Pancreatic steatosis: harbinger of type 2 diabetes in obese rodents.

OBJECTIVE: The aim of this study was to determine if the fat accumulation in the exocrine pancreas fat of obese Zucker diabetic fatty (ZDF) rodents, like that in their endocrine pancreas, precedes the onset of type 2 diabetes mellitus (T2DM). As the fat content of whole pancreas, but not islets, can now be measured in humans by magnetic resonance spectroscopy (MRS), such measurements could be used as a predictor of impending T2DM and an indication for preventive intervention. ANIMALS: Obese ZDF (fa/fa) rats and lean (+/+) controls on a 6% fat diet were killed at time points from 6 to 16 weeks and total pancreatic fat was measured biochemically and electronmicroscopic examination of tissue for fat droplets was carried out. RESULTS: Compared to lean ZDF controls, pancreatic fat was elevated above lean controls from 6 to 16 weeks of age, peaking at 10 weeks of age when hyperglycemia first appeared. The pancreatic profile of fat content in whole pancreas paralleled that of islets. Electronmicroscopic examination identified the acinar location of the fat droplets and ruled out a major contribution of intrapancreatic adipocytes. CONCLUSION: The almost identical pattern of triglyceride overaccumulation in the exocrine and endocrine pancreas of obese rodents before the onset of T2DM suggests that MRS of the human pancreas might predict T2DM in obese subjects and permit timely interventions to prevent the disease.

Protein tyrosine phosphatase 1B: a potential leptin resistance factor of obesity.

Indirect evidence implicates leptin resistance in the pathogenesis of the lipotoxicity that complicates obesity and results in the metabolic syndrome. In this issue of Developmental Cell, two groups identify protein tyrosine phosphatase 1B (PTP1B) as a cause of leptin resistance through dephosphorylation of Jak2.

Diabetic hyperglycemia: link to impaired glucose transport in pancreatic beta cells.

Glucose uptake into pancreatic beta cells by means of the glucose transporter GLUT-2, which has a high Michaelis constant, is essential for the normal insulin secretory response to hyperglycemia. In both autoimmune and nonautoimmune diabetes, this glucose transport is reduced as a consequence of down-regulation of the normal beta-cell transporter. In autoimmune diabetes, circulating immunoglobulins can further impair this glucose transport by inhibiting functionally intact transporters. Insights into mechanisms of the unresponsiveness of beta cells to hyperglycemia may improve the management and prevention of diabetes.

Circulating somatostatin acts on the islets of Langerhans by way of a somatostatin-poor compartment.

Somatostatin perfused in canine pancreases at 10 to 20 picograms per milliliter or 10 to 20 percent of the pancreatic vein somatostatin concentration inhibited insulin and glucagon secretion. This suggests that the high local concentration of endogenous somatostatin is not in contact with somatostatin receptors of the islets. The integrity of this separation may determine the sensitivity of islet cells to circulating somatostatin.

Splanchnic somatostatin: a hormonal regulator of nutrient homeostasis.

Free (approximately 1600 daltons) somatostatin-like immunoreactivity was identified in arterial plasma of dogs that had received a test meal. Neutralization of circulating somatostatin while the dogs were consuming a fatty meal increased the plasma concentrations of triglycerides, gastrin, pancreatic polypeptide, and insulin after the meal. It is concluded that, in the dog, somatostatin is a true hormone that regulates the movement of nutrients from the gut to the internal environment.

Pronase effect on pancreatic beta cell secretion and morphology.

Pronase at low concentration (4 micrograms per milliliter) produces a reversible increase of glucose-stimulated insulin release in isolated islets of Langerhans. Pronase also affects the ultrastructure of the beta cells by inducing extensive development of tight junctions as well as the accumulation of secretory product within the extracellular spaces.

Underexpression of beta cell high Km glucose transporters in noninsulin-dependent diabetes.

