Dual elimination of the glucagon and GLP-1 receptors in mice reveals plasticity in the incretin axis (original) (raw)
Glp1r is not required for pancreas enlargement or α cell hyperplasia in Gcgr–/– mice. Body weight, food intake, physical activity, and energy expenditure were comparable in WT, single incretin receptor knockout, and Gcgr–/–Glp1r–/– mice (Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI43615DS1). As Gcgr–/– mice exhibit increased pancreatic mass and marked islet and α cell hyperplasia (25) and GLP-1R activation promotes expansion of islet and pancreatic mass (29–31), we assessed the contribution of Glp1r to the development of these abnormalities in Gcgr–/– mice. Consistent with previous findings, Gcgr–/– mice exhibited very high circulating levels of GLP-1 (Supplemental Figure 2A), significantly increased pancreas weight, and an approximately 4-fold increase in islet area (Figure 1); however, islet area and pancreatic mass remained significantly increased to a similar extent in Gcgr–/–Glp1r–/– mice (Figure 1, A and B). Immunohistochemical analysis revealed that the increased islet area was predominantly due to α cell hyperplasia, with most Gcgr–/– and Gcgr–/–Glp1r–/– islets containing a core of β cells surrounded by an expanded mantle of hyperplastic α cells (Figure 1C). Hence, Glp1r is not required for development of increased pancreatic mass and islet hyperplasia following loss of Gcgr action.
Glp1r is not required for development of increased pancreas weight or α cell hyperplasia in Gcgr–/– mice. (A) Pancreas weight of 20- to 24-week-old mice shown as percentage of the final body weight (n = 7–20 mice per group). (B) Islet area shown as a percentage of total pancreas area (n = 4–12 mice per group) (C) Representative histological sections of pancreas stained for insulin or glucagon alone. Final magnification, ×80. Values are expressed as mean ± SEM. *P < 0.05, Gcgr–/– mice versus WT littermate controls; #P < 0.05, Gcgr–/–Glp1r–/– mice versus WT littermate control mice; †P < 0.05, Glp1r–/– versus Gcgr–/–Glp1r–/– mice.
Disruption of Glp1r leads to increased fasting glycemia in Gcgr–/– mice. Both fasting and random glucose were reduced in Gcgr–/– mice (Figure 2, A and B), consistent with the central role of glucagon in the maintenance of euglycemia (25). Basal GLP-1R signaling also regulates fasting glycemia (32), classically through suppression of glucagon secretion (33). Surprisingly, a significant increase in fasting glucose was observed in Gcgr–/–Glp1r–/– compared with Gcgr–/– mice (Figure 2A) and elimination of the Glp1r normalized random-fed glycemia in Gcgr–/–Glp1r–/– mice (Figure 2B). Hence, loss of Glp1r substantially attenuates improvements in both ambient and fasting glycemia in Gcgr–/– mice (25).
Glp1r controls fasting and fed glycemia in Gcgr–/– mice. (A) Blood glucose following 5 or 16 hours of fasting in 8- to 12-week-old WT, Gcgr–/–, Glp1r–/–, and Gcgr–/–Glp1r–/– mice (n = 5–30 mice per genotype). (B) Weekly random-fed blood glucose levels in 8- to 20-week-old Gcgr–/–Glp1r–/–, Gcgr–/–, Glp1r–/–, and littermate control WT mice (n = 3–20 mice per group). Values are expressed as mean ± SEM. *P < 0.05, Gcgr–/– versus Gcgr–/–Glp1r–/– mice; #P < 0.05, Gcgr–/– versus WT mice; ‡P < 0.05, Glp1r–/– versus WT mice; †P < 0.05, Glp1r–/– versus Gcgr–/–Glp1r–/– mice.
