Defective Glucose-Dependent Insulinotropic Polypeptide Receptor Expression in Diabetic Fatty Zucker Rats

Synthetic human GIP and GLP-1 were purchased from Bachem California (Torrance, CA), and 3-isobutyl-1-methylxanthine (IBMX) was purchased from Research Biochemicals International (Natick, MA). All chemicals of reagent or molecular biology grade were from Sigma (Oakville, ON, Canada) or Fisher Scientific International (Pittsburgh, PA).

VDF rats spontaneously developed from a Zucker strain kept in our laboratory. These diabetic rats were homozygous recessive for a mutation in the leptin receptor gene, fa (Gln₂₆₉Pro). Rats carrying one normal Fa allele were phenotypically lean and displayed normal glucose tolerance. Only male animals aged 14–16 weeks were used in this study. All animals tested displayed severe glucose intolerance and decreased first-phase insulin response characteristic of VDF rats. Lean littermates were used as control animals in these experiments.

Zucker rats were anesthetized with an intraperitoneal (IP) injection of sodium pentobarbital (65 mg/kg) (Somnotol; MTC Pharmaceuticals, Cambridge, ON, Canada). The right jugular vein was then exposed and cannulated with heparinized polyethylene tubing (PE50; Becton Dickinson, Sparks, MD). GIP (4 pmol · min^–1 · kg^–1) or physiological saline was infused (30 μl/min) via the cannula using an infusion pump (Harvard Apparatus, South Natick, MA) for 5 min before an IP glucose injection (40%, 1 g/kg). Blood samples (0.5 ml) were collected from the tail vein in heparinized Caraway/Natelson collecting tubes (Fisher Scientific International) 5 min before (basal) and 10, 20, 30, and 60 min after glucose injection. Concomitantly, blood glucose measurements were taken before the infusion started (basal) and then every 10 min after glucose administration using a handheld blood glucose meter (SureStep; LifeScan, Burnaby, BC, Canada). Plasma was then separated from red cells by centrifugation at 10,000g for 20 min at 4°C and then stored at –20°C until GIP and insulin radioimmunoassays could be carried out (32).

Anesthesia was established using Somnotol, and pancreases were isolated as previously described (16). The perfusate consisted of a modified Krebs-Ringer bicarbonate buffer containing 3% dextran and 0.2% bovine serum albumin (BSA) (Fraction V, RIA grade; Sigma) gassed with 95% O₂/5% CO₂ to achieve a pH of 7.4. The abdominal aorta was perfused at a rate of 4 ml/min, and portal venous outflow was collected at 1-min intervals. After a 10-min equilibration period, the pancreatic perfusion continued with 4.4 mmol/l glucose for 4 min followed by 8.8 mmol/l glucose for the remainder of the experiment. GIP (10 pmol/l), GLP-1 (50 pmol/l), or saline was introduced into the perfusion system from 20 to 40 min. The different doses of GIP and GLP-1 were chosen because they had equal insulinotropic activities in the lean Zucker perfused pancreas.

Rat pancreatic islets were isolated as previously described (33). Briefly, the rat was anesthetized, and a midline incision was made. The common bile duct was cannulated, and the pancreas was inflated with collagenase (320 mg/l, Type XI; Sigma) in Hank’s balanced salt solution supplemented with 10 mmol/l HEPES, 2 mmol/l l-glutamine, and 0.2% BSA (HBSS+) (Life Technologies, Burlington, ON, Canada). The pancreas was then removed from the rat and macerated with scissors before collagenase digestion. The pancreatic tissue was initially digested in a shaking 37°C water bath for 20 min and 10 min for pancreases from lean and fat rats, respectively; a second digestion was then carried out for 10 and 7 min for the lean and fat rats, respectively. After the collagenase digestions, the islets were passed through a 1-mm nylon screen and separated from exocrine tissue via centrifugation (1,000g at 4°C) through a discontinuous dextran gradient. Finally, islets were picked under a dissecting microscope, washed in HBSS+, and used for mRNA isolation or cultured in RPMI 1640 with 8.8 mmol/l glucose, 10% fetal calf serum (Cansera, Rexdale, ON, Canada), antibiotics (50 U/ml each penicillin G and streptomycin), 0.07% human serum albumin, 0.0025% human apotransferrin, 25 pmol/l sodium selenite, and 20 μmol/l ethanolamine hydrochloride for 20–24 h in 10-cm plastic culture dishes (Falcon; Becton Dickinson) in a humidified 5% CO₂ environment.

