Genetics, Natural History, and Pathophysiology of Behcet's Disease

National Institute Of Arthritis And Musculoskeletal And Skin Diseases

Investigators

Linked publications & trials

Abstract

In collaboration with Dr. Ahmet Gul from Istanbul University, we have now performed the largest genome-wide association study to date in search of additional variants contributing to Behcet's disease. In our study we genotyped patients with Behcet's disease and geographically matched controls from Turkey using a high-throughput SNP genotype bead array (Illumina, San Diego). After stringent quality control, the study included 1228 cases and 1303 controls, and 311,637 SNPs with call rates over 95%, overall P-values for deviation from Hardy-Weinberg equilibrium greater than 10 to the negative fifth power, and minor allele frequencies greater than 1%. After correcting for population stratification, we identified 305 SNPs with corrected P-values for association less than 10 to the negative fourth power. Of these, 248 were located on chromosome 6 within the MHC, with a minimum P-value = 1.14 x 10 to the negative 42. Regression analysis, correcting for the association due to HLA-B51, suggested at least one additional independent genetic factor located within the MHC region. Of the 57 non-MHC SNPs with P <10 to the negative fourth power, one, located in the 5 region of CPLX1, exhibited genome-wide significance (P = 5.92 x 10 to the negative eighth power;the threshold for genome-wide significance is P <1.6 x 10 to the negative seventh power). CPLX1 encodes complexin-1, which is known to be a regulator of exocytosis during synaptic vesicle membrane fusion. We have found that this gene is also highly expressed in leukocytes and may therefore play a role in exocytosis in leukocytes. Future experiments will address the role of the CPLX1 variants in Behcet's disease. The second most strongly disease-associated SNP (with P = 3.97 x 10 to the negative seventh power) was an intronic variant of the immunoregulatory cytokine, IL-10. Its minor allele frequency was 38.2% in cases, 31.3% in controls, and the allelic odds ratio was 1.36 (95% CI=1.21-1.53). We genotyped additional SNPs from the CPLX1 and IL10 gene regions in the same individuals and identified additional SNPs that supported both associations identified in the original genome-wide association study. In the IL10 gene region, another intronic SNP reached genome-wide significance (P = 5.74 x 10 to the negative eighth power). The IL10 disease-associated SNPs are part of a haplotype that includes promoter region variants that have been previously suggested to influence IL-10 gene expression. We genotyped healthy blood donors and measured IL-10 production from their stimulated monocytes. We found significantly lower IL-10 production in donors homozygous for the Behcet's disease-associated alleles compared with donors without the disease-associated alleles. These data suggest that a genetically encoded-deficient IL-10 response to inflammation contributes to Behcet's disease. We are currently preparing a manuscript describing the results of this genome-wide association study.