It has previously been shown that the incretin effect accounts for ≈50% of the insulin response to oral glucose in normal mice. Now, I have proceeded and studied the contribution of glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1) to the insulin response to oral glucose in female mice by using receptor antagonists. A specific GIP receptor antagonist (mGIP(3-30); 50 or 500 nmol/kg), a specific GLP-1 receptor antagonist (exendin(9-39); 3 or 30 nmol/kg), the combination of mGIP (500 nmol/kg) and exendin(9-39) (30 nmol/kg), or saline was given intravenously four minutes after administration of glucose (50 mg) through a gastric tube in anesthetized C57/BL6J mice (n = 95) with samples obtained before glucose administration and after 15, 30 and 60 min. The insulinogenic index, determined as the area under the 60 min curve for insulin (AUC insulin) divided by the AUC glucose, was used to reflect the insulin response. It was found that the insulinogenic index was reduced by 67 ± 4% by mGIP(3-30) (p< 0.001), by 60 ± 14% by exendin(9-39) (p = 0.007) and by 61 ± 14% by the combination of mGIP(3-30) and exendin(9-39) (p = 0.043), both at their highest doses, compared to animals injected with glucose in the same experimental series. It is concluded that both GIP and GLP-1 are required for a normal incretin effect in female mice, that they contribute similarly to the insulin response, and that it is unlikely that there is another incretin hormone in this species.
As demonstrated in 1964 in human studies, the insulin response is markedly augmented after oral glucose compared to intravenous glucose at identical glucose levels. This is the basis for the incretin effect which is an important physiological process to assure a sufficient insulin response not only after oral glucose but also after meal ingestion, as reviewed several decades ago. The incretin effect is responsible for ≈40–65% of the insulin response to oral glucose in humans. It is mainly mediated by the two gut incretin hormones, glucose-dependent insulinotropic polypeptide (GIP) and glucagon-like peptide-1 (GLP-1); these hormones are both released after the ingestion of oral glucose or meal and potentiate glucose-stimulated insulin secretion at dose levels achieved after meal ingestion.
The relative contribution of the two incretin hormones to the incretin effect has been examined using different strategies. A study in healthy humans used the specific GIP receptor antagonist GIP(3-30) and the specific GLP-1 receptor antagonist exendin(9-39). It was found that the insulin response to oral glucose was reduced more by the GIP antagonist than by the GLP-1 antagonist, suggesting that GIP is more important than GLP-1. This was also concluded in a recent review. Another study on healthy humans used infusions of physiological levels of the two hormones during stepwise glucose clamp and meal ingestion. This study showed that the two hormones nearly equally contribute to the incretin effect.
Since model experiments in mice are important in biomedical research, it is of relevance to examine the incretin effect in this species also. It has thus been demonstrated that the incretin effect indeed exists in mice. This was demonstrated in a study in which glucose was infused intravenously at a variable rate to match glucose levels achieved after the oral administration of 25 mg glucose. The results showed that the insulin response was double after the oral administration of glucose compared to the intravenous infusion. This would suggest that the incretin effect accounts for ≈50% of the insulin response to oral glucose in mice. It is also known that GIP and GLP-1 levels increase after oral glucose in mice. Furthermore, both GIP and GLP-1 have robust effects to stimulate insulin secretion in mice when administered intravenously in mice. Moreover, in mice with double deletion of the GIP and GLP-1 receptors, the insulin response to oral glucose is reduced by ≈65%. It is thus clear that GIP and GLP-1 are important incretin hormones in mice also.
However, the contribution of each of the incretin hormones for the incretin effect in mice is not known. In this study, I have used the approach of administering GIP or GLP-1 receptor antagonists, alone or together, to study the relative contribution of GIP and GLP-1 to the insulin response to oral glucose in female mice. I used exendin(9-39), which has been used in previous studies in my laboratory in mice. As GIP receptor antagonist, I used the recently established antagonist GIP(3-30), which in its human form has been successful in inhibiting the actions of GIP, and which in its murine form (mGIP(3-30)), has been characterized in mice. These antagonists were administered alone or together and in association with gastric glucose administration. The glucose and insulin responses were determined. Since the glucose levels after glucose administration were altered by the receptor antagonists and therefore not matched between experimental groups and controls, the indirect measure of insulinogenic index was used to estimate the beta cell response.
Female C57BL/6J mice from Taconic, Skensved, Denmark were used. They were 4–6 months of age, maintained in a temperature-controlled room (22 °C) on a 12:12 h light-dark cycle (light on at 7:00 A.M.) and fed a standard pellet diet (energy 14.1 MJ/kg with 14% from fat, 60% from carbohydrate and 26% from protein; SAFE, Augy, France) and tap water ad libitum. Only female mice were used, to avoid the stress of single housing, which is used in male mice.
