Comparative study on the effects of a hypoglycemic 2-substituted-2-imidazoline derivative (DG-5128) and tolbutamide on insulin secretion from and insulin synthesis in the isolated rat pancreatic islets

2[2-(4,5-Dihydro-1H-imidazol-2-yl)-1-phenylethyl]pyridine dihydrochloride sesquihydrate (DG-5128) was found to stimulate the glucose-primed insulin secretion from the isolated rat pancreatic islets throughout the incubation period, unlike tolbutamide which stimulated it only in the initial phase of incubation. The effect of DG-5128 was more pronounced at a higher glucose concentration (5 mg/ml). In the islet perifusion study, DG-5128 was also found to stimulate the glucose-induced insulin secretion in both the first and the second phases of the reaction, in contrast to tolbutamide which stimulated only the first phase of insulin secretion from the perifused islets. DG-5128 gave no significant effect on the glucose-stimulated increase in incorporation of [³H]leucine into the pro-insulin and insulin fractions, while tolbutamide significantly inhibited the incorporation especially at a low glucose concentration (1 mg/ml). These and the previous findings indicate that DG-5128 is a new class of hypoglycemic agent with a unique mode of action different from the known hypoglycemics ever reported.     

Mechanisms involved in stimulation of insulin secretion by the hypoglycaemic alpha-adrenergic antagonist, DG-5128.

The selective alpha 2-antagonist DG-5128 provoked a dose-dependent stimulation of insulin release from isolated rat islets. DG-5128 was only weakly effective as an antagonist of noradrenaline-induced inhibition of insulin secretion but, surprisingly, was able to reverse the suppression of secretion and increase in 86Rb efflux from preloaded islets, mediated by diazoxide. These effects were not reproduced with more effective alpha-antagonists, suggesting that stimulation of insulin secretion by DG-5128 is independent of alpha-receptor blockade.     

Somatostatin-induced inhibition of insulin secretion from isolated islets of rat pancreas in presence of glucagon.

In order to study the oeffect of somatostatin on the endocrine pancreas directly, islets isolated from rat pancreas by collagenase were incubated for 2 hrs 1) at 50 and 200 mg/100 ml glucose in the absence and presence of somatostatin (1, 10 and 100 mg/ml) and2) at 200 mg/100 ml glucose together with glucagon (5 mug/ml), with or without somatostatin (100 ng/ml). Immunologically measurable insulin was determined in the incubation media at 0, 1 and 2 hrs. Insulin release was not statistically affected by any concentration stomatostatin. On the other hand, somatostatin exerted a significant inhibitory action on glucagon-potentiated insulin secretion (mean +/- SEM, mu1/2 hrs/10 islets: glucose and glucagon: 1253 +/- 92; glucose, glucagon and somatostatin: 786 +/- 76). The insulin output in th epresence of glucose, glucagon and somatostatin was also significantly smaller than in thepresence of glucose alone (1104 +/- 126) or of glucose and somatostatin (1061 +/- 122). The failure of somatostatin to affect glucose-stimulated release of insulin from isolated islets contrasts its inhibitory action on insulin secretion as observed in the isolated perfused pancreas and in vivo. This discrepancy might be ascribed to the isolation procedure using collagenase. However, somatostatin inhibited glucagon-potentiated insulin secretion in isolated islets which resulted in even lower insulin levels than obtained in the parallel experiments without glucagon. It is concluded that the hormone of the alpha cells, or the cyclic AMP system, might play a part in the machanism of somatostatin-induced inhibition of insulin release from the beta-cell.     

Effect of DG5128 on epinephrine and glucagon induced glucose output from the isolated perfused rat liver

The effect of a specific alpha 2-adrenergic antagonist 2-[2-(4,5-dihydro-1.H-imidazol-2-yl)-1-phenyl-ethyl] pyridine dihydrochloride sesquihydrate (DG5128), on the glucose output by epinephrine and/or glucagon was studied using the perfused rat liver. The administration of DG5128 alone did not affect the glucose output. However, DG5128 produced a significant inhibition of the increased glucose output when induced by 10(-6) M epinephrine alone or 10(-6) M epinephrine plus 1.4 x 10(-10) M glucagon. There were no significant changes of the glucose output by 1.4 x 10(-10) M or 7.0 x 10(-11) M glucagon alone. On the other hand, addition of 1 mU/ml insulin to the perfusate suppressed the 7.0 x 10(-11) M glucagon-induced glucose output, but failed to decrease the 1.4 x 10(-10) M glucagon effect. DG5128 suppressed further the glucagon (7.0 x 10(-11) M)-induced increase of glucose output in the presence of insulin. These results suggest that DG5128 produces a hypoglycemic effect partly through an inhibition of the increased hepatic glucose output elicited by epinephrine and glucagon.     

