β Cell Gαs signaling is critical for physiological and pharmacological enhancement of insulin secretion

Abstract

The incretin peptides glucose-dependent insulinotropic polypeptide and glucagon-like peptide-1 receptors coordinate β cell secretion that is proportional to nutrient intake. This effect permits consistent and restricted glucose excursions across a range of carbohydrate intake. The canonical signaling downstream of ligand-activated incretin receptors involves coupling to Gαs protein and generation of intracellular cAMP. However, recent reports have highlighted the importance of additional signaling nodes engaged by incretin receptors, including other G proteins and β-arrestin proteins. Here, the importance of Gαs signaling was tested in mice with conditional, postdevelopmental β cell deletion of Gnas (encoding Gαs) under physiological and pharmacological conditions. Deletion of Gαs/cAMP signaling induced immediate and profound hyperglycemia that responded minimally to incretin receptor agonists, a sulfonylurea, or bethanechol. While islet area and insulin content were not affected in Gnasβcell-/-, perifusion of isolated islets demonstrated impaired responses to glucose, incretins, acetylcholine, and IBMX In the absence of Gαs, incretin-stimulated insulin secretion was impaired but not absent, with some contribution from Gαq signaling. Collectively, these findings validate a central role for cAMP in mediating incretin signaling, but also demonstrate broad impairment of insulin secretion in the absence of Gαs that causes both fasting hyperglycemia and glucose intolerance.

Keywords: Diabetes; Endocrinology; G proteins; Insulin; Metabolism.

Figure 1. Characterization of Gnas^{βcell–/–} mouse islets.

(A) Gnas and Gnaq expression in β cell– and α cell–enriched populations (n = 4–5). (B) Ambient fed glycemia over time in 6- to 8-week-old control (n = 29) and Gnas^{βcell–/–} (n = 23) mice at start of tamoxifen delivery (day 0). (C) Body weight of control (n = 24) and Gnas^{βcell–/–} (n = 15) mice and its correlation with fed glycemia. (D) Average islet size and its correlation with blood glucose at the time of sacrifice in control (n = 7) and Gnas^{βcell–/–} (n = 9) mice and its correlation with fed glycemia. (E) Insulin-positive area per total pancreas area in control (n = 7) and Gnas^{βcell–/–} (n = 9) mice. (F) Insulin granule number (localizations/μm²) from control and Gnas^{βcell–/–} mice, with representative images of insulin granules (n = 41 cells from 3 mice per group). Dashed box represents the area selected for zoom, shown in right-hand panel. Scale bars: 18.9 μm (left panels), 1.98 μm (right panels). (G) Proinsulin levels at baseline (t = 0) and 10 minutes after meal challenge with Ensure (t = 10). Data are shown as mean ± SEM, *P < 0.05 as indicated. Data were analyzed by 2-tailed Student’s t test (A and C–G), 2-way ANOVA (B and G), or linear regression (C–E). dSTORM, direct stochastic optical reconstruction microscopy.

Figure 2. Gnas^{βcell–/–} mice are hyperglycemic and do not secrete insulin in response to glucose or meal challenge.

(A) Intraperitoneal glucose tolerance test (IPGTT) (1.5 g/kg) in control (n = 24) and Gnas^{βcell–/–} mice (n = 15) and insulin at baseline (t = 0) and 10 minutes after injection (t = 10) in control (n = 20) and Gnas^{βcell–/–} mice (n = 13). (B) Mixed-meal tolerance test with Ensure (10 μL/g) in control (n = 23) and Gnas^{βcell–/–} mice (n = 15) and insulin at baseline (t = 0) and 10 minutes after injection (t = 10) in control (n = 23) and Gnas^{βcell–/–} mice (n = 15). (C) Insulin levels in 5-hour-fasted mice, (D) glycemia after 5-hour fast, and (E) the insulin/glucose ratio for 5-hour-fasted mice (n = 22, 15). Data are shown as mean ± SEM, *P < 0.05 as indicated. Data were analyzed by 2-way ANOVA of glycemia data (A and B) or 2-tailed Student’s t test (A–E).

Figure 3. Gnas^{βcell–/–} islets have an impaired insulin secretory response to both cAMP-dependent and -independent stimulation.

