Possible involvement of A2A and A3 receptors in modulation of insulin secretion and β-cell survival in mouse pancreatic islets

General and Comparative Endocrinology 187 (2013) 86–94

Contents lists available at SciVerse ScienceDirect

General and Comparative Endocrinology journal homepage: www.elsevier.com/locate/ygcen

Possible involvement of A2A and A3 receptors in modulation of insulin secretion and b-cell survival in mouse pancreatic islets M. Ohtani a,⇑, T. Oka b, K. Ohura a a b

Department of Pharmacology, Osaka Dental University, 8-1 Kuzuhahanazono-cho, Hirakata, Osaka 573-1121, Japan Central Research Institute, Wakunaga Pharmaceutical Co. Ltd., 1624 Shimokotachi, Koda-cho, Akitakata, Hiroshima 739-1195, Japan

a r t i c l e

i n f o

Article history: Received 12 November 2012 Revised 7 February 2013 Accepted 12 February 2013 Available online 27 February 2013 Keywords: Adenosine receptor Cell viability Insulin secretion Pancreatic b-cells

a b s t r a c t Adenosine A1, A2A, A2B and A3 receptor mRNAs were found to be expressed in mouse pancreatic islets and Beta-TC6 cells but their physiological or pharmacological actions are not fully clarified. We showed that adenosine (100 lM) augmented insulin secretion by islets in the presence of either normal (5.5 mM) or a high concentration of glucose (20 mM). The augmentation of insulin secretion in the presence of high glucose was blocked by an A2A antagonist, but not by A2B and A3 antagonists, while an A1 antagonist potentiated the adenosine effect. An adenosine analogue 50 -N-ethylcarboxamidoadenosine (NECA) as well as A1, A2A and A3 receptor agonists also produced stimulation. On the other hand, an A3 agonist markedly reduced Beta-TC6 cell proliferation and the islet cell viability, while adenosine and NECA did not. The effect of A3 agonist was partially blocked by the A3 antagonist. In addition, treatment with the A3 agonist produced a small but significant extent of apoptosis in Beta-TC6 cells as judged by terminal transferasemediated deoxyuridine triphosphate nick end labeling (TUNEL) assay. These results combined together suggested that like the A1 receptor, activation of A2A receptors by adenosine results in augmented insulin secretion, while the A3 receptor is involved in modulation of the survival of pancreatic b-cells. Ó 2013 Published by Elsevier Inc.

1. Introduction Adenosine exerts its diverse physiological effects via adenosine receptors on the cell membrane in a variety of tissues (Fredholm et al., 2011; Jacobson, 2009; Jacobson et al., 1999). For example, it has been shown that adenosine receptors are involved in cardioprotective or cerebroprotective effects in response to stress (Jacobson, 2009) as well as in the survival or apoptosis of muscle and immune cells (Jacobson et al., 1999). Adenosine receptors are subdivided into A1, A2A, A2B and A3 receptor subtypes and are coupled to G-proteins. Activation of the A1 and A3 receptors linked to Gi proteins results in the reduction of adenylate cyclase activity, leading to a decrease in the intracellular level of cyclic AMP (cAMP), while stimulation of the A2A and A2B receptors coupled to Gs proteins increases the cAMP level (Jacobson, 2009). In addition, it has been shown that adenosine A2B and A3 receptors are also coupled to Gq/11 proteins and can activate the phospholipase C (PLC)/ inositol 1,4,5-triphosphate (IP3) pathway to mobilize Ca2+ from intracellular stores such as the endoplasmic reticulum (Ralevic and Burnstock, 1998). The selective ligands for each adenosine receptor subtype are promising candidates for the development of therapeutic drugs (Jacobson, 2009). For instance, A3 receptor agonists such as Cl-IB-MECA suppressing cell proliferation and ⇑ Corresponding author. Fax: +81 72 864 3158. E-mail address: [email protected] (M. Ohtani). 0016-6480/$ - see front matter Ó 2013 Published by Elsevier Inc. http://dx.doi.org/10.1016/j.ygcen.2013.02.011

inflammation have been proposed for the treatment of cancer and autoimmune inflammatory diseases such as rheumatoid arthritis (Jacobson, 2009; Jacobson et al., 1999). It has been reported that mRNAs of all adenosine receptor subtypes are expressed in rat pancreatic ducts (Novak et al., 2008) and mouse pancreas (Németh et al., 2007), although their presence in pancreatic islets is unknown. It has been also shown that A1 and A2 adenosine receptors are involved in modulation of acute pancreatitis (Satoh et al., 2002), exocrine secretion (Iwatsuki, 2000) and endocrine secretion of insulin (Hillaire-Buys et al., 1987, 1989; Zywert et al., 2011) or glucagon (Chapal et al., 1985). In addition, using mouse pancreatic islets Bertrand et al. (1989) showed that a high concentration of adenosine (500 lM) augmented insulin secretion and that its effect was blocked by p-nitrobenzylthioguanosine (NBTG), an inhibitor of adenosine transporters. On the other hand, it has been reported that the A1 receptor mediated the inhibitory action of adenosine analogues, L- and D-phenylisopropyladenosine (PIA) on insulin secretion by rat pancreas (Hillaire-Buys et al., 1987, 1989). The findings obtained with A1 receptor null mice (Johansson et al., 2007; Salehi et al., 2009) supported the involvement of A1 receptors in inhibiting insulin secretion. However, it is not clear whether activation of other adenosine receptor subtypes (A2A, A2B and A3 receptors) influences insulin secretion. It has also been demonstrated that adenosine modulates proliferation of various types of normal and tumor cells via activation of the adenylate cyclase/cAMP or PLC/IP3 pathway (Ohana

