three domains of \'the question of research, knowledge production and knowledge transfer\' had an ICC and Cronbach\'s alpha higher than 0.70, whereas the figure was lower than 0.70 for \'promoting the use of evidence\' (Table [1](#T1){ref-type="table"}). A low Cronbach\'s alpha was likely considering the small number of questions and the nature of the statements in the fourth domain. A small ICC, however, showed the low reliability of the statements in the very domain. Considering the importance of these statements, they were kept in the tool in spite of their low reliability. For better clarification of the statements, we specified them with an asterisk in the guide, and asked the focal points to give further explanations to the participants. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### The tool\'s reliability indicators ::: Domain Cronbach\'s alpha Intra class correlation ------------------------------- ------------------- ------------------------- The research question 0.79 0.94 Knowledge production 0.70 0.87 Knowledge transfer 0.86 0.90 Promoting the use of evidence 0.27 0.48 ::: Assessment of research centres and departments ---------------------------------------------- All the selected centres (12 research centres and eight departments) completed and delivered the self-assessment questionnaire. The mean score of each statement was calculated for the centres and departments. The minimum and maximum mean scores obtained in various statements were 1.15 and 3.95 respectively. Tables [2](#T2){ref-type="table"}, [3](#T3){ref-type="table"}, [4](#T4){ref-type="table"} and [5](#T5){ref-type="table"} represent the mean scores and standard deviations of each of the statements in the four domains. The statements which gained the highest and lowest scores respectively have been specified in these tables. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### The mean score and standard deviation for each statement in \'question of research\' ::: Statement (resources) Mean (SD) N = 20\* ---- --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------------- 1 In our organization there is a comprehensive list of organizations that can use our research results. 2.00 (1.21) 2 The particulars of each unit\'s researchers and their capabilities are made available to other organizations through a databank. 2.15 (1.23) 3 A website and/or data bank is available in our organization for notifying the research priorities of other organizations 2.21 (1.23) 4 Compared to the organization\'s internal budget for research, the amount of external funding is such that researchers are encouraged to use external funding. 2.00 (1.37) **Statement (strategies)** 5 Regular meetings are held for the exchange and identification of research priorities of individuals and/or research-using organizations. 2.10 (1.21) 6 Individuals and decision-maker organizations know which fields our organizations\' research capacities cover. 2.50 (0.95) 7 For preparing grounds for performing related research and strengthening research utilization, our organization holds regular and purposeful meetings with decision-makers (managers and policy makers) for extending cooperation and using mutual capacities (establishment of knowledge network) 1.89 (1.10) 8 Our organizations\' research priorities are determined through meetings with executive organizations\' representatives and/or users of research results (like community representatives, patients etc) 1.50 (0.83) ↓ 9 Our organizations\' research priorities are compiled and its up-to-date list is available to the organizations\' researchers. 3.15 (1.23) ↑ 10 Compared to the internal process, the external grant securing process is such that researchers are encouraged to use external funding. (the extra-organizational part of the process) 1.95 (1.39) 11 In case of external funding, researchers can use these for research matters easily and in a short period of time. (the intra-organizational part of the process) 3.18 (1.33) ↑ 12 Incentives exist for our researchers for securing external funding. 2.50 (1.10) ↓ The statements which obtained the lowest scores ↑ The statements which obtained the highest scores \*Each of the twenty centres and departments that participated in the study agreed upon a single score which was used in calculating the mean score. ::: ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### The mean score and standard deviation for each statement in \'knowledge production\' ::: Statement Mean (SD) N = 20 --- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ------------------ 1 Researches that result in production of \'actionable messages\' with a high level of evidence (such as regular systematic reviews and/or clinical guideline development activities) are considered priorities of research and granted funds. 2.45 (1.43) 2 The groups which will use the results of research participate in its conduction and/or design. 2.15 (0.90) 3 Our impression is that the users of research results trust the quality of the researches done in the organization. 3.95 (0.83) ↑ 4 Quality assurance program is required for each research (data gathering protocol and/or training the research workers) 3.20 (0.89) ↑ 5 Quality control is carried out while research is being conducted (internal monitoring of the executive program by the research group and/or external supervision) 2.95 (1.10) ↑ 6 The gap between \'presentation of the research proposal\' and \'beginning of the research\' is reasonable (the process of reviewing the research proposal) 2.85 (1.27) ↑ 7 While designing the research proposal and performing the projects researchers are aware that applied projects should reach results in good time (the projects duration and absence of delay in performing them) 3.75 (0.91) ↑ 8 The gap between \'end of research\' and \'finalization of results in the form of a report\' is reasonable (the process of presentation of research results) 3.60 (0.82) ↑ 9 In research project proposals (projects whose users are service providers, managers, policy makers, patient groups and/or people) budget is considered for disseminating the results (other than being published in peer-review journals and/or attending conferences) 1.39 (0.98) ↓ ↓ The statements which obtained the lowest scores ↑ The statements which obtained the highest scores ::: ::: {#T4 .table-wrap} Table 4 ::: {.caption} ###### The mean score and standard deviation for each statement in \'knowledge transfer\' ::: Statement (resources) Mean (SD) N = 20 ---- ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ------------------ 1 Researchers are familiar with the topic of knowledge translation and how to perform it. 2.63 (1.21) ↑ 2 Our researchers have communication skills for knowledge transfer. 2.95 (1.36)↑ 3 Our researchers can use the services of those familiar with knowledge transfer skills (the presence of individuals in our organization who work with this objective; and/or make contracts with individuals and institutions outside our organization) 2.00 (1.25) 4 Our researchers have the necessary financial resources for preparing content appropriate to the target audience. 1.39 (0.92)↓ 5 Our researchers have the necessary equipment for preparing content appropriate to the target audience. 2.16 (1.21) 6 The necessary structure (like office and/or organizational unit) and/or manpower is available for strengthening knowledge transfer in our organization, considering the produced amount of research-based knowledge transferable to the decision makers 1.45 (0.76)↓ 7 The framework of research projects\' final reports is such that decision makers can easily point out the actionable message. 2.50 (1.05) 8 Intellectual property rights exist which support researchers who help disseminate research results prior to their publication in journals. 1.15 (0.37)↓ 9 There are criteria for evaluation of researchers\' knowledge transfer activities in our organization. 1.55 (0.83) ↓ **Statement (strategies)** 10 In our organization there is a process that determines which research results can be transferred (keeping in mind the fact that not every research result is transferable) to the target audiences (apart from other researchers and funders) 1.42 (0.69)↓ 11 In our organization, all research results are peer reviewed prior to knowledge dissemination or transfer. 3.70 (1.17)↑ 12 Our researchers convert their research results into actionable messages appropriate to the target audience. 2.35 (1.35) 13 Our researchers have adequate time for preparing content appropriate to the target audience. 2.16 (0.90) 14 Our researchers have the necessary incentives for performing knowledge transfer (rewards, appropriate promotion rules) 2.42 (1.17) 15 Knowledge transfer and utilization of research results exist in the general program of research methodology training 1.67 (0.84) 16 A list of all the (research result users) is prepared for each research project. 2.35 (1.14) 17 Our organizations\' research managers are aware of the researchers needs (separately for each study field-group etc) in the field of knowledge transfer, and perform proper interventions for them. 1.50 (0.71)↓ 18 The format of peer review journals which publish research results is such that the decision makers are easily informed of the actionable message when necessary. 2.45 (1.19) 19 The gap between sending the article and its publication in journals is such that the interventions that result from research can be implemented in reasonable time (considering the need for prompt availability of research results to decision makers). 2.35 (1.18) 20 Researchers can provide the results of their research through the web and/or electronic banks. 1.95 (1.23) 21 Meetings are held for presentation of research results to decision makers. 1.47 (1.02)↓ 22 Our organization has regular communications with public and private media and target audiences (like publications related to women and youth) for transfer of research-based evidence. 1.55 (1.00)↓ 23 Evidence-based decision making (based on domestic and/or foreign research) is among the subjects of research in our organization. 1.80 (1.24) 24 Our researchers study the extent to which decision makers utilize our organizations\' research results. 1.55 (0.89)↓ 25 Our researchers identify the potential barriers of behavioral change in decision makers for utilizing their research results. 1.60 (0.94) ↓ The statements which obtained the lowest scores ↑ The statements which obtained the highest scores ::: ::: {#T5 .table-wrap} Table 5 ::: {.caption} ###### The mean score and standard deviation for each statement in \'promoting the use of evidence\' ::: Statement Mean (SD) N = 20 --- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ ------------------ 1 We conduct education programs such as \'evidence-based medicine\' or \'evidence-based decision making\' for service providers and/or managers 1.80 (1.15) 2 Systematic reviews and clinical guidelines\...etc that strengthen evidence-based decision making are produced in our organization. 1.95 (1.28) 3 Our researchers play an active role in technical committees that help in decision making (executive organizations\' decision making, hospital management and also groups supporting the health of patients and people) 2.53 (1.39)↑ 4 We send decision makers reminders to follow the research results that we\'ve previously sent them 1.30 (0.73) ↓ ↓ The statements which obtained the lowest scores ↑ The statements which obtained the highest scores ::: The statements which obtained the lowest scores represented issues that were believed to be in a poorer condition and needed interventions. The contrary was true for the statements which obtained the highest scores. The items with a mean score lower than 1.6 (the first quartile) included setting research priorities with research users, financial resources, infrastructures, researchers\' intellectual rights, researchers\' need assessment and evaluation of their knowledge translation activities, direct communication with media and research users and follow-up of research utilization. The following statements reflect weaknesses in the university due to their low scores (less than 1^st^quartile): determining research priorities through meetings with stakeholders (statement 8 from 1^st^domain), considering budgets in proposals and securing funds for knowledge translation (statement 9 from 2^nd^domain and statement 4 from 3^rd^domain), communication between researchers and research target audiences and follow-up of utilization of research results (statements 21, 22 and 24 from 3^rd^domain and statement 4 from 4^th^domain), infrastructure (statement 6 from 3^rd^domain) and supportive regulations and measures for knowledge translation activities (statements 8, 9,10 and 17 from 3^rd^domain). On the other hand, dissemination of research priorities (statement 9 from 1^st^domain), facilitating the receipt of grants from other organizations (statement 11 from 1^st^domain), quality of research (statements 3, 4 and 5 from 3^rd^domain), timeliness of granting, conducting and providing research results (statements 6,7 and 8 from 1^st^domain) and being acquainted with knowledge translation and having communication skills (statements 1 and 2 from 3^rd^domain) were in the 4^th^quartile of the scores and reflect strengths of the university. Statements which obtained the lowest scores pertained to knowledge transfer strategies (5 statements), knowledge transfer resources (4 statements), strategies for developing research questions (1 statement), knowledge production (1 statement) and promoting the use of evidence (1 statement). On the other hand, statements which obtained the highest scores (greater than the third quartile) respectively, were, from knowledge production (6 statements) strategies for developing research questions (2 statements), knowledge transfer resources (2 statements) and strategies (1 statement), and promoting the use of evidence (1 statement). The mean domain score was also calculated. Knowledge production obtained the highest score with a mean score of 2.92 ± 0.83, whereas promoting the use of evidence obtained the lowest (1.89 ± 0.55). The mean score of the research question domain was 2.26 ± 0.50, and the knowledge transfer domain obtained a mean score as low as 2.00 ± 0.59. Discussion ========== The SATORI tool (SATORI: a Japanese Buddhist term for enlightenment, literally meaning \'understanding\') provides a way to operationalize the TUMS knowledge translation model. This tool consists of 50 statements about requisites, resources and strategies for facilitating knowledge translation in research institutes. Use of the tool enables research managers and researchers to identify strengths and weaknesses of knowledge translation within their institution and to subsequently develop interventions that could improve their organization\'s KT infrastructure and capacity. This tool was developed to assess knowledge translation activities from the \"push side\" perspective, meaning activities which are undertaken by researchers or research organizations to transfer research results to target audiences. Research managers and researchers can identify the strengths and weaknesses of their organization regarding knowledge translation upon using this 50-statement tool, and work toward identifying solutions for the improvement of the organization\'s infrastructure and capacity. Actually, this tool is a complement to the \"Is research working for you?\" which is a self-assessment tool and discussion guide designed by the Canadian Health System Research Foundation (CHSRF) to examine \"pull side\" activities; activities performed by health services management and policy organizations to benefit from research evidences \[[@B16]-[@B19]\]. In many countries, evaluation of research outputs in academic units is used as a method of allocating funds, it is also used as a management tool to monitor the performed activities \[[@B20]\]. Nonetheless, methodologies for assessing the KT capacity of research organizations are still in their infancy and until the execution of this study the authors had not come across a tool that could evaluate the capacities of research organizations (university, faculty, public and private research centres and groups), nor the obstacles faced in knowledge translation. To our knowledge, existing questionnaires for assessing knowledge translation activities are completed by health researchers individually \[[@B21],[@B22]\]. A framework has, however, been proposed by Lavis to evaluate \'linking research to action\' measures from \'Push, Pull, and Exchange\' aspects at national levels \[[@B19]\]. A questionnaire has also been designed by Tugwell et al to assess the capacity of low and middle-income countries for performing equity-oriented research at national levels \[[@B23]\]. Other studies conducted in this regard have mostly qualitatively assessed the quality of knowledge translation at organizational levels \[[@B4]\]. The comparison of organizational strengths and weaknesses, based on the scores gained through this questionnaire can help officials define intervention priorities. Prioritization, however, could not be done solely by comparing the statements\' raw scores, since all these statements have equal weights but different significance and generalization values. For example, sending a reminder regarding the research results to decision makers is a simple process in need of few resources. Although the establishment of an organizational unit for improving knowledge translation is a costly strategy; it may be beneficial for developing strategies to promote KT, such as designing a guideline for publishing the results, and the development of regulations to support the intellectual rights of the researchers. The following steps can help prioritize KTE interventions: (1) preparing a list of the statements which have obtained low scores on the self-assessment test, (2) formulating intervention options, and (3) assessing the organizational context and different aspects such as feasibility, cost and chance of success for each intervention. The pilot study conducted across 20 TUMS centres and departments showed that while conducting research and producing knowledge are performed with appropriate quality and timeliness in the university, there are significant weaknesses in the interactions between researchers and their research target audiences (both in selecting research priorities *and*transferring the results), securing the financial resources and following supportive regulations for knowledge translation activities. The comparison of the mean scores obtained in different domains confirmed that research capacity and knowledge production is acceptable but there are certain weaknesses in the aptitude of the last two domains of knowledge transfer and promoting the use of evidence. In view of the weaknesses identified in our study, the main interventions needed for TUMS include: 1\) Facilitating knowledge translation activities through provision of financial resources required for these activities; 2\) Facilitating knowledge translation activities through creating supportive and facilitating infrastructures for these activities; 3\) Facilitating interactions between researchers and target audiences to exchange questions and research findings. The challenges faced by TUMS in knowledge translation issues are somewhat similar to those faced by other countries \[[@B24]-[@B26]\]. Though many investors in developed countries financially support knowledge translation activities \[[@B27]\], and have established certain structures and regulations to strengthen knowledge translation, the interaction between researchers and target audiences is still a major concern. Most of the interventions proposed to strengthen knowledge translation, even in developed countries, focus on facilitating and enhancing the interaction between researchers and research users \[[@B28]-[@B30]\]. The aim of this study was to develop a tool to identify the obstacles (weaknesses) in knowledge translation in TUMS-affiliated faculties and research centres. It seems the knowledge translation-related issues are somewhat similar in research organizations located in developed and developing countries. As a result, the standardized-version of the tool can be used in other research organizations. The SATORI has some particular aspects. First of all, this tool has been prepared on the basis of the \"Knowledge Translation Model at Tehran University of Medical Sciences\" on the grounds of domestic studies and a comprehensive literature review on knowledge translation barriers. And in addition to being used in Iran\'s academic environment, we believe it could be applied to other countries\' research organization settings as well. This tool covers the most important activities, necessary resources and facilitating strategies for knowledge translation at the organizational level, and similar to its Canadian counterpart, this tool can also be used for re-evaluation of the organization\'s promotion in knowledge translation \[[@B17]\]. In addition, discussion and dialogue is possible concerning each statement of the tool that could lead to identification of intervention(s) regarding weaknesses of the organization. Finally, different perspectives could be elaborated upon and considered during the prioritization of interventions, since the discussion group consists of a variety of different stakeholders including members of research councils, research managers, researchers and research users. Conclusions =========== In addition to identifying the weaknesses in the KT capacities of research centres and organizations, the tool can help develop interventional priorities to solve these barriers and difficulties. It seems that strengthening knowledge translation in TUMS will take place through the provision of financial support for knowledge translation activities, creating supportive and facilitating infrastructures, and facilitating interactions between researchers and target audiences. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= RM proposed the idea and designed it. He also participated in the tool design, statistical analysis and preparing the manuscript. JG participated in the tool design, data collection management, statistical analysis, initial draft preparation and subsequent manuscript corrections. SN participated in the study design, tool design and manuscript correction. KM translated the manuscript and assisted in interpreting the statistical analysis and manuscript correction. MA assisted in the tool design, data collection and manuscript correction. BY contributed to the development of the tool as well as the development of the report. All authors approved the final manuscript. Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Knowledge Translation Self Assessment Tool for Research Institutes (SATORI) in English**. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 2 **Knowledge Translation Self Assessment Tool for Research Institutes (SATORI) in Farsi**. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ This paper has been written as a result of a project approved by TUMS through contract no. 5806-74-02-86. The authors would like to extend their gratitude to the Vice Chancellor of Research in TUMS for the moral support they have received.

PubMed Central

2024-06-05T04:04:19.615989

2011-2-22

PMC3053267

Background ========== Cholangiocarcinoma (CC) is a highly lethal adenocarcinoma arising from bile duct epithelial cells. CC accounts for approximately 15% of the total liver cancer cases worldwide, and its incidence is rising \[[@B1],[@B2]\]. The prognosis for CC is quite poor because of difficulties in early diagnosis, and relative resistance of the tumors to chemotherapy \[[@B3],[@B4]\]. At the time of diagnosis, approximately 70% of CC patients have an occult metastasis or advanced local disease that precludes curative resection. Of candidates for curative resection, 30% develop recurrent disease at the anastomotic site or within the intrahepatic biliary tree, and succumb to disease progression or cholangitis \[[@B5]\]. Established risk factors for ductal cholangiocarcinomas include primary sclerosing cholangitis, infection with *Clonorchis sinensis*or *Opisthorchis viverrini*(liver flukes), Calori\'s disease, congenital choledochal cysts, and chronic intrahepatic lithiasis \[[@B6]\]. However, for most CCs, the cause is unknown. Recently, molecular investigations have provided evidence that CC carcinogenesis involves a number of genetic alterations, including activating point mutations in the *K-ras*oncogene, and in *p53*and *BRAF*\[[@B7]-[@B9]\]. The deregulated expression of a number of other genes has also been reported, and cyclooxygenase-2 and c-erbB-2 are frequently overexpressed in CCs, suggesting an involvement in early biliary carcinogenesis \[[@B10]\]. In addition, increased expression of interleukin-6 is frequently observed in CC \[[@B11]\]. CC also develops after the liver-specific targeted disruption of the tumor suppressors *SMAD4*and *PTEN*\[[@B12]\]. The incidence of sarcomatoid changes in CC is estimated to be approximately 5% \[[@B13]\], and sarcomatoid cells are thought to result from de-differentiation of ordinary carcinomatous CC cells. Sarcomatoid neoplasms are highly aggressive and have a poorer survival rate than ordinary CCs \[[@B14]\], but the underlying molecular alterations, which may be related to the epithelial-mesenchymal transition (EMT), remain unclear. Little extensive genome-wide information about altered gene expression in CCs is available, and only a few published studies have reported a comprehensive analysis of gene expression among biliary tract cancers in general \[[@B15],[@B16]\]. The advancement of microarray technology now enables us to analyze genome-wide gene expression in a single experiment, opening avenues for the molecular classification of tumors, detection of the biological nature of tumors, and prediction of prognosis and sensitivity to treatments. In this study, we generated genome-wide gene expression profiles of 10 cell lines (9 CC cell lines and 1 immortalized cholangiocyte line), and 19 CC tissues using a BeadChip oligonucleotide technology containing 48,000 genes. This procedure allowed us to observe a comprehensive pattern of gene expression in CC compared to cultured normal biliary epithelia (NBE). In addition, we identified a set of genes associated with sarcomatoid transdifferentiation. These data are useful not only because they provide a more profound understanding of cholangiocarcinogenesis and transdifferentiation, but also because they may help to develop diagnostic tools and improve the accuracy of CC prognosis. Methods ======= Cell lines and cultures ----------------------- Tumor tissues were obtained from surgical specimens and biopsy specimens in Korean cholangiocarcinoma patients. Tumor tissues were washed three times in Opti-MEM I (Gibco, Grand Island, NY) containing antibiotics. Washed tissue was transferred to a sterile Petri dish and finely minced with scalpels into 1- to 2-mm^3^fragments. Tissue fragments in culture medium were seeded in T25 culture flasks (Corning, Medfield, MA) in Opti-MEM supplemented with 10% fetal bovine serum (FBS, Gibco), 30-mM sodium bicarbonate and antibiotics. Tumor cells were cultured undisturbed and passaged as described \[[@B17]\]. Near the 20th passage, the medium was changed from Opti-MEM I to DMEM supplemented with 10% FBS and antibiotics. NBE cells were isolated from mucosal slices of normal bile ducts, with informed consent from liver transplantation donors, and *ex-vivo*cultured in T25 culture flasks in Opti-MEM supplemented with 10% FBS, 30 mM sodium bicarbonate and antibiotics at 37°C with 5% CO~2~in air. Near-confluent NBE cells were harvested and stored at -80°C until use. Cells were routinely tested for mycoplasma and found to be negative using a Gen Probe kit (San Diego, CA). CC cell lines are in Table [1](#T1){ref-type="table"}. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Clinicopathological features of nine patients with intrahepatic cholangiocarcinomas used to generate CC cells lines. ::: Patient No Cell line Age/Sex T/N/M Stage CA19-9 (U/ml) Differentiation Tumori genicity Comments ------------ ------------------- --------- ------------- --------------- ----------------- ----------------- ------------------- 1 CK-Choi (Choi-CK) M/68 IVB 184 WD \+ 2 CK-Cho (Cho-CK) M/82 IVA 500 MD \+ 3 CK-J (JCK) M/72 IVA 125 PD \+ *C. sinensis* 4 CK-S (SCK) M/68 IVA 235.6 PD \+ Sarcomatoid 5 CK-L1 M/46 IVA 0.01 PD \+ Combined with HCC 6 CK-L2 M/65 III 2050.1 MD \+ 7 CK-P1 M/66 IVA 23.7 MD \- 8 CK-P2 F/66 IVA 121.4 MD \+ 9 CK-Y1 M/52 IVA 0.01 PD \+ Combined with HCC M, male; F, female; C. sinensis, clonorchis sinensis; HCC, hepatocellular carcinoma; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated. \*International Hepato-Pancreato-Biliary Association classification. ::: 5-Aza-2\'-deoxycytidine (Aza) treatment --------------------------------------- Choi-CK, Cho-CK, and JCK cells were seeded at 1 × 10^6^cells/ml. After overnight culture, cells were treated with 5 μM of the DNA methylating agent Aza (Sigma-Aldrich, St. Louis, MO) for 4 days, and then harvested. Patients and tissue samples --------------------------- CC tissues were obtained with informed consent from Korean patients who underwent hepatectomy and common bile duct exploration at Chonbuk National University Hospital. All tumors were clinically and histologically diagnosed as cholangiocarcinoma. Detailed clinocopathological data of the 19 samples are in Table [2](#T2){ref-type="table"}. All samples were immediately frozen in nitrogen tanks. Patient information was obtained from medical records. Clinical stage was determined according to the International Hepato-Pancreato-Biliary Association (IHPBA) classification \[[@B18]\]. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Clinicopathological features of 19 CC samples used for microarray analysis. ::: Sample No Age/Sex Loca tion Size (cm) T N M Stage Gross appearance Differentiation Sub-class Comment ------------- --------- ----------- ----------- --- --- --- ------- ------------------ ----------------- ----------- ----------------------- 1 (CC-GHS) 68/F L 8.7 × 5.4 3 1 0 IVA MF PD A 2 (CC-CYS) 57/M L NA 3 1 0 IVA MF +PDI MD A 3 (CC-LJS) 42/M A NA 1 0 0 I ID WD B Intraductal papillary 4 (CC-BJP) 62/M P 7.8 × 5.6 1 0 0 I ID WD B Intraductal papillary 5 (CC-HSR) 66/M AP 7.3 × 6 2 0 0 II MF MD B 6 (CC-HSW) 59/M AP 9 × 6.8 2 0 0 II MF MD B 7 (CC-CSB) 60/M L 4 × 4.5 3 0 0 III MF MD B 8 (CC-SJS) 71/M A 2.1 × 1.9 1 0 0 II MF WD B 9 (CC-HDS) 63/M CBD 1.1 × 0.9 1 1 1 IVB ID MD B 10 (CC-KHC) 47/M L 14 × 10 4 0 0 IVA MF PD C 11 (CC-LHG) 42/M L 5.6 × 3.9 3 1 0 IVA MF PD C Combined with HCC 12 (CC-LSH) 40/F P 8.6 × 4 2 0 1 IVB MF PD C Combined with HCC 13 (CC-KHS) 70/F L 8.5 × 4.8 3 0 1 IVB MF +PDI PD C Combined with HCC 14 (CC-LMS) 38/F AP 5 × 3.7 4 0 1 IVB MF PD C Combined with HCC 15 (CC-KJA) 39/F LP 6 × 5 4 0 0 IVA MF MD C Combined with HCC 16 (CC-JSJ) 64/M L 0.5 × 0.5 2 0 1 IVB ID PD C 17 (CC-YCU) 53/M L 4.3 × 2.1 3 1 0 IVA MF MD C 18 (CC-GMG) 50/M L 7 × 3.5 2 0 0 II MF MD C 19 (CC-BSD) 67/M L 2.9 × 2.7 4 0 0 IVA MF+P야 PD C HCC, hepatocellular carcinoma; M, male; F, female; A, anterior segment; P, posterior segment; Med, medial segment; L, lateral segment; MF, mass forming type; PDI, periductal infiltrating type; IDG, intraductal growth type; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; NA, not available. \*International Hepato-Pancreato-Biliary Association classification. ::: Primer labeling and Illumina Beadchip array hybridization --------------------------------------------------------- Total RNA from CC samples was isolated using TRIzol reagent (Invitrogen, CA) according to the manufacturer\'s instructions. RNA quality was determined by gel electrophoresis, and concentrations were determined using an Ultrospec 3100 pro spectrophotometer (Amersham Bioscience, Buckinghamshire, UK). Biotin-labeled cRNA samples for hybridization were prepared according to Illumina\'s recommended sample-labeling procedure: 500 ng of total RNA was used for cDNA synthesis, followed by an amplification/labeling step (*in vitro*transcription) to synthesize biotin-labeled cRNA using the Illumina TotalPrep RNA Amplification kit (Ambion Inc., Austin, TX). cRNA concentrations were measured by the RiboGreen method (Quant-iT RiboGreen RNA assay kit; Invitrogen-Molecular Probes, ON, Canada) using a Victor3 spectrophotometer (PerkinElmer, CT), and cRNA quality was determined on a 1% agarose gel. Labeled, amplified material (1500 ng per array) was hybridized to Illumina Human-6 BeadChips v2 containing 48,701 probes for 24,498 genes, according to the manufacturer\'s instructions (Illumina, San Diego, CA). Array signals were developed by Amersham fluorolink streptavidin-Cy3 (GE Healthcare Bio-Sciences, Little Chalfont, UK) following the BeadChip manual. Arrays were scanned with an Illumina Bead-array Reader confocal scanner (BeadStation 500GXDW; Illumina) according to the manufacturer\'s instructions. Array data processing and analysis were performed using Illumina BeadStudio software. The BeadStudio Gene Expression Module is a tool for analyzing gene expression data from scanned microarray images generated by the Illumina BeadArray Reader. Data analysis ------------- Normalization algorithms were used to adjust sample signals to minimize the effects of variation from non-biological factors. To reduce variation between microarrays, the intensity values for samples in each microarray were rescaled using a quartile normalization method in the BeadStudio module. Measured gene expression values were log2-transformed and median-centered across genes and samples for further analysis. To generate an overview of the gene expression profile and to identify major relationships in cell lines, we used unsupervised hierarchical clustering analysis. Genes with an expression ratio of at least a two-fold difference relative to the median gene expression level across all samples in at least 10% of samples were selected for clustering analysis. Average linkage hierarchical cluster analysis was carried out using a Pearson correlation as the similarity metric, using the GeneCluster/TreeView program (<http://rana.lbl.gov/EisenSoftware.htm>). Expression profiles for the differentially expressed genes were selected by *t*-test with false discovery rate (FDR) and q-values as gene significance measures, using R software (version 2.5). Because of varying significance in the analyzed comparisons, using a fixed FDR (or q-value) cut-off value was not practical. Therefore, we used *t*-test *P*= 0.01. To ascertain biological relevance, a fold-change cut-off value of 2 or 4 from the mean was chosen. The gene ontology (GO) program (<http://david.abcc.ncifcrf.gov/>) was used to categorize genes in subgroups based on biological function. Values for each GO group were calculated as a percentage of total mRNA change. For example, the Fisher exact test was used to determine whether the proportions of genes in each category differed by group. The microarray data were registered with the Gene Expression Omnibus (GEO) database (Accession No. GSE22633) Immunoblotting -------------- Extracted protein (30 μg) from cell lysates was resolved by SDS-PAGE and transferred to a nitrocellulose membrane. Membranes were incubated for 1 h at room temperature with primary antibody at 1:1000 dilution. After incubation, blots were washed three times in TBS/0.1% Tween 20. Immunoreactivity was detected using alkaline phosphatase-conjugated goat anti-rabbit IgG or a commercial chemiluminescence detection kit (Amersham), according to the manufacturer\'s instructions. Immunohistochemistry -------------------- Immunohistochemical staining was performed on formalin-fixed, paraffin-embedded 4-μM tissue sections, as described preciously \[[@B19]\]. Briefly, a deparaffinized section was pretreated by microwave epitope retrieval (750 W during 15 min in citrate buffer 10 mmol; pH 6.0) after rehydration. Before applying primary antibodies, the endogenous peroxidase activity was inhibited with 3% hydrogen peroxide, and a blocking step with biotin and bovine albumin was performed. Primary monoclonal or polyclonal antibodies were detected using a secondary biotinylated antibody and a streptavidin-horseradish peroxidase conjugate according to the manufacturer\'s instructions (DAKO, Glostrup, Denmark). Counterstaining was performed using Meyer\'s hematoxylin. Tumors were evaluated for the percentage of positive cells and the staining intensity. Negative controls were samples incubated with either PBS or mouse IgG~1~instead of primary antibody. Real-time RT-PCR ---------------- RNA prepared from dissected tissues was precipitated with isopropanol and dissolved in DEPC-treated distilled water. Reverse transcription (RT) was performed using 2 μg total RNA, 50 μM decamer and 1 μl (200 units) and RT-PCR Superscript II (Invitrogen) at 37°C for 50 min, as previously described. Specific primers for each gene were designed using the Primerdepot website (<http://primerdepot.nci.nih.gov/>) and are in Additional file [1](#S1){ref-type="supplementary-material"}. The control *18S*ribosomal RNA primer was from Applied Biosystems (Foster City, CA) and was used as the invariant control. The real-time RT-PCR reaction mixture consisted of 10 ng reverse-transcribed total RNA, 167 nM forward and reverse primers, and 2 × PCR master mixture in a final volume of 10 μl PCR, was in 384-well plates using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems). Animal model of cholangiocarcinoma ---------------------------------- The hamster CC model was modified from a previous study \[[@B20]\]. On the first day of the experiment, hamsters in the experimental group were infected with 15 metacercariae of the liver fluke, *C. sinensis*. One day after parasite infestation, hamsters received 15 ppm of dimethylnitrosamine (DMN; Kasei, Japan) in the drinking water for 4 weeks with a normal diet. Thereafter, hamsters were given tap water with a normal diet for the rest of the study. An interim stage of cholangiocarcinogenesis was confirmed at 8 weeks after experiment initiation. Control and CC model hamsters were maintained for a total of 27 weeks for CC to develop. Results ======= Gene expression patterns distinguish CC cells from cultured NBE cells --------------------------------------------------------------------- Using BeadChip microarray analysis, we compared the gene expression profiles of nine CC cell lines, an immortalized biliary epithelial cell line, and four types of NBE cells. We selected 828 unique genes with a 2-fold or greater expression difference from the mean, with a *P*\< 0.01 by *t*-test. Unsupervised hierarchical clustering analysis of all samples was based on the similarity in the expression pattern of all genes (Figure [1](#F1){ref-type="fig"}). Cell samples were separated into two main groups, the NBE cluster, and the transformed and immortalized biliary epithelial cells (CCC cluster). Each distinctive gene cluster was identified by delineation using a hierarchical clustering dendrogram. Cluster I consisted of genes upregulated in CC cells, which included tumor-related genes such as *LGR4, AGR2*, *PCAF, TMEM97, FRAT2*, *EFNB2*and *ZIC2*\[[@B21]-[@B27]\]. Cluster II included genes underexpressed in CC cells. These were mainly tumor suppressor genes such as *GREM1, THY1, STC2, SERPINE1, SPARC*and *TAGLN*\[[@B28]-[@B33]\]. Cluster III was genes upregulated in NBE cells, and contained the *PDGFRA*, *CD248*, and *BDKRB1*genes. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Unsupervised hierarchical clustering of four biliary epithelial cells, one immortalized cholangiocyte cell line and nine CC cells**. Unsupervised hierarchical clustering separated the samples into two main groups, normal biliary epithelial cells (NBE) isolated from mucosal slices of normal bile ducts and *ex-vivo*cultured as described in Methods, and cholangiocarcinoma cells (CCC). Data are in matrix format, with columns representing individual cell lines and rows representing each gene. Red, high expression; green, low expression; black, no significant change in expression level between the mean and sample. A hierarchical clustering algorithm was applied to all cells and genes using the 1 - Pearson correlation coefficient as a similarity measure. Raw data for a single array were summarized using Illumina BeadStudio v3.0 and output to the user was as a set of 43,148 values for each individual hybridization. We selected 828 unique genes with a two-fold or greater difference from the mean and *P*\< 0.01 by *t*-test, for hierarchical clustering analysis. Specific gene clusters (Cluster 1 through Cluster III) were identified in the hierarchical cluster of the genes differentially expressed in CCC compared with NBE. CC, cholangiocarcinoma; IMC, immortalized cholangiocytes. :::  ::: Gene expression patterns distinguish CC tissues from cultured NBE cells ----------------------------------------------------------------------- Using BeadChip microarrays, gene expression profiles of 19 CC tissues and 4 types of NBE cells were compared. We selected 1798 unique genes with a 2-fold or greater differences from the mean difference with a *P*\< 0.01 by *t*-test. Unsupervised hierarchical clustering analysis was as described above (Figure [2A](#F2){ref-type="fig"}). All samples separated into two main groups, NBE and CC tissues (CCT). Each distinctive gene cluster was identified using a hierarchical clustering dendrogram as above. Intriguingly, the CC sample cluster was divided into two subclasses by tumor differentiation: differentiated (Df) and undifferentiated (Udf). Clustering data for the CC group revealed three clusters. Cluster I had genes upregulated in NBE and downregulated in CCT including *SERPINB2, PAPPA, LRRC17*, and *GREM1*. Cluster II contained genes upregulated in the Df CCT and downregulated in NBE. Cluster III included genes upregulated in poorly differentiated or Udf CCT, and downregulated in NBE. A supervised hierarchical clustering analysis was performed between the NBE class, and the Df and the Udf subclasses based on the similarity of expression pattern of all genes (Figure [2B](#F2){ref-type="fig"} and [2C](#F2){ref-type="fig"}). We selected 420 differentially expressed genes in the Df subclass, and 646 genes in the Udf subclass for comparison with the NBE class (Additional files [2](#S2){ref-type="supplementary-material"} and [3](#S3){ref-type="supplementary-material"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Unsupervised hierarchical clustering of 4 biliary epithelial cells and 19 CC tissues**. (A) Unsupervised hierarchical clustering separated the samples into two main groups. We selected 1798 unique genes with two-fold or greater difference from the mean with *P*\< 0.01 by *t*-test for hierarchical clustering analysis. Specific gene clusters (Cluster 1 through Cluster III) were identified of differentially expressed in CCT compared to NBE. (B) Supervised hierarchical clustering of four biliary epithelial cells and seven differentiated CC tissues. We selected 420 unique genes with four-fold or greater difference from the mean and *P*\< 0.01 by *t*-test for hierarchical clustering analysis. (C) Supervised hierarchical clustering of 4 biliary epithelial cells and 10 undifferentiated CC tissues. We selected 646 unique genes with the criteria in B for hierarchical clustering analysis. :::  ::: Differential expression and verification of CC-related genes ------------------------------------------------------------ We compared the gene lists from the cell-based and tissue-based databases, and selected 342 commonly regulated genes, including 53 commonly upregulated genes and 289 commonly downregulated genes (Figure [3A](#F3){ref-type="fig"}). The top 25 commonly regulated genes in both CCC and CCT compared to NBE are in Additional file [4](#S4){ref-type="supplementary-material"}. To verify the microarray data, we examined the mRNA levels of the identified genes using real-time RT-PCR in human CC tissues. We selected five up-regulated genes from the commonly upregulated genes of both the cell and tissue sample classes (Figure [3B](#F3){ref-type="fig"}). We also chose the *IRX3*, *PTTG1*, and *PPARγ*genes, which were highly upregulated in only the cell sample class. These genes were preferentially expressed in CC cells and tissues. We also examined the expression of the commonly downregulated *KRT17*and *UCHL1*genes, as well as the cellular downregulated *IGFBP7*and *SPARC*genes using real-time RT-PCR in human CC. The human NBE showed substantial expression of *CK-17*, *UCHL1*, *IGFBP7* and *SPARC*, which were barely detected in CC tissues (Figure [3C](#F3){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Differentially regulated genes in human CC tissues compared to NBE cells**. (A) Venn diagram of genes commonly regulated in the cell and tissue samples. The 342 genes included 53 upregulated and 289 downregulated genes, selected from the cell- and tissue-based microarray databases. (B) Real-time RT-PCR analysis of upregulated genes selected from the list of top 25 genes commonly upregulated in both CC cells (C) and tissues (T) compared to cultured NBE cells (N). \*, selected from only the cell-based microarray database. (C) Real-time RT-PCR analysis of downregulated genes selected from the list of top 25 genes commonly downregulated in both CC cells (C) and tissues (T) compared to cultured NBE cells (N). \*, selected from only the cell-based microarray database. :::  ::: Immunohistochemical analysis of CC-related genes ------------------------------------------------ To confirm the reliability of the microarray data and the robustness of the strategy for identifying genes with altered expression, we examined the protein levels of the identified genes using immunohistochemical analysis of human tissues (Figure [4A](#F4){ref-type="fig"}). We selected three upregulated genes from the genes that were upregulated in both cell and tissue samples. The SPP1, EFNB2 and E2F2 proteins were abnormally overexpressed in the CC cell cytoplasm, and weakly or barely expressed in HCC. We also examined the IRX3, PTTG1, and PPARγ proteins, which were highly upregulated in only the cell samples. IRX3 was the most highly upregulated, and we was strongly expressed in the nucleus of CC cells in the tissue sections, but was barely detectable in the NBE nuclei. PTTG1 and PPARγ were abnormally overexpressed in the CC cell cytoplasm, and their expression was attenuated in poorly differentiated CC. Next, we also used immunohistochemical staining of human CC to examine the KRT17 and UCHL1 proteins, whose genes were both downregulated in CC cells and tissues, and the IGFBP7 and SPARC proteins, which were downregulated in CC cells only. Human NBE showed substantial expression of the CK-17, UCHL1, IGFBP7, and SPARC proteins, but these were barely detectable in CC tissue. However, KRT-17 was clearly positive in HCC (Figure [4B](#F4){ref-type="fig"}). ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Immunohistochemical staining of differentially expressed proteins in the CC tissues**. (A) Immunohistochemical staining with anti-SPP1, anti-EFNB2, anti-E2F2, anti-IRX3, anti-PTTG1 or anti-PPARγ in NBE, human CC tissues with good differentiation (well), moderate differentiation (mod) or poor differentiation (poor), and HCC tissues. The representative genes were selected from the list of top 25 commonly upregulated genes, according to antibody available for immunohistochemistry. \*, selected from only cell-based microarray database. The representative picture is one of three immunohistochemical staining replicates of different specimens. (B) Immunohistochemical staining with anti-KRT-17, anti-UCHL1, anti-IGFBP7 or anti-SPARC in the human CC and HCC tissues. \*, selected from only cell-based microarray database. The representative picture is one of three immunohistochemical staining replicates of different specimens :::  ::: Immunohistochemical analysis in hamster model of CC --------------------------------------------------- Although it is unknown whether antibodies raised to human proteins recognize hamster proteins, we examined the protein levels of the identified genes using immunohistochemical analysis of hamster CC tissues (Additional file [5](#S5){ref-type="supplementary-material"}). As in humans, the SPP1, EFNB2, and E2F2 proteins were abnormally overexpressed in the hamster CC cell cytoplasm. IRX3 was also similarly expressed in the CC cell nucleus, and PTTG1 was differentially expressed in the CC cell cytoplasm. Interestingly, in contrast to human CC cells, PPARγ was preferentially expressed in the hamster CC cell nuclei. Therefore, the immunoreactivity of identified gene proteins in hamster CC seemed to be substantially consistent with that in human CC. Gene expression patterns distinguish the SCK cell line from three CC cell lines ------------------------------------------------------------------------------- Previously, we established four human CC cell lines and characterized one with a typical sarcomatoid phenotype of SCK. We classified the other cell lines according to tumor cell differentition, as a poorly differentiated JCK, a moderately differentiated Cho-CK, and a well-differentiated Choi-CK cell line \[[@B34]\]. Two-way unsupervised hierarchical clustering analysis of quadruplicate samples for each cell line was conducted, based on the similarity of expression patterns of all genes (Figure [5](#F5){ref-type="fig"}). We selected 559 unique genes whose expression differed from the mean by four-fold or more with *P*\< 0.005 by *t*-test. Cell samples were separated into two main groups, sarcomatoid (SC) and ordinary or adenocarcinomatous CC (AC), by the gene axis. The SC group contained 292 differentially upregulated genes (\>four-fold change), and 267 downregulated genes (\<0.25-fold change), compared to the AC group. The top 25 genes that were differentially expressed in the sarcomatoid SCK cells compared to the three adenocarcinomatous CC lines are in Additional file [6](#S6){ref-type="supplementary-material"}. Clustering data within groups revealed that the core clusters I and II were associated with transdifferentiation. Genes in cluster I appeared to be downregulated in the SCK cells, compared to ordinary CC cells. In contrast, the genes in cluster II were upregulated in the SCK cells and downregulated in the ordinary CC cells. Cluster I contained the *GSTT1*, *TACSTD*, *BST2*, *RAB25*, and *MAL2*genes. Cluster II contained genes associated with tumor progression and metastasis, including *HOXA9, MUC13*, and members of the *GAGE*and *CT-45*families \[[@B35]-[@B38]\]. Expression of methylation-silenced genes, such as *LDHB, BNIP3, UCHL1*, and *NPTX2*\[[@B39]-[@B42]\], was barely detectable in the AC group, but appeared in this cluster. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Unsupervised hierarchical cluster analysis of differentially expressed genes illustrated in a heat-map**. Unsupervised hierarchical clustering separated the samples into two main groups: SC and AC. The samples were independently prepared from the cultured cells four times, and four kinds of CC cells were used: Choi-CK, Cho-CK, JCK and SCK cells. Samples were clustered closer within their own group than in samples from other groups. We selected 559 unique genes with a four-fold or greater difference from the mean and a *P*\< 0.005 by *t*-test for hierarchical clustering analysis. Cluster I included genes differentially downregulated in the sarcomatoid CC cells compared to three adenomatous CC lines. Cluster II contained genes differentially upregulated in the sarcomatoid CC cells compared to three adenomatous CC lines. :::  ::: Expressions of transdifferentiation-related genes ------------------------------------------------- From 559 genes that were differentially regulated between SCK cells and the three ordinary CC lines, we selected six upregulated genes and six downregulated genes, and examined their mRNA expression using real-time RT-PCR (Figure [6A](#F6){ref-type="fig"}), which verified the differential expression. We examined protein expression by Western blot analysis of the four CC lines. LDHB, Bnip3, HO-1, and UCHL1 were overexpressed exclusively in SCK cells. The expression of VIM and TWIST1 increased according to tumor dedifferentiation and was highest in SCK cells (Figure [6B](#F6){ref-type="fig"}, left). In contrast, LCN2, S100P, KRT7, KRT19, GPX1, and EFNA1 were preferentially expressed in Choi-CK, Cho-CK and JCK cells, but minimally expressed in SCK cells (Figure [6B](#F6){ref-type="fig"}, right). Because *LDHB*, *BNIP3*, and *UCHL1*are well-known methylation-silenced genes in tumors \[[@B39]-[@B41]\], and are highly expressed in SCK cells, this suggested that DNA demethylation was involved in CC. To confirm this hypothesis, we treated the AC cells with the demethylating agent Aza, which dramatically restorated expression of the silenced *UCHL1*gene in these cells (Figure [6C](#F6){ref-type="fig"}). In addition, we performed immunohistochemical examination of protein expression according to tumor dedifferentiation in human CC tissue (Figure [6D](#F6){ref-type="fig"}). As expected, HO-1 was exclusively overexpressed in SC, while TWIST1 was overexpressed in the poorly differentiated and SC cells. In contrast, LCN2 was exclusively downregulated in SC, while EFNA1 expression decreased with tumor dedifferentiation. Therefore, expression of these proteins clearly correlated with clinicopathological features such as tumor differentiation and EMT change, in CC tissues. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Genes and proteins differentially expressed in sarcomatoid CC and adenomatous CC cells**. (A) Real-time RT-PCR analysis of upregulated (left) and downregulated (right) genes selected from the list of top 25 genes differentially expressed in sarcomatoid SCK cells and three adenocarcinomatous CC cell lines. (B) Immunoblot of upregulated (left) and down-regulated (right) proteins in sarcomtoid SCK cells compared to three adenocarcinomatous CC cell lines. \*, selected from the top 100 genes differentially expressed in sarcomatoid SCK cells compared to three adenocarcinomatous CC cell lines. Lane 1, Choi-CK cells. Lane 2, Cho-CK cells. Lane 3, JCK cells, Lane 4, SCK cells. (C) Immunoblot of UCHL1. Expression was restored in adenocarcimatous CC cells by treatment of 5 μM of Aza for 4 days, compared to vehicle control (VC). (D) Immunohistochemical staining of down- or upregulated proteins according to tumor dedifferentiation. The representative picture is one of three immunohistochemical staining replicates of different specimens. Well, well-differentiated CC. Mod, moderately differentiated CC. Poor, poorly differentiated CC. Sar, sarcomatoid CC. \*, selected from the top 100 genes differentially expressed in sarcomatoid SCK cells compared to three adenocarcinomatous CC cell lines. :::  ::: Discussion ========== In this study, our experimental design primarily investigated the gene expression profiles of 10 cell lines and 19 CC tissues, and compared these profiles to those from four cultured NBE cell line using genome-wide BeadChip microarray analysis. Transdifferentiation-related genes were analyzed by same method. Using unsupervised hierarchical clustering analysis, we found that the *SPP1*, *EFNB2*, and *E2F*genes were commonly upregulated in both cell and tissue samples. *IRX3*, *PTTG1*, and *PPARγ*were upregulated in the cell samples, and were immunohistochemically verified in human and hamster CC tissues. SPP1 (osteopontin), a secretory adhesive alycoprotein, was identified as a highly overexpressed gene in CC lines and tissues. SPP1 is a ligand of CD44 that binds to αV-containing integrins and is important in malignant cell attachment and tumor invasion \[[@B43]\]. It was a highly overexpressed gene in HCC, and its expression correlated with earlier recurrence, poorer prognosis, and metastasis \[[@B44]\]. Consistent with our findings, a recent oligonucleotide microarray study reported that SPP1 was the most highly expressed gene in intrahepatic cholangiocarcinoma \[[@B45]\]. EFNB2 was identified as a preferentially expressed genes in CC. EFNB2 overexpression is reported to be significantly correlated with the number of lymph node metastases and clinical stage in esophageal cancer \[[@B46]\]. Several reports have examined concomitant expression of the ligand EFNB2 and its receptor EphB4 in leukemia-lymphoma cell lines \[[@B47]\], and in endometrial cancer \[[@B48]\]. E2Fs 1-3 are characterized as \"activator E2Fs\" since their binding to promoters results in increased transcription, while E2Fs 4 and 5 are \"repressor E2Fs\" since they form complexes with p130, HDACs, and other factors to block transcription \[[@B49]\]. During hepatocarcinogenesis in c-myc/TGFalpha double-transgenic mice, expression of E2F-1 and E2F-2 increases, and putative E2F target genes are induced \[[@B50]\]. For immunohistochemical verification, the representative genes were selected from the list of top 25 commonly upregulated genes, according to antibody available for immunohistochemistry. In addition, other genes were selected from only cell-based microarray database. The same immunohistochemical staining in hamster CC tissues induced by *Clonorchiasis*infestation was compared with control staining in normal hamster livers. IRX3 is involved in dorsal-ventral patterning in spinal cord development and coordination with other homeobox genes \[[@B51]\]. IRX3 is preferentially expressed in the examined CC tissues and localized to the nucleus of human and hamster malignant biliary epithelial cells, independent of cell differentiation. A methylated CpG island was detected in exon 2 of the IRX3 locus, rather than in the promoter, and is responsible for IRX3 overexpression in brain tumor cells and tissues \[[@B52]\]. PTTG1, a critical mitotic checkpoint protein, is a known proto-oncogene that is highly expressed in HCC \[[@B53]\]. Our data showed that PTTG1 was preferentially expressed in the cytoplasm of the human and hamster CC cells. PPAR-*γ*, a member of the nuclear receptor superfamily, functions as a ligand-activated transcription factor \[[@B54]\]. It is overexpressed in a variety of cancers, including HCC and pancreatic cancer \[[@B55],[@B56]\]. Positive immunostaining was localized in the cytoplasm and nuclei of human CC cells. However, positive immunostaining was exclusively detected in the nuclei of the hamster CC cells. Our data also immunohistochemically validated the downregulation of proteins KRT17, UCHL1, IGFBP7, and SPARC. Our hamster model showed the similar expression patterns of human CC related genes and therefore might be a relevant model to study human CC. Analysis of genes involved in the transdifferentiation of CC cells showed two clusters in the gene axis, with genes that were upregulated (cluster II), and downregulated (cluster I) in the SC group as compared to the AC group. The mesenchymal antigen VIM and the transcriptional factor TWIST1 were upregulated in JCK and SCK cells by tumor dedifferentiation. The overexpression of these proteins is reported to be associated with the EMT \[[@B57],[@B58]\]. Intriguingly, genes silenced by promoter hypermethylation during CC development were restored at the point of sarcomatous transdifferentation, which implied that the demethylation may be involved in the EMT progression of CC. In addition to tumor-related genes known to be overexprssed in intrahepatic CC, we identified other strongly and consistently dysregulated genes in CC that are known to be involved in other human cancers. Our data support a correlation between the expression of these genes and CC tumor differentiation, and the gene expression patterns found in this study are consistent with those associated with a poor clinical prognosis for this cancer. gene expression profiling appears to be a useful diagnostic tool, especially for differentiating CC from other liver masses, as well as for the subclassification of intrahepatic CC compared to histopathological findings. Conclusions =========== Gene expression profiling appears to be a useful diagnostic tool, especially for differentiating CC from other liver masses, as well as for the subclassification of intrahepatic CC compared to histopathological findings. The most consistently overexpressed genes are candidate therapeutic targets, and related genes can be used for predicting survival and outcomes for different therapeutic modalities. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= MS and IC performed most of the experiments and drafted the manuscript. ML and GY carried out the tissue collection and the establishment of cell lines. XC participated in the immunohistochemical analysis. BC and IK participated in the design and coordination of the study and helped to draft the manuscript. EA and SL participated in the array data processing and analysis. DK conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/11/78/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **Supplementary Table S1:** Sequences and accession numbers for the forward (FOR) and reverse (REV) primers used in real-time RT-PCR. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file2 **Supplementary Table S2**: List of genes differentially expressed between differentiated cholangiocarcinoma and normal biliary epithelium (NBE). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 3 **Supplementary Table S3**: List of genes differentially expressed between undifferentiated cholangiocarcinoma and normal biliary epithelium (NBE). ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 4 **Supplementary Table S4**: Top 25 genes commonly regulated in both CC cells and tissues compared with cultured biliary epithelial cells. ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 5 **Supplementary Figure S1**: Immunohistochemical staining with anti-SPP1, anti-EFNB2, anti-E2F2, and anti-IRX3 in hamster CC tissues induced by Clonorchiasis infestation. Control stainings were performed in normal hamster livers. \*, selected from only cell-based microarray database. 35 ::: ::: {.caption} ###### Click here for file ::: ::: {.caption} ###### Additional file 6 **Supplementary Table 5**. Top 25 genes differentially expressed in the sarcomatoid SCK cells compared with 3 adenocarcinomatous CC cell lines. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ We gratefully thank Prof. Yoon B-I (Kangwon National University, South Korea) for providing the hamster CC tissues. This study was supported by grants from the Korean Association of Internal Medicine (Chungram Research Fund, 1997), a Korea Research Foundation grant from the Korean government (Basic Research Promotion Fund, KRF-2008-313-E00434), and the National R&D Program for Cancer Control (0620220) and the Korean Health Technology R&D Project (A101834), Ministry for Health, Welfare and Family affairs (0620220), Republic of Korea.