Suppression of insulin release by galanin and somatostatin is mediated by a G-protein. An effect involving repolarization and reduction in cytoplasmic free Ca2+ concentration

The effects of galanin and somatostatin on insulin release, membrane potential, and cytoplasmic free Ca2+ concentration [( Ca2+]i) were investigated using beta-cells isolated from obese hyperglycemic mice. Whereas insulin release was measured in a column perifusion system, membrane potential and [Ca2+]i were measured with the fluorescent indicators bisoxonol (bis-(1,3-diethylthiobarbiturate)trimethineoxonol) and quin 2, in cell suspensions in a cuvette. Galanin (16 nM) and somatostatin (400 nM) suppressed glucose-stimulated insulin release in parallel to promoting repolarization and a reduction in [Ca2+]i. The reduction in [Ca2+]i comprised an initial nadir followed by a slow rise and the establishment of a new steady state level. The slow rise in [Ca2+]i was abolished by 50 microM D-600, a blocker of voltage-activated Ca2+ channels. Both peptides suppressed insulin release even when [Ca2+]i was raised by 25 mM K+. Under these conditions the inhibition of insulin release was partly reversed by an increase in the glucose concentration. Addition of 5 mM Ca2+ to a cell suspension, incubated in the presence of 20 mM glucose and either galanin, somatostatin, or the alpha 2-adrenergic agonist clonidine (10 nM), induced oscillations in [Ca2+]i, this effect disappearing subsequent to the addition of D-600. The effects of galanin, somatostatin, and clonidine on [Ca2+]i were abolished in beta-cells treated with pertussis toxin. In accordance with measurements of [Ca2+]i, treatment with pertussis toxin reversed the inhibitory effect of galanin on insulin release. The inhibitory action of galanin and somatostatin on insulin release is probably accounted for by not only a repolarization-induced reduction in [Ca2+]i and a decreased sensitivity of the secretory machinery to Ca2+, but also by a direct interaction with the exocytotic process. It is proposed that these effects are mediated by a pertussis toxin-sensitive GTP-binding protein.     

Glucose regulation of arginine-induced pancreatic somatostatin release from the isolated perfused rat pancreas

The effects of glucose alone, combinations of glucose with arginine or tolbutamide and either arginine or tolbutamide alone, on somatostatin, insulin, and glucagon secretion were investigated using the isolated perfused rat pancreas. When glucose alone was raised in graded increments at 15-min intervals from an initial concentration of 0 mM to a maximum of 16.7 mM, somatostatin as well as insulin in the perfusate increased with the glucose, while glucagon decreased. The similarity of the glucose stimulated somatostatin and insulin release was especially evident when the perfusate glucose was increased from an initial dose of 4.4 mM rather than 0 mM to 8.8 mM or 16.7 mM. In addition, glucose at concentrations varying from 4.4 mM to 11 mM dose-dependently enhanced arginine-induced somatostatin and insulin release and suppressed glucagon release dose-dependently as before. Arginine in the absence of glucose was not capable of stimulating somatostatin secretion whereas tolbutamide, in contrast, was capable of stimulating somatostatin secretion even in the absence of glucose.     

Antidiabetic action of somatostatin after oral glucose loading: due to suppression of glucagon and growth hormone or of intestinal carbohydrate absorption?

