An immunochemical assay of total extractable insulin in

For the precise assay of minute quantities of insulin in biological fluids, techniques using fibril formation (1) or paper chromatography (2), al-though specific for the insulin molecule, lack the required sensitivity. Methods based on the meas-urement in vivo of a fall in blood sugar in surgi-cally modified rats (3) afford sensitivity, but lack both specificity and precision. In zvitro biological assays, in which glucose metabolism is measured in rat diaphragm or epididymal fat, while perhaps more precise than in vi*o methods, are subject to interference by other hormones such as somato-tropin and epinephrine (4, 5), by unknown "in-sulin-like " substances whose extraction charac-teristics differ from those of insulin (6) and by numerous inhibitors of insulin activity (7-10). In contrast, an immunological method should be both chemically specific and, especially after application of an isotopic method for the im-munologic quantitation, reasonably sensitive and precise. Berson and co-workers have originally reported (11) that when insulin-'131 is added to an excess of antibodies, it rapidly becomes bound. Further-more, addition of unlabeled insulin in increasing quantities results in gradual saturation of the bind-ing sites, reflected by an increase in the percentage of unbound insulin-I"'. We have subsequently confirmed these observations (12). The change in the percentage of unbound labeled insulin forms the basis of methods recently described for the measurement of injected insulin in rabbit plasma (13) and extracted insulin from mouse pancreas (14). Both of these methods utilized hydrody-namic flow techniques for the determination of the percentage of unbound insulin. The modification of this approach for the assay of circulating lhumiian insulin is the subject of thi