PubMed Central

2024-06-05T04:04:19.619833

2011-2-19

PMC3053268

Background ========== Nuclear receptor families play a pivotal role in regulating genes involved in drug metabolism and disposition. Pregnane X receptor (PXR, also termed SXR, PAR, and *NR1I2*as its gene name) is a crucial regulator of various phase I and phase II drug metabolizing enzymes and drug transporters. PXR is expressed in liver, small intestine and other organs. PXR, with a number of therapeutic drugs and other xenobiotics as its ligands, dimerizes with retinoid X receptor α (RXRα). The ligand-PXR-RXRα complex binds to promoter and enhancer elements located upstream of cytochrome P450s (CYPs) 3A and 2C family members, UDP-glucuronosyltransferases (UGTs), sulfotransferases (SULTs), glutathione *S*-transferases (GSTs), and ATP binding cassette (ABC) drug transporters (reviewed in \[[@B1]-[@B3]\]). Wide inter-individual variability has been documented in the expression of hepatic CYP3A4 with respect to basal and PXR-inducible activities. The genetic variability of PXR and CYP3A4 is not sufficiently frequent to explain the apparent inter-individual variability \[[@B4]\]. The inter-individual variability in basal hepatic CYP3A4 expression may include variability in PXR expression, as PXR is activated by endogenous steroid hormones and bile acids. Inter-individual variability of CYP3A4 expression is also observed in human intestine. Interestingly, a report using paired tissue samples of livers and small intestines indicated no observed correlation between the hepatic and small intestine CYP3A4 expression levels \[[@B5]\]. Another research group reported that a majority of CYP3A4 resided in the proximal region of the small intestine, and that the CYP3A4 protein levels decreased dramatically in the distal small intestine \[[@B6]\]. These observations suggest the CYP3A4 is regulated through tissue specific epigenetic regulation in normal tissue. The aim of the present study is to find possible and not yet fully elucidated mechanisms which regulate heterogeneous basal PXR and CYP3A4 expression and activity in cancerous tissues as well. Indeed, several studies previously found a fraction of genes that exhibited inter-individual differences in transcript levels associated with DNA methylation status \[[@B7],[@B8]\]. DNA methylation of the CpG-rich sequence around exon 3 of the *PXR*gene is involved in the epigenetic regulation of PXR in human neuroblastoma \[[@B9]\]. However, epigenetic regulation of PXR and CYP3A4 in human gut is poorly understood. In order to determine whether epigenetic mechanisms function in PXR and CYP3A4 regulation and intestinal metabolism, we examined DNA methylation and mRNA expression of several candidate genes on the PXR/CYP3A4 regulatory pathway in human colon cancer cell lines and tissues. Methods ======= Cell lines and tissue samples ----------------------------- Human colorectal cancer cell lines LS180, Caco-2, HT29, HCT116, DLD-1, LoVo, SW48 and SW620 were purchased from DS Pharma Biomedical Co., Ltd. (Osaka, Japan). LS180 cells were cultured in E-MEM medium (Invitrogen Corp., Carlsbad, CA) at 37°C under an atmosphere of 5% CO~2~. The other cells were cultured under conditions described elsewhere \[[@B10]\]. Eighteen pairs of cancerous and adjacent normal mucosa were excised from surgical specimens of colorectal cancers. The cancerous and normal epithelia were separated from stroma using crypt isolation \[[@B11]\]. All samples were selected from the same series of cancers as we used in a previous study \[[@B10]\]. All 18 patients with colorectal cancers did not receive chemotherapy before surgical resection. The study protocol was approved by ethics committee of Iwate Medical University (molecular analysis of gastrointestinal tumors and the surrounding mucosa; reference number, H21-140). Treatment with 5-aza-2\'-deoxycytidine -------------------------------------- LS180, LoVo, Caco-2, HCT116, HT29 and SW48 cells were seeded at a concentration of 1 × 10^5^cells on a 100 mm dish. The next day, treatment of cells with 0, 0.5 or 5 μM 5-aza-2\'-deoxycytidine (5-aza-dC) (Sigma Chemical, St. Louis, MO) was started, and 5-aza-dC was removed by changing the medium 24 h later. The cells were harvested 4 days after removal of 5-aza-dC for DNA and RNA extraction. Quantitative real-time PCR analysis of basal CYP3A4, PXR and VDR mRNA levels ---------------------------------------------------------------------------- Total RNA extraction, cDNA synthesis and real-time PCR were carried out on the cells prepared above, by the same methods as described previously \[[@B10]\]. mRNA levels of the CYP3A4 (exon 3-4, Hs01546612\_m1), PXR (exon 5-6, Hs00243666\_m1) and vitamin D receptor (VDR) (exon 10-11, Hs01045840\_m1) were evaluated by TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA). In addition, the mRNA level of the PXR splicing variants (exon 1a-2) was also examined by SYBR Green assays using the following primers: 5\'-GATTGTTCAAAGTGGACCCC-3\'(forward) and 5\'-TCCAGGAACAGACTCTGTGT-3\'. The mRNA level of the above target genes was normalized to β-actin mRNA \[[@B10]\]. All samples were analyzed in duplicate and average quantities of the gene transcripts were used for calculation. Deviation of the mRNA level of each sample was within 7% of the average. DNA methylation analysis of the 6 colon cancer cell lines and 18 colon cancer samples ------------------------------------------------------------------------------------- We found CpG islands within the *PXR*(around exon 3 region), *VDR*(promoter region) and protein arginine methyltrasferase 1(*PRMT1*) (promoter region) genes using the CpG Island Searcher program \[[@B12],[@B13]\]. A CpG island was also detected in the 5\' untranslated region (UTR) of the *CYP3A4*gene (approximately 25 kb distal to the transcription start site). We also found a CpG-rich sequence in the promoter region of the *PXR*gene, although this sequence did not strictly satisfy the criteria for a CpG island \[[@B13],[@B14]\]. The location of all CpG sequences examined in this study are shown in Figure [1](#F1){ref-type="fig"}. Genomic DNA extracted from the 6 cell lines and the 18 pairs of normal and colon cancer tissue samples was modified by sodium bisulfite, and then each segment including a CpG island or CpG-rich sequence was amplified by PCR and subjected to direct sequencing. The methylation status of the PXR promoter sequence was also estimated by bisulfite sequencing on at least 7 individual DNA strands after subcloning of PCR products into the pCR4-TOPO vector using the TOPO TA Cloning Kit for Sequencing (Invitrogen). A relative methylation level of the PXR promoter was visually determined by the density of each *HpyCH4IV*-digested band using combined bisulfite restriction analysis (COBRA) \[[@B15]\]. The methylation status of the *PXR*exon 3 region was also examined by the COBRA assay using an *HhaI*digestion. All primer sequences are listed in Table [1](#T1){ref-type="table"}. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Location of the CpG sequences examined**. CpG sites (vertical bars), CpG islands identified by CpG island searcher (horizontal thick lines) and transcription start sites within the 5\' prime region (curved arrows) are shown in the (a), *PXR*; (b), *CYP3A4*; (c), *VDR*; and (d), *PRMT1*genes. The segments indicated by double-pointed arrows were examined for DNA methylation by bisulfite direct sequencing. :::  ::: ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Primers used for DNA methylation analysis. ::: Gene Segment Primer sequence (5\'-3\') Annealing Product ------------ --------- ---------------------------------------------- ----------- --------- 1F GAAGATAATTGTGGTTATTTTTTGGTA 55 574 1R CCACCTCCCTAAATAATATTACT 2F GTTGTTTTTAGTGGTAAAGGATAGA 55\* 602 2R CACACACATCTTTTAACTAAAACT *PXR* 3F AGGATTTATTATTTTAAGGAGGGGTT 55 230 *promoter* 3R CCTTAAAACAATACCTCTAACCAT 4F GTAAGATTTGGAGATTTTTTATATTTG 55 318 4R CTATCCTTCTCTACTAATAAAAATAC COBRA GTAGGGAGAATATAATGAGAATAA 55\* 209 ACTAAAATAAAAACAATACTTCCTCTTC[A]{.underline}C *PXR* F ATTTTTTTATAGGAGGGTTATGAAA 55 292 *exon 3* R TACACACRAACACCAACTCACATAT *CYP3A4* F GAGTTATGGTGGGTTTTATTTAG 55 438 *5\'UTR* R TCTACATTTCCATCTAAAATACC *VDR* 1F ATAATTTTAGGTTTTAGGAGGTAG 60 400 promoter 1R CCTAAACTAACCAAACCAAAACTT 2F GGGTTGTTTTTGTTTGTTAAAAGG 60 373 2R CTTATTACCCAAATACTAAACACT *PRMT1* F AGGAGAAAGGGGGGGTTTTGGT 55\* 273 *promoter* R AACCCTTAAAAACTAAAAAACC F, forward primer; R, reverse primer \*Dimethyl sulfoxide (5%) was added to the PCR mixture. ::: Results ======= Basal mRNA levels of the CYP3A4, PXR and VDR genes in the 6 colon cancer cell lines ----------------------------------------------------------------------------------- Real-time PCR analyses revealed that the basal levels of CYP3A4, PXR (exon 5-6), PXR (exon 1a-2) and VDR mRNA were heterogeneous among the 6 cell lines examined (Figure [2](#F2){ref-type="fig"}). The 6 cell lines were then classified into two groups (high or low expression cells) based on the basal level of PXR and CYP3A4 mRNA, because the 6 cell lines always showed either high PXR and CYP3A4 expression or low PXR and CYP3A4 expression. The levels of CYP3A4 and PXR transcripts on high expression cells (LS180 and LoVo) were 7- to 35-fold and 40- to 5,000-fold higher than those on lower expression cells (Caco-2, HT29, HCT116 and SW48), respectively. These two groups also exhibited a difference in the basal level of the VDR transcript, although this difference was not marked (3.5- to 6-fold). There was strong correlation of the levels of the transcripts between PXR (exon 5-6) and PXR (exon 1a-2) throughout all analyses. ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Basal expression profile of colon cancer cell lines**. Basal levels of (a), PXR exon 5-6; (b), PXR exon 1a-2; (c), CYP3A4; and (d), VDR transcripts in LS180, LoVo, Caco-2, HT29, HCT116 and SW48 cells. The vertical axis indicates a relative transcript level (ratio to LS180 cells). The transcripts indicated by \* (PXR exon 5-6 transcript in SW48 cells and CYP3A4 transcript in HCT116 cells) were not detected and we estimated the minimum detectable levels among gradually diluted calibration samples (0.00257 and 0.0399, respectively). :::  ::: Increased mRNA expression by 5-aza-dC treatment ----------------------------------------------- In order to determine whether DNA methylation is involved in the transcriptional regulation of these genes, the 6 cell lines were treated with DNA demethylating agent (5-aza-dC). The treatment with 5-aza-dC induced a clear increase in CYP3A4 (28- to 116-fold) and PXR (3- to 10-fold) transcripts in a dose-dependent manner in the low expression cells, but not in the high expression cells (Figure [3](#F3){ref-type="fig"}). In particular, the CYP3A4 transcript of the low expression cells eventually reached the levels seen in the high expression cells by 5-aza-dC treatment. In contrast, 5-aza-dC had no marked effect on VDR expression in any of the cell lines (0.5- to 1.2-fold increase). These results suggested that DNA methylation is involved in the regulation of CYP3A4 and PXR, but not VDR, in the low expression cell lines. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Expression profile of colon cancer cell lines after 5-aza-dC treatment**. Levels of PXR, VDR and CYP3A4 transcripts in (a), LS180; (b), Caco-2; (c), HCT116; (d), LoVo; (e), HT29; and (f), SW48 cells. Cells were treated with 5-aza-2\'-deoxycytidine (0, 0.5 or 5 μM). The vertical axis indicates a relative transcript level (ratio to cells without 5-aza-dC treatment). Numbers *in italics*indicate relative transcript levels that were over the maximum scale on the vertical axis. The transcripts indicated by \* (PXR exon 5-6 transcript in SW48 cells and CYP3A4 transcript in HCT116 cells) were not detected and we estimated the minimum detectable levels among gradually diluted calibration samples. :::  ::: DNA methylation status of the colon cancer cell lines ----------------------------------------------------- The bisulfite direct sequencing detected no DNA methylation in the *VDR*or *PRMT1*promoter sequences in the 6 cell lines. Partial methylation of the *CYP3A4*5\'-distal region and full methylation of the *PXR*exon 3 region were equally observed among the 6 cell lines. Therefore, the different expression profiles of the two groups cannot be explained by methylation of these sequences. On the other hand, the CpG-rich sequence of the *PXR*promoter showed a different methylation status among the 6 cell lines. Interestingly, the degree of methylation of the *PXR*promoter (segments 1 and 2) in the high expression cells (LS180 and LoVo) was lower than that in the low expression cells (Caco-2, HT29, HCT116 and SW48) (Figure [4](#F4){ref-type="fig"}). Therefore, the details of the methylation status of the *PXR*promoter were estimated on individual DNA strands after subcloning (Figure [5](#F5){ref-type="fig"}). We found that an *HpyCH4IV*site within segment 2 was a suitable marker for the COBRA assay to assess *PXR*promoter methylation, because the degree of methylation of this site showed inverse correlation with the levels of CYP3A4 and PXR expression. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **DNA methylation profile of PXR as detected by direct sequencing**. Methylation status of the *PXR*gene in the 6 colon cancer cell lines and a cancerous tissue sample (255C) and its paired adjacent normal tissue sample (255N) detected by bisulfite direct sequencing. Open and closed circles represent unmethylated and fully methylated CpG sites, respectively. Half-closed circles represent partially methylated CpG sites. The methylation status of the CpG island (around exon 3) and CpG-rich promoter sequence (segments 1, 2, 3 and 4) are shown in the upper and lower panels, respectively. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Detailed DNA methylation profile of PXR segment 2**. Detailed methylation profile of the *PXR*gene in the 6 colon cancer cell lines and a cancerous tissue sample (255C) and its paired adjacent normal tissue sample (255N). Methylation of each CpG site was estimated by bisulfite sequencing on 7 or 8 individual DNA strands after subcloning. Open and closed circles represent unmethylated and methylated CpG sites, respectively. A \'TATA\' indicates a putative TATA box. A curved arrow indicates a transcription start site. \'*HpyCH4IV*\' indicates restriction site using the COBRA assay. :::  ::: The COBRA assay demonstrated that the treatment with 5-aza-dC resulted in decreased amounts of methylation of the *PXR*promoter in a dose-dependent manner in the low expression cells (Figure [6a](#F6){ref-type="fig"}). Therefore, the magnitude of methylation was likely to be associated with the decreased levels of *CYP3A4*and *PXR*gene expression in all 6 cell lines. These results suggest that the *PXR*gene was transcriptionally silenced by methylation of the promoter CpG sites and that the downregulation of the PXR protein resulted in decreased expression of the *CYP3A4*gene. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Methylation status of the *PXR*promoter and exon 3 regions, as detected by the COBRA assay**. Unmethylated and methylated DNAs are shown as U and M, respectively. PCR products that were not cut by the restriction enzymes are shown as UC. (a) *PXR*promoter methylation was examined in the 6 cell lines after treatment with 5-aza-2\'-deoxycytidine (0, 0.5 or 5 μM). (b) Methylation status of the CpG island of the *PXR*exon 3 region in 6 of the 18 primary colorectal cancers (numbers are those for particular cases). DNA samples from normal and cancerous tissue are shown as N and C, respectively. Note that a high degree of methylation was detected in cancerous, but not normal, tissue. (c) Methylation status of the CpG-rich sequence of the *PXR*promoter region in 8 primary colorectal cancers. Note that a lower degree of methylation was detected in cancerous tissue compared to normal tissue. :::  ::: The *PXR*promoter methylation was not associated with the profile of microsatellite instability (MSI) or other methylated genes (Table [2](#T2){ref-type="table"}). This suggested that altered *PXR*methylation was accumulated during colorectal tumorigenesis, independent of these genetic and epigenetic events. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Profile of Microsatellite instability (MSI), mismatch repair (MMR) deficiency and promoter methylation in 6 colon cancer cell lines. ::: ---------------------------------------------------------- MMR\ Methylation\*\* deficiency\* -------- ---- -------------- ----------------- --- --- --- LS180 \+ \+ \- M ? U LoVo \+ \- \+ U M U Caco-2 \- \- \- ? ? M HT29 \- \- \- U M M HCT116 \+ \+ \- U M M SW48 \+ \+ \- M M M ---------------------------------------------------------- Summary of previous studies \[[@B27],[@B28]\] and the present study (PXR methylation). \*Mismatch repair deficiency was associated with mutations or defects in mRNA transcripts. \*\*Methylation status was defined as unmethylated (U) or methylated (M). ::: DNA methylation status of colon cancer tissue samples ----------------------------------------------------- No or slight methylation of the *PXR*exon 3 region was detected in the normal colon tissue samples by direct sequencing and the COBRA assay. The levels of methylation in the cancer tissues were mostly higher than those in the paired adjacent normal tissues (Figure [6b](#F6){ref-type="fig"}). By contrast, the CpG-rich sequence of the *PXR*promoter was partially methylated in normal tissues, and the degree of methylation was decreased in the paired cancer tissues (Figure [6c](#F6){ref-type="fig"}). The decreased level of the *PXR*promoter methylation suggested increased expression of the *PXR*gene during colorectal carcinogenesis. There were no differences in the clinicopathological findings between the colorectal cancers with *PXR*methylation and those without methylation. Discussion ========== In the present study, 6 colon cancer cell lines showed heterogeneous mRNA expression profiles and were able to be classified into two groups with respect to their basal levels of the PXR/CYP3A4 transcripts (high expression cells, LS180 and LoVo; low expression cells Caco-2, HT29, HCT116, and SW48). These results are consistent with previous studies, in which LS180 and Caco-2 cells were characterized as PXR-sufficient and PXR-deficient cells, respectively \[[@B16],[@B17]\]. Genetic polymorphisms in the regions that regulate transcription are often a major cause of inter-individual variability in the levels of transcripts. However, such polymorphisms have not been frequently observed in the human *PXR*or *CYP3A4*genes, implying that certain epigenetic mechanisms are involved in the regulation of PXR and CYP3A4 expression. We found that the CpG-rich sequence within the *PXR*promoter region is methylated to different levels in high and low expression cells. Importantly, the magnitude of this promoter methylation was inversely associated with the levels of PXR and CYP3A4 expression. Furthermore, the levels of the PXR and CYP3A4 transcripts in low expression cells were mostly restored when DNA methylation was reversed by treatment with 5-aza-dC. Although this CpG-rich sequence did not strictly satisfy the criteria for a CpG island, the most affected CpG sites were located in a highly restricted region (segments 1 and 2) and these CpG sites were proximal to several putative transcription factor binding sites (such as Sp1 and hepatocyte nuclear factor 4 alpha) \[[@B18]-[@B20]\]. Therefore, *PXR*gene expression is most likely transcriptionally regulated by methylation of these promoter CpG sites. CYP3A4 is transactivated by functional interplays with VDR-RXRα or PXR-PRMT1 \[[@B20]-[@B22]\]. CpG-island methylation of the *VDR*or *PRMT1*promoter was not detected in these cell lines and the mRNA expression of VDR was not affected by 5-aza-dC treatment. These observations imply that DNA methylation of *PXR*, but not *VDR*or *PRMT1*, resulted in downregulation of the CYP3A4 mRNA in these colon cancer cells. It is still uncertain whether re-expression of the CYP3A4 by 5-aza-dC treatment was due to the re-expression of some other genes than *PXR*. However, *PXR*must be a candidate for methylation and reduced expression of the PXR by promoter methylation, even if partially, contributes to downregulation of the CYP3A4. Indeed, several studies demonstrated that selective downregulation of the PXR by siRNA reduces the basal level of the CYP3A4 transcripts in a dose-dependent manner \[[@B23]\]. CpG islands in the exon 3 region were fully methylated throughout the cancer cell lines and most cancer tissues. Even after treatment of the cell lines with 5-aza-dC, no increase in the PXR mRNA levels was observed in the high-PXR expressing cell lines, LS180 and LoVo. This strongly suggests that in human colon cancer cells, the methylated CpG islands in the exon 3 play a much less role in the epigenetic regulation of PXR, instead, promoter methylation plays a pivotal role in its regulation. In contrast, Misawa *et al*. previously demonstrated a distinct methylation profile of neuroblastoma cells, in which mRNA expression of the PXR splicing variant (exon 1a-2) was specifically regulated by the methylation of the exon 3 region rather than promoter methylation \[[@B9]\]. We found no marked difference in the levels of the PXR (exon 5-6) and PXR (exon 1a-2) transcripts in the colon cancer cells. Therefore, a tissue-specific DNA methylation profile is most likely involved in the transcriptional regulation of the *PXR*gene. DNA methylation of the *PXR*promoter was detected in only 1 of the 18 colorectal cancer tissue samples. The results reflect the genuine DNA methylation status, because we examined pure cancerous and normal epithelia using crypt isolation and directly compared the DNA methylation status between paired epithelia. Therefore, a low level of *PXR*promoter methylation, which was observed in the high expression cells, appears to be a common feature of colorectal cancers. We also demonstrated that the level of *PXR*promoter methylation is decreased during carcinogenesis, since paired adjacent normal tissues mostly showed higher levels of *PXR*promoter methylation. We could not directly compare DNA methylation status with the PXR mRNA expression, because crypt isolation provided ethanol-fixed epithelia and it was difficult to obtain fresh mRNA samples. However, most cancer tissues exhibited a pattern of promoter methylation quite similar to that observed in cultured cells with high expression (LS180 and LoVo) (Figures [5](#F5){ref-type="fig"} and [6c](#F6){ref-type="fig"}). Therefore, the association between promoter methylation and transcriptional silencing of the *PXR*gene is most likely applicable to primary colorectal cancers. As observed in the colon cancer cell lines, the decreased level of *PXR*promoter methylation most likely led to increased expression of PXR mRNA in the colorectal cancer tissues. These results are consistent with a recent study that showed strong expression of PXR mRNA in colon cancers, with great variability \[[@B24]\]. In contrast, Ouyang *et al*. found that PXR expression was lost or greatly diminished in many colon cancers using histochemical analysis \[[@B25]\]. Although the role of the altered PXR expression in colorectal carcinogenesis remains to be clarified, Zhou *et al*. demonstrated that PXR plays an antiapoptotic role in colon carcinogenesis by induction of multiple antiapoptotic genes \[[@B26]\]. We cannot rule out the possibility that alterations of the PXR methylation levels play direct roles in tumorigenesis, because certain oncogenes or tumor suppressor genes may be trascriptionally regulated by PXR. Partial methylation of the *PXR*observed in adjacent normal mucosa may be associated with \"field defect\" for carcinogenesis. However, numerous studies have demonstrated that ligand-binding activation or siRNA-mediated silencing of the PXR can affect the activity of metabolic enzymes including CYP3A4, without changes in the cell proliferation capacity. Therefore, we think that altered level of the *PXR*methylation does not provide a selective growth advantage during colorectal cancer progression. Interestingly, overexpression of PXR in the colorectal cancer tissue samples was correlated with an increase in UDP glucuronosyl transferases UGT1A1, UGT1A9 and UGT1A10, and led to a marked chemoresistance to the active metabolite of irinotecan (CPT-11) \[[@B24]\]. In addition, CYP3A4 and p-glycoprotein, which are transcriptionally activated by PXR, play important roles in intestinal first-pass metabolism and determine a drug\'s bioavailability. We hypothesized that PXR may play a key role in the colon cancer cell response to anticancer drugs by modulating expression of drug metabolizing enzymes and transporters including UGT1A, CYP3A4 and p-glycoprotein. Therefore, DNA methylation of the *PXR*promoter might be a good predictor of chemotherapy outcome and toxicity in colorectal cancers. Conclusions =========== *PXR*promoter methylation is involved in the regulation of intestinal PXR and CYP3A4 expression. This methylation might be associated with the inter-individual variability of the drug response of colon cancer cells. Abbreviations ============= PXR: pregnane X receptor; CYPs: cytochrome P450s; VDR: vitamin D receptor; PRMT1: protein arginine methyltrasferase 1; 5-aza-dC: 5-aza-2\'-deoxycytidine; COBRA: combined bisulfite restriction analysis Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= WH designed this study and carried out the cell culture, molecular genetic studies and drafted the manuscript. GT and JT participated in the data analysis. TS performed crypt isolation and pathological diagnosis. KO and GW performed the surgeries and obtained informed consent from the patients. SO participated in the design of this study and helped to draft the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/11/81/prepub> Acknowledgements ================ This work was supported by a Grant-in-Aid for Scientific Research (C) (21590402) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

