Article Text


DDF inflammatory bowel disease symposium
OC-010 Detailed analysis of ATG16L1 demonstrates gene-wide extent of association with crohn's disease susceptibility
  1. J Van Limbergen1,2,3,
  2. B Kabakchiev4,
  3. J Stempak4,
  4. P Schumm5,
  5. W Xu6,
  6. P Henderson1,
  7. A Griffiths2,
  8. S Girardin7,
  9. D Philpott3,
  10. M Silverberg4
  1. 1Child Life and Health, University of Edinburgh, Edinburgh, UK
  2. 2Division of Gastroenterology, Hepatology and Nutrition, Toronto sick kids, Toronto, Canada
  3. 3Department of Immunology, University of Toronto, Toronto, Canada
  4. 4Division of Gastroenterology, Mount Sinai Hospital, Toronto, Canada
  5. 5Department of Health Studies, University of Chicago, Chicago, USA
  6. 6Dalla Lana School of Public Health, University of Toronto, Toronto, Canada
  7. 7Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Canada


Introduction ATG16L1 has been implicated in the susceptibility to Crohn's disease (CD), notably the T300A (rs2241880, exon 9). In ATG16L1-deficient and hypomorphic mice, autophagy, Paneth-cell homeostasis and IL-1b secretion were dependent on ATG16L1 (Saitoh et al Nature 2008, Cadwell et al Cell 2010). In contrast, studies focusing on T300A have shown conflicting results (Kuballa et al PLoS ONE 2008, Fujita et al JBC 2009). The association at other susceptibility loci (eg, NOD2 and IL23R) consists of common and rare variants (Rivas et al Nat Genet 2011). Exon-sequencing studies have not demonstrated rare variants within the 18 exons of ATG16L1. Our aim was to analyse the ATG16L1 association signal, specifically focusing on common variants.

Methods 38 single nucleotide polymorphisms (SNPs) spanning the ATG16L1 gene were imputed for 1735 subjects (800 CD/935 controls) (Franke et al Nat Genet 2010). Single SNP and haplotype frequency (>1%, solid spine of Linkage disequilibrium (LD), D’>0.8) association and permutation (n=10 000) analyses were performed (Haploview). Logistic regression (using SNPs p<0.05 on permutation analysis) was performed (SPSS). Monte Carlo simulation (n=10 000, R) was used to assess the difference in the number of haplotype blocks, based on D'>0.8, between CD and controls.

Results Single SNP permutation analysis yielded association of 16 markers (p<0.001), from intron 1 (rs6752107) to the 3'UTR (rs1045100). Analysis of D' and r2 characteristics, showed 12 SNPs were in complete LD with rs2241880 (D' and r2 0.98−1). rs6758317 (r2=0.25, intron 2), rs3792106 (r2=0.76, intron 11), rs4663396 (r2=0.25, intron 12) and rs1045100 (r2=0.84) demonstrated strong association, independent of rs2241880. Regression analysis retained rs3792106, rs6754677 (intron 14) and rs1045100 (3'UTR) (p<0.05). A Monte Carlo simulation showed no significant difference in the number of haplotype blocks between CD and controls (p=1). A strong association with CD was shown for the haplotype block containing the rs2241880 and for the block containing rs1045100, both p<10−5.

Conclusion We demonstrated that the rs2241880 alone is not sufficient to explain the strong ATG16L1 association. Additional variants, independent of rs2241880, could implicate any of the coiled-coil domain, the WD domain and/or the 3' UTR, in CD susceptibility.

Competing interests None declared.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.