OP11 Expanded genome-wide association study of Inflammatory Bowel Disease identifies 174 novel loci and directly implicates new genes in disease susceptibility

Fachal, L.(1);the International IBD Genetics Consortium, O.B.O.(1);

(1)Wellcome Sanger Institute, Human Genetics, Hinxton, United Kingdom; on behalf of the International IBD Genetics Consortium

Background

Genome-wide association studies (GWASs) have identified 243 loci associated with inflammatory bowel disease (IBD). However, the mapping of additional disease loci and causal variants is still limited by sample size. Larger GWAS can provide further insights into causal biology.

Methods

We performed a GWAS meta-analysis of 33 cohorts, totalling 54,439 IBD patients (N=30,574 with Crohn’s disease (CD), 21,193 with Ulcerative Colitis (UC)) and 37,054 European controls. Genotype imputation was undertaken using the TOPMed diverse population panel and association tests were performed using REGENIE. These results were meta-analysed with summary statistics from 4 additional studies, a Danish cohort, the UK Biobank, deCODE, and FinnGen, totalling 73,030 IBD patients and 1 million controls, Fig1.

Results

We identified 174 novel genome-wide significant signals (32 associated with CD, 36 with UC, 106 with IBD, Fig2).

Of these, 79 are located >1Mb from any known GWAS loci. We also identified two new population specific genetic associations, Fig3.


Six new loci contain genes implicated in monogenic syndromes that include colitis: CARMIL2, DOCK8, G6PC3, HPS4, NCF1, PIK3CD. Several new loci alter the expression of nearby genes, suggesting that aberrant expression of these genes underpins the association. For instance, a new variant associated with decreased risk of IBD increases the expression of VSIR in colon Fig4.

VSIR knockout mice have increased expression of IL23 and develop chronic inflammation in multiple tissues. Fine-mapping analyses identified likely causal missense variants at three new loci. DOK2 (ORCD=1.3, 27p=2x10-12) encodes a protein expressed in macrophages and T cells. Loss of Dok2 in mice causes severe DSS-induced colitis with reduced IL17A and IL22 expression. The same variant has been recently associated with CD in a sequencing study. SHARPIN (ORCD=1.2, p=1x10-16) is part of the linear ubiquitin chain assembly complex that modulates activation of the NF-κB pathway. Loss-of-function (LOF) mutations in other proteins in the complex are associated with immunodeficiency and systemic autoinflammation. CARMIL2 (ORUC=1.2, p=1x10-10) is required for NF-κB signalling in both B and T cells. LOF mutations are associated with primary immunodeficiencies and paediatric forms of IBD.

Conclusion

The greatly expanded sample size in our latest GWAS meta-analysis has enabled the identification of low frequency variants with larger effects on IBD susceptibility than the more common variants typically identified by previous GWAS. The biological overlap between Mendelian and complex forms of IBD is demonstrated by the discovery of common non-coding variants associated with complex forms of IBD that dysregulate the function of a Mendelian IBD gene.