Regulation of serum potassium during insulin-induced hypoglycemia.

Counterregulatory secretion of epinephrine occurs during severe insulin-induced hypoglycemia. Under these conditions (minimal plasma glucose 27.4 +/- 1 mg/dl) the decrease of serum potassium concentration (0.9 mVal/L) is mediated by two mechanisms: insulin-induced (0.48 mVal/L) and epinephrine-induced (0.42 mVal/L) cellular uptake of potassium. Epinephrine-induced serum potassium uptake appears to be more sensitive to beta-adrenoceptor blockade than glucose production. The intensification of insulin-induced hypokalemia by epinephrine is of clinical significance.

Insulin and anti-insulin antibody interaction. Evidence for the formation of 7 S and 10 S structures.

The clearing of monoclonal and polyclonal and anti-insulin antibodies from homogeneous solutions at 100,000 X g was used to estimate the size of soluble insulin-antibody complexes at physiologic concentrations. Monoclonal antibodies cleared as a uniform population of 6.6 S independent of the insulin concentration. Polyclonal antibodies cleared as 6.6 S monomers at saturation and as 10 S particles when the amount of insulin bound decreased, suggesting that a soluble complex with two antibodies was formed. An increase of the affinity and a decrease of antibody valency can be related to the complex formation. The binding affinity of polyclonal sera depends on the composition of the affinities of the IgG monomers and on their ability to form 10 S complexes. The formation of insulin-antibody dimers precludes cross-linking and precipitation. Both types of insulin-antibody complexes have been found in the sera from patients treated with bovine insulin.

Recognition of human insulin and proinsulin by monoclonal antibodies.

High-affinity monoclonal antibodies (MAB) were obtained from lymph node cell fusions. Affinities ranging from 0.8 X 10(9) L/M to 5.2 X 10(9) L/M were calculated from binding studies with monoiodinated human, bovine, and porcine insulins and human proinsulin. Two monoclonal antibodies were specific for human insulin, recognizing an epitope involving the amino acid B-30 (Thr). Another two monoclonal antibodies were bound to the C-terminal end of the B-chain near B-30. The B-chain-specific monoclonal antibodies did not bind human proinsulin. One monoclonal antibody recognized the A-chain loop in the positions A-8 to A-10. This antibody bound also to human proinsulin. It was concluded that the A-chain loop is exposed on the surface of proinsulin, while the C-terminal B-chain is not available for binding. The study shows that monoclonal antibodies can be used to characterize structures of insulin and proinsulin. In contrast to x-ray studies, the molecules can be used at low concentrations in soluble form. It is suggested to use monoclonal antibodies for the screening of atypical insulins in the serum of diabetic patients and for the further refinement of insulin and proinsulin measurements.

Insulin secretion in vitro and insulin binding to isolated hepatocytes in congenic mice with different H-2 complexes.

Congenic male mice with differences in the H-2 complex have been used to investigate insulin secretion in vitro, insulin binding to isolated hepatocytes, plasma glucose, and serum insulin. Plasma glucose and serum insulin did not show consistent differences in the B10.BR, B10.D2, B10.A, B10.G, B10.M, B10.S, C57/10SCSN, and C3H.OH strains. Isolated islets of Langerhans responded to stimulation with 400 mg/dl glucose with a 3-5-fold increase in insulin secretion rates (2P less than 0.01): B10.BR greater than B10.M greater than C57BL/10SCSN greater than B10.G greater than C3H.OH, B10.D2, B10.A, B10.S. The biphasic pattern of insulin secretion was less distinct in B10.M, B10.G, and C3H.OH mice. The high-affinity constants of insulin binding to isolated hepatocytes at 37 degrees C varied between 4.5 X 10(7) L X mol-1 and 4.5 X 10(8) L X mol-1 (2P less than 0.01): B10.A greater than B10.BR greater than C57BL/10SCSN, B10.S, B10.D2 greater than B10.M, B10.G. The glucose-stimulated insulin secretion from isolated islets of Langerhans and the binding of insulin to isolated hepatocytes correlate to the H-2 complex independently.

Maturation of immune response to insulin. Recruitment of new epitopes by antibodies.