(A) Insulin secretion from control or Gnas^{βcell–/–} islets in response to increasing glucose doses, acetylcholine (Ach; 10 nM), IBMX (100 μM), and KCl (30 mM) and iAUC of each treatment (n = 3). (B) cAMP traces in control and Gnas^{βcell–/–} islets (n = 55, 34). (C) Representative blot of pPKA substrates in control- or IBMX-treated islets from control or Gnas^{βcell–/–} mice. (D) The independent and combined effects of Gnas deletion or treatment with Ex9 on insulin secretion in response to an IBMX ramp (n = 3). (E) Insulin secretion from an FSK ramp in control (n = 4) or Gnas^{βcell–/–} (n = 7) islets and iAUC of each treatment (n = 3). (F) Insulin secretion in control and Gnas^{βcell–/–} islets in response to ramping concentrations of Sp-8-BnT-cAMPS (n = 4, 5). Data are shown as mean ± SEM, *P < 0.05 as indicated. Data were analyzed by 2-way ANOVA of iAUC (A and D–F).

Figure 4. Loss of Gnas in β cells partially impairs incretin-stimulated insulin secretion.

(A) Insulin secretion in response to GIP (3 nM) stimulation in control or Gnas^{βcell–/–} islets in the presence or absence of GIPR antibody (300 nM) and iAUC of GIP-stimulated insulin secretion (n = 1–6). (B) Insulin secretion in response to GLP-1 (10 pM) stimulation in control or Gnas^{βcell–/–} islets in the presence or absence of Ex9 (1 μM) and iAUC of GLP-1–stimulated insulin secretion (n = 5–7). (C) Insulin secretion in response to GIP (3 nM) stimulation in control or Gnas^{βcell–/–} islets in the presence or absence of the Gq inhibitor YM254890 (100 nM) and iAUC of GIP-stimulated insulin secretion (n = 5–6). (D) Insulin secretion in response to GLP-1 (10 and 300 pM) stimulation in control or Gnas^{βcell–/–} islets in the presence or absence of the Gq inhibitor YM254890 (100 nM) or Ex9 (1 μM) and iAUC of GLP-1–stimulated insulin secretion (n = 5–6). Data are shown as mean ± SEM, *P < 0.05 as indicated. Data were analyzed by 2-way ANOVA of iAUC.

Figure 5. β Cell Gnas expression is necessary for incretin-stimulated insulin secretion in vivo.

(A–D) Control (n = 34) or Gnas^{βcell–/–} (n = 23) mice were treated with PBS, D-Ala2-GIP (A and B), or Ex4 (C and D) at t = –10 minutes. Mice were then challenged with i.p. glucose (1.5 g/kg) and iAUC presented from t = 0. Insulin secretion in D-Ala2-GIP–challenged (B, n = 23,14) and Ex4-challenged (D, n = 21,13) mice are shown at baseline (t = 0) and 10 minutes after glucose challenge (t = 10). Data are shown as mean ± SEM, *P < 0.05 as indicated. Data were analyzed by 2-way ANOVA of glycemic curves and insulin levels or 2-tailed Student’s t test of the iAUCs.

Figure 6. β Cell Gnas is necessary for glucose lowering in response to Gαs-independent, islet-targeted therapies.

(A) Glycemia and iAUC from control (n = 28–34) and Gnas^{βcell–/–} mice (n = 16–23) treated with PBS or bethanechol (Beth; 2 mg/kg) by i.p. injection 10 minutes before (t = –10) i.p. glucose challenge (t = 0; 1.5 g/kg). (B) Insulin at baseline (t = 0) and 10 minutes after glucose injection (t = 10) in control (n = 18) and Gnas^{βcell–/–} mice (n = 7). (C) Glycemia, percent of baseline glycemia, and integrated area above the curve (iAAC) from tolbutamide (100 mg/kg) gavage in control (n = 28) and Gnas^{βcell–/–} (n = 16) mice. (D) Glycemia, percent of baseline glycemia, and iAAC from insulin tolerance test (1 U/kg) in control (n = 8) and Gnas^{βcell–/–} (n = 7) mice. (E) Glycemia after dapagliflozin (10 mg/kg) gavage in control (n = 19) and Gnas^{βcell–/–} (n = 14) mice. SGLT2i, SGLT2 inhibitor. Data are shown as mean ± SEM, *P < 0.05 as indicated. Data were analyzed by 1-way ANOVA of glycemia (E), 2-way ANOVA of glycemic curves and insulin levels, or 2-tailed Student’s t test of the iAUCs.