Glucose controls glucagon secretion by directly modulating cAMP in alpha cells
Glucagon is critical for normal glucose homeostasis and aberrant secretion of the hormone aggravates dysregulated glucose control in diabetes. However, the mechanisms by which glucose controls glucagon secretion from pancreatic alpha cells remain elusive. The aim of this study was to investigate the role of the intracellular messenger cAMP in alpha-cell-intrinsic glucose regulation of glucagon release.
Subplasmalemmal cAMP and Ca2+ concentrations were recorded in isolated and islet-located alpha cells using fluorescent reporters and total internal reflection microscopy. Glucagon secretion from mouse islets was measured using ELISA.
Glucose induced Ca2+-independent alterations of the subplasmalemmal cAMP concentration in alpha cells that correlated with changes in glucagon release. Glucose-lowering-induced stimulation of glucagon secretion thus corresponded to an elevation in cAMP that was independent of paracrine signalling from insulin or somatostatin. Imposed cAMP elevations stimulated glucagon secretion and abolished inhibition by glucose elevation, while protein kinase A inhibition mimicked glucose suppression of glucagon release.
Glucose concentrations in the hypoglycaemic range control glucagon secretion by directly modulating the cAMP concentration in alpha cells independently of paracrine influences. These findings define a novel mechanism for glucose regulation of glucagon release that underlies recovery from hypoglycaemia and may be disturbed in diabetes.
KeywordsCa2+ Cyclic AMP Glucagon release Hypoglycaemia Insulin Pancreatic alpha cell Protein kinase A Somatostatin
Subplasma membrane Ca2+ concentration
Subplasma membrane cAMP concentration
Fluorescence resonance energy transfer
Protein kinase A
Somatostatin receptor type 2
Total internal reflection fluorescence
Yellow fluorescent protein
Glucagon is released from pancreatic alpha cells in response to hypoglycaemia, amino acids, adrenaline (epinephrine), other hormones and neurotransmitters . Its main effect is to mobilise glucose from the liver, thereby preventing dangerous reductions in the blood glucose concentration. Individuals with diabetes often show chronically increased glucagon secretion, which contributes to hyperglycaemia . Hyperglucagonaemia may even be more important for the clinical presentation of diabetes than insulin deficiency . Diabetes is also characterised by impaired glucagon secretion in response to hypoglycaemia, which increases the risk of therapy-related glucopenia . Despite the importance of glucagon for normal glucose homeostasis and the aberrant secretion in diabetes, there is limited understanding of the mechanisms by which glucose controls glucagon release [1, 5, 6].
While extra-pancreatic glucose sensors, including neurons in the brain, contribute to the regulation of glucagon secretion in vivo [7, 8], it is evident that glucose is also able to control glucagon release in denervated preparations, such as isolated islets [9, 10]. Such non-neuronal control may be indirect and mediated by glucose-regulated release of paracrine factors from beta and delta cells that influence alpha cell function. Accordingly, insulin, Zn2+ and the neurotransmitters serotonin, γ-aminobutyric acid and its metabolite γ-hydroxybutyrate from beta cells have been found to suppress glucagon secretion in some [3, 11, 12, 13, 14, 15] but not all [16, 17, 18] studies, and somatostatin from delta cells potently inhibits glucagon release [16, 19, 20]. However, a strong argument against the involvement of beta cell factors is that insulin and glucagon secretion are regulated by glucose in different concentration ranges. Glucagon release is stimulated by hypoglycaemia and maximally suppressed at around 7 mmol/l glucose, which corresponds to the threshold for stimulation of insulin secretion in mouse islets . Somatostatin secretion, on the other hand, is stimulated at the low glucose concentrations that control glucagon secretion . However, glucose elevation also inhibits glucagon secretion in islets from somatostatin knockout mice  and when somatostatin receptor signalling is inhibited [9, 16, 21], indicating somatostatin-independent effects of the sugar.
