Figures
Abstract
Genetic susceptibility is an important contributor to the pathogenesis of Crohn's disease (CD). We investigated multiple CD susceptibility genes in an Australian paediatric onset CD cohort. Newly diagnosed paediatric onset CD patients (n = 72) and controls (n = 98) were genotyped for 34 single nucleotide polymorphisms (SNPs) in 18 genetic loci. Gene-gene interaction analysis, gene-disease phenotype analysis and genetic risk profiling were performed for all SNPs and all genes. Of the 34 SNPs analysed, four polymorphisms on three genes (NOD2, IL23R, and region 3p21) were significantly associated with CD status (p<0.05). All three CD specific paediatric polymorphisms on PSMG1 and TNFRSF6B showed a trend of association with p<0.1. An additive gene-gene interaction involving TLR4, PSMG1, TNFRSF6B and IRGM was identified with CD. Genes involved in microbial processing (TLR4, PSMG1, NOD2) were significantly associated either at the individual level or in gene-gene interactive roles. Colonic disease was significantly associated with disease SNP rs7517847 (IL23R) (p<0.05) and colonic and ileal/colonic disease was significantly associated with disease SNP rs125221868 (IBD5) and SLC22A4 & SLC22A4/5 variants (p<0.05). We were able to demonstrate genetic association of several genes to CD in a paediatric onset cohort. Several of the observed associations have not been reported previously in association with paediatric CD patients. Our findings demonstrate that CD genetic susceptibility in paediatric patients presents as a complex interaction between numerous genes.
Citation: Wagner J, Sim WH, Ellis JA, Ong EK, Catto-Smith AG, Cameron DJS, et al. (2010) Interaction of Crohn's Disease Susceptibility Genes in an Australian Paediatric Cohort. PLoS ONE 5(11): e15376. https://doi.org/10.1371/journal.pone.0015376
Editor: Stefan Bereswill, Charité-University Medicine Berlin, Germany
Received: August 10, 2010; Accepted: August 24, 2010; Published: November 8, 2010
Copyright: © 2010 Wagner et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The project was supported by research grants from the Jack Brockhoff Foundation, Australia, the Murdoch Children Research Institute, Australia, and from the Broad Medical Research Program, The Eli and Edythe Broad Foundation, USA. C. Kirkwood is supported by a Career Development Awards fellowship from the National Health and Medical Research Council, Australia (607347). J Ellis is supported by a National Health and Medical Research Council Capacity Building Grant in Population Health. The funders had no role in study design, data collection and analysis, decision to publish, or preparation to the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Crohn's disease (CD) is a chronic relapsing inflammatory disease occurring anywhere in the gastrointestinal tract, although it most commonly affects the small intestine [1]. CD is a major cause of morbidity throughout the world with an escalating epidemic of CD recorded globally in children and adults during the past few decades [2]. A worldwide study reported an incidence per 100,000 population as low as 0.3 in China to as high as 20.2 cases in Canada [2]. A ten-fold increase in the incidence of paediatric CD over a 31 year period was reported from the Royal Children's Hospital (RCH) in Melbourne, Australia [3]. Approximately 30 new cases of CD in children (age 2–16 years) are now diagnosed and treated at the RCH each year compared with approximately 3 new cases reported annually in 1975. European studies report a similar dramatic increase in the incidence of paediatric CD [4], [5]. It is widely accepted that CD is mediated by a dysfunctional immunological response of T-lymphocytes which is primarily induced in genetically susceptible individuals by the presence of an environmental stimulus [6], [7].
Genetic factors that affect susceptibility to CD have been identified using genetic linkage and population based association studies. Genetic susceptibility to CD has been extensively studied since the identification of the first CD susceptibility gene NOD2 [8], [9]. The NOD gene family is proposed to function as an intracellular pattern-recognition receptor that senses microbial muramyl dipeptide, a degradation product of peptidoglycan from bacterial cell wall and the function of a cytosolic sensor for the induction of apoptosis [10]. In the last decade several genome-wide association studies (GWAS) have discovered an increasing number of novel genes and single nucleotide polymorphisms (SNPs) associated with CD, including 21 novel loci identified in 2008 alone [11].
Paediatric-onset CD patients have a higher rate of gene mutations compared with adult patients [12]. Three mutations were reported to be specifically associated with paediatric-onset inflammatory bowel disease. One of these (rs2836878) resides in a region that harbours no gene, but is most closely located to the proteasome assembly chaperone 1 gene (PSMG1). The other two SNPs (rs4809330 and rs2315008) are located within a region containing several genes including the tumour necrosis factor receptor superfamily member 6B gene (TNFRSF6B) [13]. All three paediatric specific CD mutations were recently confirmed to be associated with CD in a Canadian study [14].
The large number of genetic variants implicated in CD requires multiple SNPs to be investigated simultaneously in CD patients to understand the individual contributions of loci in single genes and gene-gene interactions. In our study we investigated the occurrence of 34 SNPs simultaneously in a paediatric onset CD cohort.
Results
Hardy-Weinberg equilibrium testing
Three SNPs (rs2836878, rs2066845 and rs5743289 present on PSMG1 and NOD2) were found to deviate from Hardy Weinberg Equilibrium (HWE) at the P = 0.05 level. However, these SNPs were retained in the analyses as in each case the deviation from HWE was observed in the cases only. Such distortions in case genotype frequency can be an indication of association [15], [16]. Thus, alleles and genotypes from all SNPs were compared between the case and control phenotypic groups.
