Search

The mechanisms by which elevated plasma nonesterified fatty acid (NEFA) levels induce skeletal muscle insulin resistance remain unclear. A NEFA-induced defect in the activation of PI3K, which plays a key role in insulin's stimulation of glucose transport, has been invoked. We sought to examine the effects of elevated plasma NEFA (approximately 1 mmol/liter) on muscle PI3K activity, insulin receptor substrate (IRS)-1 (important for activation of PI3K), and Akt, which is downstream of PI3K and activated by phosphorylation on serine and threonine in a PI3K-dependent manner. Ten normal men [age, 37 +/- 9 yr (mean +/- SD); body mass index, 25.2 +/- 3.8 kg/m(2)] underwent two 5-h hyperinsulinemic (80 mU/m(2) x min) euglycemic clamps with basal and end of clamp biopsies of the vastus lateralis muscle. Plasma NEFAs were increased in one study by infusion of 20% Intralipid (1 ml/min) and heparin (900 U/h) throughout and for 2.5 h beforehand. Skeletal muscle protein levels were quantified by Western blotting. Elevated plasma NEFA reduced whole-body insulin-stimulated glucose disposal by 24% (42.1 +/- 4.0 vs. 54.8 +/- 3.6 micromol/kg x min; P < 0.001). Basal muscle IRS-1 was the same in the two studies. IRS-1 levels decreased by 40% in the control glucose clamps (P < 0.005), but did not change during the Intralipid study. Total tyrosine phosphorylated IRS-1 increased by 29% during the control clamps (P < 0.05), but by only 18% (NS) during the Intralipid studies. Total levels of p85alpha subunit of PI3K and Akt were not influenced by plasma NEFA levels either in the basal state or during the glucose clamps. The insulin-induced increase in IRS-1-associated PI3K activity was impaired by elevated NEFA, so that activity at the end of the clamps with Intralipid was 35% lower than in the control clamps (P < 0.05). The percentage reduction in PI3K activation correlated with the reduction in insulin-stimulated glucose disappearance rate that was induced by elevated NEFA (r = 0.70; P < 0.05). Basal P-ser- and P-thr-Akt levels were very low and unaffected by NEFA levels. The glucose clamps resulted in a marked increase in P-ser and P-thr Akt levels. Despite the decrease in PI3K in the Intralipid study, no defect in Akt phosphorylation was found. In summary, NEFA-induced insulin resistance is associated with an impairment of IRS-1 tyrosine phosphorylation and IRS-1-associated PI3K activation. Down-regulation of IRS-1 levels is also impaired. The NEFA-induced defect in muscle glucose uptake appears to be a consequence of a defect in the insulin-signaling pathway leading to impaired PI3K activation. This in turn may lead to impaired glucose transport through an Akt-independent pathway because Akt phosphorylation was unaffected by elevated NEFA levels.

Cellular insulin resistance in polycystic ovary syndrome (PCOS) has been shown to involve a novel postbinding defect in insulin signal transduction. To find possible mechanisms for this defect, adipocytes were isolated from age- and weight-matched obese normal cycling (NC) and PCOS subjects. Insulin sensitivity for glucose transport stimulation was impaired in PCOS adipocytes (EC50 = 290 +/- 42 pmol/L) compared to that in NC cells (93 +/- 14; P < 0.005). The lipolytic responses to isoproterenol as well as maximal suppression by insulin were similar in NC and PCOS adipocytes. However, PCOS cells were less sensitive to the antilipolytic effect of insulin (EC50 = 115 +/- 33 pmol/L) compared to NC cells (42 +/- 8; P < 0.01). Treatment of adipocytes from NC subjects with the adenosine receptor agonist N6-phenylisopropyl adenosine had no effect on either insulin responsiveness or sensitivity for glucose transport stimulation. However, N6-phenylisopropyl adenosine treatment was able to normalize insulin sensitivity in PCOS cells (EC50 = 285 +/- 47 vs. 70 +/- 15 pmol/L, before and after treatment; P < 0.05). In conclusion, our results suggest that insulin resistance in PCOS, as accessed in the adipocyte, occurs at an early step in insulin signaling that is common for glucose transport and lipolysis. In addition, this insulin resistance involves an impairment of the system by which adenosine acts to modulate insulin signal transduction.

To assess the role of insulin receptor (IR) tyrosine kinase in human insulin resistance, we examined the kinase activity of IR of skeletal muscle biopsies from eight lean and five obese nondiabetics and six obese subjects with noninsulin-dependent diabetes mellitus (NIDDM). Biopsies were taken during euglycemic clamps at insulin infusion rates of 0, 40, 120, and 1200 mU/m2.min. IRs were immobilized on insulin agarose beads, and autophosphorylation and histone 2B phosphorylation were measured. Phosphatase and protease inhibitors preserved the in vivo phosphorylation state of the IRs. Glucose disposal rates (GDR) were reduced according to insulin dose by 23-30% in the obese (P < 0.05) and 43-64% in the NIDDM subjects (P < 0.0005). IR autophosphorylation was increased up to 9-fold in controls and was reduced (P = 0.04) in NIDDM compared to obese subjects. Histone-2B kinase was increased up to 6-fold in controls and was reduced by 50% in NIDDM. Kinase values by both methods were similar in lean and obese controls. In vivo stimulation of kinase was well correlated to the increase in GDR, as was the decrement in kinase in NIDDM to the decrement in GDR. These results suggest that defects in muscle IR kinase are significant in the in vivo insulin resistance of NIDDM, but not that of obesity.

