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

Available online at www.sciencedirect.com Regulatory Peptides 144 (2007) 43 – 49 www.elsevier.com/locate/regpep Pulses of somatostatin release are slightly delayed compared with insulin and antisynchronous to glucagon Albert Salehi a , Saleem S. Qader a , Eva Grapengiesser b,⁎, Bo Hellman b a Department of Clinical Science, CRC (UMAS), University of Lund, SE-20502 Malmö, Sweden b Department of Medical Cell Biology, University of Uppsala, SE-75123 Uppsala, Sweden Received 10 October 2006; received in revised form 11 April 2007; accepted 14 June 2007 Available online 21 June 2007 Abstract It was early proposed that somatostatin-producing δ-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 δ-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 μmol/l of the P2Y1 receptor antagonist MRS 2179. © 2007 Elsevier B.V. All rights reserved. Keywords: Somatostatin pulse; MRS 2179; Pancreas perfusion; Purinoceptor 1. Introduction The temporal and spatial control of biological processes is often accomplished in an oscillatory manner [1,2]. Many cellular events, including the release of pancreatic islet hormones, are triggered by increase of the cytoplasmic Ca2+ concentration ([Ca2+]i). Oscillatory rise of [Ca2+]i is an intrinsic property of the cells producing insulin [3], glucagon [4,5] and somatostatin [4,6,7]. These oscillations are mediated by repetitive depolarization with subsequent entry of Ca2+ via voltage-activated channels. A prerequisite for pulsatile release of the islet hormones into the portal vein is entrainment of the secretory Ca2+ signal into a common rhythm. In the case of insulin, the intra-islet synchronisation of the β-cells is mediated both via gap junctions [8] and diffusible messengers like ATP [9]. The coordinating action of ATP has been attributed to IP3-induced activation of ⁎ Corresponding author. Tel.: +46 18 4714424; fax: +46 18 4714059. E-mail address: eva.grapengiesser@mcb.uu.se (E. Grapengiesser). 0167-0115/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.regpep.2007.06.003 short-lived Ca2+ transients, which temporarily interrupt the voltage-dependent entry of Ca2+ by activating a repolarizing K+ current [10]. Pulsatile release of insulin requires both coordination of the β-cells within the islets and that the cells appear in the same oscillatory phase in all islets of the pancreas. The entrainment of the islets into a common rhythm is supposed to be mediated by neural activity, critically dependent on ATP activation of P2Y1 receptors [11]. It was early proposed that δ-cells (equivalent to α1-cells) have local inhibitory effects on the release of insulin from adjacent β-cells [12]. Subsequently, the active factor was identified as somatostatin and the paracrine influence supposed to involve both the insulin-producing β-cells and the glucagonproducing α-cells [13,14]. Although much attention has been paid to intra-islet effects of somatostatin, it is still uncertain how the release of somatostatin is temporally related to insulin and glucagon [15–19]. We recently reported that pulses of insulin and glucagon have similar duration but appear with a phase shift of one half [20]. Due to this antisynchrony the maximal glucagon effect on liver cells is manifested during periods with minimal exposure to insulin. 44 A. Salehi et al. / Regulatory Peptides 144 (2007) 43–49 pulses induced by glucose and 5 + 5 for studying the effects of 1 and 10 μmol/l MRS 2179, respectively). The animals were allowed free access to food. Before perfusion of pancreas, the rats were anesthetized with 5% chloral hydrate (1 ml/100 g body weight). The experiments were approved by a local ethics committee. 2.2. Chemical and solutions Reagents of analytical grade and deionized water were used. Roche Diagnostics (Mannheim, Germany) supplied collagenase. The P2Y1 receptor antagonist 2-deoxy-N-methyladenosine 3,5-bisphosphate (MRS 2179) was a product of Tocris Cookson (Bristol, U.K.). The studies were performed at Fig. 1. Comparison of the initial somatostatin response (filled symbols) with that for insulin (panel A) and glucagon (panel B) after raising the glucose concentration from 3 to 20 mmol/l during perfusion of rat pancreas. Perfusate from the portal vein was collected as 30 s fractions. Mean values ± SEM for 15 experiments. The purpose of the present study was to examine whether glucose-induced pulses of somatostatin release depend on P2Y1 receptors and temporally relate to those of insulin and glucagon. After measuring somatostatin in perfusates of rat pancreas, previously analyzed for insulin [21] and glucagon [20], we now report that somatostatin pulses are essentially in phase with insulin but antisynchronous to glucagon. Moreover, it was found that MRS 2179, an antagonist of the P2Y1 receptor, under certain conditions suppresses the somatostatin and glucagon release pulses with maintenance of the insulin rhythmicity. 