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

et al., 2001). These studies suggested the possibility that adenosine receptors may play an important role in the regulation of secretory activities and proliferation of islet endocrine cells under physiological or pathological conditions. We have used a mouse pancreatic b-cell line Beta-TC6 cells as a model to investigate the functional roles of various substances in insulin secretion (Ohtani et al., 2006;, 2008, 2009a, 2009b). We found that several types of receptor such as nicotinic (Ohtani et al., 2006, 2009a) and purinergic (Ohtani et al., 2008) receptors as well as polyamines (Ohtani et al., 2009b) are involved in modulation of a cytoplasmic Ca2+ concentration and glucose-induced insulin release in these cells. Recently, we found that ATP stimulated insulin secretion in mouse islets at low concentrations and inhibited proliferation, DNA synthesis and viability of pancreatic b-cells at a high concentration via the purinergic (P2Y1 and/or P2X) receptors (Ohtani et al., 2011). As an extension of our previous studies, we found the existence of adenosine receptor subtypes in pancreatic b-cells and obtained the evidence indicating that the A2A and A3 receptors were involved in modulation of insulin secretion and the survival of b-cells, respectively. 2. Materials and methods 2.1. Materials Male C57 BL/6 mice (7–10 week old) were used for all experiments and the study protocol using these animals was approved by the Ethics Committee of Osaka Dental University. Adenosine, 50 -N-ethylcarboxamidoadenosine (NECA), 8-cyclopentyl-1,3-dipropylxanthine (DPCPX), 5-amino-7-(2-phenylethyl)-2-(2-furyl)pyrazolo[4,3-e]-1,2,4-triazolo[1,5-c]-pyrimidine (SCH58261), N-(4cyanophenyl)-2-(4-(1,3-dipropylxanthin-8-yl)phenoxy)acetamide (MRS1754), N-(2-methoxyphenyl)-N-[2(3-pridyl)quinazolin-4-yl]urea (VUF5574), dipyridamole and S-(4-Nitrobenzyl)-6-thioinosine (NBTI) were purchased from the Sigma Chemical Company (St. Louis, MO, USA). N6-cyclopentyladenosine (CPA), 2-[p-2-(carbonylethyl)phenethylamino]-50 -N-ethylcarboxamidoadenosine (CGS21680) and 2-chrolo-N6-(3-iodobenzyl)-50 -N-methylcarboxamidoadenosine (Cl-IB-MECA) were gifts from Dr. Kenneth A. Jacobson (NIDDK, NIH). Dulbecco’s modified Eagle’s medium (DMEM), RPMI1640 medium, penicillin, streptomycin and 0.05% trypan blue staining solution were from Invitrogen Corp. (Carlsbad, CA, USA). Fetal bovine serum (FBS) was from BioWest (Nuaill, France). Polyclonal antibodies for Western blot analysis were from Cell Signaling Technology (Danvers, MA, USA). All reagents were dissolved in dimethyl sulfoxide (DMSO) and used at less than 0.4% volume for all experiments. 2.2. Reverse transcription-polymerase chain reaction (RT-PCR) The total RNA extraction from mouse islets or Beta-TC6 cells, reverse-transcription of RNA, amplification of complementary DNA (cDNA) by PCR and agarose gel electrophoresis were performed as described previously (Ohtani et al., 2008). The RNA (0.5 lg) was reverse-transcribed into cDNA and amplified by PCR with 40 (islets) or 35 (Beta-TC6 cells) cycles. The oligo primers for A1, A2A, A2B and A3 receptors were designed based on GenBank nucleotide sequences and synthesized commercially. The primers for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as an internal control were synthesized according to the sequences shown by Ohtani et al. (2011). These primer sequences were as follows: A1 (GenBank accession number; BC079624): forward 50 -CATTCCTCTGTCCACCCACT-30 , reverse 50 -CTGCCCACTCCTCTCTGTTC-30 , A2A (BC110692): forward 50 -CTATAGGGCCCCGAGTTAGG-30 , reverse 50 -GAGAGGACCGTCTGCAATTC-30 , A2B (BC116416): forward

87

50 -CCTTTGGCATTGGATTGACT-30 , reverse 50 -AAAATGCCCACGATCATAGC-30 , A3 (BC100416): forward 50 -ATGGCTATTCTTGGGCCTTT-30 , reverse 50 -AGGGTTCATCATGGAGTTCG-30 and GAPDH: forward 50 AGCCTCGTCCCGTAGACAAA-30 , reverse 50 -GAGATGATGACCCG TTTGGC-30 . Annealing temperatures for A1, A2A, A2B, A3 and GAPDH in PCR were 59, 59, 55, 61 and 62 °C, respectively. We found that mRNAs of all adenosine receptor subtypes were expressed in a mouse brain (data not shown), indicating that the primers designed in this study were operative. The level of expression of each adenosine receptor subtype in islets (Fig. 1C) and Beta-TC6 cells (Fig. 1D) was determined by semi-quantitative RT-PCR as described previously (Ohtani et al., 2011). Since the amount of all amplified PCR products reached the plateau with 45 (islets) or 40 (Beta-TC6) cycles, the density of each PCR product separated on agarose gel after reaction with 40 or 35 cycles, respectively, was analyzed by a luminoimage analyzer. 2.3. Insulin secretion assay The effect of adenosine receptor agonists and antagonists and nucleoside transporter inhibitors on insulin secretion by mouse pancreatic islets was examined as described previously (Ohtani et al., 2011), except that isolated islets were cultured for approximately one day in RPMI1640 medium containing 11 mM glucose before assay. Briefly, groups of 15 islets were incubated for 60 min at 37 °C in the presence of 5.5 or 20 mM glucose and the indicated agent. After incubation, the amount of insulin released from islets was determined by enzyme immunosorbent assay (EIA). 2.4. Cell proliferation assay Mouse Beta-TC6 cells, hamster HIT-T15 cells and rat RINm5F cells purchased from ATCC (Manassas, VA, USA) were cultured as described previously (Ohtani et al., 2011). Cells were seeded into a 24-well plate at a density of 1  105 cells per well and cultured for 24 h in medium containing 10% FBS. The cells were then treated with an adenosine receptor agonist for the indicated time at concentrations ranging from 0.1 to 300 lM. The A3 receptor antagonist was added 20 min prior to treatment with Cl-IB-MECA. The number of cells in each well was counted using a haemocytometer with the trypan blue exclusion method after cells were harvested by incubation with 0.05% trypsin/EDTA solution (Invitrogen) for 10 min at 37 °C. We ascertained that 0.5% DMSO per se did not affect the Beta-TC6 cell growth (n = 4). 2.5. DNA synthesis assay After Beta-TC6 cells (2  104 cells per well) were cultured in the absence (control) or presence of the indicated concentration of reagent for 72 h, the rate of DNA synthesis was measured as described previously (Ohtani et al., 2011), using a 5-bromo-20 deoxyuridine (BrdU) ELISA kit (Roche Diagnostic, Penzberg, Germany). In brief, the cells were treated with BrdU for 2 h and then fixed with 70% ethanol. Incorporated BrdU into cells was measured colorimetrically at a 370 nm wavelength with the anti-BrdU monoclonal antibody conjugated with peroxidase and its substrate tetramethylbenzidine. 2.6. Cell viability assay Beta-TC6 cells (1  104 cells) were seeded into a 96-well plate and cultured for 72 h in the presence of Cl-IB-MECA, 10% FBS and 25 mM glucose as described previously (Ohtani et al., 2011). Isolation of mouse islets, preparation of islet cell mixture and measurement of the viability of islet cells were performed as described