There are several hypotheses for how glucose elevation inhibits glucagon secretion by alpha-cell-intrinsic mechanisms. For example, glucose has been found to depolarise alpha cells by ATP-sensitive K+ channel closure [10, 18]. The depolarisation causes voltage-dependent inactivation of Na+ and Ca2+ channels, which reduces Ca2+ influx via P/Q-type Ca2+ channels that have been claimed to be particularly important for exocytosis . Other studies have instead indicated that glucose reduces Ca2+ influx by hyperpolarising the alpha cells after glucose-induced inactivation of a store-operated current , by stimulating the Na/K pump  or possibly by activating other channels [25, 26]. Although Ca2+ is regarded as the main trigger of glucagon exocytosis [27, 28], recordings from alpha cells within intact islets have shown that Ca2+ signalling is only modestly or transiently reduced under conditions in which glucose strongly suppresses glucagon release , raising the possibility that Ca2+ plays a permissive role and that exocytosis is regulated by other factors.
cAMP amplifies Ca2+-dependent exocytosis in alpha cells , and a reduction in cAMP was recently suggested to account for glucose-induced paracrine inhibition of glucagon release by insulin, somatostatin  and serotonin . In the present study, we tested the hypothesis that cAMP mediates alpha-cell-intrinsic glucose sensing. Little is known about cAMP regulation in alpha cells. Using total internal reflection fluorescence (TIRF) microscopy and fluorescent reporters combined with measurements of glucagon release, we investigated how cAMP and Ca2+ concentrations in the sub-membrane space ([cAMP]pm and [Ca2+]pm) of alpha cells change in response to glucose and other modulators of glucagon secretion.
HEPES, poly-l-lysine, diazoxide and glutamate were obtained from Sigma (St Louis, MO, USA). Penicillin, streptomycin, glutamine and FBS were from Invitrogen (Carlsbad, CA, USA). The somatostatin receptor type 2 (SSTR2) antagonist PRL2903 was provided by DH Coy (Tulane University, New Orleans, LA, USA). Another SSTR2 antagonist, CYN 154806, and pertussis toxin were obtained from Bio-Techne (Abingdon, UK). A cAMP translocation biosensor was generated as previously described [31, 32]. The biosensor encodes a truncated and membrane-anchored protein kinase A (PKA) regulatory RIIβ subunit, unlabelled or tagged with cyan fluorescent protein, and a catalytic Cα subunit tagged with yellow fluorescent protein (YFP) or the red fluorescent protein mCherry. The red version combined with the unlabelled regulatory subunit was used when simultaneously recording Ca2+. The fluorescence resonance energy transfer (FRET)-based cAMP sensor Epac-SH188  was used for the experiments in Fig. 4. Superfusion and batch incubation of islets were conducted with experimental buffer containing 138 mmol/l NaCl, 4.8 mmol/l KCl, 1.2 mmol/l MgCl2, 1.3 or 2.6 (hormone release) mmol/l CaCl2, 1–20 mmol/l glucose, 25 mmol/l HEPES and 0.5 mg/ml albumin with pH adjusted to 7.40 with NaOH.
Pancreatic islet isolation and culture
Islets of Langerhans were isolated from 6- to 9-month old, normal-weight, C57B16J mice of both sexes . The mice were obtained from Taconic (Ry, Denmark) and housed in ventilated cages (up to 5 animals/cage) with a 12 h dark/light cycle and with free access to water and a standard mouse chow. All procedures for animal handling were approved by the Uppsala Animal Ethics Committee. After isolation, the islets were cultured for 1 day in RPMI 1640 medium containing 5.5 mmol/l glucose, 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin at 37°C in a 5% CO2 humidified air atmosphere. For some experiments, the islets were dispersed to single cells by pipetting in cell dissociation buffer containing 10% (vol./vol.) TrypLE (Thermo Fisher Scientific, Waltham, MA, USA). When all islets were disintegrated, the enzyme digestion was interrupted by adding serum-containing RPMI 1640 medium, followed by centrifugation (5 min at 160 g) and resuspension of the cells in the culture medium.
Human islets from three normoglycaemic cadaveric organ donors (two men, one woman, ages 30–72 years) were obtained via the Nordic Network for Clinical Islet Transplantation in Uppsala. All experiments with human islets were approved by the Uppsala Human Ethics Committee. The isolated islets were cultured for up to 7 days at 37°C in an atmosphere of 5% CO2 in CMRL 1066 culture medium containing 5.5 mmol/l glucose, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mmol/l glutamine and 10% FBS.
