The effect of viral infections on insulin binding in vivo was evaluated by measuring the binding of 125I-insulin to several different tissues. We found that splenic leucocytes from mice infected with either the diabetogenic (D) or non-diabetogenic (B) variants of encephalomyocarditis virus, herpes simplex virus, or lactic dehydrogenase virus showed up to a 130% increase in insulin binding. As much as a 300% increase in the binding of 125I-insulin to splenic leucocytes was observed in mice given bacterial lipopolysaccharide. In neither virus-infected nor lipopolysaccharide-treated mice was there any substantial change in insulin receptors on thymocytes, liver membranes, or peripheral erythrocytes. Thus, the increased binding of insulin appears to be limited to leucocytes and does not appear to represent a generalized metabolic alteration. These experiments suggest that during infection, the binding of insulin to leucocytes, which is widely used to measure insulin receptors, may not always accurately reflect the insulin receptor status of other tissues.

Insulin receptors were extracted from human erythrocytes contained in 100 ml of blood with the nonionic detergent Triton X-100 with almost 100% yield. The solubilized receptor had binding characteristics similar to those of the intact cell. Using 125I-monoiodoinsulin as tracer and a computer-assisted statistical curve-fitting program, a cooperative model gave values of 1.7 X 10(9) M-1 for the Ke (affinity of the empty receptor) and of 1.6 X 10(8) M-1 for Kf (affinity of the filled receptor). Bovine desalanine-desasparagine insulin inhibited tracer binding with 3-6% the potency of porcine insulin. Serum (B-8) containing anti-insulin receptor antibodies inhibited binding by 70% at a dilution of 1:100. Receptor autophosphorylation reaction was studied by incubation of the Triton extract with [gamma-32P]ATP and Mn2+ in the presence or absence of insulin, and the receptor was identified by immunoprecipitation with anti-receptor antibodies and sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Porcine insulin stimulated 4- to 5-fold the incorporation of 32P in a protein of Mr = 95,000, corresponding to the beta-subunit of the insulin receptor. Phosphoamino acid analysis revealed phosphorylation of the tyrosyl residues exclusively. The dose-response curve for insulin stimulation was sigmoidal; some effect of insulin was observed at 1 ng/ml but maximal effect was observed at 10 micrograms/ml. Bovine desalanine-desasparagine insulin, a noncooperative analogue of insulin, was able to fully stimulate the phosphate incorporation, but the dose-response curve was shifted to the right and steeper, consistent with the intrinsic affinity of this analogue for the insulin receptor. When insulin binding was performed under the same conditions as the phosphorylation, half-maximal stimulation of phosphate incorporation occurred with 6-29% of the fractional occupancy of the receptor. These data suggest that the insulin receptor of the human erythrocyte, as in other cells, is a tyrosine-specific protein kinase. Coupling between the receptor occupancy and kinase activation is complex. Furthermore, sufficient quantities of receptor can be easily obtained from a single individual to study the binding and kinase properties of the receptor opening the opportunity to a wide field of applications in human pathology.

The chemical synthesis of two porcine insulin analogues is described. Leucine in position B17 of the native molecule was substituted by its D-enantiomer and by L-norleucine, respectively. Both B-chain derivatives were synthesized by fragment condensation and purified as di-S-sulphonates by gel filtration followed by ion exchange chromatography on SP-Sephadex at pH3. Combination with native sulphhydryl A-chain yielded [DLeuB17]insulin and [NleB17]insulin. Both insulin analogues were isolated by gel filtration followed by ion exchange chromatography on CM-cellulose at pH 4.0. Biological activities of the analogues were determined relative to native pork insulin: 1) glucose oxidation in rat epididymal adipocytes was 6% for [DLeuB17]insulin and 16% for [NleB17]insulin, 2) receptor-binding affinity tested with cultured human fibroblasts and with rat adipocytes was 3% for [DLeuB17]insulin and 26% for [NleB17]insulin, and 3) thymidine incorporation into DNA of human fibroblasts was 35% for [DLeuB17]insulin and 100% for [NleB17]insulin.