2. Materials and methods 2.1. Animals Fifteen female Sprague–Dawley rats, weighing 250–300 g, were included in the study (5 for analysis of repetitive hormone Fig. 2. Repetitive pulses of somatostatin release generated by increase of glucose from 3 to 20 mmol/l during perfusion of pancreas. Oscillatory patterns for each of 5 rats are shown. Perfusate from the portal vein was collected as 30 s fractions and duplicate samples taken for analysis. The vertical bars represent internal standard deviations calculated from the assay duplicates within the experiment. The “ Cluster Analysis” program revealed sustained rhythmicity (P b 0.01) in all experiments. The marks on the ordinate indicate 0 pmol/l somatostatin for the trace above and/or 100 pmol/l for the trace below. A. Salehi et al. / Regulatory Peptides 144 (2007) 43–49 37 °C with a basal medium of Krebs Ringer bicarbonate buffer (pH 7.4), supplemented with 10 mmol/l HEPES and gassed with 5% CO2 + 95% O2 to obtain constant pH and oxygenation. Medium concentrations of somatostatin were measured in duplicate with radioimmunoassay, using rabbit antiserum raised against synthetic cyclic somatostatin 14 (Euro-Diagnostica AB, Malmö, Sweden). The coefficients of intra- and inter-assay variations were 4.4 and 7.1% respectively. 45 Unrecycled medium supplemented with 2% BSA (wt/vol) was infused at a rate of 0.4 ml/min. After 5 min of equilibration, the glucose concentration was raised from 3 to 20 mmol/l. The effluent from the portal vein was collected during 20 min as 30 s fractions in heparinized vials, followed by centrifugation and storage of the plasma at − 20 °C. Testing the reversibility of the effect of MRS 2179 the perfusion was prolonged. After omission of the inhibitor, sampling was interrupted for 5 min followed by collection of 4 fractions. 2.3. Perfusion of pancreas 2.4. Analysis of data Perfusion was performed with a previously described protocol [20], which allows contribution of blood from the donor (25%, estimated by comparing the glucose concentration in the effluent with that in the circulating blood). Statistically significant hormone pulses were identified with the “Cluster Analysis” program, a computerized pulse analysis algorithm [22]. Each sample was assigned a dose-dependent Fig. 3. Comparison of glucose-induced pulses of somatostatin release with pulses of insulin (panel A) and glucagon (panel B) in rat 2 from Fig. 2. The somatostatin pulses follow those of insulin. Due to antisynchrony the rise of somatostatin is related to decrease of glucagon and the decrease of somatostatin to rise of glucagon. 46 A. Salehi et al. / Regulatory Peptides 144 (2007) 43–49 standard deviation calculated from the actual duplicates within the experiment. A sliding pooled t-test was then performed to identify data points within the time series that correspond to statistically significant increases and decreases of the hormone concentrations. Test cluster sizes for the nadir and the peak widths were assigned to 2. The minimum t-statistics was specified to 4.1 for upstrokes and downstrokes respectively. These settings detected peaks with b1% false positive errors. The repetitive pulses of somatostatin after the initial peak were compared with those of insulin and glucagon using cross correlation analysis of data from the 5 rats studied in Fig. 2. The r-values were determined in each subject at each of multiple lags. The standard deviation of r was estimated by 1/(n − k)1/2 where n is the number of sample pairs and k is the number of lag units [23]. Results are presented as mean values ± SEM. Differences were statistically evaluated with t-test. 3.2. Glucose induction of repetitive somatostatin pulses During continued exposure to 20 mmol/l glucose the initial somatostatin peak was followed by repetitive pulses with an amplitude of 33 ± 4 pmol/l and a periodicity of 4.6 ± 0.1 min. Oscillatory