PW03-011 – New Behçet's loci and gene-gene interactions

Abstract

Introduction: We previously identified disease-associated common variants in IL10 and IL23R, as well as HLA-B*51, in a Behcet's disease (BD) genome-wide association study (GWAS) performed with 311,459 SNPs in 1,215 cases and 1,278 controls from Turkey, but the disease-associated variants in these genes do not fully account for the estimated genetic contribution to disease risk. Objectives: To discover novel common BD susceptibility variants and to evaluate disease loci for evidence of gene-gene interactions. Methods: We used the Turkish collection GWAS genotypes to impute genotypes of 779,000 markers in the GWAS subjects and then evaluated the imputed markers for disease association. We also searched for new disease associated loci by analyzing patients with uveitis and by specifying different genetic models. We replicated the new BD loci in additional Turkish samples (838 cases, 630 controls) and if polymorphic, in Japanese samples (612 cases, 740 controls). Gene-gene interactions were evaluated by testing the significance of a multiplicative interaction term in a logistic regression model. Results: Imputation implicated three new BD susceptibility loci (CCR1, STAT4, and KLRC4). Validation, fine-mapping, and replication confirmed these associations and meta-analyses identified variants with genome-wide significance (p<5x10-8) in each. The variants in CCR1, CC-chemokine receptor 1, and STAT4, signal transducer and activator of transcription 4, were associated with gene expression differences. PBMCs with the diseaseassociated CCR1 variant exhibited reduced migration to the CCR1 ligand, MIP1a. Two disease-associated variants in KLRC4, which encodes an NK receptor family member, encoded missense changes (I29S and N104S). The BD-associated HLA-B*51 haplotype includes MICA, an NK receptor ligand. A statistically significant interaction (p=0.03) was identified between HLAB* 51 (presumably tagging MICA variation) and KLRC4 N104S. Analysis of BD patients with uveitis identified two non-synonymous variants (D575N and R725Q) in ERAP1 that recessively conferred BD risk (p=4.7x10-10). ERAP1 is an endoplasmic reticulum-expressed aminopeptidase that trims peptides and loads them onto MHC Class I. We found strong evidence for an interaction between the BD-associated Class I allele HLA-B*51 and ERAP1 genotype (p=9x10-4). Conclusion: This study identified four new genetic loci (CCR1, STAT4, KLRC4, and ERAP1) and two gene-gene interactions (ERAP1 with HLA-B*51 and KLRC4 with HLA-B*51, presumably via its LD with MICA) that contribute to BD susceptibility. Shared genetic associations of MHC Class I, IL23R, and ERAP1, and the strong interactive effect of the disease-associated Class I allele and ERAP1 support an emerging concept that BD, ankylosing spondylitis, and psoriasis share pathogenic mechanisms.