Genotype and allele frequencies
All SNPs were initially analysed using allelic and genotypic χ2 tests (Table S 1). For SNPs where the minor allele homozygote counts were less than 5 (rs11209026, rs3792876, rs13361189, rs6958571, rs17327442, rs4986790, rs2241136, rs2289310, rs1248696, rs1793004, rs3135932, rs2066844, rs2066845, rs5743289, rs5743293, rs2836878), we applied Fisher's exact test to obtain a genotypic P value. Four SNPs, rs2066845 and rs5743289 on NOD2 gene, rs11209026 on IL23R gene, and rs9858542 on region 3p21, demonstrated evidence of association with CD (Table 1 and Table S1).
a) Additive and genotypic logistic regression analyses.
At the individual level, allelic χ2 and genotypic Fisher's comparisons of SNP rs2066845 (NOD2) were highly significant at p<0.05 (p = 0.00016, 0.0076 respectively). However, no minor allele homozygotes or heterozygotes were observed for controls, and therefore further analyses by logistic regression were not performed for this SNP.
Allelic χ2 and genotypic Fisher's comparisons of SNP rs5743289 (NOD2) were significant at p<0.05 (p = 0.027, 0.0088 respectively). This SNP remained associated by additive, but not genotypic, logistic regression (Additive: OR = 1.9; 95% CI 1.1, 3.3; p = 0.020, Genotypic (2df): p = 0.998; OR for heterozygotes compared with major allele homozygotes = 1.02; 95% CI 0.49, 2.1 (p = 0.952), no minor allele homozygotes observed in controls).
Allelic χ2 and genotypic Fisher's comparisons of SNP rs11209026 (IL23R) were significant at the p<0.05 level (p = 0.025, 0.023 respectively). This SNP remained associated by additive logistic regression (OR = 0.26, 95% CI 0.07–0.97, p = 0.045). Genotypic logistic regression was not performed due to the lack of minor allele homozygotes.
Allelic and genotypic χ2 comparisons of SNP rs9858542 on region 3p21 were significant at the p<0.05 level (p = 0.021, 0.010 respectively). This SNP remained associated by additive and genotypic logistic regression (Additive: OR = 1.8; 95% CI 1.1–2.9; p = 0.02, Genotypic (2df): p = 0.0087).
b) Further analyses.
Further analyses were performed by comparing the disease associated allele distribution between CD patients and controls (Table 1). Eleven percent of CD patients had at least one disease-associated allele of SNP rs2066845 (NOD2) compared to none in the control group (p = 0.00081). Thirteen percent of CD patients had both disease associated alleles of SNP rs5743289 (NOD2) compared to none in the control group (p = 0.00039). Analysis of all 3 main NOD2 variants (rs2066845, rs2066844, and rs5743293) revealed that 28% of CD patients had at least one NOD2 variant compared to 11% in the control group (OR = 3.1, 95% CI 1.39–6.9, p = 0.005). One patient had a triple mutation in the NOD2 gene (heterozygote for rs2066844 and rs5743293 and homozygote for rs5743289). Three patients had a double mutation in the NOD2 gene and four patients had a single mutation in the NOD2 gene. Sixty six percent of CD patients had at least one disease associated allele of SNP 9858542 (3p21) compared to 43% in the control group (OR = 2.56, 95% CI 1.36–4.81, p = 0.003).
Three SNPs on paediatric specific CD susceptibility genes (rs2836878 on PSMG1 and rs4809330 and rs2315008 on TNFRSF6B) demonstrated some evidence of association at the p<0.1 level of significance. Fifty four percent of CD patients had at least one disease associated allele of SNP rs2836878 (PSMG1) compared to 41% in the control group (p = 0.089). Disease associated alleles of SNPs rs4809330 and rs2315008 (TNFRSF6B) were observed at lower frequency in CD (44%) patients compared to 57% in the control group (p = 0.083 and p = 0.068 for rs4809330 and rs2315008, respectively).
Gene-gene interaction
Several significant gene-gene interactions were detected for all three disease associated genes (Tables S2a–S2d). There was an under-representation of wildtype allele combinations in CD patients compared to controls harbouring gene combinations of the main NOD2 variants with four other genetic variants (PSMG1, NOD2 rs5753289, TLR4, and 3p21 (Figure 1). There was an under-representation of wildtype allele combinations in CD patients compared to controls harbouring gene combinations of 3p21 variants with four other genetic variants (PSMG1, NOD2 rs5753289, TLR4, and IRGM) (Figure 1). Interestingly, PSMG1, TLR4, and IRGM were not associated individually with paediatric CD. Conversely, wildtype allele combinations of TNFRSF6B variants with NOD2 rs5743289 or IL23R rs11209026 variants were significantly higher in CD patients compared to controls (Figure 1).
Percent values represent the proportion of wildtype gene-gene combination. Four wildtype genes were detected in combinations with main NOD2 wildtype and 3p21 wildtype significantly more often in controls compare to CD patients. TNFRSF6B wildtype gene was detected in combination with wildtype SNP rs5743289 (NOD2) and wildtype SNP rs11209026 (IL23R) significantly more often in CD patients compared to controls. All differences between CD and controls were significant at P<0.05.
Genotype-Phenotype interaction
The stratification of CD patients according to phenotype is outlined in Table 2. The majority of patients (76%) had ileal/colonic disease with or without upper gastrointestinal tract involvement. We also looked for possible correlation between genotype and disease location and disease behaviour (Table S3). Four disease SNPs from three genes had significant association with disease location. Disease SNP rs7517847 (IL23R) was found more often in CD patients with colonic disease (L2±L4) (14% (10/69)) compared to the wildtype form of this SNP in the same phenotype (1% (1/69)) (p = 0.04) (Table S3). Disease SNP rs12521868 (IBD5) was found more often in CD patients with colonic and ileal/colonic disease (L2±L4 & L3±L4) (69% (47/69)) compared to the wildtype form of this SNP in the same phenotype (24% (17/69)) (p = 0.027) (Table S3). Disease SNPs rs3792876 and rs1050152 (SLC22A4 & SLC22A4/5) were found more often in CD patients with L2±L4 & L3±L4 phenotype (71% (48/68)) compared to the wild type form of the SNPs in the same phenotype (22% (15/69)) (p = 0.019) (Table S3).