Insulin receptors and Type I insulinlike growth factor (IGF) receptors have a similar structure with a major binding subunit of Mr approximately 130,000 linked by disulfide bonds to other membrane proteins to form a Mr greater than 300,000 complex. Both insulin and Type I IGF receptors also interact with both insulin and IGF, although with different binding affinities. We used a panel of human and rabbit sera containing antibodies to insulin receptors to determine whether these sera also interact with Type I IGF receptors. Immunoglobulins from five of five human sera inhibited binding of 125I-insulin and 125I-IGF-I to insulin receptors and Type I IGF receptors in human placenta and human lymphocytes. The rank order of reactivity with both receptors was the same; two sera, however, appeared to be selectively less reactive with the Type I IGF receptor, especially in placenta. Sera from five of seven patients and from a rabbit immunized with purified insulin receptor effectively immunoprecipitated both placental insulin receptors and Type I IGF receptors. Of the remaining sera, one had only a low titer against the insulin receptor and did not immunoprecipitate the IGF receptor, whereas the second serum effectively immunoprecipitated cross-linked and surface-iodinated insulin receptors, but had negligible reactivity against the Type I IGF receptor. These results suggest that most antisera to the insulin receptor also contain antibodies to Type I IGF receptors. Whether both specificities are inherent in the same or different antibody molecules remains to be determined. These data support the hypothesis that the insulin and IGF-I receptors are separate but related molecules, although there remains a small possibility that both receptors are domains on the same protein.

Hepatoma cells in culture exhibit a range of differentiated functions. Well differentiated hepatoma cells retain most of the functions of the adult liver, whereas dedifferentiated cells have lost most of them. In the present study, insulin binding and insulin effects were studied in four differentiated and two dedifferentiated cell lines derived from the Reuber hepatoma and in a partially differentiated cell line, HTC, derived from the Morris 7288C hepatoma. Specific insulin binding was lower in dedifferentiated cells than in partially differentiated and differentiated ones. In all cell lines, analysis of insulin binding yielded linear Scatchard plots. Although some variations in affinity of insulin for its receptor were observed, most of the differences in binding were accounted for by differences in insulin receptor number. In differentiated hepatoma cells, tyrosine aminotransferase (TAT)-specific activity was easily detectable, and insulin produced a 2- to 3-fold increase in enzyme activity within 4-6 h. By contrast, TAT activity in the dedifferentiated cells was low and did not respond to insulin. In the partially differentiated hepatoma HTC, insulin stimulated TAT only after basal TAT activity was induced by glucocorticoid treatment. Glycogen synthase (I and D forms) activities were detectable in all cell lines. In both differentiated and dedifferentiated Reuber hepatoma cells, insulin increased the I-form of glycogen synthase within 15 min. This effect of insulin was observed at lower insulin concentrations than stimulation of TAT activity. By contrast, insulin at any concentration was totally ineffective in stimulating glycogen synthase in glucocorticoid-treated and untreated HTC cells. These results indicate that in hepatoma cells, 1) insulin receptor number and insulin effect on TAT are modulated by the degree of cell differentiation; 2) the pathways of insulin action on TAT and glycogen synthase diverge in some postreceptor step(s) which is under independent control in differentiation; and 3) receptor and postreceptor defects exist in hepatoma cell lines which may be useful in dissecting the pathways of insulin action.