PubMed Central

2024-06-05T04:04:19.624110

2011-2-22

PMC3053269

Background ========== Esophageal squamous cell carcinoma (ESCC), the major histological form of esophageal cancer, is one of the most aggressive malignancies with poor prognosis in the world, especially in the Northern part of China \[[@B1]\]. Like other types of solid tumors, the development of ESCC is also the accumulation of the abnormal expression of oncogenes and tumor suppressor genes (TSGs). Several genetic alterations have been associated with the development of ESCC including mutations of p53 and p16, amplification of cyclin D, c-myc, and EGFR, as well as allelic loss on chromosomes 3p, 5q, 8p, 9p, 9q, 13q, 17p, 18q, and 21q \[[@B2]-[@B5]\]. Our previous studies have characterized the common deletion regions at 3p and candidate TSGs within frequently deleted regions including *PLCD1*and *PCAF*\[[@B6],[@B7]\]. However, many genes associated with the development and progression of ESCC have not been characterized. To better understand the molecular mechanisms that underlie the ESCC development and progression, cDNA microarray was used to compare the gene expression profiles between 10 primary ESCC tumors and their paired non-tumorous tissues. Among the 185 up-regulated genes, one gene named *GPR39*drew our attention. GPR39 belongs to the G protein-coupled receptors (GPCRs) superfamily, which is the largest family of cell-surface molecules involved in signal transmission. It has been reported that GPR39 plays an important role in the regulation of gastrointestinal and metabolic function \[[@B8]\]. GPR39 receptor is now thought to be activated by Zn^2+^signals and may have other, as yet unidentified, cognitive ligands \[[@B9]\]. Moreover, GPR39 receptor also displays a strong ligand-independent signaling activity through Gα~12/13~as well as Gα~q~\[[@B10],[@B11]\]. A recent study suggests that overexpression of GPR39 may inhibit cell death induced by oxidative stress, endoplasmic reticulum (ER) stress, and activation of the caspase by Bax overexpression \[[@B12]\]. Emerging evidence indicates that G protein-coupled receptors are crucial players in cancer progression and metastasis \[[@B13],[@B14]\], however, the role of GPR39 in cancer development remains unclear. In this study, we studied GPR39 expression pattern in ESCC. The tumorigenic function of GPR39 was demonstrated by both *in vitro*and *in vivo*assays. The tumorigenic mechanism of GPR39 was also addressed. In addition, the clinical significance of GPR39 overexpression in ESCC was investigated. Methods ======= ESCC cell lines and specimens ----------------------------- Chinese ESCC cell line HKESC1 was kindly provided by Professor Srivastava (Department of Pathology, The University of Hong Kong, Hong Kong, China), and two Chinese ESCC cell lines (EC18 and EC109) were kindly provided by Professor Tsao (Department of Anatomy, The University of Hong Kong). Six Japanese ESCC cell lines (KYSE30, KYSE140, KYSE180, KYSE410, KYSE510 and KYSE520) \[[@B15]\] were obtained from DSMZ (Braunschweig, Germany), the German Resource Centre for Biological Material. Fifty pairs of primary ESCCs and their surrounding non-tumorous esophageal tissues were collected immediately after surgical resection at Linzhou Cancer Hospital (Henan, China). Samples used in this study were approved by the Committees for Ethical Review of Research involving Human Subjects at Zhengzhou University and Sun Yat-Sen University. Semiquantitative RT-PCR ----------------------- Total RNA was extracted from cell lines and frozen ESCC tissues using the Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacture\'s instruction. Reverse transcripation of total RNA (2 μg) was done using SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA), and cDNA was subjected to PCR for a 30-cycle amplification with primers for GPR39Fw: 5\'-GCCACCGGGGTCTCACTTGC-3\' and GPR39Rv: 5\'-GGCCGCAGCCATGATCCTCC-3\'. *GAPDH*(Fw: 5\'-CATGAGAAGTATGACAACAGCCT; Rv: 5\'-AGTCCTTCCACGATACCAAAGT) was used as an internal control. Tissue Microarrays (TMA) and Immunohistochemistry (IHC) ------------------------------------------------------- A total of 300 formalin-fixed and paraffin-embedded ESCC tumor specimens were kindly provided by Linzhou Cancer Hospital (Henan, China). TMAs containing 300 pairs of primary ESCC tumor samples and their corresponding nontumourous tissues were constructed as described previously \[[@B16]\]. Standard streptavidin-biotin-peroxidase complex method was used for IHC staining \[[@B16]\]. Briefly, TMA section was deparaffinized, blocked with 10% normal rabbit serum for 10 min, and incubated with rabbit anti-human GPR39 polyclonal antibody (Abcam, 1:100 dilution) overnight at 4°C. The TMA section was then incubated with biotinylated goat anti-rabbit immunoglobulin at a concentration of 1:100 at 37°C for 30 min. All of the IHC staining results were reviewed independently by two pathologists. Positive expression of GPR39 was defined as the brown staining in the cytoplasm. The staining results for GPR39 were scored semiquantitatively. Intensity was estimated in comparison to the control and scored as follows: 0, negative staining; 1, weak staining; 2, moderate staining; and 3, strong staining. Scores representing the percentage of tumor cells stained positive were as follows: 0, \<1% positive tumor cells; 1, 1-10%; 2, 10-50%; 3, 50-75%; and 4, \>75%. A final score was calculated by adding the scores for percentage and intensity, resulting in scores of 0 and 2-7. A score of 0 was considered negative; 2-3 was considered weak; 4-5 was considered moderate; and 6-7 was considered strong. For statistical analysis, 0-3 were counted as low expression of GPR39, while 4-7 were counted as overexpression of GPR39. Tumorigenic function of GPR39 ----------------------------- To test the tumorigenic function of GPR39, full-length GPR39 was PCR amplified, subcoloned into pcDNA3.1(+) vector (Invitrogen, Carlsbad, CA) and stably transfected into ESCC cell line KYSE30. Stable GPR39-expressing clones (GPR39-c1 and GPR39-c4) were selected for further study. Empty-vector transfected KYSE30 cells (Vec-30) were used as control. For foci formation assay, 1 × 10^3^GPR39-expressing cells or Vec-30 cells were seeded into 6-well plate. After 7 days culture, surviving colonies (\>50 cells/colony) were counted with 1% crystal violet staining. Triplicate independent experiments were performed. Colony formation in soft agar was performed by growing 1 × 10^4^cells in 0.4% Seaplague agar on a base of 0.6% agar in a 6-well plate. After 3 weeks, colonies consisted of more than 80 cells were counted and expressed as the means ± SD of triplicate within the same experiment. To perform cell growth assay, *GPR39*-expressing cells and control Vec-30 cells were seeded in 96-well plate at a density of 800 cells per well. The cell growth rate was measured using cell counting kit-8 kit (Dojindo, Japan) according to the manufacturer\'s instruction. Triplicate independent experiments were done. Flow cytometry assay -------------------- GPR39-c4 or Vec-30 cells were cultured in DMEM medium containing 10% FBS. Serum was withdraw from the culture medium when cells were 70% confluent. After 72 hrs, 10% FBS was added in the medium for an additional 8 hrs, Cells were fixed in 70% ethanol, stained with propidium iodide, and DNA content was analyzed by Cytomics FC (Beckman Coulter, Fullerton, CA). Tumor formation in nude mice ---------------------------- For *in vivo*experiment, stable GPR39-expressing KYSE30 cells or control Vec-30 cells (1 × 10^6^) in 200 μL serum-free DMEM (Life Technologies) were injected s.c. into the right and left flank of 4 week-old nude mice (5 mice for *GPR39*-c1 cells and 5 for *GPR39*-c4 cells), respectively. The tumor volume was calculated by the formula V = 0.5 × L × W^2^\[[@B17]\]. All experiments were done in accordance with institutional standard guidelines of Sun Yat-Sen University for animal experiments. Migration and invasion assays ----------------------------- For cell migration assay, GPR39-c4 cells or Vec-30 cells were grown to confluence and then mechanically scratched with a sterile pipette tip. Cells were rinsed with PBS and grown in culture medium for additional 24 hrs. The cell motility in terms of wound closure was measured by photographing at three random fields at time points 0 and 24 hr. For invasion assay, GPR39-c4 cells or Vec-30 cells were starved with serum free medium for 24 hrs before the assay. Cells (5 × 10^4^) were suspended in 0.5 ml serum-free medium and loaded on the upper compartment of invasion chamber coated with Matrigel (BD Biosciences). The lower compartment was filled with complete medium as chemoattractant. After 24 hrs, invasive cells were fixed, stained, and counted under a microscope. Triplicate independent experiments were done. F-actin staining ---------------- Cells grown on coverslips were washed three times in PBS, fixed in 4% paraformaldehyde for 20 min, and permeabilized with 0.1% Triton X-100 for 10 min. Cells were then stained with rhodamin-labeled phalloidin (Molecule Probes) in PBS containing 1% bovine serum albumin at room temperature for 30 min. After additional PBS washes, cells were counterstained with DAPI and photographed with a *Leica DMRA*fluorescence microscope (Rueil-Malmaison, France). RNA interference ---------------- Small interfering RNA (siRNA) (20 μM) against *GPR39*(s6073; Ambion) was transfected into KYSE180 cells in 6-well plates using Lipofectamine 2000 Reagent (Invitrogen) according to the manufacturer\'s instructions. At 48 hrs after transfection, the effects of gene silencing were measured via RT-PCR. Western blot analysis --------------------- Western blot analysis was performed with the standard method with antibodies to GPR39, N-cadherin and GAPDH (Abcam, Cambridge Science Park, Cambridge, UK), cyclin D1, p21, CDK4 and CDK6 (Cell Signalling Technology, Frankfurt, Germany), and E-cadherin (Santa Cruz Biotechnology, Santa Cruz, CA). Statistical analysis -------------------- Statistical analysis was performed with the SPSS standard version 16.0 (SPSS Inc., Chicago, IL). The relationship between the expression of GPR39 protein and clinicopathologic characteristics was assessed by χ^2^test. Results expressed as mean ± SD were analyzed using the Student *t*test. Differences were considered significant when *P*\< 0.05. Results ======= GPR39 is frequently overexpressed in ESCC ----------------------------------------- Semi-quantitative RT-PCR was used to study the expression status of GPR39 in 50 primary ESCCs and 9 ESCC cell lines. Compared with their paired non-tumorous tissues, overexpression of GPR39 was detected in 27/50 (54%) of primary ESCCs (Figure [1A](#F1){ref-type="fig"}). Overexpression of GPR39 was also frequently detected in ESCC cell lines (HKESC1, KYSE140, KYSE180, KYSE410, KYSE510 and KYSE520; Figure [1B](#F1){ref-type="fig"}). GPR39 expression in protein level was further studied in 300 primary ESCCs by IHC using a tissue microarray. Informative IHC results were obtained from 207 pairs of ESCCs. Non-informative samples included lost samples, unrepresentative samples, samples with too few tumor cells, and samples with inappropriate staining; such were not used in data complication. The expression of GPR39 in normal epithelial cells was always negative or weak whereas strong positive staining of GPR39 was observed in 121/207 (58.5%) of informative ESCCs (Figure [1C](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Overexpression of GPR39 in ESCC**. *GPR39*was frequently overexpressed in primary ESCCs (A) and ESCC cell lines (B) detected by RT-PCR. For primary ESCCs, expression of *GPR39*in tumor tissues (T) was compared with their paired non-tumorous tissues (N). Normal esophageal tissue was used as a normal control. 18S rRNA was used as an internal control. (C) Representative of GPR39 expression in a pair of ESCC (right) and adjacent normal tissue (left) detected by immunostaining with anti-GPR39 antibody (brown). The slide was counterstained with hematoxylin (original magnification × 200). :::  ::: Clinical significance of GPR39 overexpression in ESCC ----------------------------------------------------- The correlation between the GPR39 overexpression and clinicopathologic features of ESCC including age (≤60 versus \>60), gender (male versus female), tumor invasion (T stage: tumor depth; T3, T4 versus T1, T2), lymph nodes metastasis (N stage; N0 versus N1), TNM stage (I, IIa versus IIb, III-IV), was studied (Table [1](#T1){ref-type="table"}). The results showed that overexpression of GPR39 was significantly associated with lymph node metastasis (*P*= 0.008) and advanced clinical stage (*P*= 0.004). No correlation was observed between GPR39 overexpression and age (*P*= 0.735), gender (*P*= 0.887), tumor differentiation (*P*= 0.846) and tumor invasion (*P*= 0.085). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Association between GPR39 expression and clinical characteristics of ESCC patients (n = 207) ::: Clinicopathologic characteristics GPR39 expression no. (%) *P* ----------------------------------- -------------------------- ----------- --------- Age (y)  ≤60 69 (59.5) 47 (40.5) 0.735  \>60 52 (57.1) 39 (42.9) Sex  Male 68 (57.6) 50 (42.4) 0.887  Female 53 (59.6) 36 (40.4) Tumor location  Upper 22 (53.7) 19 (46.3) 0.762  Middle 82 (59.4) 56 (40.6)  Lower 16 (61.5) 10 (38.5) Tumor cell differentiation  Well 15 (53.6) 13 (46.4) 0.846  Moderate 76 (58.9) 53 (41.1)  Poor 30 (60.0) 20 (40.0) Tumor invasion (T)  T1 2 (25) 6 (75) 0.085  T2 44 (65.7) 23 (34.3)  T3 75 (57.3) 56 (42.7)  T4 1 (100) 0 (0) Lymph node metastasis (N)  N0 59 (50.4) 58 (49.8) 0.008\*  N1 62 (68.9) 28 (31.1) TNM stage  I 1 (14.3) 6 (85.7) 0.004\*  IIa 57 (52.3) 52 (47.7)  IIb 15 (75.0) 5 (25.0)  III-IV 48 (67.6) 23 (32.4) \* Statistically significant (*P*\< 0.05) ::: Tumorigenic function of GPR39 ----------------------------- To investigate the tumorigenic potential of *GPR39*, *GPR39*-expression vector was stably transfected into KYSE30 cells with silenced *GPR39*. GPR39 mRNA and protein expression in these clones were confirmed by RT-PCR and Western blot analysis (Figure [2A](#F2){ref-type="fig"}). The tumorigenic function of *GPR39*was assessed by both *in vitro*and *in vivo*assays including foci formation, colony formation in soft agar, cell growth rate assays and tumor xenograft experiment. Foci formation assay showed that the frequency of foci formation was significantly increased (*P*\< 0.01) in *GPR39*-transfectants compared with control cells (Figure [2B](#F2){ref-type="fig"}). A similar result was shown in soft agar assay (*P*\< 0.01, Figure [2C](#F2){ref-type="fig"}). Cell growth assay also revealed that the cell growth rates in *GPR39*-c1 and *GPR39*-c4 cells were significantly enhanced by *GPR39*compared with Vec-30 cells (*P*\< 0.01, Figure [2D](#F2){ref-type="fig"}). To further explore the *in vivo*tumorigenic ability of *GPR39*, tumor formation in nude mice was tested by injection of *GPR39*-c1 cells (n = 5) or *GPR39*-c4 cells (n = 5), whereas Vec-30 cells were used as controls. Tumor formation was observed in all tested animals. The results showed that the tumor growth curve of *GPR39*-overexpressing cells was significantly increased compared to Vec-30 cells (*P*\< 0.01, Figure [2E](#F2){ref-type="fig"}). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Tumorigenic function of GPR39 in ESCC cells**. (A) Expression of GPR39 in *GPR39*-transfected KYSE30 cells was confirmed by RT-PCR (left) and Western blot analysis (right). c1 and c4 are two independent *GPR39*-expressing clones. Vec-30 represents empty vector-transfected KYSE30 cells. (B) Representative of foci formation in monolayer culture. Quantitative analyses of foci numbers were shown in the right panel. Values were the mean ± SD of at least three independent experiments. \*\**P*\< 0.01; independent Student\'s *t*-test. (C) Representative of colony formation in soft agar. Percentage of colonies formed was summarized in the right panel. Values were the mean ± SD of at least three independent experiments. \*\**P*\< 0.01. (D) Growth curves of *GPR39*-expressing cells were compared with Vec-30 cells by cell growth assay. The results were expressed as mean ± SD of at least three independent experiments. \*\**P*\< 0.01. (E) Tumor growth curves of *GPR39*-expressing cells in nude mice were compared with Vec-30 cells by tumor xenograft experiment. The average tumor volume of *GPR39*-expressing cells vs Vec-30 cells was expressed as mean ± SD in 10 inoculated sites for each group of cells. \*\**P*\< 0.01. (F) Representative examples of tumors formed in nude mice following injection of *GPR39*-expressing KYSE30 cells (right) and Vec-30 cells (left). :::  ::: GPR39 promotes G~1~/S transition -------------------------------- To explore the mechanism underlying growth promotion by *GPR39*, the cell cycle distributions of *GPR39*-c4 and Vec-30 cells were determined by flow cytometry. Before treatment, the percentage of *GPR39*-c4 cells in G1 phase was obviously reduced in comparison with Vec-30 cells (38.37 ± 1.02% versus 45.87 ± 0.47%, *P*\< 0.05; Figure [3A](#F3){ref-type="fig"}). After 3 days\' serum starvation followed by addition of 10% serum for 8 hrs, the percentage of cells in S phase was significantly increased in *GPR39*-c4 cells compared to Vec-30 cells (26.43 ± 0.71% versus 8.97 ± 0.31%, *P*\< 0.05; Figure [3A](#F3){ref-type="fig"}), suggesting that GPR39 was able to promote G1/S transition. To reveal the potential molecular mechanism of *GPR39*in cell cycle promotion, expressions of several key cell cycle regulators including p21, cyclin D1, CDK4 and CDK6 were compared between *GPR39*-c4 and Vec-30 cells. Increased expression of cyclin D1 and CDK6, but not p21 and CDK4, were detected in *GPR39*-c4 (Figure [3B](#F3){ref-type="fig"}). ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **GPR39 promotes G1/S transition and enhances cell motility**. (A) DNA content between *GPR39*-expressing cells and control Vec-30 cell were compared by Flow-cytometry. Untreated, cells were cultured in DMEM medium with 10% FBS; Withdraw serum, cells were cultured in DMEM medium without serum for 3 days; Add serum, cells were cultured again in DMEM medium with 10% FBS for 8 hr. (B) Expression of p21, cyclin D1, CDK4, and CDK6 were compared between *GPR39*-expressing cells (c4) and control Vec-30 cells by Western blot analyses. GAPDH was used as loading control. (C) The effect of *GPR39*on cell migration was determined by wound-healing assay. During a period of 16 hr, the spreading speed of *GPR39*-expressing cells along the wound edge was faster than that in control Vec-30 cells. (D) Representative images showed the *GPR39*-expressing cells and Vec-30 cells that invaded through the matrigel. Number of invaded tumor cells was quantified in the right panel. *Columns*, mean of triplicate experiments; \*\**P*\< 0.01. :::  ::: GPR39 enhances cell motility and invasiveness of ESCCs ------------------------------------------------------ As the TMA result showed that overexpression of GPR39 was closely associated with ESCC metastasis, the effects of *GPR39*on cell migration and invasion were studied by wound-healing and cell invasion assays. Wound-healing assay showed that that the ectopic expression of *GPR39*could significantly increase cell migration ability in *GPR39*-transfected cells compared with empty-vector control (*P*\< 0.05, Figure [3C](#F3){ref-type="fig"}). Matrigel invasion assay also found that the ectopic expression of *GPR39*could significantly enhanced the invasiveness of ESCC cells, as demonstrated by a significant increase in the number of invaded cells (*P*\< 0.01, Figure [3D](#F3){ref-type="fig"}), in *GPR39*-transfected cells compared with empty-vector control. GPR39 induces partial epithelial-mesenchymal transition (EMT) ------------------------------------------------------------- In this study, we found that the cell morphology changed obviously after the transfection of *GPR39*. *GPR39*-transfected cells showed spindle shape and fibroblastic changes in monolayer culture, whereas empty vector-transfected cells, like KYSE30 parental cells, kept their cobblestone-like phenotype (Figure [4A](#F4){ref-type="fig"}). To determine whether the effect of *GPR39*on cell motility was associated with EMT, expressions of several epithelial markers (E-cadherin, N-cadherin) and mesenchymal markers (vimentin, and fibronectin) were compared between GPR39-c4 and Vec-30 cells by RT-PCR and Western blot analysis. The results showed that E-cadherin was obviously down-regulated in *GPR39*-c4 cells; however, no obvious difference was observed in the expression of N-cadherin, vimentin and fibronectin between *GPR39*-c4 and Vec-30 cells (Figure [4B](#F4){ref-type="fig"}). These findings indicated that *GPR39*increased cell motility was partially through the EMT. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **GPR39 promotes cell mobility and invasion by inducing partial EMT and remodeling cytoskeleton**. (A) Representatives of cell morphology of GPR39-expressing cells and Vec-30 cells (original magnification × 200). (B) Expressions of epithelial markers E-cadherin and mesenchymal markers fibronectin, N-cadherin, and vimentin, were compared by RT-PCR or Western blotting analysis between GPR39-expressing cells and Vec-30 cells. GAPDH was used as loading control. (C) Representative images of F-actin staining. Formation of lamellipodia (indicated by arrows) was stimulated by GPR39 compared to control cells (magnification × 400). :::  ::: Overexpression of GPR39 induced lamellipodia formation ------------------------------------------------------ To further explore the molecular mechanism of *GPR39*in regulating cancer invasion and metastasis, the role of *GPR39*in the polymerized actin was investigated by phalloidin staining. The results showed that GPR39-expressing cells exhibited enhanced lamellipodia formation compared with control cells (Figure [4C](#F4){ref-type="fig"}), indicating that *GPR39*could induce cytoskeleton remodeling to facilitate esophageal cancer cell migration and invasion. Silencing GPR39 expression by RNA interference (RNAi) ----------------------------------------------------- ESCC cell line KYSE180, which expresses a high level of endogenous *GPR39*, was used in the siRNA experiment. Two siRNAs targeting *GPR39*(GPR39-si1 and GPR39-si2) were tested and the efficiency of *GPR39*gene silencing was detected by RT-PCR. The result showed that the GPR39-si1 had a better silencing effect (Figure [5A](#F5){ref-type="fig"}). Silencing of *GPR39*resulted in a significant inhibition of the cell growth rate (*P*\< 0.01, Figure [5B](#F5){ref-type="fig"}) and migration (Figure [5C](#F5){ref-type="fig"}). DNA content analysis by flow cytometry showed that GPR39-si1 was able to inhibit the cell cycle at the G1/S checkpoint (Figure [5D](#F5){ref-type="fig"}). The percentage of cells in the S phase was significantly reduced in GPR39-si1-treated cells (27.23 ± 1.26%) compared with that in control-si-treated cells (35.13 ± 1.12%; *P*\< 0.05). These findings further supported that the tumorigenic function of *GPR39*was through its role in promoting cell proliferation and motility. ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Silencing of GPR39 expression suppresses tumorigenic ability of GPR39**. (A) GPR39 expression was efficiently decreased by the treatment of siGPR39 by RT-PCR. Relative expression level was measured by densitometer and summarized in the right panel. \*\**P*\< 0.01. (B) Growth curve of KYSE180 cells treated with GPR39 siRNA was compared with control siRNA treated cells by cell growth assay. \*\**P*\< 0.01. (C) Cell migration assay was used to compare the frequency of migratory cells between KYSE180 cells treated with control siRNA and GPR39 siRNA. (D) DNA content between control siRNA and GPR39 siRNA treated cells were compared by Flow-cytometry. \**P*\< 0.05. :::  ::: Discussion ========== Many G protein-coupled receptors (GPCRs) have been found to play critical roles in the development and progression of cancer, including malignant transformation \[[@B18],[@B19]\], tumor growth and survival \[[@B20],[@B21]\], as well as invasion and metastasis \[[@B22],[@B23]\]. Herein, we report that one of the G protein-coupled receptors, GPR39, is frequently overexpressed in human ESCC. To our knowledge, this is the first illustration that GPR39 contributes to the development and progression of ESCC. In the present study, the tumorigenic function of GPR39 was demonstrated by both in vitro and in vivo assays. Functional studies showed that GPR39 could effectively promote ESCC cancer cell growth, increase foci formation and colony formation and enhance tumor formation in nude mice. A recent study suggested that zinc could be a ligand capable of activating the GPR39 receptor \[[@B11]\]. Interestingly, zinc deficiency along with its associated increased cell proliferation can be tumorigenic in the rat esophagus \[[@B24],[@B25]\]. Our study also provided evidence that ectopic expression of GPR39 increased ESCC cancer cell growth, indicating involvement of the GPR39 receptor in the tumorigenesis of esophageal cancer. However, whether GPR39 signaling is activated by zinc in esophageal carcinogenesis needs to be further investigated. Further study revealed that overexpression of GPR39 in esophageal cancer cells KYSE30 promoted G1/S phase transition. We showed for the first time that GPR39 controls cell cycle progression through the activation of CDK6 and its activating protein, cyclin D1. G1/S phase transition is a major checkpoint for cell cycle progression and cyclin D1-CDK6 complex is one of the critical positive regulators during this transition \[[@B26],[@B27]\]. On the other hand, we found that silencing of GPR39 expression could inhibit tumorigenicity in KYSE180 cells through the cell cycle arrest at G1/S checkpoint. Another interesting finding of this study is the promoting effect of GPR39 on tumor metastasis in ESCC. Our data showed that overexpression of GPR39 could promote cell motility and invasiveness of ESCC cells *in vitro*. This mirrored the findings of GPR39 overexpression in human ESCC samples and its association with advanced clinical stage and lymph node metastasis of ESCC. Conversely, when we knocked down the endogenous GPR39 by RNAi in ESCC cells, the mobility of ESCC cells was significantly reduced, suggesting that GPR39 is closely involved in ESCC invasion and metastasis. Moreover, the observation of overexpression of GPR39 resulting in cell morphological alteration promoted us to further investigate its effect on EMT. We found that GPR39 has some impact on the EMT as shown by decreasing the epithelial molecule E-cadherin, an event critical in tumour invasion and a \'master\' regulator of EMT. E-cadherin provides a physical link among adjacent cells and is crucial for the establishment and maintenance of polarity and the structural integrity of epithelia. Indeed, due to the physical and functional link between E-cadherin based complexes and cytoskeletal components, a change in the E-cadherin mediated adhesiveness leads to rearrangement of the cytoskeleton \[[@B28]\]. In view of this, we further explored the role of GPR39 in reorganization of the actin cytoskeleton. As expected, our result showed that GPR39 led to significant alterations on cytoskeleton by inducing the lamellipodia formation in *GPR39*-transfected ESCC cells. This finding was consistent to previous studies that some G protein-coupled receptors (GPCRs) were able to promote actin reorganization and result in cell shape changes and enhanced cell migration \[[@B13],[@B29]\], indicating that GPR39 might directly alter the cytoskeleton to favor the tumor cell invasion and metastasis in ESCC. In this study, we have also provided evidence that targeting of *GPR39*with specific RNAi will reduce the oncogenic characteristics of ESCC tumor cells. To date, some G protein-coupled receptors (GPCRs) provide important practical options for preclinical research, clinical trials, and cancer treatment \[[@B30]\]. Therefore, consideration should be given to the development of novel therapeutics targeting GPR39 for use in GPR39-expressing ESCC tumors. Conclusions =========== In summary, our findings demonstrate that GPR39 plays an important role in ESCC development and progression via promoting cell proliferation, enhancing cell motility and invasiveness, regulating cytoskeleton and inducing EMT. A better understanding of the molecular mechanism of GPR39 in ESCC development and progression would provide novel therapeutic strategies to ESCC cancer patients. Abbreviations ============= EMT: epithelial mesenchymal transition; ESCC: esophageal squamous cell carcinoma; GPCR: G protein-coupled receptor; siRNA: small interfering RNA; TMA: tissue microarray; TSG: tumor suppressor gene; L: length; V: volume; W: width. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= FX and HL performed the experimental procedures with support from YZ, YQ, YD, TZ, LC, CN, TH and YL. FX, LF and XYG were responsible for experimental design, interpretation of the results and writing the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2407/11/86/prepub> Acknowledgements ================ This work was supported by Grants from National Natural Science Foundation of China (30772475, 30700820 and 30971606), Sun Yat-Sen University \"Hundred Talents Program\" (85000-3171311), Grant from the Major State Basic Research Program of China (2006CB910104), Research Fund for the Doctoral Program of Higher Education of China (20070558272) and Research Grant Council Central Allocation (HKUST 2/06C).

PubMed Central

2024-06-05T04:04:19.626719

2011-2-25

PMC3053306

Background ========== In rheumatoid arthritis (RA), bone damage on radiographs is visible as erosions and periarticular osteoporosis. Substantial data support that both erosions and osteoporosis in RA share a common cellular pathway which involves stimulation of the osteoclast. This osteoclast activation depends on stimulation from receptor activator of nuclear factor-κ ligand (RANKL) which binds to the receptor activator of nuclear factor-κ (RANK) on the osteoclast. The expression of RANKL is stimulated by pro-inflammatory cytokines (i.a. TNF-α, interleukin-1 (IL-1), IL-6 and IL-17). In addition recent data also suggest decreased osteoblast activation through the Wnt system \[[@B1]\]. In comparison to disease modifying anti-rheumatic drugs (DMARDs) including methotrexate (MTX), anti-TNF therapy has been shown to be superior in reducing the rate of both radiographic joint damage \[[@B2]-[@B4]\] and hand bone loss \[[@B5],[@B6]\]. Recently, the rate of radiographic joint progression was reported to be reduced independent of a patient\'s clinical response to anti-TNF therapy \[[@B7],[@B8]\]. This may suggest an additional positive effect of anti-TNF therapy on bone in RA independent of its anti-inflammatory effect. This has previously not been examined for periarticular bone loss. The objective of this study was to examine if treatment with the TNF-α inhibitor adalimumab also could reduce periarticular bone loss in RA patients independent of disease activity. Methods ======= The PREMIER study cohort was used to examine the relationship between periarticular bone loss and clinical response in RA patients treated with MTX and anti TNF-therapy. In this cohort, radiographic joint progression has recently been reported to be reduced independently of patients\' clinical responses to anti-TNF therapy with adalimumab \[[@B7]\]. The clinical, radiographic and bone density data from this 2-year, multi-centre, double-blind, randomised controlled study has previously been described in detail \[[@B6],[@B9]\] In short, the efficacy and safety of adalimumab plus MTX was compared with adalimumab monotherapy and with MTX monotherapy in 799 adult patients with early (\< 3 years, mean disease duration 9.1 month), aggressive RA (inclusion criteria: ≥8 swollen joint; erythrocyte sedimentation rate ≥28 or C-reactive protein (CRP) ≥1.5 mg/dl; erosions or rheumatoid factor positive), who previously had not been treated with MTX \[[@B9]\]. Digital X-ray radiogrammetry (DXR) (Sectra, Linköping, Sweden) was used to measure hand metacarpal cortical index (MCI) on the same digitised hand X-rays used for assessment of radiographic joint damage. DXR-MCI is defined as the combined cortical thickness divided by the bone width and is a relative bone measure independent of bone size and bone length \[[@B10],[@B11]\]. In the literature short-time in-vivo precision (CV%) has been reported to range from 0.31-0.64% for DXR-MCI \[[@B10],[@B12],[@B13]\]. DXR-BMD (def: *cxVPA~comb~x*(1-*p*), where c is a density constant, VPA is volume per area, and p is porosity) was intended to be the main outcome measure in this study. However, many radiographs could not be analysed for BMD because of unknown image resolution. The equation for DXR-BMD is based on volume per area and requires a known resolution. Thus, DXR-MCI, which is a relative measure less dependent of image resolution, was used as the primary outcome measure \[[@B6]\]. DXR-MCI has been shown to be highly correlated with hand bone mineral density, both measured as DXA hand and DXR-BMD \[[@B14]\]. Detailed information on the DXR-MCI measurement of the PREMIER RA cohort has previously been reported \[[@B6]\]. This sub analysis involved 188 patients from the MTX group and 214 patients from the combination group with available DXR data and disease activity measured by DAS28 at 52 weeks \[[@B15]\]. The combination group received adalimumab 40 mg subcutaneously every other week plus weekly oral MTX (rapidly increased to 20 mg/week), and the monotherapy group received weekly oral MTX plus placebo injections. Disease activity was assessed by the disease activity score using a 28 joint count (DAS28) at 52 weeks. Levels achieved for DAS28 scores were stratified according to the EULAR improvement criteria \[[@B16]\] for remission (DAS28 ≤ 2.60), low disease activity (DAS28 2.61-3.20), moderate disease activity (DAS28 3.21-5.10), and high disease activity (DAS28\> 5.10). DXR-MCI bone loss at 52 weeks follow- up were compared for patients in remission and low disease activity vs. moderate and high disease activity according to treatment groups, as well as for patients in remission, low, moderate and high disease activity. DXR-MCI loss dependent on inflammation, assessed by CRP during follow-up, was also analysed according to treatment groups. Mean CRP was calculated from values at baseline, 26 weeks and 52 weeks follow-up, and a mean CRP of \<10 was defined as low inflammation. Statistics ---------- Since the data were skewed, non-parametric analyses were conducted. No imputations were performed. Baseline values were compared between treatment groups with the Mann-Whitney method for continuous variables and the chi-squared method for categorical variables. DXR-MCI loss is described as negative values. Group analyses were performed with the Mann-Whitney method for two independent samples and Kruskall-Wallis for more than two independent samples. Spearman correlation analyses were conducted in an attempt to correlate changes in DXR-MCI with disease activity measured by DAS28. Analyses that came out with a p-value ≤ 0.05 were considered as statically significant. Study ethics ------------ The PREMIER study was approved a central institutional review board or independent ethics committee at each participating site approved the study, and all patients provided written informed consent \[[@B9]\]. Results ======= Baseline characteristics are shown in Table [1](#T1){ref-type="table"}. There were no statistically significant differences between the MTX and the combination group. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Baseline characteristics for the examined early rheumatoid arthritis patients. Results are given in mean (± standard deviation) for continuous variables and numbers (percentages) for categorical variables. ::: -------------------------------------------------------------------- Adalimumab +\ Methotrexate\ Methotrexate\ Monotherapy\ (N = 214) (N = 188) ------------------------------------ --------------- --------------- **Demographic characteristics** Age (years) 51.8 (14.3) 52.6 (13.2) Female % 152 (71.0) 137 (72.9) **Clinical characteristics** Disease duration (years) 0.7 (0.8) 0.8 (0.9) Previously taken DMARDs % 69 (32.2) 56 (29.8) Previously taken corticosteroids % 79 (36.9) 64 (34.0) Tender joint count (0-66) 30.5 (14.6) 31.9 (14.3) Swollen joint count (0-66) 21.4 (11.4) 22.5 (12.1) C-reactive protein (mg/l) 40.2 (42.4) 40.4 (40.9) HAQ (0-3) 1.5 (0.6) 1.5 (0.7) DAS28 6.3 (0.9) 6.3 (0.9) DXR-MCI (mg/cm^2^) 0.45 (0.09) 0.45 (0.09) -------------------------------------------------------------------- DMARDs = disease modifying anti-rheumatic drugs; HAQ = Health Assessment Questionnaire; DAS28 = 28-joint disease activity score; DXR = Digital X-ray radiogrammetry; MCI = metacarpal cortical index. ::: DXR MCI loss and DAS28 level ---------------------------- In the MTX group there were a significantly greater DXR-MCI loss in the patients with elevated disease activity as assessed by DAS28 than patients in remission and low disease activity, median (interquartile range) (-3.3% (-6.0% to -1.4%) vs. -2.2% (-4.4% to -0.7%), p = 0.01). In the combination group there were no differences in DXR-MCI loss between patients with moderate and high disease activity vs. patient in remission or with low disease activity (-1.9% (-4.7% to -0.7%) vs. -2.4% (-4.1 to -0.5%), p = 0.99), and the levels of loss were closer to those observed for patients in the MTX group in remission (Figure [1](#F1){ref-type="fig"}). ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **DXR-MCI loss in the methotrexate (MTX) plus adalimumab and MTX groups dependent on disease activity measured by DAS28**. :::  ::: Among the patients in remission or with low disease activity there were no difference between median DXR-MCI loss according to treatment groups (-2.4% in the combination group and -2.2% in the MTX group, p = 0.88). In the group who still had active disease, there was significantly greater median loss in the MTX group (-3.3% vs. -1.9%, p = 0.02) as compared to the combination group. The DXR-MCI loss across subgroups in remission or with low, moderate and high disease activity is shown in Table [2](#T2){ref-type="table"}. Among RA patients receiving MTX + adalimumab the bone loss was similar across the DAS28 subgroups (p = 0.97). For the MTX group there was a trend that the patients with high disease activity lost more DXR-MCI than patients with low disease activity (p = 0.10). The correlation (correlation coefficient r) between DAS28 and percentage DXR-MCI change was -0.14 (p = 0.06) in the MTX group and -0.07 (p = 0.33) for the combination group. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### DXR-MCI loss in the methotrexate plus adalimumab group and the MTX group stratified for disease activity measured by DAS28 ::: ---------------------------------------------------------------------------------------- Adalimumab + methotrexate Methotrexate ----------------------- --------------------------- --------------- ---- --------------- **Remission**\ 118 -2.43 (-2.73) 53 -2.15 (-3.03) **DAS28 ≤2.60** **Low**\ 36 -1.98 (-2.59) 30 -2.09 (-2.90) **disease activity**\ **DAS28 2.61-3.20** **Moderate**\ 51 -2.01 (-3.17) 81 -3.33 (-4.65) **disease activity**\ **DAS 28 3.21-5.10** **High**\ 9 -1.63 (-2.72) 24 -3.02 (-4.64) **disease activity**\ **DAS28\> 5.10** ---------------------------------------------------------------------------------------- DAS28 = 28-joint disease activity score; DXR = Digital X-ray radiogrammetry; MCI = metacarpal cortical index ::: DXR MCI loss and CRP level -------------------------- In the MTX group 184 patients had CRP measures at baseline, 26 and 52 weeks and in the combination group the number was 206. In the MTX group there were 57 patients with mean CRP \< 10 mg/L and 127 with CRP ≥10 mg/L. In the combination group the respective values were 86 and 120. The median DXR-MCI loss in the MTX group was significantly higher among patients with high CRP than low CRP (-3.1% (-6.0 to -1.4) vs. -1.9% (-3.9 to -0.2), p \<0.01). In the combination group there was no difference regarding median DXR-MCI loss dependent on CRP level (-2.4% (-4.6% to -0.7%) vs. -2.0% (-4.1% to 0.8%), p = 0.48), Figure [2](#F2){ref-type="fig"}. The patients with high CRP treated with MTX lost significant more DXR-MCI than patients treated with MTX and adalimumab (-3.1 vs. - 2.4, p = 0.02), while there was no difference in bone loss in patient with low CRP (- 1.9 vs. -2.0, p = 0.78). ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **DXR-MCI loss in the methotrexate (MTX) plus adalimumab and MTX groups according to CRP level achieved**. :::  ::: Discussion ========== The main finding in this study was that adalimumab in combination with MTX reduces hand bone loss independently of clinically assessed disease activity. This has previously been shown for radiographic joint damage \[[@B7],[@B8]\], but the present study is the first to show this for hand bone loss. This disconnection between inflammation and bone loss was not seen in MTX-treated patient. Patients with a poor clinical response or elevated CRP are at high risk of developing bone damage when treated with MTX monotherapy, while the bone will be protected independent of clinical response among users of MTX in combination with anti-TNF-α therapy. These results support the hypothesis that TNF-α can influence bone loss not only by stimulating RANKL by inflammation. The most probable additive mechanism is that TNF-α activates the osteoclast directly by binding to osteoclasts precursors through TNF-α receptor. Blocking of this receptor will down regulate the osteoclast formation \[[@B1],[@B17]\]. Another theoretical mechanism of TNF-α could be inflammation independent stimulation of RANKL, however this has not yet been clinically proven. The fact that TNF-α does have more than one way of stimulating the osteoclast may explain the positive effect on bone despite any clinical improvement \[[@B7],[@B8]\]. Despite that bone loss was independent of disease activity in the combination group, there were still a bone loss observed. This loss may have been a result of the patient selection in the PREMIER study. The included patients had high disease activity and poor prognosis in terms of bone damage as Rheumatoid Factor-positive and erosive disease \[[@B9],[@B18]\]. This is a retrospective study and we did not have any influence on the technical conditions regarding the radiographs. Due to difficulties analysing DXR-BMD the relative measure DXR-MCI was used, as described in detail previously \[[@B6]\]. However, DXR has improved the precision of MCI, and there is a strong correlation between DXR-BMD and DXR-MCI (r \> 0.9) \[[@B6],[@B14]\]. Another limitation was our inability to retrieve information on the use of bisphosphonates. This may be of importance as treatment with bisphosphonates increases DXR bone density \[[@B19]\]. Conclusion ========== We conclude that adalimumab in combination with MTX reduces hand bone loss independent of disease activity. This study highlights the bone protecting effect of anti-TNF-α therapy. List of abbreviations ===================== BMD: bone mineral density; CRP: C-reactive protein; DAS28: disease activity score based on 28 joint count; DMARDs: disease modifying anti-rheumatic drugs; DXR: digital X-ray radiogrammetry; ESR: erythrocyte sedimentation rate; HAQ: health assessment questionnaire; IL: interleukin; MCI: metacarpal cortical index; MTX: methotrexate; RA: rheumatoid arthritis; RANK: receptor activator of nuclear factor-κ; RANKL: receptor activator of nuclear factor-κ ligand; TNFα: tumour necrosis factor alpha Competing interests =================== M. Hoff, T. K. Kvien, A. Kavanaugh and G. Haugeberg have received consulting fees as speakers from Abbott Laboratories. Tore K. Kvien, A. Kavanaugh and G. Haugeberg have received founding for independent research from Abbott Laboratories. Aake Elden is employed by Abbott Laboratories. Johan Kälvesten is employed by Sectra. Authors\' contributions ======================= MH performed statistical analyses and prepared the manuscript. TKK contributed to statistical analysis and interpretation of results. JK analysed the radiographs. AE prepared the collection of radiographs. AK contributed to statistical analysis and interpretation of results. GH contributed to statistical analysis and interpretation of results and was the main responsible investigator of this project. All the authors substantially contributed to the manuscript. All authors read and approved the final manuscript. Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2474/12/54/prepub> Acknowledgements ================ We acknowledge Abbott Laboratories for financial support regarding collection of the radiographs and dxr-analysis. No other funding was received.

PubMed Central

2024-06-05T04:04:19.629480

2011-2-27

PMC3053307

Background ========== Back pain is a common complaint in pregnancy. The incidence of significant back pain during pregnancy varies widely in different populations from 20-60% \[[@B1]-[@B3]\]. While the causative factors of back pain during pregnancy are likely to be multi-factorial and heterogeneous, a positive association between bone loss and pregnancy related back and pelvic pain symptoms has been proposed. Decreased femoral bone density was associated with hip pain in the immediate postpartum period \[[@B4]\], while a greater fall in os calcis BMD has been shown to be correlated to back pain symptoms during pregnancy \[[@B5]\]. The incidence of persistent back pain symptoms after pregnancy varied from the disappearance of pain in over 60% within 2 days after delivery \[[@B6]\], to as high as 82% having persistent pain at 18 months in those with recurrent back pain from previous pregnancies \[[@B7]\]. An overall incidence of around 21% at 2 years after delivery has been reported \[[@B3]\]. Various risk factors have been ascribed to the persistence of pain after the index pregnancy, including history of back pain \[[@B8]\] or other epidemiological factors such as smoking or occupation \[[@B9]\], but the role of postpartum BMD loss or osteoporosis \[[@B10]\] still remains controversial. This study aims at observing the postpartum BMD changes in a longitudinal cohort of pregnant women who had reported significant back pain during pregnancy and correlating such changes with the incidence of persistent back pain symptoms 24 months or more after pregnancy using quantitative USG at the os calcis. The findings would help to evaluate whether the extent of recovery of the BMD loss that occurred during pregnancy would be protective against persistent back pain symptoms. Methods ======= Pregnancy Cohort ---------------- In the index pregnancy cohort, consecutive patients booked at a general obstetric clinic were prospectively recruited for the study over a twelve-month period. Routine antenatal care was offered in accordance with our service protocol. Written consent was obtained at the time of recruitment, and basic epidemiological data, including early pregnancy weight and height were recorded. Quantitative ultrasound bone density measurements were performed at the os calcis bilaterally at booking between 16 to 20 weeks, and in the third trimester between 36-38 weeks. All measurements were done using the Sahara Clinical Bone Sonometer system (Hologic), a waterless portable system that involved direct contact of the probe with the heel through elastomer pads and a specific ultrasound coupling gel. The patient was positioned as recommended by the manufacturer, seated in an upright position in a stable chair without wheels. The patient foot was stabilized using a specific foot guard to ensure that the focus between the ultrasound probes corresponded to the region of interest at the os calcis. All patients were allowed 30 minutes in the clinic before measurement to allow ambient skin temperature to be attained at the heel. Measurements were made bilaterally. The system measures BUA, SOS and then used these parameters to generate a simulated BMD. This computer calculated BMD was used in the calculations. The coefficient of variation of the system was quoted as 2-3% by the manufacturer and was in accordance with the data of the investigators using the system in previous similar studies. In addition to weight, body fat percentage assay was also performed in each of these occasions using a Tanita 500 bio-impedance system. Patients who had known medical conditions or who were on long-term medications known to affect bone density values, such as steroids or thyroid drugs, and those who delivered preterm before 36 weeks, were excluded from the final analysis. In addition, patients who have known chronic back pain that required regular medical follow-up or treatment, known spinal deformities and previous surgical intervention for back pain were also excluded. The patients were then surveyed for back pain symptoms during the index pregnancy by means of a structured self-administered questionnaire in the early postpartum period within the first 3 days before their discharge from hospital. Women who reported positive pain symptoms at any stage in her pregnancy were requested to fill in a pain distribution chart from recall. A visual analog scale of pain intensity was also provided to classify mild, moderate or severe back pain symptoms. Patients who reported at least moderate pain for more than 3 consecutive days within the pregnancy, or who required additional medical consultation, sick leave or formal treatment because of back pain during their pregnancy were considered as positive. Those who complained of only mild pain of transient durations were considered negative for back pain symptoms. The correlation of presence/absence of back pain in pregnancy and the interval BMD changes in pregnancy has previously been published \[[@B5]\]. Two Year Reassessment Cohort ---------------------------- All patients were resurveyed for back pain symptoms between 24-28 months after delivery using a mailed questionnaire, with a format similar to the early postnatal version. Those who reported at least moderate pain for more than 3 consecutive days within the past 6 months of the questionnaire, or who required medical consultation, sick leave or treatment because of back pain during the past 6 months were considered as positive. Those with only mild or transient symptoms that did not require medical consultation or treatment were regarded as negative. Those who already had further pregnancies at the time of the survey were excluded from the analysis. Patients who responded to the 24-28 month survey were invited to attend a special clinic session for repeat BMD measurements to compare with their pregnancy values. At this 2-year post delivery evaluation, the patients were given a standard interview to record their menstrual status, last menstrual dates and breast feeding status. Any other remarkable medical conditions or the need for long term medications were also noted. Care was taken to exclude the possibility of pregnancy in these subjects, and if confirmed, these were excluded from further investigations and analysis. Anthropometric and quantitative ultrasound measurements were performed using an identical protocol as the assessments during pregnancy. The current study cohort consisted of women who had reported positive back pain symptoms in their index pregnancy, and who completed the 2-year post delivery questionnaire and BMD assessment. Interval changes in body weight, body fat percentage, and os calcis BMD were calculated and correlated with the presence and absence of persistent back pain, as well as with previous pregnancy changes. A regression model was established to evaluate the inter-correlation of these parameters and persistent back pain symptoms. A p-value of \<0.05 was considered significant. Data was analyzed using the SPSS version 13.0 (SPSS, Chicago, IL, USA). The study was approved of by the Ethics Committee of the cluster hospital board. Results ======= Of 463 patients recruited in pregnancy in the original cohort, 230 (49.8%) reported one or more episodes of significant back pain during pregnancy. Of these, 143 (62%) with no further pregnancies completed the 24-28 months questionnaire follow-up survey, and persistence of significant back pain symptoms was seen in 33 (23.2%). Of those that completed the 2-year survey, 60 (41.9%) were available to complete the follow-up BMD assessment, which included 24 categorized as positive for persistent back pain, and 36 without further back pain symptoms (Figure [1](#F1){ref-type="fig"}). This final cohort of 60 women was used for further analysis in the current study. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Flow chart of women with or without persistent back pain included in the two-year longitudinal study**. PBK, persistent back pain; NBP, no persistent back pain. :::  ::: The mean loss in BMD from early to late gestation in the index pregnancy of this cohort was 0.373 g/cm^2^(SD 0.029), representing around 5% of the early pregnancy value. A marginally loss of BMD was seen when the 2-year post delivery measurement was compared with the early pregnancy value (- 0.022 g/cm^2^, SD 0.04). On the other hand, body weight, body fat percentage and body mass index significantly increased from early to late pregnancy, but fell again at the 2-year post delivery survey. A positive gain was seen in all these parameters when the 2-year post delivery measurement was compared with the early pregnancy value (Table [1](#T1){ref-type="table"}). ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### Changes in basic anthropometric parameters during pregnancy and 24-28 months after delivery (n = 60) ::: ----------------------------------------------------------------------------------------------------------------------------- Early Pregnancy\ Late Third Trimester\ Two years post delivery (SD) P-value by ANOVA (\< 20 weeks) (SD) (36-38 weeks) (SD) ------------------------------ -------------------- ----------------------- ------------------------------ ------------------ Weight (kg) 56 (7.38) 65.5 (7.67) 59.3 (7.43) \< 0.001 Body Mass Index (kg/cm^2^) 22.7 (2.84) 26.6 (2.79) 24.1 (2.91) \< 0.001 Body Fat Composition (%) 29.4 (4.75) 37.3 (4.52) 32 (4.93) \< 0.001 Mean os calcis BMD (g/cm^2^) 0.638 (0.13) 0.601 (0.116) 0.616 (0.123) \< 0.001 ----------------------------------------------------------------------------------------------------------------------------- SD = standard deviation ::: The final cohort was then divided into the group with significant persistent back pain (PBP group, n = 24)) and the group without further back pain (NBP group, n = 36) at the 24-28 months assessment. The PBP group had higher early pregnancy BMD (0.686 Vs 0.605 g/cm^2^, p = 0.02) but also had higher BMD loss during pregnancy (0.047 Vs 0.030 g/cm^2^, p = 0.03) when compared to the NBP group. The PBK group also had higher weight gain at 2 years (3.68 Vs 2.86 kg, p = 0.031), and a higher net loss of BMD (-0.025 Vs -0.007 g/cm^2^, p = 0.023) at the 24-28 month assessment when compared to early pregnancy values. There was no difference in the duration of lactation in the index pregnancy between the two groups (Table [2](#T2){ref-type="table"}). The NBP group appeared to have almost completely recovered their BMD loss in pregnancy, the BMD values at 2 years after delivery was nearly identical to the early pregnancy values (Table [2](#T2){ref-type="table"}). ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Comparison of anthropometric and BMD differences in those with or without persistent pain at 24-28 months ::: -------------------------------------------------------------------------------------------------------- PBP group\ NBP group\ p-value; MD\ (n = 24) (n = 36) (95% CI) ----------------------------------------------- ----------------- ----------------- -------------------- Age (years) 33.3 (3.14) 32.2 (2.82) 0.22; 0.95\ (-0.59 to 2.51) Height (cm) 157 (5.2) 156 (6.2) 0.67; 0.66\ (-2.43 to 3.75) Early pregnancy weight (kg) 54.5 (6.83) 57 (7.66) 0.21; -2.44\ (-6.32 to 1.42) Early pregnancy BMI (kg/m^2^) 22 (2.55) 23.2 (2.96) 0.11; -1.19\ (-2.67 to 0.28) Early pregnancy body fat percentage 28.7 (4.27) 29.8 (5.01) 0.66; -0.30\ (-3.53 to 1.5) Early pregnancy BMD (g/cm^2^) 0.686 (0.135) 0.605 (0.124) 0.021; - 0.081\ (0.001 to 0.12) Pregnancy weight gain (kg) 9.55 (2.35) 9.43 (3.37) 0.88; 0.11\ (-1.47 to 1.70) Pregnancy body fat accumulation (%) 7.75 (1.89) 8.05 (3.02) 0.66; -0.30\ (-1.69 to 1.08) Pregnancy BMD loss (g/cm^2^) 0.0472(0.018) 0.0306(0.033) 0.034; 0.0165\ (-0.031 to -0.013) Lactation duration in index pregnancy (weeks) 8.8 (4.5) 8.1 (3.6) 0.50; 0.72\ (-1.4 to 2.85) Weight change at 2 yrs post delivery (kg) 3.68 (1.84) 2.86 (1.02) 0.031; 0.82\ (0.076 to 1.57) Body fat change 2 yrs post delivery (%) 2.33 (1.34) 2.68 (1.85) 0.42; -0.35\ (-1.23 to 0.53) BMD change 2 yrs post delivery (g/cm^2^) -0.0257 (0.029) -0.0007 (0.028) 0.023; -0.017\ (-0.033 to-0.0025) -------------------------------------------------------------------------------------------------------- MD = mean difference CI = confidence interval ::: The correlation between the early BMD values and the 24-28 month post delivery values were highly significant and reliable (Pearson correlation coefficient 0.95, p \< 0.001 with 2-tailed analysis), as was the correlation between the late pregnancy BMD values and the 24-28 month post delivery value (Pearson correlation coefficient 0.97, p \< 0.001 with 2 -tailed analysis). A logistic regression model was constructed using the presence or absence of persistent back pain symptoms at the 24-28 months assessment as the dependent variable against all likely confounding continuous variables. Of the 4 variables found to be significant on univariate analysis, only weight gain at 2 years post delivery (p = 0.03) and BMD changes at 2 years post delivery (p = 0.03) remained in the equation, while the effects of early pregnancy BMD values and pregnancy BMD loss disappeared (Table [3](#T3){ref-type="table"}). In qualitative terms, more weight gain at 2 years after delivery are associated with persistent back pain, while a net positive balance in BMD at 2 years was protective against persistent back pain symptoms. ::: {#T3 .table-wrap} Table 3 ::: {.caption} ###### Stepwise logistic regression using persistence of significant post delivery back pain as dependent variable ::: Variable B **S.E**. Wald Significance Odds ratio 95% CI -------------------------------------- -------- ---------- ------- -------------- ------------ -------------- ***Significant variables*** Weight Gain at 2 years post delivery -0.653 0.303 4.62 0.03 1.92 1.05 to 3.48 BMD change at 2 years post delivery -21.9 10.26 4.55 0.03 0.11 0.01 to 0.96 ***Excluded variables*** Age 0.0529 0.113 0.203 0.65 1.05 0.84 to1.32 Early pregnancy BMI -0.406 0.238 2.90 0.08 0.66 0.41 to 1.06 Early pregnancy fat percentage 0.131 0.125 1.09 0.29 1.14 0.89 to 1.45 Early pregnancy BMD 0.208 3.944 0.002 0.95 1.23 0.03 to 5.7 Weight gain in pregnancy -0.031 0.129 0.058 0.80 0.96 0.75 to 1.24 Fat gain in pregnancy -0.253 0.211 1.434 0.23 0.77 0.51 to 1.17 Pregnancy BMD loss 25.26 16.97 2.21 0.13 0.37 0.06 to 8.91 Fat change at 2 years post delivery -0.406 0.267 2.32 0.13 0.67 0.39 to 1.13 CI = confidence interval ::: Discussion ========== The data presented in this study confirmed a demonstrable progressive fall in BMD at the os calcis as measured by quantitative ultrasound from early to late pregnancy. The mean decrease in BMD was around 5.5% of the early pregnancy BMD value, and was consistent with previous studies utilizing various means to measure BMD loss in pregnancy \[[@B11],[@B12]\], including quantitative ultrasound measurements using different \[[@B13],[@B14]\] or similar systems \[[@B5],[@B15],[@B16]\]. The incidence of back pain symptoms of around 50% that was found in the current study was in line with what was reported in the literature \[[@B1]-[@B3]\]. The incidence of persistent back pain symptoms of around 20% at 2 years was also compatible with the data in the literature \[[@B3],[@B17]-[@B19]\]. In addition, the current data was able to support an association between BMD loss during pregnancy, the degree of recovery of BMD loss after pregnancy and the persistence of back pain symptoms after the delivery. Previous studies evaluating the risk factors for persistence of back pain in pregnancy focused on history of back pain \[[@B3],[@B18]\], older age \[[@B3]\], younger age and higher weight \[[@B8]\], maternal smoking \[[@B9]\], the pattern of pain during the index pregnancy \[[@B20]\], as well as other psychosocial factors \[[@B8]\], but the role of postpartum BMD changes to persistence of back pain symptoms have yet to be studied in detail. Pregnancy has been documented to be a state of marked enhancement of bone turnover \[[@B21]\], during which a significant loss in BMD could be clearly demonstrated by direct methods, including standard dual-energy X-ray absorptiometry (DXA) \[[@B11],[@B22]\] and quantitative ultrasound \[[@B13]-[@B16]\]. This BMD loss is thought largely to be reversible in the long run \[[@B23],[@B24]\]. While the role of BMD changes in relation to back pain symptoms during pregnancy have been explored \[[@B5],[@B10]\], the longer term effects of such BMD loss in relation to persistence of back pain remains controversial. Previous studies have observed that a significant proportion of women who had documented back pain symptoms in pregnancy will be predisposed to have continued symptoms in subsequent years \[[@B18],[@B19]\]. Our findings of higher BMD loss during pregnancy and inability to attain complete recovery of this loss after 2 years in the group with persistent back pain would suggest a relationship between persistent BMD loss and persistent back pain. It would be of interest to evaluate whether the group of women who would continue to have severe back pain symptoms in later life would be more prone to develop clinical osteoporosis than those without. The possible mechanisms relating BMD loss and back pain remains elusive, as even in women with severe persistent back pain, the symptoms were only rarely associated with vertebral fractures or demonstrable radiological abnormalities \[[@B25]\]. Others would ascribe the back pain to biomechanical factors rather than to BMD loss \[[@B20]\], and that the BMD loss could theoretically be the result of immobilization or reduction in exercise levels because of persistent pain symptoms. However, quantitative BMD loss short of demonstrable fractures has also been associated with back, pelvic pain, as well as with hip pain symptoms \[[@B5],[@B10]\]. Lower BMD values during pregnancy have been associated with a higher incidence of back and pelvic pain symptoms \[[@B14]\]. An association between decreased femoral bone density or transient osteoporosis of the hip and hip pain during pregnancy and in the immediate postpartum period has been reported \[[@B26]-[@B28]\]. Thus, mild forms of pregnancy osteoporosis might pass undiagnosed clinically, but could be associated with pain symptoms. In addition, in this cohort, we have not studied calcium intake or vitamin D status and the impact of these parameters on BMD loss during or after pregnancy. Further studies to address these issues would help to explain the pathophysiology underlying BMD recovery after pregnancy and delivery and the relationship to back pain. In this cohort, we have observed that the persistent back pain group had higher early pregnancy BMD, but also higher BMD loss during pregnancy, as compared to those with no persistent pain. In our previous study \[[@B5]\], we have found that those with higher BMD loss in pregnancy actually tend to have higher BMD to start off with in early pregnancy, while those with borderline low BMD in early pregnancy apparently preserved BMD better and thus have lower BMD loss in pregnancy. As those with persistent back pain after pregnancy were also more likely to have higher BMD loss during pregnancy, epidemiologically, this could lead to the observation that those with persistent back pain having significantly higher early pregnancy BMD. There were certain limitations to this study. While we have been able to survey the incidence of persistent back pain symptoms in around 55% of the original cohort, we were able to obtain BMD findings in only around 26% of our original cohort that reported back pain in pregnancy (60/230). It could be seen that only around 33% of those without further pain were available for the follow-up BMD assessment, while up to 73% of those with further pain underwent the BMD assessment. However, secondary analysis of indicative parameters such as basic epidemiological characteristics, BMD loss in pregnancy and the incidence of persistent back pain between the group that completed the follow-up study and those that defaulted did not show any significant differences. We thus believe that the data of the group presented here should be representative of the entire cohort. In addition, while the relatively small sample size in the final cohort could be underpowered to show differences in secondary parameters such as body fat changes at two years follow-up, the current cohort was already able to show consistent and significant differences in primary outcome parameters such as BMD loss during pregnancy and at 2 years post delivery between the two groups. Quantitative ultrasound measurements of BMD have in general demonstrated good correlation with DXA measurements and are comparable to DXA in the prediction of clinical osteoporosis and fractures. Serial longitudinal comparisons could be affected by a relatively large coefficient of variation of 2-3% inherent in these quantitative ultrasound systems, particularly when the absolute difference to be measured is of magnitudes smaller than the coefficient of variation. However, as the magnitude of measurable BMD loss during pregnancy was substantial larger (5-7%) than the projected least significant change (LSC) that could be measured with the method, we believe that such measurements should be valid. Previous studies and our own data have demonstrated that the results of such quantitative ultrasound systems appear to be consistent and reproducible. \[[@B13]-[@B16]\]. In addition, when comparing the pregnancy BMD values and the 2-year follow-up values, we were able to find very high correlation coefficients despite the long time intervals between measurements This should be good evidence to support the reproducibility of such BMD measurements over time for any single individual. On the other hand, it could also be argued that after pregnancy, other measurement methods such as standard DXA for the axial skeleton or peripheral quantitative computerized tomography for the appendicular skeleton, which should have lower coefficients of variation, could provide more precise data. However, due to the theoretical risks of radiological exposure during pregnancy, such methods could only be used after delivery and direct correlation with data on BMD changes of the same skeletal site during pregnancy would not be possible. We have thus chosen to use the same method of measurement after pregnancy in order to compare directly with pregnancy values despite the limitations of such measurement methodology. Our data have shown that quantitative ultrasound is a viable method for monitoring the recovery of BMD after delivery to its pre- or early pregnancy states. Conclusions =========== In summary, the findings in this study supported a correlation of BMD loss as well as the degree of recovery of this loss as measured by quantitative ultrasound and persistent back pain symptoms in pregnancy. Future larger scale studies involving serial measurements of BMD at different skeletal sites using methods to correlate with persistent back pain symptoms should be warranted. The long term implications of the ability to recover the BMD loss in pregnancy in terms of menopausal bone health and risks of osteoporosis would also need to be explored. Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= Both authors conceived of the study, participated in the design and coordination of the study. Both authors participated in the data analysis and preparation of the manuscript. Both authors read and approved the final manuscript Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2474/12/55/prepub> Acknowledgements ================ We thank all participants who spent time and effort to complete the survey and the clinical assessments. We thank the nursing staff of the Postnatal Ward of United Christian Hospital for assisting with various parts of the study.