We have characterized the insulin-dependent increase in glucose transport in human adipocytes using subcellular fractionation and antibodies specific for the two isoforms of the glucose transporter that are expressed in these cells. Plasma membranes isolated from untreated human fat cells contain the erythroid/GLUT1 isoform of the glucose transporter almost exclusively whereas the muscle-fat/GLUT4 transporter isoform is most abundant in intracellular microsomal membranes in resting cells. After exposure of adipocytes to insulin, the muscle-fat isoform is dramatically increased in the plasma membrane whereas the erythroid isoform barely changes in response to insulin. Thus, the total insulin-mediated increase in plasma membrane glucose transporters, confirmed by affinity labeling of both transporter isoforms, must be due to the increase in the muscle-fat/GLUT4 transporter. The two isoforms exist in different vesicle populations as shown by immunoadsorption of the muscle fat isoform-containing vesicles which are essentially devoid of the erythroid transporter. These data indicate that the insulin-mediated increases in glucose transport in human fat cells is a result of the translocation of vesicles uniquely containing the muscle-fat glucose transporter isoform.

The insulin receptor is expressed as two isoforms that differ by a 12-amino acid region at the carboxy-terminus of the alpha-subunit encoded by exon 11. These isoforms are produced by tissue-specific alternate splicing of the insulin receptor mRNA. To determine whether the relative expression of the isoforms is altered in skeletal muscle in two insulin-resistant states, NIDDM and obesity, relative mRNA levels were measured using a polymerase chain reaction technique. There were no differences in the relative amounts of skeletal muscle mRNA encoding the exon 11-containing form compared to the exon 11-lacking form of the insulin receptor among lean normal (30 +/- 2% Ex11-), obese nondiabetic (32 +/- 2%), and NIDDM (31 +/- 1%) subjects. We conclude that altered expression of insulin receptor isoform mRNAs does not account for skeletal muscle insulin resistance in NIDDM and obesity.

Insulin resistance is a predominant feature in women with polycystic ovarian syndrome (PCO). The cellular mechanisms for this insulin resistance have not been defined. In this study, major steps in the insulin action cascade, receptor binding, kinase activity, and glucose transport activity were evaluated in isolated adipocytes prepared from PCO subjects (n = 8) without acanthosis nigricans and in a group of age and weight-matched controls [normal cycling (NC) n = 8]. The PCO group was hyperinsulinemic and displayed elevated insulin responses to an iv glucose load. The binding of 125I-insulin to adipocytes was similar in cells from PCO and NC subjects. In PCO, autophosphorylation of the insulin receptor-subunit in the absence of insulin was normal but a significant decrease (30% below control) in maximal insulin stimulated autophosphorylation was observed. However, receptor kinase activity measured against the exogenous substrate poly glu:tyr (4:1) was normal. Cells from PCO subjects transported glucose at the same rate, in both the absence and presence of a maximal insulin concentration, as those from NC subjects. Strikingly, there was a large rightward shift in the insulin dose-response curve for transport stimulation in PCO cells (EC50 = 87 +/- 14 pmol in NC vs. 757 +/- 138 in PCO, P less than 0.0005); 8-fold greater insulin concentrations were required to attain comparable glucose transport rates in cells from PCO against NC. In conclusion, our results suggest that insulin resistance in PCO, as assessed in the adipocyte, is accompanied by normal function of insulin receptors, but involves a novel postreceptor defect in the insulin signal transduction chain between the receptor kinase and glucose transport.

We recently found that postprandial hyperinsulinemia does not compensate for the insulin resistance of obese subjects and proposed that postprandial hyperglycemia might be more important in promoting glucose disposal via the mass action effect of glucose. To test this idea we perform oral glucose tolerance tests (OGTT) in six lean and eight obese subjects, measuring glucose and insulin levels. Afterward two insulin infusion studies were performed. During infusion study I, insulin was infused in a dynamic square wave fashion to mimic the individual post-OGTT insulin levels at content euglycemic glucose levels. During study II, glucose and insulin infusions were varied to mimic post-OGTT levels in each subject. Overall glucose turnover was measured isotopically by infusion of [3-3H] glucose. During the OGTT the obese subjects exhibited significantly higher insulin (P less than 0.005) and glucose levels (P less than 0.002). Insulin-stimulated glucose disposal rates and total incremental glucose disposal (IGD) over 4 h during study I at euglycemia were significantly lower in obese compared to lean subjects (area under the curve, 824 +/- 166 vs. 1222 +/- 161 mmol/L.m2; P less than 0.01) despite higher post-OGTT insulin levels in obese subjects. When insulin plus glucose levels were matched to the individual OGTT levels, IGD was not significantly different between obese and control subjects (1712 +/- 253 vs. 1617 +/- 444 mmol/L.m2; P = NS). A significant inverse correlation (r = -0.73; P less than 0.05) existed between the degree of glucose intolerance (OGTT) and the decrease in IGD during the phasic hyperinsulinemic euglycemic study (infusion study I). These data suggest that with increasing insulin resistance, hyperinsulinemia is less effective in compensating for this decrease in insulin action, and hyperglycemia becomes more important in augmenting overall glucose disposal values.

