MICA is Associated with Type 1 Diabetes in the Belgian Population, Independent of HLA-DQ

To ascertain association of MICA with type 1 diabetes (T1D) in the Belgian population, well-characterized antibody-positive patients were analyzed for MICA transmembrane gene polymorphism in both an association study and a nuclear family study. The frequency of MICA5 was significantly increased in the T1D patient group (18%) compared with the control population (12%, OR = 1.6, pc < 10−3), whereas MICA9 was decreased (11% versus 16%, OR = 0.7, pc < 0.01). A p value <10−3 for the association of MICA conditional on HLA class II and p = 0.01 for the conditional extended transmission disequilibrium test were obtained, indicating that MICA is associated with type 1 diabetes, independent of HLA-DQ. Analysis of estimated extended HLA-DQ–MICA haplotypes revealed individual effects of MICA alleles. The most significant effect was seen for MICA5 on the HLA-DQA1*03–DQB1*0302–MICA haplotype (OR = 2.5, p < 10−3). A significant protective effect was seen for the combination of DQA1*01–DQB1*0602/3 and MICA5.1 (OR = 0.3, p < 10−3). However, patients stratified according to the presence or absence of the different MICA alleles did not differ in terms of age at onset, sex, or other diabetes-related clinical and epidemiological data. In conclusion, MICA is associated with type 1 diabetes in the Belgian population and the observed association does not result from the HLA-DQ associated risk.