88

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

Fig. 1. RT-PCR analysis of adenosine receptor subtype mRNAs in mouse islets and Beta-TC6 cells. The total RNA extracted from (A) the islets or (B) Beta-TC6 cells was reversetranscribed into complementary DNA (cDNA). PCR reactions were performed with 40 (islets) or 35 (Beta-TC6 cells) cycles using specific primers for each subtype. The expected length of PCR product corresponding to the A1, A2A, A2B, or A3 receptor and GAPDH was 349, 375, 331, 457 and 379 bp, respectively. The lane, M shows a 100 bp DNA ladder. The relative amount of PCR product for each adenosine receptor subtype expressed in (C) islets or (D) Beta-TC6 cells was determined by semi-quantitative RT-PCR. The amount of each product corresponding to the A1, A2A, A2B or A3 receptor was expressed as the ratio relative to that of GAPDH. Each bar represents the mean ± S.E.M. (n = 3). Each experiment was repeated at least three times and gave similar results.

previously (Ohtani et al., 2011). Briefly, islet cells were cultured for 72 h in the presence of the indicated concentration of adenosine receptor agonist, 10% FBS and 5.5 or 25 mM glucose. The A3 receptor antagonist was added 20 min prior to treatment with Cl-IBMECA. After culture, the viability of the cells was measured with WST-1 reagent (Roche) containing tetrazolium salts, which were changed into formazan by mitochondrial succinate–tetrazolium reductase in viable cells and the amount of formazan was measured colorimetrically at a 450 nm wavelength.

Cells were covered with a cover slip and the TUNEL- or PI-positive fluorescent staining was detected with a confocal laser microscope (FV300, Olympus, Tokyo, Japan). As a positive control, apoptosis was induced by treating cells for 24 h with a sarcoplasmic/endoplasmic reticulum Ca2+-ATPase (SERCA) inhibitor thapsigargin (Sigma), which caused apoptosis in other insulinoma cells (Srinivasan et al., 2005; Tonnesen et al., 2009). We counted the number of TUNEL-positive cells in a microscopic area (0.22 mm2) picked up randomly from the total area (36 mm2) in Fig. 7B.

2.7. Western blot analysis

2.9. Statistics

Extraction of proteins with RIPA Lysis buffer (Santa Cruz Biotechnology, CA, USA) from Beta-TC6 cells treated with the indicated agent, SDS–PAGE, immunoblotting and detection of phosphorylated extracellular signal-regulated kinase (ERK) and caspase-3 proteins were performed as described previously (Ohtani et al., 2011). Briefly, a PVDF membrane blotted with the total proteins (20 lg per lane) was incubated with the rabbit anti-phosphorylated ERK1 and 2 (Thr 202/Tyr 204), total ERK1 and 2, caspase-3 or b-actin antibody each at 1:2,000 dilution overnight at 4 °C. After washing, the membrane was incubated with the anti-rabbit IgG antibody conjugated with horse radish peroxidase (simple stain MAX-PO, Nichirei, Tokyo, Japan) diluted at 1:10,000 for 60 min. The immunoreactive band was visualized with an ECL Advance™ detection reagent and VersaDoc 5000 (Bio-Rad). The density of each band was analyzed by the Quality One analysis software.

Data were expressed as mean ± S.E.M. Statistical significance was evaluated by Mann–Whitney test between two groups or one-way analysis of variance (ANOVA), followed by Bonferroni’s post hoc test among more than three groups. A P value < 0.05 was statistically significant in comparison to control or agonist alone.

2.8. A terminal transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay TUNEL assay was performed with an in situ cell death detection kit (Roche). Beta-TC6 cells (2  105 cells) cultured on the glass base dish (Asahi Techno Glass Corp., Tokyo, Japan) were untreated (control) or treated with Cl-IB-MECA (30 lM) for 72 h and then fixed with 4% paraformaldehyde for 60 min. After washing, the cells were incubated in PBS containing 0.1% sodium citrate and 0.1% Triton-X-100 on ice for 2 min and then treated with TUNEL reaction solution for 60 min at 37 °C. After washing, cells were treated both with propidium iodide (PI, 20 lg per ml, Sigma) and ribonuclease A (100 lg/ml, Sigma) for 15 min and the nuclei were stained with PI.

3. Results 3.1. Adenosine receptor mRNAs were expressed in mouse islets and Beta-TC6 cells We first examined the expression of genes encoding four adenosine receptor subtypes, A1, A2A, A2B and A3 in mouse islets and Beta-TC6 cells by RT-PCR. As shown in Fig. 1A and B, mRNAs of all adenosine receptor subtypes were present in the islets and Beta-TC6 cells. The expression level of receptor subtype in the islets was found to be similar to one another (Fig. 1C), whereas the level of the A3 receptor was relatively lower than that of other subtypes in Beta-TC6 cells (Fig. 1D). 3.2. Adenosine receptor agonists augmented insulin secretion by mouse islets We examined the effect of adenosine and its structurally related agonists on insulin secretion by mouse pancreatic islets. As shown in Fig. 2A, adenosine augmented the release of insulin in a concentration-dependent manner in the presence of a high concentration of glucose (20 mM). The stimulatory effect of adenosine was also

