Glucagon Receptor Knockout Mice Display Increased Insulin Sensitivity and Impaired β-Cell Function

Gcgr^−/− mice were previously generated by homologous recombination in embryonic stem cells on a WW6 × C57BL/6J background (14). Mice were bred at Taconic M&B, Ry, Denmark, and male and female mice of the 10th backcross onto the C57Bl/6J strain were used for the experiments, which were performed at Lund University. Mice were 13 to 16 weeks old at the time of the clamp and intravenous glucose tolerance test (IVGTT) experiments and 20 weeks old when pancreatic islets were isolated. Age-, sex-, and body weight–matched C57BL/6J mice from the same breeding facility were used as controls. The study was approved by the animal ethics committee at Lund University, Sweden, and by the Danish Animal Experiments Inspectorate.

Mice were anesthetized with an intraperitoneal injection of midazolam (12.5 mg/kg) (Dormicum; Hoffman-La Roche, Basel, Switzerland) and a combination of fluanison (25 mg/kg) and fentanyl (0.78 mg/kg) (Hypnorm; Janssen, Beerse, Belgium). This anesthesia persisted for at least 1 h. In experiments lasting longer, administration of anesthetics was repeated after 60 min. Animals were kept on a heating pad during the entire procedure.

Hyperinsulinemic-euglycemic clamp studies were performed in anesthetized female Gcgr^−/− and C57BL/6J mice. The clamp experiments were carried out as previously described by Ahren and colleagues (15). Briefly, before the experiment, the right jugular vein (infusion) and left carotid artery (blood sampling) were catheterized. Thirty minutes after introduction of the catheters, synthetic human insulin (Actrapid; Novo Nordisk, Bagsværd, Denmark) was infused as a primed (40 mU) continuous (20 mU · kg⁻¹ · min⁻¹) infusion. The volume load was 4 μl for the 1st minute, followed by 2 μl/min thereafter. Blood glucose levels were determined at 5-min intervals for 90 min by the glucose dehydrogenase technique with the use of a Hemocue glucometer (Hemocue, Ängelholm, Sweden). Sample volume was 5 μl whole blood. A variable rate of glucose (solution of 40 g/dl) was infused to maintain blood glucose levels at 5–6 mmol/l. At 0 and 90 min, a blood sample (75 μl) was taken from the arterial catheter; plasma was separated by centrifugation (5,000g, 5 min) and kept at −20°C until determination of plasma insulin. Glucose disappearance was represented by the average glucose infusion rate (GIR) during steady state (the final 30 min of the clamp).

Mice were anesthetized as above, and a blood sample (50 μl) was taken from the retrobulbar intraorbital capillary plexus into a 100-μl glass pipette prerinsed in a heparin solution (100 units/ml in 0.9% NaCl; Lövens, Ballerup, Denmark). Blood was dispersed into a centrifuge tube and plasma separated by centrifugation (5,000g, 5 min). Plasma was kept at −20°C until analysis. At time 0, d-glucose (10 g/dl; British Drug Houses, Poole, U.K.) was injected intravenously over 3 s at a dose of 1 g/kg in a tail vein. In other experiments, the synthetic cholecystokinin octapeptide (CCK-8; 18 μg/kg), arginine (0.25 g/kg), or the cholinergic agonist carbachol (30 μg/kg) was intravenously injected. In one series of experiments, exendin-3 [exendin-3(9-39); 30 nmol/kg; Sigma], or saline in control mice, was intraperitoneally administered 5 min before intravenous injection of 1 g/kg glucose. Exendin-3 is a GLP-1 receptor antagonist (16), which has been shown to inhibit insulin secretion in response to exogenous GLP-1 (17) and oral glucose (18) in mice. The volume load in each administration was 10 μl/g body wt. Blood samples (50 μl each) were collected before and at 1, 5, 10, 20, 50, and 75 min after the intravenous injection, as described above. Blood samples were centrifuged and plasma kept at −20°C until measurement of glucose and insulin concentrations. At the end of the experiment, mice were given 1 ml 0.9% NaCl s.c. and were kept on a heating blanket until they regained consciousness.