To elucidate the mechanism by which somatostatin lowers blood glucose concentration and insulin requirement following carbohydrate ingestion in insulin dependent diabetic patients (IDDM; n = 6), the amount of insulin required for the assimilation of a 50 g glucose load was determined by means of an automated glucose-controlled insulin infusion system with and without concomitant somatostatin infusion. During the 3 hour period following glucose loading plasma concentrations of glucagon and growth hormone were diminished by somatostatin, as were the rise in blood glucose and insulin requirement (4.0 +/- 1.2 U) when compared with the control study (11.3 +/- 1.5 U; p less than 0.01). With cessation of somatostatin blood glucose levels and insulin requirement rose during the following 2 hour observation period (7.5 +/- 1.2 U) but remained basal during the control study (0.7 +/- 0.6 U; p less than 0.0005). Thus the integrated amounts of insulin required for glucose hormone were temporarily suppressed by somatostatin. It is concluded that the diminished insulin requirement and delayed rise in blood glucose during somatostatin administration after an oral glucose load is not due to its "antidiabetic" action by suppressing glucagon and growth hormone release. Our findings favour inhibition of intestinal carbohydrate absorption as the determining cause for the "antidiabetic" action of somatostatin.     

Effect of somatostatin on glucose-induced 45Ca uptake in the pancreatic islets.

This study was designed in an attempt to elucidate a mechanism of somatostatin inhibition of glucose-induced Ca+ uptake by rat pancreatic islets. Rat pancreatic islets were perifused with Krebs-Ringer bicarbonate (KRB) buffer containing 16.7 mM of glucose with somatostatin (2 micrograms/ml) or/and diltiazem HCl (2 x 10(-5) M). Somatostatin inhibited preferentially the early phase of glucose-induced insulin release, whereas diltiazem HCl inhibited the late one. And the concomitant presence of the submaximal concentration of somatostatin (2 micrograms/ml) and diltiazem HCl (2 x 10(-5 M) provided the completely additive inhibition of glucose-induced insulin release. Rat pancreatic islets were incubated with KRB buffer supplemented with 16.7 mM of glucose and 45CaCl2 (10 muCi/ml) for 5--60 min and the biphasic 45Ca uptake by pancreatic islets was obtained. Somatostatin (500 ng/ml-4 micrograms/ml) gave the suppressive effect on the early phase of glucose-induced 45Ca uptake, but the higher concentration (2 micrograms/ml) of somatostatin did not impair the late phase of 45Ca uptake by pancreatic islets. On the other hand, diltiazem HCl did suppress the late phase of glucose-induced 45Ca uptake dose-dependently, but did not suppress the early phase (2 x 10(-5) M). These data indicate that somatostatin suppresses the early phase of glucose-induced Ca2+ uptake preferentially to the late one and has a different action mechanism from Ca antagonist on glucose-induced insulin release.     

Exendin-4 dose-dependently stimulates somatostatin and insulin secretion in perfused rat pancreas.

We have investigated the effect of exendin-4, a GLP-1 analogue, on somatostatin and insulin secretion in perfused rat pancreas. At constant glucose concentration within the type 2 diabetic range (9 mM), exendin-4 stimulated somatostatin and insulin secretion in a dose-dependent manner. Dose-response curves were sigmoidal (R (2) = 0.9954 and R (2) = 0.9973, respectively; p < 0.01) and the EC (50) was 4.3 nM for somatostatin secretion and 1.4 nM for insulin secretion. Exendin-4 stimulated somatostatin output at low (3.2 mM), normal (5.5 mM) and high (9 mM) glucose concentrations, while the insulinotropic effect of exendin-4 was not found at low glucose levels. On the other hand, exendin-4 potentiated somatostatin and insulin responses to an increase in perfusate glucose levels, and to arginine and carbachol. Finally, the insulinotropic effect of exendin-4 was maintained in the absence of a somatostatin response as induced by cysteamine pretreatment, indicating a direct effect of exendin-4 on the B-cell. In summary, exendin-4 behaves as a general stimulatory agent of both insulin and somatostatin release in the perfused rat pancreas. Given that exendin-4 has also been shown to increase gastric somatostatin secretion, it is tempting to speculate that exendin-4 might behave as a general stimulator of D-cell function in other tissues, a point worthy of further investigation.     

