Linkage Analysis between the Major Histocompatibility System and Insulin-dependent Diabetes in Families with Patients in Two Consecutive Generations

Abstract
We have histocompatibility (HLA) genotyped 28 families with insulin-dependent diabetics in two or more consecutive generations, usually parent and child. This strategy of ascertainment was used to maximize the likelihood of obtaining a homogeneous type of disease within a family, and an autosomal dominant mode of inheritance. 76 diabetics and 169 nondiabetics were studied in these families. The frequencies of the antigens Dw3 and Dw4, and the genotype Dw3/Dw4 among the diabetics are 59, 68, and 30%, respectively, as compared with 15, 12, and 2% in normal controls, and 43, 41, and 10% in the nondiabetic relatives of the diabetics. Dw2 is present in only one diabetic (4%), as compared with 18% in normal controls and 17% in nondiabetic relatives. HLA haplotype concordance was analyzed for sib pairs in relation to the haplotype shared by the affected parent/child pair, and for the diabetic sib pairs within each sibship. The results failed to reveal deviations in the expected HLA haplotype assortment. Assuming an autosomal dominant mode and several penetrance levels, linkage analysis between the HLA and diabetes was performed. The total lod score is 0.37 for a recombination fraction of 0.29 at 50% penetrance. Although the linkage and concordance analysis results are inconclusive, they seem to be different from those reported by us for families with normal parents and two or more diabetic sibs. Because ascertainment biases may have influenced these results in an unquantifiable manner, it is not certain whether the two types of families are genetically different. However, the marked difference in the lod scores for the 50% penetrant autosomal recessive model between the two types of families is compatible with a genetic dissimilarity between them. The high frequency of the Dw3 and Dw4 antigens, the Dw3/Dw4 genotype, and the decreased frequency of Dw2, however, indicate the existence of two or more important diabetic genetic factors associated with the D region of the HLA in these families.