Genetic risk analysis for CD and control patients
We performed a genetic risk analysis using either all CD SNPs, or only SNPs associated with disease in this study. Quantitative analysis of all CD SNPs revealed that the proportions of patients having 7, 8, 11, 15, 16, 18, 19 and 21 CD associated SNPs were higher in the CD group than the control group (Figure S1a), whereas analysis of wildtype alleles revealed that control patients had 19, 20, 21, 22, 24, and 25 alleles (Figure S1b). However, the overall difference between CD and controls was not significant.
When taking only disease associated SNPs into consideration the overall differences in SNPs between CD patients and controls was significant (p = 0.009) (Figure 2). A significantly lower proportion of CD patients had no disease associated SNPs compared to controls (15.3% versus 28.6%, p = 0.042) whereas, a significantly higher proportion of CD patients had two disease associated SNPs compared to controls (34.7% versus 14%, p = 0.002) (Figure 2 and Table 3).
The proportion of patients carrying between none and four diseases associated SNPs was calculated for the CD and control group. P values were calculated by Chi Square analysis.
Genetic risk analysis for CD phenotype showed that one and two SNPs, respectively, were most commonly implicated in disease location “L3±L4” and disease behaviour “B1±P” (Figure 3). Location phenotype analysis revealed that rs9858542 (3p21) SNP was the most common single SNP in the most common location (L3±L4) (63.6%) followed by rs5743289 (NOD2) SNP (27.3%) (Table 4). Together these SNPs also represented the most common grouping within the most common phenotype L3+L4 (65.2%) (Table 4). Behaviour phenotype analysis revealed that the rs9858542 (3p21) SNP was the most common single SNP in the most common behaviour group (B1±P) (50%), this was followed by SNP rs5743289 (NOD2) (40%) (Table 4). Together these SNPs also represented the most common grouping within the most common behaviour group (B1±P) (71.4%) (Table 4).
The proportion of patients carrying between none and four diseases associated SNPs was stratified by CD phenotype. L1±4 = ileal disease with or without upper gastrointestinal tract (GI) involvement, L2±L4 = colonic disease with or without upper GI involvement, L3±L4 = ileal/colonic disease with or without upper GI involvement, L4 = upper GI disease, B1±P = inflammatory appearance with or without perininal disease, B2±P = stricturing appearance with or without perininal disease, B3P = penetrating appearance with perianal disease.
Discussion
This study analysed a paediatric-onset CD population for the prevalence of 34 SNPs present on 18 genes, to investigate their gene-gene interaction and to perform genetic risk profiling. Four SNP variants present on NOD2, IL23R and on a 3p21 chromosomal region were significantly associated with our CD population. At the individual level, these SNPs have been reported previously, but no studies have investigated their interaction in a paediatric CD cohort. Three CD specific paediatric SNP variants present on gene PSMG1 and TNFRSF6B were also included in our investigation [13] and while none showed a significant association, all three showed a trend towards association (p<0.1). This represents the second independent confirmation in a case-control study of a possible role for these SNPs in development of CD. There was a higher representation of PSMG1 SNP variant in CD patients, while a higher representation of TNFRSF6B SNP variants was observed for controls. The higher representation of TNFRSF6B SNP variants in controls is in contrast to the original study and to a Canadian case/control study [13], [14]. This study also demonstrated that a small patient cohort was sufficient for inferences of CD predisposing gene-gene interactions in association with paediatric-onset disease.
Our combined heterozygous/homozygous detection rate of the main NOD2 variants in CD patients was 28% compared with 11% in the control group. Previous studies have reported a detection rate between 11%–41% in CD patients and 3–11% in controls [12], [17], [18]. These genetic differences possibly reflect regional and ethnic differences in study populations, highlighted by the virtual absence of NOD2 variants in a Japanese study [19]. SNP rs5743289 of NOD2 was previously identified in GWA studies using paediatric and adult cohorts [13], [20]. The significant association of the minor homozygote variant with our CD group confirms the earlier GWAS findings and strengthens the role of NOD2 in paediatric CD patients. One patient with a triple mutation in the NOD2 was a 13 year old girl with a L3+4/B1 phenotype. At the time of initial CD diagnosis she had presented with oesophagitis, focal active gastritis, granulomatous colitis consistent with CD in the colon, and chronic active proctitis, however, the role of NOD2 triple mutation in this multiple disease presentation is not clear.
IL23R variants were first described in 2006 [21]. In our study the protective minor A allele of SNP rs11209026 was detected in 15% of controls and in 6% of CD patients. In two previous paediatric studies, the detection rate was 5.5% and 6% in controls and 3% and 2% in CD patients [22], [23]. Intronic IL23R SNP variants (rs7517847 and rs1004819) were not associated with our CD group, which is contrary to the Canadian paediatric study [22]. However, colonic disease appeared to be significantly more common in our CD patients with disease associated SNP rs751787 compared to wild type genotype, suggesting that genetic alterations might play a role in CD phenotypic appearance.
The intronic synonymous SNP rs9858542 on 3p21 in close proximity to the Basson (BSN) gene was first reported to be associated with CD by the Welcome Trust Case Control Consortium [24]. The significant association of the heterozygote/homozygote genotype identified in our study (66% in CD and 43% in controls) is similar to two other studies (60% and 61.5% in CD and 45% and 52.4% in controls) [25], [26]. We also confirmed the involvement of the minor risk allele (A) as reported by a German and Spanish study [26] but not detected in a US paediatric study [27]. The role of synonymous SNPs in CD should not be underestimated. It was reported that synonymous SNPs can alter mRNA stability, gene expression or can act in linkage disequilibrium with other important SNPs [28].