The role of defective glucose transport in the pathogenesis of noninsulin-dependent diabetes (NIDDM) was examined in Zucker diabetic fatty rats, a model of NIDDM. As in human NIDDM, insulin secretion was unresponsive to 20 mM glucose. Uptake of 3-O-methylglucose by islet cells was less than 19% of controls. The beta cell glucose transporter (GLUT-2) immunoreactivity and amount of GLUT-2 messenger RNA were profoundly reduced. Whenever fewer than 60% of beta cells were GLUT-2-positive, the response to glucose was absent and hyperglycemia exceeded 11 mM plasma glucose. We conclude that in NIDDM underexpression of GLUT-2 messenger RNA lowers high Km glucose transport in beta cells, and thereby impairs glucose-stimulated insulin secretion and prevents correction of hyperglycemia.

Relationship of glucagon suppression by insulin and somatostatin to the ambient glucose concentration.

The glucagon-suppressing activity of insulin and somatostatin were compared at high and low glucose concentrations. In normal dogs made hyperglucagonemic by phloridzin pretreatment, insulin and somatostatin suppressed glucagon at rates of 47 +/- 8 and 35 +/- 8%/h (NS), respectively, despite profound hypoglycemia. In severely hyperglycemic alloxan-diabetic dogs, insulin and somatostatin suppressed glucagon at rates of 48 +/- 13 and 54 +/- 6%/h, respectively, not different from the nondiabetic dogs. After phloridzin pretreatment to eliminate hyperglycemia in the diabetic dogs, insulin and somatostatin suppressed 51 +/- 8 and 31 +/- 10%/h (NS), respectively. Glucose infused in the phloridzin-pretreated insulin-deprived group suppressed glucagon only partially; insulin was required to reduce it further. We conclude that insulin and somatostatin suppress glucagon at similar rates irrespective of ambient glucose levels, and that diabetic hyperglucagonemia represents the summation of stimulation by insulin lack minus suppression by the associated hyperglycemia.

Somatostatin-induced changes in insulin and glucagon secretion in normal and diabetic dogs.

In conscious dogs intravenously infused somatostatin (3.3 mug per min for 1 h) caused prompt and sustained declines in mean plasma insulin and glucagon, even during alanine infusion and intraduodenal casein hydrolysate feeding; plasma glucose declined, but not significantly. 6.7 mug per min of somatostatin significantly lowered pancreatoduodenal vein glucagon and insulin within 2.5 min and profoundly suppressed their secretion throughout the infusion. Consistent bihormonal suppression occurred at rates as low as 24 ng per kg per min, but was variable at 12 and 2.4 ng per kg per min. When somatostatin-induced (3.3 mug per min) hypoglucagonemia was corrected by exogenous glucagon, hyperglycemia occurred. In dogs with long-standing insulin-requiring alloxan diabetes 3.3 mug per min of somatostatin suppressed glucagon to 55 pg per ml throughout the 30-min infusion and lowered glucose by 36.4+/-6.1 mg per dl, about 1 mg per dl per min. Glucagon suppression was maintained despite alanine infusion, and glucose, which rose 29 mg per dl during alanine infusion without somatostatin, declined 58 mg per dl in the somatostatin-treated diabetic dogs despite alanine. Continuous infusion of somatostatin for 24 h in five insulin-requiring alloxan-diabetic dogs suppressed glucagon and lowered glucose significantly, usually to below normal. It is concluded that in normal dogs pharmacologic doses of somatostatin virtually abolish insulin and glucagon secretion in the basal state and during hyperaminoacidemia. Hyperglycemia occurs during somatostatin-induced insulin lack only if hypoglucagonemia is corrected. Somatostatin suppresses glucagon in diabetic dogs and lowers their plasma glucose approximately 1 mg per dl per min, even when the gluconeogenic substrate alanine is abundant. Glucagon suppression can be maintained for several hours in such dogs and hyperglycemia is thereby reduced.

Effect of insulin-glucose infusions on plasma glucagon levels in fasting diabetics and nondiabetics.

