The autoimmune response to insulin

Diaz, Jose-Luis (1988) The autoimmune response to insulin. University of Southampton, Doctoral Thesis.

Abstract

An enzyme linked immunosorbent assay (ELISA) was optimised and validated for the measurement of insulin antibodies and insulin autoantibodies. Using this assay, insulin autoantibodies were detected in 39% of newly diagnosed insulin dependent diabetes mellitus patients examined before treatment with therapeutic insulin, and in 47% of non-diabetic identical twins of diabetic patients. Insulin autoantibodies were also observed in 3.2% of non-diabetic patients referred to an autoimmune profile laboratory. Ninety eight percent of insulin autoantibodies in diabetes-related groups were cross-reactive, binding human, procine and bovine insulins. In contrast, 60% of the autoantibodies in the polyautoimmune group bound human insulin alone. The binding profiles of insulin autoantibody and insulin antibody containing sera were examined more fully using highly purified species variants and fragments of insulins in direct ELISA, and in indirect absorption experiments, using insulins covalently bound to Sepharose beads. The data suggested that the binding of the human-insulin-specific autoantibody sera was dependent on the presence of threonine at the B30 residue. The other autoimmune sera, and sera from insulin treated diabetics, cross-reacted with a wide panel of insulins, and appeared to bind predominantly to A chain or conformational determinants on the insulin molecule. Further direct and indirect binding studies were carried out with insulin analgoues substituted at the B30 residue. The human-insulin-specific sera bound to the B30 threonine insulin, but not the B30 variants other than the structurally closely related B30 serine and B30 valine insulins, albeit with much reduced affinity. Thus, in these sera, a small change in a single amino acid side chain at the B30 residue had a major effect on the binding affinity. Many of the insulin autoantibodies gave high binding signals in ELISA, but were barely detectable in radioimmunoassay, whereas the majority of insulin antibodies wer readily detectable in both assays. The effect of the iodination site on binding of radiolabelled insulin by antibodies and autoantibodies was examined using human and porcine insulins each mono-iodinated at one of the four tyrosine residues (A14, A19, B16, B26). The influence of iodination site on the binding of labelled insulin appeared to be dependent on the proximity of the labelled tyrosine to the antibody binding site and the clonal diversity, or epitope restriction, of the binding antibodies in the test serum. Insulin autoantibodies were also found in a high proportion of low risk and high risk BB rats, both diabetic and non-diabetic, suggesting that insulin autoantibodies are not exclusively associated with the expression of overt diabetes. An increase in both frequency and titre of the autoantibodies was observed with age in these animals. However, autoantibody frequency and levels bore no relation to the degree of β cell degranulation or lymphocytic infiltration of the pancreas, suggesting that insulin autoantibodies occur in the absence of insulitis or attendant β cell damage. Insulin autoantibodies appear to provide a marker solely for genetic susceptibility and not for underlying pancreatic damage. (DX85689)

This record has no associated files available for download.

Contributors

Download statistics