Long-term pancreatic duct occlusion impairs the entero-insular axis in the dog--failure of plasma VIP to respond as "incretin"

The response of VIP to either an oral glucose load (OGT) or intravenous glucose (IV glucose), aimed at reproducing the plasma glucose level after OGT, was studied in trained, conscious, sham-operated (Sham; n = 6) dogs, and dogs having initially (12 months before the glucose experiments) undergone occlusion of the pancreatic duct by the prolamine glue technique (Occ; n = 5). As a result, prior to glucose studies, the exocrine pancreas function was found subtotally reduced, as indirectly evaluated by the para-aminobenzoic acid (PABA) test, but no signs of diabetes were detected. The two studies with glucose administration designed to demonstrate the release of insulin, VIP, somatostatin into plasma as modified by enteric signals (represented by the difference of plasma peptide concentration during OGT minus peptide concentration during IV glucose) revealed the following: (1) basal plasma glucose, insulin, VIP, somatostatin did not differ between Sham and Occ dogs; (2) after OGT in Occ dogs the plasma glucose was elevated, whereas plasma insulin was markedly reduced, and VIP, somatostatin were largely unchanged; (3) the integrated output of insulin only was impaired when considering the so-called entero-insulin axis, while integrated VIP, somatostatin were unaltered. It was concluded (a) the Occ procedure in the dog has the capacity to subtotal destruction of the pancreatic acinar tissue, and of the entero-insular axis of insulin, the latter through yet unknown pathways, (b) the Occ technique may be a useful tool for investigation of the nature of "incretin," (c) VIP and somatostatin do not respond to elevated blood glucose and may have no role in the "incretin" concept of enteric modulation of the B-cell.     

Attenuation of insulin secretion by insulin-like growth factor binding protein-1 in pancreatic beta-cells

IGFBP-1 is involved in glucohomeostasis, but the direct action of IGFBP-1 on the beta-cell remains unclear. Incubation of dispersed mouse beta-cells with IGFBP-1 for 30min inhibited insulin secretion stimulated by glucose, glucagon-like peptide 1 (GLP-1) or tolbutamide without changes in basal release of insulin and in cytosolic free Ca(2+) concentration ([Ca(2+)](i)) and NAD(P)H evoked by glucose. In contrast, IGFBP-1 augmented glucose-stimulated insulin secretion in intact islets, associated with a reduced somatostatin secretion. These results suggest a suppressive action of IGFBP-1 on insulin secretion in isolated beta-cells through a mechanism distal to energy generating steps and not involving regulation of [Ca(2+)](i). In contrast, IGFBP-1 amplifies glucose-stimulated insulin secretion in intact islets, possibly by suppressing somatostatin secretion. These direct modulatory influences of IGFBP-1 on insulin secretion may imply an important regulatory role of IGFBP-1 in vivo and in the pathogenesis of type 2 diabetes, in which loss of insulin release is an early pathogenetic event.     

Insulin release from collagenase-isolated islets of rat pancreas in the presence of cyclic AMP and somatostatin.

In order to study the role of cyclic AMP in the inhibition by somatostatin of glucose-induced insulin release, the effect of somatostatin on the potentiation by dibutyryl-cyclic AMP (db-cAMP) of insulin release from isolated pancreatic islets of rats was examined. Isolated islets were obtained from the rat pancreas by the collagenase method. Ten islets were incubated for periods of 30 min in Krebs-Ringer bicarbonate buffer containg albumin and glucose 2.0 mg/ml in the presence or absence of somatostatin (1 microgram/ml or 100 ng/ml) and/or db-cAMP 1 mM. Glucose-induced insulin release was reduced by somatostatin in concentrations of 1 microgram/ml. Somatostatin in a concentration of 100 ng/ml significantly abolished the potentiation by db-cAMP of insulin release (p less than 0;01), in spite of exerting no inhibition of glucose-induced insulin release. However, in the presence of theophylline 5 mM, somatostatin 100 ng/ml did not show that inhibitory effect on the potentiated insulin release.     

Overnight inhibition of insulin secretion restores pulsatility and proinsulin/insulin ratio in type 2 diabetes

Impaired insulin secretion in type 2 diabetes is characterized by decreased first-phase insulin secretion, an increased proinsulin-to-insulin molar ratio in plasma, abnormal pulsatile insulin release, and heightened disorderliness of insulin concentration profiles. In the present study, we tested the hypothesis that these abnormalities are at least partly reversed by a period of overnight suspension of beta-cell secretory activity achieved by somatostatin infusion. Eleven patients with type 2 diabetes were studied twice after a randomly ordered overnight infusion of either somatostatin or saline with the plasma glucose concentration clamped at approximately 8 mmol/l. Controls were studied twice after overnight saline infusions and then at a plasma glucose concentration of either 4 or 8 mmol/l. We report that in patients with type 2 diabetes, 1) as in nondiabetic humans, insulin is secreted in discrete insulin secretory bursts; 2) the frequency of pulsatile insulin secretion is normal; 3) the insulin pulse mass is diminished, leading to decreased insulin secretion, but this defect can be overcome acutely by beta-cell rest with somatostatin; 4) the reported loss of orderliness of insulin secretion, attenuated first-phase insulin secretion, and elevated proinsulin-to-insulin molar ratio also respond favorably to overnight inhibition by somatostatin. The results of these clinical experiments suggest the conclusion that multiple parameters of abnormal insulin secretion in patients with type 2 diabetes mechanistically reflect cellular depletion of immediately secretable insulin that can be overcome by beta-cell rest.     

