D. Hillaire-Buys

Purinergic receptors on insulin-secreting cells.

Summary— The insulin secreting B cell is fitted with the two types of purinergic receptors: P2 (for ATP and/or ADP) and P1 (for adenosine). The activation of P2 purinoceptors by ATP or ADP evokes a biphasic stimulation of insulin secretion from isolated perfused rat pancreas; this stimulation is dose‐dependent between 10−6 and 10−4 M. Non hydrolysable structural analogues are also effective, and the relative potency of various agonists (2‐methylthio ATP ≫ ATP = ADP = α, β‐methylene ATP ≫ AMP) gave evidence for a P2y purinoceptor subtype. Proposed mechanisms include both an increased Ca2+ uptake and an increased intracellular Ca2+ mobilization via the hydrolysis of polyphosphoinositides. ATP (or ADP) potentiates physiological insulin‐secreting agents (glucose and acetylcholine) and P2 purinoceptors could play a physiological role in the stimulation of insulin secretion. The activation of P1 purinoceptors (adenosine receptors) decreases insulin secretion. Using structural analogues of adenosine, the receptor was characterized as an A1 subtype; it is coupled to a pertussis toxin sensitive G protein and it inhibits adenylate cyclase. It is of physiological relevance that the B cell has the two types of purinoceptors with opposite effects. Recently, a metabolically stable structural analogue of ADP, adenosine‐5′‐0‐(2‐thiodiphosphate) or ADPßS, has been described as a potent secretory agent, effective at nanomolar concentrations on isolated perfused rat pancreas. In vivo, this substance is able to increase insulin secretion and to improve glucose tolerance after IV administration in rats and oral administration in dogs. Furthermore in streptozotocin‐induced diabetes, ADPßS retains its insulin secreting effects. These results suggest that P2y purinoceptors could be a new target for antidiabetic drugs.

Evidence for an inhibitory A1 subtype adenosine receptor on pancreatic insulin-secreting cells

The effects of L- and D-phenylisopropyladenosine (L- and D-PIA) were studied on glucose-induced insulin secretion from the isolated perfused rat pancreas. L-PIA at the low dose of 16.5 nM inhibited insulin secretion by 50%. In contrast, D-PIA at 16.5 and 82.5 nM was ineffective. D-PIA used at a 100-fold higher concentration (1.65 microM) than L-PIA induced a similar inhibition of insulin secretion. The inhibitory effect of L-PIA was abolished by 8-phenyltheophylline (1 microM), a potent P1 purinoceptor antagonist. The present experiments provide evidence for an adenosine receptor of the A1 subtype on the insulin-secreting pancreatic cell of rats.

Stimulation of insulin secretion and improvement of glucose tolerance in rat and dog by the P2y‐purinoceptor agonist, adenosine‐5′‐O‐(2‐thiodiphosphate)

1 In vivo effect of a P2y‐purinoceptor agonist, adenosine‐5′‐O‐(2‐thiodiphosphate) (ADPβS), on insulin secretion and glycaemia were studied both in rats and dogs. 2 In anaesthetized rats, i.v. administered ADPβS (0.2 mg kg−1) produced an insulin response dependent on the nutritional state of the animals, since we observed only a transient increase in overnight‐fasted rats and a sustained insulin secretion followed by a reduction in plasma glucose levels in fed rats. During an i.v. glucose tolerance test, ADPβS enhanced insulin release and thus increased the glucose disappearance rate. 3 In anaesthetized fasted dogs, i.v. administered ADPβS (0.1 mg kg−1) increased pancreaticoduodenal insulin output and slightly decreased blood glucose levels. 4 In conscious fasted dogs, orally administered ADPβS (0.1 mg kg−1) transiently increased insulinemia and punctually reduced glycaemia. Furthermore, during an oral glucose tolerance test, orally administered ADPβS at the same dose markedly enhanced insulin secretion and consequently reduced the hyperglycaemia. 5 In conclusion, the P2y‐agonist, ADPβS, is a potent insulin secretagogue in vivo, improves glucose tolerance and is effective after oral administration. Thus, the P2y‐purinoceptors of the β cell may be a target for new antidiabetic drugs.

