Autophagy gene-IRGM polymorphisms confer susceptibility to CD but not UC, especially in Europeans

[kiny]

full: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0080602

Xiao Cheng Lu,Yi Tao,Chen Wu,Peng Lai Zhao,Kai Li,Jin Yu Zheng,Li Xin Li

Background

Polymorphisms in immunity-related GTPase family M (IRGM) gene may be associated with inflammatory bowel disease (IBD) by affecting autophagy. However, the genetic association studies on three common variants in IRGM gene (rs13361189, rs4958847 and rs10065172) have shown inconsistent results.

Methodology/ Principal Findings

The PubMed and Embase were searched up to June 5, 2013 for studies on the association between three IRGM polymorphisms and IBD risk. Data were extracted and the odd ratios (ORs) and 95% confidence intervals (95% CIs) were calculated. Finally, we performed a meta-analysis of 25 eligible studies in 3 SNPs located at IRGM gene by using a total of 20590 IBD cases and 27670 controls. The analysis showed modest significant association for the rs13361189, rs4958847 and rs10065172 variants in Crohn’s disease (CD): the risk estimates for the allele contrast were OR=1.306 (1.200-1.420), p=5.2×10-10, OR=1.182 (1.082-1.290), p=0.0002, and OR=1.248 (1.057-1.473), p=0.009 respectively (still significant when the p value was Bonferroni adjusted to 0.017). When stratified by ethnicity, significantly increased CD risk was observed in Europeans, but not in Asians. Conversely, there was no association of rs13361189 or rs4958847 variant with risk of ulcerative colitis (UC).

Conclusions/ Significance

These results indicated that autophagy gene-IRGM polymorphisms appear to confer susceptibility to CD but not UC, especially in Europeans. Our data may provide further understanding of the role of autophagy in the pathogenesis of CD.

 

[kiny]

Discussion

IRGM gene, located on chromosome 5q33.1, plays an important role in autophagy. Autophagy has a central function in physiological and pathological processes, involving in innate and adaptive immunity by delivering intracellular pathogens and other antigens. Singh et al demonstrated that IRGM can induce autophagy to eliminate intracellular mycobacterium tuberculosis [41]. Moreover, IRGM 1 -deficient mice have a reduced defense against intracellular pathogens such as Toxoplasma gondii and Listeria monocytogenes [42]. The presence of bacterial flora is essential for IBD formation in animal models [43]. In addition, autophagy plays an important role in the elimination of apoptotic bodies, and the failure of which might contribute to persistent inflammation in CD [44]. In the recent GWAS meta-analysis, Frank provided strong evidence for the association between IRGM rs7714584 and the risk of IBD (P<10-8, meeting genome-wide significance). Parkes et al [18] found two immediately flanking IRGM variants (rs13361189 and rs4958847) associated with CD in two different British cohorts. Thus, IRGM may appear to be a good candidate for IBD. However, the results of associations between three IRGM variants (rs13361189, rs4958847 and rs10065172) and risk of IBD were contradictory. Therefore, we saw the need to perform pooled analyses with larger sample size by summarizing previous case-control studies in order to understand the association between IRGM variants and IBD risk better.

Our meta-analysis showed significant susceptibility of CD from rs13361189 in the overall population in all genetic contrasts. When stratified by ethnicity, a significant association with rs13361189 was observed in European population, but not in Asians. Similar results were also found between the rs4958847 or rs10065172 variant and risk of CD in overall and European population. It is widely accepted that genetic markers in predisposition to IBD vary across ethnic groups. For instance, nucleotide oligomerization domain 2 (NOD2) polymorphisms have been strongly association with CD in Europeans [45,46], but not in Asian population [47-49]. These results suggested that rs13361189 and rs4958847 variants might be ethnic population-specific risk factors for CD. However, the lack of association in the Asian population from this study might not be very conclusive owing to the relatively small number of Asian populations used in the analysis (only 2 studies for rs4955847 and 3 studies for rs13361189 in Asian population). Therefore, further studies in Asian populations with larger sample sizes might need to be performed to clarify possible roles of IRGM polymorphisms in CD. Since these SNPs are close to each other, we used 1000 Genomes Pilot sequence data to identify whether these SNPs were in linkage disequilibrium (LD) (r2>0.8). The results showed that rs13361189 was in perfect LD with 2 other IRGM SNPs (rs10065172 and rs1000113, r2=1.000) in Europeans. However, rs4958847 was not in LD with 3 other IRGM SNPs (rs13361189, rs10065172 and rs1000113, r2=0.304).

Overall, no significant association between rs13361189 or rs4958847variant and susceptibility to UC was found in this meta-analysis in any genetic model. To date, there was lack of association of these two SNPs with UC in all the individual studies. Recently, twin studies and familial clustering of cases suggested that genetic factors were likely to play a more prominent role in CD than in UC [3]. This observation was also supported by the finding that both NOD2 and ATG16L1 were associated with CD, but not with UC [50,51].

Heterogeneity was significant for the most comparisons of rs13361189 polymorphism in overall population. To identify the source of heterogeneity, meta-regression and subgroup analysis were carried out. We found that ethnicity was identified as a potential source of between-study heterogeneity. Meta-regression indicated that ethnicity could explain 41.37% of τ2. Moreover, the heterogeneity was remarkably decreased among Asian and European population, (CC vs TT: P=0.763, P=0.336, respectively), which may be attributed that IBD is a complex disease and different genetic backgrounds or different environments existed among different ethnicities. Moreover, the sample size could explain 63.28% of τ2 under dominant model. In addition, the pooled OR did not change in the sensitivity analysis by excluding studies departed from HWE.

Our meta-analysis significantly increased statistical power by pooling data from different studies, while several limitations should be considered in the present meta-analysis. First, only 2 and 3 studies were performed in Asians for rs4955847 and rs13361189 variants, respectively. Therefore validation of association is required in other population. Second, significant heterogeneity between studies was detected in the current meta-analysis, whereas difference in ethnicity was identified as potential sources of heterogeneity. Third, gene–environment and gene-gene interactions were not analyzed because of insufficient data.

In conclusion, despite these limitations, our meta-analysis still yields statistically significant results. The present data synthesis indicated that rs13361189, rs4958847 and rs10065172 were considered to be risk factors of CD in Europeans but not of UC. In addition, subgroups analysis suggested that this increased risk may be ethno-specific. Further studies in other ethnic groups (e.g. Asians and Africans) are needed to clarify possible roles of IRGM polymorphisms in CD or UC. To identify the exact role of IRGM polymorphisms in the pathogenesis of CD, more studies such as animal disease modeling are of great importance.