Elimination of Glp1r reverses improvements in i.p. glucose tolerance in Gcgr–/– mice. To elucidate the contribution of enhanced GLP-1 receptor signaling to improved β cell function and glucose tolerance in Gcgr–/– mice (25), we first assessed clearance of i.p. glucose in mice of different genotypes. Intraperitoneal glucose tolerance was significantly enhanced and plasma insulin levels increased in Gcgr–/– mice; conversely, i.p. glucose tolerance was impaired in Glp1r–/– mice (Figure 3A), consistent with previous studies (25, 32). Furthermore, disruption of Glp1r in Gcgr–/– mice reversed the improvements in i.p. glucose tolerance and normalized plasma insulin levels in Gcgr–/–Glp1r–/– mice (Figure 3, A and B), whereas insulin sensitivity, approximated by insulin tolerance, was comparable among all genotypes (Figure 3C). Hence, the elevated levels of GLP-1 leading to increased GLP-1R signaling is primarily responsible for enhanced β cell function and improved glucose clearance after i.p. glucose challenge in Gcgr–/– mice.
Loss of Glp1r reverses improvements in i.p. glucose tolerance without altering insulin sensitivity in Gcgr–/– mice. (A) IPGTT in 8- to 10-week-old WT, Gcgr–/–, Glp1r–/–, and Gcgr–/–Glp1r–/– mice (n = 9–24 mice per group). (B) Area under the curve and plasma insulin levels at 0 and 15 minutes following i.p. glucose challenge (n = 4–8 mice per group). (C) Insulin tolerance test in 12- to 14-week-old mice; values are normalized to basal glucose, with right graph showing area under the glucose curve (n = 5–20 mice per group). Values are expressed as mean ± SEM. *P < 0.05, Gcgr–/– versus Gcgr–/–Glp1r–/– mice; #P < 0.05, Gcgr–/– versus WT mice; ‡P < 0.05, Glp1r–/– versus WT mice; †P < 0.05, Glp1r–/– versus Gcgr–/–Glp1r–/– mice.
The GLP-1 receptor mediates reduced gastric emptying; however, oral glucose tolerance remains improved independent of Glp1r in Gcgr–/– mice. As GLP-1–mediated reduction in gastric emptying may substantially account for the improved oral glucose tolerance in Gcgr–/– mice (25, 26), we quantified gastric emptying with two complementary methods. Both liquid-phase gastric emptying, assessed via measurement of plasma acetaminophen levels, and solid-phase gastric emptying were significantly reduced in Gcgr–/– mice and normalized in Gcgr–/–Glp1r–/– mice (Figure 4, A and B). We hypothesized that normalization of gastric emptying would be associated with deterioration of oral glucose tolerance in Gcgr–/–Glp1r–/– mice (25). Unexpectedly, oral glucose tolerance remained significantly improved in Gcgr–/–Glp1r–/– mice to an extent comparable to that in Gcgr–/– mice alone (Figure 4, C and D). Furthermore, in contrast to the normalization of plasma insulin levels seen following i.p. glucose challenge in Gcgr–/–Glp1r–/– versus Gcgr–/– mice (Figure 3B), Gcgr–/–Glp1r–/– mice continued to exhibit significantly increased levels of plasma insulin following oral glucose challenge (Figure 4D).
Glp1r mediates reduced gastric emptying but not improved oral glucose tolerance in Gcgr–/– mice. (A) Liquid-phase gastric emptying (as determined by the appearance of acetaminophen in the circulation after 15 minutes) in 10- to 11-week-old mice (n = 4–14 mice per group). (B) Solid-phase gastric emptying in 20-week-old mice (n = 4–10 mice per group). Values are expressed as mean ± SEM. *P < 0.05, Gcgr–/– versus Gcgr–/–Glp1r–/– mice; #P < 0.05, Gcgr–/– versus WT mice. (C) Blood glucose levels during an OGTT in 10- to 11-week-old mice (n = 11–22 mice per group). (D) Area under the glucose curve and plasma insulin levels 0 and 15 minutes following oral glucose challenge (n = 4–9 mice per group). Values are expressed as mean ± SEM. In C and D: *P < 0.05, Gcgr–/– versus WT littermate control mice; #P < 0.05, Gcgr–/–Glp1r–/– versus WT mice; †P < 0.05, Glp1r–/– versus Gcgr–/–Glp1r–/– littermate control mice; ‡P < 0.05, Glp1r–/– versus WT littermate control mice.