Recent results from in vivo studies have shown that the kinetics of insulin action are impaired in lean and obese noninsulin-dependent diabetes mellitus (NIDDM) subjects as well as in obese nondiabetic subjects. We have measured the onset and loss of insulin action on glucose transport in adipocytes obtained from obese nondiabetic and obese NIDDM subjects to determine the contributions of obesity and diabetes to these cellular defects in insulin action. Basal and maximally insulin-stimulated rates of 3-O-methylglucose transport in adipocytes from obese and obese NIDDM subjects were reduced to 50% of the values in cells from normal subjects (P less than 0.05). The activation of glucose transport by insulin (4.3 nmol/L) was slower in cells from obese NIDDM patients. Half of the maximal insulin effect (A50) was reached by 23.0 +/- 5.0 min compared to 9.4 +/- 1.1 min in normal cells (P less than 0.05). Conversely, the deactivation of insulin-stimulated glucose transport upon removal of insulin was more rapid in adipocytes from the obese and obese NIDDM subjects. Half of the maximal insulin effect (D50) was lost by 12.4 +/- 1.7 min in obese NIDDM cells and by 8.9 +/- 1.9 min in obese subjects compared to 25.3 +/- 1.9 min in adipocytes from normal subjects (P less than 0.01). In conclusion, 1) basal and insulin-stimulated rates of glucose transport are similarly reduced in adipocytes from obese and obese NIDDM subjects; and 2) adipocytes from obese and obese NIDDM subjects display defects in the kinetics of insulin action, slower activation and accelerated deactivation, that mirror the defects measured in vivo. Both impairments in the kinetics of insulin action may contribute to the insulin resistance in these subject groups.

In an attempt to elucidate the cellular mechanisms of the insulin resistance associated with aging, insulin-mediated glucose transport was studied in isolated adipocytes obtained from 14 elderly subjects (mean age +/- SE, 69 +/- 2 yr) and 11 nonelderly (40 +/- 4 yr) subjects using the nonmetabolized glucose analogue 3-O-methylglucose. In elderly subjects with normal oral glucose tolerance tests, maximal insulin-stimulated glucose transport was reduced compared to nonelderly control levels [1.26 +/- 0.17 (+/- SE) vs. 1.96 +/- 0.26 pmol/2 X 10(5) cells X 15 sec; P less than 0.025). Elderly subjects with nondiagnostic oral glucose tolerance tests had more marked reductions in basal, submaximal, and maximal stimulated rates of glucose transport compared to the nonelderly group. The elderly group with the greatest decrease in in vitro glucose transport also had the greatest decrease in in vivo insulin-stimulated glucose disposal, whereas the elderly subjects with the mildest in vitro defect also had the smallest reduction in in vivo glucose disposal. A significant positive relationship existed between the maximal rate of in vitro glucose transport and the maximal rate of in vivo glucose disposal in the various subjects (Spearman rank correlation coefficient = 0.54; P less than 0.05). We conclude that aging is associated with a significant defect in the glucose transport system in isolated adipocytes, which may account for the post-receptor defect in in vivo insulin action in aging.

Isolated circulating mononuclear leukocyte preparations were obtained from 19 normal and 11 adult diabetic patients. No differences in monocyte, T cell, or B cell content were found when preparations from the two study groups were compared. The amount of insulin bound to cells from the diabetic patients was found to be decreased regardless of the cell type to which the data were normalized. This indicates that the previously reported decrease in the ability of mononuclear cells of diabetic patients to bind insulin was not a result of differences in mononuclear cell subpopulation.

A 14-yr-old woman presented with fasting hyperglycemia (269 mg/dl), fasting hyperinsulinemia (45 microU/ml), acanthosis nigricans, and insulin resistance. The patient's circulating insulin was normal by physical and biological criteria, and insulin receptor antibodies were not detected. Both the patient's in vivo dose-response curve for insulin-stimulated glucose transport in isolated adipocytes were shifted to the right and showed marked decreases in the maximal insulin response. Basal hepatic glucose output was significantly increased, and the in vivo dose-response curve for insulin-mediated suppression of basal hepatic glucose output was shifted to the right. Insulin binding to the patient's erythrocytes, monocytes, and adipocytes was markedly decreased. To confirm that the severe reduction in cellular insulin receptors was a primary rather than an acquired defect, similar studies were conducted using cultured fibroblasts. No detectable binding of insulin to these cells was observed. Further studies showed that the patient's mother and two sisters were hyperinsulinemic and insulin resistant, and had comparable, although less severe, changes in insulin binding. The patient was also demonstrated to have an insulin secretory defect both to both oral and iv glucose challenges. We thus conclude that this family demonstrates a genetic deficiency of insulin receptors, resulting in insulin resistance and, in this patient, severe diabetes mellitus.

