Anti-BSA antibodies are a major cause of non-specific binding in insulin autoantibody radiobinding assays

Abstract

Insulin autoantibodies (IAA) are usually the first risk-markers detected during the type 1 diabetes prodrome, but precise measurement is difficult as insulin binding is often low. Non-specific binding (NSB) of (125)I-labelled insulin necessitates competitive displacement with unlabelled insulin to demonstrate specificity. NSB varies with different batches of label, suggesting that it is caused by impurities in the label. Addition of bovine serum albumin (BSA) can reduce NSB, so we investigated whether BSA antibodies cause lack of specificity in IAA assays. Samples from patients with newly-diagnosed type 1 diabetes, healthy schoolchildren previously found to have raised (125)I-insulin binding (≥ 0.4 units) and IAA-negative schoolchildren were re-assayed for IAA by radiobinding microassay using commercial (125)I-insulin with and without 1g/dl BSA added to the buffer. Of 100 patients, 68 were IAA-positive on re-assay with BSA compared to 72 without BSA (p=0.125). Of 154 schoolchildren who previously had raised (125)I-insulin binding, only 45 had (125)I-insulin binding ≥ 0.4 units on re-assay with BSA compared to 90 without BSA (p<0.001). Following competitive displacement with unlabelled insulin, 40 were IAA-positive with BSA compared to 48 without BSA (p=0.02). No IAA-negative schoolchildren were IAA-positive on re-assay. Levels of NSB were associated with antibodies binding (125)I-BSA and purification of labelled insulin reduced NSB. Addition of BSA to assay buffer improves the screening efficiency of the IAA assay without reducing disease sensitivity in patients. High titre BSA antibodies interfere with IAA measurement because of (125)I-BSA present in some insulin labels. Improved purification of insulin labels should obviate the need for competitive displacement.