Gene-gene interaction analysis performed by stratification of disease associated SNPs both within, and between, the candidate genes examined in this study have revealed some very interesting findings. In particular the apparent interaction of PSMG1 and TLR4 with the main NOD2 variants, and PSMG1, TLR4, and IRGM with 3p21, are of interest because these genes were not associated individually with paediatric CD in our cohort (Figure 1). This finding illustrates the complex genetic architecture of CD, in that it is unlikely to be dependent on a single gene but probably is polygenic in nature. A number of genes in combination are likely to affect immunological response and microbial detection, and hence CD risk.
Several of the identified gene-gene variants have been implicated in microbial detection and interaction. Best known is NOD2, the product of which is important for the innate recognition of bacterial lipopolysaccharides and peptidoglycans [10], [29], [30]. TLR4 SNP variant was not associated individually in our CD cohort. However, its significant gene-gene association with NOD2 and the lipopolysaccharide-signalling role of cell surface toll-like receptors [31], [32], provides strong evidence of a microbial role in CD genetically susceptible individuals.
The role of PSMG1 SNP variant which showed a trend towards an association with CD, and its significant association with NOD2 and 3p21 variants suggests an influence on chaperone-driven proteasome assembly which is important in degradation of proteins [33]. Up to a three-fold increase in expression of the proteasome subunit (LMP2), which plays a role in the formation of immunoproteasome, has been reported in the inflamed gut of patients with CD and ulcerative colitis [34], [35]. Bacterial lipopolysaccharides have been shown to trigger the formation of immunoproteasomes in vivo mouse cell culture models [36] and play a role in the generation of active NF-kappaB subunits [37], [38]. The significantly larger double mutation rate of NOD2 and PSMG1 variants in the CD group (60%) compared with the control group (18%) may also suggest that NOD2 mutations affect NF-kappaB signalling and PSMG1 mutations may potentiate microbial-triggered inflammation in CD patients.
IRGM gene expression regulates cellular autophagy of internalized bacteria, a process implicated in CD [39]. Studies have shown that IRGM gene mutation was not confirmed in children but in ileal CD in the adult population [40]–[42]. However, its role in gene-gene association with 3p21 as identified in our study requires further evaluation. A recent study reported an association between 3p21 variants and another variant of IRGM gene (rs10000113) which was not investigated in our study [43].
Another interesting finding of our study is the negative association of TNFRSF6B mutations with CD. The trend towards association with controls at the individual level and significant associations with SNPs rs5743289 (NOD2) and rs11209026 (IL23R) in controls; point towards a possible role in protection against development of CD. The role of mutations in tumour necrosis factor receptor genes in CD is not well known. A case/control study investigating genetic variants of TNFRSF1A and 1B in association with the three main NOD2 variants, reported that one out of two SNPs from each gene was significantly implicated in the CD cohort [44].
The non-significant differences considering all SNPs, between CD patients and controls, demonstrate that large genotypic variation occurs in the general population. As a result, a clear IBD wildtype genotype is difficult to define, but is rather a mixture of major homozygote and heterozygote combination.
This study illustrates that an association between the number of SNPs and disease status can be established. A high occurrence of SNP rs9858542 (3p21) with ileal/colonic disease and inflammatory behaviour with SNP rs5743289 (NOD2) were identified. Other studies have been equivocal about the association of rs9858542 with CD phenotype [25], [26].
The advantage of this study is that our analysis combined 34 CD susceptibility SNPs in a single paediatric onset cohort. As a result we report novel findings of associations between disease-associated SNPs and paediatric CD phenotypes. A limitation of this study is the relatively small sample size compared to adult studies. Currently, we are not able to repeat the study by recruiting another paediatric cohort to confirm our findings.
In conclusion, this study has shown that CD susceptibility genes are likely act in a complex interactive manner in paediatric-onset CD. Several genes involved in microbial processing (TLR4, PSMG1, NOD2) were significantly associated either at the individual level or synergistically with other genes. A possible novel protective effect of TNFRSF6B genetic variants, in combination with two other genes, was suggested however, this was not confirmed by a larger cohort study [14]. Many of the genetic interactions identified have not been reported previously. The results are important to understanding the pathogenesis of CD, however, need to be confirmed in future studies.
Methods
Study population
In this study 72 paediatric CD patients and 98 paediatric control patients were analysed. All patients were admitted through the Department of Gastroenterology at the Royal Children's Hospital, Melbourne, Australia. The clinical diagnosis of CD was established using standard clinical, endoscopic, and histopathological criteria according to the Montreal classification [45]. Patients in the control group had been admitted for investigation of symptoms of inflammatory bowel disease (IBD) but were diagnosed either with gastritis, oesophagitis or no pathological condition. All the patients were recruited at initial diagnosis. UC patients were not included in this study due to the low number of patients available. The mean age in the CD group and control group were 11.6 years (2.2–17.2) and 11.9 years (1.7–19.8), respectively. The male/female ratio in the CD group and control group was 46/26 and 45/53, respectively. The CD phenotype subgroups are present in Table 2.
Ethics Statement
Ethics approval for the study was obtained from the Human Research Ethics Committee of the Royal Children's Hospital (EHRC no.23003). Written informed consent was obtained from each individual, parent or guardian prior to enrolment in the study.
Genotype Analysis
34 SNPs from 18 genes were selected for analysis. The SNPs and genes were selected from published data that implicated these SNPs in children and/or adults (Table 5).