Guinea pigs were immunized with regular insulin in Freund's adjuvant weekly (N = 6) or with NPH insulin daily (N = 10). The concentrations, the affinities, and the sizes of the insulin-immunoglobin G (IgG) antibody complexes were determined by ultracentrifugation. During the immune response to insulin, 7-S complexes were observed after 3 wk, and 10-S complexes were observed after 4 wk (2P less than .05 in each group). "7-S" antibodies were transferred into immunized guinea pigs before the formation of measurable antibody concentrations. The development of 10-S structures was enhanced in outbred Pirbright white guinea pigs (2P less than .05, N = 6) and in inbred strain II l.b.m. guinea pigs (2P less than .05, N = 7). Insulin may act as a monovalent antigen (7 S) after 3 wk and become bivalent after 4 wk (10 S). The change in valency is enhanced by antibody transfer. It reflects the formation of antibodies recognizing new epitopes on insulin.

Specificity of monoclonal anti-human insulin antibodies.

To define the specificity and epitope of five monoclonal antibodies (MoAbs) to human insulin, binding studies with artificially modified insulins and a number of native insulins were done. Epitopes on the A-chain (A4, A8-A10) and on the end of the B-chain (B30) could be identified. For two MoAbs, substructures of the amino acid B30 were found, which were essential for binding (hydroxyl and methyl groups of B30). In contrast to most antisera, MoAbs to human insulin show high specificity. However, as the study shows, the specificity is not absolute. With suitable artificial epitope modifications, cross-reaction can be seen. Two of the MoAbs used here show sufficient specificity to discriminate between insulin and proinsulin.

Receptor binding studies and clinical effects of human insulin (recombinant DNA): studies in patients with newly diagnosed type i diabetes, type II diabetes, insulin resistance (type A and type B), insulin antibodies, insulin allergy, and "brittle" diabetes.

Thirty-eight insulin-dependent diabetic subjects were treated for periods ranging from 1 to 14 mo with human insulin (recombinant DNA) in order to investigate the clinical effects of human insulin in comparison with pork insulin. Human insulin was well tolerated and no side effects were detected. The following differences between human and pork insulin were observed: reduced blood glucose oscillations associated with a reduction in hypoglycemic symptoms in patients with "brittle" diabetes and type I diabetes, decreased concentrations of antibodies against pork insulin related to a reduction of insulin requirement of approximately 15%, increased specific receptor binding in patients with type I diabetes and insulin resistance: type A, possibility for treating patients with pork insulin allergy, and an increased biologic activity in a patient with polyclonal antireceptor antibodies. No difference was detected between pork and human insulin treatment in patients with type II diabetes and in a patient with insulin resistance: type B with monoclonal antireceptor antibodies. Human insulin was used safely and successfully in the treatment of diabetic patients.

Less pronounced changes in serum potassium and epinephrine during hypoglycemia induced by human insulin (recombinant DNA).

Human insulin (recombinant DNA) and purified porcine insulin were compared in healthy men. Intravenous insulin tolerance tests showed identical effects on plasma glucose when a bolus of 0.1 U/kg was injected. Following human insulin, hypokalemia and epinephrine secretion were significantly less pronounced. The differences in serum potassium concentrations are caused by a lower epinephrine response to hypoglycemia induced by human insulin in comparison to purified porcine insulin.

Binding of biosynthetic human insulin to human antibodies and receptors.

The binding of biosynthetic human insulin (BHI) and pork insulin to anti-pork insulin antibodies was tested at an insulin concentration of 4 microunits/ml. Identical binding data were obtained. The binding of the two insulins to mononuclear lymphocytes was also identical. The data are compared to previous results obtained with synthetic human insulin. Previously, we investigated the in vitro properties of fully synthetic human insulin from Dr. Rittel, Ciba-Geigy, Basel. Receptor binding of this human insulin was identical to pork insulin. The binding properties of the new BHI from Lilly (Indianapolis) were studied with human antibodies and freshly isolated human monocytes.

Different counterregulatory responses to human insulin (recombinant DNA) and purified pork insulin.

The biologic effect of human insulin (recombinant DNA) and purified pork insulin (PPI) was compared during insulin-induced hypoglycemia at two intravenous dosages 0.075 and 0.1 U/kg body wt in healthy volunteers. Serum insulin concentrations and plasma glucose curves were identical. PPI induced a significantly (P less than 0.05) higher output of epinephrine, growth hormone, and cortisol at both doses. Less inhibition (P less than 0.05) of endogenous insulin secretion was observed for human insulin at 0.1 kg body wt. An elevated incidence of sweating during hypoglycemia was related to epinephrine secretion. The results indicate that homologous insulin produces in vivo effects which are different from those produced by heterologous insulin.