Recordings of [cAMP]pm and [Ca2+]pm
Islets and cells were infected with cAMP biosensor adenoviruses during 1.5 h incubation in 50 μl medium (approximately 105 fluorescence forming units/islet), followed by washing and culture for 16–20 h before use. Islets expressing the cAMP biosensor were preincubated for 20–30 min in experimental buffer. For measurements of [Ca2+]pm, the preincubation buffer was complemented with 1.2 μmol/l of the acetoxymethyl ester of the Ca2+ indicator Fluo-4 (Life Technologies, Carlsbad, CA, USA). After washing in indicator-free buffer, the islets were allowed to attach onto poly-l-lysine coated coverslips during 5–10 min. Coverslips with the islets or cells were used as exchangeable bottoms of an open 50 μl chamber that was superfused with experimental medium (0.12–0.2 ml/min) at 37°C. [cAMP]pm and [Ca2+]pm were measured using TIRF or confocal microscopy (see electronic supplementary material [ESM] Methods for details).
Measurements of glucagon and insulin release
Quantification and statistical analysis
Image analysis was conducted using MetaFluor (Molecular Devices Corp, Sunnyvale, CA, USA). Fluorescence intensity changes are presented in regions of interest corresponding to single cells with the signal normalised (after background subtraction) to the mean level during the initial condition. In single-wavelength cAMP recordings, the signal was inverted to achieve a positive relationship between normalised intensity and [cAMP]pm. In ratiometric cAMP recordings, the ratio changes (positively related to [cAMP]pm) were normalised to the initial ratio. Single-cell recordings representative of a given number of cells and experiments (islet donors) are shown, along with quantification of responses as scatter plots or means ± SEM. Statistical comparisons were made using Student’s paired t tests.
Glucose-induced modulation of [cAMP]pm in alpha cells parallels changes in glucagon secretion
TIRF imaging of mouse islets expressing a cAMP biosensor and exposed to 1–3 mmol/l glucose showed that [cAMP]pm was stable in most cells. An increase in the glucose concentration to 20 mmol/l resulted in decrease of [cAMP]pm in cells identified as alpha cells by their positive [cAMP]pm response to 10 μmol/l adrenaline (Fig. 1a–c). The [cAMP]pm lowering started immediately or after a delay of up to 2.3 min. Half-maximal decrease was observed 1.9 ± 0.2 min after the start of the decline. The beta cells within the same islet usually responded with a [cAMP]pm increase after glucose elevation and often, but not always, with adrenaline-induced lowering (Fig. 1b, c). The effect of 7 mmol/l glucose on alpha cell [cAMP]pm was close to maximal and was often characterised by an initial nadir followed by a somewhat less pronounced sustained reduction (see ESM Results and ESM Fig. 2). Some cells showed additional decrease at 20 mmol/l glucose, but the mean effect did not reach statistical significance (Fig. 1d, e). Alpha cells within human islets showed similar [cAMP]pm reductions in response to glucose elevation (Fig. 1f–h). When the glucose concentration was instead lowered from 7 to 1 mmol/l, mouse alpha cells responded with a rise in [cAMP]pm (Fig. 1i) and perifusion experiments revealed stimulated glucagon secretion with strikingly similar kinetics (Fig. 1j). Control experiments in cAMP biosensor-expressing islet alpha cells loaded with the pH indicator BCECF ascertained that the cAMP responses to glucose did not reflect a pH effect on the biosensor (see ESM Results and ESM Fig. 3).