The effect of the intravenous infusion of insulin plus glucose on plasma glucagon levels was studied in hyperglycemic fasting adult-type and juvenile-type diabetics and compared with fasting nondiabetics. Adult-type diabetics were given insulin for 2 h at a rate of 0.03 U/kg-min, raising their mean insulin to between 25 and 36 muU/ml; glucagon declined from a base-line value of 71+/-2 (SEM) to 56+/-1 pg/ml at 120 min (P less than 0.001). In juvenile-type diabetics given the same insulin-glucose infusion, glucagon declined from a base-line level of 74+/-8 to 55+/-5 pg/ml at 120 min (P less than 0.05). The absolute glucagon values in the diabetic groups did not differ significantly at any point from the mean glucagon levels in nondiabetics given insulin at the same rate plus enough glucose to maintain normoglycemia. When glucagon was expressed as percent of baseline, however, the normoglycemic nondiabetics exhibited significantly lower values than adult-type diabetics at 90 and 120 min and juvenile-type diabetics at 60 min. In nondiabetics given insulin plus glucose at a rate that caused hyperglycemia averaging between 134 and 160 mg/dl, glucagon fell to 41+/-7 pg/ml at 120 min, significantly below the adult diabetics at 90 and 120 min (P less than 0.01 and less than 0.05) and the juvenile group at 60 min (P less than 0.01). The mean minimal level of 39+/-2 pg/ml was significantly below the adult (P less than 0.001) and juvenile groups (P less than 0.05). When insulin was infused in the diabetic groups at a rate of 0.4 U/kg-min together with glucose, raising mean plasma insulin to between 300 and 600 muU/ml, differences from the hyperglycemic nondiabetics were no longer statistically significant. It is concluded that, contrary to the previously reported lack of insulin effect in diabetics during carbohydrate meals, intravenous administration for 2 h of physiologic amounts of insulin plus glucose is accompanied in unfed diabetics by a substantial decline in plasma glucagon. These levels are significantly above hyperglycemic nondiabetics at certain points but differ from normoglycemic nondiabetics only when expressed as percent of the baseline. At a supraphysiologic rate of insulin infusion in diabetics, these differences disappear.

Opposing actions of glucagon and insulin on splanchnic d cell function.

The effect of glucagon at various infusion rates on plasma levels of somatostatin-like immunoreactivity (SLI) was examined in conscious normal and chronic alloxan diabetic dogs. In normal dogs glucagon infused at 6, 36, and 120 ng/kg per min did not affect the peripheral venous plasma SLI levels. In diabetic dogs, however, peripheral venous plasma SLI levels in inferior vena cava rose significantly from a mean base-line value of 181+/-9 pg/ml to a peak value of 279+/-38 pg/ml during the infusion of 120 mug/kg per min of glucagon, which raised plasma immunoreactive glucagon to >5,000 pg/ml. This glucagon-mediated increase was completely abolished by coinfusion of 7 mU/kg per min of insulin, a rate that maintained the ratio of insulin to glucagon at approximately the baseline value. In anesthetized normal dogs the concentration of SLI in the venous effluent of the pancreas, the gastric fundus and the antrum increased significantly with each infusion rate of glucagon using including the lowest rate of 4 ng/kg per min, which raised the plasma level of glucagon to 395+/-19 pg/ml. This stimulatory effect on SLI secretion was completely abolished by insulin coinfusion at a rate designed to maintain the insulin to glucagon ratio at approximately the baseline value, but the effect of a high 90-ng/kg per min infusion on pancreatic and gastric SLI release was not suppressed by coinfusion of 10 mU/kg per min insulin. These results suggest that glucagon stimulates splanchnic D cells unless insulin secretion is proportionally stimulated and suggests that the splanchnic D cell is a common target upon which the two hormones exert opposing actions. The loss of insulin inhibition of glucagon-mediated somatostatin secretion may account for the hypersomatostatinemia of severe diabetes.

Role of nitric oxide in obesity-induced beta cell disease.

Here we report that free fatty acid-induced suppression of insulin output in prediabetic Zucker diabetic fatty (ZDF) rats is mediated by NO. When normal islets were cultured in 2 mM FFA, NO production and basal insulin secretion increased slightly. In cultured prediabetic ZDF islets, FFA induced a fourfold greater rise in NO, upregulated mRNA of inducible nitric oxide synthase (iNOS), and reduced insulin output; both nicotinamide and aminoguanidine, which lower NO, prevented the FFA-mediated increase in iNOS mRNA, reduced NO, and minimized the loss of insulin secretion. In vivo nicotinamide or aminoguanidine treatment of prediabetic ZDF rats prevented the iNOS expression in islets and decreased beta cell dysfunction while blocking beta cell destruction and hyperglycemia. We conclude that NO-lowering agents prevent adipogenic diabetes in obese rats.