Islets from Gcgr–/–Glp1r–/– mice display increased sensitivity to GIP. As Glp1r–/– mice exhibit enhanced GIP secretion and increased sensitivity to GIP (28), we explored whether GIP-related mechanisms underlie the enhanced enteral glucose-stimulated insulin secretion in Gcgr–/–Glp1r–/– mice. Although levels of GIP were modestly elevated in Glp1r–/– mice, GIP levels were not significantly increased in Gcgr–/–Glp1r–/– mice (Supplemental Figure 2B). We next examined control of insulin secretion from WT, Glp1r–/–, Gcgr–/–, and Gcgr–/–Glp1r–/– islets. No significant differences across genotypes were detected in response to 16.7 mM glucose with a modest but nonsignificant reduction in insulin secretion observed with Gcgr–/–Glp1r–/– islets (Figure 5A). Consistent with the loss of the Glp1r, the insulinotropic response to exendin-4 was absent in Glp1r–/– and Gcgr–/–Glp1r–/– islets (Figure 5A). Although GIP sensitivity was not enhanced in Glp1r–/– or Gcgr–/– islets, the insulinotropic response to GIP was significantly increased in Gcgr–/–Glp1r–/– islets (Figure 5A).
Function of GPCRs in isolated islets. Islet insulin secretion was assessed by preincubation of islets in KRB for 60 minutes at 2.8 mM glucose at 37°C before distribution in batches of 10 islets per condition into wells containing 16.7 mM glucose with or without (A) exendin-4 (Ex-4, 10 nM), [d-Ala2]GIP (GIP, 10 nM), (B) PACAP (10 nM), tolbutamide (Tol, 100 μM), or l-arginine (L-arg, 10 mM) for 1 hour at 37°C. Levels of insulin in the secretion medium were normalized to levels of islet insulin content and are expressed as a fold change in insulin secretion relative to WT high-glucose treatment. Insulin content values averaged approximately 30–40 ng/islet for Glp1r–/– and WT mice and 15–25 ng/islet for Gcgr–/– and Gcgr–/–Glp1r–/– mice (n = 3 mice per group). Data shown are representative of 2–3 independent experiments, each with 3 replicates per condition. (C) Total cellular and media cAMP in islets from WT, Glp1r–/–, Gcgr–/–, and Gcgr–/–Glp1r–/– mice was quantified following treatment of the islets with 0, 1, 3, or 10 nM [d-Ala2]GIP. Levels of cAMP in the secretion medium were normalized to levels of islet insulin content and are expressed as a fold change in islet cAMP levels relative to WT high-glucose treatment (n = 3 mice per group). Values are expressed as mean ± SEM. §P < 0.05, Glp1r–/– versus Gcgr–/–Glp1r–/– mice; #P < 0.05, Gcgr–/– versus Gcgr–/–Glp1r–/– mice; ‡P < 0.05, Gcgr–/–Glp1r–/– versus WT mice; †P < 0.05, Glp1r–/– versus WT mice; *P < 0.05, Gcgr–/– versus WT mice.
To explore the selectivity of the enhanced response to GIP, we tested a range of other insulin secretagogues. In contrast to the enhanced response to GIP, Gcgr–/–Glp1r–/– islets exhibited a normal response to pituitary adenylate cyclase–activating peptide (PACAP) but significantly reduced insulin secretory responses to tolbutamide and l-arginine (Figure 5B). Consistent with the increased GIP sensitivity demonstrated for insulin secretion (Figure 5A), Gcgr–/–Glp1r–/– islets also exhibited enhanced cAMP accumulation in response to GIP (Figure 5C). To evaluate GIP sensitivity in vivo, we administered i.p. glucose in the presence or absence of submaximal doses of exogenous GIP to WT, Glp1r–/–, Gcgr–/–, and Gcgr–/–Glp1r–/– mice. Both Glp1r–/– and to a greater extent Gcgr–/–Glp1r–/– mice exhibited enhanced sensitivity to GIP, as revealed by reduced glycemic excursions and increased circulating levels of plasma insulin in response to exogenous GIP administration (Figure 6, A–D, and Supplemental Figure 3, A–D).