Glucose-induced changes in alpha cell [cAMP]pm show little correlation with [Ca2+]pm
[cAMP]pm lowering by glucose elevation occurs independently of paracrine insulin and somatostatin signalling
Somatostatin was more efficient and lowered [cAMP]pm at 100 pmol/l in three out of 12 cells from four experiments. Of the remaining alpha cells, seven responded to 1 nmol/l, whereas two cells showed [cAMP]pm reductions only after an increase in somatostatin to 100 nmol/l (data not shown). We therefore used 1 nmol/l somatostatin in the remaining experiments and found that 13 out of 16 alpha cells from six experiments responded at this concentration with a reduction in [cAMP]pm (Fig. 3c). The SSTR2 antagonist PRL2903 (5 μmol/l) often increased [cAMP]pm irrespective of the glucose concentration (7/11 alpha cells in three experiments at 1 mmol/l glucose and 7/9 cells in two experiments at 20 mmol/l glucose), but prevented neither [cAMP]pm lowering by increasing glucose from 1 to 7 and 20 mmol/l (Fig. 3d, e), nor inhibition of glucagon secretion by 5 and 20 mmol/l glucose (Fig. 3f). The experiment in Fig. 3f also shows that glucagon release was suppressed by 5 mmol/l glucose without any simultaneous stimulation of insulin secretion, reinforcing that paracrine signalling from beta cells is not important under those conditions.
Another SSTR2 antagonist, CYN 154806, also increased [cAMP]pm at 1 mmol/l glucose in many alpha cells, but did not prevent [cAMP]pm reduction induced by a subsequent elevation in the glucose concentration (Fig. 3g). To exclude the involvement of somatostatin receptors other than SSTR2, we treated the islets with pertussis toxin (200 ng/ml for 18 h) to inhibit the action of Gαi, which transduces signals from all somatostatin receptor subtypes. Pertussis toxin did not affect the [cAMP]pm increase induced by a reduction in the glucose concentration from 7 to 1 mmol/l, but prevented the [cAMP]pm-lowering effect of somatostatin (Fig. 3h–j).
Glucose-induced cAMP modulation involves an alpha-cell-intrinsic mechanism
In an additional approach to prevent paracrine signalling, experiments were performed after dispersion of islets into single cells. Individual alpha cells responded to a 7 to 1 mmol/l reduction in glucose or the addition of adrenaline with increases in [cAMP]pm similar to those in islets (Fig. 4e, f). In addition, dispersed beta cells showed islet-similar responses to glucose reduction and adrenaline, with lowering of [cAMP]pm (Fig. 4g).
Glucose-induced inhibition of glucagon secretion occurs independently of somatostatin and is prevented by fixing cAMP at high levels
PKA inhibition mimics glucose inhibition of glucagon secretion
Since many effects of cAMP are mediated by PKA, we investigated the potential involvement of this kinase in glucagon secretion. Changing glucose from 7 to 1 mmol/l induced a marked stimulation of glucagon secretion (Fig. 6b). Subsequent introduction of the PKA inhibitor Rp-8-CPT-cAMPS (100 μmol/l) reduced secretion by more than 70%, and there was some further reduction at 7 mmol/l of the sugar (Fig. 6b).
Previous studies have provided evidence that Ca2+ is a main trigger of exocytosis in alpha cells [27, 28], and alpha cells show electrical activity and Ca2+ signalling in the presence of low glucose concentrations that stimulate glucagon release in hypoglycaemia [27, 28, 29, 35]. However, increases in glucose concentration, which strongly inhibit glucagon secretion, have modest or transient effects on alpha cell Ca2+ signalling [29, 35]. Therefore, messengers other than Ca2+ may be more important in regulating glucagon release. Exocytosis in alpha cells is known to be strongly dependent on cAMP . The present study provides evidence for a key role of cAMP in glucose-regulated glucagon secretion mediated by a direct effect of the sugar on alpha cells.
We found that glucose modulated [cAMP]pm in both mouse and human alpha cells. Since mouse alpha cells have been reported to express glucagon receptors , and glucagon, at least in some mouse alpha cells, increases [cAMP]pm , glucose-induced decrease of [cAMP]pm might simply be the consequence rather than the cause of inhibited glucagon secretion. Indeed, [cAMP]pm and secretion showed strikingly similar kinetics. However, the glucose-induced changes in alpha cell cAMP persisted when glucagon secretion was inhibited by Ca2+-deficient medium. Moreover, the glucose-induced reduction in glucagon secretion was prevented when the intracellular cAMP concentration was fixed at a high level, either by membrane-permeable cAMP or a phosphodiesterase inhibitor. These observations strongly indicate that lowering of [cAMP]pm underlies the suppression of glucagon secretion.