Effects of homocysteine on acetylcholine‐ and adenosine‐induced vasodilatation of pancreatic vascular bed in rats

1 Epidemiological and experimental data have shown that homocysteine may provoke vascular lesions and that moderate homocysteinaemia may constitute an independent risk factor for vascular disease. It is now documented that homocysteine damages human endothelial cells in culture, possibly by producing hydrogen peroxide in an oxygen‐dependent reaction. 2 In this study, we have examined the direct effect of this sulphur amino acid on pancreatic vascular resistance. Experiments were performed on the vascular bed of the rat isolated pancreas perfused at constant pressure; thus, any change in pancreatic vascular resistance resulted in a change in the flow rate. D,L‐Homocysteine perfused for one hour at three different concentrations (200 μM, 2 mM, 20 mM) did not induce any significant change in the flow rate per se. Homocysteine infusion for 30 min at a concentration of 200 μM or 2 mM abolished the endothelium‐dependent vasodilatation induced by acetylcholine (0.05 μM), but did not modify adenosine (1.5 μM)‐induced vasodilatation. 3 The effect of D,L‐homocysteine (200 μM or 2 mM) cannot be ascribed to a direct antimuscarinic effect since 30 min pretreatment of rat ileum with these concentrations did not significantly change the contractile effect of increasing concentrations of acetylcholine (0.015–15 μM). 4 Preincubation of human umbilical vein endothelial cells with D,L‐homocysteine (0.2–5.0 mM) had no significant effect on overall cell number or viability during 18 h of incubation; the endothelial cells exposed to concentrations up to 5 mM exhibited a spindle‐shaped, whirled pattern. This pattern was reversed 48 h after the removal of homocysteine. A cytotoxic effect was seen after 18 h incubation in 10 mM D,L‐homocysteine. 5 In conclusion, an acute infusion of homocysteine altered acetylcholine endothelium‐induced vasodilatation, whereas the adenosine vasodilatator effect was insensitive to the deleterious action of homocysteine in vitro.

Alterations of insulin response to different β cell secretagogues and pancreatic vascular resistance induced by Nω-nitro-L-arginine methyl ester

1 We studied a possible interplay of pancreatic NO synthase activity on insulin secretion induced by different β cell secretagogues and also on pancreatic vascular bed resistance. 2 This study was performed in the isolated perfused pancreas of the rat. Blockade of NO synthase was achieved with Nω‐nitro‐L‐arginine methyl ester (l‐NAME); the specificity of the antagonist was checked by using its D‐enantiomer as well as by substitutive treatments with sodium nitroprusside (SNP) as a NO donor in studies of glucose‐induced insulin secretion. 3 Arginine (5 mM) induced a monophasic insulin response which was, in the presence of L‐NAME at equimolar concentration, very strongly potentiated and converted into a 13 times higher biphasic one. D‐NAME (5 mM) was only able to induce a 3 times higher response, but provoked a similar vasoconstrictor effect. 4 The small biphasic insulin secretion induced by L‐leucine (5 mM) was also strongly enhanced, by 8 times, in the presence of L‐NAME (5 mM) vs 2 times in the presence of D‐NAME (5 mM). 5 β cell responses to KC1 (5 mM) and tolbutamide (0.185 mM) were only slightly increased by L‐NAME (5 mM) to values not far from the sum of the effects of L‐NAME and of the two drugs alone. D‐NAME (5 mM) was totally ineffective on the actions of both secretagogues. 6 L‐NAME, infused 15 min before and during a rise in glucose concentration from 5 to 11 mM, was able in the low millimolar range (0.1‐0.5 mM) to blunt the classical biphasic pattern of β cell response to glucose and, at 5 mM, to convert it into a significantly greater monophasic one. In contrast, D‐NAME (5 mM) was unable to induce similar effects. 7 SNP alone at 3 μm was ineffective but at 30 μm substantially reduced the second phase of insulin response to glucose; however, at both concentrations the NO donor partly reversed alterations in insulin secretion caused by L‐NAME (5 mM) and restored a biphasic pattern of insulin response. At a high (300 μm) concentration, SNP drastically reduced the second phase of β cell response, but in the presence of L‐NAME, provoked a significantly greater biphasic response. 8 When L‐NAME was infused only for the 15 min before high glucose, an exaggerated first phase of β cell response was followed by an abortive second one. SNP, at a low concentration (30 nM), given simultaneously with L‐NAME, restored a biphasic pattern and prevented the vasoconstrictor effect induced by the inhibitor. 9 L‐NAME, when infused only during high glucose, markedly enhanced the second phase of insulin response which could be significantly reduced by SNP (3 μm). The NO donor induced a dilator effect significantly greater in L‐NAME‐treated pancreata than in non‐treated ones. 10 In conclusion our data bring evidence that NO synthase activity exerts an inhibitory control on pancreatic β cell response to various nutrient secretagogues and may, at least partly, be implicated in the generation of the biphasic pattern of insulin response to glucose.