Recent studies from our laboratory have demonstrated that human biosynthetic proinsulin has a much slower MCR than insulin as well as a greater effect on hepatic than on peripheral tissues. Since both of these features make proinsulin potentially useful as an adjunct to insulin, the present study was undertaken to characterize the biological effects of combined infusions of insulin and proinsulin. As an initial study, multiple euglycemic clamp studies were performed in six normal subjects to construct dose-response curves for insulin- and proinsulin-mediated glucose disposal. Insulin infusion rates of 0.63, 1.67, 5, and 10 micrograms/M2 X min were compared to proinsulin infusion rates of 2.75, 7.5, 22.5, and 45 micrograms/M2 X min. Analysis of these dose-response curves indicated that proinsulin-mediated glucose disposal was approximately 7% that of insulin. Each subject also received a combined infusion of insulin (0.63 micrograms/M2 X min) and proinsulin (2.75 micrograms/M2 X min). The mean (+/- SE) glucose disposal rate obtained during the combination study (319 +/- 21 mg/M2 X min) was similar to that predicted by adding the results from the individual hormone infusions (307 +/- 21 mg/M2 X min). Analysis of the dose-response curves for insulin and proinsulin was also useful in predicting the glucose disposal rate during the combination study (290 +/- 19 mg/M2 X min). Thus, using either approach to predict the glucose disposal rates, insulin and proinsulin appeared to have additive effects on total body glucose disposal.

With the recent availability of biosynthetic human proinsulin there has been a renewed interest in evaluating its metabolic effects, either alone or in combination with insulin. It has been suggested that pretreatment with proinsulin enhances the hypoglycemic response to subsequently administered insulin. On the other hand, the simultaneous administration of proinsulin and insulin has additive, not synergistic, effects. To clarify this question we used the euglycemic glucose clamp technique in 10 normal subjects to compare the steady state effects on glucose disposal of combined infusions of insulin (0.54 microgram/M2 . min, equivalent to 15 mU/M2 . min) and proinsulin (2.75 micrograms/M2 . min) given both simultaneously and sequentially. The mean +/- SEM steady state glucose disposal rates were similar whether the two hormones were given simultaneously (7.2 +/- 0.7 mg/min . kg), after proinsulin pretreatment (7.7 +/- 0.7 mg/min . kg), or after insulin pretreatment (7.1 +/- 0.7 mg/min . kg). The serum proinsulin concentration of 5.39 +/- 0.3 pmol/ml during the infusion of proinsulin alone was unchanged by the simultaneous infusion of insulin, suggesting that in the doses used, insulin did not affect proinsulin clearance. We conclude that in normal subjects there is no enhancement of the combined action of insulin and proinsulin to stimulate glucose disposal by pretreatment with proinsulin or insulin.

Duchenne muscular dystrophy (DMD) is an X-linked inherited neuromuscular disease characterized by progressive weakness and severe muscle wasting. Alterations in carbohydrate metabolism are often associated with neuromuscular disorders. We performed oral glucose tolerance tests and insulin binding studies on erythrocytes from 17 DMD and 8 normal males. Furthermore, we measured insulin binding to erythrocytes from 12 normal males and from 11 mothers and 10 sisters of affected males. As a group, DMD patients had mild glucose intolerance and both fasting and postabsorptive marked hyperinsulinemia (insulin resistance). Levels of glucose and insulin, expressed as incremental areas under their respective curves, were significantly elevated in the wheelchair-ridden patients. Incremental areas of glucose (0-2 h) and insulin (0-5 h) were 42 +/- 5 mg/dl X h (mean +/- SEM) and 96 +/- 18 microU/ml X h, respectively, in normal subjects and 71 +/- 6 (P less than 0.05) and 206 +/- 30 (P less than 0.05), respectively, in the wheelchair-confined DMD patients. All of the ambulatory DMD males had normal oral glucose tolerance tests. Insulin binding to erythrocytes was 20-30% lower (P less than 0.01) in all DMD patients than in normal males appropriately matched for age and degree of sexual development. This difference in binding was a result of lower affinity of the insulin receptor in DMD erythrocytes. On the other hand, insulin binding to fibroblasts was the same in normal males and DMD patients, suggesting that the abnormality of erythrocyte binding in DMD is probably not genetically induced. Insulin binding to erythrocytes and monocytes was the same in all females studied, regardless of whether they were carriers of the DMD gene. Our results suggest that abnormal insulin binding in DMD erythrocytes is an acquired rather than genetic abnormality, but insulin binding is not helpful in the identification of carrier females. The defect in insulin binding in DMD is present before the development of insulin resistance, which occurs only in severely immobilized patients. Thus, the cause of the insulin resistance in DMD may reside at steps beyond the binding of insulin to its receptor.

