Physiologic and Pharmacologic Modulation of Glucose-Dependent Insulinotropic Polypeptide (GIP) Receptor Expression in β-Cells by Peroxisome Proliferator–Activated Receptor (PPAR)-γ Signaling: Possible Mechanism for the GIP Resistance in Type 2 Diabetes

All protocols were in accordance with the principles of laboratory animal care and were approved by the University of Vermont Institutional Animal Care and Use Committee.

The 5-week-old male Sprague-Dawley rats, and Zucker fatty (ZF, fa/fa) or lean controls (ZL, fa/+ or +/+), underwent 60% Px or sham Px as previously described (11,13).

Mice with PPARγ deficiency restricted to pancreatic epithelium were generated by crossing Pdx-1 Cre mice and mice with two floxed PPARγ alleles as previously detailed (12,14). Controls were littermate Cre negative PPARγ floxed mice.

INS-1 (832/13) cells were cultured as previously described (11,12).

See the online appendix, available at http://diabetes.diabetesjournals.org/cgi/content/full/db09-1655/DC1.

Data are presented as mean ± SEM. Each data point from the animal studies was obtained from an individual rat or mouse. Statistical significance was determined by the unpaired Student's t test or two-way ANOVA.

We previously identified a functional PPRE in the mouse pdx-1 promoter (12). Gene Bank blast for this nucleotide sequence found a 100% identical sequence in the rat GIP-R promoter at positions −871 to −883 from the transcription start site (Gene Bank: AF050667) (supplemental Fig. 1A, available in the online appendix). MatInspector software (15) uncovered analogous sequences in the mouse and human GIP-R promoters (supplemental Fig. 1B).

Studies were performed in rat-derived INS-1 cells to confirm functionality of this putative PPRE. PPARγ binding was determined with the chromatin immunoprecipitation assay. Flanking primer pairs for a 213-bp PCR product that included the GIP-R PPRE (schema in supplemental Fig. 2A) generated the correct-sized PCR band with input DNA and PPARγ antibody–precipitated DNA, whereas only faint bands were observed with nonimmune serum (Fig. 1 A). Representative negative and positive controls are shown in supplemental Fig. 2.

PPARγ regulation of GIP-R expression was tested using four pooled siRNA duplexes against PPARγ in INS-1 cells that cause a 75% decrease in PPARγ protein (11). Cells were treated with troglitazone or vehicle for 72 h (Fig. 1 B). GIP-R mRNA band intensity was markedly lowered in the PPARγ siRNA cells (lanes 1 and 4). Troglitazone doubled it in control cells (lane 2), whereas the increase was eliminated when troglitazone and the PPARγ siRNA duplexes were used together (lane 3).

PPARγ regulation of GIP-R transcription was confirmed with a luciferase reporter gene assay that used a wild-type rat GIP-R promoter fragment and also those containing mutations in the DR1 and DR2 hexamers of the GIP-R PPRE (Mut-1 and Mut-2 fragments), subcloned into the pTAL luciferase reporter vector. The DR1 and DR2 mutations both lowered basal luciferase activity (Mut-1 63 ± 9% of wild-type GIP-R, P < 0.015; Mut-2 38 ± 8% of wild-type GIP-R, P < 0.002) (Fig. 1,C). Also, 24-h incubation with troglitazone (Fig. 1 D) stimulated expression of the wild-type GIP-R construct (162 ± 8% of vehicle-treated wild-type cells, P < 0.002), whereas this effect was absent with both mutations.

In vivo testing of PPARγ regulation of β-cell GIP-R expression was carried out using PANC PPARγ^−/− mice that have a pancreas-specific deletion of PPARγ from a Cre/loxP recombinase system with Cre driven by the pdx-1 promoter (14). We previously reported that 8-week-old male PANC PPARγ^−/− mice are modestly hyperglycemic, with a normal β-cell mass and normal-appearing pancreas histology and islet cytoarchitecture (12). The current studies confirmed the expected defect in GIP-R expression (Fig. 2,A, mRNA; Fig. 2,B, protein 29 ± 5% of control, P < 0.0001) in isolated islets from these mice. Immunofluorescence studies of pancreas sections showed markedly lowered GIP-R staining in islet β-cells and non–β-cells of PANC PPARγ^−/− mice versus control mice (Fig. 2 C).

Analogous studies were performed in normoglycemic Sprague-Dawley rats 14 days after 60% Px when islet nuclear PPARγ expression is 2.5-fold increased (11). Comparable increases in GIP-R mRNA (Fig. 3,A) and protein (Fig. 3,B, 1.9 ± 0.2-fold of sham, P < 0.02) were found in Px versus sham rat islets. Also, GIP-R staining intensity was increased in islet β-cells and non–β-cells versus the sham rats (Fig. 3 C).