We previously reported that very high glucose concentrations induce increases in [cAMP]pm with oscillations in a small fraction of alpha cells . The responses sometimes involved alternating increases and decreases of [cAMP]pm above and below the baseline, resembling the pattern of glucagon secretion under similar conditions . The reason for the discrepancy with the presently observed [cAMP]pm reduction by glucose is unknown, but the lowering effect is now extensively substantiated.
Based on immunohistochemical detection of cAMP, it has been suggested that glucose elevation induces paracrine lowering of cAMP in alpha cells by stimulating the secretion of somatostatin and insulin . In other cell types, including adipocytes, insulin is known to promote cAMP degradation by activating phosphodiesterase 3B . However, we found that insulin itself or blockade of its receptor lacked an effect on alpha cell cAMP, indicating that paracrine insulin signalling is not involved. Alpha cells also express somatostatin receptors with domination of SSTR2 [39, 40]. Somatostatin receptors activate Gi/o to suppress adenylyl cyclase activity, and the present study consequently showed that exogenously applied somatostatin lowers [cAMP]pm in most alpha cells, an effect that was prevented when Gi/o signalling was inhibited with pertussis toxin. However, the toxin did not interfere with the glucose reduction-induced increase in [cAMP]pm. The SSTR2 antagonist PRL2903 increased [cAMP]pm, which indicates that endogenous somatostatin affects alpha cell cAMP, but the drug did not prevent glucose-induced [cAMP]pm reduction. Another SSTR2 inhibitor, CYN 154806, also did not influence the glucose-induced reduction in [cAMP]pm, although it strongly increased glucagon secretion. These observations are consistent with previous conclusions that somatostatin has a tonic inhibitory effect on glucagon secretion [9, 16, 21]. More importantly, the data clarify that glucose elevation reduces [cAMP]pm in alpha cells independent of paracrine influences from insulin or somatostatin. It is also unlikely that serotonin  or other factors secreted from islet cells mediate cAMP regulation by glucose, since influx of Ca2+ is required for exocytosis in all major islet cell types [21, 28, 41] and the effect of glucose was maintained under Ca2+-deficient conditions. We therefore conclude that glucose controls cAMP and glucagon release by a direct effect on the alpha cell.
It remains to be elucidated how glucose lowers cAMP. In beta cells, glucose increases [cAMP]pm by activating adenylyl cyclases via elevations of Ca2+  and the cAMP precursor ATP . There are also adenylyl cyclases that are inhibited by Ca2+  and Ca2+-stimulated phosphodiesterases degrading cAMP [44, 45]. Little is known about the adenylyl cyclase and phosphodiesterase protein expression profiles in alpha cells, but several observations indicate that Ca2+ does not mediate the effect of glucose on [cAMP]pm. Although simultaneous recordings of the two messengers indicated that Ca2+ sometimes seems to promote increases in alpha cell [cAMP]pm, there was generally no strong correlation between glucose-induced changes in [cAMP]pm and [Ca2+]pm. Moreover, in accordance with previous studies [29, 35], [Ca2+]pm showed modest glucose dependence in alpha cells. Most importantly, the glucose effect on [cAMP]pm remained when Ca2+ entry was prevented by Ca2+-deficient conditions.
There are alternative hypotheses for the Ca2+-independent effects of glucose on [cAMP]pm. In beta cells, glucose has been found to activate phosphodiesterase 3B in a Ca2+- and insulin-independent manner, associated with changes in enzyme phosphorylation . A similar mechanism may operate in alpha cells. Our previous work has emphasised a role for store-operated Ca2+ entry in regulating alpha cell electrical activity, Ca2+ signalling and glucagon secretion [9, 23]. Interestingly, it has been suggested that cAMP production could be regulated by a store-operated mechanism independently of Ca2+ . Since glucose reduces store-operated signalling in alpha cells , such a mechanism can be expected to result in reduced cAMP production.