PubMed Central

2024-06-05T04:04:19.631130

2011-2-28

PMC3053308

Background ========== Best practice in health care should be guided by the results of research on the safety and effectiveness of different courses of clinical action. This evidence needs to be assembled, justified and presented in the form of health advice for multiple stakeholders including health professionals, decision makers and consumers of health care. Clinical practice guidelines are recognised as one of the best ways to present recommended courses of action based on research evidence, although recommendations are often presented inconsistently \[[@B1]\]. Where such evidence is not available, guidelines may use consensus-based practice points and/or identify areas requiring further research. Both format and content can adversely affect the adoption and integration of guidelines into clinical practice \[[@B2]\]. The National Health and Medical Research Council (NHMRC) of Australia has been a world leader in developing and supporting the development of evidence-based health advice, including clinical practice guidelines. As early as 1999, the NHMRC commissioned and published \'Guidelines for Guideline Development\' \[[@B3]\], anticipating the need for a comprehensive set of resources to help guideline developers produce high quality guidelines. This was followed by a more detailed series of handbooks on different aspects of finding and reviewing clinical research \[[@B4]\]. Australian guideline developers must comply with NHMRC standards in order to gain NHMRC approval. These standards (such as rigorous evidence-based methods, multidisciplinary panels and public consultation processes) have resulted in NHMRC approved guidelines being of higher quality than those developed outside NHMRC processes \[[@B5]\]. By 2004, it had become clear that the NHMRC standards required expansion and revision in response to the rapid growth and diversification of clinical practice guidelines in Australia and elsewhere. There were two main areas where a need for revision was identified. The first was the need to develop a set of levels (or hierarchy) of evidence which would cover the different individual study designs used to address the different types of questions formulated by guideline development panels. This work (covering interventions, diagnostic accuracy, prognosis, aetiology and screening) is outlined in Merlin et al \[[@B6]\]. The second area was the need to develop a new system, or adaptation of an existing system, of formulating and grading recommendations for clinical practice guidelines that incorporated an assessment of the \'body of evidence\'. The concept of a body of evidence --------------------------------- Many guideline recommendations have been rated solely according to the level of evidence of the individual studies contributing to that recommendation. In the late 1990 s and early 2000 s, NHMRC prepared a series of handbooks to assist clinical practice guideline developers. These handbooks stated that other elements such as study quality, size and precision of study results, and relevance to local practice were also important \[[@B3],[@B4]\]. They did not, however, go as far as providing a transparent logical framework for assessing these elements when formulating recommendations. What was needed was a method for considering all of these elements across all of the research studies addressing the clinical question as a whole (the \'body of evidence\') like some other guideline development methodologies (such as those used by the Scottish Intercollegiate Guidelines Network or the National Institute for Health and Clinical Excellence). Recommendations based on the body of evidence could then be graded according to the degree of confidence that implementing the suggested course of action would lead to improved patient health outcomes. In recognition of this need, and in response to requests from methodological experts that consult for the NHMRC on guideline development (Guidelines Assessment Register \[GAR\] consultants) (see Appendix 1), the NHMRC undertook to revise and update its methodological approaches. This paper reports on the production and trial of a methodology and associated processes to assist Australian guideline developers in considering a body of evidence and grading the resulting guideline recommendations. Methods ======= In 2004, the NHMRC commissioned a review of existing frameworks for assessing evidence internationally \[[@B7]\]. This internal report provided a resource for a working party (comprising GAR consultants and NHMRC personnel - see Appendix 1 for members) to review existing practice, design and/or adapt a framework for grading a body of evidence and pilot this process with Australian guideline developers. The report identified nine possible systems for use in developing clinical practice guidelines. Of these, three were considered to be most useful for informing the development of an Australian guideline recommendation process. These frameworks were the Scottish Intercollegiate Guidelines Network (SIGN) system and considered judgement statement (SIGN50, revised 2008)\[[@B8]\]; the Strength of Recommendation Taxonomy (SORT) \[[@B9]\]; and the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) \[[@B10]\]. These systems were discussed at a face-to-face meeting of the working party with respect to their advantages and disadvantages and compatibility with the existing advice in the NHMRC \'Guidelines for Guideline Development\' handbooks. A consensus was reached about how these frameworks could be adapted in the new process. From the three systems, we combined elements to achieve our objectives, which were: to have a system that matched and complemented the current NHMRC evidence dimensions and documents as closely as possible; simplicity and clarity of approach; and to provide transparent method/s of formulating and documenting judgments to give a graded set of recommendations. The working party drafted a new framework for grading recommendations and this was refined by extensive email consultation and iteration within the group. The resulting draft framework was piloted by GAR consultants working with guideline developers between 2005 and 2009. There were five main methods to gather feedback: • Known experts in the international guideline field were approached by NHMRC directly for comment on the draft system - this was a formal request and responses were semi-structured in that the experts were free to review in their own style, • Key evidence-based assessment organisations in Australian and New Zealand were invited to register feedback on the website where the system was posted, • All guideline development groups working within the NHMRC endorsement framework during this period used FORM under the guidance of the GARs - they were all invited to offer feedback during and after the process. • The draft process was presented at key conferences and interactive workshops (eg International Cochrane Colloquium \[[@B11]\]), • The website was open for the 5 years (passive seeking) and included a structured feedback form. Following this initial period of consultation (up until 2007) the FORM\'s framework was further refined, taking account of the feedback received, and the public consultation period was extended to June 2009. During the development, trialing and refinement period from 2004 to 2009, the international guideline community continued to debate and evolve other systems of guideline production - these developments were monitored and helped to inform the Australian process. The revised version of FORM was subsequently endorsed by the Council of the NHMRC. Results ======= The new FORM framework was loosely based on the SIGN considered judgement form \[[@B8]\]. It provides guideline developers with a structured process for considering the whole body of evidence relevant to a particular clinical question, in the context of the setting in which it is to be applied. FORM recognises that ascribing a *level of evidence*to each study that reflects the risk of bias in its design, is only one small part of assessing evidence for a guideline recommendation. FORM provides a framework for assessing all the studies relevant for a recommendation against five criteria: the *evidence base*(i.e. number, level and risk of bias in included studies); the *consistency*of findings between studies; the *clinical impact*suggested by the evidence base; the *generalisability*of the results to the population for whom the guideline is intended; and the *applicability*of the results to the Australian (and/or local) health care setting. Under FORM, these five key components are individually assessed for each clinical question giving a picture of both the internal and external validity of the evidence base under consideration. Key components of FORM ---------------------- ### 1. Evidence base The evidence base is assessed in terms of the quantity and quality of the studies identified by a systematic literature review for the clinical question concerned (\'included studies\'). Study quality relates to an assessment of the risk of bias inherent in the conduct, design and reporting of results in the included studies. The guideline developers are free to choose the most relevant process or tool to assess risk of bias. To ensure that consideration is given to the full range of study designs required to assess the breadth of clinical questions in a guideline, the GAR consultants also developed levels of evidence to address different clinical questions (prognosis, diagnostic accuracy, aetiology etc). This has been comprehensively addressed by Merlin et al \[[@B6]\] (see also, NHMRC website: <http://www.nhmrc.gov.au/guidelines/developers.htm>) ### 2. Consistency The consistency component of the \'body of evidence\' assesses the extent to which the findings are consistent across the included studies (including across a range of study populations and study designs). This allows users to assess whether the results are likely to be replicable or only likely to occur under certain conditions. Consistency may be assessed where appropriate as statistical heterogeneity (applying an I-squared statistic for example) or more likely will require the users to make a judgment about the overall direction of effects across multiple studies with reference to clinical heterogeneity. Possible sources of inconsistency (heterogeneity) in the results of studies may be differences in the study design, the quality of the studies (risk of bias), the population studied, and varying definitions of outcomes being assessed. Should results differ for certain subpopulations, this could then be reflected in the development of the recommendation. ### 3. Clinical impact Clinical impact is a measure of the likely benefit that application of the guideline would have across the target population, and involves a clinical judgement. Factors that need to be taken into account when estimating clinical impact include: the relevance of the evidence to the clinical question; the statistical precision and size (and clinical importance) of the effect reported in the evidence-base; the relevance of the effect to patients, compared to other management options; the duration of therapy required to achieve the effect; and the balance of risks and benefits to the patient group, including potential harm. A hypothetical example of incorporating both clinical importance and potential harm may be for the use of statins in the control of dyslipidaemia where there is a very large body of evidence with low risk of bias indicating *a substantial*reduction in risk of cardiovascular events. In this case a qualifying recommendation could be made to differentiate the small group of people who may experience adverse events as a result of statin therapy. Clinical impact is arguably the most subjective of the five evidence components rated in the evidence statement. However, we have found in assisting many guideline development groups to produce clinical practice guidelines using the FORM process that it is often clearer for clinicians than it is for methodological experts. Clinicians seem to grasp the net benefit concept quite easily, although often robust discussions occur before a consensus is reached regarding the rating of this component. A strength of FORM is that these discussions contribute to formulating appropriate recommendations, and the final conclusion can be documented so that users of the guideline can see how the developers arrived at the recommendation. ### 4. Generalisability The assessment of generalisability involves determining how precisely the available body of evidence answers the clinical question that was asked. Issues to be considered include: how well the participants and settings of the included studies match the patient population being targeted by the guideline; the clinical setting where the recommendation will be implemented; and other factors such as the stage of the disease (e.g. early versus advanced), the duration of illness and (for diagnostic accuracy questions) the prevalence of the disease in the study population as compared to the target population for the guideline. ### 5. Applicability This component addresses whether the evidence base is relevant to the Australian health care system generally, or to more local settings for specific recommendations (such as rural areas or cities). Factors that may reduce the direct application of study findings to the Australian or more local settings include organisational factors (e.g. availability of trained staff, clinic time, specialised equipment, tests or other resources) and cultural factors (e.g. attitudes to health issues, including those that may affect compliance with the recommendation). The FORM Matrix and Evidence Statement Form ------------------------------------------- The FORM matrix forms part of the overall process which is detailed in Additional file [1](#S1){ref-type="supplementary-material"}. Each of the components in the FORM matrix can be rated from A to D. The body of evidence supporting a recommendation rarely consists of the same rating for each of the five components. There may be a large number of studies with a low risk of bias and consistent findings, but which have only a limited clinical impact, and are not directly generalisable to the target population or applicable to the local (e.g. Australian) healthcare context. Alternatively, a body of evidence may consist of one or two randomised trials with small sample sizes that have a moderate risk of bias but have a very large clinical impact and are directly applicable to the local healthcare context and target population. By rating each of the five components separately, FORM allows for this mixture of components, while still reflecting the overall body of evidence supporting a guideline recommendation. The FORM Matrix provides guidance for users about how to rate each component of the body of evidence (see Table [1](#T1){ref-type="table"}). The accompanying Evidence Statement Form is provided for guideline developers to complete for each clinical question with room for additional information and dissenting opinions to be recorded. A recommendation to answer the clinical question is developed in two stages. First, a rating is assigned for each of the five components described above and an evidence statement is written in passive voice to reflect the findings of the evidence base. Second, an overall recommendation or action statement is developed on the basis of the evidence statement and an overall grade is assigned to this recommendation that reflects the level of confidence in the evidence supporting the recommendation. ::: {#T1 .table-wrap} Table 1 ::: {.caption} ###### NHMRC Body of evidence matrix ::: ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- Component A\ B\ C\ D\ Excellent Good Satisfactory Poor ------------------ --------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- ---------------------------------------------------------------------------------------------------------------------------------------------------------- Evidence base^1^ One or more level I studies with a low risk of bias or several level II studies with a low risk of bias One or two level II studies with a low risk of bias or an SR/several level III studies with a low risk of bias One or two level III studies with a low risk of bias, or level I or II studies with a moderate risk of bias Level IV studies, or level 1 to II studies/SRs with a high risk of bias Consistency^2^ All studies consistent Most studies consistent and inconsistency may be explained Some inconsistency reflecting genuine uncertainty around clinical question Evidence is consistent Clinical impact Very large Substantial Moderate Slight or restricted Generalisability Population/s studied in body of evidence are the same as the target population in the guideline Population/s studied in the body of evidence are similar to the target population for the guideline Population/s studied in the body of evidence differ to the target population guideline but it is clinically sensible to apply this evidence to the target population^3^ Population/s studied in the body of evidence differ to the target population and hard to judge whether it is sensible to generalize to target population Applicability Directly applicable to Australian healthcare context Applicable to Australian health care context with few caveats Probably applicable Australian healthcare context with some caveats Not applicable to Australian healthcare context ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- SR = systematic review; several = more than two studies ^1^Level of evidence determined from the NHMRC Evidence Hierarchy ^2^If there is only one study, rank this component as \'not applicable\' ^3^For example, results in adults that are clinically sensible to apply to children OR psychosocial outcomes for one cancer that may be applicable to patients with another cancer ::: Evidence statements may be developed by outcome measures for each intervention and then the multiple evidence statements for a single question can be collapsed into a single recommendation. Guideline developers can produce a combined recommendation taking into account the balance of benefits and harms or separate recommendations for benefits and harms, if this is more appropriate. The FORM process allows considerable flexibility in developing the recommendation. The overall grades for recommendations should indicate the strength of the body of evidence underpinning the recommendation. This assists users of the clinical practice guidelines to make appropriate and informed clinical judgments. Grade A or B recommendations are generally based on a body of evidence that can be trusted to guide clinical practice, whereas Grade C or D recommendations must be applied carefully to individual clinical and organisational circumstances and should be interpreted with caution (see Table [2](#T2){ref-type="table"}). A recommendation cannot be graded A or B unless the evidence base and consistency of the evidence are both rated A or B. In some cases, lower-graded evidence statements may not provide sufficient confidence to support an evidence-based recommendation at all. However, the framework allows *Good Practice Points (GPP)*to be included when developers feel it is important to provide non-evidence-based guidance. ::: {#T2 .table-wrap} Table 2 ::: {.caption} ###### Definition of NHMRC grades of recommendations ::: Grade of recommendation Description ------------------------- ---------------------------------------------------------------------------------------------------------- A Body of evidence can be trusted to guide practice B Body of evidence can be trusted to guide practice in most situations C Body of evidence provides some support for recommendation(s) but care should be taken in its application D Body of evidence is weak and recommendation must be applied with caution ::: In formulating the recommendation users are advised to address the specific clinical question and to use action statements. The wording of the recommendation should reflect the strength of the body of evidence. Words such as \'must\' or \'should\' or \'use\' are included when the evidence underpinning the recommendation is strong, and words such as \'might\' or \'could\' or \'consider\' are used when the evidence base is weaker. The following recommendations illustrate these points and are taken from the NHMRC Clinical Practice Guidelines for the Management of Melanoma in Australia and New Zealand (NHMRC 2008). These show that the evidence base, consistency and impact were high for dermoscopy, but not so high for total body photography (also indicated by the use of the verb \'recommended\' in the first case and \'consider\' in the second): • Training and utilisation of dermoscopy is recommended for clinicians routinely examining pigmented skin lesions: Grade A; • Consider the use of baseline total body photography as a tool for the early detection of melanoma in patients who are at high risk for developing primary melanoma: Grade C (p xxii \[[@B12]\]). Developers are also asked to consider how the guideline will be implemented at the time that the guideline recommendations are being formulated. The Evidence Statement Form requests developers to consider whether: the recommendation will result in changes in usual care; there are any resource implications associated with implementing the recommendation; the implementation of the recommendation will require changes in the way care is currently organised; and the guideline development group are aware of any barriers to the implementation of the recommendation. This information is used to inform the implementation plan for the Guideline. Feedback, piloting and users\' experiences ------------------------------------------ Over the trial and consultancy period for the FORM grading process, we obtained feedback from invited experts (see acknowledgements), from current guideline developers and from the public. These issues and suggestions were carefully considered at the face-to-face meeting of the GAR consultants in 2007 (see methods). Where appropriate, we amended the FORM methodology and/or supporting documents to incorporate the suggestions or address problems. This iterative process ensured that the development of FORM was responsive to the needs of its core user group - guideline developers - and was as clear and comprehensible as possible, even for developers with limited methodological expertise. It also allowed the FORM development process to keep abreast of the sometimes rapidly changing methodology underpinning guideline development internationally and incorporate changes into FORM as appropriate. As developers of FORM and also methodological experts assisting guideline developers we (the authors) have been able to field-test the FORM process and gain first-hand feedback and direct experience about problems and issues encountered. This has been invaluable in modifying FORM to be more effective and useful. The following issues were identified in the first consultation and addressed in the second iteration of FORM where appropriate: • deciding between grades - but this has become easier with time and familiarity • determining and extracting relevant information from synthesised sources (such as existing systematic reviews) which are incompletely reported • insufficient funding, human resources and/or time for the rigorous systematic literature reviews needed to underpin the evidence statements • need to accommodate subjectivity in the interpretation of the components and the final recommendation/s In response to specific suggestions made in the first consultation period, we made the following modifications to the FORM supporting documentation: • revision of the notes, matrix and form to be more user friendly • the addition of \'explanatory notes\' sections for developers to document reasons for particular decisions within the matrix • the addition of a \'dissenting opinions\' and \'unresolved issues\' sections to the Evidence Statement Form to keep decision making transparent and informed • a flowchart to assist in navigation Feedback from the second stage of consultation showed that the modifications were a major improvement and that guideline developers agreed that the FORM system of grading was an improvement on the previous system where recommendations were \'graded\' according to the level of evidence from the NHMRC evidence hierarchy \[[@B3],[@B6]\]. They also reported that the framework offers an opportunity to develop guidelines that improve dissemination and uptake in clinical practice. With increasing familiarity users have found the framework fairly simple to use. As methodological experts assisting guideline developers, we have found the framework provides additional flexibility, especially when handling evidence with more than one outcome measure (for example overall survival, pain, readmission rates). Variable results/evidence statements for multiple outcomes can be captured by a single recommendation. Furthermore, the framework also allows a recommendation to be developed that balances the benefits and harms of an intervention (i.e. safety and effectiveness), but with enough flexibility to keep them separate if it is felt to be important. More than 20 NHMRC guidelines have now been completed using FORM. Discussion ========== The formulation and inclusion of recommendations is one of the defining differences between clinical practice guidelines and other evidence syntheses such as systematic reviews. A recent review of the adequacy of guideline recommendations has highlighted that over half of the recommendations (52.7%) give no indication of the strength of that recommendation \[[@B1]\]. The FORM process for formulation and grading of recommendations in clinical practice guidelines is logical, simple to use and intuitive. Its concurrent development with Australian levels of evidence \[[@B6]\] means that NHMRC can provide Australian (and other) guideline developers with an integrated framework for producing high-quality recommendations that represent best-practice and are implementable, acceptable and appropriate for the local health care system. The framework is also generic - the same processes can be used to formulate and grade recommendations for any type of clinical question, despite the differences in the type of evidence required to address that question (e.g. questions of diagnostic test accuracy, risk factors for disease progression or poor prognosis). Furthermore, health service providers can implement the evidence-based course of action with appropriate modification in light of the individual patient\'s values and preferences. In areas like public health where there may never be high-level evidence supporting the use of different interventions, practice recommendations developed using other grading systems would consistently rate a lower grade than is felt appropriate by experts in those fields. Examples of such areas include large-scale dietary questions, passive smoking or exposure to environmental chemicals. This does not occur using the FORM methodology. Using the NHMRC levels of evidence for aetiology questions as an alternative to the levels for intervention questions \[[@B6]\] allows the evidence base component of our grading system to be rated higher than would otherwise occur and this would be reflected in the overall grade of recommendation. The extensive pilot of FORM and subsequent uptake by both new and experienced guideline developers has shown that the framework is feasible and accepted. The component approach allows transparency in how recommendations are formulated, with users of the guidelines able to explicitly see the various contributions of factors such as quality of the evidence and clinical impact. A further strength is that implementation and resourcing issues are considered separately, which means that effective but potentially costly interventions are not penalised with a downgraded recommendation as the developers of this system felt that users\' willingness to pay will vary according to the context of use. Arguably the greater ability to differentiate strength of recommendation (four levels) in FORM offers more precision for developers. Limitations ----------- The UK National Institute for Health and Clinical Excellence (NICE) has decided to discontinue summary grades for recommendations, on the grounds that their previous grading system was being misinterpreted. They have stated that they are not sure that the GRADE system\'s approach to summary labels overcomes this \[[@B13]\]. We are not aware of this sort of misinterpretation occurring with FORM, and believe that the benefits of grading outweigh the harms as clinicians are striving for clear-cut health advice to assist with their individual decision-making. However, ongoing monitoring and periodic review of the application and use of FORM needs to be considered. Recommendation formulation and grading can be particularly challenging when the evidence is scant and/or poor, or conflicting. NICE has outlined some strategies to address these challenges, including using consensus when no evidence is found for a particular clinical question and highlighting gaps in the evidence where evidence is scant or poor.\[[@B14]\] NICE reminds us that whenever guidelines are unable to rely on a solid evidence base other methods used for formulating recommendations must be transparent and set out clearly in the guideline. A particular strength of an explicit process such as FORM is that the path from evidence to recommendation is made clear. Current evidence frameworks are grappling with how to integrate other forms of evidence needed to answer qualitative questions such as optimal quality of life, and we anticipate that FORM will need to be periodically reassessed in the light of international debate about levels of evidence and grading recommendations. The purpose of clinical practice guidelines is to change or guide health professionals\' behaviour and to improve quality of care. Therefore, the ultimate test of guidelines and the processes used to develop and implement guidelines will be improved health outcomes and improved systems. One way of facilitating this is by developing recommendations that are transparently produced through a process that is user-friendly, weighs up multiple concepts when formulating a course of action (much as the clinician does for an individual patient), and provides clear advice on the confidence or uncertainty associated with the recommended course of action. Conclusion ========== FORM provides a contemporary and internationally relevant structure within which clinical guideline developers can consider current literature related to specific clinical questions. It has been developed through a unique partnership of government, academic, private consultancy and clinical personnel with considerable experience in evidence-based practice and development of clinical practice guidelines. Our work with over 20 guideline developers during the piloting of the FORM process has demonstrated it to be a logical, simple to use and intuitive system for formulating and grading recommendations in clinical practice guidelines. Competing interests =================== The authors declare that they have no competing interest. Meeting attendance fees were paid to the authors (or their institutions) by the NHMRC - a not-for-profit research organisation funded by the Australian Government. Authors\' contributions ======================= All authors were involved in the process of drafting and revising the grading process as described and all authors were involved in the preparation and final approval of the current manuscript. Appendix 1 ========== History of NHMRC Guidelines Assessment Register (GAR) and members of the Levels and Grades Working Party In 2002, the NHMRC convened a register of methodological experts (Guidelines Assessment Register \[GAR\]) to assist external guideline developers in Australia through the process of identifying and synthesising evidence for guidelines in a way that complied with NHMRC specified requirements and would assist them in gaining NHMRC endorsement for their work. The main role of the GAR consultants was to oversee the methodological processes in external development of guidelines, particularly reviewing and classifying the quality of the evidence, and how these classifications correlated to the resultant recommendations. The expected outcome of the involvement of the GAR consultants was that consistently high quality guidelines would be submitted to HAC for approval, and that problems identified *post hoc*in guideline development could be pre-empted. Kristina Coleman, Sarah Norris, Adele Weston (Health Technology Analysts Pty Ltd) Karen Grimmer-Somers, Susan Hillier (iCentre for Allied Health Evidence, University of South Australia) Tracy Merlin (Adelaide Health Technology Assessment, Discipline of Public Health, University of Adelaide) Philippa Middleton, Rebecca Tooher (ARCH, University of Adelaide) Janet Salisbury (Biotext Pty Ltd) Pre-publication history ======================= The pre-publication history for this paper can be accessed here: <http://www.biomedcentral.com/1471-2288/11/23/prepub> Supplementary Material ====================== ::: {.caption} ###### Additional file 1 **NHMRC Evidence Statement *(landscape Document)***. This file includes the FORM documents discussed - the matrix and trigger questions for guideline developers to complete the Evidence Statement. ::: ::: {.caption} ###### Click here for file ::: Acknowledgements ================ Feedback on levels/grades has been provided during the development phase from: Marita Broadstock - New Zealand Health Technology Assessment Suzanne Dyer - NHMRC Clinical Trials Centre Paul Glasziou - Oxford University, United Kingdom Sally Green - Australasian Cochrane Centre Brian Haynes - McMaster University, Canada Paul Ireland - National Institute of Clinical Studies Nicki Jackson -Deakin University Sally Lord and Les Irwig - University of Sydney Skye Newton and Janet Hiller - University of Adelaide Andrew Oxman - Oslo, Norway (GRADE Working Group) Janine Keough and Vesna Cvjeticanin (NHMRC representatives)

