Oscillations of circulating plasma insulin concentrations in the rat

In previous studies, we found that insulin is secreted in a pulsatile fashion in vitro in isolated rat pancreatic islets. This study evaluated whether similar plasma insulin fluctuations occur in the rat in vivo. Freely moving rats were implanted with a chronic jugular catheter and serial blood samples were obtained 48-72 hrs post surgery. Blood was sampled at 3 min intervals for 60 mins with volume replacement using a red cell preparation. Plasma insulin concentrations were observed to fluctuate around a mean of 10.6 +/- 1.1 uU/ml, with an amplitude of 4.7 +/- 0.5 uU/ml and a period of 13.3 +/- 1 mins (n = 6). This was similar to the cycling observed in isolated islets at similar glucose concentrations. Sampling during the dark phase of the light-dark cycle in the rat was associated with an increase in the mean plasma level, amplitude and period of insulin oscillations compared with values obtained during the light phase (n = 3). These data are the first in vivo demonstration of oscillatory circulating insulin concentrations in the rat and show that the pulsatility in this species is similar to that observed in other mammals including man. We conclude that the chronically catheterised rat is a useful model for the evaluation of oscillating insulin concentrations in vivo, and may provide interesting insights by comparison with in vitro data in the same species.

Fetal catecholamine responses to maternal hypoglycemia

The present studies were designed to determine the fetal catecholamine and metabolic responses to insulin-induced maternal hypoglycemia. Maternal hypoglycemia was induced by a primed constant infusion of insulin and glucose administered to pregnant ewes to maintain maternal glucose at 20-25 mg/dl. Maternal and fetal samples for measurement of catecholamine, glucose, and free fatty acid levels and arterial blood gas analysis were collected before insulin infusion and at intervals thereafter for 6 h. Maternal and fetal plasma catecholamine levels increased significantly in response to hypoglycemia (analysis of variance, P less than 0.01). Fetal insulin and glucagon levels did not change despite a 50% reduction in fetal whole blood glucose concentration. Fetal free fatty acid levels increased significantly during hypoglycemia (P less than 0.05). There were no significant changes in maternal or fetal heart rate, blood pressure, or arterial blood gases during hypoglycemia. These results suggest that the fetus is capable of responding to hypoglycemia with an increase in catecholamine and free fatty acid levels. These results are consistent with the interpretation that the fetus is able to mobilize alternative energy substrates in response to maternal insulin-induced hypoglycemia.

Sulphonylurea effects on insulin secretion in islets desensitized to glucose

Prolonged, continuous exposure of the islets of Langerhans to high glucose concentrations results in desensitization of the beta cell to glucose stimulation. This study tested the ability of a sulphonylurea to stimulate insulin secretion in this setting. Normal isolated rat islets were cultured for 18-20 h in RPMI-1640 with 300 mg/dl glucose to induce desensitization or with 100 mg/dl as a control. Islets were then placed into a perifusion system and perifused with 60 mg/dl glucose followed by a stimulus. After preincubation at 300 mg/dl, a significant 50% suppression of glucose-induced insulin secretion compared with secretion in the control group preincubated at 100 mg/dl glucose was observed (p less than 0.025-0.001). This confirmed the occurrence of desensitization to glucose in this in vitro model. In contrast, stimulation of insulin secretion by glyburide (500 ng/ml) was unaffected compared with control. We also tested whether glyburide corrects the defective response to glucose stimulation in glucose-desensitized islets. Control islets (preincubated at 100 mg/dl) were stimulated with 300 mg/dl glucose or with this glucose concentration plus glyburide. Peak incremental insulin responses were similar (0.81 +/- 0.07 and 0.77 +/- 0.12 microU/ml.islet). After preincubation at 300 mg/dl, perifusion with 300 mg/dl glucose alone or with glyburide was associated with smaller, but similar, peak insulin responses (0.53 +/- 0.13 and 0.62 +/- 0.06 microU/ml.islet). In conclusion, islets in which the insulin-secretory response is compromised by desensitization to glucose are nevertheless completely responsive to the direct stimulatory effects of a sulphonylurea. However, the sulphonylurea does not correct the defect in glucose-induced insulin secretion.