Evaluation of a solid-phase monoclonal antibody-based enzyme immunoassay for insulin in human serum.

A 'sandwich-type' enzyme immunoassay for the measurement of serum insulin is described in which a monoclonal antibody-alkaline phosphatase conjugate and an antibody-immobilized polystyrene solid phase are used. Serum samples of 50 microliters can be analyzed and the enzyme immunoassay is as sensitive as the conventional radioimmunoassay for insulin. The results obtained with ELISA correlate well with those of the radioimmunoassay (r = 0.9844) and the between-assay and within-assay coefficients of variation are less than 15% over the useful ranges of the assay (2-200 microIU/ml). The sensitivity is 2 microIU/ml and this can be increased by longer incubation times. The crossreaction with porcine insulin is 45%, with bovine insulin 30% and with human proinsulin 20%.

A time-resolved fluoroimmunoassay for human insulin based on two monoclonal antibodies.

The insulin time-resolved fluoroimmunoassay (TRFIA) described here is a 'sandwich type' immunoassay using two monoclonal antibodies recognizing different epitopes of insulin. One of the antibodies is fixed to a solid phase and the other is biotinylated. After sandwich formation the biotinylated antibody is detected by time-resolved fluorometry with the fluorophore europium which is coupled to streptavidin. Serum samples of 100 microliters can be analyzed within a concentration range of 4 microIU/ml up to 500 microIU/ml. The detection limit of this assay, 3.9 microIU/ml, is comparable to the conventional radioimmunoassay and the results overall correlate well with those of the radioimmunoassay (r = 0.99). There is no detectable crossreaction with human proinsulin even using a proinsulin concentration of 2000 microIU/ml. The assay is performed with the DELFIA system (Pharmacia).

The detection of single cells forming antibodies to defined epitopes on insulin.

A solid-phase immunoenzymatic technique was modified to permit the detection and enumeration of antibody-secreting cells recognizing the amino or carboxy terminal of the insulin B chain. The procedure involves coating well surfaces with avidin and binding insulins specifically labelled with biotin at the B1 or B30 residue. On day 15 after immunization with human insulin, 20 outbred NMRI mice had generated cells secreting anti-insulin antibodies. The recognition of B1- or B30-related epitopes differed between individuals, suggesting that there was genetic determination of the epitopes expressed early in the immune response. The method can be used to find strains of mice with a preferential immune response to defined areas on the surface of small peptide molecules. Such strains could then be used to produce specific monoclonal antibodies.

Gamma irradiation as a substitute for H2O2 in peroxidase-labelled enzyme immunoassays.

The ability of gamma irradiation to substitute for H2O2 in peroxidase-labelled enzyme immunoassays was investigated. Colouring of TMB (3,3',5,5'-tetramethylbenzidine), ABTS [2,2'-azino-di(3-ethylbenzthiazoline-6-sulphonic acid)], and OPD (o-phenylenediamine) was produced with doses of 10 and 20 Gy. Addition of superoxide dismutase enhanced TMB and ABTS, and inhibited OPD. The reaction was terminated when the irradiation was stopped. The use of denaturing compounds to stop the reaction was unnecessary and coated antigens could be re-used for a second immunoassay.

In vivo labelling of the spleen with a red-fluorescent cell dye.

A method for labelling mouse spleen cells in situ is described. Spleen vessels were clamped before intrasplenic injection of a red-fluorescent cell dye (PKH-26). The labelling rate was 11.8% of all cells and 13.9% of B lymphocytes 30 min after injection as determined by FACS. 3 days later, the proportions of labelled cells were reduced to 7.4% (P < 0.01) and to 10.7% (P < 0.05), respectively. Only 10% of cells detected by FACS could be detected by fluorescence microscopy. Labelled cells were not found in peripheral lymph nodes 30 min after spleen injection, but were present 3 days later (FACS: 2.8% of all cells and 5.4% of B lymphocytes, P < 0.05 each), indicating migration of stained cells. Neither adverse nor toxic effects of in situ staining were observed. Isolated stained B lymphocytes from spleens of naive mice and from lymph nodes after immunisation with insulin showed proper function when tested in an ELISA-spot assay. The labelling procedure was used to follow splenic B lymphocytes producing natural anti-insulin antibody. These cells were found to be recruited for the early immune response in lymph nodes immunised with insulin.