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

89

Fig. 7. Detection of DNA fragments and activated caspase-3 proteins in Beta-TC6 cells treated with Cl-IB-MECA by TUNEL assay and Western blot analysis. (A) A double staining with propidium iodide (PI) (nuclei) and TUNEL (fragmented DNA) was carried out in Beta-TC6 cells in the absence (control) or presence of thapsigargin (0.1 lM) for 24 h or Cl-IB-MECA (30 lM) for 48 h. The fluorescent color intensity was detected in PI (red) or TUNEL (green) staining. Scale bars, 50 lm. (B) The number of TUNEL-positive cells in the randomly selected area. Each bar represents the mean ± S.E.M. of three independent experiments (n = 9–12). (C) Upper panels, the proteins (20 lg per lane) extracted from cells cultured in the absence (0 h) or presence of thapsigargin for 24 h or Cl-IB-MECA for 24 or 48 h were separated by SDS–PAGE and immunoblotted with the anti-both total and cleaved (activated) caspase-3 antibody or the anti-b-actin antibody. The expected molecular weights of the activated caspase-3 and b-actin protein were approximately 18 and 45 kDa, respectively. The other details were described in ‘‘Materials and methods’’. Lower panels, the relative amount of total immunoreactive cleaved caspase-3 bands was expressed as cleaved caspase-3/b-actin of the chemiluminescent intensity. Each bar represents the mean ± S.E.M. (n = 3–4). Each experiment was repeated at least three times. Statistical significance was calculated with (B and C, right) ANOVA followed by Bonferroni’s post hoc test or (C, left) Mann–Whitney test. ⁄ P < 0.05 in comparison to control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

observed in the presence of normal glucose (5.5 mM) (Fig. 2B), although it was smaller. As shown in Fig. 3A and B, an A1 receptor agonist CPA and an A3 receptor agonist Cl-IB-MECA at concentrations higher than 100 lM also significantly enhanced insulin secretion in the presence of high glucose (20 mM). The stimulatory effect of an adenosine analogue NECA was concentration-dependent such that stimulation was small at a low concentration (1 lM) and much stronger at high concentrations (P100 lM) (Fig. 3C). Although the effects were not statistically significant, an A2A receptor agonist CGS21680 tended to elicit a slight stimulatory effect at 0.1 and 1 lM and an inhibitory effect at 10 and 100 lM (Fig. 3D, left). We also found by the independent experiments that the A2A agonist at lower concentrations (0.01 and 0.001 lM) produced greater stimulation (Fig. 3D, right). Among various agonists tested, the effect of Cl-IB-MECA was the most potent at high concentrations (Fig. 3E) and the effect of all agonists except CGS21680 was concentration-dependent at concentrations ranging from 10 to 300 lM.

3.3. A2A receptors mediated the stimulatory effect of adenosine on insulin secretion by mouse islets Since adenosine binds to all adenosine receptor subtypes, we examined which adenosine receptor subtype(s) mediated the ef-

Fig. 2. Effect of adenosine on insulin secretion by mouse islets. (A) Groups of 15 islets were incubated for 60 min at 37 °C in the absence (0 lM, control) or presence of different concentrations (0.1, 1, 10, 100 or 300 lM) of adenosine in incubation buffer containing high concentration of glucose (20 mM). The amount of insulin released into the buffer was measured by enzyme immunosorbent assay (EIA). (B) Islets were incubated in the absence (control) or presence of adenosine (100 lM) in incubation buffer containing normal (5.5 mM, G5.5) or high glucose (G20). Each bar represents the mean ± S.E.M. (n = 4–5). Each experiment was repeated at least 3 times. Statistical significance was calculated with (A) ANOVA followed by Bonferroni’s post hoc test or (B) Mann–Whitney test. ⁄P < 0.05 in comparison to control.

fect of adenosine. As shown in Fig. 4A, the stimulatory effect of adenosine (100 lM) was blocked when islets were pretreated with

90

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

Fig. 3. Effect of adenosine receptor agonists on insulin secretion by mouse islets. Groups of 15 islets were incubated for 60 min at 37 °C in the absence (0 lM, control) or presence of different concentrations (0.001, 0.01, 0.1, 1, 10, 100 or 300 lM) of (A) an A1 receptor agonist CPA, (B) an A3 receptor agonist Cl-IB-MECA, (C) a nonselective adenosine receptor agonist NECA or (D) an A2A receptor agonist CGS21680 in incubation buffer containing high concentration of glucose (20 mM). The amount of insulin released into the buffer was measured by enzyme immunosorbent assay (EIA). (E) Effect of adenosine and its structurally related agonists on insulin secretion was expressed as % of control. Each point or bar represents the mean ± S.E.M. (n = 3–7). Each experiment was repeated at least 3 times. Statistical significance was calculated with ANOVA followed by Bonferroni’s post hoc test or Mann–Whitney test. In C and D, the significance between control and the effect of NECA (1 lM) or CGS21680 (0.01 lM) was calculated with Mann–Whitney test as indicated by #. ⁄,#P < 0.05 in comparison to control.

an A2A receptor selective antagonist SCH58261 (1 lM). An A2B receptor antagonist MRS1754 (control: 100%; adenosine alone: 258 ± 52%; MRS1754 alone: 124 ± 26%; adenosine + MRS1754: 245 ± 49%, n = 7) and an A3 receptor antagonist VUF5574 (control: 100%; adenosine alone: 194 ± 21%; VUF5574 alone: 131 ± 14%; adenosine + VUF5574: 217 ± 27%, n = 5) each at 1 lM had no effect. On the other hand, the effect of adenosine was potentiated by approximately 40% when islets were pretreated with an A1 receptor antagonist DPCPX (1 lM) (Fig. 4B). Since the effect of adenosine was not completely blocked by the A2A antagonist, the possibility existed that equilibrative nucleoside transporters (ENTs), which allow adenosine to pass bidirectionally across the cell membrane (Molina-Arcas et al., 2009), were involved in the stimulatory effect of adenosine. We found that the effect of adenosine was not affected when pretreated with two inhibitors of ENT subtype 1 (ENT1), whose mRNA was expressed in mouse islets and Beta-TC6 cells (data not shown), dipyridamole (control: 100%; adenosine alone: 241 ± 36%; dipyridamole alone: 147 ± 30%; adenosine + dipyridamole: 240 ± 55%, n = 7) and NBTI (control: 100%; adenosine alone: 152 ± 11%; NBTI alone: 93 ± 13%; adenosine + NBTI: 175 ± 7%, n = 3) each at 1 lM. These results suggested that A2A receptors played a major role in modulation of insulin secretion by extracellular adenosine. 3.4. A3 receptor agonist Cl-IB-MECA inhibited Beta-TC6 cell proliferation via A3 receptor We examined whether adenosine and its structurally related agonists affected proliferation or the survival of Beta-TC6 cells. As shown in Fig. 5A, we found that the number of cells was decreased by treating with an A3 receptor agonist Cl-IB-MECA at concentrations higher than 10 lM, whereas nonselective adenosine receptor agonists, adenosine and NECA, had no effect. Similar re-

