The insulin-suppressive effect of resveratrol - an in vitro and in vivo phenomenon

Resveratrol, a naturally occurring phytoalexin, is known to exert numerous beneficial effects in the organism. Literature data indicate that this compound may, among other effects, play a role in prevention of diabetes and diabetic complications. Resveratrol was recently found to affect insulin secretion in vitro and to change blood insulin concentrations. These effects are, however, not fully elucidated. In the present study, 1, 10 and 100microM resveratrol incubated for 90min with pancreatic islets isolated from normal rats failed to affect basal insulin release, but substantially impaired the secretory response to physiological and maximally effective glucose. In depolarized islets exposed to resveratrol, succinate-induced insulin secretion was also diminished. The blockade of somatostatin receptors substantially enhanced insulin secretion induced by 6.7mM glucose and simultaneously suppressed the inhibitory effect of 1microM resveratrol, but at 10 and 100microM, resveratrol was still able to attenuate hormone secretion. Acetylcholine clearly increased the insulin-secretory response to 6.7mM glucose and canceled the inhibitory effect of 1microM resveratrol. However, resveratrol at concentrations 10 and 100microM strongly decreased insulin secretion. The direct activation of protein kinase C totally suppressed the inhibitory influence of 1 and 10microM resveratrol on hormone secretion. However, activation of this enzyme appeared to be insufficient to cancel the insulin-suppressive effect of 100microM resveratrol. These data indicate that resveratrol-induced inhibition of insulin secretion may be partially mitigated by suppression of somatostatin action, activation of protein kinase C or the presence of acetylcholine. The in vivo experiment revealed that resveratrol, administered to normal rats at the dose 50mg/kg body weight, diminished blood insulin concentrations at 30min, without concomitant changes in glycemia. These observations point to the direct insulin-suppressive action of resveratrol in the rat.     

Effects of muscarinic receptor type 3 knockout on mouse islet secretory responses

The impact of muscarinic type 3 receptor knockout (M3KO) on the cholinergic regulation of insulin secretion and phospholipase C (PLC) activation was determined. Islets isolated from control, wild-type mice or heterozygotes responded with comparable insulin secretory responses to 15 mM glucose. This response was markedly amplified by the inclusion of 10 microM carbachol. While 15 mM glucose-induced release remained similar to wild-type and heterozygote responses in M3KO mice, the stimulatory impact of carbachol was abolished. Stimulation with 15 mM glucose plus 50 microM carbachol increased fractional efflux rates of myo-[2-3H]inositol from control wild-type and heterozygote islets but not from M3KO islets. Fed plasma insulin levels of M3KO mice were reduced 68% when compared to values obtained from combined wild-type and heterozygote animals. These studies support the conclusion that the M3 receptor in islets is coupled to PLC activation and insulin secretion and that cholinergic stimulation of the islets may play an important role in the regulation of plasma insulin levels.     

Insulin secretion reactivation of pancreatic islets from starved rats in vitro.

Starvation of Wistar rats induced a shift of glucose threshold for insulin secretion of isolated islets above 5 mM, which can be restored by pretreatment of the tissue with glucose, mannose, glyceraldehyde, an theophylline, but not with acetylcholine or lactate. The improved insulin secretion is not connected with an enhanced glucose utilization.     

5-amino-imidazole carboxamide riboside acutely potentiates glucose-stimulated insulin secretion from mouse pancreatic islets by KATP channel-dependent and -independent pathways

AMP-activated protein kinase (AMPK) is an important signaling effector that couples cellular metabolism and function. The effects of AMPK activation on pancreatic beta-cell function remain unresolved. We used 5-amino-imidazole carboxamide riboside (AICAR), an activator of AMPK, to define the signaling mechanisms linking the activation of AMPK with insulin secretion. Application of 300 microM AICAR to mouse islets incubated in 5-14 mM glucose significantly increased AMPK activity and potentiated insulin secretion. AICAR inhibited ATP-sensitive K(+) (K(ATP)) channels and increased the frequency of glucose-induced calcium oscillations in islets incubated in 8-14 mM glucose. At lower glucose concentration (5mM) AICAR did not affect K(ATP) activity or intracellular ([Ca(2+)](i)). AICAR also did not inhibit (86)Rb(+) efflux from islets isolated from Sur1(-/-) mice that lack K(ATP) channels yet significantly potentiated glucose stimulated insulin secretion. Our data suggest that AICAR stimulates insulin secretion by both K(ATP) channel-dependent and -independent pathways.     

