Abstract

Genetic associations between type 1 diabetes and alleles at the HLA class II locus DPB1 have been previously reported. Observed associations could be due to variation in the DPB1 locus itself or to linkage disequilibrium (LD) between DPB1 alleles and other susceptibility loci. One measure of whether the association of an allele with a disease reflects a true effect of the locus or is simply due to LD is the observation of that association in multiple ethnic groups. Previous type 1 diabetes associations have been reported for DPB1*0301 and DPB1*0202 (predisposing) and for DPB1*0402 (protective). In this study, results are reported from testing these associations in three different sample sets: 1) Puerto Rican case and control subjects, 2) Mexican-American simplex families, and 3) high-risk (DR3/DR4) individuals with and without an affected relative. DPB1*0301 was associated in all three groups, even after accounting for LD with DRB1-DQB1. DPB1*0202 and DPB1*0402 were positively and negatively associated, respectively, in two of the three populations. These results suggest that the observed DPB1 associations, especially that of the DPB1*0301 allele, with type 1 diabetes are likely to be true associations. This supports the concept that multiple genes in the HLA region can affect type 1 diabetes susceptibility.