Genomic DNA was extracted from gut biopsies or blood according to protocols in our laboratory [46], [47]. The SNP site flanking regions were retrieved from NCBI SNP reference assembly database (Build 131). The Sequenom genotyping tools (www.mysequenom.com) designed PCR amplification, extension primers and grouped the 34 SNPs into two multiplex assays - 19 multiplex and 16 multiplex assays. The multiplex PCR and extension reactions were carried out using the Sequenom iPLEX Gold reaction protocol. Genotyping was performed using the matrix-assisted laser desorption ionization time-of-flight (MALDI TOF) mass spectrometry platform [48]. Briefly, the assays were performed in 5 µl volume and contained 1 µl DNA (10–20ng), 1× PCR buffer, 2mM MgCl2, 500 uM dNTPs, 100nM of PCR primer mix, 0.5 unit PCR enzyme and nuclease free water. The PCR cycling conditions were: 94°C for 4 minutes, and 45 cycles at 94°C for 20 seconds, 56°C for 30 seconds, and 72°C for 1 minute, and a final extension of 72°C for 3 minutes. The second extension PCR reaction was performed by adding 2 µl iPLEX Gold reaction mix to the cleaned up primary PCR product. The iPLEX termination mix contained 0.2 µl iPLEX buffer plus 0.2 µl iPLEX termination mix, 0.94 µl extension primer mix, 0.041 µl iPLEX enzyme, and nuclease free water. The iPLEX cycling conditions were: 94°C for 30 seconds, and 40 cycles at 94°C for 5 seconds followed by 5 cycles at 52°C for 5 seconds and 80°C for 5 seconds. The final extension was at 72°C for 3 minutes. The iPLEX Gold reaction was purified up by adding 16 µl of nuclease free water and 6 mg resin to each well of the 384 well plates and rotated for 5 minutes and centrifuged at 3200 g for 5 minutes. The products were then transferred to a Sequenom SpectroCHIP and analysed on a MALDI-TOF mass spectrometer (Sequenom MassARRAY) and the SNP calls were viewed using the MassARRAY TYPER 4.05 analyser software. In the majority of SNPs (32/34) a base call was obtained for at least 99% CD samples and control samples. Eighty sequencing reactions were performed to retrieve missing base calls, mainly from the TLR4, NOD2, and DLG5.
Statistical Analyses
Statistical analyses were performed using PLINK version 1.05 statistical software package [49] (http://pngu.mgh.harvard.edu/~purcell/plink/). Differences between cases and controls were assessed using allelic and genotypic χ2 analyses. The relationship between case/control status and each SNP showing some evidence of association by χ2 testing was also considered using additive and genotypic regression methods with adjustment for the covariates age and sex. Gene-gene interaction analysis was calculated using χ2 analyses. For this the wildtype form (major homozygote) of any disease associated genes was stratified with wildtype form of all other genes included in this study. Genotype-phenotype interaction analysis was calculated using Fisher exact test. STATA version 11 was used for χ2 and Fisher exact test analyses. Where the number was 5 or below 5 Fisher exact test was used. Multiple testing adjustments were not performed because the study is in essence replication of previous studies regarding individual SNPs (not discovery) (Table 5). This study investigated SNPs that had been previously associated with CD, significantly reducing the risk of false positive findings [50].
Supporting Information
Figure S1.
https://doi.org/10.1371/journal.pone.0015376.s001
Genetic risk profile analysis in CD patients and controls for all disease associated SNPs. The proportion of patients carrying between the minimum number (n=7) and maximum number of CD associated SNPs (n=21) were calculated for the CD and control group (Figure S1a). Genetic risk profile analysis in CD patients and controls for all wildtype SNPs. The proportion of patients carrying between the minimum number (n=12) and maximum number of wildtype SNPs (n=27) were calculated for the CD and control group (Figure S1b). (PDF)
Table S1.
https://doi.org/10.1371/journal.pone.0015376.s002
Genotypic and allelic distribution for all SNPs. Genotype (GENO) frequencies expressed as minor allele homozygote/heterozygote/major allele homozygote and allele frequencies expressed as minor allele/major allele for all SNPs are outlined. (PDF)
Table S2.
https://doi.org/10.1371/journal.pone.0015376.s003
Gene‐gene interaction with main NOD2 variants, NOD2 rs5743289 variant, with IL23R rs11209026 variant, and 3p21 rs9858542 variant, respectively. (PDF)
Table S3.
https://doi.org/10.1371/journal.pone.0015376.s004
Genotype frequency in CD cases stratified by CD phenotype. (PDF)
Acknowledgments
We thank the children and their families for their participation in this study. We thank Don Vicendese (Murdoch Childrens Research Institute, Australia) for assistance with STATA.
Author Contributions
Conceived and designed the experiments: JW EKO RB CDK. Performed the experiments: JW EKO. Analyzed the data: JW JAE. Wrote the paper: JW RB CDK. Collection of patient data and extraction of DNA, WS. Patient recruitment, clinical sample collection and critical manuscript reviewing: AGC-S DJWC.
References
- 1. Fiocchi C (1998) Inflammatory bowel disease: etiology and pathogenesis. Gastroenterology 115: 182–205.
- 2. Economou M, Pappas G (2008) New global map of Crohn's disease: Genetic, environmental, and socioeconomic correlations. Inflamm Bowel Dis 14: 709–720.
- 3. Phavichitr N, Cameron DJ, Catto-Smith AG (2003) Increasing incidence of Crohn's disease in Victorian children. J Gastroenterol Hepatol 18: 329–332.
- 4. Cosgrove M, Al-Atia RF, Jenkins HR (1996) The epidemiology of paediatric inflammatory bowel disease. Arch Dis Child 74: 460–461.
- 5. Vind I, Riis L, Jess T, Knudsen E, Pedersen N, et al. (2006) Increasing incidences of inflammatory bowel disease and decreasing surgery rates in Copenhagen City and County, 2003–2005: a population-based study from the Danish Crohn colitis database. Am J Gastroenterol 101: 1274–1282.
- 6. Shanahan F (2002) Crohn's disease. Lancet 359: 62–69.