Islet amyloid polypeptide (IAPP;amylin) influences the endocrine but not the exocrine rat pancreas.

The effect of synthetic rat amylin (10,100,1000 pmol/l) on glucose (10 mmol/) and arginine (10 mmol/l) -stimulated islet hormone release from the isolated perfused rat pancreas and on amylase release from isolated pancreatic acini was investigated. Amylin stimulated the insulin release during the first (+76%) and the second secretion period (+42%) at 1 nmol/l. The first phase of the glucagon release was inhibited concentration dependently by amylin and completely suppressed during the second phase. Amylin diminished the somatostatin release in a concentration dependent manner. This effect was more pronounced at the first than the second secretion period (1 nmol amylin: 1 phase: -60%, 2.phase: -22%). Amylin was without any effect on basal and CCK stimulated amylase release from isolated rat pancreatic acini. Our data suggest amylin, a secretory product of pancreatic B-cells, as a peptide with approximately strong paracrine effects within the Langerhans islet. Therefore, amylin might be involved in the regulation of glucose homeostasis.     

Effects of exogenous somatostatin and insulin on islet amyloid polypeptide (amylin) release from perfused rat pancreas.

This study examined the effects of exogenous somatostatin and insulin on the release of islet amyloid polypeptide (IAPP), or amylin, from the isolated perfused rat pancreas. Somatostatin inhibited the release of both amylin and insulin from the perfused pancreas to the same extent. The infusion of 10 nM somatostatin resulted in 40% inhibition of the secretion of both amylin and insulin induced by 11.1 mM glucose and 10 mM arginine, and this inhibition was significantly increased to 70% by the infusion of 100 nM somatostatin (p less than 0.05). The amylin/insulin molar ratios remained constant at 0.8% and were not changed by the infusion of somatostatin. On the other hand exogenous insulin at a concentration of 1.8 nM did not affect the release of amylin induced by 11.1 mM glucose and 10 mM arginine, whereas 180 nM insulin slightly, although not significantly, inhibited the release of amylin by 15%. These findings suggest that the release of amylin may be negatively regulated by somatostatin and that circulating insulin may have no direct effect on the release of amylin at least at a physiological concentration.     

Pulses of somatostatin release are slightly delayed compared with insulin and antisynchronous to glucagon

It was early proposed that somatostatin-producing delta-cells in pancreatic islets have local inhibitory effects on the release of insulin and glucagon. Recent observations that pulses of insulin and glucagon are antisynchronous make it important to examine the temporal characteristics of glucose-induced somatostatin release. Analysis of 30 s fractions from the perfused rat pancreas indicated that increase of glucose from 3 to 20 mmol/l results in initial suppression of somatostatin release followed by regular 4-5 min pulses. During continued exposure to 20 mmol/l glucose, the pulses of somatostatin overlapped those of insulin with a delay of 30 s. Somatostatin and glucagon pulses were coupled in antisynchronous fashion (phase shift 2.4+/-0.2 min), supporting the idea that the delta-cells have a local inhibitory effect on glucagon release. It was possible to remove the pulses of somatostatin and glucagon with maintenance of the insulin rhythmicity by addition of 1 micromol/l of the P2Y(1) receptor antagonist MRS 2179.     

The effect of calcium on somatostatin inhibition of insulin release and cyclic AMP production in mouse pancreatic islets.

