{"id":614971,"date":"2024-06-14T20:00:00","date_gmt":"2024-06-15T00:00:00","guid":{"rendered":"https:\/\/platohealth.ai\/intra-islet-%ce%b1-cell-gs-signaling-promotes-glucagon-release-nature-communications\/"},"modified":"2024-06-15T11:29:59","modified_gmt":"2024-06-15T15:29:59","slug":"intra-islet-%ce%b1-cell-gs-signaling-promotes-glucagon-release-nature-communications","status":"publish","type":"post","link":"https:\/\/platohealth.ai\/intra-islet-%ce%b1-cell-gs-signaling-promotes-glucagon-release-nature-communications\/","title":{"rendered":"Intra-islet \u03b1-cell Gs signaling promotes glucagon release – Nature Communications","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"
To explore the metabolic outcome of selectively activating Gs<\/sub> signaling in pancreatic \u03b1-cells, we used DREADD technology to generate a mutant mouse strain that selectively expressed the Gs-DREADD (GsD) in \u03b1-cells of adult mice. Specifically, we intercrossed mice harboring the GsD<\/i> allele preceded by a loxP-STOP-loxP<\/i> (LSL<\/i>) sequence (CAG-LSL-GsD<\/i> mice)33<\/a><\/sup> with Gcg-CreER<\/i>T2<\/i><\/sup> mice34<\/a><\/sup> to generate heterozygous CAG-LSL-GsD<\/i> mice containing one copy of the Gcg-CreER<\/i>T2<\/i><\/sup> transgene. These mutant mice were then treated with tamoxifen (TMX), thus promoting Cre activity and GsD expression in pancreatic \u03b1-cells and endocrine L cells of the gastrointestinal tract. Because intestinal L cells turn over rapidly, Cre-modified L-cells are known to be replaced by wild-type (WT) L-cells 4 weeks after TMX treatment34<\/a><\/sup>. Four weeks after the last TMX injection, the CAG-LSL-GsD Gcg-CreER<\/i>T2<\/i><\/sup> mice expressed GsD selectively in \u03b1-cells (hereafter referred to as \u03b1-GsD mice). TMX-treated CAG-LSL-GsD<\/i> mice that did not harbor the Gcg-CreER<\/i>T2<\/i><\/sup> transgene and thus did not express the GsD receptor served as control animals throughout the study.<\/p>\n The expression of the GsD designer receptor was detected using an antibody specific to the HA tag that had been fused to the N-terminus of GsD33<\/a><\/sup>. Immunoblotting analysis revealed that \u03b1-GsD mice expressed GsD in pancreatic islets but not in any other tissues including intestinal tissue known to contain proglucagon-producing L-cells (Supplementary Fig. 1a<\/a>). Immunofluorescence staining of slices prepared from pancreatic islets from \u03b1-GsD mice confirmed the expression of GsD in glucagon-expressing \u03b1-cells and the lack of GsD expression in insulin-containing \u03b2-cells or other islet cells (Supplementary Fig. 1b<\/a>). As expected, GsD was not detectable in pancreatic slices prepared from control mice (Supplementary Fig. 1b<\/a>). Moreover, immunofluorescence staining of brain slices prepared from \u03b1-GsD mice failed to detect GsD expression in proglucagon-producing neurons of the nucleus tractus solitarius (NTS; Supplementary Fig. 1c<\/a>). These observations indicate that GsD is selectively expressed in pancreatic \u03b1-cells of \u03b1-GsD mice.<\/p>\n The expression of GsD in pancreatic \u03b1-cells did not affect body weight, islet size, or \u03b1- and \u03b2-cell mass (Supplementary Fig. 1d\u2013g<\/a>). \u03b1-GsD mice (males) and their control littermates consuming regular chow were then injected with either saline (i.p.) or the selective DREADD agonist, DCZ (10\u2009\u03bcg\/kg, i.p.)27<\/a><\/sup>, followed by the monitoring of changes in plasma glucagon, plasma insulin, and blood glucose levels. Similar to previous observations13<\/a>,35<\/a><\/sup>, i.p. injection of control and \u03b1-GsD mice with saline or of control mice with DCZ resulted in reduced plasma glucagon levels 15 and 30\u2009min after injection (p<\/i>\u2009<\u20090.0001; one-way repeated measures ANOVA, followed by post-hoc Bonferroni adjustment) (Fig. 