patterns for each of 5 pancreatic glands are shown in Fig. 2. The pulsatile release of somatostatin was closely related to that of insulin (Fig. 3A). Comparison with cross correlation analyses of hormone concentrations in 30 s samples, taken during the period of continued rhythmicity, indicated that the somatostatin pulses lag slightly behind those of insulin (Fig. 4A). The cross correlation coefficients (r) were higher at 30 s than at zero lag in each of 5 experiments. The temporal relation between repetitive somatostatin and glucagon pulses is shown in Fig. 4B. The pulsatile somatostatin release was antisynchronous to that of glucagon with a delay of 2.4 ± 0.1 min. 3. Results 3.1. Initial somatostatin response to glucose 3.3. Disappearance of somatostatin pulses after inhibition of purinoceptors The somatostatin concentration was 11 ± 1 pmol/l (n = 9) in the presence of 3 mmol/l glucose. Increase of the glucose concentration to 20 mmol/l resulted in lowering of somatostatin to 6 ± 1 pmol/l after 30–60 s (P b 0.001), followed by a steep rise to 56 ± 9 pmol/l (P b 0.001) after 2 min (Fig. 1). The initial reduction of somatostatin coincided with increase of both insulin (Fig. 1A) and glucagon (Fig. 1B). Whereas insulin peaked together with somatostatin after 2 min, the glucagon release was maximal after 1 min and then decreased dramatically with rise of somatostatin. Our previous studies have shown that pulsatile release of insulin and glucagon is critically dependent on P2Y1 receptors. It was therefore tested how the purinoceptor antagonist MRS 2179 affected the continued rhythmicity of somatostatin. In the presence of 1 μmol/l there was a disappearance of somatostatin and glucagon pulses with maintenance of the insulin pulses (Fig. 5). Added at a concentration of 10 μmol/l, MRS 2179 removed the pulsatility of all 3 hormones (not shown). The suppressive action of MRS 2179 on somatostatin release was reversible (Fig. 5). Comparing the highest and lowest Fig. 4. Median cross correlation coefficients (r-values) between somatostatin release pulses and the pulses of insulin (panel A) and glucagon (panel B) recorded after the initial peak in the 5 rats shown in Fig. 2. Somatostatin is slightly delayed compared to insulin with the highest positive cross correlation at − 30 s. The highest negative correlation between somatostatin and glucagon is seen at zero time. Asterisks above or below the series of r-values indicate P b 0.001. A. Salehi et al. / Regulatory Peptides 144 (2007) 43–49 Fig. 5. Effects of MRS 2179 on pulses of somatostatin (panel A), insulin (panel B) and glucagon (panel C) release simultaneously recorded during perfusion of a rat pancreas with 20 mmol/l glucose. Perfusate from the portal vein was collected as 30 s fractions and duplicate samples were taken for analysis. After omission of MRS 2179 the sampling was interrupted for 5 min followed by collection of 4 fractions. ⁎Pulses statistically verified by the “Cluster Analysis” program (P b 0.01). Addition of 1 μmol/l MRS 2179 removed somatostatin and glucagon pulses with maintenance of insulin rhythmicity. Representative for all of 5 experiments. concentration of somatostatin during 2 min periods before and 5 min after omission of 1 μmol/l MRS 2179, the range was found to increase from 2.8 ± 0.3 to 15.3 ± 2.6 pmol/l (P b 0.01). 4. Discussion Somatostatin is a polypeptide with a short biological half-life (about 30 s) produced both in brain and peripheral organs. Somatostatin, released from the islet δ-cells, has been proposed to act as a paracrine inhibitor of the secretory activity in 47 glucagon-producing α-cells and insulin-producing β-cells [12,14]. The effects of somatostatin are mediated by high affinity interactions with a family of G-protein coupled receptors. Immunofluorescence studies have shown that all five receptor subtypes known so far are expressed in rat islets [24]. However, the expression is not uniform. It is possible to inhibit either insulin or glucagon release using combinations of agonists selective for different subtypes of the somatostatin receptor [25]. In variance with the report of independent release cycles for the major islet hormones [19], we now demonstrate that somatostatin pulses are related both to insulin and glucagon. Absence of previous studies with sufficient time resolution precludes a direct comparison of the present data with observations made in man. It