Insulin autoantibody polymorphisms with greater discrimination for diabetes in humans

Insulin autoantibodies, like islet cell antibodies, are found not only in the sera of newly diagnosed Type 1 (insulin-dependent) diabetic patients and their relatives, but also in patients with other autoimmunities who do not develop diabetes. Insulin autoantibodies are oligo/monoclonal and frequently binding-site restricted. As determinant selection is genetically determined, we questioned whether certain polymorphisms of insulin autoantibodies, identified by their binding site on the insulin molecule, could better discriminate for Type 1 diabetes, which is also HLA determined. First, we raised monoclonal antibodies to human insulin by classic fusion methods in order to determine the range of antibody polymorphism, and identified five distinct types by their binding profiles to a panel of insulin variants, using an enzyme-linked immunosorbent assay. Two of these polymorphisms, type A and type B, were subsequently found in insulin autoantibody positive human sera using the same panel of insulin variants, and successfully distinguished diabetes-related from diabetes-unrelated individuals. Thus, the type B polymorphism was responsible for binding in 60% of 41 insulin autoantibody positive individuals with polyautoimmune disease but no personal or family history of diabetes (diabetes unrelated), but in only 2% of a group which comprised 17 newly-diagnosed insulin autoantibody positive Type 1 diabetic patients, 19 insulin autoantibody positive discordant twins of Type 1 diabetes and six insulin autoantibody positive healthy siblings of Type 1 diabetic patients (diabetes related) (p<0.01). Isolation of the type A polymorphism alone reduced the proportion of false negatives in the insulin autoantibody test for diabetes relatedness from 49% to 20% without diminishing its specificity. Thus, insulin autoantibody polymorphisms are more discriminating than the ‘nominal’ antibody, due possibly to linkage between immune response genes determining response to the type A epitope on the one hand, and susceptibility to Type 1 diabetes on the other.