For insulin secretion studies, pancreatic islets were isolated from male mice with the collagenase digestion technique, as modification of the original technique described by Lacy and Kostianovsky (19). Briefly, after a midline laparotomy, the common bile duct was cannulated and ligated at the papilla of Vater. The pancreas was filled with 3 ml ice-cold Hanks’ balanced salt solution (Sigma) supplemented with 0.4 mg/ml collagenase P (Boehringer Mannheim, Mannheim, Germany) before removal and incubation at 37°C for 19 min. After washing three times in Hanks’ balanced salt solution, the islets were handpicked under a stereomicroscope. The isolated islets were then incubated overnight in an RPMI medium supplemented with 2.06 mmol/l l-glutamine (Life Technologies, Täby, Sweden), 10% fetal bovine serum, 100 units/ml penicillin, and 0.5 mg/ml streptomycin (Kebo Laboratory, Spånga, Sweden) at 37°C in 5% CO₂ and 95% O₂. After the overnight incubation, islets were washed in a HEPES buffer containing 125 mmol/l NaCl, 5.9 mmol/l KCl, 1.28 mmol/l CaCl₂, 1.2 mmol/l MgCl₂, 25 mmol/l HEPES, 3.3 mmol/l glucose, and 0.1% fatty acid–free serum albumin (pH 7.36) and preincubated in the same medium for 60 min. Thereafter, three islets were incubated in 100 μl of the medium for 60 min at 37°C in the presence of glucose at varying concentrations (2.8, 5.6, 8.3, 11.1, 16.7, or 22.2 mmol/l) or in the presence of 11.1 mmol/l glucose and either CCK-8 (100 nmol/l), arginine (10 mmol/l), GLP-1 (100 nmol/l), carbachol (100 μmol/l), glucagon (1 μmol/l), glucose-dependent insulinotropic polypeptide (GIP; 100 μmol/l), or forskolin (2.5 μmol/l) (all secretagogues purchased from Sigma). After incubation, aliquots of 25 μl in duplicate were collected and stored at −20°C until analysis of insulin concentration.

Glucose oxidation was measured in isolated islets from Gcgr^−/− and wild-type mice. Batches of 30 islets in quadruplicates were incubated in a reaction mixture containing 0.1 μCi or 0.7 μCi [¹⁴C]glucose (specific activity 310 mCi/mmol; NEN, Boston, MA) as tracer at final concentrations of 2.8 or 16.7 mmol/l glucose. The reaction was terminated after incubation of the samples for 2 h in 37°C, and the amount of released ¹⁴CO₂, trapped with benzetonium hydroxide, was determined by scintillation counting.

Insulin concentration in plasma and incubation media was determined by a double-antibody radioimmunoassay using guinea pig anti-rat insulin antibody, ¹²⁵I-labeled human insulin, and, as standard, rat insulin (Linco Research, St. Charles, MO). Plasma glucose was measured by the glucose oxidase technique.

Statistical analysis was performed using GraphPad Prism (version 4.02; Graphpad, Monrovia, CA). The rate of glucose disappearance after bolus glucose administration was evaluated by the glucose tolerance index K_G (percent per minute), calculated as the slope of the regression versus time of the logarithmically transformed glucose concentration values from 1 to 20 min. The acute insulin response (AIR) to intravenous administration of the secretagogues was calculated as the suprabasal mean 1- and 5-min insulin levels. Data are represented as means ± SE and were compared using Student’s t test or two-way ANOVA when applicable. A P value <0.05 was considered statistically significant.

To investigate the insulin sensitivity of Gcgr^−/− mice, hyperinsulinemiceuglycemic clamp studies were performed. Basal blood glucose levels were lower in Gcgr^−/− compared with wild-type mice (5.3 ± 0.4 vs. 9.4 ± 0.7 mmol/l, respectively; P < 0.001), and basal insulin levels were also reduced in Gcgr^−/− compared with wild-type mice (57 ± 9 vs. 224 ± 67 pmol/l, respectively; P < 0.01). Figure 1A shows the mean blood glucose levels during the hyperinsulinemic-euglycemic clamp experiments and demonstrates that both groups of mice were clamped at a blood glucose level of 6 mmol/l. The mean steady-state GIR required to clamp blood glucose at 6 mmol/l during the insulin infusion was elevated in Gcgr^−/− compared with wild-type mice (29.6 ± 3.1 vs. 18.8 ± 1.3 mg · kg⁻¹ · min⁻¹, respectively; P < 0.01) (Fig. 1B). Insulin levels at the end of the clamp experiments were similar in Gcgr^−/− and wild-type mice (9,280 ± 1,740 vs. 9,970 ± 1,200 pmol/l; P = NS).

