Effects of Non-HLA genes on islet autoimmunity and type 1 diabetes in a population with high-risk HLA genotypes

Abstract

HLA-DRB1 and DQB1 are the major genetic determinants of type 1 diabetes (T1D) while non-HLA genes have much smaller effects. In children carrying at least one high-risk HLA allele, we analyzed the effects of four non-HLA T1D susceptibility polymorphisms. The Diabetes Autoimmunity Study in the Young (DAISY) population included 737 first-degree relatives and 1325 general population children identified as having at least one high-risk HLA haplotypes: DR3,DQB1∗0201 or DR4,DQB1∗0302. Of those, 109 developed islet autoimmunity (IA) and 50 progressed to T1D during the 8 year mean prospective follow-up. IA was defined as presence of >1 of the autoantibodies to insulin, GAD65, IA-2 or ZnT8 on at least 2 consecutive visits. Study participants were genotyped for rs689 (INS), rs2476601 (PTPN22), rs231775 (CTLA4), and rs11203203 (UBASH3A), previously confirmed as T1D risk polymorphisms. In Cox regression models, the number of susceptibility alleles in PTPN22, INS, CTLA-4, and UBASH3A (as a continuous variable ranging from 0-8 as there was no evidence for non-linearity) was evaluated as a predictor of IA or T1D. Table 1 reports hazard ratios (HR) for one additional non-HLA risk allele, stratified by HLA genotypes, adjusted for ethnicity, gender and family history of T1D. Interestingly, the effect of non-HLA risk alleles was strongest in children with the HLA-DR3/4 genotype. The interaction between the number of non-HLA risk alleles and HLA-DR,DQ genotypes was significant for IA (p=0.02) but not for T1D (p=0.35). For >6 risk alleles present in 2.4% of the population, HR was 4.6 for IA and 15.7 for T1D. The results did not differ by family history of T1D and did not change when limited to non-Hispanic whites (71% of the cohort). Additional risk conferred by non-HLA susceptibility alleles appears to be strongest in those subjects already carrying the highest-risk HLA-DR3/4 genotype. These findings require replication in different populations.(Table persented).