Effects of combined secretagogues and extracellular calcium on neonatal insulin release

The insulin response of 3-day old neonatal rat islets was evaluated following a 1 h incubation with glucose alone and in the presence of 30 nM sulfated cholecystokinin octapeptide (CCK) and/or 20 microM carbachol (CCh). Insulin secretion was found to be incrementally increased from the lowest glucose concentration and enhanced several fold in the presence of CCK and/or CCh. In combination, CCK and CCh increased glucose-stimulated insulin secretion by an amount equivalent to the sum of their individual increases. The presence of either CCK alone or CCK plus CCh increased phosphoinositide hydrolysis by the same relative amount that they increased insulin secretion when compared to 8.3 mM glucose. Glucose-stimulated insulin secretion was totally inhibited when calcium was omitted from the incubation buffer; this effect was partially negated by CCK alone and more so by CCK combined with CCh. Insulin secretion in response to 8.3 mM glucose alone was unchanged when calcium in the incubation buffer was increased from 1 to 5 mM; however, the insulin response to 16.7 mM glucose alone and 8.3 mM glucose in the presence of CCK and/or CCh was increased under this condition. Thus, we have shown that, even at 3 days postpartum, insulin secretion from isolated islets is a complex response capable of being molded by several secretagogues at once and ultimately determined by interplay of different signaling systems activated.     

Effects of antisomatostatin and antiglucagon sera on insulin release from isolated Langerhans' islets of rats

Effects of antisomatostatin and antiglucagon sera on insulin secretion were studied in isolated pancreatic islets of rats. Addition of antisomatostatin or antiglucagon serum to the incubation medium containing 5.5 mM glucose significantly increased insulin secretion from the islets. In contrast, these effect were not apparent when the glucose concentration was elevated to 20 mM. In the presence of antiglucagon serum the free glucagon concentration in the medium was decreased, while the total glucagon was markedly increased, suggesting that the antiglucagon serum caused an increase in glucagon secretion. These results suggest that A- and D-cells may play an important role in regulating insulin release at physiologic glucose concentration.     

Characteristics of the biotin enhancement of glucose-induced insulin release in pancreatic islets of the rat

Perifused isolated rat islets were used to show that biotin plus 16.5 mM glucose evoked more insulin secretion than 16.5 mM glucose alone. Whether or not this reinforcement of glucose-induced insulin secretion by biotin is unique was studied by using perifused islets stimulated with 16.5 mM glucose plus 100 microM of one of various components of the vitamin B group. No effect of any of these vitamins was found on glucose-induced insulin secretion. These results indicate that biotin is unique among the members of the vitamin B group in enhancing glucose-induced insulin secretion. Static incubation experiments showed that biotin did not potentiate insulin release when the islets were incubated with an experimental solution containing either no or 2.8 mM glucose. The addition of biotin to 27.7 mM glucose, which is the maximal concentration for stimulating insulin release, did not significantly enhance the effect of the glucose on insulin release (although it did at 16.5 mM glucose). These findings indicate that biotin, by itself, does not stimulate insulin secretion, and does not enhance glucose-induced insulin secretion beyond the ability of glucose itself to stimulate insulin secretion.     

Ghrelin activates neuronal constitutive nitric oxide synthase in pancreatic islet cells while inhibiting insulin release and stimulating glucagon release

In view of our previous data, showing that ghrelin and nitric oxide (NO) display apparently parallel effects on insulin secretion (inhibitory) and glucagon secretion (stimulatory), we have now investigated the effect of ghrelin on islet hormone secretion in relation to its effect on NO synthase (NOS) isoenzymes in isolated rat pancreatic islets. Dose-response studies revealed that ghrelin at concentrations of 0.01-1 micromol l-1 inhibited insulin secretion stimulated by 8.3 mmol l-1 glucose, while ghrelin at concentrations lower than the physiological range (0.01 pmol l-1 to 1 nmol l-1) were without effect. In contrast, glucagon secretion was stimulated by 1.0 nmol l-1 to 1 micromol l-1 ghrelin. These effects of ghrelin on insulin and glucagon secretion were accompanied by increased NO production through activation of neuronal constitutive NOS (ncNOS). Ghrelin had no appreciable effect on the activity of inducible NOS (iNOS) in the islets. Addition of an NO scavenger (cPTIO) or the NOS inhibitor L-NAME to the incubation medium prevented the effects of ghrelin on hormone secretion from isolated islets. The present results confirm our previous data showing that ghrelin inhibits insulin and stimulates glucagon secretion from pancreatic islets of the mouse and we now show similar effects in rat islets. The effects of ghrelin were accompanied by an increased rate of NO production. Conceivably, ncNOS activation partly accounts for to the inhibitory effect of ghrelin on insulin secretion and the stimulatory effect of ghrelin on glucagon secretion.     

