Abstract
Insulin antibodies (IA) are detectable in the sera of most insulin-treated patients with diabetes mellitus. Antibodies to exogeneous insulin sometimes cause clinical symptoms of insulin resistance, allergy, and local lipoatrophy. Although the frequency of these complications has diminished with the use of highly purified porcine insulin or recombinant human insulin, there are some patients with high titer of IA. Autoantibodies to insulin (IAA) are also described. IAA has been reported to be in association with both insulin-dependent diabetes mellitus (IDDM) and polyendocrine autoimmune disease. For many years these antibodies have been measured by radiobinding assay (RBA) in which the complexes are precipitated non-specifically by polyethylene glycol. In the present study we developed a rapid and quantitative enzyme-linked immunosorbent assay (ELISA) method for measuring IA and IAA using recombinant human insulin antigen. We applied this method to the samples obtained from patients with diabetes mellitus and autoimmune thyroid disease and then compared the results with those obtained from the RBA method. The calibration curve for ELISA was derived from the dilution curve of a single serum from a patient positive for insulin antibody, and the results were expressed arbitrarily as ELISA UNIT. The calibration curve was approximately linear on the log-log scale within the range of 0.1-2.0 at optical density (OD)450nm, (6.25-200 ELISA UNIT). The intra-assay (CV=2.3-3.1%) and inter-assay (CV=2.8-7.2%) precisions were acceptable. Recovery rate varied from 74.5%to 118.5%and dilution experiments showed good linearity. Specificity was demonstrated by substituting purified human IgG for the test serum and glucagon for insulin. Except for hemoglobin, coexisting substances in serum had almost no effect on ELISA. The range of ELISA UNIT (Mean±SD) of 83 normal sera was 12.7±4.6. Positivity for IA by ELISA (>normal Mean+3SD) was 11 out of 58 (19.0%) and 26 out of 55 (47.3%) in patients with IDDM and with non-insulin-dependent diabetes mellitus (NIDDM) who were treated with insulin, respectively. Positivity for IAA by ELISA was 5 out of 173 (2.8%) and 1 out of 20 (5.0%) in patients with NIDDM without insulin therapy and hyperthyroidism due to Graves' disease, respectively. However, by RBA, we detected 4 other cases positive for IAA in NIDDM without insulin therapy and one case in Graves' disease. The present study demonstrates that the newly developed method of ELISA using recombinant human insulin antigen is clinically useful for measuring IA and IAA.
