Disease-associated autoantibodies and HLA-DQB1 genotypes in children with newly diagnosed insulin-dependent diabetes mellitus (IDDM). The Childhood Diabetes in Finland Study Group.

The possible relation between HLA-DQ genotypes and both frequencies and levels of autoantibodies associated with IDDM was assessed by examining HLA-DQB1 alleles and antibodies to islet cells (ICA), insulin (IAA), glutamic acid decarboxylase (GADA) and the protein tyrosine phosphatase-related IA-2 molecule (IA-2A) in 631 newly diagnosed diabetic children under the age of 15 years. ICA were found in 530 children (84.0%), while close to half of the subjects (n = 307; 48.7%) tested positive for IAA. GADA were detected in 461 index cases (73.1%), with a higher frequency in those older than 10 years (78.9% versus 69.2% in the younger ones; P = 0.006). More than 85% of the children (n = 541; 85.7%) tested positive for IA-2A. Altogether there were only 11 children (1.7%) who had no detectable autoantibodies at diagnosis. There were no differences in the prevalence of ICA or GADA between four groups formed according to their HLA-DQB1 genotype (DQB1*0302/02, *0302/X (X = other than *02), *02/Y (Y = other than *0302) and other DQB1 genotypes). The children with the *0302/X genotype had a higher frequency of IA-2A and IAA than those carrying the *02/Y genotype (93.8% versus 67.3%, P < 0.001; and 49.0% versus 33.6%, P = 0.002, respectively). The children with the *02/Y genotype had the highest GADA levels (median 36.2 relative units (RU) versus 14.9 RU in those with *0302/X; P = 0.005). Serum levels of IA-2A and IAA were increased among subjects carrying the *0302/X genotype (median 76.1 RU versus 1.6 RU, P = 0.001; and 50 nU/ml versus 36 nU/ml, P = 0.004) compared with those positive for *02/Y. Only three out of 11 subjects homozygous for 02 (27.3%) tested positive for IA-2A, and they had particularly low IA-2A (median 0.23 RU versus 47.6 RU in the other subjects; P < 0.001). The distribution of HLA-DQB1 genotypes among autoantibody-negative children was similar to that in the other patients. These results show that DQB10302, the most important single IDDM susceptibility allele, is associated with a strong antibody response to IA-2 and insulin, while GAD-specific humoral autoimmunity is linked to the *02 allele, in common with a series of other autoimmune diseases as well as IDDM. We suggest that IA-2A may represent beta cell-specific autoimmunity, while GADA may represent a propensity to general autoimmunity.