HIGH-RISK HLA-DR DQ GENOTYPES FOR IDD CONFER SUSCEPTIBILITY TO AUTOANTIBODIES BUT DQB1-ASTERISK-0602 DOES NOT PREVENT THEM/

Although HLA class II genes are important in insulin-dependent diabete s (IDD), their influence on the expression of IDD-associated autoantib odies (aAb) is unclear. We compared HLA-DRB1 and DQB1 gene frequencies in several Caucasian groups: 191 normal controls, 378 IDD patients, a nd 357 non-diabetic relatives of which 250 had no aAb, 107 had at leas t one aAb (79 ICA(+), 31GAD(65)(+) and 49 IAA(+)), and 23 had both ICA (+) and IAA(+). We found that the frequencies of DR3/4 or DQB10201/03 02 heterozygotes were significantly higher in aAb(+) relatives compare d to aAb(-) relatives. The frequencies of DR4/4 or DR4/X (X=non 3 or 4 ) and DQB10302/X (X=0201 or 0302) in aAb(+) relatives were not differ ent from the aAb(-) relatives (which were enriched for these haplotype s), but were significantly higher than normal controls. The frequencie s of DR3/X or DQB10201/X were decreased in both aAb(+) relatives and IDD patients. Interestingly, the dominant IDD-protective DQB10602 all ele allowed the development of individual aAbs (10% of ICA(+) and 8% I AA(+) relatives had the allele), but was not observed in any high risk double aAb(+), or GAD(65)Ab(+) relatives. The latter finding was simi lar to that in our patients with IDD, in that only two of them (0.5%) had a DQB10602 allele. In conclusion, HLA-encoded susceptibilities to disease-relevant autoantibody production and IDD are concordant with the susceptibility alleles, but discordant for the protective DQB1060 2. Thus HLA genotyping for DQB10602 would impact on the selection of aAb(+) relatives for disease prevention trials.