A glucose reduction challenge in the differential diagnosis of fasting hypoglycemia: a two-center study

Investigation of patients with suspected or proven hypoglycemia is often a time-consuming and expensive process. We describe a glucose reduction challenge test which may be useful as an out-patient screening procedure. Insulin is infused for 3 h at 40 mU/kg.h. Plasma glucose was monitored at the bedside during the test, and blood samples were collected for measurement of C-peptide. Responses were examined in 17 normal controls, and 6 patients with insulinomas. In normal subjects, mean plasma glucose fell to a plateau value of 3.2 +/- 0.2 mmol/L (57 +/- 2.6 mg/dL) and remained at that level with few symptoms. In contrast, five of six patients with insulinomas developed severe hypoglycemia, with plasma glucose levels between 1.9 (34 mg/dL) and 2.2 mmol/L (39 mg/dL). Plasma C-peptide concentrations were suppressed to 0.08 pmol/mL or less in normal subjects, but in insulinoma patients remained at 0.32-1.6 pmol/mL i.e. outside the normal range, and diagnostic of nonsuppressible insulin secretion. These data demonstrate that moderate reduction of serum glucose maintained for a prolonged period results in marked suppression of plasma C-peptide, permitting improved discrimination between normal subjects and patients with insulinomas. This glucose reduction challenge can, therefore, be used as a test of glucose-regulating ability, where failure (hypoglycemia) per se represents a measurable abnormality. C-Peptide measurements will determine whether the cause of hypoglycemia is due to hyperinsulinemia.

Pulsatile insulin secretion in isolated rat islets

The pancreas secretes insulin in an oscillatory fashion, but the precise site of the pacemaker for pulsatile insulin secretion has not been identified. These studies were designed to determine whether islets also secrete insulin in a pulsatile fashion if they are isolated from their pancreatic milieu. Isolated rat islets (80-100) were perifused 8 h in culture medium after overnight incubation, and samples were collected at 3.3-min intervals. Insulin secretion was evaluated for pulsatility with the Clifton Cycle Detection Program. Perifusion of islets was associated with a spontaneous, persistent, and regular pulsatility of insulin secretion, which was observed in all conditions tested. Perifusion with medium containing 5.5 mM glucose (n = 11) demonstrated oscillations with a mean periodicity of 17.6 +/- 1.1 min and a mean amplitude of 4.8 +/- 0.4 microU/ml when overall mean insulin concentration was 16.7 +/- 2.4 microU/ml. When the glucose concentration was 16.7 mM (n = 9), overall mean insulin concentration was 54.4 +/- 2.6 microU/ml, with increases in periodicity (22.0 +/- 1.3 min) and amplitude (10.7 +/- 0.5 microU/ml). All measurements were significantly different from those observed during perifusion with 5.5 mM glucose (P less than 0.02-0.001). Theophylline (1 mM) also enhanced the overall mean insulin concentration and amplitude (69.4 +/- 10.4 and 14.2 +/- 1.2 microU/ml, respectively) compared with control studies without theophylline (16.7 +/- 5.3 and 4.3 +/- 0.5 microU/ml) (P less than 0.01). The period of the cycle was also increased from 17.5 +/- 1.1 to 26.4 +/- 6.3 min, but this was not significantly different from the control group.(ABSTRACT TRUNCATED AT 250 WORDS)