may well exist that species differences in the order of the pulses are dependent on the position of the somatostatin-producing δ-cells within the islet. In mice, but not in humans, the δ-cells are situated between peripheral α-cells and more centrally situated β-cells [26]. Although it has been reported that fluorescent dye is transferred within seconds between the δ- and β-cells [27], imaging of [Ca2+]i does not support the existence of heterologous gap junctional coupling [28]. Electrophysiological patch clamp recordings in intact mouse islets have convincingly demonstrated that coupling via gap junctions is essentially restricted to the β-cells [29]. Discussing the mechanisms for intra-islet effects of somatostatin, attention should be paid to observations that δ-cells have processes extending towards the capillaries [30]. Different opinions have been expressed whether islet microcirculation promotes paracrine flow from β-cells to non-β-cells [18,31], or whether the order is the opposite [32]. There is also a third alternative where the contraction of capillary endothelial cells shunts the blood to islet regions containing a mixed population of endocrine cells [33]. The latter model allows different types of autoregulatory feed back loops. Apart from a local inhibitory action of δ-cells on the secretory activity of α- and β-cells, it has been reported that glucagon stimulates [34] and insulin inhibits [35,36] somatostatin release. Moreover glutamate, released from α-cells during exocytosis of glucagon, can stimulate the somatostatin secretion in the presence of low concentrations of glucose [37]. We now observe that increase of glucose from 3 to 20 mmol/ l results in pronounced suppression (45%) of somatostatin release followed by rise to a peak coinciding with that of insulin. The mechanism for this reduction is open for discussion. Under certain conditions increase of the glucose concentration results in temporary suppression of insulin release, a phenomenon regarded to reflect lowering of [Ca2+]i [38]. Interestingly, the initial suppression of somatostatin release was simultaneous with increase of glucagon, supporting a role for δ-cells for paracrine inhibition of the secretory activity of α-cells. Like other hormone-producing islet cells, the δ-cell has an intrinsic ability to generate oscillations of cytoplasmic Ca2+[4,6]. Entrained into a common rhythm these oscillations will account for the pulsatility of somatostatin release reported in dogs [15–17], monkeys [16,18,19] and man [16]. In one of these 48 A. Salehi et al. / Regulatory Peptides 144 (2007) 43–49 studies 9–10 min cycles of somatostatin were reported to coincide with pulses of insulin [17]. With improved time resolution we now demonstrate that somatostatin pulses have a duration of 4–5 min and are phase shifted compared with insulin and glucagon. In relation to insulin the delay was only 30 s. Infusion of exogenous somatostatin inhibits the pulsatile component of insulin secretion from dog pancreas [39]. Nevertheless, we now observe that the somatostatin pulses mainly overlap those of insulin. The phase difference between the somatostatin and glucagon pulses (2.4 ± 0.2 min) corresponded to half the peak interval. Due to this antisynchrony, periods with rise of somatostatin will be coupled to decrease of glucagon. The existence of such a temporal relationship adds to the arguments that endogenous release of somatostatin is a key regulator of pancreatic α-cells. Activation of P2 purinoceptors is known to enhance the release of somatostatin from dog pancreas during perfusion with medium containing low or intermediate concentrations of glucose [40,41]. The non-hydrolysable P2Y1 receptor agonist ADP-β-S (1 μmol/l) stimulates somatostatin secretion with maintenance of basal insulin and glucagon release [41]. Moreover, the P2Y1 receptor antagonist MRS 2179 [42,43] effectively inhibits (IC50 ≈ 0.3 μmol/l) the transients of [Ca2+]i supposed to entrain the pancreatic β-cells into a common rhythm [9]. We now observe that MRS 2179 suppresses the release of somatostatin into the portal vein. This purinoceptor antagonist has previously been found to counteract the pulsatility of insulin [21] and glucagon [20] release. However, the present study reveals marked differences in the sensitivity to MRS 2179. At a concentration of 1 μmol/l both the somatostatin and glucagon pulses disappeared with