After the culture period, 40 healthy islets were selected and sandwiched between two layers of Cytodex-3 beads (Amersham Pharmacia, Baie d’Urfé, PQ, Canada) in 0.2-ml polyethylene perifusion chambers (Endotronics, Coon Rapids, MN). The chambers were then perifused in an Acusyst-s perifusion apparatus (Endotronics) under a humid 37°C 5% CO₂ environment at a flow rate of 0.5 ml/min with 10 mmol/l HEPES-buffered Krebs-Ringer bicarbonate buffer (KRBH) supplemented with 0.2% BSA. Perifusion experiments were carried out for 80 min after a 60-min equilibration period in 2.8 mmol/l glucose (low glucose) perifusate. After 20 min, the perifusate glucose concentration was switched to 16 mmol/l with or without GIP. Samples were collected every 2 min, and insulin levels were determined by radioimmunoassay as previously described (32).

After an overnight culture, 40 healthy islets were selected, washed twice with 0.5 ml of KRBH supplemented with 0.2% BSA, and allowed to equilibrate for 30 min in a humidified 5% CO₂ environment. The islets were then stimulated with either vehicle, 10 μmol/l forskolin, or 10 nmol/l GIP for 30 min in the presence of 0.5 mmol/l IBMX. The islets were then lysed by boiling for 5 min in 0.05N hydrochloric acid. Samples were then dried by vacuum centrifugation (Speed-Vac; Sorvall, Farmingdale, NY) and stored at –20°C for cAMP radioimmunoassay (Biomedical Technologies, Stoughton, MA).

Rat islet RNA was isolated immediately after islet isolation using Trizol and the standard protocol supplied by the manufacturer (Life Technologies). Specifically, 1 ml Trizol reagent was used per 100 islets, and all A₂₆₀/A₂₈₀ ratios of isolated RNA were ≥1.80. After RNA isolation, 1 μg islet RNA was subjected to reverse transcription (RT). Total RNA was reverse transcribed in a volume of 10 μl containing 0.5 mmol/l deoxynucleotide triphosphate, 15 pmol gene-specific primer targeted at the COOH-terminus of the rat GIP receptor (GTT CTG GAG TAG AGG TCC GTG TA), 75 pmol random hexamers (Amersham Pharmacia), 100 U Superscript II RNase H⁻ Reverse Transcriptase (Life Technologies), 10 U RNase inhibitor (RNA Guard; Amersham Pharmacia), 1 mmol/l dithiothreitol, 50 mmol/l Tris-HCl, pH 8.3, 75 mmol/l KCl, and 3 mmol/l MgCl₂. After RT, 100 ng rat islet tissue cDNA was used in the real-time polymerase chain reaction (PCR) to measure GIP receptor expression, whereas 10 ng cDNA was used in the GAPDH control PCR. The PCR mix consisted of 1× TaqMan Buffer A (PE Applied Biosystems, Foster City, CA), 10 mmol/l MgCl₂; 200 μmol/l dATP, dCTP, and dGTP and 400 μmol/l dUTP; 200 nmol/l rat GIP receptor 5′ forward primer (5′-CCG CGC TTT TCG TCA TCC-3′), 200 nmol/l rat GIP receptor 3′ reverse primer (5′-CCA CCA AAT GGC TTT GAC TT-3′), 200 nmol/l GIP receptor probe colabeled with the fluorescent dyes FAM and TAMRA (5′-CCC AGC ACT GCG TGT TCT CGT ACA GG-3′), 0.01 U/μl AmpErase uracil-DNA glycosylase (UNG) (PE Applied Biosystems), and 0.025 U/μl of AmpliTaq Gold (PE Applied Biosystems). The GAPDH reactions included the above reaction conditions with the exception of the primers and probe, which were purchased from PE Applied Biosystems and were directed toward rodent GAPDH. PCRs were carried out in triplicate in the PE Applied Biosystems 7700 sequence detection system. The reaction profile included a 10-min preincubation at 50°C to allow the UNG to degrade any uracil-containing nucleic acids and a further 10 min incubation at 94°C to activate the AmpliTaq Gold. After these preincubations, a two-step PCR protocol was carried out, which included a denaturation step at 94°C for 15 s followed by a 1-min annealing/extension step at 60°C. Fluorescence was measured during the annealing/extension steps over 40 cycles and used to calculate a cycle threshold, i.e., the point at which the reaction is in the exponential phase and is detectable by the hardware. All reactions followed the typical sigmoidal reaction profile, and cycle threshold was used as a measure of amplicon abundance (34).