[![Gcgr–/–Glp1r–/– mice exhibit enhanced sensitivity to [d-Ala2]GIP.](https://dm5migu4zj3pb.cloudfront.net/manuscripts/43000/43615/small/JCI43615.f6.gif)
](/articles/view/43615/figure/6)Figure 6
Gcgr–/–Glp1r–/– mice exhibit enhanced sensitivity to [d-Ala2]GIP. An IPGTT was performed in 20- to 22-week-old (A) WT, (B) Gcgr–/–, (C) Glp1r–/–, and (D) Gcgr–/–Glp1r–/– mice following treatment with 1 nmol/kg [d-Ala2]GIP or saline (vehicle [Veh]). Insets depict the area under the glucose excursion curve (AUC) in mM×min and plasma insulin levels at 0 and 15 minutes following glucose challenge (n = 5–8). Values are expressed as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, [d-Ala2]GIP–treated versus saline-treated group.
Plasticity of the incretin axis revealed through reduction of Gcgr action in Glp1r–/–Gipr–/– mice. The available data strongly suggested that preservation of improved glucose tolerance and enhanced insulin secretion despite loss of GLP-1 action reflects increased GIP sensitivity in Gcgr–/–Glp1r–/– islets. To more rigorously test this hypothesis, we reduced Gcgr expression using ASOs in mice lacking both functional incretin receptors, i.e., Glp1r–/–Gipr–/– (double incretin receptor knockout [DIRKO]) mice (34). Hepatic Gcgr mRNA transcripts were markedly decreased and plasma levels of GLP-1 progressively increased in WT and DIRKO mice following ASO treatment (Figure 7, A and B, respectively). WT mice treated with Gcgr ASOs showed improved i.p. glucose tolerance and increased insulin levels following i.p. glucose challenge (Figure 7C). In contrast, DIRKO mice treated with Gcgr ASOs showed no improvement in glucose tolerance or insulin levels following i.p. glucose challenge (Figure 7D), consistent with the importance of the GLP-1R for improved β cell function following reduction of Gcgr expression (21, 22, 32). WT mice treated with Gcgr ASOs also showed improved oral glucose tolerance and significantly higher plasma insulin levels than vehicle-treated controls (Figure 7E). Remarkably, despite loss of both classical incretin receptors, DIRKO mice treated with Gcgr ASOs also exhibited improved oral glucose tolerance and significantly increased plasma insulin levels (Figure 7F). Hence, preferential improvement of glucose tolerance and enhanced β cell function following glucose administration in the gut can be achieved despite loss of both incretin receptors.
Enteroinsular axis is maintained in DIRKO mice treated with Gcgr ASOs. (A) mRNA expression of Gcgr in the liver (n = 3 per group) following treatment with 6 injections of 25 mg/kg Gcgr ASOs. (B) Total plasma GLP-1 levels following 2, 4, or 6 injections of 25 mg/kg saline or Gcgr ASOs (n = 5 per group). (C and D) An i.p. glucose challenge was performed on 13- to 14-week-old male (C) WT mice and (D) DIRKO mice that had been treated with 3 injections of vehicle or 25 mg/kg Gcgr ASOs (n = 5 per group). (E and F) An OGTT was performed on 15- to 16-week-old (E) WT mice and (F) DIRKO mice that had been treated with 4 injections of vehicle or 25 mg/kg Gcgr ASOs (n = 5 per group). Insets depict plasma insulin levels at 0 and 15 minutes following glucose challenge for saline- or Gcgr ASO–treated mice (n = 5 per treatment group). Values are expressed as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, vehicle- versus Gcgr ASO–treated WT or DIRKO mice.
To identify mechanisms responsible for improvement of oral glucose tolerance and β cell function despite absence of both GLP-1 and GIP receptors, we assessed the expression of insulinotropic receptors in islets from (a) DIRKO mice treated with Gcgr ASOs and (b) Gcgr–/–Glp1r–/– mice. Remarkably, levels of mRNA transcripts for the insulinotropic receptors gastrin-releasing peptide receptor (Grpr), Cckar, and Gpr119 were significantly increased in islet RNA from Gcgr–/–Glp1r–/– mice (Figure 8A). Similarly Gpr119 and Cckar mRNA transcripts were also significantly increased in islet RNA from DIRKO mice following Gcgr ASO administration (Figure 8B). These findings raised the possibility that increased activity of related insulinotropic receptors may compensate for the loss of GLP-1 and GIP action on islet β cells.