Augmented Insulinotropic Action of Arachidonic Acid through the Lipoxygenase Pathway in the Obese Zucker Rat

Objective: The metabolism of arachidonic acid (AA) has been shown to be altered in severe insulin resistance that is present in obese (fa/fa) Zucker rats. We examined the effects and mechanism of action of AA on basal and glucose‐stimulated insulin secretion in pancreatic islets isolated from obese (fa/fa) Zucker rats and their homozygous lean (Fa/Fa) littermates. Research Methods and Procedures: Islets were isolated from 10‐ to 12‐week‐old rats and incubated for 45 minutes in glucose concentrations ranging from 3.3 to 16.7 mM with or without inhibitors of the cyclooxygenase or lipoxygenase pathways. Medium insulin concentrations were measured by radioimmunoassay, and islet production of the 12‐lipoxygenase metabolite, 12‐hydroxyeicosatetraenoic acid (12‐HETE), was measured by enzyme immunoassay. Results: In islets from lean animals, AA stimulated insulin secretion at submaximally stimulatory glucose levels (< 11.1 mM) but not at 16.7 mM glucose. In contrast, in islets derived from obese rats, AA potentiated insulin secretion at all glucose concentrations. AA‐induced insulin secretion was augmented in islets from obese compared with lean rats at high concentrations of AA in the presence of 3.3 mM glucose. Furthermore, the inhibitor of 12‐lipoxygenase, esculetin (0.5 μM), inhibited AA‐stimulated insulin secretion in islets from obese but not lean rats. Finally, the islet production of the 12‐HETE was markedly enhanced in islets from obese rats, both in response to 16.7 mM glucose and to AA. Discussion: The insulin secretory response to AA is augmented in islets from obese Zucker rats by a mechanism related to enhanced activity of the 12‐lipoxygenase pathway. Therefore, augmented action of AA may be a mechanism underlying the adaptation of insulin secretion to the increased demand caused by insulin resistance in these animals.     

Successful cultivation of isolated islets of Langerhans without attachment: relationship between Glucose- and theophylline-induced insulin release and insulin content in rats islets after cultivation.

The effect of various inhibitors of insulin secretion such as mannoheptulose (20 mM), atropine (1 mM), diphenylhydantoin (20 microng/ml), high concentration of Mg++ (5.3 mM) in the presence of 20 mM glucose (control) on insulin content and secretion from collagenase-isolated rat pancreatic islets was studied in vitro by cultivation of islets up to 5 or 9 days in glass Petri dishes without attachment. In a following short-term incubation for 60 min the glucose-induced insulin release without and with theophylline (5 mM) was investigated. Islets cultivated at 5 mM glucose and at 20 mM glucose with the inhibitors mannoheptulose or atropine lost the responsiveness to glucose and theophylline whereas such islets cultivated at 20 mM glucose alone or with diphenylhydantoin (DPH) or 5.3 mg Mg++ showed a stimulation of insulin secretion by glucose and theophylline. Compared, however, with freshly isolated islets all cultivated islets were restricted in their maximal glucose response and this defect was not evoked alone by quantitative changes in islet insulin content. Nevertheless, culture conditions which facilitate a net increase of insulin (content and release) during cultivation influenced also positively the glucose-induced insulin release without and with 5 mM theophylline in the following short-term experiments.     

Parallel effects of arachidonic acid on insulin secretion, calmodulin-dependent protein kinase activity and protein kinase C activity in pancreatic islets.

