Population-based analysis of the HLA associated risk for IDDM

Sammanfattning: Insulin dependent diabetes (IDDM) is one of the most common chronic illnesses among childrenand young adults In Sweden. The disease has a complex pattern of inheritance and envlronmentalfactors are important. The strongest linkage to IDDM is found on chromosome 6p21 in the MajorHistocompatibility Complex (MHC) region. A strong linkage disequilibrium In the MHC havemade identiflcation of the alleles responsible for the primary association difQcult. The aim ofthis project was to Identify the primary assoclation of Human leukocyte antigens (HLA) to IDDMin Swedish population-based case-control studies and to analyse the interaction of these factorsage, gender and islet cell autoantibody levels. In addition the transmission of IDDM associatedHLA haplotypes were investigated in patient and control families. Patients and controls from three population-based case-control studles were HLA typed forDQA1, DQB1 and DRB. A comparison of allele frequencies between patients and controls showedthat two haplotypes DR3-DQA1'0501-DQB1'0201 and DR4-DQA1'0301-DQBl-0302, werepositively associated with IDDM. The genotype associated with the highest risk for IDDM wasDR3-DQA1'0501-DQB1'0201/DR4-DQAl'0301-DQB1'0302. All other positively associatedgenotypes contained either the DR3-DQA1'0501-DQB1'0201 or DR4-DQA1'0301-DQB1'0302haplotype. A total of 6 haplotypes showed a negative association: DR15-DQA1-0102-D9B1'0602,DR4-DQA 1'0301 -DQB1'0301, DR13-DQA1'0103-DQB1'0603, DR11-DQA1'0501-DQB1'0301, DR14-DQA1'0101-DQB1'0503 and DR7-DQA1'0201-DQB1'0303. The relative predisposing allele test,employed to asses if all negatively associated haplotypes were protective for IDDM, showed thatonly DR15-DQA1'0102-DQB1'0602 was protective in this analysis. This was further conflrmed bythe observation that all genotypes with a negative association contained DR15-DQA1'0102-DQB1'0602 or alleles on this haplotype. The 95% confidence intervals of the OR estimates of alleles occurring on the same haplotypeoverlapped extensively and could therefore not be used to distinguish which allele was responsilblefor the association. Stratification analysis and the predisposing allele test revealed thatDQB1'0302 on the DR4-DQA1'0301-DQB1'0302, DR3 on the DR3-DQA1'0501-DQB1'0201 andDQBl'0602 on the DR15-DQAl'0102-DQBl'0602 haplotype showed the strongest association toIDDM. Additional risk to that associated with DQB1'0302, was conferred by DRB1'0401 on theDR4-DQA1'0301-DQB1'0302 haplotype. The highest absolute risk (1/54) was observed for thecombination of DR3 and DQBl'0302. Kendalls rank order correlation demonstrated a correlation between the rank of ORassociated with different genotypes and age at onset. This correlation remained after correctingfor the correlation between age and islet cell antibodies (ICA) or glutamic acid decarboxylaseantibodies (GAD65Ab). The DR3-DQA1'0501-DQB1'0201 haplotype and the DR3-DQA1'0501-DQB1'0201/DR4-DQA1'0301-DQB1'0302 genotype showed a significantly lower frequency amongpatients diagnosed before the age of 7 compared with those diagnosed after 28. DR15-DQA1'0102-DQB1'0602 was not found among patients diagnosed before the age of 8 but showed a significanttrend in association with age. DR15-DQA1'0102-DQBl'0602 was more common among male thanfemale patients. Logistic regression modelling showed an interaction between DR3 and gender,DQB1'0602 and age and DQB1'0302 and season at onset while GAD65Ab showed an interactionwith season and age at onset. Hence not only age but also season at onset as well as gender affectthe risk for IDDM conferred by HLA. Offspring to diabetic fathers have a higher risk for developing IDDM than offspring todiabetic mothers. This has been suggested to be due to an increased transmission of IDDM riskHLA haplotypes from fathers to their offspring compared with mothers, Transmisslon rates wereanalysed in both IDDM and control families. In the control families there was no deviation fromexpected transmisslon rates. In IDDM families, DR4 haplotypes show an increased and DR2haplotypes a decreased transmission to patients. No signiflcant differences in transmission ratesfrom mothers and fathers were detected. The haplotype that the mother do not transmit to heroffspring, the non-inherited maternal haplotype (NIMH) showed a decrease in positivelyassociated haplotypes, while the non-inherited paternal haplotype (NIPH), did not differ comparedwith control haplotypes. Hence development of tolerance to antigens coded for on the NIMH mayaffect the risk of developing IDDM later in life. Swedish population-based case-control studies have identified HLA high-risk haplotypesand uncovered significant effects of age and gender. Studies of the 0-35 year olds show that theprotective effect of DQB1'0602 decreases with Increasing age at onset.Key words: Insulin-dependent diabetes mellitus, HLA-DR, HLA-DQAl, HLA-DQBl,predispositional allele test, absolute risk, relative predispositional effect, transmission rates,non-inherited maternal haplotypes.