1a, d<\/a>). Although the precise mechanism underlying this phenomenon remains unclear, it is possible that the stress of the i.p. injection is confounding basal hormonal levels. Although there was a trend towards higher basal plasma glucagon level in \u03b1-GsD mice (time 0; Fig. 1d<\/a>), this effect failed to reach statistical significance (p<\/i>\u2009=\u20090.085; Student\u2019s t<\/i> test at time 0). Importantly, DCZ treatment of \u03b1-GsD mice led to a statistically significant increase in plasma glucagon levels (Fig. 1d<\/a>), thus overcoming the inhibitory effect on glucagon release caused by the injection stress.<\/p>\n Freely fed \u03b1-GsD mice and control littermates were injected with saline (a<\/b>\u2013c<\/b>) or DCZ (10\u2009\u03bcg\/kg, i.p.) (d<\/b>\u2013f<\/b>). Plasma glucagon (a<\/b>, d<\/b>), plasma insulin (b<\/b>, e<\/b>), and blood glucose (c<\/b>, f<\/b>) levels were measured at the indicated time points. g<\/b>\u2013j<\/b> Pancreatic islets prepared from control and \u03b1-GsD mice were perifused with the indicated glucose concentrations in the presence of DCZ (10\u2009nM) and (alanine (3\u2009mM)). Glucagon ((g<\/b>); insert: glucagon from 23 to 70\u2009min) and insulin secretion (i<\/b>) were measured in the presence of low and high glucose levels (3\u2009mM [G3] and 12\u2009mM [G12], respectively). AOC values were calculated for glucagon (h<\/b>) and insulin (j<\/b>) secretion calculated for different stimulation periods. All experiments were carried out with male littermates (12\u201316 weeks old). Data are given as means\u2009\u00b1\u2009SEM (in vivo studies: control, n<\/i>\u2009=\u20099; \u03b1-GsD, n<\/i>\u2009=\u20097; in vitro studies: 3 independent perifusion experiments with 75\u2013100 islets per perifusion chamber). Data were analyzed via two-way repeated measures ANOVA for time with Bonferroni post hoc test for comparison of individual time points (d<\/b>\u2013f<\/b>) or two-tailed Student\u2019s t<\/i> test (h<\/b>, j<\/b>). Numbers of above data points or horizontal lines in the bar graphs represent p values. AOC, area of the curve. Source data are provided as a Source Data<\/a> file.<\/p>\n<\/div>\n<\/div>\n Saline treatment of \u03b1-GsD mice and control littermates resulted in statistically significant increases in blood glucose levels (p<\/i>\u2009<\u20090.0001 at 15 and 30\u2009min after injection (time 0); one-way repeated measures with time ANOVA for each group) (Fig. 1c<\/a>). While this effect persisted in DCZ-treated control mice, blood glucose levels remained unchanged after DCZ injection of \u03b1-GsD mice (Fig. 1f<\/a>), most likely due the increase in plasma insulin levels observed with DCZ-treated \u03b1-GsD mice (Fig. 1e<\/a>) that \u201cneutralized\u201d the hyperglemic effect of the injection stress.<\/p>\n Acute DCZ (10\u2009\u03bcg\/kg, i.p.) treatment of \u03b1-GsD and control mice had no significant effect on the plasma levels of somatostatin and the two major incretin hormones, GIP and GLP-1 (note that GLP-1 is a cleavage product of proglucagon and is primarily secreted from intestinal L cells) (Supplementary Fig. 1h\u2013j<\/a>). These data support the concept that chemogenetic activation of \u03b1-cell Gs<\/sub> signaling promotes the secretion of glucagon which can then act on adjacent \u03b2-cells to stimulate the release of insulin11<\/a>,12<\/a>,13<\/a>,14<\/a>,15<\/a><\/sup>.<\/p>\n To confirm that the DCZ-induced changes in hormone secretion observed with \u03b1-GsD mice in vivo were indeed due to altered Gs<\/sub> signaling in pancreatic \u03b1-cells, we conducted perifusion experiments using islets prepared from \u03b1-GsD mice (\u03b1-GsD islets) and control littermates (control islets). Basal glucagon release at both G3 and G12 was significantly higher in \u03b1-GsD islets, as compared to control islets (Fig. 1g, h<\/a> and Supplementary Fig. 1k<\/a>). This observation is consistent with previous findings that the GsD designer receptor shows a certain degree of constitutive activity under distinct experimental conditions25<\/a>,33<\/a><\/sup>.