Genetics of type 1 diabetes

Abstract

Listen

Translate article

The Review article by She on the relative contributions of HLA-DR and -DQ to type 1 diabetes (Ref. [1]), suffers in my opinion from the paucity of in-depth studies in the field. Whenever there have been sufficient data, HLA-DQ alleles have always shown a much stronger association with susceptibility or resistance to type 1 diabetes than DR alleles. Furthermore, except in one case (see below), DR alleles do not have an independent contribution either to susceptibility or resistance in this disease. Two very large studies are relevant in this regard: (1) part of the 11th International HLA Workshop, with 981 unrelated patients and 2228 healthy controls worldwide (reviewed in Ref. [2]); and (2) the Swedish childhood diabetes study, of 425 children consecutively diagnosed with type 1 diabetes (ascertainment rate >95%), and 367 unrelated controls[3]. Both studies show HLA-DQ as the true susceptibility locus, while any association with DR alleles arises from linkage disequilibrium of such alleles with the DQ alleles. The Swedish study went even further, in testing for the association with disease of the nine HLA-DR4 alleles found in the general Swedish population. Of these DR4 alleles, DRB1*0401 is most often linked to type 1 diabetes (62% of patients and 25% of controls with an odds ratio of 4.95). The odds ratio for the high risk DQ allele A1*0301/B1*0302 (DQ8) is 7.47, higher than that of DRB1*0401. Furthermore, the DRB1*0401 allele is linked to susceptibility when in linkage with the susceptible DQ8 allele (odds ratio of 7.69), and linked to resistance, when in linkage with the resistant allele DQA1*0301/B1*0301 (DQ7) (odds ratio 0.52). Other DR4 alleles fare even worse by this comparison. In addition, a test of strongest association as developed by Svejgaard and Ryder[4]was performed in this study, showing that indeed DQ is by far the strongest linkage locus and all linkages to DR are secondary and due to linkage disequilibrium alone. By contrast, a recent study has shown that whereas HLA-DQ is the primary susceptibility locus, the allele DRB1*0403 significantly protected from type 1 diabetes, even in the presence of the highest risk heterozygous genotype (A1*0301/B1*0302–A1*0501/B1*0201), by being present in 5/49 control subjects and 0/171 type 1 diabetics[5]. This difference between the two studies is certainly worthy of further investigation. Furthermore, the fact that certain DQ alleles susceptible or resistant to type 1 diabetes, are obtained by trans-complementation and correlate with susceptibility to type 1 diabetes better than DR alleles[6], can only mean one of two things: either DQAB is the true susceptibility gene for the disease; or two genes, one closely linked to DQA and another to DQB, constitute such a gene—a rather unlikely possibility.