A potential role of arachidonic acid in the modulation of insulin secretion was investigated by measuring its effects on calmodulin-dependent protein kinase and protein kinase C in islet subcellular fractions. The results were interpreted in the light of arachidonic acid effects on insulin secretion from intact islets. Arachidonic acid could replace phosphatidylserine in activation of cytosolic protein kinase C (K0.5 of 10 microM) and maximum activation was observed at 50 microM arachidonate. Arachidonic acid did not affect the Ca2+ requirement of the phosphatidylserine-stimulated activity. Arachidonic acid (200 microM) inhibited (greater than 90%) calmodulin-dependent protein kinase activity (K0.5 = 50-100 microM) but modestly increased basal phosphorylation activity (no added calcium or calmodulin). Arachidonic acid inhibited glucose-sensitive insulin secretion from islets (K0.5 = 24 microM) measured in static secretion assays. Maximum inhibition (approximately 70%) was achieved at 50-100 microM arachidonic acid. Basal insulin secretion (3 mM glucose) was modestly stimulated by 100 microM arachidonic acid but in a non-saturable manner. In perifusion secretion studies, arachidonic acid (20 microM) had no effect on the first phase of glucose-induced secretion but nearly completely suppressed second phase secretion. At basal glucose (4 mM), arachidonic acid induced a modest but reproducible biphasic insulin secretion response which mimicked glucose-sensitive secretion. However, phosphorylation of an 80 kD protein substrate of protein kinase C was not increased when intact islets were incubated with arachidonic acid, suggesting that the small increases in insulin secretion seen with arachidonic acid were not mediated by protein kinase C. These data suggest that arachidonic acid generated by exposure of islets to glucose may influence insulin secretion by inhibiting the activity of calmodulin-dependent protein kinase but probably has little effect on protein kinase C activity.     

Effect of Dexamethasone on Insulin Secretion: Examination of Underlying Mechanisms

ObjectiveTo study the mechanism of increased insulin secretion in response to short-term administration of dexamethasone.MethodsMale Wistar rats were injected intraperitoneally with dexamethasone (dexamethasone; 200 mcg/kg body weight per day) or saline for 3 consecutive days. Insulin secretion in response to glucose, ionomycin, and KCl was quantified in islets isolated from the animals, and the amount of glucokinase was measured by Western blot.ResultsDexamethasone-treated animals had 1.18-fold higher fasting blood glucose concentration and 6.5-fold increase in fasting serum insulin concentration compared with findings from animals injected with saline. Compared with islets isolated from control rats, islets from dexamethasone-treated rats secreted more insulin at 60 minutes in response to 5.5 mM glucose (416.4 vs 115.6 fmoles/10 islets, P = .011) and in response to 16.6 mM glucose (985.5 vs 520.6 fmoles/10 islets, P = .014); no change in insulin secretion was observed at 10 minutes. Insulin secretion from islets of dexamethasone-treated rats and control rats was not differentially augmented in response to either ionomycin or potassium chloride. Glucokinase expression was not altered by treatment with dexamethasone.ConclusionsAugmentation of insulin secretion in response to glucose in the pancreatic islets from dexamethasone-treated rats is preserved in islets studied in vitro. The increase in glucose-stimulated insulin secretion appears to be mediated by steps upstream to β -cell membrane depolarization and the attended increase in intracellular calcium in the signaling pathway of insulin secretion. (Endocr Pract. 2010;16:763-769)     

Oxygen and temperature dependence of stimulated insulin secretion in isolated rat islets of Langerhans

The effects of lowered O2 tension on insulin secretion and changes in cellular energy parameters were investigated in isolated rat pancreatic islets perifused with buffers equilibrated with 21, 9, 5, and 1% oxygen and containing 5 mM glucose. Decreasing the external [O2] reduced the amount of insulin released in response to 16 mM glucose, 20 mM alpha-ketoisocaproic acid, and 40 mM KCl. Secretion elicited by high glucose or KCl had declined significantly at 9% oxygen, whereas that caused by alpha-ketoisocaproic acid became inhibited below 5% O2. Lowering the oxygen tension also decreased the ability of islets to respond with a rise in [ATP]/[ADP] upon stimulation with metabolic secretagogues. This reduction in the evoked increase in the nucleotide ratios paralleled the inhibition of stimulated insulin secretion. Addition of 2 mM amytal markedly decreased the islet energy level and eliminated the secretory response to 16 mM glucose. The results suggest that enhancement of B-cell energy production and a consequent rise in [ATP] (or [ATP]/[ADP]) are a necessary event for the hormone release elicited by high glucose and alpha-ketoisocaproic acid. A decrease in temperature inhibited insulin secretion with all three secretagogues tested. The energies of activation were similar for high glucose and KCl-induced secretion, about 20 kcal/mol, but were higher for alpha-ketoisocaproic acid, about 35 kcal/mol. At 28 degrees C, the [ATP]/[ADP] was larger than that at 38 degrees C (8 versus 5) and was not increased further upon addition of 16 mM glucose. It is suggested that a decrease in the rate of energy production at lowered temperatures may contribute to the inhibition of insulin release caused by metabolic secretagogues.     