- 7. Sartor RB (2006) Mechanisms of disease: pathogenesis of Crohn's disease and ulcerative colitis. Nat Clin Pract Gastroenterol Hepatol 3: 390–407.
- 8. Hugot JP, Chamaillard M, Zouali H, Lesage S, Cezard JP, et al. (2001) Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature 411: 599–603.
- 9. Hugot JP, Laurent-Puig P, Gower-Rousseau C, Olson JM, Lee JC, et al. (1996) Mapping of a susceptibility locus for Crohn's disease on chromosome 16. Nature 379: 821–823.
- 10. Inohara N, Nunez G (2003) NODs: intracellular proteins involved in inflammation and apoptosis. Nat Rev Immunol 3: 371–382.
- 11. Barrett JC, Hansoul S, Nicolae DL, Cho JH, Duerr RH, et al. (2008) Genome-wide association defines more than 30 distinct susceptibility loci for Crohn's disease. Nat Genet 40: 955–962.
- 12. de Ridder L, Weersma RK, Dijkstra G, van der Steege G, Benninga MA, et al. (2007) Genetic susceptibility has a more important role in pediatric-onset Crohn's disease than in adult-onset Crohn's disease. Inflamm Bowel Dis 13: 1083–1092.
- 13. Kugathasan S, Baldassano RN, Bradfield JP, Sleiman PM, Imielinski M, et al. (2008) Loci on 20q13 and 21q22 are associated with pediatric-onset inflammatory bowel disease. Nat Genet 40: 1211–1215.
- 14. Amre DK, Mack DR, Morgan K, Fujiwara M, Israel D, et al. (2009) Investigation of reported associations between the 20q13 and 21q22 loci and pediatric-onset Crohn's disease in Canadian children. Am J Gastroenterol 104: 2824–2828.
- 15. Deng HW, Chen WM, Recker RR (2000) QTL fine mapping by measuring and testing for Hardy-Weinberg and linkage disequilibrium at a series of linked marker loci in extreme samples of populations. American journal of human genetics 66: 1027–1045.
- 16. Lee WC (2003) Searching for disease-susceptibility loci by testing for Hardy-Weinberg disequilibrium in a gene bank of affected individuals. American journal of epidemiology 158: 397–400.
- 17. Ferraris A, Knafelz D, Torres B, Fortina P, Castro M, et al. (2005) Analysis of CARD15 gene variants in Italian pediatric patients with inflammatory bowel diseases. J Pediatr 147: 272–273.
- 18. Tomer G, Ceballos C, Concepcion E, Benkov KJ (2003) NOD2/CARD15 variants are associated with lower weight at diagnosis in children with Crohn's disease. Am J Gastroenterol 98: 2479–2484.
- 19. Yamazaki K, Takazoe M, Tanaka T, Kazumori T, Nakamura Y (2002) Absence of mutation in the NOD2/CARD15 gene among 483 Japanese patients with Crohn's disease. J Hum Genet 47: 469–472.
- 20. Libioulle C, Louis E, Hansoul S, Sandor C, Farnir F, et al. (2007) Novel Crohn disease locus identified by genome-wide association maps to a gene desert on 5p13.1 and modulates expression of PTGER4. PLoS Genet 3: e58.
- 21. Duerr RH, Taylor KD, Brant SR, Rioux JD, Silverberg MS, et al. (2006) A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science 314: 1461–1463.
- 22. Amre DK, Mack D, Israel D, Morgan K, Lambrette P, et al. (2008) Association between genetic variants in the IL-23R gene and early-onset Crohn's disease: results from a case-control and family-based study among Canadian children. Am J Gastroenterol 103: 615–620.
- 23. Van Limbergen J, Russell RK, Nimmo ER, Drummond HE, Smith L, et al. (2007) IL23R Arg381Gln is associated with childhood onset inflammatory bowel disease in Scotland. Gut 56: 1173–1174.
- 24. (2007) Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature 447: 661–678.
- 25. Latiano A, Palmieri O, Corritore G, Valvano MR, Bossa F, et al. (2009) Variants at the 3p21 locus influence susceptibility and phenotype both in adults and early-onset patients with inflammatory bowel disease. Inflamm Bowel Dis.
- 26. Marquez A, Cenit MC, Nunez C, Mendoza JL, Taxonera C, et al. (2009) Effect of BSN-MST1 locus on inflammatory bowel disease and multiple sclerosis susceptibility. Genes Immun 10: 631–635.
- 27. Peterson N, Guthery S, Denson L, Lee J, Saeed S, et al. (2008) Genetic variants in the autophagy pathway contribute to paediatric Crohn's disease. Gut 57: 1336–1337; author reply 1337.
- 28. Duan J, Wainwright MS, Comeron JM, Saitou N, Sanders AR, et al. (2003) Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA stability and synthesis of the receptor. Hum Mol Genet 12: 205–216.
- 29. Bonen DK, Ogura Y, Nicolae DL, Inohara N, Saab L, et al. (2003) Crohn's disease-associated NOD2 variants share a signaling defect in response to lipopolysaccharide and peptidoglycan. Gastroenterology 124: 140–146.
- 30. Inohara N, Ogura Y, Fontalba A, Gutierrez O, Pons F, et al. (2003) Host recognition of bacterial muramyl dipeptide mediated through NOD2. Implications for Crohn's disease. J Biol Chem 278: 5509–5512.
- 31. Medzhitov R (2001) Toll-like receptors and innate immunity. Nat Rev Immunol 1: 135–145.
- 32. Hausmann M, Kiessling S, Mestermann S, Webb G, Spottl T, et al. (2002) Toll-like receptors 2 and 4 are up-regulated during intestinal inflammation. Gastroenterology 122: 1987–2000.
- 33. Rosenzweig R, Glickman MH (2008) Chaperone-driven proteasome assembly. Biochem Soc Trans 36: 807–812.