Effects of α-adrenoceptor agonists and antagonists on insulin secreting cells and pancreatic blood vessels: comparative study

The effects of alpha-adrenergic drugs were studied on glucose-induced insulin secretion and effluent flow rate on the same preparation: the isolated perfused rat pancreas. An alpha 1-adrenoceptor agonist, phenylephrine 0.05 microM slightly decreased insulin secretion (-25%); this inhibition was counteracted by an alpha 2-adrenoceptor antagonist, yohimbine 0.6 microM. Phenylephrine evoked a fall in liquid flow rate (-13%) which was reversed by an alpha 1-adrenoceptor antagonist, prazosin 6 microM, but not by yohimbine. An alpha 2-adrenoceptor agonist, clonidine 0.01 and 0.05 microM decreased insulin secretion (-80%). This inhibition was reversed by yohimbine 0.6 and 6 microM respectively. Only the concentration of 0.05 microM clonidine evoked a fall (-25%) in liquid flow rate; this fall was counteracted by yohimbine 0.6 microM. In conclusion our results show that adrenergic inhibition of insulin secretion is mediated only by alpha 2-receptors whereas both types of adrenoceptors are implicated in the vasoconstrictor effect. The insulin inhibitory effect of adrenoceptor agonists is not related to vasoconstriction.

Mechanisms involved in the effect of nitric oxide synthase inhibition on L-arginine-induced insulin secretion

1 A constitutive nitric oxide synthase (NOSc) pathway negatively controls L‐arginine‐stimulated insulin release by pancreatic β cells. We investigated the effect of glucose on this mechanism and whether it could be accounted for by nitric oxide production. 2 NOSc was inhibited by N∞‐nitro‐L‐arginine methyl ester (l‐NAME), and sodium nitroprusside (SNP) was used as a palliative NO donor to test whether the effects of L‐NAME resulted from decreased NO production. 3 In the rat isolated perfused pancreas, L‐NAME (5 mM) strongly potentiated L‐arginine (5 mM)‐induced insulin secretion at 5 mM glucose, but L‐arginine and L‐NAME exerted only additive effects at 8.3 mM glucose. At 11 mM glucose, L‐NAME significantly inhibited L‐arginine‐induced insulin secretion. Similar data were obtained in rat isolated islets. 4 At high concentrations (3 and 300 μm), SNP increased the potentiation of arginine‐induced insulin output by L‐NAME, but not at lower concentrations (3 or 30 nM). 5 L‐Arginine (5 mM) and L‐ornithine (5 mM) in the presence of 5 mM glucose induced monophasic β cell responses which were both significantly reduced by SNP at 3 nM but not at 30 nM; in contrast, the L‐ornithine effect was significantly increased by SNP at 3 μm. 6 Simultaneous treatment with L‐ornithine and L‐arginine provoked a biphasic insulin response. 7 At 5 mM glucose, L‐NAME (5 mM) did not affect the L‐ornithine secretory effect, but the amino acid strongly potentiated the alteration by L‐NAME of L‐arginine‐induced insulin secretion. 8 L‐Citrulline (5 mM) significantly reduced the second phase of the insulin response to L‐NAME (5 mM) + L‐arginine (5 mM) and to L‐NAME + L‐arginine + SNP 3 μm. 9 The intermediate in NO biosynthesis, NG‐hydroxy‐L‐arginine (150–300 μm) strongly counteracted the potentiation by L‐NAME of the secretory effect of L‐arginine at 5 mM glucose. 10 We conclude that the potentiation of L‐arginine‐induced insulin secretion resulting from the blockade of NOSc activity in the presence of a basal glucose concentration (1) is strongly modulated by higher glucose concentrations, (2) is not due to decreased NO production but (3) is probably accounted for by decreased levels of NG‐hydroxy‐L‐arginine or L‐citrulline, resulting in the attenuation of an inhibitory effect on arginase activity.