To investigate the glucose-stimulated insulin release in vivo, IVGTT was performed in Gcgr^−/− and wild-type mice (Fig. 2). Basal glucose was, again, reduced in Gcgr^−/− compared with wild-type mice (6.6 ± 0.3 vs. 8.6 ± 0.3 mmol/l, respectively; P < 0.001). Basal insulin levels were also lower in Gcgr^−/− compared with wild-type mice (73 ± 14 vs. 163 ± 16 pmol/l, respectively; P < 0.001). The AIR (mean suprabasal 1 and 5 min insulin) to intravenously administered glucose was increased in Gcgr^−/− compared with wild-type mice (2,270 ± 350 vs. 870 ± 100 pmol/l, respectively; P < 0.01). Plasma glucose increased to similar levels in Gcgr^−/− and wild-type mice, but the difference in insulin responses was reflected by the faster glucose clearance in Gcgr^−/− mice, as calculated by the K_G index, which is based on the slope of the glucose curves after logarithmic transformation of the values (3.3 ± 0.5 vs. 1.3 ± 0.1 %/min in wild-type mice; P < 0.001).

To further examine insulin secretion in Gcgr^−/− mice, we tested the in vivo insulin-stimulating effects of intravenous administration of three nonglucose secretagogues: carbachol, arginine, and CCK-8 (Fig. 3A–C). The AIR to carbachol did not differ significantly between the groups, although a tendency toward a lower insulin response was observed in the Gcgr^−/− mice (Fig. 3D). In contrast, the AIR to CCK-8 (P < 0.05) and arginine (P < 0.001) were both significantly lower in Gcgr^−/− compared with wild-type mice (Figs. 3D).

The effect of the GLP-1 receptor antagonist exendin-3 on glucose-stimulated insulin secretion was examined in Gcgr^−/− and wild-type mice (Fig. 4). Exendin-3 decreased the AIR in Gcgr^−/− mice by 60% compared with vehicle-treated Gcgr^−/− mice (1,750 ± 130 vs. 4,810 ± 640 pmol/l, respectively; P < 0.001). In contrast, exendin-3 did not significantly affect the AIR to intravenous glucose in wild-type mice compared with vehicle-treated wild-type mice (275 ± 60 vs. 430 ± 80 pmol/l, respectively; P = NS). Although insulin secretion was significantly reduced in Gcgr^−/− mice injected with exendin-3, glucose elimination was similar to that of vehicle-treated Gcgr^−/− mice (K_G 8.4 ± 0.4 vs. 8.2 ± 0.5 %/min). Also, there was no difference in glucose elimination between exendin-3 and vehicle-treated wild-type mice (3.0 ± 0.3 vs. 2.6 ± 0.2 percent/min, respectively).

To study the effects of glucose and nonglucose insulin secretagogues isolated from systemic effects, pancreatic islets were isolated from Gcgr^−/− and wild-type mice and incubated in the presence of either increasing concentrations of glucose or with different insulin secretagogues (glucagon, GLP-1, carbachol, CCK-8, arginine, and GIP) in the presence of 11.1 mmol/l glucose. Further, we investigated the effect of cAMP production on insulin secretion by incubating islets from wild-type and Gcgr^−/− mice in the presence of glucose and forskolin, a compound known to stimulate cAMP production.

In response to increasing glucose concentration in the incubation media, insulin secretion increased in pancreatic islets from both Gcgr^−/− and wild-type mice (Fig. 5A). However, the insulin response in Gcgr^−/− mice was significantly reduced compared with that of wild-type mice at the highest glucose concentrations (16.7 [P < 0.05] and 22.2 mmol/l [P < 0.01]).

The exposure of wild-type and Gcgr^−/− islets to various insulin secretagogues in the presence of 11.1 mmol/l glucose had varied effects on islet insulin secretion. In islets from wild-type mice, carbachol and GLP-1 were strong stimulators, while arginine, CCK-8, and GIP had less of an effect (Fig. 5B). Secretagogue-stimulated insulin secretion was significantly reduced in islets from Gcgr^−/− compared with wild-type mice, most markedly with GLP-1, GIP, and carbachol.