maintenance of the insulin rhythmicity. Accordingly, generation of somatostatin pulses is not a prerequisite for cyclic variations of insulin. In conclusion, glucose-induced pulses of somatostatin release closely follow insulin but are 180° out of phase with glucagon. The antisynchrony between somatostatin and glucagon reinforces previous arguments that δ-cells have inhibitory effects on α-cells. It is possible to remove both the somatostatin and glucagon pulses with maintenance of the insulin rhythmicity by antagonizing the P2Y1 receptors. Acknowledgements This study was supported by grants from the Swedish Research Council (72X-562, 12X-6240, 04X-20029), the Swedish Diabetes Association, The Novo Nordic Fund, the Albert Påhlsson Foundation, the Crafoord Foundation, and the Family Ernfors Foundation. We thank Dr Michael Johnson, Pharmacological Department, University of Virginia Health Science Center, for providing the software. The technical assistance of Britt-Marie Nilsson is gratefully acknowledged. References [1] Rapp PE. Why are so many biological systems periodic? Prog Neurobiol 1987;29:261–73. [2] Kruse K, Jülicher F. Oscillations in cell biology. Curr Opin Cell Biol 2005;17:20–6. [3] Grapengiesser E, Gylfe E, Hellman B. Glucose-induced oscillations of cytoplasmic Ca2+in the pancreatic β-cell. Biochem Biophys Res Commun 1988;151:1299–304. [4] Berts A, Gylfe E, Hellman B. Ca2+oscillations in pancreatic islet cells secreting glucagon and somatostatin. Biochem Biophys Res Commun 1995;208:644–9. [5] Liu YJ, Vieira E, Gylfe E. A store-operated mechanism determines the activity of the electrically excitable glucagon-secreting pancreatic α-cell. Cell Calcium 2004;35:357–65. [6] Berts A, Liu YJ, Gylfe E, Hellman B. Oscillatory Ca2+signaling in somatostatin-producing cells from the human pancreas. Metabolism 1997;46:366–9. [7] Berts A, Ball A, Dryselius S, Gylfe E, Hellman B. Glucose stimulation of somatostatin-producing islet cells involves oscillatory Ca2+signaling. Endocrinology 1996;137:693–7. [8] Ravier MA, Guldenagel M, Charollais A, Gjinovci A, Caille D, Sohl G, Wollheim CB, Willecke K, Henquin JC, Meda P. Loss of connexin36 channels alters β-cell coupling, islet synchronization of glucose-induced Ca2+and insulin oscillations, and basal insulin release. Diabetes 2005;54:1798–807. [9] Hellman B, Dansk H, Grapengiesser E. Pancreatic β-cells communicate via intermittent release of ATP. Am J Physiol Endocrinol Metabol Gastrointest Physiol 2004;286:E759–65. [10] Grapengiesser E, Dansk H, Hellman B. Synchronization of pancreatic β-cell rhythmicity after glucagon induction of Ca2+transients. Cell Calcium 2003;34:49–53. [11] Grapengiesser E, Dansk H, Hellman B. Pulses of external ATP aid to the synchronization of pancreatic β-cells by generating premature Ca2+oscillations. Biochem Pharmacol 2004;68:667–74. [12] Hellman B, Lernmark Å. Inhibition of the in vitro secretion of insulin by an extract of pancreatic a1-cells. Endocrinology 1969;84:1484–8. [13] Unger RH, Orci L. Possible roles of the pancreatic D-cell in the normal and diabetic states. Diabetes 1977;26:241–4. [14] Efendic S, Lins PE, Luft R, Uvnäs-Wallensten K, Szecowska J. Somatostatin: paracrine or endocrine substance? Front Horm Res 1980;7:41–51. [15] Stagner J, Samols E, Weir G. Sustained oscillations of insulin, glucagon and somatostatin from the isolated pancreas during exposure to a constant glucose concentration. J Clin Invest 1980;65:939–42. [16] Jaspan JB, Lever E, Polonsky KS, Van Cauter E. In vivo pulsatility of pancreatic islet peptides. Am J Physiol 1986;251:E215–26. [17] Matthews DR, Hermansen K, Connolly AA, Gray D, Schmitz O, Clarc A, Orskov H, Turner RC. Greater in vivo than in vitro pulsatility of insulin secretion with synchronized insulin and somatostatin secretory pulses. Endocrinology 1987;120:2272–8. [18] Samols E, Stagner JI. Islet somatostatin — microvascular, paracrine and pulsatile regulation. Metabolism 1990;39:55–60. [19] Goodner C, Koerker D, Stagner JI, Samols E. In vitro pancreatic hormonal pulses are less regular and more frequent than in vivo. Am J Physiol 1991;260:E422–9. [20] Grapengiesser E, Salehi A, Qader SS, Hellman B. Glucose induces glucagon release pulses antisynchronous with insulin and sensitive to purinoceptor inhibition. Endocrinology 2006;147:3472–7. [21] Salehi A, Qader SS, Grapengiesser E, Hellman B. Inhibition of purinoceptors amplifies glucose-stimulated