- 34. Visekruna A, Slavova N, Dullat S, Grone J, Kroesen AJ, et al. (2009) Expression of catalytic proteasome subunits in the gut of patients with Crohn's disease. Int J Colorectal Dis 24: 1133–1139.
- 35. Fitzpatrick LR, Small JS, Poritz LS, McKenna KJ, Koltun WA (2007) Enhanced intestinal expression of the proteasome subunit low molecular mass polypeptide 2 in patients with inflammatory bowel disease. Dis Colon Rectum 50: 337–348; discussion 348–350.
- 36. Stohwasser R, Giesebrecht J, Kraft R, Muller EC, Hausler KG, et al. (2000) Biochemical analysis of proteasomes from mouse microglia: induction of immunoproteasomes by interferon-gamma and lipopolysaccharide. Glia 29: 355–365.
- 37. Hayashi T, Faustman D (2000) Essential role of human leukocyte antigen-encoded proteasome subunits in NF-kappaB activation and prevention of tumor necrosis factor-alpha-induced apoptosis. J Biol Chem 275: 5238–5247.
- 38. Hayden MS, Ghosh S (2004) Signaling to NF-kappaB. Genes Dev 18: 2195–2224.
- 39. McCarroll SA, Huett A, Kuballa P, Chilewski SD, Landry A, et al. (2008) Deletion polymorphism upstream of IRGM associated with altered IRGM expression and Crohn's disease. Nat Genet 40: 1107–1112.
- 40. Amre DK, Mack DR, Morgan K, Krupoves A, Costea I, et al. (2009) Autophagy gene ATG16L1 but not IRGM is associated with Crohn's disease in Canadian children. Inflamm Bowel Dis 15: 501–507.
- 41. Roberts RL, Hollis-Moffatt JE, Gearry RB, Kennedy MA, Barclay ML, et al. (2008) Confirmation of association of IRGM and NCF4 with ileal Crohn's disease in a population-based cohort. Genes Immun 9: 561–565.
- 42. Van Limbergen J, Russell RK, Nimmo ER, Drummond HE, G D, et al. (2009) Germline variants of IRGM in childhood-onset Crohn's disease. Gut 58: 610–611.
- 43. Latiano A, Palmieri O, Corritore G (2010) Variants at the 3p21 Locus influence susceptibility and phenotype both in adults and early-onset patients with inflammatory bowel disease. Inflamm Bowel Dis.
- 44. Waschke KA, Villani AC, Vermeire S, Dufresne L, Chen TC, et al. (2005) Tumor necrosis factor receptor gene polymorphisms in Crohn's disease: association with clinical phenotypes. Am J Gastroenterol 100: 1126–1133.
- 45. Silverberg MS, Satsangi J, Ahmad T, Arnott ID, Bernstein CN, et al. (2005) Toward an integrated clinical, molecular and serological classification of inflammatory bowel disease: Report of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can J Gastroenterol 19: Suppl A5–36.
- 46. Kirkwood CD, Wagner J, Boniface K, Vaughan J, Michalski WP, et al. (2009) Mycobacterium avium subspecies paratuberculosis in children with early-onset Crohn's disease. Inflamm Bowel Dis 15: 1643–1655.
- 47. Wagner J, Short K, Catto-Smith AG, Cameron DJ, Bishop RF, et al. (2008) Identification and characterisation of Pseudomonas 16S ribosomal DNA from ileal biopsies of children with Crohn's disease. PLoS One 3: e3578.
- 48. Jurinke C, van den Boom D, Cantor CR, Koster H (2002) Automated genotyping using the DNA MassArray technology. Methods Mol Biol 187: 179–192.
- 49. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, et al. (2007) PLINK: a tool set for whole-genome association and population-based linkage analyses. American journal of human genetics 81: 559–575.
- 50. Wacholder S, Chanock S, Garcia-Closas M, El Ghormli L, Rothman N (2004) Assessing the probability that a positive report is false: an approach for molecular epidemiology studies. J Natl Cancer Inst 96: 434–442.
- 51. Cucchiara S, Latiano A, Palmieri O, Staiano AM, D'Inca R, et al. (2007) Role of CARD15, DLG5 and OCTN genes polymorphisms in children with inflammatory bowel diseases. World J Gastroenterol 13: 1221–1229.
- 52. Lesage S, Zouali H, Cezard JP, Colombel JF, Belaiche J, et al. (2002) CARD15/NOD2 mutational analysis and genotype-phenotype correlation in 612 patients with inflammatory bowel disease. Am J Hum Genet 70: 845–857.
- 53. Rigoli L, Romano C, Caruso RA, Lo Presti MA, Di Bella C, et al. (2008) Clinical significance of NOD2/CARD15 and Toll-like receptor 4 gene single nucleotide polymorphisms in inflammatory bowel disease. World J Gastroenterol 14: 4454–4461.
- 54. Franchimont D, Vermeire S, El Housni H, Pierik M, Van Steen K, et al. (2004) Deficient host-bacteria interactions in inflammatory bowel disease? The toll-like receptor (TLR)-4 Asp299gly polymorphism is associated with Crohn's disease and ulcerative colitis. Gut 53: 987–992.
- 55. Waller S, Tremelling M, Bredin F, Godfrey L, Howson J, et al. (2006) Evidence for association of OCTN genes and IBD5 with ulcerative colitis. Gut 55: 809–814.
- 56. Ogura Y, Bonen DK, Inohara N, Nicolae DL, Chen FF, et al. (2001) A frameshift mutation in NOD2 associated with susceptibility to Crohn's disease. Nature 411: 603–606.
- 57. McGovern DP, Hysi P, Ahmad T, van Heel DA, Moffatt MF, et al. (2005) Association between a complex insertion/deletion polymorphism in NOD1 (CARD4) and susceptibility to inflammatory bowel disease. Hum Mol Genet 14: 1245–1250.
