Genetic association of the major histocompatibility complex with rheumatoid arthritis implicates two non-DRB1 loci

Objective The HLA–DRB1 locus within the major histocompatibility complex (MHC) at 6p21.3 has been identified as a susceptibility gene for rheumatoid arthritis (RA); however, there is increasing evidence of additional susceptibility genes in the MHC region. The aim of this study was to estimate their number and location. Methods A case–control study was performed involving 977 control subjects and 855 RA patients. The HLA–DRB1 locus was genotyped together with 2,360 single-nucleotide polymorphisms in the MHC region. Logistic regression was used to detect DRB1-independent effects. Results After adjusting for the effect of HLA–DRB1, 18 markers in 14 genes were strongly associated with RA (P < 10−4). Multivariate logistic regression analysis of these markers and DRB1 led to a model containing DRB1 plus the following 3 markers: rs4678, a nonsynonymous change in the VARS2L locus, ∼1.7 Mb telomeric of DRB1; rs2442728, upstream of HLA–B, ∼1.2 Mb telomeric of DRB1; and rs17499655, located in the 5′-untranslated region of DQA2, only 0.1 Mb centromeric of DRB1. In-depth investigation of the DQA2 association, however, suggested that it arose through cryptic linkage disequilibrium with an allele of DRB1. Two non–shared epitope alleles were also strongly associated with RA (P < 10−4): *0301 with anti– cyclic citrullinated peptide–negative RA and *0701 independently of autoantibody status. Conclusion These results confirm the polygenic contribution of the MHC to RA and implicate 2 additional non-DRB1 susceptibility loci. The role of the HLA–DQ locus in RA has been a subject of controversy, but in our data, it appears to be spurious.