To further elucidate the mechanism behind the altered insulin secretion in islets from Gcgr^−/− mice, glucose oxidation activity was measured in isolated islets at 2.8 and 16.7 mmol/l glucose (Fig. 6). At the low glucose level, no difference in glucose oxidation was detected between Gcgr^−/− and wild-type islets. With increased glucose levels, however, glucose oxidation was significantly reduced in islets from Gcgr^−/− mice compared with that of islets from wild-type mice.

In this study, we have characterized the insulin sensitivity and β-cell function in mice with a targeted deletion in the glucagon receptor (Gcgr^−/−). Previous characterization of Gcgr^−/− mice demonstrated increased glucose tolerance as measured by intraperitoneal glucose tolerance test, while insulin tolerance tests indicated that insulin sensitivity was similar in Gcgr^−/− and wild-type mice (14). In contrast, in the present studies, we demonstrated that the mean steady-state GIR during hyperinsulinemic-euglycemic clamp was increased in Gcgr^−/− compared with wild-type mice. Thus, when measured under steady-state conditions, whole-body insulin sensitivity seems to be increased in the Gcgr^−/− mice. An apparent increase in glucose utilization in Gcgr^−/− mice may be explained by the increased lean body mass previously detected in Gcgr^−/− mice (14). A higher percentage of muscle tissue mass would increase the total amount of glucose taken up by muscle during the clamp, without necessarily affecting insulin sensitivity per se. Increased insulin sensitivity may also be due to the reduced adiposity in Gcgr^−/− mice, in view of the impairment of insulin sensitivity that often is associated with obesity (20–22). On the other hand, the difference in lean body mass between Gcgr^−/− and wild-type mice is only modest (10% increase in lean body mass [14]) and may not account for the entire difference in glucose uptake demonstrated by the clamp study. A most likely explanation is instead that the absence of glucagon signaling in Gcgr^−/− mice has resulted in reduced HGP and, during clamp conditions, to unopposed effect of insulin on suppression of HGP. In a hyperinsulinemic-euglycemic clamp experiment without tracer infusion, this effect would be seen as an increase in the GIR necessary to keep blood glucose at the clamped level, as was demonstrated in the present study. When examining islet function in Gcgr^−/− mice, we found that the AIR to intravenous glucose was increased and, consequently, glucose clearance faster than in wild-type mice. These findings are in agreement with the improved glucose tolerance found in previous studies of two different Gcgr^−/− mouse strains (14,23) and in mice with a targeted deletion in the gene encoding prohormone convertase 2 (PC2^−/−), an enzyme that participates in the processing of proglucagon to glucagon (24). In contrast, we found that the insulin response to glucose from isolated islets was reduced in Gcgr^−/− mice. This would suggest that glucagon signaling in β-cells is required for a normal insulin secretion but that in Gcgr^−/− mice, compensatory mechanisms in vivo have counteracted this to allow an augmented insulin response upon intravenous administration of glucose. One such compensatory mechanism would be GLP-1. Small amounts of GLP-1 are normally produced in the pancreas (25,26), and, as previously reported, in Gcgr^−/− mice, pancreatic and plasma levels of biologically active GLP-1 are increased in parallel with the increased production of glucagon (14). It is thus possible that GLP-1 augments and restores the impaired insulin secretion in vivo in these mice. To test this hypothesis, mice were injected with the GLP-1 receptor antagonist exendin-3, a truncated form of exendin-4 [exendin-4(9-39)-amide] (16), which previously has been shown to inhibit the insulin response to GLP-1 in mice (17). Indeed, exendin-3 significantly inhibited the insulin response to intravenous glucose in Gcgr^−/− mice. Despite the inhibition with exendin-3, insulin secretion was still threefold elevated in Gcgr^−/− compared with wild-type mice. This marked elevation in insulin secretion likely explains the maintenance of glucose elimination found in the exendin-3–treated Gcgr^−/− mice. These results demonstrate that circulating GLP-1 may be responsible for the strong augmentation of glucose-stimulated insulin secretion (GSIS) in Gcgr^−/− mice. However, other mechanisms, such as indirect neural effects (27), may also contribute to the augmented insulin secretion in response to intravenous glucose found in Gcgr^−/− mice in vivo.