PubMed Central

2024-06-05T04:04:19.633793

2011-2-28

PMC3053309

Background ========== EGb761 is a standard extract from the leaves of *Ginkgo biloba*(*Yinxing*) containing 24% ginkgo-flavone glycosides (*eg*kaempferol, quercetin and isorhamnetin derivatives) and 6% terpenoid (*eg*ginkgolides A, B, C, J and bilobalide) \[[@B1]\]. The cardioprotective effects of EGb761 have been demonstrated in various *in vivo*and *in vitro*animal models and humans. The flavonoid components of EGb761 scavenge superoxide, hydroxyl radicals and nitric oxide (NO) and protect myocardia from ischemia-reperfusion injury \[[@B2]-[@B5]\]. The terpenoid constituents of EGb761 also showed their cardioprotective effects independent from the free radical-scavenging properties \[[@B6]\]. Therefore, it is necessary to further elucidate whether the cardioprotective mechanisms of EGb761 are attributed to its flavonoids or terpenoid conponents in the prevention of myocardial ischemia-reperfusion injury. Mitochondria-dependent caspase-3 pathway is one of the critical signal pathways in apoptotic cell death during myocardial ischemia-reperfusion injury \[[@B7]\]. The mitochondrial apoptotic pathway plays a pivotal role in the apoptotic cell death \[[@B8]\]. The release of cytochrome c from mitochrondria in response to proapoptotic signals has been suggested as an initiating event in the apoptotic process \[[@B9]\]. Cytochrome c released from mitochondria is associated with apoptosis protease activating factor (Apaf-1) and pro-caspase-9, triggering the activation of caspase-3 and resulting in cell death \[[@B9]\]. Reactive oxygen species (ROS) generated from the ischemia-reperfused cardiomyocytes trigger apoptotic cell death \[[@B10]\]. ROS can lead to the release of cytochrome c and precursors of caspases from the mitochondria into the cytoplasm \[[@B11]\]. Ischemia-reperfusion initiates the release of cytochrome c within minutes and the program of apoptotic cell death within hours in the cardiomyocytes \[[@B12]\]. Therefore, antioxidant therapy targeting the mitochondrial apoptotic pathway may be an important strategy in the treatment of myocardial ischemia-reperfusion injury. Recent studies show that EGb761can effectively and extensively counteract the cardial toxicity of doxorubicin *via*preventing the activation of the p53-mediated, mitochondrion-dependent apoptotic signaling pathway \[[@B13],[@B14]\]. EGb761 also protects against mitochondrial dysfunction in platelets and hippocampi in ovariectomized rats \[[@B15]\]. However, it is unclear yet whether EGb761 can regulate mitochondria-dependent caspases pathway in ischemia-reperfused cardiomyocytes. In this study, we selected two representative constituents of EGb761, namely quercetin and bilobalide, and compared their effects on the release of cytochrome c from mitochondria, the expression of caspase 3, the cleavage activities of caspases and apoptotic cell death. Methods ======= Cell culture and drug treatment ------------------------------- Neonatal Wistar rat cardiac myocytes were isolated and cultured according to the method described in our previous report \[[@B16]\]. The rats were obtained from the Laboratory Animal Unit of the University of Hong Kong. Animal housing, care and application of experimental procedures were all in accordance with the institutional guidelines and approved by the University Committee on the Use of Live Animals in Teaching and Research for the University of Hong Kong. Briefly, after anesthesia, the hearts from Wistar rats (aged 2-3-days) were minced and dissociated with 0.06% trypsin (Sigma, USA). The dispersed cells were incubated on 100 mm culture dishes for 15 minutes at 37°C with 100% relative humidity in a CO~2~incubator. Non-attached viable cells were collected and incubated in Dulbecco\'s modified Eagle\'s medium (DMEM, Invitrogen, USA) supplemented with 10% fetal calf serum (FCS), penicillin (50 U/ml) and streptomycin (50 μg/ml) for six hours, followed by incubation in the same media supplemented with 10^-6^mol/L cytosine arabinoside (Ara C, Sigma, USA) for 48 hours to reduce the rate of non-myocytes. The purity of the isolated cardiomyocytes was identified to be more than 90%. The isolated cardiomyocytes were cultured in high glucose DMEM supplemented with 10% heat-inactivated FCS, 1% glutamine, 50 U/ml penicillin and 50 μg/ml streptomycin and incubated at 37°C with 100% relative humidity in a CO~2~incubator for three days. To mimic ischemia, we placed the medium with low glucose DMEM supplemented with 1% FCS, 1% glutamine, 50 U/ml penicillin and 50 μg/ml streptomycin. The cardiomyocytes were incubated at 37°C in an air-tight incubator from which normal air was removed by a vacuum pump (GAST DOA-P184-BN, USA) and replaced by 1% O~2~/5% CO~2~balanced by 95% N~2~. The cardiomyocytes were cultured under hypoxia for twenty-four hours. To mimic reperfusion, we placed the medium with the high glucose medium supplemented with 10% FCS again and replaced the gas with 95% air/5% CO~2~for four hours. Corresponding control cells were incubated under the same conditions but perfused with air (ambient). Prior to hypoxic treatment, EGb761 (1, 10, 100 μg/ml, IPSEN Institute, France) and its components quercetin and bilobalide (1, 10, 100 μg/ml, Sigma, USA) were added respectively into the media. To confirm whether the antioxidant therapy reduced the release of cytochrome c from the cardiomyocytes, prior to hypoxia treatment we added MnTMPyP (5 μM; A G Scientific, USA), a superoxide dismutase (SOD) mimic, to the culture medium to dismutate cellular superoxide. Electrophoretic analysis of DNA fragmentation --------------------------------------------- After hypoxia and reoxygenation treatment, the cells were washed twice with phosphate buffered saline (PBS), precipitated by centrifugation and incubated with lysis buffer \[0.2M Tris-HC1, pH8.0; 0.1 M sodium ethylenediaminetetracetic (Na~2~EDTA); 1% SDS and 100 mg/L proteinase K\] for four hours at 55°C. The nuclear lysates were extracted twice with phenol and then extracted with an equal volume of phenol-chloroform-isoamyl alcohol (25:24:1). DNA was precipitated with 5 mol/L sodium chloride and ethanol overnight at -20°C and spun down at 6,000 *g*with Beckman Coulter Avanti-JE Centrifuge (USA) for 10 minutes at 4°C. The DNA was incubated with Tris-DETA buffer (10 mM Tris-HC1, pH7.5; 1 mM Na~2~EDTA) containing 20 mg/L RNAse A for one hour at 37°C and finally extracted twice with an equal volume of phenol-chloroform-isoamyl alcohol (25:24:1). After washed with 75% ethanol, DNA samples were analyzed with electrophoresis on 1.5% agarose gels with TAE buffer (40 mM Tris-HC1, 20 mM acetic acid and 1 mM Na~2~EDTA). The gel was stained with 0.5 μg/ml ethidium bromide and photographed with UV transillumination. Detection of DNA fragmentation by ELISA --------------------------------------- Cellular DNA fragmentation was determined with the cell death detection ELISA reagents (Boehringer Mannheim, Germany) according to the manufacturer\'s instructions. The DNA fragmentation was expressed with the enrichment of histone-associated mono- and oligonucleosomes released into the cytoplasm. The enrichment factor (EF) was calculated with absorbance at 405 nm. Caspase cleavage assay ---------------------- Caspase activity was measured as previously described \[[@B17]\]. Briefly, cells were washed with PBS and suspended in 500 μl of lysis buffer containing 50 mM Tris-HCl pH7.4; 1 mM Na~2~EDTA; 10 mM sodium ethylene glycol tetraacetic (EGTA) and 10 mM digitonin at 4°C for 30 minutes. Lysates were centrifuged at 7200 g with Beckman Coulter Centrifuge (USA) for 10 minutes, and the supernatant containing 50 μg of protein was incubated with 50 μM Ac-DEVD-AMC (Calbiochem, USA) at 37°C for 30 minutes. The release of 7-amino-4-methylcoumarin (AMC) was measured in a fluorescence spectrophotometer (Hitachi F-3010, Japan; excitation 380 nm; emission 460 nm). The enzyme activity was expressed as fluorescent units per minute per milligram of protein. The protein concentration of the supernatant was determined by the Bradford method. Western blot analysis --------------------- Cytochrome c release from mitochondria into the cytosol was measured with Western blot analysis. Cells were harvested in buffer A containing 20 mM HEPES (pH7.5), 10 mM KCl, 1.5 mM MgCl2, 1.0 mM Na~2~EDTA, 1.0 mM EGTA, 1.0 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride and 250 mM sucrose, supplemented with protease inhibitors (10 μg/ml pepstatin A; 10 μg/ml leupeptin; 10 μg/ml aprotinin). After sat on ice for 15 minutes, the cells were homogenized and centrifuged at 1,000 *g*with Beckman Coulter Avanti-JE Centrifuge (USA) for 15 minutes at 4°C. Then the supernatants were centrifuged at 10,000 *g*(Beckman Coulter Avanti-JE Centrifuge, USA) for 15 minutes at 4°C, and the mitochondrial fractions (pellets) were resuspended with buffer A. The supernatants prepared at 10,000 *g*were further centrifuged at 100,000 *g*with Beckman Coulter Optima ™ L-100 XP Ultientrifuge (USA) for one hour at 4°C. The cytosol fractions (supernatants) and mitochondrial fractions were stored at -80°C. Proteins (25 μg) extracted from the cytosol and mitochondria were separated by 15% SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose membranes. Membranes were blocked with 5% nonfat dry milk in Tris-buffered saline containing 0.01% Tween 20 and incubated with mouse anti-cytochrome c monoclonal antibody (BD bioscience, USA). β-actin (1:2000, Sigma, USA) was used as internal control. Blots were washed, incubated with goat anti-mouse IgG conjugated to horseradish peroxidase and developed by incubation with enhanced chemiluminescence Western blot detection reagents (Amersham, USA). For the detection of caspase-3, the cells were lysed with buffer A. The lysed cells were centrifuged at 12 000 *g*with Beckman Coulter Avanti-JE Centrifuge (USA). Equal amounts of protein (40 μg) were used and Western blot analysis was performed following the same procedure as the detection of cytochorome c. The monoclonal antibody against caspase-3 (Calbiochem, USA) was used. β-actin (1:2000, Sigma, USA) was used as internal control. Detection of superoxide production ---------------------------------- Cardiomyocytes (2 × 10^5^cells) were plated on 24-well plates with a 12 mm glass coverslip precoated with poly-L-lysine (10 μg/mL). Hydroethidine (HEt, Polyscience), a reduced derivative of ethidium bromide, can cross cell membrane and be oxidized by superoxide (O~2~^-^) specifically, yielding red fluorescent ethidium bromide tightly binding to DNA. The intracellular superoxide production was measured from the fluorescence of HEt. After the hypoxia-reoxygenation experiments, the cells were incubated with 1 g/ml HEt for 15 minutes in dark at 37°C and then incubated for one hour in 4% (wt/vol) paraformaldehyde in PBS for fixation and washed with PBS. The cover slips were mounted onto glass slides and these slides were observed with a fluorescent microscope (Axioskop 2-plus, Zeiss, USA) at excitation wavelength between 510 nm and 550 nm and an emission wavelength at \>580 nm. Statistical analysis -------------------- All data were expressed as mean ± standard deviation (SD). One way ANOVA was used to assess the statistical significance of differences and followed by Student-Newman-Keuls (S-N-K) test for two group comparisons. SPSS 16.0 (IBM, USA) was used to perform all stastical analysis. *P*\< 0.05 was considered as statistically significant. Results ======= Effects of EGb761, quercetin and bilobalide on apoptotic cell death ------------------------------------------------------------------- We first investigated the inhibitory effects of EGb761, quercetin and bilobalide on apoptotic cell death in the hypoxia-reoxygenated cardiomyocytes using agarose electrophoretic analysis and cell death ELISA detection. As shown in Figure [1](#F1){ref-type="fig"}, the characteristic ladder pattern of the DNA fragmentation was found in the hypoxic cardiomyocytes and the DNA ladder became prodominant in the hypoxia-reoxygenated cardiomyocytes. Both EGb761 and quercetin showed inhibitory effects on the formation of the DNA fragmentation whereas bilobalide had no effect on the DNA fragmentation. We next quantitatively determined the apoptotic cell death with a cell death ELISA kit. The enrichment factor (EF) represents the enrichment of histone-associated mono- and oligo-nucleosomes released into the cytoplasm. The hypoxia treatment led to produce DNA fragmentation and the following reoxygenation treatment significantly increased the magnitude of DNA fragmentation (hypoxia: EF 1.75 ± 0.17; hypoxia-reoxygenation: EF 2.92 ± 0.21). As shown in Figure [2](#F2){ref-type="fig"}, EGb761, quercetin and bilobalide had inhibitory effects on the enrichment of DNA fragmentations in a dose-dependent manner. The inhibitory effects of EGb761 and quercetin were more pronounced than those of bilobalide. These results indicate that quercetin has better cardioprotective effects than bilobalide in the cardiomyocytes exposed to hypoxia-reoxygenation condition. ::: {#F1 .fig} Figure 1 ::: {.caption} ###### **Effects of EGb761, quecertin and bilobalide on DNA fragmentations in the hypoxia-reoxygenated cardiomyocytes**. DNA fragmentation was detected with agarose gel electrophoresis. (A) hypoxia-reoxygenation (HR); (B) HR + bilobalide (100 μg/ml); (C) HR + quercetin (100 μg/ml); (D) HR + EGb761 (100 μg/ml); (E) hypoxia (HO); (F) λ DNA/EcoRI+Hind III marker :::  ::: ::: {#F2 .fig} Figure 2 ::: {.caption} ###### **Effects of EGb761, quecertin and bilobalide on the enrichment factors (EF) of DNA fragmentations in the hypoxia-reoxygenated cardiomyocytes**. DNA fragmentation was determined with the cell death detection ELISA method. EF represented the enrichment of histone-associated mono- and oligonucleosomes released into the cytoplasm. Drug concentrations of EGb761, quecertin and bilobalide were 1, 10, 100 μg/ml respectively. Data are shown as mean ± SD (*n*= 5). :::  ::: Effects of EGb761, quercetin and bilobalide on the release of cytochrome c -------------------------------------------------------------------------- We then determined whether EGb761, quercetin and bilobalide could reduce the release of cytochrome c from mitochondria, an important apoptotic pathway triggered by caspases in the hypoxia-reoxygenated cardiomyocytes. Figure [3](#F3){ref-type="fig"} illustrates the results of immunoblot analysis on the expression of cytosol and mitochondrial cytochrome c. The expression of cytochrome c was predominantly distributed in the mitochondrial fractions instead of the cytosol fractions in the cardiomyocytes under normoxia condition; however, after the cardiomyocytes were exposed to hypoxia for 24 hours, the expression of cytochrome c in the cytosol fractions became predominant, indicating that the hypoxic treatment induces the release of cytochrome c from the mitochondria. Compared with the hypoxia group, the cardiomyocytes treated with twenty-four hours of hypoxia following four hours of reoxygenation had a significantly enhanced release of cytochrome c from the mitochondria. Importantly, the treatment of EGb761 markedly inhibited the release of cytochrome c from the mitochondria. We also compared the effects of EGb761 and its ingredients quercetin and bilobalide on the release of cytochrome c after the hypoxia-reoxygenation treatment. The results showed that EGb761 and quercetin had much better effects than bilobalide. In addition, as a positive control, MnTMPyP, a SOD mimic to dismutate cellular superoxide, was added to culture medium. Pre-treatment of MnTMPyP also inhibited the release of cytochrome c from the cardiomyocytes after the hypoxia-reoxygenation treatment. These results suggest that the inhibitory effects of EGb761 on the release of cytochrome c are mainly attributed to the antioxidant constitutions or their synergetic actions with terpenoid components in EGb761. ::: {#F3 .fig} Figure 3 ::: {.caption} ###### **Effects of EGb761, quecertin and bilobalide on the release of cytochrome c from mitochondria in the hypoxia-reoxygenated cardiomyocytes**. Cytochrome c was detected by Western blot analysis using a monoclonal antibody against cytochrome c. Anti-actin antibody was used as internal control. (A) effects of MnTMPyP (5 μM) on the release of cytochrome c in hypoxia-reoxygenated cardiomyocytes; (B) effects of EGb761, quecertin and bilobalide on the release of cytochrome c in hypoxia-reoxygenated cardiomyocytes. HO: Hypoxia; HR: hypoxia-reoxygenation. Drug concentration of EGb761, quecertin and bilobalide was 100 μg/ml. These representative data were obtained from three independent experiments. :::  ::: Effects of EGb761, quercetin and bilobalides on caspases activities ------------------------------------------------------------------- Next we investigated the cleavage activities of caspases and the expression of caspase-3 in the hypoxia-reoxyegenated cardiomyocytes. Ac-DEVD-AMC is a fluorogenic tetrapeptide substrate cleaved by caspases. The results showed that the Ac-DEVD-AMC cleaving activity was significantly elevated in the groups of both hypoxia treatment and hypoxia-reoxygenation treatment (control: 0.12 ± 0.04; hypoxia: 0.78 ± 0.06; hypoxia-reoxygenation: 1.02 ± 0.12 nmol/mg protein/min). Figure [4](#F4){ref-type="fig"} shows the effects of EGb761, quercetin and bilobalide on the cleavage activities of caspases. EGb761, quercetin and bilobalides decreased the Ac-DEVD-AMC cleaving activities in the hypoxia-reoxygenated cardiomyocytes in a dose-dependent manner. Both EGb761 and quercetin had stronger inhibitory effects on the Ac-DEVD-AMC cleaving activities than bilobalide. Moreover, the expression of caspase-3 was detected with Western blot analysis using anti-caspase-3 antibody (Figure [5](#F5){ref-type="fig"}). The expression of caspase-3 was clearly enhanced in both hypoxic and hypoxia-reoxygenated cardiomyocytes. EGb761, quercetin and bilobalide had dose-dependent inhibitory effects on the expression of caspase-3 in the hypoxia-reoxygenated cardiomyocytes. Importantly, EGb761 had stronger inhibitory effects than either quercetin or bilobalide alone, indicating that the synergetic action of flavonoids and terpenoids may contribute to inhibitions of EGb761 on the activations of caspase pathway in the hypoxia-reoxygenated cardiomyocytes. ::: {#F4 .fig} Figure 4 ::: {.caption} ###### **Effects of EGb761, quecertin and bilobalide on the expression of caspase-3 in the hypoxia-reoxygenated cardiomyocytes**. Caspase-3 was detected with Western blot analysis using the monoclonal antibody against caspase-3. Anti-actin antibody was used as internal control. HO: hypoxia; HR: hypoxia-reoxygenation. Drug concentrations of EGb761, quecertin and bilobalide were 1, 10 and 100 μg/ml respectively. These representative data were obtained from five independent experiments. :::  ::: ::: {#F5 .fig} Figure 5 ::: {.caption} ###### **Effects of EGb761, quecertin and bilobalide on the cleavage activities of caspases in the hypoxia-reoxygenated cardiomyocytes**. The extracted proteins from the cardiomyocytes were incubated with Ac-DEVD-AMC and the release of 7-amino-4-methylcoumarin (AMC) was measured. Drug concentrations of EGb761, quecertin and bilobalide were 1, 10 and 100 μg/ml respectively. HO: hypoxia; HR: hypoxia-reoxygenation. Data are shown as mean ± SD (*n*= 5). :::  ::: Effects of EGb761, quercetin and bilobalides on scavenging superoxide anions ---------------------------------------------------------------------------- We finally visualized the production of O~2~^-^in the cardiomyocytes using HEt-staining fluorescent imaging technology. As shown in Figure [6](#F6){ref-type="fig"}, hypoxia-reoxygenation treatment markedly induced the HEt fluorescence, indicating the production of O~2~^-^. Both EGb761 and quercetin significantly reduces the production of O~2~^-^in the hypoxia-reoxygenated cardiomyocytes; however, bilobalides had no significant effects on the production of O~2~^-^. Taken together, these results suggest that the antioxidant properties of EGb761 contribute to the anti-apoptotic effects on the hypoxia-reoxygenated cardiomyocytes. The anti-apoptotic effects of EGb761 are partially attributed to scavenge superoxide and inhibit the release of cytochrome c from the mitochondria in the hypoxia-reoxygenated cardiomyocytes. ::: {#F6 .fig} Figure 6 ::: {.caption} ###### **Effects of EGb761, quecertin and bilobalide on the production of O**~**2**~^-^**in the hypoxia-reoxygenated cardiomyocytes**. The production of O~2~^-^in the cardiomyocytes was visualized using HEt-staining fluorescent imaging technology. HO: hypoxia; HR: hypoxia-reoxygenation. The concentration of EGb761, quecertin and bilobalide was 100 μg/ml. These representative data were obtained from three independent experiments. :::  ::: Discussion ========== In the experiments of this study, hypoxia-reoxygenation induced the production of superoxide, the release of cytochrome c from mitochrondria, up-regulation of caspase-3, activated the cleavage activities of caspases and triggered apoptotic cell death in the cardiomyocytes. EGb761 exhibited cardioprotective effects on scavenging superoxide, inhibited the release of cytochrome c from the mitochrondria and the activation of the cleavage activities of caspases, and prevented apoptotic cell death in the hypoxia-reoxygenated cardiomyocytes. EGb761 and quercetin had stronger inhibitory effects than the terpenoid component bilobalide. MnTMPyP, a SOD mimic to dismutate cellular superoxide, also inhibited the release of cytochrome c from mitochondria in the cardiomyocytes after the hypoxia-reoxygenation treatment. Therefore, we hypothesize that the antioxidant components of EGb761 contribute to the cardioprotective effects of EGb761 *via*the regulation of the mitochondria-dependent caspases pathways. To our knowledge, the present study provides the first evidence that EGb761 and its antioxidant components have inhibitory effects on the release of cytochrome c from the mitochondria, activation of caspases and apoptotic cell death in cardiomyocytes under hypoxia-reoxygenation conditions. Exposure to hypoxia for 24 hours induced the occurrences of DNA fragmentation, a characteristic of apoptotic cell death. Treatment of hypoxia for 24 hours followed by four hours of reoxygenation further increased the magnitude of DNA fragmentation, suggesting reoxygenation exacerbates hypoxia-induced cell death. In the experiments, hypoxia was accompanied by the deprivation of glucose and serum whereas reoxygenation was co-supplied with glucose and serum, similar to the ischemia-reperfusion process. Our results obtained from the cultured cardiomyocytes are consistent with previous reports using various myocardial ischemia-reperfusion models \[[@B18],[@B19]\]. Cytochrome c is an essential component of mitochondrial respiratory chain and released from the mitochondria in response to various stimuli. Cytosolic cytochrome c activates caspase-9 cleavages, triggering the activation of caspase 3 and apoptosis \[[@B20]\]. To clarify the mitochondrial mechanisms of cell death in the ischemia-reperfused cardiomyocytes, we investigated the release of cytochrome c and the cleavage activation of caspases. The release of cytochrome c into cytosol was found in the cardiomyocytes exposed to hypoxia for twenty-four hours and low glucose and low serum; it became pronounced in the cardiomyocytes after four hours of reoxygenation. The expression of caspase-3 and the cleavage activation of caspases were concomitant with the release of cytochrome c, consistent with previous studies \[[@B21]\]. Acting as an inhibitor of several reactive oxygen species, EGb761 exhibits a wide spectrum of antioxidant activities \[[@B22]\]. Our previous studies and those of others suggest that the antioxidant properties of EGb761 contribute to the cardioprotective effects against myocardial ischemia-reperfusion injury \[[@B2],[@B3],[@B23]\]. In the present study, EGb761 inhibited the release of cytochrome c from mitochondria and the cleavage activities of caspases, down-regulated the expression of caspase-3 and prevented apoptotic cell death in the hypoxia-reoxygenated cardiomyocytes. EGb761 contains both antioxidant and non-antioxidant ingredients. Our data clearly showed that quercetin, an antioxant in EGb761, had stronger inhibitory effects than bilobalide on the release of cytochrome c, the activation of caspases and apoptotic cell death in the hypoxia-reoxygenated cardiomyocytes. Quercetin showed a similar inhibitory effect as EGb761, suggesting that ginkgo-flavone glycosides may be the major components of EGb761, contributing to its cardioprotective effects on inhibiting the release of cytochrome c and activation of caspases. These results are consistent with the report that flavonoid components exhibit better effects than Ginkgolides and Bilobalides on protections of mitochondrial membrane potential in NO-treated PC12 cells and mouse brain cells \[[@B24]\]. It is necessary to mention that bilobalide also inhibited the mitochrondria-dependent caspases activation and prevented the cardiomyocytes from apoptotic cell death even though the effects of bilobalide were much weaker than those of quercetin. A previous report suggested that ginkgolides, the non-flavon fraction of EGb761, protected PC12 cells against hypoxia-induced injury *via*p22/p44 MAPK pathway dependent elevation of HIF-1 transcription efficiency \[[@B25]\]. Therefore, the synergetic actions of flavone glycosides and terpenoids may contribute to the cardioprotective effects of EGb761 on inhibiting the release of cytochrome c and activation of caspases and preventing cell death during myocardial ischemia-reperfusion injury. Conclusion ========== The antioxidant constituents such as quercetin mainly contribute to the cardioprotective effects of EGb761 and inhibit the mitochondria-dependent caspase pathway. It is possible that the mitochondria-dependent caspase pathway may be one of the molecular targets of EGb761 against myocardial ischemia-reperfusion injury. Abbreviations ============= HO: hypoxia; HR: hypoxia-reoxygenation; ROS: Reactive oxygen species; NO: nitric oxide; Apaf: apoptosis protease activating factor; DMEM: Dulbecco\'s modified Eagle\'s medium; PBS: phosphate buffered saline; FCS: fetal calf serum; Het: hydroethidine; EF: enrichment factor; SOD: Superoxide dismutase; AMC: 7-amino-4-methylcoumarin; EGTA: sodium ethylene glycol tetraacetic; Na~2~EDTA: sodium ethylenediaminetetracetic; SD: standard deviation Competing interests =================== The authors declare that they have no competing interests. Authors\' contributions ======================= JS designed the study, performed the experiments, analyzed the data and drafted the manuscript. WL performed the experiments and analyzed the data. YG collected the background materials and drafted the manuscript. YT, PCWF and LT interpreted the results and revised the manuscript. All authors read and approved the final version of the manuscript. Acknowledgements ================ The work was supported by a grant from a Seed Funding of Basic Science, The University of Hong Kong and Dr B L Wong family donation. EGb761 (IPSEN Institute, France) was a gift from Professor Wenjuan Xin, Institute of Biophysics Academia Sinica, China

PubMed Central

2024-06-05T04:04:19.637478

2011-2-23

PMC3053310

Introduction {#s1} ============ Animal territoriality aims at excluding conspecifics from certain areas through the use of auditory, visual or olfactory signals as well as aggressive interactions [@pcbi.1002008-Brown1]. Its widespread occurence across so many different taxa has prompted the question as to whether general mechanisms for such behaviour exist [@pcbi.1002008-Stamps1], [@pcbi.1002008-Adams1]. Answering this question however has proved elusive, partly because territorial behaviour spans organizational levels from individual animals [@pcbi.1002008-Ydenberg1] to populations [@pcbi.1002008-LopezSepulcre1] and the ecosystem [@pcbi.1002008-Gompper1], but also because it requires an understanding of how conspecific avoidance processes observed at small spatial scales and short time scales generate extended and lasting territorial patterns. Such multi-scale dynamics are ubiquitous in ecology [@pcbi.1002008-Levin1] and their explanation often requires a *microscopic* level description of the processes at play. Here we provide such an approach by building a stochastic individual-based model (probabilistic cellular automata [@pcbi.1002008-Ilachinski1]) that reveals how territorial boundaries are formed, and change position, from the animal movements and interactions mediated through olfactory signals (scent marks). Although the importance of animal interactions in determining the shape and size of territories is now recognized [@pcbi.1002008-Adams1], none of the recent studies have attempted [@pcbi.1002008-Lewis1]--[@pcbi.1002008-Moorcroft1] or succeeded in quantifying [@pcbi.1002008-White2] *microscopically* how territories emerge from the movement and interaction between animals. These recent analyses in fact share the common feature of being *macroscopic* representations of the interaction processes since they do not account for the discreteness of the animal population and the interaction events [@pcbi.1002008-Durrett1], which occur locally and over short time periods. Key to our approach is in fact the recognition that conspecific avoidance, mediated through the deposition of olfactory signals, makes animal territories undergo the so-called *exclusion processes* [@pcbi.1002008-Harris1], [@pcbi.1002008-Liggett1] whose dynamics demand an individual level description. Here we explore the transient dynamics in the formation of animal territories, moving away from traditional approaches where territories are assumed to settle to a steady state [@pcbi.1002008-Lewis1]--[@pcbi.1002008-White2] and moving towards a mechanistic explanation derived from the individual animal\'s social interactions. By keeping track of the locations where each animal wanders over the terrain, we show how scent serves as a short-term cue, explaining why territorial mammals regularly renew and refresh their scent marks. We also show how to apply our findings by using data from a well-studied territorial species, the red fox (*Vulpes vulpes*), and extract information about the persistence of scent marks. Results {#s2} ======= We have used data from the so-called 'hinterland markers' [@pcbi.1002008-Macdonald1], [@pcbi.1002008-Gorman1] which, in our model, deposit scent marks throughout their territory as they move freely over a homogenous terrain as discrete time random walkers on a lattice [@pcbi.1002008-Okubo1] until they encounter a foreign scent mark, deposited by a neighbouring individual, from which they subsequently retreat in a random direction (see [Fig. 1](#pcbi-1002008-g001){ref-type="fig"} for a detailed visual illustration of the movement and interaction mechanisms). Occasionally, as a slight modification of the above model, a second version is used that incorporates a variable degree of correlation in the choice of the animal\'s direction, but only for a short time after the encounter of foreign scent. However unless otherwise stated, we assume the first model throughout the paper. ::: {#pcbi-1002008-g001 .fig} 10.1371/journal.pcbi.1002008.g001 Figure 1 ::: {.caption} ###### Graphical illustration of the scent-mediated avoidance interaction. This plot shows the possible movement of an animal inside its own territory and when it encounters a foreign scent mark. The figure represents an hypothetical snapshot in time of the position of two animals, the red and blue dots, and their own scent profile, the red and blue open circles, respectively. Wherever red (blue) open circles are present it means that the red (blue) animal has walked over that location in the past timesteps, where is the period during which scent remains active. The absence of any scent marks at coordinates (5,1) and (2,4) implies that no animal has occupied those coordinates within a time . The interaction occurs whenever an animal is occupying a site with a foreign scent as displayed for the blue animal at position (4,2). Since the blue animal has deposited scent at (4,2), this point will eventually become blue territory if the red animal does not re-scent it before the red scent becomes inactive. The subsequent allowed locations where the blue animal can move are those for which no red scent is present, i.e. towards the coordinates (5,2) or (4,3), with the actual movement picked at random from these two possibilities. On the other hand, in the absence of an interaction, an animal such as the red one at coordinates (2,2) can move randomly in any of the four possible directions. :::  ::: As the terrain gets covered with scent marks released by the different individuals, each animal is segregated within an area delimited by the locations where the scent marks of neighbours are present, that is, the territory. However, since animals only respond to fresh scent, we consider the scent no longer present after time (active scent time). As a consequence the bounded domain where animals are allowed to roam is constantly changing, and the movement of the territories depends on the past locations visited by each animal. A territory is thus a dynamic quantity whose shape and centroid location change continuously. The rate of movement of a territory depends on the value of the active scent time ; the longer the less the territorial shift (see the Supplementary video S1 and S2 for a visual illustration of territorial movement dynamics for two different values of ). The two extreme situations occur when and , corresponding, respectively, to immobile territories and to territories trivially associated with the instantaneous location of each animal. For intermediate values of , the territories are like deformable elastic objects (see [@pcbi.1002008-Huxley1] for the elastic disc hypothesis and [@pcbi.1002008-Adams1] for its interpretation in terms of territory compression) whose collective dynamics are characterized by the so-called *exclusion process* [@pcbi.1002008-Landim1], [@pcbi.1002008-Schnherr1], which prevents the instantaneous occupation of the same location by two or more territories. This exclusion comes about because animals retreat from the locations where they encounter foreign scent, making possible the spatial overlap of different scents only at the boundaries. Territories cannot freely move over the terrain and get hampered, resulting in their movement rates being qualitatively different from that of the animals, the latter being diffusive and the former being subdiffusive [@pcbi.1002008-Metzler1]. The animal/territory dynamics are thus composed of two time scales, a relatively fast one associated with the diffusive movement of the animals inside their own territory and a slower one associated with the movement of the territories themselves. This latter time scale is controlled by the neighbours\' pressure and the time necessary for the resident animal to defend its own borders by refreshing the scent marks. The qualitative difference becomes evident when comparing the variance of the occupation probability, i.e. the mean square displacement (MSD), of an animal and its territory centroid. In 2D these increase respectively as (linear in time as an ordinary diffusive process) and (sublinear in time as a subdiffusive process), the latter due to the exclusion process [@pcbi.1002008-Landim1]. An indicator of territorial behaviour is often associated with the size of an animal\'s territory but for practical convenience territories are often characterized through time-integrated measures which look at the size of an animal\'s home range [@pcbi.1002008-Burt1] and the extent of the home range overlap [@pcbi.1002008-Fieberg1]. This procedure involves tracking the animal\'s locations through time for a period of observation and then selecting a computational method, e.g. the minimum convex polygon (MCP) [@pcbi.1002008-Harris2] or the utilization distribution [@pcbi.1002008-VanWinkle1], to determine the area delimiting where each animal spends most of the time for its daily activity. In [Fig. 2](#pcbi-1002008-g002){ref-type="fig"} we show a typical contour level plot of the utilization distribution of a population of 16 individuals obtained from our model. The remarkable feature that becomes apparent in looking at such a plot is that the heterogeneity in space use of the animals emerges dynamically from their interactions without the need to consider any heterogeneity in resource distribution, supporting the idea that territoriality is tightly linked to animal behavioural traits [@pcbi.1002008-Makarieva1] and not just a response to defend potential food and other resources [@pcbi.1002008-Adams1]. The heterogeneity in the spatial patterns are in fact only due to the stochastic nature of the interactions as well as the initial animal locations. Features that are commonly observed in field observations [@pcbi.1002008-White3], such as unequal home range size, boundary areas with small size home ranges (e.g. the pink home range towards the bottom right hand side) squeezed between larger ones, and spatial regions (territorial interstices) which are rarely occupied (e.g. the area between the black, red and dark green ranges), also emerge from our model. ::: {#pcbi-1002008-g002 .fig} 10.1371/journal.pcbi.1002008.g002 Figure 2 ::: {.caption} ###### Contour level plot of the utilization distribution. 2D plot of the relative frequency distribution of 16 animals\' locations with periodic boundary conditions observed up to time (density is 0.0016 animals per site). The positions and are spatial coordinates normalized to the size of the box. On moving away from foreign scent, the animals perform a correlated random walk with turning angles drawn from a 2-sided exponential distribution with a parameter proportional to , where is the number of steps since last encountering foreign scent. The coloured crosses represent the initial animal locations from which their trajectories started to be recorded. This initial condition is obtained from a single run of the simulation up to time , starting from uniformly distributed animals with no initial scent. :::  ::: To obtain a more quantitative understanding of the relation between the territory and the animal movement rates, we studied in detail a simplified version of our model with two animals in a 1D box with periodic boundary conditions. Since this reduced system also has a time scale disparity between the subdiffusive movement rate of the territories and the diffusive movement rate of the animals, it captures the fundamental characteristics of the animal/territory dynamics in 2D, with the advantage of reducing considerably the computational time of our stochastic simulations. The dynamic nature of the emergence of animal territoriality can be appreciated by plotting in [Fig. 3a](#pcbi-1002008-g003){ref-type="fig"} the variance of the occupation probability, the so-called MSD, of the animal and territorial boundary locations in our model, respectively, and . The disparity in movement rates between the boundaries and the animal is evident, with the former being subdiffusive and proportional to . The reduced dimensionality in 1D exclusion processes hinders the movement of the territories even more than in 2D, giving rise to an even slower MSD. The animal MSD, on the other hand, first increases linearly with time and then proportionally to as well, once the increase in the MSD is only due to the random displacement of the territorial boundaries. As the boundaries may roam over the entire space, the sum of the home range overlaps eventually equals the animal\'s home range at the crossing time , as indicated in [Fig. 3a](#pcbi-1002008-g003){ref-type="fig"}. Although the animals\' behaviour is still territorial beyond , the measurement would suggest that they probably do not possess any exclusive area because the distance from each territorial center is equal to half the width of each animal probability distribution [@pcbi.1002008-Giuggioli1]. This dependence on the experimental observation time becomes particularly relevant in comparative analyses of territory sizes but it is often overlooked [@pcbi.1002008-Laver1]. Meta-analysis of animal territory size should thus be performed by maintaining the same ratio for any species. ::: {#pcbi-1002008-g003 .fig} 10.1371/journal.pcbi.1002008.g003 Figure 3 ::: {.caption} ###### Mean square displacement of the locations of the animals and territorial boundaries. Boundaries are represented by dotted lines, animal 1 by solid lines and animal 2 by dashed lines. In (a) we have plotted the time dependence of the MSD of an animal, ( represents an average over the stochastic realizations of multiple trajectories starting with the same initial conditions), and the sum of the left and right boundaries, each , adjusted to correspond to a 90% MCP estimation (see the 'Relationship between home range size and overlap and mean square displacement' section of [Materials and Methods](#s4){ref-type="sec"}). Both animals exhibit the same time-dependent MSD so only one is plotted. The choice of the observation time span, from zero up to time in the figure, determines the degree of territoriality one may infer from the data, the ratio of the adjusted boundary and animal mean square displacements being proportional to the square-ratio of the overlap to the size of a home range (see [Materials and Methods](#s4){ref-type="sec"}). The probability distribution as a function of the spatial position , relative to the box size, of the locations of the boundaries and animals at time are plotted in (b) and (c), representing the different types of reaction to the encounter of foreign scent marks corresponding to the two versions of our movement model: (b) a random walk movement after retreat and (c) a correlated random walk, where the probability of continuing straight is , where is the number of steps since the animal last encountered foreign scent. These side plots illustrate the role the type of movement performed by the animals may have on the shape of their probability distribution. :::  ::: Since a territorial boundary does not change location if an animal has moved over it within a time from when it was originally scent marked, the longer it takes for an animal to move back and forth between the foreign scented areas, the more frequently the territorial borders move. As an animal deposits fresh scent at any location it wanders through, the amount of time necessary for this back and forth movement is determined by the temporal extent of two events: traversing from one boundary location to another, and subsequently returning to the starting position. The probability of either of these events occuring is precisely the first-passage probability [@pcbi.1002008-Redner1] to reach one boundary location having started at another. We call the sum of the mean of each probability the boundary return time , and we define the ratio representing the frequency of boundary encounters relative to the frequency of boundary loss. is the fundamental quantity controlling territorial emergence and allows us to compute the movement of the territorial boundaries in terms of the 'microscopic' dynamics of the animals. For boundaries move so often that rapidly becomes equal to (as shown in [Fig. 3a](#pcbi-1002008-g003){ref-type="fig"}), giving a low crossing time . In such a scenario, unless the observation time window is less than , the territories will most likely have no area of exclusive use as shown in the inset (a) of [Fig. 4](#pcbi-1002008-g004){ref-type="fig"}, where the occupation probability at time is depicted. For , on the other hand, the movement of the territorial boundaries is reduced and areas of exclusive use are likely to emerge (see insets (b) and (c) in [Fig. 4](#pcbi-1002008-g004){ref-type="fig"}). ::: {#pcbi-1002008-g004 .fig} 10.1371/journal.pcbi.1002008.g004 Figure 4 ::: {.caption} ###### Exclusivity of space use. Cross-over from territories with an area of exclusive use to ones without in terms of versus . The situation where exclusivity arises is indicated by the closed circles, whereas the absence of exclusivity is represented by the open circles. For a fixed observation time , the insets indicate the probability distribution of the two animals as a function of the spatial position relative to the box size. The degree of overlap between territorial neighbours diminishes as increases, as indicated by inspecting the insets (a), (b) and (c) sequentially. The reaction to the neighbouring scent encountered is the one employed in [Fig. 3b](#pcbi-1002008-g003){ref-type="fig"} and the ratio as one moves from inset (a) to (c). :::  ::: To verify that meaningful active scent time values can be recovered from movement data, we apply our theory to the red fox (*Vulpes vulpes*) [@pcbi.1002008-Baker1] population in Bristol (UK), before the 1994-6 mange epizootic which decimated the fox population [@pcbi.1002008-Baker2]. The data we analyze are ideally suited to test our model assumptions since food was widely distributed and available to excess on the study area. We have attempted to use both the 2D as well as the 1D version of our model, the latter one by projecting the experimental observations onto the line connecting the centroids of each animal\'s home range. Whilst the 2D model is ostensibly more realistic, it does not reflect the fact that foxes tend to snoop around the territory boundaries on each visit. In the 1D model the boundary just consists of two points, so each visit causes half the boundary to be scented. The analyses in the two cases gave results ( days for 2D, and days for 1D), the latter of which agrees well with the time lags of 3--4 days of territorial takeover following the death of all the residents, as observed by one of us (SH) during the mange epizooty in Bristol. We have also examined a variation of our model whereby each animal snoops round a proportion of the territory perimeter on each visit. By increasing , the estimated can be reduced. Particulary, a of 3--4 days requires . This suggests that foxes are deliberately spending time scenting the boundary, in addition to the movement patterns described in our model. Discussion {#s3} ========== Mechanistic approaches to territorial behaviour may be limited in scope when mathematical predictions can only be tested at the 'macroscopic' population level [@pcbi.1002008-Adams1]. Here we have provided a quantifiable 'microscopic' testable hypothesis on the mechanism of conspecific avoidance that relates a physio-ecological animal characteristic, the active scent time, to the population distribution of territorial patterns and promises to enrich the emerging field of conservation physiology [@pcbi.1002008-Wikelski1]. To study the role of scent marking in the emergence of territories, one requires the use of modelling techniques that go beyond deterministic reaction-diffusion formalisms [@pcbi.1002008-Moorcroft1], since the latter are not viable approximations when the stochastic interaction events are rare and spatially heterogeneous [@pcbi.1002008-Durrett1]. In territorial formation and maintenance these events are precisely the random encounters of a resident individual with the scent boundaries of a neighbouring conspecific. Moreover, the type of reaction-diffusion model, recently employed to study territorial formation in wolf-packs [@pcbi.1002008-White2], contains a fundamental constraint: the *a priori* assignment of the location of the focal activity points (e.g. a den or burrow) towards which each animal is attracted [@pcbi.1002008-Borger1]. In other words, one of the outcomes of the formation process is predetermined. By borrowing concepts from non-equilibrium statistical physics, we are able to explain that the different rates of movement between the territories and the animals is the result of geometric constraints coming about because of exclusion processes and the ensuing anomalous sub-diffusive properties of the animal scent profiles. This conceptual framework has allowed us to quantify territorial dynamics in relation to the time an animal needs to move across its own territory refreshing its own scent marks, and the time the chemical signals present in the scent remain active. Although anomalous diffusion is recognized [@pcbi.1002008-Giuggioli2] as a useful framework to interpret statistically animal movement data in the context of foraging strategies (see e.g. [@pcbi.1002008-Bartumeus1]), our study is the first to show its relevance to animal social interactions. The necessary level of biological realism that our agent-based simulations introduce is at the expense of mathematical complications since an animal location and its territorial scent profile depend on the history of the other animal trajectories; in other words it is non-Markovian [@pcbi.1002008-vanKampen1] in character. As a consequence each animal perceives an environment which is being modified by the spatio-temporal trajectories of other individuals, making the animal movement dynamics context-dependent [@pcbi.1002008-Torney1]. From a single animal perspective, this context-dependence generates a fluctuating heterogeneous environment, which manifests itself in the transient territorial patterns of the individual animals. These findings lend support to the idea that interactions are key to territorial emergence and why no significant effect of resource abundance on territory size has been found in many experimental studies [@pcbi.1002008-Adams1]. Although our results are based on considering only individual animals defending a territory, it has a wide applicability since the vast majority of mammals do not form groups, and among these territorial defense is performed by one animal, usually the male, or, in those cases where both sexes play a role, the male generally takes the dominant role. In those cases where groups defend a communal territory but move together (e.g. wolves), it makes little difference whether one or all of the animals scent mark or play a role in territorial defense, since their movement behaviour will be little different from the solitary animal represented in our model. We have tested the applicability of our model by verifying that a meaningful active scent time can be obtained from urban red fox data, that is data from a territorial species moving within an environment where food is widely and abundantly available. This result and the simplicity of our assumptions in the movement of the animals implies that our study has produced a null model of animal interactions onto which one can add nutritional effects [@pcbi.1002008-Simpson1] and test optimality questions [@pcbi.1002008-Reynolds1] in cooperative search strategies [@pcbi.1002008-Torney1], [@pcbi.1002008-Couzin1]. Besides the relevance to territorial formation, the results of our model represent a benchmark to test ideas related to the role of scent in animal communication, and in particular in the context of foxes and possibly other carnivores. Fox scent marks provide a great deal of information about the fox that left the message, because chemical analyses of the volatile compounds present in fox scent marks can identify sex, season, relatedness, health and possibly social status of an animal [@pcbi.1002008-Arnold1]; whilst it is currently unknown whether foxes can also receive all of these messages, it seems highly probable that they do. This would then explain why foxes snoop: they can learn all they need to about their neighbours, both presence and more detailed information, without ever needing to meet. In this respect our model provides a useful tool to help design field experiment to study behavioural response of resident animals to alien scent marks from animals of known age, sex and social status, both in the snoop zone and further into the territory. It is relevant to note that the importance of the frequency of animal encounters has been considered in studies of the allometric scaling of the exclusivity of space use [@pcbi.1002008-Jetz1], by using an analogy between animal encounter rate to the interaction frequency of physical particles in an ideal gas [@pcbi.1002008-Hutchinson1]. In that context our results provide a spatially explicit 'microscopic' interpretation of that study [@pcbi.1002008-Jetz1]. In summary, with a systems biology type of approach we have been able to show that scent marks in territorial animals serve as a short-term cue, illustrating why territorial mammals regularly renew and refresh their scent marks. Our field observations on foxes showed that, when territories became vacant, they were rapidly taken over by neighbours and our model demonstrates this very neatly. We have also shown the practical steps required to extract active scent time from 2D animal fixes. Materials and Methods {#s4} ===================== The stochastic simulations {#s4a} -------------------------- The positions of the interacting random walkers on a lattice are updated at every time-step following the rules depicted in [Fig. 1](#pcbi-1002008-g001){ref-type="fig"}. In order to measure the walker probability distibutions over time, initial conditions in the stochastic simulations are expressed in terms of the animal locations and the scent spatial profiles, i.e. each lattice site defines the time the scent is still available before it becomes inactive. A biologically relevant initial scent profile is a curve with minimum values at the scent boundaries and a maximum at the lattice site where the animal is initially present. Since territories may move only when the scent profile at the boundary of at least one of two neighbouring animals is equal to zero, we have used an initial scent profile with such a feature. The shape of the curve, which interpolates between zero at the boundaries and the maximum corresponding to the animal\'s position, is obtained by averaging over stochastic simulations that are run starting with a spatially uniform distribution of animals and with all scent profiles equal to zero. The moment when this average is computed is at a time (see [table 1](#pcbi-1002008-t001){ref-type="table"}) longer than , corresponding to the situation when the boundary MSD has reached its asymptotic regime. ::: {#pcbi-1002008-t001 .table-wrap} 10.1371/journal.pcbi.1002008.t001 Table 1 ::: {.caption} ###### Notation glossary. ::: {#pcbi-1002008-t001-1} Symbol Explanation Input/Output -------- -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- -------------- Time-window over which data on animal positions are gathered. Input Active scent time: the amount of time that the scent of one animal is considered to be 'fresh' by other animals. Input Boundary-return time: the average time for an animal to visit every point on the boundary and return to the point it set out from. Output The quotient representing the frequency of boundary encounters relative to the frequency of boundary loss. (see , ) The position of an animal in a simulation. Output The MSD of an animal, averaged across simulated trajectories and starting from the same initial conditions. Output The position of a territorial boundary location. Output The MSD of a territory boundary. In 1D, this is simply the MSD of either of the boundary points , averaged over simulations. In 2D, the main contributor to is the displacement of the territory centroid, . However, we also added the displacement of the territory radius , defined to be the mean distance from the centroid to the boundary points at any instant in time. More precisely , where is the average over stochastic realizations along an arbitrarily chosen reference direction (north-east in our choice). Output 90% MCP home range size. Output 90% MCP home range overlap: i.e. the 90% MCP of the boundary position distribution. Output The average time span beyond which we are going to observe areas of exclusive space-use, i.e. the time when . Output Generalised diffusion constant of the territory boundaries. Output Time at which reaches its asymptotic limit (i.e. is proportional to in 1D and in 2D). Output Glossary of the various symbols used throughout the text, a brief explanation of each and whether the quantity is measured from the model (output) or a parameter of the model (input). ::: In order to measure the boundary position distributions in both the 1D and 2D models, we examined the asymptotic regime, where in 1D or in 2D, so that the movement is independent of initial conditions. For [Figs. 3](#pcbi-1002008-g003){ref-type="fig"} and [4](#pcbi-1002008-g004){ref-type="fig"} the number of simulations carried out to plot the occupation probabilities and the boundary and animal MSD is equal to . [Fig. 2](#pcbi-1002008-g002){ref-type="fig"} on the other hand is generated with only 100 simulations since utilization distributions are time-integrated quantities [@pcbi.1002008-VanWinkle1] and require less averaging to obtain smooth contour level plots. Relation between home range size and overlap and mean square displacement {#s4b} ------------------------------------------------------------------------- As territorial interactions are dynamic, and often rely on scent marking, it is hard to detect subtle changes in territorial boundaries using conventional field techniques. So size of animal home ranges is used as a surrogate for territory size. A by-product of the home range estimation for neighbouring individuals is the determination of the area of home range overlap. It is thus of interest to know how to relate the MSD, as shown in [Fig. 3](#pcbi-1002008-g003){ref-type="fig"}, with home range size and home range overlap . Since boundary movement is the result of an exclusion process, the boundary location distribution function over an extended period is Gaussian [@pcbi.1002008-Hahn1], so that the boundary MSD is , where is the width of the boundary location distribution at of the distribution\'s maximum height. For the animal, on the other hand, it is necessary to relate the MSD to the area obtained from the extreme statistics associated with the determination of the MCP [@pcbi.1002008-RandonFurling1], [@pcbi.1002008-Majumdar1]. Since the animal and boundary position distributions both have infinite tails, there is an unbounded increase of the 100% MCP as the number of samples from either distribution increases [@pcbi.1002008-Moorcroft1]. Therefore we have selected an A% MCP with whose width saturates with time. This choice is related to the shape of the probability distribution and we found that a flat-topped curve with exponential tails well approximates the shape of the animal probability distribution for time values relevant in our simulations. We noticed that by employing A = 90% we can relate the value obtained from the MCP calculation with the MSD through the expression within an error of 5%. From normal distribution tables we know that 90% MCP of the boundary distribution, in other words the 90% overlap , is a factor of 1.645 larger than for 1D from which it follows that . In [Fig. 3](#pcbi-1002008-g003){ref-type="fig"} the MSD of the animal has been plotted as is, whereas one of the boundaries has been multiplied by to ensure that the two curves cross where . For 2D, the value corresponding to 1.645 is 2.146. This is used in the data-fitting below. Analysis of the red fox *Vulpes vulpes* radiotelemetry data {#s4c} ----------------------------------------------------------- We used data from a long-term study of the red fox population in the Bristol urban area with a spatial resolution of , larger than the spatial heterogeneity perceived by the animals. We analyzed radio telemetry data from adult foxes recorded during the springs of 1978 and 1991 and the summer of 1990. Throughout each season, three foxes from different groups were radio-tracked for between 5 and 11 nights, with positional fixes taken every 5 minutes. We employed both the results in 1D and 2D in estimating an active scent time from the radio telemetry data. For the 1D model we projected the animal locations onto a line through their home range centres, calculated by taking the centroid of all position fixes. From the first 25 minutes after a fox begins to move, the diffusion constant (see e.g. [@pcbi.1002008-Giuggioli3]) of the animals was estimated to be , giving a time-step in the model corresponding to s. Since the red fox is active for just 8 hours in any 24 hour period [@pcbi.1002008-Saunders1], each 90 day data-gathering time-window corresponds to 720 hours, or an observation time time-steps. [Table 2](#pcbi-1002008-t002){ref-type="table"} shows the number of fixes used and the period when data were collected as well as the animal\'s gender. Although the dominant male is the primary territorial defender, since other adults in the group will have roughly the same home ranges as the dominant male, we have used data from both male and female foxes. However, all the data are from adult foxes since cubs and sub-adults tend to have smaller home ranges [@pcbi.1002008-Robertson1]. ::: {#pcbi-1002008-t002 .table-wrap} 10.1371/journal.pcbi.1002008.t002 Table 2 ::: {.caption} ###### Data collected on neighbouring foxes. ::: {#pcbi-1002008-t002-2} Fox Season Number of radio locations Gender Fox density ----- ------------- --------------------------- -------- ------------- A Spring 1978 410 F 0.0257 B Spring 1978 124 M 0.0257 C Spring 1978 433 F 0.0257 D Summer 1990 488 M 0.0324 E Summer 1990 673 M 0.0324 F Summer 1990 267 M 0.0324 E Spring 1991 873 M 0.0442 F Spring 1991 582 M 0.0442 G Spring 1991 932 F 0.0442 Number of radio locations, season of observation, gender and number of individuals per hectare for the foxes with overlapping home ranges observed between 1978 and 1991. ::: From extensive simulations of our 1D model we determined that the quantitative dependence of the generalized diffusion coefficient in the boundary MSD falls onto a universal curve when plotted versus where is the size of the periodic box. The fitting line is given by with To understand this universal curve we analysed the dynamics of two interacting animals in two extreme situations: and . For the case , we constructed with the help of diffusion graph transform [@pcbi.1002008-Burioni1] a discrete Master equation for the relative distance of two random walkers roaming within a lattice of sites with periodic boundary conditions. The two walkers move freely except when they meet, after which they move away from each other. By solving this master equation, we computed the mean first passage time for two walkers, starting in the same square, with walker 1 having just moved left, to meet again with walker 1 having just moved right. This quantity is half the mean boundary return time (for the case ) that we have defined in the text. It turns out that . For the other extreme case , the first-passage time is governed purely by the sizes of the two territories that initially form, since thereafter they do not move. For a random walk restricted to move on a line segment, the mean first-passage time to go from one (reflecting) edge to the other (absorbing) edge is the square of the length of the segment [@pcbi.1002008-Redner1]. Assuming that one of the territories is of size and the other of size , the average is equal to . Let be the probability that the initial territories are of sizes and . Then Using numerical simulations for various , we find so . For intermediate values of and sufficiently large values of the box size , we can expect an dependence of with a coefficient of proportionality interpolating between 4 and 2.072, explaining the universal fitting curve as described above. This dependence is also what one would expect in 2D, where first-passage times in a bounded domain are proportional to the size of the domain itself, to a first order approximation [@pcbi.1002008-Codamin1]. By projecting position fixes onto the line between adjacent home range centres, we calculated the 90% MCP home range width to be m and the overlap-to-size ratio . This corresponds to a model density of 1 animal per lattice sites and an overlap of sites. Since the boundary displacement is Gaussian, we can use the 90% home range overlap, which is , making the boundary MSD . From the theoretical values of the diffusing boundary MSD , we recover a value of . From the universal curve described above and the experimental population density, we relate to , giving days. The procedure in 2D is similar, although we have not studied in detail if there exists a universal curve of the dependence of as a function of the box size and . We considered the average population density from the data, which was 1 male dominant fox per 29.3 hectares, and ran the simulation for various . In each case, we considered the asymptotic regime of the boundary MSD and used this to calculate for time-steps. The resulting empirical relationship obtained from the stochastic simulation between and gives us days. Supporting Information {#s5} ====================== Video S1 ::: {.caption} ###### Territorial dynamics with small active scent time. Movie of the territorial dynamics of 25 animals with an active scent time time steps in a box of 100×100 sites with periodic boundary conditions. The initial movie frame is recorded after a small transient obtained from an initial condition with the animals periodically placed on the lattice and without any scent profile. The snapshots of the simulations are taken every 10,000 time steps. (GIF) ::: ::: {.caption} ###### Click here for additional data file. ::: Video S2 ::: {.caption} ###### Territorial dynamics with large active scent time. Movie of the territorial dynamics of 25 animals with an active scent time time steps in a box of 100×100 sites with periodic boundary conditions. The initial movie frame is recorded after a small transient obtained from an initial condition with the animals periodically placed on the lattice and without any scent profile. The snapshots of the simulations are taken every 10,000 time steps. (GIF) ::: ::: {.caption} ###### Click here for additional data file. ::: We acknowledge the Advanced Computing Research Centre (<http://www.bris.ac.uk/acrc/>), at the University of Bristol, for making available to us their computing resources and we thank three anonymous reviewers for their helpful comments. The authors have declared that no competing interests exist. This work was partially supported by the EPSRC grant number EP/E501214/1 (L.G. and J.R.P) and by the Dulverton Trust (S.H.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. [^1]: Provided the biological insights on territorial species: SH. Performed research with inputs from LG: JRP. Conceived and designed the experiments: LG. Performed the experiments: SH. Analyzed the data: JRP. Wrote the paper: LG JRP SH.