Acceleration by fructose of the ATP-sensitive K(+) channel-independent pathway of glucose-induced insulin secretion.

The mechanism with which fructose augments glucose-induced insulin secretion is still unclear. The present study was aimed at examining whether the ketohexose potentiates the ATP-sensitive K(+) channel-independent pathway of glucose-induced insulin secretion and, if so, how this happens. When isolated rat islets were depolarized by incubating them with 50 mM KCl in the presence of 150 microM diazoxide (an opener of ATP-sensitive K(+) channels), 10 mM glucose plus 20 mM fructose elicited significantly higher insulin secretion than 10 mM glucose alone, whereas 20 mM fructose alone did not stimulate insulin secretion. The fructose 1,6-bisphosphate and inositol trisphosphate contents were markedly higher in islets incubated with glucose plus fructose than in islets incubated with glucose alone. The results demonstrate that fructose has the ability to potentiate the ATP-sensitive K(+) channel-independent pathway of glucose-induced insulin secretion. The increase in fructose 1,6-bisphosphate content induced by the co-presence of fructose with glucose, resulting in the rise in inositol trisphosphate content, is likely to be one of the signals involved in the fructose potentiation of glucose-induced insulin secretion.     

Dual action of adiponectin on insulin secretion in insulin-resistant mice

Adiponectin is secreted by adipocytes and has been implicated as a mediator of insulin sensitivity. In this study, the acute effects of adiponectin on islets isolated from normal or diet-induced insulin resistant mice were examined. In normal islets, adiponectin (5 microg/ml) had no significant effect on insulin secretion. In contrast, in islets from mice rendered insulin resistant by high-fat feeding, adiponectin inhibited insulin secretion at 2.8 mM (P < 0.01) but augmented insulin secretion at 16.7 mM glucose (P < 0.05). The augmentation of glucose-stimulated insulin secretion by adiponectin was accompanied by increased glucose oxidation (P < 0.005), but without any significant effect on palmitate oxidation or the islet ATP/ADP ratio. Furthermore, RT-PCR revealed the expression of the adiponectin receptor AdipoR1 mRNA in mouse islets, however, with no difference in the degree of expression level between the two feeding groups. The results thus uncover a potential dual role for adiponectin to modify insulin secretion in insulin resistance.     

Effects of galnon, a non-peptide galanin-receptor agonist, on insulin release from rat pancreatic islets

Galanin is a neurotransmitter peptide that suppresses insulin secretion. The present study aimed at investigating how a non-peptide galanin receptor agonist, galnon, affects insulin secretion from isolated pancreatic islets of healthy Wistar and diabetic Goto-Kakizaki (GK) rats. Galnon stimulated insulin release potently in isolated Wistar rat islets; 100 microM of the compound increased the release 8.5 times (p<0.001) at 3.3 mM and 3.7 times (p<0.001) at 16.7 mM glucose. Also in islet perifusions, galnon augmented several-fold both acute and late phases of insulin response to glucose. Furthermore, galnon stimulated insulin release in GK rat islets. These effects were not inhibited by the presence of galanin or the galanin receptor antagonist M35. The stimulatory effects of galnon were partly inhibited by the PKA and PKC inhibitors, H-89 and calphostin C, respectively, at 16.7 but not 3.3 mM glucose. In both Wistar and GK rat islets, insulin release was stimulated by depolarization of 30 mM KCl, and 100 microM galnon further enhanced insulin release 1.5-2 times (p<0.05). Cytosolic calcium levels, determined by fura-2, were increased in parallel with insulin release, and the L-type Ca2+-channel blocker nimodipine suppressed insulin response to glucose and galnon. In conclusion, galnon stimulates insulin release in islets of healthy rats and diabetic GK rats. The mechanism of this stimulatory effect does not involve galanin receptors. Galnon-induced insulin release is not glucose-dependent and appears to involve opening of L-type Ca2+-channels, but the main effect of galnon seems to be exerted at a step distal to these channels, i.e., at B-cell exocytosis.     