insulin release with removal of its pulsatility. Diabetes 2005;54:2126–31. [22] Veldhuis JD, Johnson ML. Cluster analysis: a simple, versatile and robust algorithm for endocrine pulse detection. Am J Physiol Endocrinol Metab 1986;250:E486–93. [23] Mulligan T, Iranmanesh A, Johnson ML, Straume M, Veldhuis JD. Aging alters feed-forward and feed-back linkages between LH and testosterone in healthy men. Am J Physiol 1997;273:R1407–13. [24] Ludvigsen E, Olsson R, Stridsberg M, Janson ET, Sandler S. Expression and distribution of somatostatin receptor subtypes in the pancreatic islets of mice and rats. J Histochem Cytochem 2004;52:391–400. [25] Ludvigsen E, Stridsberg M, Taylor JE, Culler MD, Öberg K, Janson ET, Sandler S. Regulation of insulin and glucagon secretion from rat pancreatic islets in vitro by somatostatin analogues. Regul Pept 2007;138:1–9. A. Salehi et al. / Regulatory Peptides 144 (2007) 43–49 [26] Alm G, Hellman B. Distribution of the two types of A-cells in the pancreatic islets of some mammalian species. Acta Endocrinol (Copenh) 1964;40:307–16. [27] Michaels RL, Sheridan JD. Islets of Langerhans: dye coupling among immunocytochemically distinct cell types. Science 1981;214:801–3. [28] Nadal A, Quesada I, Soria B. Homologous and heterologous asynchrony between identified α-, β- and δ-cells within intact islets of Langerhans in the mouse. J Physiol 1999;517:85–93. [29] Göpel S, Kanno T, Barg S, Galvanovskis J, Rorsman P. Voltage-gated and resting membrane currents recorded from B-cells in intact mouse pancreatic islets. J Physiol 1999;521:717–28. [30] Hellerström C, Hellman B, Peterson B, Alm G. The two types of pancreatic A-cells and their relation to the glucagon secretion. In: Brolin SE, Hellman B, Knutson H, editors. The structure and metabolism of the pancreatic islets. Oxford: Pergamon Press; 1964. p. 117–30. [31] Bonner-Weir S, Orci L. New perspectives on the microvasculature of the islets of Langerhans in the rat. Diabetes 1982;31:883–9. [32] Murakami T, Fujita T. Microcirculation of the rat pancreas with special reference to the insulo-acinar portal and insulin-venous draining systems: a further scanning electron microscope study of corrosion casts. Arch Histol Cytol 1992;55:453–76. [33] Brunicardi FC, Stagner JI, Bonner-Weir S, Wayland H, Kleinman R, Livingston E, Guth P, Menger M, McCuskey R, Intaglietta M, Charles A, Ashley S, Cheung A, Ipp E, Gilman S, Howard T, Passaro E. Microcirculation of the islets of Langerhans. Long Beach Veterans Administration Regional Medical Education Center Symposium. Diabetes 1996;45:385–92. [34] Hermansen K. Secretion of somatostatin from the normal and diabetic pancreas. Diabetologia 1980;19:492–504. 49 [35] Gerber PPG, Trimble ER, Wollheim CB, Renold AE. Effect of insulin on glucose and arginine-stimulated somatostatin secretion from the isolated perfused rat pancreas. Endocrinology 1981;109:279–83. [36] Ribes G, Gross R, Chenon D, Loubatieres-Mariani MM. Effect of insulin on basal pancreaticoduodenal output of somatostatin in normal and diabetic dogs. Acta Endocrinol (Copenh) 1988;109:43–50. [37] Muroyama A, Uehara S, Yatsushiro S, Echigo N, Morimoto R, Morita M, Hayashi M, Yamamoto A, Koh DS, Moriyama Y. A novel variant of ionotropic glutamate receptor regulates somatostatin secretion from δ-cells of islets of Langerhans. Diabetes 2004;53:1743–53. [38] Hellman B. Ca2+balance as a determinant for glucose-stimulated insulin release. Diabetologia 1985;28:494–501. [39] Pørksen N, Munn SR, Steers JL, Veldhuis JD, Butler PC. Effects of somatostatin on pulsatile insulin secretion: selective inhibition of insulin burst mass. Am J Physiol 1996;270:E1043–9. [40] Bertrand G, Gross R, Ribes G, Loubatieres-Mariani MM. P2purinoceptor agonists stimulate somatostatin secretion from dog pancreas. Eur J Pharmacol 1990;182:369–73. [41] Hillaire-Buys D, Gross R, Pares-Herbute N, Ribes G, Loubatieres-Mariani MM. In vivo and in vitro effects of adenosine-5'-O-(2-thiodiphosphate) on pancreatic hormones in dogs. Pancreas 1994;9:646–51. [42] Boyer JL, Mohanram A, Camaioni E, Jacobson KA, Harden TK. Competitive and selective antagonism of P2Y1 receptors by N6-methyl 2'-deoxyadenosine 3',5'-bisphosphate. Br J Pharmacol 1998;124:1–3. [43] Brown SG, King BF, Kim YC, Burnstock G, Jacobson KA. Activity of novel adenine nucleotide derivatives as agonists and antagonists at recombinant P2X receptors. Drug Dev Res 2000;49:253–9.