PubMed Central

2024-06-05T04:04:19.639691

2011-3-10

PMC3053311

Introduction {#s1} ============ GroEL participates in the folding of 5-10% of cellular *Escherichia coli* proteins by providing an isolated chamber for non-native substrate proteins together with a heptameric ring shaped co-chaperonin, denoted GroES [@pcbi.1002004-Georgopoulos1]--[@pcbi.1002004-Stan1]. It has also been suggested that GroEL might forcefully unfold kinetically trapped misfolded intermediates [@pcbi.1002004-Shtilerman1], [@pcbi.1002004-Lin1]. GroEL is composed of two heptameric rings made up of identical subunits, each of 57 kDa. The two separate rings, denoted cis (active) and trans (inactive), are stacked back-to-back to form two folding environments (or cages) working off-phase, analogous to a two-stroke engine. Each subunit is divided into three domains; equatorial (residues 1-133, 409-548), intermediate (134-190, 377-408), and apical (191-376) domain, separated by two hinges that facilitate large conformational transitions in the complex [@pcbi.1002004-Braig1]--[@pcbi.1002004-Xu1]. Substantial mechanistic insight has been obtained through comparison of the large amount of structural data available for the GroEL complex. The first high-resolution X-ray structure was released in 1994 by Braig *et al.* [@pcbi.1002004-Braig1], and since then, numerous structural studies, including X-ray crystallography, cryo-EM, and NMR have been published of different functional states of GroEL (for reviews see [@pcbi.1002004-Horovitz1], [@pcbi.1002004-Horwich1]). On this structural background it has been possible to make predictions and educated hypothesis about the transition pathways during the protein functional cycle. This includes the ATP dependent opening of the cis cavity with the concomitant increase of its volume from to [@pcbi.1002004-Xu1]. The conformational transitions occurring on the subunit level are substantial, and its trajectory is generally explained in a sequential manner. Binding of ATP to each of the seven equatorial domains in the *cis* ring, together with a non-native polypeptide produces a counterclockwise twist of the apical domain, and a downward rotational movement of the intermediate domain [@pcbi.1002004-Ranson1], [@pcbi.1002004-Ranson2]. This structural state of the ring is denoted **R** (where each of the 7 subunits are in the **r** state) while the initial closed state is denoted **T** (each of the 7 subunits are in the **t** state). Reaching the **R** state facilitates GroES association to the apical domains of the cis ring promoting much larger conformational changes and resulting in the fully open **R′** conformation (all seven subunits in the **r′** state). This **r′** conformer is characterized by a elevation and clockwise twist of the apical domains (opposite direction to that seen upon ATP binding) [@pcbi.1002004-Xu1]. Non-native polypeptide folding takes place within the *cis* ring of the **R′** state of the reaction cycle, which is the longest lived (about 8--10 s) [@pcbi.1002004-Rye1], and continues until ATP hydrolysis induce the **R″** conformation permiting ATP binding to the opposite *trans* ring [@pcbi.1002004-Rye2], [@pcbi.1002004-Fridmann1]. This final rearrangement result in a conformer very similar to the **r′** form (RMSd of 1.46 Å). Despite this extensive structural insight, the mechanisms involved in allosteric signaling are not yet fully understood at atomic and residue level. In general, X-ray crystallography provides invaluable snapshots of different states of the protein reaction cycle, but not of the transitions between them. In this context, computational approaches have the potential to nicely complement the experimental techniques. In particular, molecular dynamics (MD) simulations provide important insight into protein dynamics at the atomic level, and allow following subsequent individual atomic interactions and fluctuations as a function of time [@pcbi.1002004-Karplus1]. Additionally, normal mode analysis (NMA) has proven to be efficient and accurate in the task of predicting and describing large scale conformational transitions in proteins [@pcbi.1002004-Skjaerven1]--[@pcbi.1002004-Yang1]. NMA analytically characterizes all possible deformations of a protein around a stable equilibria with respect to their energetic cost. Although the utilization of NMA involves a loss of time-dependent fluctuations and resolution, i.e. by the use of coarse-grained models, it has been shown that it provides functionally relevant motions, and information on allosteric mechanisms [@pcbi.1002004-Skjaerven2]--[@pcbi.1002004-Zheng1]. Various computational methods have indeed provided essential information on the GroEL subunit dynamics. Notably, the transitions between the main functional states (**T**, **R**, and **R″**) have been further refined by utilizing NMA [@pcbi.1002004-Ma2]--[@pcbi.1002004-Keskin1], targeted MD (TMD) [@pcbi.1002004-Ma3], brownian dynamics [@pcbi.1002004-Hyeon1], principal component analysis (PCA) [@pcbi.1002004-deGroot1], and MD simulations [@pcbi.1002004-Sliozberg1], [@pcbi.1002004-Abrams1]. Moreover, a number of computational studies have been dedicated to find pathways for both intra- and inter-ring communication in GroEL [@pcbi.1002004-Brocchieri1]--[@pcbi.1002004-Tehver1]. From these, several residues have been pointed out as important for the allosteric signaling thus increasing the understanding of the involved mechanisms. The concept of the two-stage transition has been strengthened by the TMD simulation of the GroEL subunit which pulls the **t** form to the fully open **r″** form along a 500 ps long MD simulation [@pcbi.1002004-Ma3]. This computational study suggested that the transition begins with a downward tilt of helix M, and the subsequent counterclockwise twist of the apical domain. Moreover, biasing the MD simulation by employing temperature acceleration to increase the conformational sampling recently showed the ability of the isolated GroEL subunit to undergo the **t** to the **r″** state transition [@pcbi.1002004-Abrams1]. However, this simulation was not able to sample the closing of the binding pocket as seen in the X-ray structures of the **r″** conformers. Finally, unbiased MD simulation of the GroEL subunit has been performed in order to investigate the ATP-driven conformational changes [@pcbi.1002004-Sliozberg1]. This simulation samples the transition from the **t** to the semi-relaxed **r** conformation during 20 ns, nicely illustrating formation and rupture of hydrogen bonds. A full scale monitoring of individual particle motions to probe the mechanistic basis for the transitions requires extensive unbiased conformational sampling at atomistic resolution. The 20 ns long MD simulations of Sliozberg and Abrams are too short to observe the full scale transitions of the subunit [@pcbi.1002004-Sliozberg1]. Moreover, significant variations between individual MD simulations have been detected [@pcbi.1002004-Skjaerven2], [@pcbi.1002004-Caves1], [@pcbi.1002004-Clarage1], thus highlighting the importance of multiple simulations to extract statistically relevant information on the conformational changes. Investigating the positive (intra-ring) and negative cooperativity (inter-ring) of ATP-binding would require MD simulations of the entire GroEL oligomer. Such simulations on relevant timescales for conformational transitions are beyond the capabilities of present computational resources. It has also been reported the surprising facility to obtain stable folded monomers of GroEL by a variety of means [@pcbi.1002004-Lissin1]--[@pcbi.1002004-Horowitz1], which can be explained by the small area buried by the monomers at subunit interfaces in the X-ray structure of the GroEL complex [@pcbi.1002004-Braig1]. Furthermore, under some particular experimental conditions it seems that monomeric GroEL exhibits a weak chaperone activity [@pcbi.1002004-Taguchi1]. Despite the fact that the chaperone activity of monomeric GroEL is a controversial issue [@pcbi.1002004-White1], the ability of the monomer to fold into a native-like conformation that can bind nucleotide, points to the subunit as a relevant structural unit to be investigated by MD simulations. In the current work we present extensive (in total long) unbiased MD simulations of the GroEL subunit starting from the closed (**t**) and open (**r″**) conformations, with and without bound nucleotide. We use PCA of 287 experimentally obtained GroEL subunits to interpret the conformational sampling of our simulations. We observe nucleotide dependent shifts in the conformational ensembles, and show that the subunit response, as observed in the oligomeric structure, is intrinsically coded in the structure-dynamics of the isolated subunit. These simulations provide so far unexplored sampling from **t** all the way to a structure close to **r″**. A nearly complete transition in the opposite direction is also sampled by removing ADP from the **r″** form. Another interesting outcome of our simulations is that the inherent motion of unliganded GroEL subunit is biased along the transition pathway towards both the **r** and **r″** states. Furthermore, the MD simulations reveal a weak stabilization of the equatorial domain upon ATP binding, resulting in a modest decrease of the configurational entropy of this domain. Conversely, a larger increase of entropy is found for the whole GroEL subunit. Finally, we decipher the underlying mechanisms for the conformational transitions by investigating the atomic interactions unique to the unliganded and nucleotide bound structures obtained from the X-ray structures and MD simulations. Several of the interactions that characterize the conformational intra-subunit effects brought about by ATP binding were not revealed in previous studies. Results {#s2} ======= Analysis of GroEL structures and dynamics {#s2a} ----------------------------------------- 27 crystal structures of the GroEL complex, with a total of 364 subunits, were collected from the RCSB protein databank. Six of the crystal structures were not considered due to missing coordinates, which leaves 21 GroEL complexes with 287 subunits for the following analysis. These include ATP and ADP bound conformers as well as apo forms from wild type and mutant structures. The 287 GroEL subunits were superimposed onto the invariant 'core' defined as the area with least structural variation [@pcbi.1002004-Grant1], and PCA was performed to investigate the major conformational differences between the collected structures. As much as 91.5% of the total variance of the atomic fluctuations was captured along the first principal component (PC), while 2 and 3 dimensions were necessary to capture 95.3% and 97.5%, respectively (see inset in [Figure 1A](#pcbi-1002004-g001){ref-type="fig"}). ::: {#pcbi-1002004-g001 .fig} 10.1371/journal.pcbi.1002004.g001 Figure 1 ::: {.caption} ###### PCA of 287 GroEL subunits. \(A) Experimentally obtained structures projected onto the two first principal components (PC). The inset shows the eigenvalue spectrum representing the percentage of the total variance captured by the corresponding eigenvector. Labels besides each point indicate the cumulative sum of the total variance accounted for in all lower number eigenvectors. (B) Visualization of PC1 of the X-ray structures, which consists of a downward tilt and rotation of the apical domain, along with a downward tilt of helix M in the intermediate domain. Internal helix motion is also observed within the equatorial domain along with flapping loop motion (residues 36-51). :::  ::: The GroEL subunits can be divided into three major groups along the two first PCs; the closed cis/trans **t** forms, open cis **r″** forms, and the semi-relaxed **r** forms ([Figure 1A](#pcbi-1002004-g001){ref-type="fig"}). The first PC is shown in [Figure 1B](#pcbi-1002004-g001){ref-type="fig"} and describes the main differences between the **r″** and **t** conformers. The motions described by PC1 consist of (1) an upward movement of the apical domain away from the equatorial domain; (2) a small rotation of the apical domain; (3) a downward tilt of helix M (residues 386-409) and (4) a translation of helices F and G (residues 141--152 and 155--169, respectively) along their axis of inertia. This region of the intermediate domain moves down to cover the nucleotide binding site in the equatorial domain while the apical domain twists upwards. In addition, modest internal fluctuations are observed in the equatorial domain. The stem loop spanning from Arg36 to Lys51 has a flapping motion opposite to the intermediate domain, and helices A (residues 8-27) and C (residues 65-85) show a rotation relative to helices D (residues 88-107), E (residue 113-134), and R (residues 496-514). Similar internal motions in the equatorial domain are also observed in PC2. In particular, helix A (residues 8-27) undergoes a translational motion along its longitudinal axis, while both helices D (residues 88-107) and R (residues 496-514) rotate along their axis. PC2 represents the largest variation between the **t** and **r** conformers, consisting of a counterclockwise twist and a modest elevation of the apical domain. Similarly, PC3 describes a rotational motion of the apical domain together with a translational motion of helix A along its longitudinal axis. The internal fluctuations within the equatorial domain reveal the presence of two sub-domains (helices A+C, and helices D+E+R) consistent with a previous study utilizing 35 subunit structures [@pcbi.1002004-deGroot1]. Moreover, it is interesting to notice that the invariant core is a relatively small part of the equatorial domain (helices E, O, and R), thus pointing out the structural variability within this domain. Conformational population shift as a response to nucleotide binding {#s2b} ------------------------------------------------------------------- To investigate the intrinsic response of nucleotide binding to GroEL we conducted 14 MD simulations of the isolated GroEL subunit starting from both ends of the reaction cycle; the closed (**t**) and open (**r″**) states, with (holo) and without (apo) bound nucleotide (ATP or ADP). Of these, we performed four 300 ns long simulations: (A) closed **t** unbound, (B) open **r″** unbound, (C) closed **t** ATP-bound, and (D) open **r″** ADP-bound. An additional ten 100 ns long simulations of the closed **t** form were also carried out: 5 (E--J) unliganded, and 5 (K--P) ATP-bound. Simulations (D) and (K) consist of the 0--100 ns interval of simulation (A) and (C), respectively. Each conformer obtained from the MD simulations was projected onto the first two PCs determined from the X-ray structures ([Figure 2](#pcbi-1002004-g002){ref-type="fig"}). These projections display the relationship between the MD conformers in terms of the conformational differences described by the two first PCs, thus enabling interpretation of the conformational space sampled in each of the simulations. ::: {#pcbi-1002004-g002 .fig} 10.1371/journal.pcbi.1002004.g002 Figure 2 ::: {.caption} ###### Conformer plots of X-ray and MD structures reveal the nucleotide dependent conformational sampling. Projection of X-ray (contoured grey dots) and MD (blue) structures onto the two most significant principal planes defined by 287 GroEL experimentally obtained structures. The 300 ns simulations are shown in the first row of panels, while the 100 ns simulations are shown in rows 2--4. The distribution of MD structures is depicted with density shaded blue points. Light grey dots represent the first 50 ns of the 100 ns simulations. Orange, red, and green dots represent the closed (**t**), semi-relaxed (**r**), and open (**r″**) X-ray structures, respectively. The conformer plots show a two dimensional representation of the conformational sampling in terms of conformational difference as described by the PC\'s. :::  ::: Remarkably, the simulations of the closed subunit all sample the space along PC2, which describes the transition between the **t** and **r** form, independently of whether ATP is bound or not. The main difference between the unliganded and ATP-bound simulations lay in the sampling along PC1; the ATP bound simulations shift the ensemble of conformations along PC1, which thus comes closer to the fully open **r″** form ([Figure 2A](#pcbi-1002004-g002){ref-type="fig"}+C). This difference is most apparent in the 300 ns simulations where the ensemble samples closest to the fully open structure (see [Video S1](#pcbi.1002004.s005){ref-type="supplementary-material"}). The difference in sampling along PC1 is also significant for the 100 ns simulations (Figure S10), however they are too short to fully observe this effect. Admittedly, 100 ns is not long enough simulation time to bring the ATP bound simulation (N) away from the **t** form ([Figure 2N](#pcbi-1002004-g002){ref-type="fig"}). A similar shift of the ensemble along PC1 and PC2 is also observed for the open simulations. Interestingly, the ADP-bound simulation samples the space in the proximity of the **r″** form ([Figure 2D](#pcbi-1002004-g002){ref-type="fig"}). Conversely, removing ADP yields a sampling of the conformational space which is significantly shifted, in particular along PC1, towards the **t** conformers ([Figure 2B](#pcbi-1002004-g002){ref-type="fig"}). Clustering analysis on each of the four 300 ns long trajectories was performed to identify the predominant conformations throughout the simulations. Each frame in the trajectories is attributed to one particular cluster depicted as color bars in [Figure 3](#pcbi-1002004-g003){ref-type="fig"} along with the resulting RMSd values. The conformations sampled in the closed apo simulation ([Figure 3A](#pcbi-1002004-g003){ref-type="fig"}) resembles its initial starting structure (**t**). Conversely, the closed holo simulation, shown in [Figure 3C](#pcbi-1002004-g003){ref-type="fig"}, comes much closer to the open **r″** form than its apo counterpart. The RMSd values for the average conformations in each of the clusters with respect to both the closed and open X-ray structures are shown in [Table 1](#pcbi-1002004-t001){ref-type="table"}. The predominant clusters of the closed apo simulation are clusters A1+A2 which have a relatively close similarity to the closed X-ray structure, with an RMSd value of 2.56 and 3.45 Å, respectively. These low RMSd values of the closed apo simulation stand in contrast to higher RMSd values of the closed holo simulation, i.e. 5.97 and 4.09 Å for the two dominating clusters C2+C3. Cluster C4 of the closed holo simulation, which consists of 92 member conformations out of a total of 1500 MD conformers, has a higher average similarity to the open (RMSd 6.86 Å) than the closed X-ray structure (9.21 Å). We thus sample relatively close to the experimental **r″** conformer (minimal deviation of 5.8 Å) with the major difference being an open nucleotide pocket, similar to what a recent temperature biased MD simulation by Abrams and Vanden-Eijnden showed [@pcbi.1002004-Abrams1]. Interestingly, both the unbound and the ATP bound simulations also show a tendency to mimic the **r** form (minimal deviation of 2.6 Å) characterized by the counterclockwise rotation of the apical domain (i.e. clusters A3 and C1). ::: {#pcbi-1002004-g003 .fig} 10.1371/journal.pcbi.1002004.g003 Figure 3 ::: {.caption} ###### RMSd variation for the 300 ns long simulations of the GroEL subunit. Black line corresponds to the RMSd with respect to the starting structure (**t** for A+C, and **r″** for B+D), while the grey line corresponds the RMSd with respect to the opposite form (**r″** for A+C, and **t** for B+D). Each plot corresponds to one specific MD simulation; (A) closed (**t**) unbound, (B) open (**r″**) unbound, (C) closed (**t**) ATP bound, and (D) open (**r″**) ADP bound. Black dashed line depicts the RMSd value between the open (**r″**) and closed (**t**) X-ray structure. Clustering analysis is performed on each of the 4 simulations, and the bar on the top of each RMSd plot is colored according to which cluster the corresponding frame belongs to. :::  ::: ::: {#pcbi-1002004-t001 .table-wrap} 10.1371/journal.pcbi.1002004.t001 Table 1 ::: {.caption} ###### Clustering analysis of MD trajectories of the GroEL subunit. ::: {#pcbi-1002004-t001-1} Simulation Cluster No. Cluster size t r r″ ----------------- ------------- -------------- ------- ------- ------- Closed apo (A) A1 172 2.56 4.10 11.31 A2 1253 3.45 4.06 12.77 A3 72 5.533 4.11 10.10 A4 3 3.94 5.27 13.64 Open apo (B) B1 112 12.06 11.98 3.11 B2 664 14.14 13.57 5.65 B3 84 14.93 13.98 7.23 B4 269 12.52 11.79 6.80 B5 160 10.19 9.96 6.10 B6 211 7.23 9.36 8.11 Closed holo (C) C1 157 4.04 3.72 10.70 C2 654 5.97 5.62 8.10 C3 464 4.09 5.05 9.66 C4 92 9.21 8.94 6.86 C5 133 5.68 4.71 11.66 Open holo (D) D1 313 11.05 11.69 4.40 D2 13 10.09 9.73 5.70 D3 374 10.63 11.01 3.91 D4 15 13.90 13.99 4.80 D5 695 8.41 9.66 4.90 D6 90 13.56 12.99 5.27 The RMSd values between the closed and open X-ray structures and the average conformations in each of the clusters in the four simulations of the GroEL subunit are shown. The RMSd between the open and closed X-ray structure (PDB ids: 1SVT and 1XCK, respectively) is 12.30 Å. ::: The RMSd values and clustering of the open (**r″**) apo and holo simulations are shown in [Figure 3B](#pcbi-1002004-g003){ref-type="fig"}+D. The apo simulation initially samples conformations around its starting point; close to the open X-ray structure. The predominant clusters of this simulation are B2 and B4 with a total of 933 member conformations, and the average conformations of these clusters have RMSd values much closer to the open than to the closed X-ray structure as shown in [Table 1](#pcbi-1002004-t001){ref-type="table"}. After about 240 ns of simulation the structure undergoes a remarkable conformational change resulting in higher similarities towards the closed (**t**) form. These structures are assigned to cluster B6, and the member conformations have an average RMSd of 7.23 Å towards the closed X-ray, and 8.11 Å towards the open X-ray structure. We thus sample relatively close to the experimental **t** conformer starting from the **r″** conformer (minimal deviation of 5.1 Å) with the major difference being a small twist of the apical domain. Conversely, the open holo simulation consistently shows an open structure with RMSd values closer to the open **r″** than to the closed X-ray structure ([Figure 3D](#pcbi-1002004-g003){ref-type="fig"} and [Table 1](#pcbi-1002004-t001){ref-type="table"}). Stabilization of equatorial domain upon nucleotide binding {#s2c} ---------------------------------------------------------- To probe the differences between the unliganded and ATP-bound simulations we investigated the residue fluctuations based on the 10--60 ns interval of the subunit simulations in the closed form. As expected, most residues show a similar fluctuation pattern in the holo and apo simulations. In the equatorial domain, residues adjacent to ATP show the most evident differences between the apo and holo simulations (see [Figure 4A--B](#pcbi-1002004-g004){ref-type="fig"}). In particular, residues Lys28, Lys34, Asp87, Asn457, Glu461, Tyr478-Glu483, and Tyr485, and its immediate neighbors are significantly () stabilized in the presence of ATP. Conversely, Glu61, Glu63, Arg421, and Asn475, show increased fluctuations in the holo simulations indicating rearrangement of these residues upon ligand binding. Modest differences in the fluctuation profiles are also observed within the apical (Figure S11A) and intermediate (Figure S11B--C) domains. Only two residues are shown to be significantly altered (); Arg350 (apical) and Lys390 (intermediate). ::: {#pcbi-1002004-g004 .fig} 10.1371/journal.pcbi.1002004.g004 Figure 4 ::: {.caption} ###### Residual fluctuations of the closed unbound and ATP bound simulations. Black bars show the RMSF values obtained from the ATP bound simulations, while red dots represent the unbound simulations. Binding site residues are depicted green, while residues that yield a significant change () between ATP-bound and ATP-free simulations are depicted with blue squares. Residues of particular interest are labeled. RMSF values are shown for the equatorial domain (residues 2-133 (A), 409-525 (B)). Secondary structure elements are indicated schematically with helices in black and strands in gray. :::  ::: Since the equatorial domain possesses the ATP binding site it is thus the epicenter for the initiation of the conformational changes. In order to relate the differences in fluctuations to binding of ATP we calculated the potential hydrogen bonds between ATP and the protein along the simulations. [Figure 5](#pcbi-1002004-g005){ref-type="fig"} shows the occupancy of each of the ATP hydrogen bonds in the closed holo simulations. Three hydrogen bonds are shown to be particularly strong with an occupancy of more than 90% of the simulation; Thr89 and Thr91 bind to the beta and gamma phosphate, respectively. The fifth single strongest bond is Asn479 which binds to the group of the adenine ring, with a occupancy of about 80% of the simulation. The negatively charged carboxyl group of Asp495 forms 4 alternating hydrogen bonds with the ribose OH, which have occupancies of approximately 40% each. Weaker bonds also exist, such as Gly32, Lys51, Thr90, and Ala480. Moreover, Asp87 is held in close contact with ATP through the tight binding to , resulting in a mean distance of 3.58 Å () between the oxygens of Asp87 and the three adjacent phosphate oxygens of ATP. ::: {#pcbi-1002004-g005 .fig} 10.1371/journal.pcbi.1002004.g005 Figure 5 ::: {.caption} ###### Schematic representation of potential hydrogen bonds between protein and nucleotide. Hydrogen bond donors and acceptors are labeled below the bars in which the height represents the occupancy in % of the simulation time. Each bar depicts the occupancy of one H-bond in one of the 6 ATP-bound simulations (simulations K-P in [Figure 2](#pcbi-1002004-g002){ref-type="fig"}). :::  ::: The identified hydrogen bonds between ATP and specific protein residues can be directly linked to the change in the fluctuation pattern observed in [Figure 4](#pcbi-1002004-g004){ref-type="fig"}. In particular hydrogen bonds involving Gly32, Thr89-Thr91, and Asn479-Ala480 might cause the stabilization of these areas. Moreover, the tight binding between Asp87 and might be a stabilizing factor for Asp87 and its neighbors. Configurational entropies of the closed simulations were estimated to investigate the entropic penalty upon ATP binding. Entropy values calculated from quasi-harmonic analysis of MD simulations are sensitive to simulation length and the number of frames in which the calculation is based on [@pcbi.1002004-Harris1]. Moreover, significant variation between individual MD has been reported [@pcbi.1002004-Caves1], [@pcbi.1002004-Clarage2]. We thus performed entropy calculations on the multiple closed MD simulations for the equatorial domain and for the whole subunit ([Figure 6](#pcbi-1002004-g006){ref-type="fig"}). We observe only small changes in entropy within the equatorial domain upon nucleotide binding. Moreover, during the first 60 ns of the simulations, no entropy difference is found for the whole subunit, while after 80--100 ns the entropy of the ATP-bound simulation is slightly higher than that of the apo simulations. ::: {#pcbi-1002004-g006 .fig} 10.1371/journal.pcbi.1002004.g006 Figure 6 ::: {.caption} ###### Configurational entropy. Configurational entropy of the equatorial domain (A) and the whole subunit (B) with respect to the simulation length. The calculation is based upon the 100 ns long simulations of the closed GroEL subunit; unbound (black), and ATP-bound (red). :::  ::: Differential atomic interactions {#s2d} -------------------------------- Comparing distance maps between different conformations has the potential of revealing unique atomic interactions potentially important for the conformational transitions within the GroEL subunit. We used difference contact maps (DCMs) to identify atomic interactions (all heavy atoms) unique to the **t**, **r**, and **r″** X-ray conformers (Figure S12A-E) complemented with DCMs obtained from the MD simulations of the closed form ([Figure 7](#pcbi-1002004-g007){ref-type="fig"}). Of particular interest are those residues which change interaction partner during the transition from **t** to **r″**. ::: {#pcbi-1002004-g007 .fig} 10.1371/journal.pcbi.1002004.g007 Figure 7 ::: {.caption} ###### Difference contact map between ATP-bound and unliganded MD simulations. The differences in side chain contacts are mapped for the MD conformers. Two residues are assumed to be in contact if the minimal distance between them . Only contacts existing in at least half of the structures are considered. Green triangles represent contacts present in the unliganded forms and not in the liganded forms. While blue dots represents contacts which are present in the liganded forms, but not in the unliganded forms. Red circles depict contacts which are found in both the X-ray and MD analysis. Secondary structure elements are indicated schematically with helices in black and strands in gray. :::  ::: A large number of contacts are found to be unique for either the **t** or the **r″** forms (Figure S12A). High density of differential atomic interactions are observed within the equatorial domain, and in particular the interaction pattern between helices B, C, and D is altered upon the **t**-**r″** transition. Contacts between helices B-C and C-D are found to be unique for the **r″** conformers. Conversely, the **t** conformers show tighter interactions between helix C and the loop connecting helix C and D. Here, Asn82 contacts with Asp87 in the **t** conformers but with Arg58 in the **r″**. Moreover, Asp87 changes its contact partners from Asn82 in the **t** state to Ser151 and Asp398 in the **r″** state, which is the main interaction between helix M of the intermediate domain and the equatorial domain. Unique for the **r″** forms is also a set of hydrophobic interactions between helices C and D (Val74-Ile100, Val77-Ala96, Ala81-Ala92, Val77-Ala92). Conversely, the **t** forms have unique contacts between helix C and R (Val74-Val510, Val77-Tyr506, Ala85-Val499). The intermediate domain also undergoes conformational rearrangements during the transition. Residues downstream of helix G initially interact with residues of helix M in the **t** form, but change interaction with helix L and K during the transition, e.g. the interaction Glu172-Arg404 in the **t** forms is changed to Glu172-Arg350 in the **r″** forms. Moreover, Asp179 contacts Lys390 at the intermediate domain in the **t** forms, and the Thr48 at the equatorial domain in the **r″** forms. While the ensemble of static X-ray structures provides the differential atomic interactions between the two end points of the reaction cycle, the MD simulations can reinforce the findings based on X-ray and further complement this by providing the differential atomic interactions at an earlier phase in the transition. Consistent with the findings from the X-ray-based DCM, the areas of high density differential contacts are situated in the equatorial and intermediate domains ([Figure 7](#pcbi-1002004-g007){ref-type="fig"}). Twenty differential contacts are consistently identified ([Table 2](#pcbi-1002004-t002){ref-type="table"}). Perhaps the most interesting of these are the interactions Val412-Asn475 and Leu134-Asn475 located near the lower hinge ([Figure 8A](#pcbi-1002004-g008){ref-type="fig"}). These interactions show similar average distance difference in the X-ray and MD distance matrices. Moreover, they are located in the equatorial domain, close to the lower hinge, making them potentially important for the interaction between the equatorial and intermediate domain. ::: {#pcbi-1002004-g008 .fig} 10.1371/journal.pcbi.1002004.g008 Figure 8 ::: {.caption} ###### Local conformational changes in the equatorial domain upon nucleotide binding. X-ray structures of **t** and **r″** are shown in green and cyan colors, respectively. (A) shows the opposite side of the binding site where nucleotide binding increases the distance between Leu134 and Asn475. The closer contact between Lys34 and Glu483 brought about by nucleotide binding is shown in (B). ADP is shown in stick and surface representation. :::  ::: ::: {#pcbi-1002004-t002 .table-wrap} 10.1371/journal.pcbi.1002004.t002 Table 2 ::: {.caption} ###### Differential atomic interactions upon nucleotide binding obtained from X-ray structures and MD simulations. ::: {#pcbi-1002004-t002-2} Interaction ------------- ------- ------ ------ D83 K327 −30.4 −4.4 \- V177 R322 −21.9 −2.2 \- V412 N475 −1.8 −2.2 −1.8 L134 N475 −1.8 −2.0 −1.8 D179 K390 −2.7 −0.9 \- T176 V376 −1.9 −0.3 \- K7 D11 −0.8 −0.9 −0.8 L23 K75 −0.6 −0.9 \- V177 L400 −1.0 −0.3 \- L17 T101 −0.7 −0.6 \- N479 M488 −0.7 −0.2 −0.7 A85 V499 −0.5 −0.2 \- V412 I489 \- −0.8 −0.4 R231 E310 0.6 0.3 \- K65 L524 0.8 0.6 \- L31 T91 0.8 0.8 0.8 A84 Y506 1.0 0.7 1.0 V77 A507 \- 0.8 0.9 V77 A92 2.2 1.0 \- V54 A78 3.5 1.4 2.2 The average difference in distance () is shown for each of the three calculations; between X-ray **t** and **r″** (column 1), MD simulations of **t** with and without ATP (column 2), and X-ray **t** and ATP bound **t** conformers (column 3). Negative values (above mid rule) depict interactions present (or closer) in the nucleotide free state (unliganded), while positive values (below mid rule) depict interactions closer in the nucleotide bound state. ::: In the intermediate domain Asp179 and Lys390 connect helix M to three sheets (, , ) close to the apical domain in the **t** forms. These residues are 0.9 Å closer in the apo simulations than in the holo ones. This is consistent with the X-ray DCMs, which shows that Lys390 interacts with Thr48 in the equatorial domain. The distance matrices can also be helpful to probe the underlying mechanisms for the changes in fluctuation pattern of the equatorial domain ([Figure 4A--B](#pcbi-1002004-g004){ref-type="fig"}). [Table 3](#pcbi-1002004-t003){ref-type="table"} summarizes the average difference in atomic distances for residues identified to have altered fluctuations. Of particular interest is Arg421 and Asn475 which show higher RMSF values in the holo simulations. Arg421 shows a tighter binding to Gly471 in the apo simulations than in the holo ones (). This is consistent with X-ray data, although the difference is about 1 Å smaller ( and ). Perhaps the most evident difference affects Asn475 which in the MD simulations appears 2.2 Å closer to Val412, 2.0 Å closer to Leu134, and 1 Å closer to Ala413. These findings are again consistent with the differences found in the X-ray structures; i.e. 1.8, 1.8, and 0.8 Å for the **t**-**r″** difference, and 1.4, 2.5, and 0.7 Å for the **t**-**r** difference. ::: {#pcbi-1002004-t003 .table-wrap} 10.1371/journal.pcbi.1002004.t003 Table 3 ::: {.caption} ###### Differential atomic interactions potentially involved in the stabilization of the equatorial domain. ::: {#pcbi-1002004-t003-3} Interaction ------------- ------ ------ ------ V412 N475 −1.8 −2.2 −1.4 L134 N475 −1.8 −2 −2.5 R421 G471 −0.5 −1.5 −0.4 V54 T89 0.5 −1.2 0.8 A413 N475 −0.8 −1 −0.7 L17 T101 −0.7 −0.6 −0.2 G53 T89 1.4 −0.5 0.8 V469 Y478 −0.1 −0.4 −0.2 Y485 M491 0.3 −0.3 −0.5 V54 S79 2.6 0.4 0.3 L17 I100 −0.2 0.5 −0.4 K28 V94 −0.5 0.6 0.1 V27 V56 0.2 0.7 0 V54 A78 3.5 1.4 1.7 K34 A480 0.4 2.2 0.7 K34 A480 0.4 2.2 0.7 K34 E483 0.8 2.7 1 See [Table 2](#pcbi-1002004-t002){ref-type="table"} for legend. ::: Interactions between 2-loop-3 and 16-loop-17, close to the ATP-binding site, are strengthened upon nucleotide binding which might explain the smaller RMSF values in this area ([Figure 8B](#pcbi-1002004-g008){ref-type="fig"}). The interaction Lys34-Glu483 is particularly affected; the average distance difference is 2.71 Å for the MD structures, while 0.8 and 1 Å for the X-ray structures. A significant difference is also found for Val54-A78 which are 1.4 Å closer in the holo simulation than in the apo simulations (3.5 and 1.7 Å for the X-ray differences). Determining the residue cross correlation to investigate whether the motions of one residue are related to the motions of another can aid in deciphering the underlying mechanisms for the observed conformational transitions. We have previously reported the cross-correlation map for the entire GroEL subunit revealing global correlations [@pcbi.1002004-Skjaerven2]. In the present work we focus on the equatorial domain which has the advantage of revealing more detailed correlations around the binding site. The corresponding residue-residue correlation map for the equatorial domain is shown in [Figure 9A](#pcbi-1002004-g009){ref-type="fig"}, and highlights 7 off-diagonal correlated areas. These areas are consistently described as correlated throughout all subsets of the simulation (10, 20, and 50 ns intervals), and are the following: (1) residues Lys16-Gly20 (helix A) and Phe66-Asn68 (helix C), including two positively charged residues in helix A, and one negatively charged residue in helix C; (2) residues Gln454+Ile455+Asn458 (helix P) and Thr31-Gly33 (loop region between helix A and 2); (3) The latter region is also correlated to the phosphates of ATP; (4) the phosphates of ATP and residues Gly89-Thr92 (helix D); (5) residues Asp115-Ile119+Ala123 and Asp435-Asn437+Ile440+Ala443; (6) Val411-Val412 (15) and Leu494-Thr497 (18) close to the lower hinge; and (7) the adenine ring of ATP and residues Asn479-Thr482 (16-loop-17 region). ::: {#pcbi-1002004-g009 .fig} 10.1371/journal.pcbi.1002004.g009 Figure 9 ::: {.caption} ###### Residue correlated motions of the equatorial domain. \(A) Cross-correlation map of the closed ATP bound simulation reveals residues that fluctuate in the same direction at the same time. The color scale runs from pink (anticorrelated motions) to blue (correlated motions). Red circles mark areas that are consistently found as correlated over all sets of 10 ns intervals of the trajectory. Secondary structure elements are indicated schematically with helices in black and strands in gray. (B) Mapping of the correlated residues onto the GroEL equatorial domain. Red lines link two correlated residues. ATP is colored blue. :::  ::: Discussion {#s3} ========== Extensive unbiased MD simulations {#s3a} --------------------------------- Several computational structural biology attempts have focused on the investigation of the ligand-induced conformational changes in GroEL [@pcbi.1002004-Zheng1], [@pcbi.1002004-Ma3], [@pcbi.1002004-Hyeon1], [@pcbi.1002004-Sliozberg1], [@pcbi.1002004-Abrams1]. Comparison of crystal structures at different stages of the functional cycle, NMA analysis and targeted MD simulations provide important information on the conformational space and motions associated to ligand-binding. Nevertheless, all-atom unbiased MD simulations have a superior potential to reveal the mechanisms involved in conformational transitions in proteins since using targeted MD simulations applying fictitious driving forces bias the motions toward the target and may drive the transitions along unrealistic deformations. However, GroEL has been considered to be beyond the analysis by unbiased MD simulations which have been held back due to the large size of the protein. Nevertheless, Sliozberg and Abrams have previously performed a 20 ns long unbiased MD simulation of one subunit, investigating the ATP-driven conformational changes [@pcbi.1002004-Sliozberg1]. This simulation manages to sample the transition from the **t** unbound, tense conformation to **r**, defined as a semi-relaxed conformation intermediate towards the open **r″** conformation. This transition includes a counterclockwise rotation of the apical domain along with closing of the nucleotide binding pocket, and was associated to the initial transition in a 2-stage activation by ATP [@pcbi.1002004-Ranson1], [@pcbi.1002004-Ma3]. The **T** to **R** transition has been associated to the intra-ring cooperative allosteric response to ATP binding, since functionally the **R** conformer shows high affinity for ATP [@pcbi.1002004-Ranson1]. Following this reasoning Sliozberg and Abrams interpreted their results on the way the **r** structure would be concertedly transmitted to all seven subunits in the cis ring. In our present simulations of the isolated subunit, we sample close to the **r** structure (minimum deviation of 2.6 Å), and in addition, these long MD simulations (300 ns) provide so far unexplored insight into the ATP-induced transition from **r** all the way to a structure that is, according to the RMSd values, close to **r″** (minimum deviation of 5.8 Å). An almost full transition in the opposite direction, from **r″** to **t**, is revealed from a similar simulation of the ligand free apo state starting from the **r″** structure (minimum deviation of 5.1 Å). This transition would correspond to the conversion occurring in the trans ring, which functionally is accompanied by ejection of the hydrolysis product ADP and the co-chaperonin GroES. We thus show for the first time with unbiased MD simulations that the conformational transitions in the GroEL oligomer are favored by the intrinsic behavior of the isolated GroEL subunit. The predictive power of the present study is reinforced by the multiple 100 ns long MD simulations. The conformer plots and the fluctations along the simulations also reveal that the **r** conformation is sampled both in the presence and absence of ATP. Thus, our results further add to the accumulating experimental and theoretical proofs of the pre-existing equilibrium between inactive and active states (see e.g. [@pcbi.1002004-Teilum1]--[@pcbi.1002004-Cui2]). A similar trend was also detected in a recent fluorescence study where the authors were able to measure the fraction of molecules in the **T** and **R** state of GroEL with increasing ATP concentrations [@pcbi.1002004-Frank1]. They determined that about 50% of the molecules were in the **T** state even with high ATP concentrations, indicating a constant cycling between the two conformations. Moreover, Chaudhry et al. showed that the inherent motions of the unliganded GroEL were "biased along the transition pathway that leads to the folding-active state" [@pcbi.1002004-Chaudhry1]. X-ray structures of the nonhydrolyzable ATP analog , in which the domain rotations are not observed [@pcbi.1002004-Boisvert1], give means to further support a **t** to **r** cycling. These unbiased simulations also show that the open structure, similar to the **r″** structure, is attained for the isolated subunit, just in the presence of ADP, and without the need of the co-chaperonin GroES. It is accepted that binding of GroES to the heptameric cis ring with the subunits in the **r** conformation is the main functional determinant for the induction of the **r″** conformer, triggering the encapsulation of bound protein substrate into the cavity. Our results thus support that the subunit conformational changes along the functional cycle of GroEL, as revealed by X-ray crystallography and cryoelectron microscopy, are largely intrinsic to the 3D structure (sequence-structure-dynamics) of the subunit (intramolecular), though stabilization of intermediate conformations should be associated to intersubunit or interprotein (intermolecular) interactions in the GroEL-GroES oligomers. Isolating a monomer from its natural biological assemblage may result in an altered dynamical behavior. However our results indicate that the dominant motion of a single GroEL subunit resembles that occurring in the ensemble of GroEL complex crystal structures. Thus, the dynamics of the subunit alone seems to be reflected in the larger biological assembly. This is seen by the RMSd analysis of the trajectories ([Figure 3](#pcbi-1002004-g003){ref-type="fig"}), as well as in the comparison between the X-ray and MD derived PCs (root mean squared inner product = 0.73) (Figure S13). The dominance of this intrinsic motion might be related to the relatively small inter-subunit contact area (6.6%, or ) [@pcbi.1002004-Braig1], which is comparatively low for an oligomeric protein [@pcbi.1002004-Miller1], [@pcbi.1002004-Janin1]. Nevertheless, each of the 7 apical domains interacts with two neighboring apical domains, as well as with one intermediate domain in the initial closed **T** form [@pcbi.1002004-Braig1]. This arrangement imposes constraints on the motion of the subunit, though at present to an unknown degree. It is therefore expected that, for similar timescales, simulating the entire GroEL assembly would lead to a more restricted conformational sampling of the subunit which would most likely be more consistent with the timescale upon which these conformational transitions are thought to occur [@pcbi.1002004-Rye1]. Conformational changes {#s3b} ---------------------- Sliozberg and Abrams described that MgATP binding to the subunit in the **t** conformation initiates a conformational change mostly due to the strong interaction of and Asp398 [@pcbi.1002004-Sliozberg1]. This interaction induces a series of H-bond ruptures (such as the Asp155-Arg395 salt bridge) and H-bond formations that were described in an induced-fit, cascade-like way [@pcbi.1002004-Koshland1]. We were however unable to reproduce this pattern of interactions. As seen in the conformer plots ([Figure 2](#pcbi-1002004-g002){ref-type="fig"}) the effect of ATP binding is rather described by causing a shift of conformational equilibria towards the **r″** conformation, with the peculiarity that in this case the ATP-bound conformation corresponds to the open conformation of the subunit, while usually ligand binding leads to a closed structure for the majority of proteins. In fact, in GroEL, ATP binding leads to a closed and less dynamic equatorial-domain structure but open and more dynamic (**r** and **r″**) subunit structure. In the case of preexisting equilibria where the ligand binds selectively to an "active" conformation the energy barrier between the conformations in equilibrium should be low [@pcbi.1002004-Teilum1], and binding should not bring a high entropic penalty, as certainly seems to be the case for the GroEL subunit ([Figure 6](#pcbi-1002004-g006){ref-type="fig"}). Further experimental support for low energy barriers is also found for Hsp70 proteins [@pcbi.1002004-Taneva1]. Moreover, a thermodynamic study on nucleotide binding to GroEL provided a positive entropy change for the binding of ATP at temperatures higher than [@pcbi.1002004-Terada1]. Notwithstanding the fact that these measurements were performed on the complete oligomer, it is noteworthy to relate the trend in entropy change to that found for the isolated monomer. Detailed changes {#s3c} ---------------- Computational studies employing advanced procedures in conjunction with NMA and brownian dynamics have probed allosteric networks in the GroEL system [@pcbi.1002004-Hyeon1], [@pcbi.1002004-Chennubhotla1], [@pcbi.1002004-Yang2], [@pcbi.1002004-Tehver1]. They have been able to determine and highlight a large set of residues responsible for the transitions in the GroEL complex. Lys80-Asp359, Asp83-Lys327, Arg58-Glu209, Pro33-Asn153, and Gly257-Arg268 are among several intra-subunit interactions that have been identified to be important during the conformational transitions in GroEL [@pcbi.1002004-Hyeon1], [@pcbi.1002004-Yang2]. Of these, Arg58, Asp83, Gly209, and Lys327 were also highlighted by Tehver et al. [@pcbi.1002004-Tehver1], and shown to be highly conserved [@pcbi.1002004-Kass1]. Our study of the MD trajectories and the large amount of X-ray data reinforces these studies as we identify many of the same interactions and residues. Moreover, we identify several other atomic differential interactions brought about by ATP-binding which have not previously been detected. Of particular interest is the tight binding between Lys34 and Glu483 in the presence of ligand ([Figure 8A](#pcbi-1002004-g008){ref-type="fig"}), which is captured in both the MD simulations and the X-ray crystallographic data. This effect appears as an important rearrangement induced by ATP-binding. These charged residues are situated close to the nucleotide binding site in two separate loops (16-loop+17 and A-loop-2), which movements interestingly are highly correlated to ATP; residues Asn479-Thr482 to the adenine ring, and Thr31-Gly33 to the phosphates of ATP. The correlation between these areas and ATP is likely to contribute to the stabilization of these loops as observed by fluctuation calculations ([Figure 4](#pcbi-1002004-g004){ref-type="fig"}), which in turn might aid in the salt-bridge formation between Lys34 and Glu483 ([Figure 8B](#pcbi-1002004-g008){ref-type="fig"}). More peripheral to the ATP-binding site in the equatorial domain, close to the lower hinge, we observe weaker contacts between Leu134-Asn475, Arg421-Gly471, and Val412-Asn475 ([Figure 8A](#pcbi-1002004-g008){ref-type="fig"}). This is possibly due to rearrangements of helix N upon nucleotide binding which disrupts the initial interactions found in the **t** conformers and along our MD simulations of the unbound **t** forms. Moreover, the DCM analysis pinpoint an altered interaction pattern for helix C in the equatorial domain, possibly explaining the movements of helices A+C in the PCA analysis of X-ray structures. Among these interactions, hydrophobic amino acids are a common denominator; Ala84-Tyr506, Val77-Ala92, Val54-Ala78, Val77-Ala507 all show a tighter binding in the nucleotide bound conformers. Experimentally, this core of hydrophobic amino acids has been shown to be important, and e.g. the Ala92Thr protein vaiant shows a low ATPase activity [@pcbi.1002004-Kovacs1]. Interestingly, the MD simulations also capture the rupture of the Asp179-Lys390 contact within the intermediate domain, and the formation of Arg231-Glu310 contact in the apical domain upon ATP-binding, both consistent with X-ray data. Realizing that a deep understanding of the function of GroEL as a molecular machine certainly requires analysis of inter-subunit and inter-ring interactions in the oligomer, MD simulations of the complete GroEL complex would be required to probe the full effects of ATP-binding. While these are expected to reveal essential details on the allosteric regulation of GroEL function, multisubunit simulations of a system of 600.000 atoms are a very computationally-demanding task. Regardless of the capacity of GroEL to assemble into oligomeric structures, various studies have also shown the surprising facility to obtain stable folded monomers of GroEL by a variety of means [@pcbi.1002004-Lissin1]--[@pcbi.1002004-Horowitz1], thay might display a weak chaperone activity of the GroEL monomer [@pcbi.1002004-Taguchi1]. In the present work we have attained detailed and statistically proven results from MD simulations of the isolated GroEL subunit, which have highlighted the importance of considering the proper dynamics and response of the subunit in the context of the large scale transitions in the GroEL complex. Since the equatorial domain holds the ATP binding site, it constitutes the epicenter of the conformational changes in the GroEL complex. By paying extra attention to the detailed mechanisms in this region of the protein during the conformational transitions observed along the simulations we have been able to map important intramolecular rearrangements which potentially are a prerequisit for the intermolecular transitions to occur. Materials and Methods {#s4} ===================== Structural analysis {#s4a} ------------------- All available GroEL structures were collected from the RCSB protein databank [@pcbi.1002004-Berman1]. Six structures were omitted due to large amount of missing coordinates (PDB-codes: 2CGT, 1GR5, 1IOK, 2C7C, 1GRL, 3CAU). A total of 21 crystal structures (287 GroEL subunits) were kept for further analysis (PDB-codes: 1PCQ, 1PF9, 1SVT, 3C9V, 1AON, 1GRU, 1MNF, 1XCK, 2C7D, 2NWC, 3E76, 2EU1, 1SS8, 1SX3, 1J4Z, 1KPO, 2C7E, 1KP8, 1OEL, 1WE3, 1WF4). Structural superposition was performed on the invariant "core" as defined by Grant et al. [@pcbi.1002004-Grant1]. These structures were collected and analyzed using the Bio3D package [@pcbi.1002004-Grant1]. Principal component analysis {#s4b} ---------------------------- The calculation of the PCA modes involves two main steps; (1) the calculation of the covariance matrix, , of the positional deviations, and (2) the diagonalization of this matrix [@pcbi.1002004-Ichiye1], [@pcbi.1002004-Amadei1]. The dimensional covariance matrix is calculated based on an ensemble of protein structures, and the elements of are defined aswhere and are atomic coordinates and the brackets denote the ensemble average. The diagonalization of the symmetric matrix involves the eigenvalue problemwhere is the eigenvectors and the associated eigenvalues. Our PCA calculations were based on the C coordinates of the ensemble of the 287 GroEL crystal structures and was performed with the Bio3D package [@pcbi.1002004-Grant1]. The X-ray and MD conformers were projected into the sub-space defined by PC1 and PC2, where the maximum variation of the conformational distribution was observed. Plotting these projections results in 'conformer plots' which displays a low dimensional representation of the conformational change in terms of the two principal components. Molecular dynamics simulations {#s4c} ------------------------------ All-atom MD simulations of the closed and open forms of the isolated GroEL subunit were performed both with and without bound nucleotide (designated holo and apo, respectively). The atomic models were prepared from the high-resolution crystal structures with PDB codes 1XCK chain A [@pcbi.1002004-Bartolucci1] and 1SVT chain A [@pcbi.1002004-Chaudhry1], for the closed and open forms, respectively. MgATP coordinates for the closed holo (with nucleotide) simulation was collected from the crystal structure with PDB id 1KP8 chain A [@pcbi.1002004-Wang1]. All atomic models were prepared with Amber10 [@pcbi.1002004-Case1] and the corresponding Amber03 forcefield [@pcbi.1002004-Duan1], [@pcbi.1002004-Lee1]. ATP and ADP parameters were obtained from Meagher et al. [@pcbi.1002004-Meagher1]. For each of the simulations, the protein was solvated in a periodic truncated octhahedron box with TIP3 water molecules [@pcbi.1002004-Jorgensen1], providing 16 Å of water between the protein surface and the periodic box edge. The solute was minimized for 10,000 steps, followed by 10,000 steps of minimization of the entire system. The protein was then heated to 100 K with weak restraints for 100 ps, and to 300 K in 200 ps. 2 ns of equilibration with constant pressure and temperature (NPT) of the system was performed prior to the production run in order to ensure correct density. The production runs were performed with constant volume and energy (NVE) with a 1 fs time step, using SHAKE constraints on hydrogen-heavy atom bonds. A total of 14 MD simulations of the GroEL subunit were carried out. Of these, we performed four 300 ns long simulations: (A) closed **t** unbound, (B) open **r″** unbound, (C) closed **t** ATP-bound, and (D) open **r″** ADP-bound. Additionally ten 100 ns long simulations of the closed **t** form were performed: 5 (E-J) unliganded, and 5 (K-P) ATP-bound. Simulations (D) and (K) consist of the 0--100 ns interval of simulations (A) and (C), respectively. Difference contact maps {#s4d} ----------------------- Distance matrices were calculated between all 3856 pairs of atoms in order to monitor the atomic interactions (between residues at least four residues apart in sequence) [@pcbi.1002004-Grant1]. The distance matrix was issued to residue grouping by only considering the minimal atomic distance between the residue pairs. Residue pairs closer than 4 Å are assumed to be in contact, and constitute the contact matrix for one particular conformation. Contact matrices were calculated for 28 closed (T) apo subunits (PDB code: 1XCK, 1SS8, 1OEL), 28 closed (T) ATP bound subunits (PDB code: 1KP8, 1SX3) and 28 open (R″) ADP bound subunits (PDB codes: 1AON, 1SVT, 1SX4, and 1PF9), and 6000 snapshots obtained from the last 50 ns of the 12 independent MD simulations on the closed GroEL subunit (6 apo and 6 holo). Only contacts with at least 50% occupancy and an average distance difference of 0.5 Å were considered. The difference of two contact maps (DCM), i.e. difference between apo and holo, then defines side-chain contacts which exist in one form, but not in the other. Additional analysis of simulated data {#s4e} ------------------------------------- Clustering analysis, correlation maps, entropy calculations, and hydrogen bond analysis were performed with the the ptraj module of AmberTools [@pcbi.1002004-Case1]. Clustering was performed on the MD conformers using the average-linkage clustering algorithm [@pcbi.1002004-Shao1]. Cross-correlation calculations were performed on several subsets of the closed ATP bound simulations (10, 20, and 50 ns intervals) in order to obtain areas of consistent correlations. All figures of GroEL are made in Pymol [@pcbi.1002004-DeLano1]. Supporting Information {#s5} ====================== Figure S1 ::: {.caption} ###### Statistics of projections of the MD conformers onto the first principal component. Each bar represents the projections (mean, median, or the last frame) of the simulations (only the 50--100 ns interval) onto the first principal component. The 6 apo and 6 holo simulations are depicted with grey and black bars, respectively. The full projections are provided in [Figure 4](#pcbi-1002004-g004){ref-type="fig"}. (EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S2 ::: {.caption} ###### Residual fluctuations of the closed unbound and ATP bound simulations for the intermediate and apical domains. Black bars show the RMSF values obtained from the ATP bound simulations, while red dots represent the unbound simulations. Binding site residues are depicted green, and the residues having the most significant changes are marked with labels. RMSF values are shown for the intermediate domain in (A) (134-190, 377-408), and apical domain (B) (191-376). (EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S3 ::: {.caption} ###### Difference contact maps between t, r, and r″ conformers. Difference in side chain contacts are mapped for the X-ray structures (A-E), and MD structures (F). Two residues are assumed to be in contact if the minimal distance between them is . Only contacts existing in at least half of the structures are considered. Green triangles represent contacts present in the unliganded forms and not in the liganded forms. While blue dots represents contacts which are present in the liganded forms, but not in the unliganded forms. Orange lines display residues in which a change in contact partner has been detected. Red circles depict contacts which are found in both the X-ray and MD analysis. Secondary structure elements are indicated schematically with helices in black and strands in gray. (EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S4 ::: {.caption} ###### Overlap map of the X-ray and MD PCs. Squared overlap between pairs of eigenvectors obtained from X-ray and MD PCA are shown. Dark grey corresponds to values close to 1 (high similarity) while white depict overlap close to 0 (orthogonal). Root mean squared inner product (RMSIP) is provided according to ref. [@pcbi.1002004-Amadei2]. (EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: Video S1 ::: {.caption} ###### Time-dependent conformer plot of the 300 ns ATP-bound simulation. Projection of X-ray (contoured grey dots) and MD (blue) structures (along the simulation time) onto the two most significant principal planes defined by 287 GroEL experimentally obtained structures. The distribution of MD structures is depicted with density shaded blue points. Orange, red, and green dots represent the closed (**t**), semi-relaxed (**r**), and open (**r″**) X-ray structures, respectively. Gray dashed line depicts the average pathway of the conformational sampling. (AVI) ::: ::: {.caption} ###### Click here for additional data file. ::: The authors have declared that no competing interests exist. The Norwegian Research Council is acknowledged for CPU resources granted through the NOTUR supercomputing program (<http://www.notur.no/>) and Bergen Center for Computational Science for providing powerful computer facilities (<http://www.bccs.uni.no/>). Work at CSIC/UPV/EHU was financed by MICINN (Grant BUF2007-64452). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. [^1]: Conceived and designed the experiments: A. Muga, J. McCammon, N. Reuter, A. Martinez. Performed the experiments: L. Skjaerven, K. Teigen. Analyzed the data: L. Skjaerven, B. Grant, K. Teigen. Contributed reagents/materials/analysis tools: J. McCammon, N. Reuter, A. Martinez. Wrote the paper: L. Skjaerven, A. Muga, N. Reuter, A. Martinez.