Role of circulating somatostatin in regulation of gastric acid secretion, gastrin release, and islet cell function. Studies in healthy subjects and duodenal ulcer patients.

Studies were designed (a) to determine whether somatostatin is released into the circulation after meals in sufficient amounts to regulate gastric or pancreatic islet function in humans and (b) to investigate the possible role of somatostatin in the pathogenesis of duodenal ulcer disease. Mean plasma somatostatin-like immunoreactivity (SLI) increased from 6.2 +/- 1.5 pg/ml to a peak level of 13.8 +/- 1.3 pg/ml in eight healthy subjects after a 1,440-cal steak meal (P less than 0.005). When somatostatin-14 was infused intravenously, basal and food-stimulated gastric acid secretion and also basal and food-simulated plasma insulin and glucagon concentrations were reduced significantly at mean plasma SLI concentrations within the range seen after a meal. Thus, the amount of somatostatin reaching the systemic circulation after a steak meal was sufficient to inhibit gastric acid secretion and islet cell function. On the other hand, basal and food-stimulated plasma gastrin concentrations were reduced by intravenous somatostatin only at plasma SLI concentrations that were several-fold greater than post-prandial SLI concentrations. Although duodenal ulcer patients had significantly higher basal, food-stimulated, and peak pentagastrin-stimulated gastric acid secretion rates than healthy controls, duodenal ulcer patients and controls had nearly identical basal and food-stimulated SLI concentrations. Moreover, food-stimulated gastric acid secretion and gastrin release were inhibited by intravenous somatostatin to the same extent in ulcer patients and controls. These studies suggest that duodenal ulcer patients release normal amounts of somatostatin into the circulation and that target cells controlling acid secretion and gastrin release are normally sensitive to somatostatin in these patients.

Plasma somatostatin-like immunoreactivity during the interdigestive period in the dog.

To study possible physiologic relationships between somatostatin and the gastric interdigestive contractions (GIC), gastric motor activity, and plasma somatostatin-like immunoreactivity (SLI) concentration were determined simultaneously in four conscious dogs, each of which was studied on two separate occasions. Plasma SLI level was highest during the GIC period and lowest 60 and 80 min after the cessation of the GIC; the mean difference in plasma SLI was 41 +/- 6 pg/ml. When synthetic motilin, a known stimulus of GIC, was infused at a physiologic rate during the period in which plasma SLI levels were low, SLI rose to approximately the same values observed during the contraction period and GIC similar to those that occur spontaneously were observed. When synthetic somatostatin, a known inhibitor of endogenous motilin release, was infused at a rate that raised the plasma SLI to approximately the levels observed during the contraction period (0.1 microgram/kg per h), the appearance of the subsequent GIC was significantly delayed. These results are consistent with a physiological role for somatostatin in the regulation of GIC in dogs and suggest an interrelationship between motilin and somatostatin.

Leptin- or troglitazone-induced lipopenia protects islets from interleukin 1beta cytotoxicity.

Interleukin 1beta (IL-1beta)-induced beta cell cytotoxicity has been implicated in the autoimmune cytotoxicity of insulin-dependent diabetes mellitus. These cytotoxic effects may be mediated by nitric oxide (NO). Since long-chain fatty acids (FFA), like IL-1beta, upregulate inducible nitric oxide synthase and enhance NO generation in islets, it seemed possible that islets might be protected from IL-1beta-induced damage by lowering their lipid content. We found that IL-1beta-induced NO production varied directly and islet cell viability inversely with islet triglyceride (TG) content. Fat-laden islets of obese rats were most vulnerable to IL-1beta, while moderately fat-depleted islets of food-restricted normal rats were less vulnerable than those of free-feeding normal rats. Severely lipopenic islets of rats made chronically hyperleptinemic by adenoviral leptin gene transfer resisted IL-1beta cytotoxicity even at 300 pg/ml, the maximal concentration. Troglitazone lowered islet TG in cultured islets from both normal rats and obese, leptin-resistant rats and reduced NO production and enhanced cell survival. We conclude that measures that lower islet TG content protect against IL-1beta-induced NO production and cytotoxicity. Leptin or troglitazone could provide in vivo protection against insulin-dependent diabetes mellitus.