<\/p>\n Basal insulin secretion was similarly low in both \u03b1-GsD and control islets at G3 (Supplementary Fig. 1l<\/a>). Treatment of \u03b1-GsD and control islets with a physiological amino acid mixture (AAM) or alanine which stimulate the secretory activity of mouse \u03b1-cells11<\/a>,36<\/a><\/sup> enhanced the secretion of both glucagon and insulin at G12 (Fig. 1g\u2013j<\/a>, Supplementary Fig. 1k, l<\/a>), consistent with the concept that the paracrine effects of glucagon requires elevated glucose concentrations to stimulate insulin secretion11<\/a>,12<\/a>,13<\/a>,14<\/a>,15<\/a><\/sup>. DCZ (10\u2009nM) treatment of \u03b1-GsD islets in the presence of 3 or 12\u2009mM glucose resulted in significant increases in glucagon secretion only in the presence of alanine (3\u2009mM) (Fig. 1g, h<\/a>). To correct for differences in basal hormone secretion, we calculated area of the curve (AOC) values by subtracting the areas under or over the baseline37<\/a><\/sup>. This type of analysis is the method of choice when baseline levels between two or more experimental groups differ37<\/a><\/sup>.<\/p>\n Under physiological conditions, pancreatic islets are exposed to high levels of circulating amino acids which increase the responsiveness of \u03b1-cells to various glucagon secretagogues36<\/a>,38<\/a><\/sup>, providing a likely explanation for the inability of DCZ to stimulate glucagon release from \u03b1-GsD islets in the absence of alanine (Fig. 1g, h<\/a>). At G12, the DCZ\/alanine-induced increases in glucagon release in \u03b1-GsD islets were accompanied by marked enhancements of glucose-stimulated insulin secretion (GSIS) (Fig. 1i, j<\/a>). Thus, except for the enhanced basal activity of GsD in vitro, the islet perifusion data are in good agreement with the in vivo results described in the previous paragraph (Fig. 1d, e<\/a>).<\/p>\n To explore the impact of acute activation of \u03b1-cell Gs<\/sub> signaling on glucose homeostasis, \u03b1-GsD mice and control littermates maintained on regular rodent chow (lean mice) were subjected to an i.p. glucose tolerance test (ipGTT). Following co-injection of DCZ (10\u2009\u03bcg\/kg, i.p.) and glucose (2\u2009g\/kg, i.p.), \u03b1-GsD mice showed a significant improvement in glucose tolerance, as compared to control littermates (Fig. 2a<\/a>). This beneficial metabolic effect was associated with pronounced increases in both plasma glucagon and insulin levels in \u03b1-GsD mice (Fig. 2b, c<\/a>), suggesting that the increase in insulin secretion following activation of \u03b1-cell Gs<\/sub> signaling causes improved glucose tolerance. Co-injection of control and \u03b1-GsD mice with insulin (0.75 IU\/kg, i.p.) (insulin tolerance test, ITT) and DCZ (10\u2009\u03bcg\/kg; i.p.) resulted in comparable decreases in blood glucose levels in both groups of mice, indicating that stimulation of \u03b1-cell Gs<\/sub> signaling does not affect peripheral insulin sensitivity (Fig. 2d<\/a>). In the absence of DCZ, control and \u03b1-GsD mice did not show any significant differences in blood glucose excursions in the ipGTT and ITT assays (Supplementary Fig. 2a, b<\/a>). Taken together, these findings indicate that activation of \u03b1-cell Gs<\/sub> signaling leads to improved glucose tolerance, most likely due to increased insulin release triggered by enhanced glucagon secretion.