Pentamidine-induced beta cell toxicity is not preventable by high glucose

The incidence of beta cell damage attributable to pentamidine treatment of pneumocystis pneumonia is increasing in frequency because of the AIDS epidemic. We carried out in vitro studies in perfused rat islets using insulin secretion as an index of beta cell damage to study the effects of pentamidine and to test whether glucose can prevent toxicity in this physiologic model. Isolated islets were cultured for 16-18 hours of static incubation, in a culture medium containing 100 mg/dl glucose, with or without pentamidine (10(-6) M, a therapeutic concentration). Islets were then perfused with media containing 60 mg/dl followed by 300 mg/dl glucose concentrations to study the insulin secretory response. Incubation of islets with pentamidine was associated with subsequent basal hypersecretion of insulin (0.40 +/- 0.05 microU/islet .5 minute vs. 0.18 +/- 0.04 microU/islet .5 minute, p less than .005), and an insulin secretory response to glucose which was completely abolished (0.05 +/- 0.04 microU/islet .5 minute versus 1.12 +/- 0.02 microU/islet .5 minute, p less than .005). To determine whether glucose may protect against the effects of pentamidine, islets were then exposed to high glucose concentrations during simultaneous incubation with pentamidine. Coincubation with high glucose did not prevent these insulin secretory defects. A more extended culture of pentamidine-treated islets in the absence of pentamidine and at a glucose concentration of 100 mg/dl did not result in any recovery of insulin secretion. We conclude that pentamidine-induced beta cell damage is irreversible, not preventable by incubation with high glucose concentrations, and may therefore result from a mechanism different to that of alloxan.

Sparing of cognitive function in mild hypoglycemia: dissociation from the neuroendocrine response

Central nervous system function during insulin-induced reductions in plasma glucose was studied by measuring plasma epinephrine concentrations and testing cognitive function. Mild glucose reduction [mean plasma glucose, 62 +/- 3 (+/- SEM) mg/dL (3.4 +/- 0.2 mmol/L)] was induced with an iv insulin infusion at the rate of 40 mU/kg.h for 180 min in 7 normal subjects. Despite a marked increase in mean plasma epinephrine concentrations, which peaked at 426 +/- 68 pm/mL (2325 +/- 371 pmol/mL; P less than 0.001), no significant differences in cognitive function occurred as determined by a series of trail-making tests compared with the results of serial tests in a group of 17 control subjects. In contrast, when hypoglycemia was induced (plasma glucose, less than 42 mg/dL; 2.3 mmol/L) by bolus injection of insulin in 4 normal subjects, cognitive function was impaired in every subject, as demonstrated by a delay in completion of the trail-making test. The mean completion time was prolonged to 107 +/- 16% of the baseline at the time of hypoglycemia vs. 74 +/- 4% in control subjects (P less than 0.01). These findings suggest that cognitive function may be spared during mild plasma glucose reductions and is dissociated from the neuroendocrine adrenergic response that is activated under these conditions. This dissociation may be part of a homeostatic process in which overall brain function is maintained during glucoprivation, although counterregulation has already been triggered to prevent a further decrease in plasma glucose.

Somatostatin impairs clearance of exogenous insulin in humans

Somatostatin has been widely used to suppress endogenous pancreatic hormone secretion in research studies. Many of these studies required the simultaneous infusion of a hormone together with somatostatin. A critical assumption for its use in metabolic investigation is that somatostatin has no effect on the action or clearance of a concomitantly infused hormone. To test whether clearance of an exogenously infused hormone is affected, we infused insulin with or without somatostatin in two sets of studies. Insulin (40 mU X kg-1 X h-1) was infused for 100 min (n = 6). Plasma glucose levels fell to 55 +/- 4.1 mg/dl with insulin alone and significantly lower, to 44 +/- 1.9 mg/dl, when somatostatin (250 micrograms/h) was also infused (P less than .01). Plasma immunoreactive insulin (IRI) rose to 57 +/- 12.5 microU/ml with insulin alone, which was significantly different from 88 +/- 15 microU/ml when insulin was infused together with somatostatin (P less than .01). When a smaller dose of insulin (30 mU X kg-1 X h-1) was infused for 100 min (n = 4), similar results were observed. When somatostatin was infused together with insulin, plasma glucose fell to lower levels (41 +/- 4.2 vs. 62 +/- 9.5 mg/dl; P less than .01) and plasma IRI rose higher (39 +/- 8.5 vs. 27 +/- 5.9 microU/ml; P less than .01) than when insulin was infused alone. C-peptide was equally suppressed by hypoglycemia regardless of whether somatostatin was administered, indicating suppression of endogenous insulin during these studies. We conclude that somatostatin infusion impairs the clearance of exogenous insulin.(ABSTRACT TRUNCATED AT 250 WORDS)