Islets were isolated as previously described. After isolation, islet GIP receptor protein was analyzed as previously described (35). Briefly, islets were lysed in ice-cold radioimmunoprecipitation assay buffer (150 mmol/l NaCl, 20 mmol/l Tris-Cl, pH 7.5, 1 mmol/l EDTA, 1% Nonidet P-40, 1% deoxycholate, 0.1% SDS, 5 mmol/l NaF, 1 mmol/l phenymethylsulfonyl fluoride, 1 mmol/l dithiothreitol [DTT], 10 μg/ml leupeptin, 10 μg/ml pepstatin A, 10 μg/ml bestatin, and 1% Trasylol [Bayer Pharmaceuticals, Etobicoke, ON, Canada]) for 30 min on ice. Protein concentration was determined using a Bicinchoninic Acid kit (Pierce, Rockford, IL). Then, 50 μg total islet protein was denatured under reducing conditions (100 mmol/l DTT) at 100°C for 5 min and run by SDS-PAGE. Proteins were then transferred to nitrocellulose membrane, blocked with 5% skim milk (in Tris-buffered saline with 0.5% Tween 20 [TBST]), and incubated with a well-characterized polyclonal anti-GIP receptor antibody (35). Membranes were then washed three times in TBST and incubated with horseradish peroxidase–conjugated goat anti-rabbit IgG secondary antibody (Jackson Immunoresearch Laboratories, West Grove, PA). After further washing, the immunoreactive bands were visualized using enhanced chemiluminescence (Amersham Pharmacia). Finally, bands were subjected to densitometry using Eagle Eye II software (Stratagene, La Jolla, CA), and molecular weight was determined using relative mobility (R_f) analysis.

All data are expressed as mean ± SE, with the sample size indicated in the appropriate figure legend. Unpaired two-tailed t tests were carried out to compare groups of animals. P values ≤0.05 were considered statistically significant. Area under the curve was determined using curve analysis software (Graphpad; Prism, San Diego, CA).

The optimal GIP dose was determined by carrying out the bioassay with various GIP concentrations ranging from 2 pmol · min^–1 · kg^–1 to 2 nmol · min^–1 · kg^–1 (data not shown). The optimum dose determined from this study (4 pmol · min^–1 · kg^–1) produced a submaximal glucose-lowering response in the lean animals. Thus, this dose was used for the remainder of the experiments. Figure 1A shows the blood glucose response to an IP glucose tolerance test with and without GIP in the lean control animals. This figure clearly shows that GIP was able to significantly improve glucose tolerance in the lean animals as early as 30 min after IP glucose injection. Furthermore, this improvement in glucose tolerance was maintained as long as the GIP infusion was continued. Figure 1B shows that the integrated glucose response (over 60 min) for the lean animals receiving GIP was significantly smaller than that in animals receiving saline.