Effects of lactate on pancreatic islets. Lactate efflux as a possible determinant of islet-cell depolarization by glucose.

The secretion of insulin from perifused rat pancreatic islets was stimulated by raising the glucose concentration from 5.6 to 20 mM or by exposure to tolbutamide. The addition of sodium lactate (40 mM) to islets perifused in the presence of glucose (5.6 mM) resulted in a small, transient, rise in the rate of secretion. The subsequent removal of lactate, but not glucose or tolbutamide, from the perifusate produced a dramatic potentiation of insulin release. The rate of efflux of 45Ca2+ was also increased when islets were exposed to a high concentration of glucose or lactate or to tolbutamide, and again subsequently upon withdrawal of lactate. Efflux of 86Rb+ was modestly inhibited upon addition of lactate and markedly enhanced by the subsequent withdrawal of lactate from islets. The output of [14C]lactate from islets incubated in the presence of [U-14C]glucose increased linearly with increasing concentrations of glucose (1-25 mM). It is proposed that the activation of islets by the addition or withdrawal of lactate is not due to increased oxidative flux, but occurs as a result of the electrogenic passage of lactate ions across the plasma membrane, resulting in islet-cell depolarization, Ca2+ entry and insulin secretion. The production of lactate via the glycolytic pathway, and the subsequent efflux of lactate from the islet cells with concomitant exchange of H+ for Na+, could be a major determinant of depolarization and hence insulin secretion, in response to glucose.     

Glucose regulates glucokinase activity in cultured islets from rat pancreas

In this study, we have used isolated pancreatic islets cultured for 7 days in 3 or 30 mM glucose to explore whether glucokinase is induced or activated by high glucose concentrations and has related enzyme activity to glucose-stimulated insulin release. Islets cultured in low glucose medium or low glucose medium plus 350 ng/ml insulin did not respond to high glucose stimulation. Islets cultured in medium containing high glucose concentrations showed a high rate of basal insulin secretion when perifused with 5 mM glucose, and the insulin release was greatly augmented in a biphasic secretion profile when the glucose concentration was raised to 16 mM. Islet glucokinase and hexokinase activities were determined by a sensitive and specific fluorometric method. Glucokinase activity was reduced to approximately 50% in islets cultured in low glucose medium with or without insulin present compared to results with fresh islets. However, islets cultured in 30 mM glucose showed that glucokinase activity was elevated to 236% compared to results with fresh islets. It is concluded that (a) glucose is the physiological regulator of glucokinase in the islet of Langerhans and that (b) the activity of glucokinase plays a crucial role in glucose-induced insulin secretion.     

Tetracaine stimulates insulin secretion through the mobilization of Ca2+ from thapsigargin- and IP3-insensitive Ca2+ reservoir in pancreatic beta-cells

The effect of tetracaine on 45Ca efflux, cytoplasmic Ca2+ concentration [Ca2+]i, and insulin secretion in isolated pancreatic islets and beta-cells was studied. In the absence of external Ca2+, tetracaine (0.1-2.0 mM) increased the 45Ca efflux from isolated islets in a dose-dependentOFF efflux caused by 50 mM K+ or by the association of carbachol (0.2 mM) and 50 mM K+. Tetracaine permanently increased the [Ca2+]i in isolated beta-cells in Ca2+-free medium enriched with 2.8 mM glucose and 25 microM D-600 (methoxiverapamil). This effect was also observed in the presence of 10 mM caffeine or 1 microM thapsigargin. In the presence of 16.7 mM glucose, tetracaine transiently increased the insulin secretion from islets perfused in the absence and presence of external Ca2+. These data indicate that tetracaine mobilises Ca2+ from a thapsigargin-insensitive store and stimulates insulin secretion in the absence of extracellular Ca2+. The increase in 45Ca efflux caused by high concentrations of K+ and by carbachol indicates that tetracaine did not interfere with a cation or inositol triphosphate sensitive Ca2+ pool in beta-cells.