Naloxone decreases centrally induced hyperglycemia in dogs. Evidence for an opioid role in glucose homeostasis

Intracerebroventricular (ICV) instillation of morphine and beta-endorphin causes centrally induced hyperglycemia. Locally active, endogenous opioids in the central nervous system may, therefore, also be involved in the elevation of blood sugar. This possibility was tested by examining the glucoregulatory response to central glucoprivation induced by ICV administration of 2-deoxy-D-glucose (2DG) in dogs. Administration of 2DG resulted in a rise in plasma glucose and immunoreactive glucagon (IRG) of 108 +/- 19 mg/dl and 70 +/- 20 pg/ml, respectively. These changes were attenuated by the simultaneous central infusion of the opiate antagonist naloxone: plasma glucose levels increased by 77 +/- 14 mg/dl and IRG by 43 +/- 3 pg/ml, both significantly different from the effect of 2DG alone (P less than 0.05-0.01). These findings suggest that opiate receptors participate in the counterregulatory response to central glucoprivation. They also provide a mechanism by which endogenous opioid peptides may play a role in the central regulation of glucose homeostasis.

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.

Central control of peripheral circulating somatostatin in dogs: effect of 2-deoxyglucose

Circulating plasma somatostatin concentrations are known to fluctuate in response to nutrients and hormones. However, little is known about neural or central nervous system (CNS) control of somatostatin secretion. To test whether peripheral circulating somatostatin is influenced by a central stimulus, 2-deoxyglucose (37.5 mg/kg) was infused into a lateral cerebral ventricle of six conscious dogs over a period of 15 min. Plasma somatostatin levels rose from a base line of 105 +/- 6 pg/ml (mean +/- SE) to a peak of 154 +/- 10 pg/ml (P less than 0.005) at 30 min after the onset of the infusion. Somatostatin levels were still significantly elevated (P less than 0.025) at 60 min (119 +/- 6 pg/ml) and thereafter gradually returned toward base line. Plasma glucose and glucagon levels increased in response to intraventricular 2-deoxyglucose. Glucose concentrations rose from 105 +/- 5 mg/dl to peak at 203 +/- 16 mg/dl (P less than 0.005) at 80 min and remained elevated to 120 min. The concentration of plasma glucagon increased from 41 +/- 6 to 92 +/- 18 pg/ml at 60 min (P less than 0.05) and then declined. In marked contrast to intraventricular 2-deoxyglucose, similar concentrations of 2-deoxyglucose administered intravenously (n = 4) resulted in a slight fall in plasma somatostatin. Intraventricular saline did not result in a change in plasma somatostatin. It is concluded that peripheral circulating somatostatin may be susceptible to central nervous system control.

Effect of somatostatin on determinants of bile flow in unanesthetized dogs

Seven dogs each underwent cholecystectomy, ligation of the accessory pancreatic duct, and insertion of a Thomas duodenal cannula opposite the ampulla of Vater. After full recovery, bile secretions were studied in the unanesthetized dogs by opening the cannula and placing a ureteric catheter through the papilla into the common bile duct. All animals received, throughout study, constant infusions of taurocholic acid to replace losses caused by interruption of the enterohepatic circulation and 14 C-erythritol for measurement of erythritol clearance. After bile flow stabilized somatostatin 800 ng/kg/minute was infused for 100 minutes and bile flow declined from 3.0 +/- 0.7 ml/10 minutes (SD) to 1.19 +/- 0.47 ml/10 minutes (p less than 0.001) and 14C-erythritol clearance fell from 3.6 +/- 1.14 to 1.77 +/- 0.43 ml/10 minutes (p less than 0.001). Bile salt output was unchanged, indicating that somatostatin inhibited bile salt-independent canalicular flow (BSICF). In other experiments animals underwent intraduodenal acidification which resulted in a marked increase in bile flow. Somatostatin infusion again causes a sharp fall in bile flow (p less than 0.05) suggesting that somatostatin also inhibited ductular flow. Infusion of somatostatin did not inhibit choleresis produced by exogenous secretin administration. Thus, somatostatin inhibits 1) ductular flow by inhibiting secretin release and 2) BSICF by a direct effect or by decreasing the release of hormones which induce canalicular flow.