Fig. 4. Effect of adenosine receptor antagonists on the adenosine-induced increase in insulin secretion by mouse islets. Groups of 15 islets were incubated for 60 min in the absence (control) or presence of adenosine (100 lM), (A) an A2A receptor selective antagonist SCH58261 or (B) an A1 receptor selective antagonist DPCPX each at 1 lM alone, or their indicated combination in incubation buffer containing high glucose (20 mM). The antagonist was added 20 min prior to treatment with adenosine. The amount of insulin released into the incubation medium was measured by EIA. Effect of agonists and antagonists on insulin secretion was expressed as % of control. Each bar represents the mean ± S.E.M. (n = 5–7). Each experiment was repeated at least 3 times. Statistical significance was calculated with ANOVA followed by Bonferroni’s post hoc test. In A, the significance between the effect of adenosine alone and of adenosine plus SCH58261 was calculated with Mann–Whitney test as indicated by #. ⁄,#P < 0.05 in comparison to control or adenosine alone.

sults using high concentrations of adenosine were observed on the rate of DNA synthesis (100 lM: 90 ± 13%, 300 lM: 73 ± 7% in comparison with control, n = 6). The effect of Cl-IB-MECA was concentration-dependent within the range from 10 to 50 lM (Fig. 5B). Its effect (30 lM) was also time-dependent during 72 h culture

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

(Fig. 5C). In addition, CPA and CGS21680 each at a higher concentration (300 lM) inhibited the cell growth approximately 30% and 40%, respectively (Fig. 5A). In order to clarify whether the A3 receptor was involved in the inhibitory action of Cl-IB-MECA on BetaTC6 cell proliferation, we examined the effect of VUF5574 and found that the effect of Cl-IB-MECA (30 lM) was partially blocked by VUF5574 (1 lM) (Fig. 5D), suggesting the involvement of the adenosine A3 receptor. The cell viability was reduced by Cl-IBMECA (control: 100%; 10 lM: 100 ± 9.7%; 30 lM: 7.3 ± 0.3%⁄; 50 lM: 7.0 ± 2.1%⁄, n = 3, ⁄P < 0.05). In addition, the DNA synthesis rate was greatly reduced by this agonist (Fig. 5E). These results suggested that the A3 receptor played a role by reducing BetaTC6 cell proliferation or the survival. In addition, we found that treatment with Cl-IB-MECA (30 lM) for 72 h decreased the cell number of other pancreatic b-cell lines; rat RINm5F cells to 24 ± 3% of control (n = 3, P < 0.05) and hamster HIT-T15 cells to 42 ± 12% of control (n = 3, P < 0.05). 3.5. Phosphorylation of ERK was reduced by Cl-IB-MECA in Beta-TC6 cells It has been demonstrated that stimulation of the adenosine receptors results in activation of the MAPK/ERK pathway, which is involved in the regulation of cell proliferation, differentiation or apoptosis (Jacobson, 2009). We thus investigated the effect of Cl-IB-MECA on the phosphorylation of ERK in Beta-TC6 cells. As shown in Fig. 6, Cl-IB-MECA at 30 lM reduced the ERK1 and 2 phosphorylation approximately 50% during 30 min incubation. The inhibitory effect of Cl-IB-MECA was not blocked by VUF5574

91

(1 lM) (control: 1.0; Cl-IB-MECA alone: 0.66 ± 0.10; VUF5574 alone: 1.07 ± 0.08; Cl-IB-MECA + VUF5574: 0.62 ± 0.09, n = 5), suggesting that the reduction in ERK phosphorylation was independent of activation of the A3 receptor. 3.6. Cl-IB-MECA induced Beta-TC6 cell death independently of apoptosis Since the cell number was markedly decreased by treatment with Cl-IB-MECA (Fig. 5A and B), the possibility existed that apoptosis of Beta-TC6 cells occurred. As shown in Fig. 7A and B, we performed TUNEL assay using a SERCA inhibitor thapsigargin as a positive control. TUNEL-positive immunostaining was observed in cells treated with thapsigargin (0.1 lM) for 24 h. Although the number was smaller than that with thapsigargin, the positive staining was also detected in cells treated with Cl-IB-MECA (30 lM) for 48 h, suggesting that apoptosis occurred in these cells. Based on the fact that activation of various proteases such as caspase-3 was involved in apoptosis of pancreatic b-cells (Hui et al., 2004), we also investigated by Western blot analysis whether ClIB-MECA affected activation of caspase-3, using an antibody that recognizes both total and activated caspase-3 proteins. As shown in Fig. 7C, the amount of cleaved (activated) caspase-3 protein (approximately 18 kDa) was not greatly changed when cells were cultured for 24 or 48 h in the presence of Cl-IB-MECA, whereas the increase in the amount of activated caspase-3 protein was clear among cells treated with thapsigargin for 24 h. These results suggested that treatment with Cl-IB-MECA induced the death of Beta-TC6 cells mainly via necrosis but not by apoptosis.

Fig. 5. Effect of adenosine receptor agonists and A3 receptor antagonist on proliferation of Beta-TC6 cells. (A) The cells were cultured for 72 h in the absence (0 lM, control) or presence of different concentrations (0.1, 1, 10, 100 or 300 lM) of adenosine, NECA, CPA, CGS21680 or Cl-IB-MECA in medium containing 10% FBS and 25 mM glucose and the resulting cell number was counted after harvesting. (B) The concentration-dependent effect of Cl-IB-MECA. The cells were cultured in the absence (control) or presence of ClIB-MECA at 10, 30 or 50 lM for 72 h. (C) The time-dependent effect of Cl-IB-MECA. Cells were cultured in the absence (control) or presence of Cl-IB-MECA (30 lM) for 24, 48 or 72 h. The cell number was counted at the indicated time. (D) Cells were cultured for 72 h in the absence (control) or presence of Cl-IB-MECA (30 lM), VUF5574 (1 lM) alone, or their indicated combination. The antagonist was added 20 min prior to treatment with Cl-IB-MECA. (E) The cells were cultured for 72 h in the absence (control) or presence of Cl-IB-MECA (10, 30 or 50 lM) and the rate of DNA synthesis was measured with BrdU as described in Section 2. Each point or bar represents the mean ± S.E.M. (n = 3–12). Each experiment was repeated at least three times. Statistical significance was calculated with ANOVA followed by Bonferroni’s post hoc test or Mann–Whitney test. In D, the significance between the effect of Cl-IB-MECA alone and of Cl-IB-MECA plus VUF5574 was calculated with Mann–Whitney test as indicated by #. ⁄,#P < 0.05 in comparison to control or Cl-IB-MECA alone.

