Genetic and immunological characteristics of type I diabetes mellitus in an Indo-Aryan population

Abstract

Aims/hypothesis. Our aim was to characterise the genetic and immunological features associated with Type I (insulin-dependent) diabetes mellitus in a cohort of Indo-Aryan children resident in the United Kingdom. Methods. Children with Type I diabetes (n = 53), unaffected first-degree relatives (n = 146) and unrelated healthy control children (n = 54) were typed for alleles of the HLA-DRB1, HLA-DQA1 and HLA-DQB1 genes. Islet cell antibodies and antibodies to glutamic acid decarboxylase, protein tyrosine phosphatase-2 (IA-2ic) and insulin were measured in the diabetic and control children. Results. The DRB1*03.DQA1*05.DQB1*02 haplotype was positively associated with the disease, occurring in 78% of diabetic children compared with 22.6% of healthy children (p(c) < 2.4 x 10-5). In simplex families, this haplotype was transmitted more frequently to the diabetic children than to their unaffected siblings (p < 1 x 10-4). The DRB1*04.DQA1*03.DQB1*0302 haplotype was also transmitted preferentially to the diabetic probands (p < 0.025) but was not associated with disease in the case control study. Islet-related autoantibodies were detected in 89.6% of diabetic patients compared with 11.8% of control children (p < 1 x 10-6). Although protein tyrosine phosphatase-2 autoantibodies were detected more frequently among DRB1*04-positive diabetic patients compared with patients lacking this allele, the overall frequency of these autoantibodies was lower than observed in Europid diabetic subjects. This could reflect the absence of a disease association with DRB1*04 in the Indo-Aryan cohort. Conclusion/interpretation. Type I diabetes in our Indo-Aryan cohort is similar to the disease observed in Anglo-Europeans but has important immunogenetic differences. The low frequency of protein tyrosine phosphatase- 2 autoantibodies among the Indo-Aryan diabetic children could have important implications for the design of future strategies for disease prediction in this population.