A total of 95 animals were allocated for the study (mean body weight (±SEM) was 20.6 ± 0.2 g). Food was removed from the cages at 7:30 A.M. At 9:30 A.M., anaesthesia was induced with Fluafent (i.e., a mixture of fluanisone and fentanyl citrate) and midazolam, as previously described. In short, 10 mg fluanisone (Key Organics, Camelford, Cornwall, UK) was dissolved in 1 mL sterile water at 70 °C for 60 min. This solution was mixed with 1 mL of fentanyl citrate (Sigma-Aldrich, St. Louis, MO, USA; 0.315 mg/mL); 100 µL of this solution were given intraperitoneally to each mouse (0.016 mg fentanyl citrate and 0.5 mg fluanisone/mouse). Midazolam (0.167 mg/mouse; Roche, Basel, Switzerland) was also given (100 µL/mouse). Fifteen minutes later, 50 mg of glucose (Sigma-Aldrich, St. Louis, MO, USA, dissolved in saline) were administered through a gastric tube (outer diameter 1.2 mm). Four minutes later, mice were injected intravenously in a tail vein (10 µL/g body weight given over 3 s) with synthetic exendin(9-39) (Sigma-Aldrich, 3 or 30 nmol/kg, dissolved in saline), synthetic mGIP(3-30) (CASLO ApS, Kongens Lyngby, Denmark; 50 or 500 nmol/kg; dissolved in saline with addition of NaOH; final concentration 60 µmol/L), the combination of exendin(9-39) (30 nmol/kg) and mGIP(3-30) (500 nmol/kg), or saline. Whole blood was sampled in heparinized pipettes from the intraorbital retrobulbar sinus plexus (40 µL) at 0, 15, 30 and 60 min. Plasma was separated by centrifugation and stored at −20 °C until analysis. The experiments were undertaken in batches of 6–8 mice on each experimental day by one experienced technician. Animals given glucose alone were always included in the experimental series with receptor antagonists.
Glucose in whole blood was detected with the glucose oxidase method using AccuChek Aviva (Hoffman-La Roche, Basel, Switzerland). Insulin was determined by ELISA (Mercodia, Uppsala, Sweden). The intra-assay coefficient of variation (CV) of the method is 4% at low and high levels, and the inter-assay CV is 5% at low and high levels. The lower limit of the assay is 6 pmol/L.
All individual results from the completer population were included in the analysis. Data are presented as means ± SEM. Suprabasal (incremental) areas under the curves (AUCs) were calculated with the trapezoid rule. The insulinogenic index was estimated by dividing AUC insulin by AUC glucose. Whether AUC glucose, AUC insulin and the insulinogenic index were normally distributed was tested by analyzing the estimates after glucose administration alone in all animals, using the Kolmogorov–Smirnov test. Since the data points for AUC glucose and AUC insulin were normally distributed and the data points for the insulinogenic index were not normally distributed, differences between groups were determined using Student’s t test for tests of significance for AUC glucose and AUC insulin and non-parametric tests were used for tests of significance for insulinogenic index. Tests were performed between animas tested in the same experimental series. For example, groups of animals given mGIP(3-30) at 50 nmol/kg had their glucose controls and animals given mGIP(3-30) at 500 nmol/kg had their glucose controls. Statistical tests were performed by SPSS, v 28 (IBM Corp., Statistics for Windows, Version 28.0. Armonk, NY, USA). Statistical significance was defined as p< 0.05.
Supplementary Figure S1 shows glucose and insulin data in all animals given glucose alone (n = 45). It is seen that glucose levels peaked at 30 min and insulin levels peaked at 15 min. The figure also shows individual data points for AUC glucose, AUC insulin and the insulinogenic index. The Kolmogorov–Smirnov test for normality showed that the individual data points for AUC glucose (p = 0.18) and AUC insulin (p = 0.20) were normally distributed, whereas the data points for the insulinogenic index were not normally distributed (p = 0.001).
The GIP receptor antagonist mGIP(3-30) or saline was given intravenously at four minutes after the oral administration of 50 mg glucose. At the low dose of 50 nmol/kg, mGIP(3-30) did not significantly affect the glucose or insulin responses to glucose, except a slightly higher glucose level in the mGIP(3-30) group at 30 min (p =0.031) (Supplementary Figure S2). At 500 nmol/kg, mGIP(3-30) significantly increased the 15, 30 and 60 min glucose levels (p = 0.008, p = 0.014 and p = 0.026, respectively) and significantly reduced the 15 min insulin levels (p = 0.010) compared to glucose alone (Figure 1). Furthermore, AUC glucose, AUC insulin and the insulinogenic index were not significantly different between mGIP(3-30) at 50 nmol/kg and the glucose control (Table 1). In contrast, AUC glucose was significantly higher after mGIP(3-30) at 500 nmol/kg compared to glucose alone (p< 0.001), whereas AUC insulin did not differ significantly between the groups (p = 0.148). The insulinogenic index was significantly lower in the mGIP(3-30) group (p< 0.001) (Table 1). The mean insulinogenic index was reduced by 67 ± 4% (p< 0.001) by mGIP(3-30) at 500 nmol/kg compared to glucose alone.