92

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

4. Discussion

Fig. 6. Western blot analysis of phosphorylated ERK proteins in Beta-TC6 cells treated with Cl-IB-MECA. The cells were cultured in the absence (control, 0 min) or presence of Cl-IB-MECA (30 lM) with the varying length of times (5, 15, 30, 60, 120 min). The equal amount of extracted proteins (20 lg per lane) were separated by SDS–PAGE and immunoblotted with the anti-phosphorylated ERK1 and 2 (pERK1/2) or total ERK1 and 2 (ERK1/2) antibody. The other details were described in Section 2. The relative amount of total immunoreactive pERK bands (pERK1 and pERK2) was expressed as pERK1 plus pERK2/ERK1 plus ERK2 of the chemiluminescent intensity. Each bar represents the mean ± S.E.M. (n = 4). Each experiment was repeated at least three times. Statistical significance was calculated with ANOVA followed by Bonferroni’s post hoc test. ⁄P < 0.05 in comparison to control.

3.7. Mouse islet cell viability was reduced by Cl-IB-MECA but not by adenosine or NECA The effect of Cl-IB-MECA on the viability of mouse islet cells was examined in culture. As shown in Fig. 8A, treatment with Cl-IBMECA reduced the cell viability in a concentration-dependent manner in the presence of either normal (5.5 mM) or high (25 mM) concentration of glucose. The viability was slightly increased in the presence of high glucose alone during 72 h culture in comparison with normal glucose (127 ± 13%, n = 9, P < 0.05), suggesting no glucose toxicity during long term incubation. The effect of Cl-IB-MECA (30 lM) was partially antagonized by pretreating cells with VUF5574 (3 lM) in the presence of high glucose (Fig. 8B). In addition, pretreatment with dipyridamole (1 lM) did not affect the effect of Cl-IB-MECA (control: 100%; Cl-IB-MECA alone: 32 ± 3%; dipyridamole alone: 109 ± 15%; Cl-IB-MECA + dipyridamole: 40 ± 3%, n = 3). These results suggested that Cl-IB-MECA reduced the survival rate of primary pancreatic islet cells partly via the A3 receptor in a glucose-independent manner. On the other hand, adenosine and NECA even at 300 lM did not affect the cell viability in the presence of either 5.5 or 25 mM glucose (Fig. 8C).

Previous studies showed that genes encoding adenosine receptor subtypes (A1, A2A, A2B and A3) were all expressed in rat pancreatic ducts (Novak, 2008) and mouse pancreas (Németh et al., 2007). The expression of A1, A2A and A2B mRNAs as well as the presence of the A1 receptor protein was also found in a rat b-cell line, INS-1 cells (Töpfer et al., 2008). The finding of these adenosine receptors raised the possibility that they play a role in the regulation of hormone secretion, cell proliferation or apoptosis in pancreatic b-cells. The observed stimulatory effect of adenosine at a high concentration (100 or 300 lM) on insulin secretion in mouse islets was in accord with the previous finding that perfusion of adenosine (500 lM) enhanced the release of insulin by b-cells (Bertrand et al., 1989). Our data suggested that action of adenosine was mainly mediated by the A2A receptor. On the other hand, an A2A selective agonist CGS21680 caused stimulation only at low concentrations (0.01 and 0.001 lM) (Fig. 3D). The difference in the effective concentrations of the two agonists may be due to the difference in the affinity for A2A receptors. Adenosine has a relatively low affinity (150 nM) compared with that of the A2A agonist (19 nM) (Fredholm et al., 2011). We found that the two types of adenosine receptor agonist, CPA (A1 receptor) and Cl-IB-MECA (A3 receptor) needed high concentrations (P100 lM) to produce stimulation of insulin secretion, despite their high affinity for A1 (2.3 nM) and A3 (1.4 nM) receptors, respectively. One possible reason why these agonists needed a high concentration is that when used at high concentrations they may augment insulin secretion nonspecifically via binding to other receptors such as the A2A receptor despite low affinity (CPA: 794 nM, Cl-IB-MECA: 5360 nM). A synthetic adenosine analogue NECA (nonselective agonist) produced weak stimulation at 1 lM but strong stimulation at high concentrations (Fig. 3C). The affinity of NECA is 5.1 nM for A1, 9.7 nM for A2A and 113 nM for A3 receptors (Fredholm et al., 2011), suggesting that A2A receptors could be activated by this agonist at low concentrations. It has been reported that low concentrations (16.5 nM or 1.65 lM) of A1 receptor agonists L- and D-PIA inhibited insulin release by rat pancreas (Hillaire-Buys et al., 1987, 1989). Since the extracellular concentration of adenosine is estimated to be 2– 9 lM in mouse islets (Yang et al., 2012), endogenous adenosine (affinity for A1: 73 nM) may bind to the A1 receptor and contribute to the regulation of insulin secretion via activation of this receptor. It was shown that treatment with an A1 selective antagonist DPCPX (1 lM) alone stimulated insulin secretion by rat islets along with the increased content of cAMP (Zywert et al., 2011), which promotes insulin exocytosis in b-cells (Yamada et al., 2002). Thus, the reason for the interaction between adenosine and DPCPX

Fig. 8. Effect of adenosine, NECA, Cl-IB-MECA and A3 receptor antagonist on the viability of mouse islet cells. Islet cells were cultured in the absence (control) or presence of (A) Cl-IB-MECA (1, 10, 30 or 50 lM), (C) adenosine or NECA each at 300 lM for 72 h in medium containing 10% FBS and 5.5 or 25 mM glucose. (B) The cells were cultured in medium containing 25 mM glucose in the absence (control) or presence of Cl-IB-MECA (30 lM), VUF5574 (3 lM) alone, or their indicated combination for 72 h. The cell viability was measured as described in Section 2. The antagonist was added 20 min prior to treatment with Cl-IB-MECA. Each bar represents the mean ± S.E.M. (n = 3–11). Each experiment was repeated at least three times. Statistical significance was calculated with ANOVA followed by Bonferroni’s post hoc test. In B, the significance between the effect of Cl-IB-MECA alone and of Cl-IB-MECA plus VUF5574 was calculated with Mann–Whitney test as indicated by #. ⁄,#P < 0.05 in comparison to control or Cl-IB-MECA alone.