- 58. Glas J, Seiderer J, Wetzke M, Konrad A, Torok HP, et al. (2007) rs1004819 is the main disease-associated IL23R variant in German Crohn's disease patients: combined analysis of IL23R, CARD15, and OCTN1/2 variants. PLoS One 2: e819.
- 59. Latiano A, Palmieri O, Valvano MR, D'Inca R, Cucchiara S, et al. (2008) Replication of interleukin 23 receptor and autophagy-related 16-like 1 association in adult- and pediatric-onset inflammatory bowel disease in Italy. World J Gastroenterol 14: 4643–4651.
- 60. Gasche C, Grundtner P, Zwirn P, Reinisch W, Shaw SH, et al. (2003) Novel variants of the IL-10 receptor 1 affect inhibition of monocyte TNF-alpha production. J Immunol 170: 5578–5582.
- 61. Stoll M, Corneliussen B, Costello CM, Waetzig GH, Mellgard B, et al. (2004) Genetic variation in DLG5 is associated with inflammatory bowel disease. Nat Genet 36: 476–480.
- 62. Newman WG, Gu X, Wintle RF, Liu X, van Oene M, et al. (2006) DLG5 variants contribute to Crohn disease risk in a Canadian population. Hum Mutat 27: 353–358.
- 63. Browning BL, Huebner C, Petermann I, Demmers P, McCulloch A, et al. (2007) Association of DLG5 variants with inflammatory bowel disease in the New Zealand Caucasian population and meta-analysis of the DLG5 R30Q variant. Inflamm Bowel Dis 13: 1069–1076.
- 64. Noble CL, Nimmo ER, Drummond H, Smith L, Arnott ID, et al. (2005) DLG5 variants do not influence susceptibility to inflammatory bowel disease in the Scottish population. Gut 54: 1416–1420.
- 65. Lakner L, Csongei V, Sarlos P, Jaromi L, Safrany E, et al. (2009) IGR2096a_1 T and IGR2198a_1 C alleles on IBD5 locus of chromosome 5q31 region confer risk for Crohn's disease in Hungarian patients. Int J Colorectal Dis 24: 503–507.
- 66. Rioux JD, Daly MJ, Silverberg MS, Lindblad K, Steinhart H, et al. (2001) Genetic variation in the 5q31 cytokine gene cluster confers susceptibility to Crohn disease. Nat Genet 29: 223–228.
- 67. Noble CL, Nimmo ER, Drummond H, Ho GT, Tenesa A, et al. (2005) The contribution of OCTN1/2 variants within the IBD5 locus to disease susceptibility and severity in Crohn's disease. Gastroenterology 129: 1854–1864.
- 68. Russell RK, Drummond HE, Nimmo ER, Anderson NH, Noble CL, et al. (2006) Analysis of the influence of OCTN1/2 variants within the IBD5 locus on disease susceptibility and growth indices in early onset inflammatory bowel disease. Gut 55: 1114–1123.
- 69. Peltekova VD, Wintle RF, Rubin LA, Amos CI, Huang Q, et al. (2004) Functional variants of OCTN cation transporter genes are associated with Crohn disease. Nat Genet 36: 471–475.
- 70. Prescott NJ, Fisher SA, Franke A, Hampe J, Onnie CM, et al. (2007) A nonsynonymous SNP in ATG16L1 predisposes to ileal Crohn's disease and is independent of CARD15 and IBD5. Gastroenterology 132: 1665–1671.
- 71. Cummings JR, Cooney R, Pathan S, Anderson CA, Barrett JC, et al. (2007) Confirmation of the role of ATG16L1 as a Crohn's disease susceptibility gene. Inflamm Bowel Dis 13: 941–946.
- 72. Hampe J, Franke A, Rosenstiel P, Till A, Teuber M, et al. (2007) A genome-wide association scan of nonsynonymous SNPs identifies a susceptibility variant for Crohn disease in ATG16L1. Nat Genet 39: 207–211.
- 73. Parkes M, Barrett JC, Prescott NJ, Tremelling M, Anderson CA, et al. (2007) Sequence variants in the autophagy gene IRGM and multiple other replicating loci contribute to Crohn's disease susceptibility. Nat Genet 39: 830–832.
- 74. Weersma RK, Stokkers PC, Cleynen I, Wolfkamp SC, Henckaerts L, et al. (2009) Confirmation of multiple Crohn's disease susceptibility loci in a large Dutch-Belgian cohort. Am J Gastroenterol 104: 630–638.
- 75. Yamazaki K, Takahashi A, Takazoe M, Kubo M, Onouchi Y, et al. (2009) Positive association of genetic variants in the upstream region of NKX2-3 with Crohn's disease in Japanese patients. Gut 58: 228–232.
- 76. Yu W, Lin Z, Kelly AA, Hegarty JP, Poritz LS, et al. (2009) Association of a Nkx2-3 polymorphism with Crohn's disease and expression of Nkx2-3 is up-regulated in B cell lines and intestinal tissues with Crohn's disease. Journal of Crohn's and Colitis 3: 189.
- 77. Franke A, Hampe J, Rosenstiel P, Becker C, Wagner F, et al. (2007) Systematic association mapping identifies NELL1 as a novel IBD disease gene. PLoS One 2: e691.
- 78. Krupoves A, Seidman EG, Mack D, Israel D, Morgan K, et al. (2008) Associations between ABCB1/MDR1 gene polymorphisms and Crohn's disease: A gene-wide study in a pediatric population. Inflamm Bowel Dis.
- 79. Latiano A, Palmieri O, Valvano MR, D'Inca R, Caprilli R, et al. (2008) The association of MYO9B gene in Italian patients with inflammatory bowel diseases. Aliment Pharmacol Ther 27: 241–248.