The fat animals displayed basal hyperglycemia, with an average of 8.3 ± 0.4 mmol/l, compared with 5.5 ± 0.2 mmol/l for the lean animals (Figs. 1A and C). The fat animals also had much higher peak glucose levels in response to the glucose challenge, with the control (saline) values peaking at 15 mmol/l (compared with 11 mmol/l in lean animals); thus, the fatty animals were glucose intolerant and hyperglycemic. The GIP infusion did not result in a decrease in circulating glucose levels in the fat animals, as no difference (P > 0.05) was observed between GIP and saline infusions at any time after IP glucose (Fig. 1C). Furthermore, the integrated glucose response for the fatty animals that received GIP was not different from the integrated response of saline-infused animals (Fig. 1D). Thus, GIP yielded no improvement in glucose tolerance in the diabetic fatty animals.

Plasma was collected at –5, 10, 20, 30, and 60 min after IP glucose and assayed for insulin and GIP content. There was no difference observed in basal circulating GIP levels between the fat (16.7 ± 1.8 pmol/l) and lean (15.5 ± 1.6 pmol/l) animals. GIP yielded a significant increase in circulating insulin levels in the lean animals with a peak of 400 pmol/l at 20 min, which was before the glucose peak, in the same test group (Figs. 1A and2A). Additionally, the integrated insulin response was significantly greater in the lean animals that received GIP (Fig. 2B). Because of the insulin-resistant state of the fatty animals, the insulin levels were much higher at all times during the infusion protocol in this group (Fig. 2C). Furthermore, there was neither a significant increase in insulin secretion elicited by GIP infusion (Fig. 2C) nor an increase in the integrated insulin response in these animals (Fig. 2D).

Pancreatic perfusions were carried out to determine if the release of insulin was a direct response to GIP in the intraperitoneal glucose tolerance test. As indicated in Figs. 3A and C, both 10 pmol/l GIP and 50 pmol/l GLP-1 in the presence of 8.8 mmol/l glucose evoked approximately equal peak 2.8-fold (6.1 AUC) and 3.4-fold (5.3 AUC) increases, respectively, in insulin secretion from the lean pancreas compared with control conditions. However, this augmentation in insulin secretion was not observed in the fatty Zucker pancreas when stimulated with GIP (Figs. 3B and D) where insulin secretion decreased from ∼8.5 nmol/l to around 3.7 nmol/l during the high glucose and GIP infusion period. In contrast, GLP-1 caused a significant insulin secretion from the pancreas of the fat animals that reached a plateau at 10 times the control perfusion level of secretion (Figs. 3B and D).

Pancreatic islet perifusion was carried out to determine whether the defect in GIP stimulation of insulin secretion was confined to the islet. As seen in Fig. 4A, 10 nmol/l GIP in the presence of 16 mmol/l glucose was able to stimulate insulin secretion from the lean Zucker islet, with a peak level 10 times the basal level of 62 pmol/l. High glucose (16 mmol/l) alone only produced a fourfold increase in insulin release from the islets. Figure 4B illustrates the effects of GIP on the islets from the fatty Zucker rat. As illustrated, 10 nmol/l GIP had little effect on insulin release from these islets in the presence of 16 mmol/l glucose. The peak GIP-stimulated insulin release from the fat islets in the perifusion system was 4.3 ± 1.5 × basal, whereas glucose alone produced about a 3.2 ± 1.5-fold increase from a basal level of 95 pmol/l; this response was similar to the response to glucose alone seen in the lean rat islets (Figs. 4A and B).

Islet cAMP studies were carried out to elucidate where the defect in the GIP signaling pathway in the fatty animals occurred. Figure 5 illustrates the effect of GIP and forskolin on cAMP production in islets from fat and lean rats. GIP (10 nmol/l) produced a significant (2.2-fold basal) response in the lean control islets; however, there was no observable cAMP response to GIP in the islets from obese animals. The basal values of cAMP production did not differ significantly between the two phenotypes. Additionally, the forskolin responses of the fat and lean animals did not differ significantly, indicating that the cAMP production machinery was not defective in the VDF rats.