Studies of muscle capillary basement membranes in normal subjects, diabetic, and prediabetic patients.

A technique is described for the measurement of muscle capillary basement membranes by electron microscopic examination of needle biopsies of the quadriceps muscle. With this procedure it has been possible to obtain an objective evaluation of the significance of capillary basement membrane hypertrophy in diabetic microangiopathy. The results of such studies of muscle capillary basement membrane thickness in 50 normal, 51 diabetic, and 30 prediabetic patients have demonstrated the following. First, that the average capillary basement membrane width of diabetic patients is over twice that of normal subjects; moreover, such basement membrane thickening is a very constant finding among overtly diabetic patients, in that approximately 98% of individual diabetic subjects demonstrated this lesion. The degree of basement membrane thickening in diabetic patients is, however, unrelated to age, weight, severity, or duration of diabetes. Second, capillary basement membrane hypertrophy has been found in approximately 50% of patients who are genetically prediabetic but who have not yet demonstrated evidence of the manifest carbohydrate disturbances of diabetes mellitus. Third, in contrast to the results obtained in genetically diabetic patients, subjects with severe hyperglycemia due to causes other than genetic diabetes only infrequently show basement membrane hypertrophy. These results indicate that thickening of the muscle capillary basement membranes is a characteristic of genetic diabetes mellitus, and further, that the hyperglycemia of diabetes is probably not the factor responsible for the microangiopathy characteristic of diabetes mellitus. Finally, the discovery of thickened capillary basement membranes in prediabetic patients suggests that basement membrane hypertrophy is a relatively early lesion of the diabetic syndrome and provides further support for the conclusion that this vascular defect is independent of carbohydrate derangements of diabetes mellitus.

Glucagon-stimulating activity of 20 amino acids in dogs.

The effect of 20 L-amino acids upon pancreatic glucagon secretion has been studied in conscious dogs. Each amino acid was administered intravenously over a 15 min period in a dose of 1 mmole/kg of body weight to a group of four or five dogs. Pancreatic glucagon and insulin were measured by radioimmunoassay. 17 of the 20 amino acids caused a substantial increase in plasma glucagon. Asparagine had the most glucagon-stimulating activity (GSA), followed by glycine, phenylalanine, serine, aspartate, cysteine, tryptophan, alanine, glutamate, threonine, glutamine, arginine, ornithine, proline, methionine, lysine, and histidine. Only valine, leucine, and isoleucine failed to stimulate glucagon secretion, and isoleucine may have reduced it. No relationship between glucagon-stimulating activity and insulin-stimulating activity was observed. The amino acids which enter the gluconeogenic pathway as pyruvate and, which are believed to provide most of the amino acid-derived glucose, had a significantly greater GSA than the amino acids which enter as succinyl CoA or as alpha-ketoglutarate. However, pyruvate itself did not stimulate glucagon secretion. The R-chain structure of the amino acid did not appear to be related to its GSA, except that the aliphatic branched chain amino acids, valine, leucine, and isoleucine, were devoid of GSA.

The effect of insulin on the alpha-cell response to hyperglycemia in long-standing alloxan diabetes.