Is the reuse of needles for insulin injection systems associated with a higher risk of cutaneous complications?

Twenty diabetic patients participated in a study to assess if multiple use of needles for insulin injection systems (Pens) is safe under normal daily conditions. The previous mean duration of Pen therapy was 16.3 months. During this time, the 20 patients carried out altogether more than 33,000 injections without any sign of local infection despite needle reuse. Patients were told to use needles if possible for 1, 3, 6, 9, and 12 injections before bacteriological assessment. Bacteriological investigation of these needles showed no contamination, except with one needle used three times, which was colonized with coagulase negative Staphylococcus. In contrast, half of the needles' plastic ground points which touched the skin were contaminated. No signs of infection were observed at the injection sites throughout the study. We conclude that, based on the bacteriostatic effects of commercially formulated insulin and on the siliconisation of needles' surfaces, bacterial growth is sufficiently prevented. Therefore, we can recommend the reuse of pen needles as a simple, safe and cost-beneficial procedure.

Lack of beta 2-adrenoceptor induced long-acting effect on glucose tolerance in type 2 diabetic patients.

The long-acting effect of a 10-min pulse infusion of the beta 2-adrenergic agonist fenoterol on oral glucose tolerance tests in controls and in normotensive patients with type 2 diabetes mellitus on diet was compared. During an oral glucose load starting 2 h after fenoterol control persons showed hyperglycemia (area: 25,950 +/- 467 vs. 22,650 +/- 410, P less than 0.01), hyperinsulinemia (area: 13,980 +/- 1050 vs. 8160 +/- 405, P less than 0.02) and a pronounced fall of serum potassium (area: 775 +/- 26 vs. 748 +/- 25, P less than 0.02). The patient group showed no late response to fenoterol: plasma glucose (area: 51,000 +/- 382 vs. 51,300 +/- 413, n.s.), serum insulin (area: 7215 +/- 233 vs. 8280 +/- 410, n.s.), serum potassium (area: 748 +/- 26 vs. 750 +/- 24, n.s.). The data show that there is a defect of the beta 2-adrenergic long-acting effect on glucose metabolism and on insulin release in type 2 diabetes mellitus.

IgE antibodies to insulin and related peptides, a result of insulin treatment?

Anti-insulin immunoglobulin E (IgE) antibodies in sera of insulin-treated diabetic patients were characterized by affinity and concentration, and by binding to proinsulin and insulin A and B chains. The affinities of IgE to insulin scattered between less than 10(7) and 0.6 x 10(9) l/mol, the concentrations between 0.3 and 3.5 ng/ml. All positive sera of type 1 and few of type 2 patients recognized proinsulin, A chain, and B chain with comparable affinities and concentrations. Elevated concentrations of total serum IgE in 12 of the 16 anti-insulin IgE-positive sera from insulin-treated patients indicate that these patients are predisposed to allergies. The incidence of elevated levels of total serum IgE in type 1 (17.7%) and in type 2 (7.8%) diabetic patients did not differ from the general population. Reagins to insulin and related antigens were also observed in sera of non-diabetic allergic persons without previous contact with exogenous insulin. The natural occurrence of insulin-specific reagins makes the use of IgE as a marker of antigenicity of insulin questionable. Discrepancies between the insulin-positive radioallergosorbent test (RAST) or skin test and clinical manifestations of insulin allergy exist, because IgE antibodies with low affinities require high concentrations of insulin for binding. Such amounts of insulin may occur at the injection site, but not in the circulation.

Insulin abuse in long-standing IDDM.

Two patients with insulin-dependent diabetes mellitus (IDDM) for 42 and 46 years respectively were diagnosed to produce factitious hypoglycemia. They had several properties in common: abnormal personality, refusal to stop incipient hypoglycemia, ideal body weight, good hemoglobin A1c (HbA1c) values, non-specific changes in the EEG, and brain atrophy sparing the periventricular areas. The addiction forced the patients to continue their habit during hospital stay. The dose of surreptitiously injected regular insulin always produced serum insulin concentrations large enough to clarify the diagnosis.

Evidence for coronary spasm during flushing in the carcinoid syndrome.

In this report we present the history of a patient with symptomatic carcinoid syndrome. During flushing he suffered from variant angina. The observation of coronary spasm due to excess 5-hydroxytryptamine (5-HT) is discussed with regard to the discharge of 5-HT from clotting platelets in the coronary arteries.