Although type I diabetic patients are clearly insulin deficient, it is unclear whether they have normal in vivo sensitivity to insulin. Recent studies which suggested that insulin resistance is a common feature of insulin-dependent diabetics have not taken into account their degree of metabolic control or the presence of circulating antibodies. In the present study, we performed multiple euglycemic glucose clamp studies to construct insulin dose-response curves in 5 well controlled and 5 poorly controlled type I diabetic patients and 21 age-matched normal subjects. Each study was performed on a separate day at insulin infusion rates of 15, 40, 120, 240, or 1200 mU/M2 X min. During the 40 and 120 mU/M2 X min infusions, steady state insulin levels of 96 +/- 8 (+/- SE) and 285 +/- 27 microU/ml respectively, were achieved within 25 min in normal subjects. In contrast, diabetic subjects did not achieve steady state insulin levels (62 +/- 8 and 212 +/- 16 microU/ml) until 90 min of infusion, and insulin antibodies were detectable in the serum of all these patients. The dose-response curve for insulin stimulation of glucose disposal in well controlled diabetic subjects was comparable to that in normal subjects, with half-maximally effective insulin levels of 84 microU/ml in the diabetic patients compared to 70 microU/ml in normal subjects and virtually identical maximal rates of glucose disposal (433 +/- 11 vs. 411 +/- 17 mg/M2 X min in controls). In contrast, the dose-response curve for poorly controlled diabetic subjects was significantly right-shifted (half-maximally effective insulin level, 112 microU/ml), with marked reduction in the maximal glucose disposal rate (324 +/- 25 vs. 411 +/- 17 mg/M2 X min in normal subjects). Basal hepatic glucose output was increased in both poorly controlled and well controlled type I diabetic patients (132 +/- 7 and 101 +/- 16 mg/M2 X min, respectively) compared to normal subjects (76 +/- 7 mg/M2 X min). However, during each insulin infusion, hepatic glucose output was virtually 100% suppressed in all 3 groups.(ABSTRACT TRUNCATED AT 400 WORDS)

We studied the dose-response characteristics of insulin's ability to modulate its own secretion in normal and type II diabetic (NIDDM) subjects by measuring suppression of serum C-peptide levels during insulin infusions with the plasma glucose level held constant. In normal subjects at euglycemia, primed continuous insulin infusion rates of 15, 40, 120, and 240 mU/M2 X min acutely raised serum insulin to steady state levels of 37 +/- 2 (+/- SE), 96 +/- 6, 286 +/- 17, and 871 +/- 93 microU/ml, respectively. During each infusion, maximal suppression of C-peptide to 30% of basal levels occurred by 130 min. At the higher insulin levels (greater than or equal to 100 microU/ml), C-peptide levels fell rapidly, with an apparent t1/2 of 13 min, which approximates estimates for the t1/2 of circulating C-peptide in man. This is consistent with an immediate 70% inhibition of the basal rate of insulin secretion. At the lower insulin level (37 +/- 2 microU/ml), C-peptide levels fell to 30% of basal values less rapidly (apparent t1/2, 33 min), suggesting that 70% inhibition of basal insulin secretion rates was achieved more slowly. In NIDDM subjects, primed continuous insulin infusion rates of 15, 40, 120, and 1200 mU/M2 X min acutely raised serum insulin to steady state levels of 49 +/- 7, 93 +/- 11,364 +/- 31, and 10,003 +/- 988 microU/ml. During studies at basal hyperglycemia, only minimal C-peptide suppression was found, even at pharmacological insulin levels (10,003 +/- 988 microU/ml). However, if plasma glucose was allowed to fall during the insulin infusions, there was a rapid decrease in serum C-peptide to 30% of basal levels, analogous to that in normal subjects. Three weeks of intensive insulin therapy did not alter C-peptide suppression under conditions of hyperinsulinemia and falling plasma glucose. The following conclusions were reached. 1) In normal subjects, insulin (40-1000 microU/ml) inhibits its own secretion in a dose-responsive manner; more time is required to achieve maximal 70% suppression at the lower insulin level (40 microU/ml). 2) In NIDDM studied at basal hyperglycemia, insulin has minimal ability to suppress its own secretion. Thus, impaired feedback inhibition could contribute to basal hyperinsulinemia. 3) Under conditions of hyperinsulinemia and falling plasma glucose, insulin secretion is rapidly suppressed in NIDDM (analogous to that in normal subjects studied during euglycemia.