Our data also show that absence of glucagon signaling in the Gcgr^−/− mice results in a reduced β-cell response not only to glucose but also to the nonglucose stimuli carbachol, arginine, CCK-8, GLP-1, and GIP. Although the islets were incubated with secretagogues in the presence of 11.1 mmol/l glucose, the insulin response in Gcgr^−/− islets was reduced to a degree that could not be explained solely by the impairment in GSIS observed in these islets. The studied secretagogues signal to insulin secretion through different pathways in the islets. Therefore, the results indicate that disruption of glucagon signaling effects late events in the insulin secretory pathways, suggesting that the presence of glucagon signaling is required for a normal β-cell secretion through several pathways. Incubation in the presence of forskolin augmented insulin secretion in islets from wild-type and Gcgr^−/− mice. Still, the insulin response to forskolin was markedly reduced in Gcgr^−/− mice, indicating that the cAMP production may be downregulated in Gcgr^−/− as a consequence of the absent glucagon signaling.

An intriguing observation in the present results was that the response to GLP-1 in isolated islets was impaired, yet GLP-1 seems to account for a major part of the augmentation of GSIS in an IVGTT. These findings indicate that factors other than GLP-1 are involved in augmenting GSIS in vivo in Gcgr^−/− mice.

Thus, while compensatory and protective mechanisms seem to preserve insulin secretion in Gcgr^−/− mice in vivo, the ablation of glucagon receptor signaling causes a more severe perturbation in isolated islets. Further studies are required to describe the nature of the compensatory mechanisms leading to the observed discrepancies between the in vivo and in vitro β-cell function in Gcgr^−/− mice.

Glucose oxidation in the β-cell precedes GSIS. To further investigate the role of glucagon signaling in GSIS, we measured glucose oxidation in pancreatic islets isolated from wild-type and Gcgr^−/− mice in the presence of low and high concentrations of glucose. We found that, at high (16.7 mmol/l) glucose concentrations, the oxidation of glucose was markedly reduced in islets from Gcgr^−/− mice. Thus, absence of glucagon receptors appears to reduce islet glucose metabolism, which may explain the blunted insulin secretion in islets.

A nonlinear inverse relation has been documented between insulin sensitivity and β-cell function, meaning that reduced insulin sensitivity is accompanied by increased insulin secretion in order to maintain normoglycemia (20,22,28). To obtain a true image of β-cell function, insulin secretion should therefore not be evaluated in isolation but instead be viewed in context of the level of insulin sensitivity in the individual. The inverse relation between insulin sensitivity and β-cell function also includes individuals with increased insulin sensitivity. In a study of professional athletes (29), insulin sensitivity was increased, while insulin secretion was reduced in the athletes compared with a group of sedentary control subjects matched for age and BMI. This indicates that reduced insulin secretion may not imply reduced β-cell function when viewed in the context of increased insulin sensitivity. This adaptation of β-cells to the improved insulin sensitivity may also be the cause of the reduced insulin secretion observed in our studies of Gcgr^−/− mice.

In summary, Gcgr^−/− mice display increased insulin sensitivity measured during hyperinsulinemic-euglycemic clamp conditions. Furthermore, the insulin response to intravenous glucose was increased by a mechanism sensitive to GLP-1 receptor antagonism and therefore probably dependent on the augmented GLP-1 levels. At the same time, the glucose-stimulated insulin response was blunted in isolated islets at high glucose concentrations. The insulin response to nonglucose secretagogues was reduced in Gcgr^−/− mice in vivo and in isolated islets. In addition, glucose oxidation in isolated islets from Gcgr^−/− mice was reduced. This suggests that glucagon signaling in pancreatic β-cells acts to augment insulin secretion. The reduced insulin secretion may also be a result of an islet adaptation to increased insulin sensitivity and constantly lowered blood glucose levels.

This study was supported by grants from the Swedish Research Council (6834), the Swedish Diabetes Association, the Påhlssons foundation, Region Skåne, and the Faculty of Medicine, Lund University.

We thank Lena Kvist, Lillian Bengtsson, and Kristina Andersson for expert technical assistance.