Expression of insulinotropic GPCRs in islets. (A) Islets were isolated from WT, Gcgr–/–, Glp1r–/–, and Gcgr–/–Glp1r–/– mice, followed by isolation of mRNA for real-time PCR of basal levels of transcripts encoding Gipr, Pacapr, Gpr40, Grpr, Cckar, and Gpr119. (B) Islets were isolated from WT or DIRKO mice following 6 injections of vehicle or 25 mg/kg Gcgr ASOs, and mRNA levels of Grpr, Gpr119, and Cckar were determined. Levels of transcripts were normalized to levels of cyclophilin for each RNA sample. n = 4 mice per genotype. Values are expressed as mean ± SEM. §P < 0.05, Glp1r–/– versus Gcgr–/–Glp1r–/– mice; #P < 0.05, Gcgr–/– versus Gcgr–/–Glp1r–/– mice; ‡P < 0.05, Gcgr–/–Glp1r–/– versus WT mice; ¶P < 0.01, WT Gcgr ASO– vs DIRKO Gcgr ASO–treated mice; *P < 0.01, WT saline- versus DIRKO Gcgr ASO–treated mice; †P < 0.01, DIRKO saline- versus DIRKO Gcgr ASO–treated mice.
Gcgr–/–Glp1r–/– mice display increased sensitivity to Gpr119 and Cckar agonists. To assess the functional significance of increased islet receptor expression, we carried out glucose tolerance tests in the presence or absence of exogenous GRP, CCK, PACAP, and the GPR119 agonist AR231453 (35) in WT, Gcgr–/–, Glp1r–/–, and Gcgr–/–Glp1r–/– mice. We did not detect enhanced sensitivity to exogenous GRP or PACAP (Supplemental Figure 4 and Supplemental Figure 5, A–D). However, although AR231453 failed to improve i.p. glucose tolerance in WT, Gcgr–/–, or Glp1r–/– mice, a robust improvement in glucose tolerance and marked stimulation of plasma insulin levels were observed following AR231453 administration in Gcgr–/–Glp1r–/– mice (Figure 9, A–D, and Supplemental Figure 6, A–D). Similarly, doses of CCK that failed to improve glucose tolerance in WT or single Gcgr–/– or Glp1r–/– mice produced a significant reduction in glycemic excursion and significantly elevated plasma insulin levels in Gcgr–/–Glp1r–/– mice, consistent with increased sensitivity to CCK (Figure 10, A–D, and Supplemental Figure 7, A–D).
Gcgr–/–Glp1r–/– mice exhibit enhanced sensitivity to the GPR119 agonist AR231453. An IPGTT was performed in 22- to 24-week-old (A) WT (B), Gcgr–/– (C), Glp1r–/–, and (D) Gcgr–/–Glp1r–/– mice 30 minutes following treatment with 5 mg/kg AR231453 or vehicle. Insets depict the area under the glucose excursion curve in mM.min and plasma insulin levels at 0 and 15 minutes following glucose challenge (n = 5–8). Values are expressed as mean ± SEM. **P < 0.01, AR231453- versus vehicle-treated mice.
Gcgr–/–Glp1r–/– mice exhibit increased sensitivity to CCK. An IPGTT was performed in 22- to 24-week-old (A) WT, (B) Gcgr–/–, (C) Glp1r–/–, and (D) Gcgr–/–Glp1r–/– mice following treatment with 9 μg/kg of CCK-8 or vehicle. Insets depict the area under the glucose excursion curve (AUC) in mM.min and plasma insulin levels at 0 and 15 minutes following glucose challenge (n = 5–8). Values are expressed as mean ± SEM. *P < 0.05, **P < 0.01, ***P <0.001, CCK-8– versus saline-treated mice.