The insulin receptor purified from human placenta by sequential affinity chromatography on wheat germ agglutinin- and insulin-Sepharose to near homogeneity retained tyrosine-specific protein kinase activity. This purified insulin receptor kinase specifically catalyzed the incorporation of 32P from [gamma-32P]ATP into not only the beta-subunit of the insulin receptor but also histone H2B, a synthetic peptide which is sequentially similar to the site of tyrosine phosphorylation in pp60src (a gene product of the Rous sarcoma virus) and antibodies to pp60src present in the sera obtained from three rabbits bearing tumors induced by the Rous sarcoma virus. In each case, phosphorylation occurred exclusively on tyrosine residues. Insulin stimulated phosphorylation of these substrates 3- to 5-fold. Kinetic analysis using the synthetic peptide indicated that insulin acted by increasing the Vmax of peptide phosphorylation from about 3.1 to 9.5 nmol X mg-1 of protein X min-1, whereas the value of the Km for the peptide, about 1.5 mM, was not significantly changed. This kinase acted weakly on casein, alpha-S-casein, actin, and a tyrosine-containing peptide analogue of a serine-containing peptide used commonly as a substrate for the cyclic AMP-dependent protein kinases. These data show that the insulin receptor kinase displays specificity toward exogenous substrates similar to the substrate specificity observed for pp60src and the protein kinase activity associated with the receptor for epidermal growth factor. The data suggest that the catalytic sites of these three tyrosine kinases are similar and that insulin activates its receptor kinase by increasing the Vmax.

Of all available liver cells in culture, only primary cultured hepatocytes are known to respond to glucagon in vitro. In the present study we investigated whether glucagon could stimulate amino acid transport and tyrosine aminotransferase (TAT;EC 2.6.1.5) activity (two well-characterized glucagon effects in the liver) in Fao cells, a highly differentiated rat hepatoma cell line. We found that glucagon had no effect on transport of alpha-aminoisobutyric acid (AIB; a non-metabolizable alanine analogue) nor on TAT activity, even though both activities could be fully induced by insulin [2-fold and 3-fold effects for AIB transport and TAT activity, respectively, after 6h; EC50 (median effective concentration) = 0.3 nM], or by dexamethasone (5-8-fold effects after 20 h; EC50 = 2 nM). Analysis of [125I]iodoglucagon binding revealed that Fao cells bind less than 1% as much glucagon as do hepatocytes, whereas insulin binding in Fao cells was 50% higher than in hepatocytes. The addition of dibutyryl cyclic AMP, which fully mimics the glucagon stimulation of both AIB transport and TAT activity in hepatocytes, induced TAT activity in Fao cells (a 2-fold effect at 0.1 mM-dibutyryl cyclic AMP) but had no effect on AIB transport. Cholera toxin stimulated TAT activity to the same extent as did dibutyryl cyclic AMP. These results indicate that the lack of glucagon responsiveness in cultured hepatoma cells results from both a receptor defect and, for amino acid transport, an additional post-receptor defect. Moreover, the results show that amino acid transport and TAT activity, which appeared to be co-induced by insulin or by dexamethasone in these cells, respond differently to cyclic AMP. This suggests that different mechanisms are involved in the induction of these activities by glucagon in liver.

Genes in the major histocompatibility complex of mice and guinea pigs control immunologic responsiveness to insulins from other animal species. In order to determine if similar genetic control exists in man, we have examined lymphocyte proliferation responses to components of therapeutic insulins by employing lymphocytes from diabetic patients that receive insulin. Distinct groups of individuals demonstrated positive lymphocyte proliferative responses to beef insulin, beef and pork insulin, beef proinsulin, pork proinsulin, and protamine. Lymphocytes from the patient population were typed for the HLA-A, B, C, and DR antigens. An increased frequency of certain HLA antigens was found in those individuals that responded to the following therapeutic insulin components: beef, HLA-DR4; beef and pork, HLA-DR3; beef proinsulin, HLA-BW4, CW2, CW5, DR2, and DR5; protamine, HLA-CW3, CW5, and DR7. The results demonstrate that the human immune system recognized the structural differences between human and beef and/or pork insulin. These differences are two amino acids in the A chain, alpha loop, of beef insulin and the single terminal amino acid, alanine, which is common to pork and beef insulins. Positive responses to both beef proinsulin and pork proinsulin demonstrated the capability of restricted recognition of more complex proteins represented by the C-peptide in these insulin preparations. Lymphocyte proliferative responses to protamine were also restricted, which suggests a genetic control to this antigen. The association of these responses with HLA alloantigens strongly suggests that genes within the human major histocompatibility complex control recognition and lymphocyte response to therapeutic insulin components.

