Protamines are cationic fish chromosomal proteins that retard absorption of isophane (NPH) insulins. Protamines are also administered in large doses for heparin neutralization in cardiac procedures. This study used a rapid enzyme-linked immunosorbent assay to examine frequency of protamine antibodies in diabetic and control populations. Antigen specificity of the IgG binding to protamine-coated plates was verified by competitive inhibition with other protamines, histone, glucagon, thyroid-stimulating hormone, arginine, and lysine. All antibodies tested cross-reacted completely with all protamines. Only 4 of 18 had any cross-reactivity with histones. None cross-reacted with the other inhibitors. In population surveys, 122 (38%) of 319 NPH insulin-treated diabetic subjects, 3 (8%) of 39 diabetic subjects treated with protamine-free lente insulins, and 5 (2.5%) of 202 normal control subjects had protamine antibody. No correlation was found between insulin and protamine antibodies. Because more than one-third of insulin-treated diabetic subjects have circulating IgG specific for protamine, they are potentially at risk for acute immunologic or anaphylactoid reactions when protamine is administered for heparin neutralization.

No abstract

We recently reported that the peak effect and duration of action of regular insulin injected sc were prolonged in diabetic patients and were not related to the presence of insulin antibodies. The results suggested that the ambient level of plasma glucose might be an important factor in determining the pharmacokinetics of regular insulin. In the present study we used a glucose clamp technique, which minimizes interference by counterregulatory phenomena, to study the pharmacokinetics of regular insulin injected sc at 2 different blood glucose concentrations [276 +/- 7 (+/- SEM) and 130 +/- 5 mg/dl] in 10 insulin-dependent diabetic patients. The patient's blood glucose concentration was maintained constant by means of a variable rate iv infusion of 20% dextrose in water after sc injection of regular insulin (0.2 U/kg) in the deltoid region of the arm. The onset of insulin action occurred at similar times at both glucose concentrations (0.6 +/- 0.1 h at 276 mg/dl vs. 0.5 +/- 0.1 h at 130 mg/dl; P greater than 0.05). Peak insulin action (determined from the time of the maximal glucose infusion rate) was delayed in the studies done at 276 mg/dl (4.7 +/- 0.2 h) compared to that in studies done at mean glucose concentrations of 130 mg/dl (4.3 +/- 0.2 h; P less than 0.05). The duration of insulin action was also significantly prolonged in the studies done at the higher glucose concentrations (9.1 +/- 0.3 h at 276 mg/dl vs. 7.7 +/- 0.2 h at 130 mg/dl; P less than 0.01). These results confirm previous reports of prolonged insulin action in diabetic patients, especially in the presence of hyperglycemia.

We previously reported that in insulin-treated diabetic subjects the time course of action of regular insulin injected sc is different from that reported in standard textbooks. The present studies evaluated the role of insulin antibodies (Abs) in the altered pharmacokinetics of regular insulin by comparing the time course of insulin action in 10 patients receiving chronic insulin therapy and having insulin Abs with that in 15 previously untreated patients without detectable Abs. After an overnight fast, the patients were given an infusion of 5% dextrose in water at 100 ml/h. Regular insulin (15 U) was then injected sc in the deltoid region of the arm. The onset of action of sc insulin, as indicated by a 10% fall in serum glucose, was similar in both patient groups [1.9 +/- 0.1 (+/- SEM) hour in Ab-negative and 1.8 +/- 0.1 h in Ab-positive patients]. The peak effect of insulin action, as determined by the nadir of serum glucose, was 4.6 +/- 0.2 h in the previously untreated patients, not significantly different from the value in the diabetic patients with insulin Abs (5.2 +/- 0.4 h). The duration of action of insulin was also similar in both groups (14.7 +/- 0.7 vs. 14.4 +/- 1.0 h). No significant correlations were found between insulin Ab levels and any of these 3 parameters of insulin action. However, the peak effect and total duration of insulin action were significantly correlated with the baseline serum glucose levels. A possible role of insulin Abs was evaluated in these patients by repeating the studies over a 2-year period. During this time, the previously untreated patients were treated with highly purified pork insulin, to which they developed low titers of insulin Abs. The diabetic patients who had been chronically treated with insulin were changed from less purified insulin to highly purified pork insulin, and all had a significant reduction in their Ab titers. No changes in insulin pharmacokinetics were found in either group. These studies demonstrate that the prolonged action of sc injected regular insulin in diabetic patients is not related to the effect of circulating insulin Abs.

Antibodies to islet antigens are useful markers of the pathological process that results in destruction of beta cells in type I diabetes. The targets of these antibodies, however, are not well characterized and their role in the pathological process remains to be established. To better understand the range of antigens recognized by autoantibodies in type I diabetes mellitus, we carried out immunoblotting on protein extracts from human islet and rat insulinomas. Using this approach higher titer antibodies specific for islet antigens were detected in 20/28 sera from recent onset type I diabetics and were infrequent (4/28) in sera from age-matched controls. Sera from diabetics reacted with multiple antigens at titers from 1/100-1/4000 while control sera usually bound a single band at lower dilutions. Although antigens of Mr 52, 84, 116 and 150 kilodalton were recognized most frequently, no antigen was bound by more than 50% of diabetic sera. Some of these antigens enriched in the membrane fraction while others were not. The data demonstrate that a heterogenous group of islet antigens is recognized by autoantibodies present at the onset of type I diabetes when islet destruction is complete. These findings contrast previous reports using immunoprecipitation with undiluted sera that principally identify a 64 kDa islet antigen. Thus, the immunoblot technique detects a different set of reactivities than previously identified by immunoprecipitation. The pattern of multiple reactivities with both surface and cytoplasmic antigens suggest that many autoantibodies may be generated by beta cell destruction and some of these may amplify the ongoing immune response.