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

(Fig. 4B) may be explained by the hypothesis that blockade of A1 receptors, which are coupled to Gi proteins, enhanced the adenylate cyclase activity and thus elevated the cAMP level, resulting in enhancement of insulin secretion induced by adenosine via A2A receptor activation. On the other hand, the reason why adenosine and CPA at low concentrations did not inhibit insulin secretion still remains unclear. Based on the above findings we proposed that the A1 receptor plays a physiological role in the negative regulation of insulin secretion, whereas the A2A receptor can be activated by adenosine under some pathological (supraphysiological) conditions such as severe ischemia in which the level of extracellular adenosine is rapidly increased to the micromolar concentration range (Jacobson, 2009). The finding that VUF5574, an A3 receptor antagonist, partially blocked the inhibitory effect of Cl-IB-MECA suggested that activation of the A3 receptor contributed to negative modulation of BetaTC6 cell proliferation. The number of Beta-TC6 cells in Fig. 5B in the presence of 30 and 50 lM Cl-IB-MECA was 1.12 ± 0.34  105 cells (control: 2.15 ± 0.5  105 cells, n = 5) and 0.15 ± 0.03  105 cells (control: 1.57 ± 0.33  105 cells, n = 3), respectively. Since the starting number of cells was 1.0  105 cells, it was possible that treatment with the A3 agonist reduced the rate of cell growth at 30 lM and was cytotoxic at concentrations higher than 50 lM. Although treatment with Cl-IB-MECA slightly increased TUNELpositive Beta-TC6 cells, activation of caspase-3 did not significantly occur, suggesting that caspase-independent apoptosis, which often shares common characteristics with caspase-dependent apoptosis (Tait and Green, 2008), may be involved. However, it was suggested that a main cause to induce the Beta-TC6 cell death was necrosis but not apoptosis, since the number of TUNEL-positive cells, one of indices for apoptosis, was considerably smaller than positive control (Fig. 7B). It was reported that interleukin-1 (IL1), which is synthesized and secreted into the extracellular space, induced apoptosis of b-cells in an autocrine or a paracrine manner when cells were exposed to high glucose for a long term period (Donath et al., 2005; Mandrup-Poulsen, 2003). Thus one possible reason for the reduction in the cell number was that the exposure to high concentrations of the A3 agonist elicited the secretion of inflammatory cytokines, which caused necrosis or apoptosis of Beta-TC6 cells. Although Cl-IB-MECA (P100 lM) augmented insulin secretion by mouse islets in the presence of high glucose (20 mM), the A3 agonist markedly reduced the viability of mouse islet cells and the effect of agonist was partly reversed by the A3 antagonist. On the other hand, adenosine and its analogue NECA, which is resistant to adenosine deaminases (Fredholm et al., 2011; Jacobson, 2009), had apparently no effect on the islet cell viability even at a high concentration (300 lM). Very recently, Andersson et al. (2012) reported that NECA lowered blood glucose and enhanced b-cell proliferation in type 1 diabetes model mice. Thus, an adenosine derivative may be useful for treatment of diabetes mellitus, since adenosine and NECA can promote insulin secretion without affecting the viability of the normal islet cells.

5. Conclusion Genes of all adenosine receptor subtypes were expressed in mouse pancreatic islets. We found that adenosine and synthetic adenosine receptor agonists markedly augmented insulin secretion by islets. Based on the studies using various pharmacological agents it was suggested that the A2A receptor mainly mediated the effect of adenosine. On the other hand, the A3 receptor agonist Cl-IB-MECA at a high concentration robustly reduced the survival of mouse islet cells and Beta-TC6 cells, while adenosine had no effect. These results suggested that the A2A receptor plays a major