The link between cAMP and glucagon secretion was studied in islets cultured overnight, which were found to have a more pronounced and consistent secretory response to repeated low-glucose challenges than freshly isolated islets. The cAMP effects seem to involve PKA, since, in line with a previous report , inhibition of this kinase suppressed glucagon secretion at 1 mmol/l glucose, whereas another study  found no effect of a different PKA inhibitor. cAMP may act in part by activating the voltage-gated Ca2+ channels that mediate exocytosis-triggering Ca2+ influx . The effects are complex and have been suggested to involve PKA-mediated inhibition of N-  or P/Q-type  channels at small cAMP elevations, and Epac-dependent stimulation of L-type channels at higher elevations . However, a quantitatively more important effect of cAMP is to accelerate the mobilisation of granules to the readily releasable pool , consistent with the poor correlation between alpha cell [cAMP]pm and [Ca2+]pm as well as the modest changes in alpha cell [Ca2+]pm that accompany glucose-regulated glucagon secretion. Future studies will clarify whether the residual stimulation of glucagon secretion at 1 mmol/l glucose in the presence of PKA inhibitor (Fig. 6b) is mediated by Epac.
We now propose that glucose concentrations in the hypoglycaemic range regulate glucagon secretion by directly modulating the cAMP concentration in alpha cells. Such a mechanism does not exclude a further role for cAMP in the paracrine regulation of glucagon release [15, 30], which may become dominating during hyperglycaemia . Although Ca2+ is a critical trigger of exocytosis in alpha cells, the ion seems to have a more permissive role and the magnitude of secretion is controlled mainly by cAMP-mediated amplification of granule exocytosis. Further studies are warranted to clarify the mechanisms by which alpha-cell-intrinsic glucose sensing controls glucagon secretion via cAMP, and whether aberrant cAMP signalling underlies the dysregulated glucagon secretion in diabetes.
Some of these data were presented as an abstract at the 52nd EASD annual meeting in 2016.
QY, HS and PA conducted experiments and analysed data. EG and AT designed the experiments, analysed data and wrote the paper. All authors critically revised and approved the final version of the manuscript. AT is the guarantor of this work and, as such, had full access to all the data and takes responsibility for the integrity of data and the accuracy of data analysis.
This study was supported by grants from the Diabetes Wellness foundation, the European Foundation for the Study of Diabetes (EFSD-MSD and EFSD-Novo Nordisk), the Family Ernfors Foundation, the Leona M. and Harry B. Helmsley Charitable Trust, the Novo Nordisk Foundation, the Swedish Diabetes Foundation and the Swedish Research Council. Human islets were obtained from the Nordic Network for Clinical Islet Transplantation, supported by grants from the JDRF and the strategic grant consortium Excellence of Diabetes Research in Sweden (EXODIAB).
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.
- 27.Gromada J, Bokvist K, Ding WG et al (1997) Adrenaline stimulates glucagon secretion in pancreatic A-cells by increasing the Ca2+ current and the number of granules close to the L-type Ca2+ channels. J Gen Physiol 110(3):217–228. https://doi.org/10.1085/jgp.110.3.217 CrossRefPubMedPubMedCentralGoogle Scholar
- 33.Klarenbeek J, Goedhart J, van Batenburg A, Groenewald D, Jalink K (2015) Fourth-generation epac-based FRET sensors for cAMP feature exceptional brightness, photostability and dynamic range: characterization of dedicated sensors for FLIM, for ratiometry and with high affinity. PLoS One 10(4):e0122513. https://doi.org/10.1371/journal.pone.0122513 CrossRefPubMedPubMedCentralGoogle Scholar
- 39.Strowski MZ, Parmar RM, Blake AD, Schaeffer JM (2000) Somatostatin inhibits insulin and glucagon secretion via two receptors subtypes: an in vitro study of pancreatic islets from somatostatin receptor 2 knockout mice. Endocrinology 141(1):111–117. https://doi.org/10.1210/endo.141.1.7263 CrossRefPubMedGoogle Scholar
- 48.Tian G, Tepikin AV, Tengholm A, Gylfe E (2012) cAMP induces stromal interaction molecule 1 (STIM1) puncta but neither Orai1 protein clustering nor store-operated Ca2+ entry (SOCE) in islet cells. J Biol Chem 287(13):9862–9872. https://doi.org/10.1074/jbc.M111.292854 CrossRefPubMedPubMedCentralGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.