The biosynthesis and carbohydrate processing of the insulin receptor were studied in cultured human lymphocytes by means of metabolic and cell surface labeling, immunoprecipitation with anti-receptor autoantibodies, and analysis on sodium dodecyl sulfate-polyacrylamide gels under reducing conditions. In addition to the two major subunits of Mr = 135,000 and Mr = 95,000, two higher molecular weight bands were detected of Mr = 210,000 and Mr = 190,000. The Mr = 210,000 band and the two major subunits were labeled by [3H]mannose, [3H]glucosamine, [3H]galactose, and [3H]fucose, and were bound by immobilized lentil, wheat germ, and ricin I lectins. On the other hand, the Mr = 190,000 band was labeled only by [3H]mannose and [3H]glucosamine and was bound only by lentil lectin. All four components could be labeled with [35S] methionine; however, in contrast with the other three polypeptides, the Mr = 190,000 band was not labeled by cell surface iodination with lactoperoxidase, suggesting that it is not exposed at the outer surface of the plasma membrane. Pulse-chase studies with [3H]mannose showed that the Mr = 190,000 was the earliest labeled component of the receptor; radioactivity in this band reached a maximum 1 h after the pulse, clearly preceded the appearance of the other components, and had a very brief half-life (t1/2 = 2.5 h). The Mr = 210,000, Mr = 135,000, and Mr = 95,000 bands were next in appearance and reached a maximum 6 h in the chase period. Monensin, an ionophore which interferes with maturation of some proteins, blocked both the disappearance of the Mr = 190,000 protein and the appearance of the Mr = 135,000 and Mr = 95,000 subunits. The mannose incorporated in the Mr = 190,000 component was fully sensitive to treatment with endoglycosidase H while that in the Mr = 210,000 band and the two major subunits was only partially sensitive. Tryptic fingerprints of the 125I-labeled Mr = 210,000 band suggested that this component contains peptides of both the Mr = 135,000 and Mr = 95,000 subunits. In conclusion, the Mr = 190,000 component appears to represent the high mannose precursor form of the insulin receptor that undergoes carbohydrate processing and proteolytic cleavage to generate the two major subunits. In addition, the Mr = 210,000 band is probably the fully glycosylated form of the precursor that escapes cleavage and is expressed in the plasma membrane.

The expression of insulin receptors and insulin action was studied in cell hybrids and cybrids produced by fusion of the BWIJ mouse hepatoma cell line with nucleated and enucleated mouse L-cells (LEA-2A) respectively. The BWIJ parent and the cybrids expressed high numbers of insulin receptors, whereas the hybrids resembled the L-cell parent with low numbers of receptors. Likewise, the hybrids resembled the LEA-2A cells with high levels of glycogen synthase, whereas the BWIJ cells and cybrids had much lower levels. Both parents, the cybrids, and the hybrids, expressed insulin stimulation of alpha-aminoisobutyric acid influx, but the dose-response curves indicated an increased insulin sensitivity in the cells with the higher receptor concentration. Insulin also stimulated 86Rb+ uptake in the hepatoma parent, hybrids and cybrids, but not in the L-cell parent. These data suggest that insulin receptors, like other hepatoma-specific properties, behave as a 'luxury function' of the hepatoma cell line and are extinguished when the hepatoma cell is fused with a less differentiated cell type. The biological activities associated with insulin action, on the other hand, are much more complex in their expression and probably the result of the interaction of multiple factors that vary in their expression in cell hybrids and cybrids.