We have studied the effects of moderate dietary fat restriction on plasma triglyceride, cholesterol, glucose, and insulin response in 27 subjects. Compared with a control diet (45% fat, 40% carbohydrate [CHO], 15% protein) the low fat (higher CHO) diet (30% fat, 55% CHO, 15% protein) produced a 41% increase in fasting triglyceride level (155 +/- 17 to 219 +/- 23 mg%) with no change in fasting plasma cholesterol level. Furthermore, this increase in triglyceride levels; induced by the higher CHO content of the low fat diet, was seen in 26 out of 27 subjects. Postprandial triglyceride, glucose, and insulin levels were also higher on the low fat (higher CHO) diet. Since hypertriglyceridemia is a significant risk factor for the development of coronary heart disease, and since our data indicate that the moderate increase in dietary CHO associated with a low fat diet will elevate plasma triglyceride levels, we believe that more caution is necessary before recommending the wide-spread use of low fat diets for heart disease prevention.

The insulin resistance of type II diabetes mellitus is due to both receptor and postreceptor defects of in vivo insulin action, with the postreceptor defect being the predominant abnormality. Diminished glucose transport has been found in adipocytes from patients with type II diabetes, suggesting that decreased cellular glucose transport activity may be responsible in part for the in vivo postreceptor defect observed in these patients. Recent studies have shown that the in vivo postreceptor defect initially present in patients with Type II diabetes is significantly reversed by insulin therapy. For these reasons, we speculated that the defect in adipocyte glucose transport might also be corrected with exogenous insulin therapy. Therefore, we measured adipocyte 3-O-methylglucose transport in cells from five type II diabetic subjects before and after a 2-week period of intensive insulin treatment. Glycemic control was significantly improved by this regimen. The mean (+/- SE) fasting serum glucose level fell from 292 +/- 24 to 135 +/- 29 mg/100 ml (P less than 0.005), and the mean integrated glucose area under a 7-h meal tolerance test curve decreased from 171,212 +/- 20,403 to 72,408 +/- 9,292 mg/min . dl. The mean 3-O-methylglucose transport activity increased after treatment at all insulin concentrations studied, including basal (before, 0.18 +/- 0.05; after, 0.45 +/- 0.09 pmol/2 X 10(5) cells . 10 sec; P less than 0.005) and maximally effective (25 ng/ml) insulin concentrations (before, 0.50 +/- 0.14; after, 1.32 +/- 0.30 pmol/2 X 10(5) cells . 10 sec; P less than 0.025), although the mean maximal glucose transport activity was still 25% decreased compared to normal values, indicating that a residual in vitro postreceptor defect remained. These results corresponded well with the degree of reversal (75%) of the in vivo postreceptor defect, as assessed by the euglycemic glucose clamp technique. These studies demonstrated that the decrease in adipocyte glucose transport activity in type II diabetes is practically reversible by intensive insulin therapy. This closely corresponds to the reversal by insulin therapy of the postreceptor defect expressed in vivo and provides further evidence that a cellular cause of the postreceptor defect in type II diabetes is a decrease in glucose transport system activity in the major insulin target tissues.

We studied the ability of isolated adipocytes from normal and type II diabetic subjects to internalize and process [125I]insulin. Adipocytes were incubated with [125I]insulin at 16 or 37 C, and at various times total cell-associated, surface-bound, and intracellular insulin were quantitated using an acid-barbital extraction technique which quickly removes cell surface insulin, leaving behind the intracellular insulin. Insulin internalization was slow in normal adipocytes at 16 C, such that only 13% of total cell-associated insulin was intracellular after 2 h of incubation. In contrast, internalization was rapid at 37 C, such that the intracellular pool of insulin was near maximal by 30 min and accounted for approximately 40% of the total cell-associated insulin. Sephadex G-50 column chromatography of the intracellular insulin demonstrated that more than 95% of this pool coeluted with native insulin. In adipocytes from the diabetic subjects, approximately 45% of total cell-associated insulin was intracellular after 30 min of incubation at 37 C. After 60 min of incubation at 37 C, the percentages of total cell-associated and surface-bound insulin were significantly lower in adipocytes from diabetic compared to normal subjects [1.81 +/- 0.31% (+/- SEM) vs. 2.92 +/- 0.24% (P less than 0.05) and 0.97 +/- 0.14% vs. 1.72 +/- 0.15% (P less than 0.01), respectively]. The percentage of insulin in the intracellular compartment was also slightly lower in adipocytes from diabetic compared to normal subjects (0.84 +/- 0.19% vs. 1.20 +/- 0.16%; P greater than 0.05). The lysosomotropic agent chloroquine increased total cell-associated insulin, and this was due entirely to an increase in intracellular insulin. In adipocytes from normal subjects, chloroquine increased intracellular insulin by 32% at 30 min, by 89% at 60 min, by 140% at 90 min, and by 178% at 120 min. In comparison to the normal adipocytes, the chloroquine-mediated increase in intracellular insulin was lower in adipocytes from the diabetic subjects (-8.1% at 30 min, 37% at 60 min, 58% at 90 min, and 63% at 120 min; P less than 0.05 at all time points). These results indicate that insulin is rapidly internalized in human adipocytes at 37 C such that approximately half of total cell-associated insulin is intracellular the intracellular insulin is largely intact; and intracellular processing of insulin by a chloroquine-sensitive pathway(s) is impaired in adipocytes from type II diabetic subjects.