93

role in the modulatory effect of adenosine on insulin secretion by islets and that the A3 receptor is involved in the survival of pancreatic b-cells. Acknowledgments This work was supported by the intramural funds from Osaka Dental University. We thank Dr. Kenneth A. Jacobson (NIDDK, NIH) for critical reading of the manuscript and useful comments and suggestions. References Andersson, O., Adams, B.A., Yoo, D., Ellis, G.C., Gut, P., Anderson, R.M., German, M.S., Stainier, D.Y., 2012. Adenosine signaling promotes regeneration of pancreatic b cells in vivo. Cell Metab. 15, 885–894. Bertrand, G., Petit, P., Bozen, M., Henquin, J.C., 1989. Membrane and intracellular effects of adenosine in mouse pancreatic b-cells. Am. J. Physiol. Endocrinol. Metab. 257, E473–E478. Chapal, J., Loubatières-Mariani, M.M., Petit, P., Roye, M., 1985. Evidence for an A2subtype adenosine receptor on pancreatic glucagon secreting cells. Br. J. Pharmacol. 86, 565–569. Donath, M.Y., Ehses, J.A., Maedler, K., Schumann, D.M., Ellingsgaard, H., Eppler, E., Reinecke, M., 2005. Mechanisms of b-cell death in type 2 diabetes. Diabetes 54, S108–S115. Fredholm, B.B., Ijzerman, A.P., Jacobson, K.A., Linden, J., Müller, C.E., 2011. International union of basic and clinical pharmacology. LXXI. Nomenclature and classification of adenosine receptors—an update. Pharmacol. Rev. 63, 1–34. Hillaire-Buys, D., Bertrand, G., Gross, R., Loubatières-Mariani, M.M., 1987. Evidence for an inhibitory A1 subtype adenosine receptor on pancreatic insulin-secreting cells. Eur. J. Pharmacol. 136, 109–112. Hillaire-Buys, D., Gross, R., Loubatières-Mariani, M.M., Ribes, G., 1989. Effect of pertussis toxin on A1-receptor-mediated inhibition of insulin secretion. Br. J. Pharmacol. 96, 3–4. Hui, H., Dotta, F., Di Mario, U., Perfetti, R., 2004. Role of caspases in the regulation of apoptotic pancreatic islet cells death. J. Cell. Physiol. 200, 177–200. Iwatsuki, K., 2000. Subtypes of adenosine receptors on pancreatic exocrine secretion in anaesthetized dogs. Fundam. Clin. Pharmacol. 14, 203–208. Jacobson, K.A., 2009. Introduction to adenosine receptors as therapeutic targets. Handb. Exp. Pharmacol. 193, 1–24. Jacobson, K.A., Hoffmann, C., Cattabeni, F., Abbracchio, M.P., 1999. Adenosineinduced cell death: evidence for receptor-mediated signaling. Apoptosis 4, 197– 211. Johansson, S.M., Salehi, A., Sandström, M.E., Westerblad, H., Lundquist, I., Carlson, P.O., Fredholm, B.B., Katz, A., 2007. A1 receptor deficiency causes increased insulin glucagon secretion in mice. Biochem. Pharmacol. 74, 1628–1635. Mandrup-Poulsen, T., 2003. Apoptotic signal transduction pathways in diabetes. Biochem. Pharmacol. 66, 1433–1440. Molina-Arcas, M., Casado, F.J., Pastor-Anglada, M., 2009. Nucleoside transporter proteins. Curr. Vasc. Pharmacol. 7, 426–434. Németh, Z.H., Bleich, D., Csóka, B., Pacher, P., Mabley, J.G., Himer, L., Vizi, E.S., Deitch, E.A., Szabó, C., Cronstein, B.N., Haskó, G., 2007. Adenosine receptor activation ameliorates type 1 diabetes. FASEB J. 21, 2379–2388. Novak, I., 2008. Purinergic receptors in the endocrine and exocrine pancreas. Purinergic Signal. 4, 237–253. Novak, I., Hede, S.E., Hanse, M.R., 2008. Adenosine receptors in rat and human pancreatic ducts stimulate chloride transport. Pflugers Archiv. 456, 437–447. Ohana, G., Bar-Yehuda, S., Barer, F., Fishman, P., 2001. Differential effect of adenosine on tumor and normal cell growth: focus on the A3 adenosine receptor. J. Cell. Physiol. 186, 19–23. Ohtani, M., Daly, J.W., Oka, T., 2009a. Co-existence of muscarinic and nicotinic receptors and their functional interaction in mouse Beta-TC6 cells. Eur. J. Pharmacol. 604, 150–157. Ohtani, M., Mizuno, I., Kojima, Y., Ishikawa, Y., Sodeno, M., Asakura, Y., Samejima, K., Oka, T., 2009b. Spermidine regulates insulin synthesis and cytoplasmic Ca2+ in mouse Beta-TC6 insulinoma cells. Cell Struct. Funct. 34, 105–113. Ohtani, M., Ohura, K., Oka, T., 2011. Involvement of P2X receptors in the regulation of insulin secretion, proliferation and survival in mouse pancreatic b-cells. Cell. Physiol. Biochem. 28, 355–366. Ohtani, M., Oka, T., Badyuk, M., Xia, Y., Kellar, K.J., Daly, J.W., 2006. Mouse beta-TC6 insulinoma cells: high expression of functional alpha3beta4 nicotinic receptors mediating membrane potential, intracellular calcium, and insulin release. Mol. Pharmacol. 69, 899–907. Ohtani, M., Suzuki, J., Jacobson, K.A., Oka, T., 2008. Evidence for the possible involvement of the P2Y6 receptor in Ca2+ mobilization and insulin secretion in mouse pancreatic islets. Purinergic Signal. 4, 365–375. Ralevic, V., Burnstock, G., 1998. Receptors for purines and pyrimidines. Pharmacol. Rev. 50, 413–492. Salehi, A., Parandeh, F., Fredholm, B.B., Grapengiesser, E., Hellman, B., 2009. Absence of adenosine A1 receptors unmasks pulses of insulin release and prolongs those of glucagon and somatostatin. Life Sci. 85, 470–476.

94

M. Ohtani et al. / General and Comparative Endocrinology 187 (2013) 86–94

Satoh, A., Satoh, K., Masamune, A., Yamagiwa, T., Shimosegawa, T., 2002. Activation of adenosine A2a receptor pathway reduces leukocyte infiltration but enhances edema formation in rat caerulein pancreatitis. Pancreas 24, 75–82. Srinivasan, S., Ohsugi, S.M., Liu, Z., Fatrai, S., Bernal-Mizrachi, E., Permutt, M.A., 2005. Endoplasmic reticulum stress-induced apoptosis is partly mediated by reduced insulin signaling through phosphatidylinositol 3-kinase/Akt and increased glycogen synthase kinase-3b in mouse insulinoma cells. Diabetes 54, 968–975. Tait, S.W.G., Green, D.R., 2008. Caspase-independent cell death: leaving the set without the final cut. Oncogene 27, 6452–6461. Tonnesen, M.F., Grunet, L.G., Friberg, J., Cardozo, A.K., Billestrup, N., Eijiriki, D.L., Størling, J., Mandrup-Poulsen, T., 2009. Inhibition of nuclear factor-jB or Bax prevents endoplasmic reticulum stress- but not nitric oxide-mediated apoptosis in INS-1E cells. Endocrinology 150, 4049–4103.

Töpfer, M., Burbiel, C.E., Müller, C.E., Knittel, J., Verspohl, E.J., 2008. Modulation of insulin release by adenosine A1 receptor agonists and antagonists in INS-1 cells: the possible contribution of 86Rb efflux and 45Ca uptake. Cell Biochem. Funct. 26, 833–843. Yamada, S., Komatsu, M., Sato, Y., Yamaguchi, K., Kojima, I., Aizawa, T., Hoshizume, K., 2002. Time-dependent stimulation of insulin exocytosis by 30 ,50 -cyclic adenosine monophosphate in the rat islet beta-cell. Endocrinology 143, 4203– 4209. Yang, G.K., Squires, P.E., Tian, F., Kieffer, T.J., Kwok, Y.N., Dale, N., 2012. Glucose decreases extracellular adenosine levels in isolated mouse and rat pancreatic islets. Islets 4, 64–70. Zywert, A., Szkudelska, K., Szkudelski T., 2011. Effects of adenosine A1 receptor antagonism on insulin secretion from rat pancreatic islets. Physiol. Res. 60, 905– 911.