In acute experimental diabetes in animals, alpha-cell unresponsiveness to hyperglycemia can be promptly corrected by insulin, but in human diabetes, even massive doses of insulin have little effect. To determine if this inability of insulin to correct the alpha-cell abnormality in man is merely the consequence of the long duration of the diabetic state (rather than of a difference in mechanism), the effect of insulin was studied in alloxan diabetes of long duration. Alloxan-diabetic dogs were maintained for 7-18 mo and treated daily with insulin. When glucose was infused without insulin, glucagon did not decline but rose paradoxically. However, when insulin was infused at a rate of 9 mU/kg/min together with glucose, a prompt decline in glucagon from a base-line average of 171 pg/ml SEM+/-34 to a nadir of 41 pg/ml SEM+/-9 was observed. This decline indicated that alpha-cell responsiveness to hyperglycemia is completely restored by large quantities of insulin. To determine if small amounts of insulin would similarly restore alpha-cell responsiveness in long-standing experimental diabetes, 1.4 mU/kg/min was infused. By the time the mean insulin level had risen 43 muU/ml, glucagon had declined significantly and ultimately fell to a nadir of 44 pg/ml. It is concluded from these studies that alpha-cell responsiveness to hyperglycemia can be fully restored in long-standing alloxandiabetic dogs as readily as in acutely diabetic dogs. Its ineffectiveness in restoring alpha-cell responsiveness to hyperglycemia in human diabetes may not, therefore, be related to duration of the diabetic state, and may reflect a primary alpha-cell defect.

The glycogenolytic activity of immunoreactive pancreatic glucagon in plasma.

Conclusions concerning the physiologic role of pancreatic glucagon in health and its contribution to disorders of carbohydrate metabolism, such as diabetes mellitus, are based entirely on measurements of plasma glucagon by radioimmunoassay. The changes in plasma immunoreactive glucagon can have the metabolic and clinical significance which has been implied, only if the glucagon detected by immunoassay has biological activity. The present study was designed to determine if a relationship between the immunoassayable glucagon and glycogenolytic activity of plasma could be demonstrated. Plasma specimens obtained from normal and diabetic subjects under widely varying circumstances of alpha cell activity were extracted by a modification of the Kenny technique and the recovery of immunoreactive glucagon was calculated. Glycogenolytic activity of each extract was determined by perfusion in the Mortimore rat liver system, modified so as to detect as little as 1 ng of crystalline glucagon.A significant correlation between the calculated quantity of immunoreactive glucagon and the glycogenolytic activity of plasma extracts was observed for both normal and diabetic subjects. Most of the glycogenolytic activity was abolished by incubating the extract with antiglucagon serum. It was concluded that the glycogenolytic activity of extractable glucagon is proportional to its immunoreactivity as calculated from its original concentration in plasma. This would tend to support the view that all or most of the immunoreactive glucagon of plasma is biologically active.

The effect of alanine on glucagon secretion.

If glucagon plays a hormonal role in the regulation of gluconeogenesis from endogenous amino acids, its secretion might be stimulated by an increase in the concentration of alanine, which has recently been identified as a principal gluconeogenic precursor. To determine if this is the case, 0.75 mmole of alanine per kilo was infused into conscious dogs immediately after a priming injection of 0.25 mmole per kg for 15 min. A uniform rise in the plasma level of pancreatic glucagon, as determined by a relatively specific radioimmunoassay for pancreatic glucagon, was observed. The rise, which averaged 90 pg per ml, was highly significant at 7(1/2) and 15 min after the start of the infusion. Insulin rose an average of only 8 muU per ml, while glucose rose an average of 10 mg per 100 ml. A lower dose of alanine, 1 mmole per kg, infused over a 1 hr period without an initial priming injection, also elicited a significant rise in glucagon measured in the pancreaticoduodenal venous plasma; glucagon rose from 350 pg per ml to 1066 pg per ml at the end of the infusion. The insulin response was modest and inconsistent, and glucose, again, rose 10 mg per 100 ml. To determine if the availability of exogenous glucose would abolish the alanine-induced rise in glucagon secretion, dogs were made hyperglycemic by a constant intravenous glucose infusion and were then given the high-dose alanine infusion. Under these circumstances, glucagon did not rise above the mean fasting concentration of 75 pg per ml, whereas mean insulin rose dramatically by more than 100 muU per ml. It was concluded that, in the fasting state, alanine does stimulate the secretion of glucagon, while having very little stimulatory effect on insulin secretion. Glucagon could, therefore, be a humoral mediator of gluconeogenesis from endogenous alanine, responding to hyperalaninemia in the fasting state, but not when exogenous glucose is available.