P2y purinoceptor responses of β cells and vascular bed are preserved in diabetic rat pancreas

1 To investigate the effect of experimental diabetes on the P2y purinoceptor responses of pancreatic β‐cells and vascular bed, we used adenosine‐5′‐O‐(2‐thiodiphosphate) (ADPβS), a potent and stable P2y agonist. This work was performed in the isolated perfused pancreas of the rat. 2 Diabetes was induced by streptozotocin (66 mg kg−1, i.p.). Five weeks after the induction of diabetes, on the day of pancreas isolation, the animals displayed marked hyperglycaemia (37.6 ± 2.7 mm). Age‐matched rats were used as controls. 3 Insulin response to a glucose stimulation from 5 to 10 mm was completely lost and stimulation of insulin release by the sulphonylurea, tolbutamide (185 μm), was drastically impaired in the diabetic pancreas (maximum responses were 1.5 ± 0.4 and 7.0 ± 1.4 ng min−1 for diabetic and age‐matched rats respectively). 4 In contrast, in the diabetic pancreas ADPβS (15 μm), infused in the presence of glucose 5 mm, elicited an immediate and significant insulin release similar to that observed in the age‐matched pancreas (maximum responses were 7.6 ± 1.5 and 6.7 ± 1.3 ng min−1 respectively). This ADPβS stimulating effect occurred independently of the glucose concentration (5, 8.3 and 28 mm) in the diabetic pancreas. On pancreatic vascular resistance, ADPβS induced a similar vasodilatation in diabetic and age‐matched rats. 5 In conclusion, ADPβS retains its insulin stimulatory and vasodilator effects in experimental diabetes; P2y purinoceptors could therefore be considered as a new target for the development of antidiabetic drugs.

Effect of pertussis toxin on A1-receptor-mediated inhibition of insulin secretion

Previous studies have provided evidence for the presence on B cell membrane of adenosine receptors (P1‐purinoceptors) of the A1‐subtype which inhibit insulin secretion. In this work we have investigated the implication of a guanosine triphosphate (GTP) binding protein (G protein) in the A1 purinoceptor‐induced inhibition of insulin secretion from the isolated perfused pancreas of the rat. A group of rats was treated with pertussis toxin (10 μgkg−1, i.v.) 28 h prior to pancreas extirpation. This treatment totally abolished the 50% decrease in insulin secretion induced by (+)‐N6‐phenylisopropyl adenosine (1.65 μm), a P1‐purinoceptor agonist. These results indicate that the A1‐receptor‐mediated inhibition of insulin secretion involves a pertussis toxin‐sensitive G protein.

Comparative effects of adenosine-5'-triphosphate and related analogues on insulin secretion from the rat pancreas

Summary— Adenosine tri‐ and diphosphate (ATP and ADP) and their structural analogues stimulate insulin secretion from the isolated perfused rat pancreas, an effect mediated by P2Y‐purinoceptor activation. Concerning the base moiety of the nucleotide, it was previously shown that purine but not pyrimidine nucleoside triphosphates were active and that substitution on purine C2 with the 2‐methylthio group greatly enhanced the potency. In this study, we further analyze the consequences of ribose and polyphosphate chain modifications. Modifications in 2′ and 3′ position on the ribose led to a decrease in insulin response when bulky substitutions were made: indeed, 2′‐deoxy ATP was similar in activity to ATP, whereas arylazido‐aminopropionyl ATP (ANAPP3) was weakly effective and trinitrophenyl ATP (TNP‐ATP) was inactive. Substitution on the /phosphorus of the triphosphate chain led to a decrease (y‐anilide ATP) or no change (y‐azido ATP) in potency; the replacement of the bridging oxygen between β and /phosphorus by a peroxide group did not significantly change the activity, whereas substitution by a methylene group completely abolished stimulation of insulin secretion. As for the phosphorothioate analogues, adenosine‐5′‐0‐(3‐thiotriphosphate) (ATP)S) induced an insulin response similar to that produced by ATP, whereas adenosine‐5′‐0‐(2‐thiodiphosphate) (ADP/JS) was about 100‐fold more potent than ATP, as previously shown. In conclusion, two structural features seem to have a strategic importance for increasing the insulin secretory activity of ATP analogues: substitution at the C2 position on the adenine ring of ATP and modifications of the polyphosphate chain at the level of the β phosphorus.