The cAMP data suggested that a decrease of GIP receptor expression could be responsible for the decreased effectiveness of GIP in the fatty Zucker rat. This hypothesis was tested by carrying out real-time RT-PCR on RNA isolated from islets of lean and fat animals. We observed a significant (75 ± 5%) decrease in GIP receptor mRNA in the islets from the fatty Zucker rats, as seen in Fig. 6A. Additionally, the reduction of GIP receptor mRNA was obtained when measured with RT-competitive PCR—an alternate means of measuring RNA expression (data not shown).

The Western blot in Fig. 6B is representative of those obtained when comparing fat and lean islet GIP receptor protein content. The posttranslationally modified GIP receptor appears to run at ∼65 kDa, which is in agreement with previous work (35). Figure 6B illustrates a marked decrease in GIP receptor protein level in islets from VDF rats. This decrease is in accordance with that seen with mRNA levels as well as insulin release and cAMP stimulation of islets with GIP.

Human type 2 diabetic subjects have been characterized by a decreased incretin response (21,26). This response has been attributed to dysfunction in the GIP portion of the enteroinsular axis because GLP-1 continues to have both insulinotropic and blood glucose–lowering actions in these individuals (22,26,36). The current study examined the expression and function of the GIP receptor on the pancreatic β-cell of the VDF rat—an animal model for type 2 diabetes. It was shown for the first time that there is a decreased level of GIP receptor mRNA and protein in the pancreatic islets of animals exhibiting characteristics of type 2 diabetes: hyperinsulinemia and insulin resistance. This decrease in GIP receptor resulted in a decreased insulin response to physiological levels of GIP.

The VDF rats used in this study were both glucose intolerant (Fig. 1) and hyperinsulinemic (Figs. 2 and 3). Additionally, these animals did not exhibit a second-phase insulin response to glucose (Fig. 3) and, therefore, have developed a syndrome similar to that observed in human type 2 diabetes. The pancreas perfusions (Fig. 3) demonstrated that there is a blunted first-phase insulin secretion in the fat animals: 125% increase compared with a 600% increase in lean animals in response to the introduction of 8.8 mmol/l glucose. Furthermore, in the normal lean animal, 10 pmol/l GIP stimulated a characteristic biphasic insulin response, whereas this hormone had no effect on second-phase insulin secretion in the fat animal (Figs. 3A and B). Interestingly, 10 pmol/l GIP and 50 pmol/l GLP-1 had similar insulinotropic effects in the perfused pancreas from the lean animals (Fig. 3C); however, only GLP-1 was able to potentiate glucose-induced insulin secretion from the VDF pancreas (Fig. 3D). The perifusion data demonstrate that GIP was able to significantly increase first-phase insulin secretion in islets from lean animals, but continued exposure to 10 nmol/l GIP may have resulted in desensitization of β-cell surface GIP receptors, and no further significant effect of GIP on insulin secretion was observed (29).

Recently, there has been considerable interest in using dipeptidylpeptidase IV (DPIV) inhibitors in type 2 diabetes therapy (37,38). Circulating DPIV inactivates both GLP-1 and GIP by cleaving the NH₂-terminal dipeptide from the parent incretin polypeptide, rendering both peptides biologically inactive and thereby decreasing the circulating half-life of the incretins (39). Currently, there is controversy as to whether GIP levels are elevated, normal, or lowered in type 2 diabetes and animal models for the disease (24,25,26,27,40). One possible explanation for the ineffectiveness of GIP in the VDF rat is an increase in DPIV in these animals, which could inactivate GIP prior to its actions on the β-cell. Because it was impossible to measure differences in intact versus NH₂-terminally truncated GIP with our radioimmunoassay, it was not possible to determine the role of DPIV in our findings, nor was it possible to determine the concentration of bioactive GIP in these animals. However, it has been demonstrated by Pederson et al. (41) that the levels of circulating DPIV in VDF rats and their lean littermates are similar; therefore, this explanation for GIP ineffectiveness can be ruled out.