The effect of somatostatin on glucose-induced insulin secretion and cyclic AMP accumulation in isolated islets from obese, hyperglycemic ob/ob mice was studied in a microperifusion system. The normal biphasic pattern of insulin release as well as the inhibitory pattern of insulin release produced by somatostatin (0.5--1 microgram/ml) was matched by similar changes in the intracellular concentration of cyclic AMP. When islets were stimulated by glucose (3 mg/ml) plus 3-isobutyl-1-methylxanthine (0.1 mM), somatostatin (0.5 microgram/ml) failed to inhibit insulin secretion or cyclic AMP formation in the second phase whereas in the first phase both parameters were significantly reduced by somatostatin (0.5 microgram/ml). In batch-type incubations it was shown that addition of excess calcium (to 6 mM) reversed this inhibition. In the second phase calcium potentiated the (glucose + 3-isobutyl-1-methylxanthine)-stimulated insulin secretion without affecting the cyclic AMP production. This potentiation was inhibited by somatostatin (0.1 microgram/ml). Somatostatin (1 microgram/ml) inhibited adenylate cyclase activity in islet homogenates. No effect of somatostatin on islet glucose utilization could be demonstrated. The results indicate a dual action of somatostatin in the inhibition of insulin release, one involving the islet adenylate cyclase and one affecting the islet uptake of calcium.     

Mechanism of action of growth-hormone-releasing hormone in stimulating insulin secretion in vitro from isolated rat islets and dispersed islet cells

Human growth-hormone-releasing hormone [(1-44)NH2] (hGHRH) was a potent stimulus for insulin release from rat islets of Langerhans in vitro; the optimum concentration used was 10(-11) M. The dose response curves for hGHRH effects on insulin secretion were notably different in intact islets of Langerhans compared to cultured dispersed islet cells. Pancreatic islets responded to a very low hGHRH concentration (10(-12) M), but at a higher hGHRH concentration (10(-9) M) no stimulation of insulin release was observed. When somatostatin antiserum was included in the incubation medium, hGHRH (10(-9) M) stimulated insulin release from intact islets. In cultured dispersed islet cells, which are principally insulin-secreting B cells, hGHRH directly and potently stimulated insulin release even at a concentration of 10(-9) M. Addition of somatostatin (10(-7), 10(-8) M) significantly reduced the hGHRH-induced insulin-secretory responses of dispersed islet cells. hGHRH (10(-11)-10(-9) M) raised islet cAMP levels; individually, hGHRH and theophylline exerted positive effects on insulin release, their combined effect was greater than that caused by either one. We conclude that hGHRH directly affects insulin secretion in vitro by a cAMP-dependent mechanism, and that the difference in responses of intact islets versus islet cells to increasing concentrations of hGHRH may be related to hGHRH-induced release of somatostatin in intact rat islets.     

The effect of calcium on somatostatin inhibition of insulin release and cyclic AMP production in mouse pancreatic islets

The effect of somatostatin on glucose-induced insulin secretion and cyclic AMP accumation in isolated islets from obese, hyperglycemic ob/ob mice was studied in a microperifusion system. The normal biphasic pattern of insulin release as well as the inhibitory pattern of insulin release produced by somatostatin (0.5–1 μg/ml) was matched by similar changes in the intracellular concentration of cyclic AMP. When islets were stimulated by glucose (3 mg/ml) plus 3-isobutyl-1-methylxanthine (0.1 mM), somatostatin (0.5 μg/ml) failed to inhibit insulin secretion or cyclic AMP formation in the second phase whereas in the first phase both parameters were significantly reduced by somatostatin (0.5 μg/ml). In batch-type incubations it was shown that addition of excess calcium (to 6 mM) reversed this inhibition. In the second phase calcium potentiated the (glucose + 3-isobutyl-1-methylxanthine)-stimulated insulin secretion without affecting the cyclic AMP production. This potentiation was inhibited by somatostatin (0.1 μg/ml). Somatostatin (1 μg/ml) inhibited adenylate cyclase activity in islet homogenates. No effect of somatostatin on islet glucose utilization could be demonstrated.The results indicate a dual action of somatostatin in the inhibition of insulin release, one involving the islet adenylate cyclase and one affecting the islet uptake of calcium.     

Difference in dose-response curves for glucose-induced insulin and somatostatin release in rat pancreas

We studied the glucose dependence of insulin and somatostatin release from rat pancreata, which were perfused in vitro in the presence of 3-isobutyl-1-methylxanthine (IBMX; 0.5 mM). Half-maximal insulin release occurred at approx. 12 mM glucose, and half-maximal somatostatin release at approx. 7 mM glucose.