Morphine-induced hyperglycemia: role of insulin and glucagon

An iv bolus injection of 0.5 mg/kg morphine, about twice the therapeutic dose, caused plasma glucose to rise more than 120 mg/dl in alloxan-diabetic conscious dogs but had little effect on conscious normal dogs. Plasma glucagon rose in the diabetic and nondiabetic groups by 30 +/- 10 and 100 +/- 29 pg/ml, respectively, but insulin levels increased significantly only in the nondiabetics. The hyperglycemic action on morphine may, at least in part, be the result of an increase in glucagon secretion without a sufficient accompanying release of insulin.

Somatostatin analogs inhibit somatostatin release

To determine if, like insulin, somatostatin inhibits its own secretion from the pancreas, nonimmunoreactive analogs of somatostatin were perfused in an isolated dog pancreaticoduodenal preparation using a nonrecirculating system. [D-Trp8-D-Cys14]somatostatin, at a concentration of 200 ng/ml, blocked the response of somatostatin-like immunoreactivity (SLI) to cholecystokinin and arginine. When perfusion of the analog was discontinued, SLI release increased. At a concentration of 0.1 ng/ml, des Asn5-[D-Trp8]somatostatin lowered SLI levels significantly without significantly reducing glucagon levels. At a concentration of 1 ng/ml, des Asn5-[D-Trp8]somatostatin significantly inhibited SLI as well as insulin and glucagon release. Perfusion of glucagon at a concentration of 10 ng/ml failed to overcome the blockade of SLI and insulin release caused by 50 ng/ml des Asn5-[D-Trp8]somatostatin. The results are compatible with a direct inhibitory effect of somatostatin analogs upon SLI release and raise the possibility of a self-inhibiting action of the native hormone.

Bombesin stimulates the release of insulin and glucagon, but not pancreatic somatostatin, from the isolated perfused dog pancreas

Synthetic bombesin, perfused in the isolated canine pancreas at a rate of 340-380 ng/min for 10 min, elicited a 4-fold rise in insulin to a peak at 2 min; a rapid decline followed discontinuation of bombesin. Glucagon rose by 50% to a peak at 6 min, but remained elevated after discontinuation of the bombesin. Somatostatin-like immunoreactivity was not significantly affected by perfusion with bombesin.

Properties of endogenous somatostatin-like immunoreactivity and synthetic somatostatin in dog plasma

Somatostatin-like immunoreactivity (SLI) in the peripheral venous plasma of dogs and in their pancreatic and gastric venous effluents was characterized and compared with synthetic somatostatin. Both endogenuous plasma SLI and somatostatin added to plasma were eluted from Sephadex gels at pH 8.8 in the 150,000--200,000-mol wt region but at pH 2.5 both appeared in the 1,500--2,000-mol wt region. The SLI released from the isolated dog pancreas perfused with plasma-free buffer was eluted entirely as a 1,600-dalton polypeptide, but when the pancreas was perfused with plasma, SLI was eluted in the 150,000--200,000-mol wt zone. Affinity chromatography of plasma samples on immobilized antibodies directed against the central portion of the somatostatin molecule (residues 5--9 and 11) removed approximately equal to 90% of both endogenous SLI and somatostatin added to plasma, but neither was removed by affinity chromatography on antibodies directed against the NH2-terminal region of somatostatin (residues 1--4). The SLI from plasma and from pancreas perfusate isolated by affinity chromatography was identical in molecular size, charge, and immunometric properties to synthetic somatostatin. It is concluded that endogenous SLI is secreted by the pancreas and stomach in a form not distinguishable from synthetic somatostatin, but circulates in plasma bound to large molecular weight components; the NH2-terminal residues of somatostatin appear to be important in this binding.