<\/p>\n \u03b1-GsD mice and control littermates consuming regular chow (lean mice) or a high-fat diet (HFD; obese mice) were subjected to a series of metabolic tests. a<\/b> Glucose tolerance test (ipGTT). Lean mice that had been fasted overnight were co-injected (i.p.) with glucose (2\u2009g\/kg) and DCZ (10\u2009\u03bcg\/kg) (control, n<\/i>\u2009=\u20098; \u03b1-GsKO, n<\/i>\u2009=\u20096). Changes in plasma glucagon (b<\/b>) and plasma insulin (c<\/b>) levels following i.p. co-injection of lean mice with glucose and DCZ (control, n<\/i>\u2009=\u20097; \u03b1-GsKO, n<\/i>\u2009=\u20098). d<\/b> Insulin tolerance test (ITT). Lean mice that had been fasted for 4\u2009h after were injected (i.p.) with a mixture of insulin (0.75\u2009U\/kg) and DCZ (control, n<\/i>\u2009=\u200910; \u03b1-GsKO, n<\/i>\u2009=\u200910). e<\/b> ipGTT. Obese mice that had been fasted overnight were co-injected (i.p.) with glucose (1\u2009g\/kg) and DCZ (10\u2009\u03bcg\/kg) (control, n<\/i>\u2009=\u20097; \u03b1-GsKO, n<\/i>\u2009=\u200911). Changes in plasma glucagon (f<\/b>) and plasma insulin (g<\/b>) levels following i.p. co-injection of obese mice with glucose and DCZ (control, n<\/i>\u2009=\u20098; \u03b1-GsKO, n<\/i>\u2009=\u20097). h<\/b> Insulin tolerance test (ITT). Following a 4\u2009h fast, obese mice were injected (i.p.) with a mixture of insulin (1\u2009U\/kg) and DCZ (10\u2009\u03bcg\/kg) (control, n<\/i>\u2009=\u20098; \u03b1-GsKO, n<\/i>\u2009=\u200911). Blood samples were collected from the tail vein at the indicated time points. All experiments were carried out with male littermates that were at least 14 weeks old. Obese mice consumed the HFD for at least 8 weeks. Data are given as means\u2009\u00b1\u2009SEM. Data were subjected to two-tailed Student\u2019s t<\/i> test (AOC bars) or to two-way repeated measures ANOVA for time with Bonferroni post hoc test for comparison of individual time points (a<\/b>\u2013c<\/b>, e<\/b>\u2013g<\/b>). AOC, area over the curve. Numbers in the bar graphs or next to specific data points data points refer to p<\/i> values. AOC, area of the curve. Source data are provided as a Source Data<\/a> file.<\/p>\n<\/div>\n<\/div>\n We next examined whether stimulation of \u03b1-cell Gs<\/sub> signaling also improved glucose homeostasis in obese, glucose-intolerant mice. To address this question, \u03b1-GsD mice and control littermates were maintained on a high-fat diet (HFD) for at least 8 weeks. Consumption of the HFD led to a similar degree of weight gain in both groups of mice (Supplementary Fig. 2c<\/a>). The two cohorts of mice were then co-injected with glucose (1\u2009g\/kg, i.p.) and DCZ (10\u2009\u03bcg\/kg, i.p.) (ipGTT). Co-injected obese \u03b1-GsD mice displayed significantly improved glucose tolerance, as compared to obese control littermates (Fig. 2e<\/a>). As observed with lean mice (Fig. 2b<\/a>), DCZ-induced activation of \u03b1-cell Gs<\/sub> signaling resulted in a striking increase in plasma glucagon levels in obese \u03b1-GsD mice, but not in obese control littermates (Fig. 2f<\/a>). Plasma insulin levels were also significantly elevated in obese \u03b1-GsD mice co-injected with glucose and DCZ (GSIS) (Fig. 2g<\/a>). Co-injection of obese \u03b1-GsD mice and their control littermates with insulin (1\u2009U\/kg, i.p.) and DCZ (10\u2009\u03bcg\/kg, i.p.) (ITT) did not reveal any differences in insulin sensitivity between the two groups of mice (Fig. 2h<\/a>). These data indicate that acute activation of \u03b1-cell Gs<\/sub> signaling results in an insulinotropic effect that improves glucose homeostasis in both lean and obese, glucose-intolerant mice.<\/p>\n Our next goal was to identify Gs<\/sub>-coupled GPCRs that are endogenously expressed by pancreatic \u03b1-cells with relatively high selectivity. Analysis of previously published scRNAseq data from human and mouse islets31<\/a>,32<\/a><\/sup> led to the identification of three Gs<\/sub>-coupled receptors that are selectively expressed in \u03b1-cells, as compared to other islet cell types. These receptors include GPR119 (gene name: Gpr119)<\/i>, the A2A<\/sub> adenosine receptor (A2AR; gene name: Adora2a<\/i>), and the \u03b21<\/sub>-adrenergic receptor (\u03b21<\/sub>-AR, gene name: Adrb1<\/i>) (Supplementary Fig. 3a, b<\/a>). Because of the availability of highly selective A2AR agonists and antagonists39<\/a><\/sup> and floxed A2AR mice40<\/a><\/sup>, we decided to explore the potential metabolic roles of \u03b1-cell A2ARs. This receptor subtype is also expressed at low to moderate levels in mouse islet \u03b4-cells (Supplementary Fig. 3a<\/a>; Fig. 4a<\/a>).