Cross-reactivity of three recombinant insulin analogs with five commercial insulin immunoassays

Abstract

Several new insulin analogs that are prepared with recombinant DNA technology are available for clinical use [for a recent review, see Ref. (1)]. These agents have altered pharmacokinetics compared with regular human insulin. Insulin aspart (NovoLogTM; Novo Nordisk Pharmaceuticals) is homologous with regular human insulin except for a single substitution of aspartic acid for proline at position B28. This single substitution reduces the molecule’s tendency to form hexamers. Therefore, insulin aspart is absorbed more rapidly after subcutaneous injection and has both a faster onset of action and a shorter duration of action than regular insulin. A second short-acting recombinant insulin is insulin lispro (Humalog®; Eli Lilly and Company), which is a human insulin analog created by reversing the amino acids at positions 28 (Pro to Lys) and 29 (Lys to Pro) of the B chain. It is absorbed more rapidly than regular human insulin when administered subcutaneously and has a shorter duration of action. A third recombinant insulin that is longer acting is insulin glargine (Lantus®; Aventis Pharmaceuticals), which differs from regular human insulin by the substitution of glycine for asparagine at position A21 and by the addition of two arginine residues to the COOH terminus of the B chain. These modifications lead to a slower, more prolonged absorption than regular insulin and a relatively stable concentration–time profile over 24 h. Insulin glargine is partially metabolized at the COOH terminus of the B chain in the subcutaneous depot to form two active metabolites, M1 and M2, with in vitro activity similar to that of insulin. Metabolite M1 is 21A-Gly-insulin, and metabolite M2 is 21A-Gly-des-30B-Thr-insulin One obvious question with these recombinant preparations is whether they are detectable by immunoassay. Immunoassays have been developed specifically for pharmacokinetic studies of insulin analogs. A sensitive RIA that is specific for insulin lispro and an ELISA that is specific for insulin aspart have been developed (2)(3). A chemiluminescent enzyme immunoassay that can quantify human insulin, proinsulin, despentapeptide insulin, porcine insulin, and insulin lispro with comparable cross-reactivity has also been described (4). Two insulin assays, one that can specifically quantify human insulin and a second with which human insulin, insulin aspart, and insulin lispro cross-react equally, have been used to estimate the concentration of insulin aspart or lispro by subtraction (5). All of these assays are used primarily for research and are not commonly used in clinical laboratories. Two reports have appeared in this journal describing the use of a commercial RIA for quantifying insulin lispro (6)(7). Only one study to date has examined the cross-reactivity of insulin lispro with an automated insulin assay on a multichannel, random-access analyzer (8). These authors found that the cross-reactivity of insulin lispro in the Elecsys insulin assay was <0.02%. The goal of this study was to quantify the cross-reactivity of insulin aspart, insulin glargine, and insulin lispro with several commercial insulin assays to determine which might be specific for human insulin and which might show cross-reactivity with the insulin analogs. Four of the five insulin assays were performed on automated, multichannel, random-access analyzers that are available in clinical laboratories. We obtained vials of each of the three insulin analogs, with a nominal concentration of 100 IU/mL and suitable for injection, from our hospital pharmacy. Each was diluted volumetrically with 60 g/L aqueous bovine serum albumin to final insulin concentrations of 30, 100, 300, and 1000 mIU/L. All dilutions of each insulin preparation were analyzed in duplicate, and the percentage cross-reactivity was calculated from the ratio of the measured and nominal concentration. Measurements were made on an Access analyzer (Beckman Coulter), an Advia Centaur analyzer (Bayer Diagnostics), an E170 analyzer (Roche Diagnostics), and an IMMULITE 2000 analyzer (Diagnostics Products Corporation), using the manufacturers’ reagents according to the instructions. Two lots of reagent for the IMMULITE 2000 were examined: lot 122, which is the old formulation, and lot 151, which is a new formulation, presumably with a change in one or both antibodies. In addition, a manual assay, Coat-A-Count (Diagnostic Products Corporation), was also used for testing. A summary of the recoveries observed for each combination of analog and assay method is presented in Table 1 . All analogs had an equivalent cross-reactivity of 80% with the Access method. This may represent cross-reactivity that is truly <100%, or assay calibration may not exactly match the nominal concentration of each insulin analog. Insulin lispro demonstrated 90% cross-reactivity on the Advia Centaur, whereas insulin aspart and glargine had mean cross-reactivities of 126% and 143%, respectively. Each of the analogs had 35–45% cro