PubMed Central

2024-06-05T04:04:19.644393

2011-3-10

PMC3053312

Introduction {#s1} ============ One important feature of embryonic and organ development is the collective migration of groups of cells [@pcbi.1002007-McMahon1]--[@pcbi.1002007-Lecaudey1]. In some cases during development, large groups of cells move in streams with each cell independently following chemotactic cues from the environment [@pcbi.1002007-Affolter1]. However, during the morphogenesis of organ systems, wound healing, and cancer metastasis, it is more common to find cells migrating as an adherent group [@pcbi.1002007-Lecaudey1]. Some examples of these motions are the movement of cells during the morphogenesis of the inner blastocyst [@pcbi.1002007-Trinkhaus1], epithelial cell migration at the rim of the optic and the invaginating thyroid placode [@pcbi.1002007-Hilfer1], tissue repair by keratinocytes moving across provisional wound matrix [@pcbi.1002007-Grose1], and the migration of border cells through the early ovary in *Drosophila* [@pcbi.1002007-StarzGaiano1]. Though the fundamental features of single cell motility are now understood at some level [@pcbi.1002007-Abercrombie1]--[@pcbi.1002007-Mitchison1], the physical underpinnings of the collective migration of groups of cells remains enigmatic. In this paper we focus on understanding the collective migration of epithelial cells during wound healing. Two separate mechanisms have been proposed to account for wound closure. In the first, a circumferential ring of actin bundles contract to draw the wound edges together [@pcbi.1002007-Martin1]. This mechanism has been demonstrated in embryonic chick wing buds; however, wound closure in adults is presumed to rely on the crawling motility of epithelial cells [@pcbi.1002007-Martin1], [@pcbi.1002007-Fenteany1]. The cellular mechanics of this latter form of wound healing remains unclear. We propose that the fundamental driving force behind this process is the generic migratory behavior of an individual cell, and, specifically, that the active contractile stress generated within polarized epithelial cells coupled with cell-cell adhesion is sufficient to explain many of the features observed during wound healing assays. Though there is significant variation in the biochemical composition of different crawling cells, the basic biophysical process of single cell crawling entails (i) cytoskeletal extension at the front of the cell; (ii) adhesion to the substrate, which is typically mediated by integrin; and (iii) advance of the rear [@pcbi.1002007-Abercrombie1]--[@pcbi.1002007-Mitchison1]. In addition to these fundamental features, crawling cells are almost always polarized [@pcbi.1002007-Lacalle1], [@pcbi.1002007-Omelchenko1] and are observed to exert a dipole-distributed traction stress on the substrate [@pcbi.1002007-Doyle1]--[@pcbi.1002007-Dembo1]. Neighboring cells can bind to one another through membrane-bound cadherin molecules [@pcbi.1002007-Chen1], [@pcbi.1002007-Lambert1]. We hypothesize that bulk cellular motions in tissue are strongly dependent on these general features and that other specific details of single cell crawling are less important. In wound healing assays, a gouge is made in a continuous monolayer of cells (often MDCK cells are used), and the rate that the cells fill in the artificial wound is measured. There are a number of features that are observed in these assays that suggest that the healing process is not solely reliant on biochemical signaling triggering the migration of cells into the denuded area. The cells in the wounded monolayer typically migrate in groups and maintain cell-cell contacts [@pcbi.1002007-Boca1]--[@pcbi.1002007-Farooqui1]. Cells many cell diameters away from the wound edge are motile [@pcbi.1002007-Boca1]--[@pcbi.1002007-Farooqui1], and the rate of migration away from the edge is observed to be inversely proportional to the distance from the margin [@pcbi.1002007-Farooqui1]. At the wound edge, cells do not always migrate perpendicular to the boundary [@pcbi.1002007-Poujade1], [@pcbi.1002007-Petitjean1], and cell division is not observed to play a strong role in closing the wound [@pcbi.1002007-Poujade1], [@pcbi.1002007-Farooqui1]. Interestingly, it is also observed that the wound border progression advances roughly proportional to time squared [@pcbi.1002007-Poujade1], [@pcbi.1002007-Rosen1]; i.e., the average boundary velocity increases proportional to time since wounding. In this study, we explore the possibility that the collective cellular migration that occurs during wound healing is largely a mechanical process. We develop a mathematical model that incorporates the bulk features of single migrating cells and cell-cell adhesions. We first apply the model to the average motion of a spreading strip of cells and identify the key physics that may drive the advance of the cells at the wound edge. We then consider the healing of circularly-symmetric wounds and show that the model can reproduce wound closure when rows of cells at the wound periphery are deactivated. Finally, we simulate two-dimensional wound healing assays and show that our model can reproduce the complex cellular flows and border advance that is observed in experiments. Taken together, these results show that wound healing may not require substantial biochemical signaling or even mechanisms to identify wounding, but rather may result from the typical dynamics of motile cells. Results {#s2} ======= A single cell based model for the collective migration of epithelial cells {#s2a} -------------------------------------------------------------------------- Inside a crawling eukaryotic cell, the actin cytoskeleton flows rearward at the front of the cell and forward at the rear of the cell [@pcbi.1002007-Watanabe1]. Nascent and/or mature focal adhesions, which include integrin, link the cytoskeleton to the substrate or extracellular matrix (ECM) [@pcbi.1002007-Choi1], and thereby convert the cytoskeletal flows into traction stresses that are applied to the substrate [@pcbi.1002007-Doyle1], [@pcbi.1002007-Dembo1], [@pcbi.1002007-Roure1]. Like the actin velocity, the force that the cell exerts on the substrate is rearward at the front and forward at the rear; i.e., it is distributed like a dipole ([**Figure 1**a](#pcbi-1002007-g001){ref-type="fig"}) shows the traction stress inside a cell that is polarized along the direction d) [@pcbi.1002007-Doyle1], [@pcbi.1002007-Dembo1], [@pcbi.1002007-Roure1]. These dipole-distributed traction stresses, σ~d,~ lead to a net thrust force F that propels the cell at roughly constant velocity ([**Figure 1**b](#pcbi-1002007-g001){ref-type="fig"}). For example, isolated MDCK cells plated on a substrate spread to be about 20 µm long and crawl at speeds of about 10 µm/hr [@pcbi.1002007-Boca1]; the magnitude of the traction stress that the cell exerts on the substrate is of order 3×10^4^ dynes/cm^2^ [@pcbi.1002007-Roure1]. The turnover rate of integrin inside focal adhesions is on order of a minute [@pcbi.1002007-Cluzel1], and, therefore, integrin turnover is fast compared to the crawling speed of the cell, which allows us to treat the interaction between the cytoskeletal flows and the substrate as a resistive drag force that is proportional to the velocity, with drag coefficient ζ [@pcbi.1002007-Leibler1]. ::: {#pcbi-1002007-g001 .fig} 10.1371/journal.pcbi.1002007.g001 Figure 1 ::: {.caption} ###### Schematic of the model depicting the cell orientation and forces. \(a) Cells are aligned along the direction **d** and move with velocity **v**. Differences in velocity between neighboring cells produce a viscous stress. Neighboring cells preferentially align with differences in the orientation producing a torque on neighboring cells. Cells exert a dipole-distributed stress on the substrate and also on neighboring cells. (b) Each cell exerts a net force -**F** on the substrate. An equal but opposite force is exerted back on the cell. The substrate also exerts a drag force -ζA**v** on each cell. The thrust force and the drag force are offset, which produces the effective dipole stress on the cell. :::  ::: Epithelial cells that are in close contact can adhere to one another through cadherin molecules [@pcbi.1002007-Chen1], [@pcbi.1002007-Lambert1]. The turnover of cadherin molecules in cell-cell adhesions is on the order of tens of minutes to an hour, which is significantly slower than the turnover rate of integrin in focal adhesions [@pcbi.1002007-Lambert1]. For timescales less than this turnover time τ, neighboring cells are effectively stuck together. A tissue of cells should therefore behave like an elastic solid on short timescales. On longer timescales, though, cadherin turnover allows the cells to slide with respect to each other, and the bulk tissue should behave more like a fluid with viscosity η. Therefore, the stress between cells is maintained on times shorter than τ, but dissipates on longer timescales. Cells in monolayers overlap [@pcbi.1002007-Farooqui1] and the initial 10--12 hours of the dynamics of wound healing are not dependent on cell division [@pcbi.1002007-Poujade1], so we do not track the density in our model. We assume, though, that changes in density are resisted by a different effective viscosity than shear displacements and define a volumetric viscosity (λ - η/2). The intercellular stress σ~c~ can then be described with the Maxwell model,which is a simple model for viscoelastic fluids. In (1), denotes the identity matrix. Our choice of this cell-cell interaction model is justified by analyzing the behavior of two solid objects connected by spring-like adhesion molecules, which is based on a model for muscle cross-bridges developed by Lacker and Peskin [@pcbi.1002007-Lacker1] (See supplemental **[Text S1](#pcbi.1002007.s001){ref-type="supplementary-material"}**B for a complete description of how this model leads to the Maxwell model). For crawling cells, the resistive drag forces are large compared to the inertial terms. Therefore, the sum of all of the forces acting on a cell must be equal to zero. In our model, we consider four types of forces that act throughout the monolayer. First is the force produced by the intercellular stress that is described above (Eq. 1). The second force is due to the internal stresses that are generated inside single cells. This stress, which we denote by σ~d~, includes the viscoelastic stress of the cytoskeleton, as well as the active stresses from actin dynamics and molecular motors, such as myosin. For our model, we consider that this stress is largely dipole-distributed along the polarization direction of the cell and set it equal to its average value *f* ~0~ *b* **dd**, where *f* ~0~ is the dipole force and *b* is the dipole length. The actin flow inside a cell interacts with the substrate through adhesions and produces the thrust force **F** against the substrate. Finally, motion of the cell with respect to the substrate is resisted by drag forces, which are also due to the cell-substrate adhesions. We average the internal forces that are generated by a cell and balance these with the average external applied forces on the cell, which provides a mean-field dynamic equation governing the flow of the cells (for complete details, see [Text S1A](#pcbi.1002007.s001){ref-type="supplementary-material"}):where **f** = **F**/*A* is the thrust force per unit area, and *A* is the area of a cell. In this model, we assume that the magnitude of the thrust force is a constant. The velocity v in Eq. 2 defines the average local velocity of the cells in the monolayer. To complete the biophysical description of the epithelial cell layer, we must define the dynamics of the cell polarization. We assume that changes in cell orientation are driven by the mechanical interactions between cells as they move in the monolayer. We consider two torques that act to determine cellular orientation. First, the polarization of the cells combined with the cell elasticity favor alignment of neighboring cells ([**Figure 1**](#pcbi-1002007-g001){ref-type="fig"}). When neighboring cells are not aligned, there is a restoring torque that acts to align them. Therefore gradients in the orientation produce an elastic torque similar to the torque on a nematic liquid crystal. For this model, we use a single Franck constant, *K*, to describe the magnitude of the elastic restoring torque. Second, a resistive drag torque impedes the reorientation of the cells and is proportional to the time rate of change of the orientation vector. The re-orientational dynamics are then similar to that for nematic liquid crystals [@pcbi.1002007-DeGennes1]: Here ζ~r~ is a drag coefficient, and v is the velocity field for the cells. The second term on the lefthand side represents changes in orientation due to advection. The third term represents rotation of the polarization due to the motion of the cells (see **[Text S1](#pcbi.1002007.s001){ref-type="supplementary-material"}**C for more details). Eqs. 2 and 3 are similar to equations that have been used to describe the collective swimming of bacteria [@pcbi.1002007-Wolgemuth1]. In these systems, complex flow patterns are observed that are characterized by transient vortices and jets [@pcbi.1002007-Dombrowski1]. Parameter estimation from experimental data {#s2b} ------------------------------------------- The model that is described in the previous section has a total of 7 parameters (the dipole stress strength, the substrate drag coefficient, the magnitude of the thrust force, the cadherin turnover rate, the viscoelastic shear and volumetric viscosity, and the ratio of the Franck constant to the substrate drag coefficient). Because the model is based on the biophysics of single cell motility, many of these parameters can be estimated from experimental data. The magnitude of the dipole stress is given by the length of the cell (∼10 µm [@pcbi.1002007-Boca1]) times the traction stress measured in experiments (∼ 10^4^ dynes/cm^2^ [@pcbi.1002007-Roure1]). The traction stress is also equal to the substrate drag coefficient, ζ, times the local cytoskeletal velocity. At the rear of the cell, the velocity is equal to the average single cell crawling speed (10 µm/hr [@pcbi.1002007-Boca1]), and, therefore, ζ is approximately 100 pN·hr/µm^3^. At the front of the cell, the retrograde flow rate (1.5 µm/hr [@pcbi.1002007-Watanabe1]) and, therefore, ζ is approximately 500 pN·hr/µm^3^. The average value of ζ should be between these two values. The magnitude of the thrust force, F, is the substrate drag coefficient times the average cell speed multiplied by the area of the cell, which is approximately 10^5^ pN. As mentioned previously, the cadherin turnover rate, τ, is between 15 minutes and 1 hour [@pcbi.1002007-Lambert1]. Much less is known about the final three parameters in the model. We treat the cell-cell shear viscosity as a free parameter and make the assumption that the volumetric viscosity is about ten to one hundred times larger than this value. To estimate the Franck constant, we use the results from wound healing assays where the polar order parameter, which describes the orientation of the velocity field with respect to the border, was measured as a function of time [@pcbi.1002007-Petitjean1]. In these experiments, it is observed that the order parameter asymptotes to a fixed value in about 20 hours. If we assume that this orientational ordering arises due to the preference for neighboring cells to align, then this timescale should be approximately equal to the elastic relaxation timescale for the orientational dynamics. Therefore, ζ~r~ *L* ^2^/4π^2^ *K* should be approximately 20 hrs. Using the observed velocity correlation lengthscale, *L* ∼ 200 µm [@pcbi.1002007-Poujade1], [@pcbi.1002007-Petitjean1], we get that *K*/ζ~r~ is 50 µm^2^/hr. If we then assume that the orientational drag coefficient is predominantly due to sliding against the substrate, we expect that ζ~r~ ∼ ζ*b*/24  = 40 pN·hr/µm^2^, and *K* is then approximately 2×10^3^ pN. A simplified look at the advance of an initially-straight border {#s2c} ---------------------------------------------------------------- In typical wound healing assays, a relatively straight gouge is made through a monolayer of cells, and the rate that the cells fill in the denuded area is measured. To gain insight into the fundamental workings of our model, we consider a long strip of cells with vacant substrate bounding either side of the strip ([**Figure 2**a](#pcbi-1002007-g002){ref-type="fig"}). If we now average the dynamics of the cells over the entire length of the strip, assuming that the cellular orientations are isotropically distributed, then the orientational dynamics (Eq. 3) and the thrust force **F** average to zero. In addition, the net affect of the dipole-distributed stress is an over all compressive pressure that is exerted on to the substrate. An expansive pressure is, therefore, exerted back on the cell strip, and it is this pressure that drives the expanse of the cells into the wounded area. ::: {#pcbi-1002007-g002 .fig} 10.1371/journal.pcbi.1002007.g002 Figure 2 ::: {.caption} ###### Healing of a one-dimensional wound. \(a) Schematic of the 1D model for wound healing. A strip of cells (shown in green) aligned along the y-direction crawls to fill in the denuded area adjacent to them. Though the border advance is irregular, we consider the average advance of the border as a function of time (depicted by the red dashed line). (b) The average stress profile inside the band of cells. The stress is highest at the edge and the profile moves as a travelling front with the motion of the cells. (c) The net advance of the edge as a function of time. (d) The border progress exponent as a function of the viscosity and viscoelastic time constant. As observed in experiments, the border expands nonlinearly with time. At later times (∼20 hrs), the exponent has decreased to closer to one. :::  ::: The averaged dynamics in this simplified system is given by the balance of the forces due to the viscoelasticity of the cell-cell adhesions and the drag with respect to the substrate, driven out at the boundary by the effective dipole-induced pressure. The derivation of the resulting mathematical system is given in **[Text S1](#pcbi.1002007.s001){ref-type="supplementary-material"}**D. We solve the one-dimensional model with a free boundary that moves with the average velocity of the cells at the boundary. The 1D equations are integrated using a semi-implicit, finite difference scheme that is described in detail in **[Text S1](#pcbi.1002007.s001){ref-type="supplementary-material"}**F. All simulations used a time step of 0.001 hr and solved on a domain with 512 grid points. To compare to experimental data, we compute the net displacement of the boundary as a function of time. In addition, a number of experiments have observed that the net displacement of the border advances roughly proportional to time squared, when the width of the cell strip is larger than 200 µm [@pcbi.1002007-Poujade1], [@pcbi.1002007-Rosen1]. This "acceleration" of the border is somewhat counter-intuitive, as random cellular motions should lead to diffusive behavior that scales like the square root of time, and if the cells crawl at constant velocity, then one would expect the border advance to scale linearly with time. Interestingly, for cell strips with initial widths less than 200 µm, the border advance does scale roughly proportional to time [@pcbi.1002007-Poujade1]. In order to explore whether our model can explain this interesting bi-phasic behavior, we also compute the exponent of the time dependence of the border progression, *E*, as a function of time. In general, the model produces a uniformly distributed stress away from the boundary of the cell strip, with a sharp transition at the boundary, which reflects the stress generated by the expansive pressure. This stress profile is stable and moves along with the boundary, similar to a traveling front ([Figure 2b](#pcbi-1002007-g002){ref-type="fig"}). In this simplified system, the model depends on two parameters, the cadherin turnover rate τ and the volumetric viscosity (λ+η/2). When the turnover rate is fast or the viscosity is high, then the transition region of the stress is larger than when the turnover rate is slower or the viscosity is lower. We find that the border progression exponent is greater than one (i.e., the border advances supra-linearly in time), regardless of our choice of parameters; however, for times over about 20 hours, the exponent asymptotes to one ([**Figure 2**c](#pcbi-1002007-g002){ref-type="fig"}). During the first ten hours, the exponent is maximal. The value of the maximum exponent and the value at 16 hours are shown in [**Figure 2**c](#pcbi-1002007-g002){ref-type="fig"}. The cadherin turnover time strongly influences the intermediate value of the border progression exponent, with larger turnover times increasing the exponent. The exponent *E* is not influenced strongly by the cell-cell viscosity, as the viscosity determines the overall speed of the cellular flow, but does not limit the expansion at early times ([**Figure 2**c](#pcbi-1002007-g002){ref-type="fig"}). We also find that the maximal border progression exponent is not dependent on the initial width of the cell strip. The healing of circular wounds {#s2d} ------------------------------ Though a standard wound healing assay makes a long, relatively straight gouge through an intact monolayer of cells, many wounds in vivo are more localized and have a circular or ellipsoidal geometry. In fact, these roughly circular-shaped wounds can allow actin contraction at the wound periphery to close the wound (i.e., the purse-string mechanism [@pcbi.1002007-Martin1]). Wounds with these geometries can be made in vitro by scraping a micro-injection needle over the surface using a micro-positioner [@pcbi.1002007-Fenteany1]. MDCK cell monolayers wounded in this fashion with an ellipsoidal-shaped wound with minor axis of ∼100 µm close in approximately 10 hours [@pcbi.1002007-Fenteany1]. Interestingly, inhibiting Rac, a protein that is associated with formation of lamellipodia, in a single row of cells at the wound edge does not stall wound closure; however, inhibiting Rac in the first three rows does. Therefore, the purse-string mechanism is not required for wound healing in this system, and the crawling motility of cells away from the wound edge can drive wound closure. To test whether our model can reproduce these findings, we consider a circular-shaped wound with an initial radius, R~0~, which we set to be 100 µm ([**Figure 3**a](#pcbi-1002007-g003){ref-type="fig"}). By averaging the equations about the circumferential direction, we can solve a one-dimensional model that describes the average closure dynamics of this circular wound (See **[Text S1](#pcbi.1002007.s001){ref-type="supplementary-material"}**E for the mathematical details). The equations are integrated using a semi-implicit, finite difference scheme that is described in detail in **[Text S1](#pcbi.1002007.s001){ref-type="supplementary-material"}**F. All simulations used a time step of 0.001 hr and solved on a domain with 512 grid points. We use a random initial condition, and determine how the closure time τ~c~ depends on the model parameters. We find that the parameter that influences the closure time the most is the viscoelastic time constant τ ([**Figure 3**b](#pcbi-1002007-g003){ref-type="fig"}), with larger values of τ leading to longer closure times. Varying the Franck constant *K* does not produce a statistically significant effect on the closure time ([**Figure 3**b](#pcbi-1002007-g003){ref-type="fig"}), and neither does varying the viscosities η and λ. ::: {#pcbi-1002007-g003 .fig} 10.1371/journal.pcbi.1002007.g003 Figure 3 ::: {.caption} ###### Healing circular wounds. \(a) Schematic of a circular wound. A circular-shaped wound with radius R~0~ is made in an intact cell monolayer. In a small region of size Δ about the wound, cellular actin dynamics is deactivated by Rac inhibition. The time to close the wound, τ~c~, can be predicted from the model. (b) When Δ = 0, the model predicts that the closure time is strongly dependent on the viscoelastic timescale τ, which is set by the turnover rate of cell-cell adhesions. The Franck constant *K*, which defines the preference for neighboring cells to align, does not produce a statistically significant effect on the closure time (*K* = 0.064 (solid line), *K* = 0.256 (dashed line), and *K* = 1.024 (dotted line)). (c) Deactivation of the cells in a small region about the wound boundary leads to an increase in the closure time. For deactivation zones that are between 30--50 µm, we find that the wounds close intermittently. Above 50 µm, the wounds consistently fail to close. Error bars show one standard deviation in the closure time from simulations that were started with random initial conditions. :::  ::: In our model, we can also "de-activate" the actin dynamics in a band of cells that borders the wound by setting the thrust force and dipole stress terms to zero. We define a region of size Δ about the wound edge, where the actin dynamics are deactivated. By varying the width of this region, we simulate the closure of the wound and measure the average closure time as a function of the width. For widths of the deactivated zone up to 0.3 R~0~, we find that the average closure time increases ([**Figure 3**c](#pcbi-1002007-g003){ref-type="fig"}). For widths between 0.3--0.5 R~0~, it is not possible to define an average closure time, as the wounds do not close for all initial conditions. We define this as the intermittent closure regime. When the width is above 0.5 R~0~, we find that the wound always fails to close. Since typical experiments examined wounds that were of order of 100 µm, our model is in good agreement with the observation that deactivation of 3 rows of cells prevents wound closure. The two-dimensional dynamics of a wounded epithelial monolayer {#s2e} -------------------------------------------------------------- In the preceding sections, we have shown that our model can capture the motion of the boundary that accompanies wound healing for simple geometries. These simulations allow us to determine reasonable values of the unknown model parameters, such as the Franck constant and the effective viscosities that are due to cell-cell adhesion. Experiments that monitor the motion of cells in the two-dimensional plane of the substrate are able to visualize the complex flows of cells that accompany wound healing and the traction stresses that are exerted on the substrate during wound closure. As has been already mentioned, these experiments show vortical motion of the cells in the monolayer with long-range correlations in the velocity field over lengths of roughly 100 µm [@pcbi.1002007-Petitjean1], fingering at the wound boundary, supra-linear advance of the boundary with respect to time with an exponent that depends on the initial width of the epithelial monolayer [@pcbi.1002007-Poujade1], and high traction stress in the immediate vicinity of the boundary [@pcbi.1002007-Roure1]. Using the force and stress parameters that were estimated from experiments and the Franck constant and viscosities that were determined from our 1D simulations (see [**Table 1**](#pcbi-1002007-t001){ref-type="table"}), we solve the two-dimensional, free boundary problem of an infinite strip of cells with initial width *L* ~0~. The dynamic equations were discretized and solved using the Moving Boundary Node Method [@pcbi.1002007-Wolgemuth2]. This method is a level set-based, finite volume algorithm (further details of the numerical routine are given in **[Text S1](#pcbi.1002007.s001){ref-type="supplementary-material"}**). For these simulations, we used a time step of 0.001 hr and a grid spacing that was 1/40 of the initial width of the domain (i.e., for a monolayer with an initial width of 300 µm, the grid spacing was 7.5 µm). We initialize our simulations with zero stress and a random orientational field. We used multiple simulations to explore the variation that is caused by the random initial condition. These simulations show many of the features observed in the experiments. For example, the motion of the cells is spatially nonuniform; however, there exists long-range correlations in the velocity field over distances of 100--200 µm ([**Figure 4**a](#pcbi-1002007-g004){ref-type="fig"}). Transient vortices are also observed. Near the boundary the cells do not always move perpendicular to the boundary and the boundary shows characteristics of a fingering instability ([**Figure 4**a, d--f](#pcbi-1002007-g004){ref-type="fig"}). However, the fingering of the border that we observe in our simulations is not as pronounced as is sometimes observed in experiments. As in the 1D simulations, there is an increase in the traction stress that is exerted on the substrate, and this increased stress dies off within 10--20 µm from the boundary (see colorscale in ([**Figure 4**a, d--f](#pcbi-1002007-g004){ref-type="fig"}). ::: {#pcbi-1002007-g004 .fig} 10.1371/journal.pcbi.1002007.g004 Figure 4 ::: {.caption} ###### Complex flows and border progression in two-dimensional wound healing assays. \(a) A characteristic time course from one of our simulations with an initial width of 200 µm showing the local velocity of the cells (black arrows) and the traction force exerted against the substrate (colormap). Inside the cell-filled region, the cells move with complex dynamics, which include vortices and long-range correlations in the velocity field. The border advance is non-uniform and shows characteristics of a fingering-type instability. (b) The average advance of the border matches well data from experiments by Poujade, et al. [@pcbi.1002007-Poujade1] when the initial width of the cell-filled region is above 200 µm (red points). Also shown is a simulation with an initial width of 100 µm (blue points). As seen in experiments, smaller initial widths lead to a smaller exponent for the advance of the border. (c) The border advance exponent *E* increases with the initial width of the cell-filled region. The error bars show one standard deviation (*N* = 10). (d)--(f) Other characteristic internal flows and border shapes from simulations, highlighting the appearance of vortices and border fingering. All simulations use an initial width of 200 µm (horizontal blue lines). These simulations use the parameter values given in [**Table 1**](#pcbi-1002007-t001){ref-type="table"}. :::  ::: ::: {#pcbi-1002007-t001 .table-wrap} 10.1371/journal.pcbi.1002007.t001 Table 1 ::: {.caption} ###### Model Parameters. ::: {#pcbi-1002007-t001-1} Parameter Symbol Value Source ----------------------------- -------- ---------------------- ------------------------------------------------------------------------------------- Viscoelastic time scale τ 0.25 hr E.E.[\*](#nt101){ref-type="table-fn"} [@pcbi.1002007-Lambert1] Effective shear viscosity η 10 dynes×hr/cm E.S.[†](#nt102){ref-type="table-fn"} Volumetric viscosity λ 10^3^ dynes×hr/cm E.S.[†](#nt102){ref-type="table-fn"} Substrate drag coefficient ζ 10^7^ dynes×hr/cm^3^ E.E.[\*](#nt101){ref-type="table-fn"} [@pcbi.1002007-Boca1], [@pcbi.1002007-Roure1] Average cell crawling speed V~0~ 10 µm/hr [@pcbi.1002007-Boca1] Traction stress f~0~ 10^4^ dynes/cm^2^ [@pcbi.1002007-Roure1] Dipole length *b* 10 µm E.E.[\*](#nt101){ref-type="table-fn"} [@pcbi.1002007-Roure1] Rotational drag coefficient ~ζr~ 400 dynes×hr/cm^2^ E.E.[\*](#nt101){ref-type="table-fn"} [@pcbi.1002007-Kole1] Franck constant *K* 2×10^−4^ dynes E.E.[\*](#nt101){ref-type="table-fn"} [@pcbi.1002007-Petitjean1] \*E.E. (estimated from experiments. See text for more details.) **†:** E.S. (estimated from simulations): value determined by matching simulation results to existing experimental data. ::: We tracked the average advance of the boundary as a function of time and compared it to data that was published previously [@pcbi.1002007-Poujade1] ([**Figure 4**b](#pcbi-1002007-g004){ref-type="fig"}). We find that the average border progression scales supra-linearly with time, and our simulations match the experiments by Poujade, et al. when the initial width of our simulated region was greater than 200 µm. When the initial width was smaller than 200 µm, the rate of advance of the border decreases. As in our 1D simulations, we define the border progression exponent *E* and measure the dependence of *E* on the initial width of the monolayer. For initial widths between 100--300 µm, we see an increase in the border progression exponent, which increases from 1.2 to 1.6 ([**Figure 4**c](#pcbi-1002007-g004){ref-type="fig"}). This result is also consistent with what has been measured [@pcbi.1002007-Poujade1]. Cell morphology and motility are known to depend on the stiffness of the substrates [@pcbi.1002007-Discher1]. It is therefore interesting to ask how substrate stiffness would affect wound closure. It is likely that substrate stiffness affects the magnitude of the traction stress, the thrust force, and the resistive drag force. If all three of these parameters change in a similar fashion with respect to changes in substrate stiffness, then our model suggests that the results will be equivalent to changing the cell-cell viscosity. For example, if traction stress, thrust force, and resistive drag increase with increases in the substrate stiffness, then this would behave like decreasing the cell-cell viscosity in our model, and we would therefore only expect small changes in the overall efficiency of wound closure with changes in substrate stiffness. We also explored how wound healing would be affected by the magnitude of the dipole-distributed stress, *f* ~0~, leaving all other parameters fixed. For this case, we find that wound closure is more efficient (i.e., the border progression is faster on average with larger traction stress) and the border exponent also increases. We find that increasing the traction stress by a factor of three leads to an increase in the border exponent from 1.3 to 1.8. Note that it is possible to determine the affect of changing substrate stiffness on the parameters in our model by measuring the traction stress, average speed of isolated cells, and cytoskeletal flow rates as a function of substrate stiffness, which allows a method for testing these predictions. Discussion {#s3} ========== The cooperative cellular behavior that accompanies wound healing is astonishing. Even in a simple in vitro monolayer of cells, wounds are "repaired" as cellular movements fill in a denuded region. These movements have been attributed to processes such as the purse-string mechanism, where cells along the periphery of the wound concentrate actin and myosin at the wound edge. Contraction of the actin cortex by myosin can then drive wound closure, and, indeed, in chick embryos, this process is likely to play a significant role [@pcbi.1002007-Martin1]; however, healing of larger wounds is presumed to rely on cell crawling. It is also possible that wounding triggers a biochemical response that signals cells to move to fill in the wound. For example, Matsubayashi et al. observed waves of phosphorylation of Mitogen Activated Protein Kinase after wounding a cell monolayer [@pcbi.1002007-Matasubayashi1]. Release of reactive oxygen species were identified as potential upstream cues to these activation waves [@pcbi.1002007-Nikolic1]. The experiments by Poujade et al. [@pcbi.1002007-Poujade1], which do not damage the cells during wounding, suggest that release of signaling factors at the wound site are not required for wound healing. These results suggest that a mechanical mechanism may be driving wound healing. In this paper, we have shown that the same mechanical process that drives single cell crawling, augmented by cell-cell adhesion, is sufficient to drive wound healing. The principal driving force in this model comes from the polarization of crawling cells; i.e., single crawling cells exert a dipole-distributed force distribution on the substrate. At the edge of the wound, this force distribution acts like a pressure that pulls the cells out into the denuded region. Within the cell-filled region, the force distribution causes instabilities that lead to the observed complex flow fields of the cells, which include vortices and jets and also a fingering-like appearance of the moving boundary. Cell-cell adhesion keeps the monolayer cohesive, which produces long-range correlations in the cellular velocity field and also causes the cell monolayer to act like a viscoelastic fluid that is fairly rigid on short timescales but flows on longer timescale. This viscoelastic behavior, consequentially, produces the accelerative advance of the cells out into the wounded region. Therefore, this conceptually simple model with reasonable choices for the parameters captures quantitatively most of the observed features of wound healing. The parameters that are used in our model are physical. Therefore, they are all, in principle, measureable. Indeed, many of the parameters have already been measured or there exist experiments that can be used to estimate the parameters. In fact, only two parameters are largely undetermined, the two effective viscosities due to cell-cell adhesions. However, the cell-cell shear viscosity in tissue has been estimated previously to be on order of 10^5^ Poise in embryonic tissue [@pcbi.1002007-Forgacs1]. This viscosity is the three-dimensional viscosity; however, our model uses a two-dimensional that contains a factor of the thickness of the monolayer. If we assume that this thickness is around 1 µm, then our estimates based on our simulations suggest a shear viscosity of around 10^8^ Poise, which is significantly larger than the previous finding. Our results suggest, though, that the overall dynamics of wound healing are only weakly dependent on the viscosities, and, therefore, our predictions of the viscosities are probably only good to an order of magnitude. In recent years there have been a few other models that have been proposed to describe the mechanical process of wound healing. Two separate groups have examined wound healing driven by cell proliferation and diffusive motion [@pcbi.1002007-Cai1]--[@pcbi.1002007-Bindschadler1]. These models predict linear dependence of the border progression on time and do not capture the complex cellular motions that are observed in MDCK wound healing assays. However, they are able to fit well experiments using fibroblasts, which do not form adherens junctions [@pcbi.1002007-Bindschadler1]. Cellular sheets lacking adherens junctions should have a reduced viscoelastic timescale, and therefore our model would also predict a more linear advance of the border. A recent paper by Mark et al. [@pcbi.1002007-Mark1] showed that treating the cells near the wound boundary as an active membrane bounding an elastic monolayer of cells can explain the fingering instability of the wound boundary. In this model, the active membrane is treated as an elastic contour that satisfies the Helfrich energy functional with surface-tension; i.e., there are energetic costs for bending and for increases in contour length. Cell migration defines an outward force at the boundary which drives the cells out into the void. Finally, there is a restoring force from the cells that are away from the boundary. It is interesting to note that the dipole-distributed traction stress in the model presented here naturally accounts for the active driving force of the Mark et al. model. It should be noted, however, that the fingering instability that we observe is not as pronounced as what is observed experimentally. In these experiments, a leader cell is often observed at the tip of the finger. These leader cells arise directly from the general population of cells and have a more-spread appearance with an active lamellipodium. Once the finger reaches the distal side of the wound, the leader cells revert back to a typical epithelial morphology. We suggest that these leader cells may be a result of reduced cadherin binding to adjacent cells; i.e., that given sufficient space to spread, that the epithelial cells will naturally spread and take on this new appearance. It is likely that a more spread cell will exert a different traction stress and thrust force than the standard cells in the population. Therefore, our model may not completely reproduce the boundary fingering because we assume that all cells in the population are equivalent. Though the model developed here shows that it is possible that wound healing can be driven by a purely mechanical mechanism, it does not imply that cell signaling does not play an important role in this process, too. Indeed, it is definitely true that biochemical regulation is required for controlling the mechanical processes that underlie our model, specifically the actin dynamics that produce the dipole stress and thrust force, and therefore it is likely that inter-cellular signaling may modify and enhance force production in the cells that are closest to the boundary. Methods {#s4} ======= Numerical methods {#s4a} ----------------- The two dimensional free boundary problem was solved using the Moving Boundary Node Method [@pcbi.1002007-Wolgemuth2]. This method is a level set-based, finite volume algorithm. For these simulations, we used a time step of 0.001 hr and a grid spacing that was 1/40 of the initial width of the domain (i.e., for a monolayer with an initial width of 300 µm, the grid spacing was 7.5 µm). Supporting Information {#s5} ====================== Text S1 ::: {.caption} ###### Supplemental text that provides more complete details on the mathematical and computational aspects of the wound healing model presented in this paper. (DOC) ::: ::: {.caption} ###### Click here for additional data file. ::: The authors have declared that no competing interests exist. This research was supported by the NSF (DMS 0920279). The funders had no role in study design, data collection and analysis, decision to publish, or prepartation of the manuscript. [^1]: Conceived and designed the experiments: CWW PL. Performed the experiments: CWW PL. Analyzed the data: CWW PL. Contributed reagents/materials/analysis tools: CWW PL. Wrote the paper: CWW PL.