We have studied insulin binding to isolated erythrocytes from 28 nondiabetic, pregnant women and have compared these results to similar data obtained in nonpregnant, nondiabetic women. In the nonpregnant subjects insulin binding was 35% higher during the follicular stage of the menstrual cycle as compared to the luteal phase. In the pregnant women insulin binding was comparable to that seen in the nonpregnant women during the follicular phase but was higher than that seen during the luteal phase of the menstrual cycle. Insulin binding was measured in 5 pregnant women during the third trimester and in the postpartum period. In these subjects insulin binding to erythrocytes declined by 31% after delivery. These results demonstrate that erythrocyte insulin receptors are not decreased in normal human pregnancy despite the presence of hyperinsulinemia and insulin resistance. Providing erythrocyte insulin receptors are reflective of insulin receptors in other tissues in human pregnancy, these findings suggest that some factor exists in pregnancy that modulates insulin receptors in a positive manner and are consistent with the possibility that the insulin resistance in pregnancy is related to a postreceptor abnormality.

To assess the effects of very low caloric (VLC) diets on glucose homeostasis in noninsulin-dependent diabetes mellitus, 30 obese subjects with NIDDM were studied for 40 days while eating a 330 Cal/day diet, with a subgroup of 12 subjects further evaluated during 40 days of refeeding. All subjects successfully lost weight, with an average weight loss of 4.6 +/- 0.2 kg (+/- SEM) after 10 days, 7.1 +/- 0.3 kg after 20 days, and 10.5 +/- 0.4 kg after 40 days of VLC diet therapy. Thus, weight loss was steady and progressive throughout the diet period. In contrast, the majority (87%) of the reduction in mean fasting plasma glucose (FPG) levels (297 +/- 13 to 158 +/- 10 mg/dl; P less than 0.001) occurred after 10 days of VLC diet therapy, with a further reduction in glucose levels to 138 +/- 9 mg/dl on day 40. The FPG response measured after 10 days of VLC diet was unrelated to the degree of obesity, rate or extent of weight loss, or prevailing insulin levels, but did correlate significantly with the initial FPG level (r = 0.37; P less than 0.05) and duration of diabetes (r = 0.42; P less than 0.05). After discontinuation of the VLC diet and refeeding of an isocaloric (weight maintenance) diet in 12 subjects, a variable increase in the FPG occurred, with an average increase of 80% after 40 days of refeeding. However, the mean FPG level after 40 days of refeeding was still markedly lower than that before VLC diet therapy (254 +/- 20 vs. 167 +/- 14 mg/dl; P less than 0.02) despite withdrawal of antidiabetic medication in all subjects. The basal hepatic glucose output (HGO) fell rapidly from 149 +/- 13 to 81 +/- 5 mg/M2 X min (P less than 0.001) after 10 days of VLC diet and rose from 67 +/- 4 to 88 +/- 7 mg/M2 X min (P less than 0.001) after 10 days of refeeding. Basal HGO demonstrated a highly significant positive correlation with FPG levels (r = 0.89; P less than 0.001) before and during both VLC diet therapy and refeeding. A significant correlation was also found between the change in FPG level and the change in basal HGO (r = 0.84; P less than 0.001) during both VLC diet and refeeding. Compared to that before the VLC diet, glucose tolerance to mixed meals was markedly improved during the refeeding period, with no change in circulating insulin levels.(ABSTRACT TRUNCATED AT 400 WORDS)

In vivo glucose uptake (Rd) occurs via two mechanisms: 1) insulin-mediated glucose uptake (IMGU), which occurs in insulin-sensitive tissues, and 2) noninsulin-mediated glucose uptake (NIMGU), which occurs in both insulin-sensitive and insulin-insensitive tissues. Thus, in the postabsorptive (basal) state Rd = IMGU + NIMGU. To determine whether these two pathways for in vivo glucose disposal are regulated independently, we studied the effect of stress levels of epinephrine (EPI) on IMGU and NIMGU in seven normal men after an overnight fast. To study NIMGU, somatostatin (600 micrograms/hr) was infused to suppress endogenous insulin secretion, and glucose turnover was measured isotopically while the serum glucose level was clamped at about 200 mg/dL for 240 min. Separate studies were done during the infusion of saline or EPI (0.2 microgram/kg X min). The final 120 min of each study were used for data analysis. Under these conditions insulin action is absent and Rd = NIMGU. NIMGU was 210 +/- 15 (+/- SEM) and 200 +/- 17 mg/min during saline and EPI treatment, respectively (P = NS). Therefore, EPI has no ability to modulate NIMGU. To measure the effect of EPI on Rd, hyperglycemic (200 mg/dL) hyperinsulinemic clamp (30 mU/M2 X min) studies were performed during the infusion of saline and EPI. EPI decreased Rd by 46 +/- 6% (751 +/- 85 to 405 +/- 43 mg/min; P less than 0.01). When the effect of EPI on IMGU (Rd - NIMGU) was considered separately, the inhibitory effect of EPI was more potent, as indicated by a 61 +/- 12% decrease in IMGU. In conclusion, 1) EPI inhibits IMGU, but has no effect on NIMGU; 2) when NIMGU is taken into account, EPI has a more potent ability to inhibit IMGU than previously found; and 3) the systems responsible for NIMGU and IMGU are independently regulated.