<\/p>\n A previous study using an enzyme-coated electrode biosensor demonstrated that the extracellular levels of adenosine in rodent islets are inversely correlated with glucose levels in the surrounding medium41<\/a><\/sup>. We therefore speculated that adenosine-mediated activation of \u03b1-cell A2ARs might play an important role in promoting glucagon release when glucose levels are low. To test this hypothesis, we perifused islets from WT mice with a selective A2AR agonist, UK 432097 (50\u2009nM). As expected for an agonist acting on an \u03b1-cell Gs<\/sub>-coupled receptor, UK 432097 treatment of WT islets resulted in significant increases in glucagon release at both low and high glucose levels (G3 and G12, respectively), resembling the pattern observed with DCZ-treated \u03b1-GsD islets (Fig. 3a, b<\/a>). This UK 432097 effect was completely abolished by pretreatment of WT islets with SCH 442416 (0.5 \u03bcM), a selective A2AR antagonist, confirming the involvement of A2ARs (Fig. 3a, b<\/a>). Strikingly, at low, but not at high glucose concentrations, application of SCH 442416 alone caused a pronounced decrease in glucagon release (Fig. 3a<\/a>), suggesting that the \u03b1-cell A2AR signaling is required to maintain sufficient glucagon release under hypoglycemic conditions.<\/p>\n a<\/b>, b<\/b> Measurement of glucagon secretion from perifused mouse WT islets. Experiments were carried out at low and high glucose levels (3\u2009mM [G3] and 12\u2009mM [G12], respectively) either in the presence of UK432097 (A2AR-selective agonist: 50\u2009nM at G3, 10\u2009nM at G12) or SCH442416 (A2AR-selective antagonist, 0.5\u2009\u03bcM) alone or in the presence of both ligands (n<\/i>\u2009=\u20094 mice per group). cAMP production in \u03b1-cells from \u03b1-CAMPER mice at G3 (c<\/b>) or G12 (d<\/b>) in the presence of UK432097 (100\u2009nM) or SCH442416 (0.5\u2009\u03bcM) or in the presence of both ligands. e<\/b>\u2013h<\/b> Glucagon release studies with islets lacking G\u03b1s<\/sub> or A2ARs in their \u03b1-cells. Islets were prepared from \u03b1-GsKO and \u03b1-A2AR-KO mice and their corresponding littermates. In (e<\/b>), islets were treated with ADA (5\u2009U\/ml) at G3 to enzymatically remove extracellular adenosine (n<\/i>\u2009=\u20093 or 4 mice per group). In (f<\/b>, g<\/b>), \u03b1-GsKO and control islets were treated with UK432097 at G3 and G12 (n<\/i>\u2009=\u20093 or 4 mice per group). AOC values (h<\/b>) for glucagon release data shown in (f<\/b>) and (g<\/b>) (time period: 8\u201330\u2009min). i<\/b>, j<\/b> A2AR activation stimulates glucagon secretion from human islets. Islets from human donors were perifused with G3 (i<\/b>) or G12 (j<\/b>), respectively, either in the presence of vehicle (DMSO), UK432097 or SCH442416 alone, or in the presence of both UK432097 and SCH442416 (n<\/i>\u2009=\u20093 donors per group). Islets were obtained from male or female mice that were 14\u201324 weeks old. AOC values were calculated for different stimulation periods. Data are shown as means\u2009\u00b1\u2009SEM (3 or 4 independent perifusions with 75\u2013100 islets per perifusion chamber). Data were analyzed via two-tailed Student\u2019s t<\/i> test (AOC values in a<\/b>, b<\/b>, e<\/b>, h<\/b>\u2013j<\/b>) or two-way repeated measures ANOVA with time (c<\/b>, d<\/b>). Numbers above horizontal lines in the bar graphs represent p values. ADA, adenosine deaminase. AOC, area of the curve. Source data are provided as a Source Data<\/a> file.