Exposure of the rat islet to GIP leads to desensitization of the islet to GIP (29). Currently, it is not clear whether this is due to a reversible phosphorylation of the GIP receptor, a decrease in GIP receptor mRNA expression, or both. However, it was demonstrated that there was a rapid homologous desensitization of GIP-stimulated insulin secretion in mouse β-cells (βTC3) that occurred at the receptor level as well as further downstream in the signaling cascade. In contrast, in the current study, the cAMP response to GIP in the islets of the VDF animals suggests there is a defect in the GIP receptor–adenylyl cyclase segment of the GIP receptor signaling pathway. Furthermore, we were unable to measure elevated ambient GIP levels in the VDF rat, indicating that GIP is not causing a loss of functional cell surface receptors, either by desensitization or downregulation, in these animals at this level of glucose intolerance. Expression of other G-protein–coupled receptors (such as the glucagon receptor) in the superfamily are regulated by glucose as well as the adenylyl cyclase activator forskolin (42). Thus, it is likely that GIP receptor downregulation and dysfunction are a result of inappropriate stimulation of the β-cell by abnormal levels of metabolites and hormones in this animal model.

The current study demonstrates for the first time that GIP receptor expression is downregulated at the mRNA and protein levels in the pancreatic β-cell of the diabetic Zucker rat (Figs. 6A and B). These observations suggest that there is a decrease in receptor expression on the pancreatic β-cell and that this decrease in expression leads to the loss of islet sensitivity to GIP. Furthermore, we have recently demonstrated in clonal β-cells that changes in mRNA expression also produce changes in cell surface GIP receptor expression using radioligand saturation binding curves (F.C.L., unpublished data). Therefore, we believe that there is a decrease in cell surface GIP receptor concomitant to the decrease in total GIP receptor mRNA and protein in these islets.

Recently, Béguin et al. (43) demonstrated that stimulation of the β-cell by GIP caused phosphorylation of the Kir6.2 (K_ATP) channel via protein kinase A on serine 372. Phosphorylation of this serine residue led to an increased open probability of the channel. This article first demonstrates that GIP stimulation of the β-cell leads to protein phosphorylation. However, the physiological basis for this phosphorylation event is still unclear since Béguin et al. believe that Kir6.2 is maximally phosphorylated in the basal state. It is tempting to speculate that decreased levels of GIP receptor on the β-cell surface result in a decrease in the phosphorylation state of the K_ATP channel and decreased open probability. This in turn would lead to membrane depolarization and insulin secretion. If this receptor deficit was great enough, there could be uncoupling of glucose-stimulated insulin secretion and concomitant β-cell decompensation, as observed in fatty rats.

In conclusion, glucose tolerance and insulin responses were studied after GIP infusion in the Vancouver Zucker diabetic fatty rat. In these animals, GIP did not potentiate glucose-induced insulin secretion, either in vivo or from the perfused rat pancreas and isolated perifused rat islets, whereas the GLP-1 response remained intact. Moreover, GIP failed to stimulate cAMP production in isolated fa/fa islet static incubations. Finally, GIP receptor mRNA and protein levels were shown to be downregulated in the islets of these animals, and this was hypothesized to be the basis for their insensitivity to GIP. This study suggests that the decreased GIP response observed in this animal model and in type 2 diabetes may be a result of a downregulation of the pancreatic GIP receptor.

This work was supported by the Medical Research Council (MRC) of Canada (Grant 590007), the Canadian Diabetes Association, the Canada Foundation for Innovation, and the British Columbia Health Research Foundation. F.C.L. is funded by the MRC and the Killam Trusts.

We would like to thank Cuilan Nian for technical assistance on this project.