We have studied the effects of insulin on glucose transport in cultured fibroblasts obtained from six normal subjects and six patients with type II or noninsulin-dependent diabetes mellitus. Initial rates of glucose transport were determined by measurement of 2-deoxy-D-glucose uptake. In both normal and diabetic fibroblasts, insulin stimulated the uptake of 2-deoxy-D-glucose in a dose-dependent manner. The maximal level of insulin stimulation of 2-deoxy-D-glucose uptake (approximately 35% over basal uptake values) and the half-maximally effective insulin concentration (approximately 1 nM) were essentially the same for both the normal and diabetic fibroblasts. Furthermore, the absolute basal and maximally insulin-stimulated 2-deoxy-D-glucose uptake values were also the same. These results demonstrate that cultured fibroblasts from patients with noninsulin-dependent diabetes mellitus have normal basal and insulin-stimulated glucose transport capacities. Thus, the postreceptor defects in the glucose transport system previously demonstrated in vivo and in freshly isolated adipocytes from type II diabetic patients are most likely due to in vivo environmental factors rather than to an intrinsic (genetic) cellular defect(s).

We have investigated the previously described time-dependent increase in specific insulin binding to freshly isolated human erythrocytes incubated at 37 C. We found that at 37 C, specific insulin binding to erythrocytes rose to and remained at equilibrium for the first 90 min of incubation; thereafter, it rose in a rapid linear fashion, directly related to the increase in the degradation of unbound insulin and paralleling the intensity of visible hemolysis. The rise in specific binding was intensified by conditions in which hemolysis was enhanced and attenuated by conditions designed to limit hemolysis or by agents capable of inhibiting the degradation of insulin despite hemolysis. Thus, gentle handling of the cells prevents hemolysis, the degradation of insulin, and the increase in apparent insulin binding. Furthermore, 5% albumin, 2.5 mM N-ethylmaleimide, or excess unlabeled insulin (100 micrograms/ml) inhibited insulin degradation (even in the presence of hemolysis) and prevented the rise in insulin binding. The rise in cell-associated radioactivity after 90 min of incubation at 37 C was due to cellular uptake of products of insulin degradation, since degraded [125I]insulin rapidly associated with freshly prepared erythrocytes. Acid extraction studies suggested that about 60% of the cell-associated degraded material was intracellular, while the remaining approximately 40% was bound to the cell surface. The data suggest that the rise in binding in erythrocytes incubated at 37 C is a result of insulin degradation products which associate with the cells. The generation of degraded insulin is due to hemolysate released from leaky cells, and this phenomenon is unique to the in vitro situation.

To assess the significance of deficiency of circulating proinsulin in patients with type I diabetes mellitus, we studied the metabolic effects of biosynthetic human proinsulin in 24 patients. After withdrawing insulin, an infusion of proinsulin to physiological plasma levels did not prevent elevations of plasma glucose or beta-hydroxybutyrate. During steady state infusions of insulin and proinsulin, 13.7 times the steady state plasma level of proinsulin compared to insulin was required to maintain euglycemia. This finding indicates that proinsulin is approximately 7.3% as biologically active as insulin on a molar basis in maintaining glucose control. The MCRs of insulin and proinsulin during these steady state infusions were 12.5 +/- 2.2 (+/- SEM) and 2.62 +/- 0.33 ml/kg X min, respectively. After maintaining euglycemia overnight with an infusion of insulin or proinsulin and then acutely stopping these infusions, the rise in plasma glucose after proinsulin was delayed significantly compared to insulin, consistent with proinsulin's slower clearance. Further studies are necessary to determine whether biosynthetic human proinsulin has a specific role in the treatment of diabetes.

An insulin-requiring diabetic patient with intermittent periods of increased insulin requirements and insulin resistance was studied. The patient was found to have high titers of antiinsulin antibodies; subfractionation of the patient's serum revealed several populations of antiinsulin antibodies with differing affinities and titers for insulin. The ability of one of the insulin antibody fractions to bind [125I]iodoinsulin was markedly inhibited by the patient's serum (insulin depleted) and by purified total immunoglobulin G from which antiinsulin antibodies and insulin were removed. These findings suggested an antiidiotypic antibody in the patient's immunoglobulin G fraction reacting specifically with the antiinsulin antibody subfraction. Finally, the patient's serum contained an antiinsulin receptor antibody, as demonstrated by the ability of serum to specifically immunoprecipitate covalently labeled soluble insulin receptors. In conclusion, these results suggest that this patient generated a widespread polyclonal response to insulin, with the development of several populations of antiinsulin antibodies. An antiidiotypic antibody to a specific insulin antibody subfraction was present in the patient's serum which we believe had structural similarity to the binding site of the insulin molecule, accounting for the reactivity of the antiidiotypic antibody with the insulin receptor.