<\/p>\n<\/div>\n<\/div>\n A2AR-mediated activation of Gs<\/sub> is predicted to increase intracellular cAMP levels via Gs<\/sub>-induced activation of adenylyl cyclase. To monitor A2AR-stimulated cAMP accumulation in \u03b1-cells, we employed islets from \u03b1-CAMPER mice that express a cAMP biosensor exclusively in \u03b1-cells36<\/a><\/sup>. We found that A2AR agonist treatment (UK 432097, 5 or 20\u2009nM) of \u03b1-CAMPER islets resulted in a small reduction in cAMP levels at G3 but caused a significant increase in cAMP accumulation at G12 (Fig. 3c, d<\/a>). The A2AR agonist-induced decrease in cAMP levels at G3 is probably due to the facts that A2ARs are already strongly stimulated by high endogenous adenosine levels41<\/a><\/sup> and that \u03b1-cell cAMP levels are already high in a low glucose environment42<\/a><\/sup>. In agreement with these observations, addition of the A2AR antagonist SCH 442416 (0.5\u2009\u03bcM) led to a very robust reduction of cAMP levels at G3 (Fig. 3c<\/a>), raising the possibility that the inhibitory effect of the A2AR agonist at G3 was caused by A2AR desensitization or the activation of other, yet unknown, signaling pathways that interfere with adenosine-induced cAMP production. On the other hand, at G12, \u03b1-cell A2AR signaling is predicted to be reduced due to low extracellular adenosine levels41<\/a><\/sup>, explaining why the A2AR agonist (UK 432097) promoted cAMP production and the A2AR antagonist (SCH 442416) caused only a minor reduction in cAMP levels, as compared to basal levels prior to the addition of ligands (Fig. 3d<\/a>).<\/p>\n To confirm the involvement of intraislet adenosine in stimulating \u03b1-cell A2ARs at low glucose levels (G3), we treated control islets with adenosine deaminase (ADA, 5\u2009U\/ml) which leads to the conversion of adenosine to inosine, a metabolite that is unable to activate A2ARs43<\/a>,44<\/a><\/sup>. As shown in Fig. 3e<\/a>, ADA treatment of control islets resulted in a pronounced decrease in glucagon secretion at G3. Strikingly, this effect was absent in islets prepared from mice that selectively lacked A2ARs receptors or the \u03b1-subunit of Gs<\/sub> (G\u03b1s<\/sub>) in \u03b1-cells (\u03b1-A2A-KO mice and \u03b1-GsKO mice, respectively; see below) (Fig. 3e<\/a>). Moreover, basal glucagon secretion at G3 was drastically reduced in \u03b1-A2A-KO islets (Fig. 3e<\/a>). Taken together, these data strongly suggest that \u03b1-cell A2AR\/Gs<\/sub> signaling plays a key role in stimulating sufficient glucagon release under hypoglycemic conditions.<\/p>\n Previous studies have shown that increases in intracellular Ca2+<\/sup> levels resulting from the activation of \u03b1-cell Gq<\/sub>-coupled receptors can also trigger glucagon release from \u03b1-cells35<\/a>,45<\/a>,46<\/a><\/sup>. Prompted by this finding, we also studied islets from \u03b1-GCaMP6s mice (\u03b1-GCaMP6s islets) that express a Ca2+<\/sup> reporter exclusively in \u03b1-cells36<\/a><\/sup>. Treatment of \u03b1-GCaMP6s islets with UK 432097 (A2AR agonist) or SCH 442416 (A2AR antagonist) had no significant effect on intracellular Ca2+<\/sup> levels at G3 or G12 (Supplementary Fig. 3c, d<\/a>), confirming that G proteins of the Gq<\/sub> family do not contribute to \u03b1-cell A2AR-mediated glucagon secretion.<\/p>\nStimulation of \u03b1-cell Gs<\/sub> signaling leads to hyperglucagonemia and hyperinsulinemia in vivo<\/h3>\n
<\/a><\/div>\n
Stimulation of \u03b1-cell Gs<\/sub> signaling promotes glucagon release from perifused mouse islets<\/h3>\n
Activation of \u03b1-cell Gs<\/sub> signaling improves glucose tolerance in both lean and obese mice<\/h3>\n
<\/a><\/div>\n
Key role of \u03b1-cell adenosine A2A<\/sub> receptors in regulating \u03b1-cell function<\/h3>\n
<\/a><\/div>\n