PubMed Central

2024-06-05T04:04:19.649925

2011-3-10

PMC3053313

Introduction {#s1} ============ Our motor commands generally produce two kinds of consequences: a sensory consequence in terms of activity in our primary sensory organs (e.g., vision, proprioception), and a rewarding consequence in terms of forming a subjective measure of utility or usefulness of these sensations (e.g., release of dopamine). For example, while dancing, the motor commands that move our body produce proprioceptive feedback, while internal evaluation of that feedback indicates a pleasurable experience. These two consequences of the motor command form the basis for two kinds of prediction error: a sensory prediction error, and a reward prediction error. In principle, learning from sensory prediction error should alter an internal model that predicts the sensory consequences of motor commands, i.e., a forward model [@pcbi.1002012-Synofzik1], [@pcbi.1002012-Synofzik2]. In contrast, learning from reward prediction error should alter the valuation of the sensory states that are the consequence of those motor commands, i.e., a value function. Motor adaptation studies often focus on learning from sensory prediction error [@pcbi.1002012-Synofzik1], [@pcbi.1002012-Synofzik2], [@pcbi.1002012-Baddeley1], [@pcbi.1002012-Berniker1], [@pcbi.1002012-Kording1], [@pcbi.1002012-Sing1], [@pcbi.1002012-vanBeers1], despite the fact that people are also rewarded for each movement. Similarly, studies that focus on learning from reward prediction error (e.g., decision making tasks) often do not consider potential sensory prediction errors [@pcbi.1002012-Frank1], [@pcbi.1002012-Schonberg1], [@pcbi.1002012-Trommershauser1]. It seems rational that most learning would rely on both kinds of error. Here, we focus on a simple motor adaptation task and consider a mathematical framework in which both reward and sensory prediction errors could contribute to the trial-to-trial change in the motor commands. We attempt to ask whether learning from these two distinct signals can be behaviorally dissociated. Our idea is that while motor commands might change because of sensory or reward prediction errors, only in the former case would there also be a change in the map that predicts the sensory consequences of the motor command. We focus on a well studied motor adaptation protocol: reaching in the context of visuomotor perturbations. While there have been numerous models of motor adaptation [@pcbi.1002012-Berniker1], [@pcbi.1002012-Kording1], [@pcbi.1002012-Kawato1], [@pcbi.1002012-Kawato2], [@pcbi.1002012-Thoroughman1], [@pcbi.1002012-Smith1], to our knowledge all current models assume that the process of motor adaptation is driven by sensory prediction errors. Our objective is to test the hypothesis that during motor adaptation, learning from sensory prediction errors leaves a behavioral signature that is distinct from learning from reward prediction errors. Results {#s2} ======= Consider a typical adaptation task in which the learner experiences a perturbation. The limb is covered by a screen to prevent direct observation of the hand, and a cursor that represents hand position undergoes a kinematic rotation so that when the hand moves straight ahead, the cursor moves slightly to the left ([Fig. 1A](#pcbi-1002012-g001){ref-type="fig"}). Reward is provided if the cursor passes through the target area. In this reach adaptation task there are two kinds of error: the difference between the expected and observed visual feedback of the hand (i.e. visual cursor), and the difference between the expected and observed success of the reach. Our hypothesis is that learning mechanisms engaged by the two types of error may be behaviorally dissociable. ::: {#pcbi-1002012-g001 .fig} 10.1371/journal.pcbi.1002012.g001 Figure 1 ::: {.caption} ###### Experimental setup. (**A**) In the reaching task, subjects held a handle of a robotic arm and made 'shooting' movements to move a cursor through a target at 10 cm. The arm was covered by a screen. During adaptation, the cursor-hand relationship was perturbed so that the cursor position was rotated around the center at the start position. The coordinate system is drawn on the left side of the robot (invisible for subject) where the clockwise rotation around the start is positive. The cumulative score of each block was provided to the subject. In the localization task, subjects pointed with their left hand over the screen to the remembered location of their right hand as it crossed the (unseen) target area in the previous trial. In the localization task, the start box was not visible. (**B**) Experimental paradigms. In ERR, full visual feedback about the cursor position was provided as well as the animation and the sound indicating target explosion regarding success or failure of the task. In EPE, while the cursor was unseen during the shooting movement, it was presented for 200 ms as the hand crossed an imaginary circle with the radius equal to the target, providing endpoint error with respect to the target. The reward signal was also provided as in the ERR condition. In RWD, no visual feedback about the cursor was provided. All information that subjects were able to use was the success or failure of the task. (**C**) Reach angles of three representative subjects during the adaptation phase. The yellow line in the ERR group is the ideal reach angle, which shifted gradually up to 8 degrees by the visual rotation. The gray area indicates the reward region, which shifted with the same schedule in the three groups. (**D**) Reach variability in the final 100 trials for each group. There are the significant differences between ERR and EPE (t-test, p\<0.003) as well as between EPE and RWD (t-test, p\<0.001). (**E**) Results of the localization task for the three subjects. The reach trajectory is plotted for the POST condition. Red line is for the RWD subject, blue line is for the ERR subject, and green line is for the EPE subject. The circle around the reach trajectory is the averaged pointing location in the localization trial. :::  ::: To examine this hypothesis, we recruited two groups of subjects in Experiment 1. One group (RWD) was provided only with information regarding whether they succeeded or failed at each trial (reward r = 1 or 0), indicated by explosion of the target, and received no other visual feedback regarding their movement ([Fig. 1B](#pcbi-1002012-g001){ref-type="fig"}). Another group was provided with full visual feedback of the cursor as well as the reward so that they were able to use both potential error signals (ERR). We asked two questions: 1) In the ERR paradigm in which sensory consequences of motor commands were available, would adaptation of the motor commands accompany a change in the motor-sensory map (i.e., a change in the perceived sensory consequences of motor commands), and 2) in the RWD paradigm in which sensory consequences of motor commands were unavailable, would adaptation of the motor commands take place but without a change in the motor-sensory map. [Fig. 1C](#pcbi-1002012-g001){ref-type="fig"} shows data from representative subjects in the ERR and RWD paradigms. In this figure, the yellow line in the ERR group is the ideal reach angle (shifts gradually up to 8°). The gray area indicates the region that provided reward, which shifts with the same schedule in both groups. The subjects were provided with different kinds of error feedback, but updated their motor commands by roughly the same amount (group data, mean change in reach direction, 7.49° for RWD and 7.63° for ERR, not significantly different from each other p\>0.8, t-test). The total amount of adaptation of the two groups was comparable. However, the variability of reach angles was larger for the RWD subject ([Fig. 1C](#pcbi-1002012-g001){ref-type="fig"}), and this was consistent across the entire group ([Fig. 1D](#pcbi-1002012-g001){ref-type="fig"}). Before and after this adaptation task (PRE and POST adaptation), we measured how subjects predicted the sensory consequences of their motor commands. In this localization part of the task, after subjects completed a reach with their right hand, their hand was returned to the center location, and they were then asked to estimate the location of their right hand in the previous trial by pointing with their left hand over the screen ([Fig. 1A](#pcbi-1002012-g001){ref-type="fig"}). During the localization neither the cursor nor the target was projected. The localization data for representative subjects are shown in [Fig. 1E](#pcbi-1002012-g001){ref-type="fig"}. As a consequence of adaptation, the subject in the ERR group had a sensory remapping in which she estimated her hand to be to the left of its actual position. In contrast, the subject in the RWD group had little or no sensory remapping, suggesting that the changes in the motor commands did not accompany a change in the motor-sensory map. [Fig. 2A](#pcbi-1002012-g002){ref-type="fig"} shows the group data for the localization task. We compared the change in the estimate of hand position from the PRE to the POST adaptation condition and found that the subjects in the ERR group estimated their hand position to have changed by 8.8°+/−0.6° to the left of actual position. In contrast, in the POST condition of the RWD group, the subjects had no significant change in their sensory estimates (there was a significant difference between PRE and POST in the ERR group p\<0.0001, whereas the difference in the RWD group was not significant p = 0.8). ::: {#pcbi-1002012-g002 .fig} 10.1371/journal.pcbi.1002012.g002 Figure 2 ::: {.caption} ###### The sensory remapping and the generalization function. (**A**) The average estimated localization of hand position in PRE and POST conditions. Error bars are SEM. (**B**) Generalization of adaptation from the learned target direction (at 0°) to neighboring target directions. (**C**) Illusion index (change in estimated location of the hand from PRE to POST adaptation), as a function of generalization index in subjects in EPE condition. Each dot indicates individual subject\'s data. There are significant negative correlation in these two indices (R = −0.68, p = 0.02). :::  ::: If the sensory and reward prediction errors engage learning in distinct neural structures, then adaptation might result in distinct generalization patterns [@pcbi.1002012-Pearson1], [@pcbi.1002012-Shadmehr1], [@pcbi.1002012-Haswell1], [@pcbi.1002012-Bedford1]. To test this idea, we recruited subjects for Experiment 2 and quantified the patterns of generalization that accompanied adaptation. In the adaptation session, the target was projected at 0° (straight ahead). In the pre and post adaptation periods the target appeared randomly at various angular displacements (−30 to 30 deg). For these generalization targets, we provided neither the cursor nor reward information. [Fig. 2B](#pcbi-1002012-g002){ref-type="fig"} plots the average reach angle across subjects for each target direction. We found that the RWD group had a narrower generalization function than the ERR group (ANOVA, F(1,126) = 9.632, p = 0.005). In summary, in the RWD condition the learning that produced changes in the motor commands accompanied a narrow generalization function and no change in the map that predicted the sensory consequences of motor commands. In contrast, in the ERR paradigm the learning that produced changes in the motor commands accompanied a broad generalization function and a significant change in the perceived sensory consequences of motor commands. In the RWD paradigm the binary feedback signal carried much less information than the continuous sensory error signal available in the ERR paradigm. This may have forced the subjects to adopt a completely new strategy, making the learning that we see in the RWD paradigm irrelevant for a typical adaptation paradigm. In Experiment 3 we considered a paradigm (EPE) in which the visual cursor was available only at the endpoint of the movement and was otherwise invisible during the reach. In this new experiment we measured the localization change (as in Exp. 1) and the generalization (as in Exp. 2), attempting to test the results of experiments 1 and 2 in the same population. [Fig. 1C](#pcbi-1002012-g001){ref-type="fig"} shows the reach angles of a representative subject in the EPE group. The adaptation in the EPE group was comparable with the ERR group (mean change in reach direction, t-test, p = 0.64), i.e., the motor commands in the three groups adapted by approximately the same amount. Interestingly, in the localization task the subject in the EPE group had a sensory illusion that was in between the ERR and RWD groups ([Fig. 1E](#pcbi-1002012-g001){ref-type="fig"}). In the group data in the POST adaptation condition, the strength of the localization illusion in the EPE group was weaker than in the ERR group (t-test, p\<0.007), but stronger than the RWD group (t-test, p\<0.006) ([Fig. 2A](#pcbi-1002012-g002){ref-type="fig"}). The generalization of the EPE group appeared to be in between ERR and RWD (we did not see a significant difference from either ERR or RWD, [Fig. 2B](#pcbi-1002012-g002){ref-type="fig"}). In Experiment 2 we had found that learning from reward produced a narrow generalization, while in Experiment 1 we had found that learning from error produced a motor-sensory remapping. In Experiment 3 we had the means to test a crucial prediction: across subjects, individuals who relied more on reward (narrow generalization) should show a smaller motor-sensory remapping. Indeed, we found a significant correlation between the amount of generalization and the localization illusion across subjects ([Fig. 2C](#pcbi-1002012-g002){ref-type="fig"}). That is, it appeared that when a subject had a larger sensory illusion (suggesting that learning was driven more by sensory prediction errors), they also had a wider generalization. To explore the mechanism behind these findings, we considered a model of adaptation that relied on both sensory and reward prediction errors ([Fig. 3A](#pcbi-1002012-g003){ref-type="fig"}). Suppose that the brain generates a motor command *u*, resulting in a change in the state of the hand *h*, which also depends on a perturbation *p*. The nervous system senses the resulting motion of the limb *y* as well as whether that motion was rewarded *r*. Here, we considered a learner who updates motor command *u* to maximize reward. In theory, producing the motor commands that maximize probability of reward may rely on two kinds of learning: forming an optimal action selector, and forming an optimal state predictor ([Fig. 3B](#pcbi-1002012-g003){ref-type="fig"}). On trial *k*, action selector outputs motor commands . This depends on the estimated perturbation (which depends on sensory prediction error ), as well as the reward prediction error . Therefore, in theory the trial-to-trial change in the motor commands is driven by two different error signals: the state estimator updated by the sensory prediction error, and the action selector updated by the reward prediction error. ::: {#pcbi-1002012-g003 .fig} 10.1371/journal.pcbi.1002012.g003 Figure 3 ::: {.caption} ###### The theoretical problem of learning motor control. (**A**) A generative model of the motor adaptation task. Motor commands are corrupted by a perturbation, which result in a hand position that is sensed via a cursor, and may also result in reward. The objective of the learner is to find the motor commands that maximize reward. White circles are hidden variables and gray circles are observed variables. Arrows indicate conditional probabilities. (**B**) Model of optimal learner. The learning system is composed of two compensatory mechanisms: action selector and internal forward model. At the trial k, the action selector outputs the motor command to make a transition of the state of the body and task from to . The state variable includes three elements: hand position *h*, perturbation *p*, and the position *t*. The brain observes the part of the state of the body . At the same time, the learner predicts the transition of the body state from the efference copy of the motor command. Kalman filtering correct the prediction to minimize the sensory prediction error to have the updated state . The action selector selects the optimal action as a function of the updated state at the next trial. (**C**) Sample disturbance and the response of the model. The task is to control the reach angle. Clockwise (CW) direction is positive and the target is at 0°. The uncertainty of the visual feedback was controlled to modulates the Kalman gain. The simulations predict a remapping regarding estimated hand position modulated by the level of visual uncertainty. :::  ::: An important prediction from this model is that reliance on the sensory prediction error is modulated by the Kalman gain, which is the ratio of estimation uncertainty to observation uncertainty. Therefore, if the uncertainty of visual feedback is large, the credit on the sensory prediction error becomes small, which makes the credit on the reward prediction error larger. [Fig. 3C](#pcbi-1002012-g003){ref-type="fig"} shows results of simulations for different uncertainty levels of visual feedback. When the learner is provided with high quality visual feedback (analogous to ERR condition, [Fig. 3C](#pcbi-1002012-g003){ref-type="fig"} left column), it updates its estimate of perturbation , resulting in a motor-sensory remapping. As a result, the estimated hand position is near the location of the cursor and different from actual hand position . In contrast, when the learner is provided with uncertain visual feedback (analogous to EPE condition, middle column in [Fig. 3C](#pcbi-1002012-g003){ref-type="fig"}), the learner alters the motor commands using both the sensory prediction error and the reward prediction error. In this case, the adaptation produces a partial sensory remapping ( is not very different from in the middle column of [Fig. 3C](#pcbi-1002012-g003){ref-type="fig"}). Finally, when the learner is provided with extremely poor visual feedback (analogous to RWD condition, right column of [Fig. 3C](#pcbi-1002012-g003){ref-type="fig"}), all that is available to the learner is success or failure (  = 0 or 1). The learner still alters the motor commands to compensate for the perturbation, but the adaptation does not produce a sensory remapping ( is not different from in the right column of [Fig. 3C](#pcbi-1002012-g003){ref-type="fig"}). These three different patterns of sensory remapping generated by the model help explain the reason why we observed different patterns of sensory remapping in the three different paradigms. In the ERR condition in which high quality sensory feedback was available, adaptation produced large change in the state predictor, producing the sensory remapping. In RWD condition in which the visual feedback of the cursor was not available, adaptation focused on the action selector, which was updated by reward prediction error. Because this process did not involve a sensory remapping, we did not observe a change in the localization behavior of the subjects. In the EPE condition in which partial visual feedback was provided, learning depended on both an updating of the state predictor and the action selector. As a result, we observed the partial sensory remapping. To validate our model, we used it to estimate how much of the change in the motor commands that we observed in our subjects was due to each type of error. We imagined that the motor commands were generated by the sum of two states with a search noise, , where represents the estimate of the perturbation as updated by sensory prediction error and the is updated by reward prediction error. Using a nonlinear optimization algorithm, we fit the model to the trial-to-trial behavior of each subject (reach direction on each trial), and the state of reward on that trial. In the RWD paradigm, the only feedback available was reward prediction error, i.e., . The results of our model fit are shown in [Fig. 4](#pcbi-1002012-g004){ref-type="fig"} via the average of estimated parameters and , and their sum. These estimated values were superimposed on the average of subjects\' trial-to-trial reach angle (black line) with SEM across subjects. In the ERR condition, by the end of adaptation the contributions of these two states were \[, \] = \[7.82,0.26\]+/−\[0.18,0.31\]. Despite the fact that we used the exact the same model to fit the data for ERR and EPE, the best fit estimates of these two states in EPE were \[, \] = \[4.53,3.33\]+/−\[0.59,0.69\], which were significantly different from those of ERR (ANOVA, F(1,18) = 18.93,p\<0.001). ::: {#pcbi-1002012-g004 .fig} 10.1371/journal.pcbi.1002012.g004 Figure 4 ::: {.caption} ###### Estimated contribution of reward and sensory prediction errors to change in motor output during adaptation. When subjects experienced the ERR and EPE condition, we assumed that the motor commands were produced by the sum of two memories, , where was updated by the sensory-prediction error and was updated by the reward prediction error. The best fit parameters predict the update of the two memories. The black think line is the averaged subject\'s reach angle during the adaptation period. The gray shadow is SEM. The superimposed purple line is the estimated reach angle from the model which is a combination of (red) and (blue). In the RWD condition, the motor commands are updated by only the reward-prediction error: . :::  ::: By fitting the model to the data, we were able to estimate the search noise . We found that the variance of the search noise in ERR was , which was significantly smaller than that of EPE (, p\<0.001), and RWD (, t-test, p\<0.01). Our estimate of a significantly smaller search noise in the ERR condition is consistent with our inference that with high quality sensory feedback, the change in the motor commands is driven almost entirely by sensory prediction errors. This is also consistent with the fact that in the ERR condition, there was a scarcity of reward prediction error: In ERR, more than 95% of trials were rewarded, whereas the probability of reward in EPE was 83% and that in RWD was 76%. Therefore, our analysis suggests that in the ERR paradigm the change in the motor commands was due primarily to adaptation of the state estimator (accounting for the sensory remapping), whereas in the RWD paradigm the change was due to adaptation of the action selector (accounting for the lack of sensory remapping). In the EPE paradigm the change was due to both the state estimator and the action selector. Discussion {#s3} ========== Our goal was to determine whether during motor adaptation one could dissociate between learning from reward prediction errors vs. learning from sensory prediction errors. We considered a reaching task in which visual feedback regarding cursor position was altered. The quality of this feedback was manipulated so that in one group the sensory feedback was of high quality (available throughout the reach, ERR group), in another group the sensory feedback was of low quality (available only at the end of the reach, EPE group), and in a third group the sensory feedback was unavailable (RWD group). All groups had access to reward (success or failure) at the end of their movement. We found that after a long period of training, all three groups adapted their motor commands. In the ERR group this adaptation accompanied a wide pattern of generalization and a significant change in the perceived sensory consequences of motor commands. In contrast, in the RWD group the adaptation accompanied a narrow pattern of generalization and no change in the perceived sensory consequences of motor commands. In the EPE group, generalization and sensory remapping were intermediate. Interestingly, in the EPE group individuals who demonstrated a larger sensory remap also had a wider generalization function. Increasing the uncertainty in the sensory prediction error altered both the width of generalization function and the amount of sensory remapping, while it did not affect the level of adaptation. While previous models of motor adaptation have relied exclusively on sensory prediction errors to form an estimate of the perturbation [@pcbi.1002012-Berniker1], [@pcbi.1002012-Kording1], [@pcbi.1002012-Wolpert1], [@pcbi.1002012-Barnes1], the comparable levels of motor adaptation in our groups (ERR, RWD, and EPE) suggest that the brain relied on another source of error, the reward prediction error, when the sensory prediction error was not informative. In fact, it has been shown that the reward may modulate motor planning [@pcbi.1002012-Trommershauser2], [@pcbi.1002012-Takikawa1]. Thus, it seems more rational that the purpose of learning is not merely to estimate the magnitude of a perturbation, but to produce motor commands that maximize reward [@pcbi.1002012-Doya1]. We formulated this adaptation as a reward maximization process by assuming an "optimal learner". The optimization relied on two update equations: one was the optimal estimator that inferred the state of the body, and the other was the optimal policy that selected the action as a function of the estimated state [@pcbi.1002012-Izawa1], [@pcbi.1002012-Izawa2], [@pcbi.1002012-Todorov1]. Based on this theory, our model of the optimal learner was composed of two components: reinforcement learning for action selection, and state estimation for identifying the sensory consequences of motor commands [@pcbi.1002012-Shadmehr2]. In this model, the objective of state estimation was to estimate the perturbation in the environment and the hand position as a consequence of the motor command, while the objective of the reinforcement learning was to update how to select the action to maximize reward probability [@pcbi.1002012-Izawa3]. The simulation showed that the learner relied mostly on the sensory prediction error in ERR paradigm. As a result, the learner updated the parameter associated with the sensory consequence of the motor command, which predicted the illusion that we observed in Experiment 1. In contrast with the ERR paradigm, the RWD paradigm did not provide the sensory prediction error. Thus, the simulation with the RWD paradigm showed that the reward-prediction error updated the action but did not change the estimate of hand position. Thus, high quality sensory feedback produced learning that depended primarily on sensory prediction errors. While our model was not designed to account for the distinct generalization patterns in the ERR and the RWD paradigms, previous studies have speculated that generalization patterns are a reflection of the neural encoding of information during learning [@pcbi.1002012-Shadmehr1], [@pcbi.1002012-Poggio1]. For example, generalization patterns during reach adaptation in force fields appear consistent with an encoding in which the neurons have activity fields that resemble those in the primary motor cortex [@pcbi.1002012-Thoroughman1], [@pcbi.1002012-Hwang1]. In contrast, generalization patterns in visuomotor rotations appear more consistent with an encoding similar to cells in the posterior parietal cortex [@pcbi.1002012-Tanaka1]. In this framework, the two different generalization patterns seen in RWD and ERR paradigms suggest engagement of two different neural mechanisms that each learn from reward and sensory prediction error. Another possibility, however, is that the two forms of prediction error converge on a single neural structure that guides motor learning. By presenting the optimal learner model that includes two forms of prediction error, we built a connection between two disparate areas of research that has focused on different parts of the brain. Motor adaptation has focused on tasks that typically depend on the integrity of the cerebellum [@pcbi.1002012-Tseng1], [@pcbi.1002012-Smith2]. Habit learning [@pcbi.1002012-Yin1], visuomotor sequence learning [@pcbi.1002012-Nakahara1], or action selection [@pcbi.1002012-Tanaka2], [@pcbi.1002012-Samejima1] have focused on tasks that depend on the integrity of the basal ganglia [@pcbi.1002012-Packard1], [@pcbi.1002012-Wickens1]. In fact, goal directed action in habitual learning is mediated by two representations: a representation of the instrumental contingency between the action and the outcome, and a representation of the outcome as a goal for the agent [@pcbi.1002012-Dickinson1]. Because motor adaptation is also a goal directed action, the two learning mechanisms observed in this paper might be the general systems involved in a broad category of procedural learning. For example, these two distinct memories might be mediated by parallel cortico-basal ganglia mechanisms with different sensory domains [@pcbi.1002012-Nakahara1]. Patients with basal ganglia disorders show little or no deficits in motor adaptation paradigms like force fields [@pcbi.1002012-Smith2] or visuomotor perturbations [@pcbi.1002012-Gabrieli1], [@pcbi.1002012-Agostino1] (although patients with PD appear to show a deficit in consolidation of the memory [@pcbi.1002012-Marinelli1]). Why is this? Our theory provides a potential answer: in the typical force field or visuomotor tasks, high quality sensory feedback is available, making it likely that sensory prediction errors play a dominant role. Because learning from sensory prediction errors likely depends on the integrity of the cerebellum [@pcbi.1002012-Synofzik2], [@pcbi.1002012-Tseng1], [@pcbi.1002012-CriscimagnaHemminger1], the implication is that the ability of basal ganglia patients to adapt to visuomotor and force field perturbations is not evidence for normal motor adaptation, but rather evidence for the idea that changes in motor output in these tasks are primarily driven by sensory prediction errors. The other implication of the theory is that the inability to adapt the sensory consequences of motor commands did not prevent adaptation of the motor commands in response to reward prediction errors. Indeed, when we altered the adaptation paradigm and made it so that changes in the motor output were driven by reward prediction errors, we found that in response to the reward prediction error subjects altered their motor commands. This theory predicts that by providing rewards appropriately during a motor adaptation task, the cerebellar patients may be able to update their motor commands without sensory recalibration. Another implication of the theory is that the active search noise to explore the motor commands plays an important role in updating the action selector. Indeed, we found that trial-to-trial variability was modulated depending on types of error with significantly larger variability in the RWD than in ERR and in EPE. In previous studies, movement variability is generally thought to be due to signal dependent noise in the neuronal structures that generate motor commands [@pcbi.1002012-Harris1], [@pcbi.1002012-Jones1], [@pcbi.1002012-Faisal1]. However, noise is present even in the planning stage of movements [@pcbi.1002012-Churchland1]. Here, we found that during adaptation variability in movements was not due to meaningless noise, but an inherent part of a search that the brain engaged in to find motor commands that acquired a more rewarding state. In summary, changes that take place in motor commands during adaptation are likely to be driven by both sensory and reward prediction errors. Learning from sensory prediction error alters the predicted sensory consequences of motor commands, leaving behind a sensory remapping. During motor adaptation, the reliance on reward prediction errors can be increased by degrading the quality of the sensory feedback. Learning from reward prediction error does not accompany a sensory remapping. It is likely that the neural basis of learning from sensory and reward prediction errors are distinct because they produce different generalization patterns. Methods {#s4} ======= Experimental Procedures {#s4a} ----------------------- Subjects sat in front of a robotic arm and held its handle [@pcbi.1002012-Izawa2]. A video projector painted the screen that covered the manipulandum and the subject\'s arm. A trial began by the robot positioning the subject\'s hand in a start box, at which point a target of 6° width appeared at 10 cm. Subjects were instructed to perform a 'shooting' motion so that their hand crossed within the target area, at which point the target was animated to show an explosion, and a score was increased by one point. In the error-based learning (ERR) paradigm, the cursor position was displayed during the movement toward the target. In the reward-based learning (RWD) paradigm, the cursor position was not displayed. For both groups, target explosion indicated success of the trial. The cursor was not displayed during the return of the hand to the start position. Ethics Statement {#s4b} ---------------- Protocols were approved by the local IRB and all subjects signed a consent form. ### Experiment 1: Learning from sensory prediction errors {#s4b1} Volunteers (n = 14, 26±4.7 years old) were assigned to the ERR (n = 7) or the RWD group (n = 7). After a familiarization session, the experiment was composed of a visuomotor adaptation phase and two localization phases (PRE and POST). In the localization phase ([Fig. 1A](#pcbi-1002012-g001){ref-type="fig"}), the subjects performed four shooting trials followed by one localization trial. For the first 4 trials, the cursor was visible for the ERR learning group but invisible for the RWD group. For the 5^th^ trial, the cursor was invisible for both groups. In the localization trial, neither the cursor nor the target was projected. In this trial, subjects pointed with their left hand (over the screen) to the estimated position of their right hand as it crossed the target area in the previous trial. That is, the subjects were asked to estimate the location of their right hand in the previous trial. These five trials (four shooting and one localization) were repeated 10 times for the PRE phase, and 10 times for the POST phase. The PRE localization phase was followed by an adaptation phase in which subjects experienced zero-rotation with 40 trials and then the perturbation increased by 1° every 40 trials until it reached 8° ([Fig. 1C](#pcbi-1002012-g001){ref-type="fig"}). The 8° perturbation lasted 80 trials. After a short break, subjects experienced 96 additional trials with the 8° perturbation and then were tested in the POST localization task. ### Experiment 2: Generalization {#s4b2} The idea behind this experiment was to test whether adaptation in response to sensory prediction errors (ERR paradigm) vs. reward prediction errors (RWD paradigm) differed in their generalization patterns. Volunteers (n = 27, 24±4.4 years old) were assigned to the RWD (n = 18) or ERR groups (n = 9). Both groups were provided with a familiarization session. Subsequently the subjects experienced two baseline blocks composed of 80 trials. In the baseline block, the target position was selected randomly from \[−30°, −20°, −10°, 0°, 10°, 20°, 30°\] with respect to the trained target. The frequency of the center target (0°) was 32/80 trials and that of each peripheral target was 8/80. The objective of the peripheral targets was to test generalization. During these trials the cursor was not displayed and the target did not explode. For the center target, an explosion was provided for both groups but the cursor was displayed for only the ERR group. The baseline phase was followed by an adaptation phase. In the adaptation phase, the target appeared at only the center direction (0°) and the subjects experienced zero-rotation with 40 trials and then the perturbation shifted every 40 trials by −1° until it reached −8° and was held at this level for 80 trials. After a short break, subjects experienced another 3 blocks of 48 trials with −8° perturbation. Finally, we tested the generalization of this adaptation via a protocol that was the same as pre-adaptation. ### Experiment 3: End point error paradigm {#s4b3} Volunteers (n = 11, 26.1±5.2 years old) were recruited for the end-point error group (EPE). The target was located at one of seven position \[−30 −20 −10 0 +10 +20 +30\] degree with respect to the center line, along a boundary circle with a 10 cm radius. At the moment the cursor passed through the boundary circle, the boundary pass point (endpoint) was marked by the cursor for 200 ms. After a familiarization block, the subjects experienced the baseline block for the generalization task which is the same as Experiment 2, followed by the PRE localization block which is the same as the Experiment 1. Then, subjects experienced the adaptation blocks which is the same as Experiment 1, where the perturbation was gradually increased up to −8 degree which was followed by another 2 blocks of 48 trials with −8° perturbation. Next, we tested the generalization of this adaptation via a protocol that was the same as pre-adaptation. Finally, the subjects experienced the POST phase of the localization task. ### Data analysis {#s4b4} The endpoint of the movement was defined as the intersection between the hand path and a 10 cm radius circle centered at the start position. The reach angle was calculated as the angle between the center of the target and the line that connects the start location and the reach endpoint. With respect to the midline, a clockwise rotation was defined as positive. ### An optimal learner {#s4b5} Let us cast the problem of adaptation in a framework in which the brain predicts the sensory and reward consequences of motor commands, and then learns from prediction errors in both modalities. Hand position depends on the motor command (initial reach direction) and is influenced by noise : The units of all variables in Eq. (1) are degrees. The hand position controls the cursor position , in which the perturbation is imposed during the trial: On trial *k*, subjects observe their hand position via a visual cursor at but cannot observe the perturbation directly: where represents perceptual noise. Because subjects observe the hand position indirectly, we suppose that they predict hand position using the efference copy of the motor command where is the estimate of the perturbation. As subjects are repeatedly exposed to the perturbation, they build a prior knowledge of the characteristics of the perturbation: perturbations are correlated from trial to trial, and are also affected by noise [@pcbi.1002012-Berniker1], [@pcbi.1002012-Burge1], [@pcbi.1002012-Kording2]: Set the extended state of the system as . We then have a state update equation that relates motor commands with changes in state: where , , , , and the observation equation is: where . In summary, Eqs. (6) and (7) represent the relationship between motor commands and their sensory consequences. We assume that the objective for the learner is to maximize the rewards and minimize the cost. Under this assumption, for a linear dynamical system, optimal feedback control theory suggests that two interacting mechanisms are necessary: the optimal estimator and the optimal policy [@pcbi.1002012-Izawa1], [@pcbi.1002012-Todorov1]. The optimal estimator is composed of a forward model and a Kalman filter: where is the Kalman gain and is the sensory prediction error. The Kalman gain is a function of the uncertainty of the estimated state and the measurement noise such that where is the uncertainty of the state estimation and is the variance of the observation noise. The optimal policy outputs motor commands as a function of the estimated state. In optimal control theory, the policy is computed from the end of the learning period backward [@pcbi.1002012-Izawa1]. However, in a learning problem, the learner updates the policy on every trial and the backward computation is not plausible. Thus, we used Actor-Critic architecture that enables it to find the optimal policy without backward computation [@pcbi.1002012-Sutton1]. Here we represent this policy with where represents the active search noise to explore the motor commands and represents changes to the motor commands to maximize reward. Suppose that the expected cost-to-go function is of the form for a general reward function and discount rate of reward . We used a standard temporal learning algorithm to solve this optimization problem [@pcbi.1002012-Izawa3], [@pcbi.1002012-Sutton1]. In this algorithm, the policy is updated to minimize the reward prediction error: where . We used Temporal Difference (TD) error learning algorithm to updates the policy and the value. For our simulations ([Fig. 3C](#pcbi-1002012-g003){ref-type="fig"}), we used the following definition of the reward function: where is the scaling parameter of the motor cost. In summary, the learner has two kinds of prediction errors: a sensory prediction error (Eq. 8), and a reward prediction error (Eq. 13). The sensory prediction error updates an estimate of state produced by the motor commands (the sensory consequences of the action). The reward prediction error updates an estimate of the value of the states, and the policy that describes the 'best' motor commands to maximize reward. ### Fitting the model to data {#s4b6} The previous section described how, in principle, one might alter the motor commands from trial to trial based on sensory and reward prediction errors. Here, we wished to fit this model to people\'s data and then test the predictions of the model. In the ERR and EPE paradigm, subjects were provided with both types of error, whereas in the RWD paradigm they were provided with only reward information. Our objective was to estimate contributions of each form of error to the change in motor commands during these three paradigms. Our data from each subject consisted of the following: reach angle , visual cursor (both in units of degrees), and success or failure on that trial (reward) . If a subject generated hand position on a given trial, we assumed that this was related to three hidden variables: their estimate of perturbation , the accumulated change in the motor commands due to reward prediction errors , and an active search noise to find more rewarding motor commands : The problem is to estimate the variables of the right hand side from the measured sequence of hand positions. This requires solving an optimization problem. A rational cost is to minimize the squared difference between the observed sequence of hand positions and the sequence predicted by the model . This is equivalent to minimizing the summation of magnitude of the active search noise . The constraint equations of this optimization process are Eqs. (8) and (13). From Eq. (15) we have where is the experimenter\'s observation of subject\'s hand position, and are the memory the optimal learner model updated. We will substitute Eq. (16) into Eq. (13) to update and . We would also estimate the sensory prediction error is: where is the visual rotation that the experimenter imposed and is the estimation of the perturbation that the optimal learner updated. Then, we substitute Eq. (17) into Eq. (8) in order to update . Starting with initial conditions , , and , if we knew the unknown parameters \[, , \], we could use the sensory prediction error in Eq. (8) to update , and the reward prediction error to update while updating the estimation of the value though updating . We searched for these three unknown parameters (using lsqnonlin in Matlab 6.5) in order to minimize the squared sum of difference between the model generated sequence of hand positions and the measured hand positions for each subject. We found that in the ERR paradigm, the average of the estimated parameters were \[, , \] = \[6, 0.15, 0.04\] and in EPE paradigm, \[69, 0.39, 0.03\]. In the RWD paradigm, because no visual feedback of the hand position was provided, we assumed that motor commands were updated only by the reward prediction error. We set the Kalman gain to be zero and the average of the estimated parameters were \[, \] = \[0.13, 0.14\]. In the main document, we report the evolution of two memories and the sum of them: , , and . We wish to thank Andy Barto for helpful discussions. The authors have declared that no competing interests exist. This work was supported in part by a grant from the National Institutes of Health (NS37422) and a grant from the by Japan Society of Promotion of Science (Grants-in-Aid for scientific research 22800019). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. [^1]: Conceived and designed the experiments: JI RS. Performed the experiments: JI. Analyzed the data: JI. Contributed reagents/materials/analysis tools: RS. Wrote the paper: JI RS.

PubMed Central

2024-06-05T04:04:19.654091

2011-3-10