Thiazolidinedione (TZD) stimulation of GIP-R expression was tested by incubating normal mouse islets for 72 h with troglitazone or vehicle. Troglitazone caused a near-tripling of the GIP-R protein level (supplemental Figs. 3A and B, 2.7 ± 0.4-fold of DMSO islets, P < 0.02). Also the insulin response to 16.7 mmol/l glucose/10 nmol/l GIP peptide was nearly doubled (P < 0.05), whereas insulin responses to 2.8 mmol/l glucose with or without GIP peptide, and to high glucose alone, were unchanged (supplemental Fig. 3C).

We speculated that impaired PPARγ expression caused the β-cell GIP-R depletion in animal models of type 2 diabetes (6,7). Accordingly, we performed PPARγ and GIP-R Western blots on islet extracts from 3-week postsurgery 60% Px and sham-operated ZF and ZL rats, based on our report that Px ZF rats are hyperglycemic (∼15 mmol/l) with the same degree of obesity and serum levels of nonesterified fatty acids and GIP as normoglycemic sham ZF rats (13). Figure 4 shows that islet PPARγ and GIP-R protein levels are doubled in the sham ZF versus sham ZL rats (1.9 ± 0.1-fold, P < 0.001; 2.2 ± 0.3-fold, P < 0.01, respectively). In contrast, both are lowered by 60% in Px (hyperglycemic) ZF islets versus the sham (normoglycemic) ZF islets (0.7 ± 0.1-fold and 0.8 ± 0.1-fold of sham ZL islets, respectively).

We used in vitro and in vivo techniques to show for the first time that GIP-R transcription in β-cells is regulated by PPARγ. Also, that pharmacologic activation of PPARγ by TZDs induces the same effect resulting in greater GIP potentiation of glucose-induced insulin secretion. These results add to our prior studies of PPARγ signaling in β-cells that showed 40% of adult mouse islet pdx-1 expression is PPARγ regulated (11,12) by our now finding a second gene with the identical PPRE and functional characteristics.

There are several noteworthy implications to these results. First, Pdx-1 is considered the most important transcription factor in β-cells related to its essential role for normal β-cell function, viability, and compensation capacity (16). Also, studies of incretin receptor knockout mice and in vivo usage of GIP-R peptide agonists have demonstrated a necessary role for GIP in normal mealtime glucose tolerance and insulin secretion (17,18). Clinical trials have found a particularly high success rate of TZDs in pre-diabetes and early type 2 diabetes, with the mechanism believed to be the peripheral insulin sensitization lowering the drive for insulin secretion (so-called “β-cell rest”) (19). However, given the importance of the identified PPARγ-regulated genes in β-cells, our results raise the possibility that direct PPARγ-mediated effects on β-cells also account for some of the clinical benefits. Indeed, others have shown a direct effect of TZDs to restore Pdx-1 levels and reduce endoplasmic stress in islets from diabetic rats (20).

Second, there is the possibility we have uncovered an unknown feature of incretin physiology, i.e., variable GIP-R expression, related to our finding increased islet GIP-R expression in two rat models of β-cell compensation from unrelated causes: reduced β-cell mass in 60% Px Sprague-Dawley rats (21) and mutated leptin receptors resulting in obesity, insulin resistance, and hyperlipidemia in ZF rats (13). This shared observation is particularly interesting, since there is no precedence for GIP-R hyperexpression except in the adrenocortical tumors of some individuals with Cushing's syndrome (22). Also, our finding that the GIP-R hyperexpression in both models was paralleled by increased β-cell expression of PPARγ (11; Fig. 4) is consistent with the main conclusions of this study regarding PPARγ regulation of β-cell GIP-R expression. On the other hand, an in vitro study reported PPARα regulates β-cell GIP-R expression and speculated this effect occurs in vivo related to the β-cell fatty acid load (23). This mechanism seems unlikely in the 60% Px Sprague-Dawley rats, since serum triglyceride and free fatty acid levels are unchanged post-Px in lean rats (13,21), plus islet PPARα expression is reduced (13). In contrast, ZF rats are markedly hyperlipidemic (13), making this a potential mechanism. Also the lack of a post-Px increase in PPARγ and GIP-R expression in ZL rats (Fig. 4) may reflect the different times of study—3 weeks post-Px in ZL rats versus 2 weeks post-Px in Sprague-Dawley rats.

The third implication of our results is our speculation that aberrant PPARγ signaling is a new mechanism for the lowered β-cell GIP-R expression and GIP resistance in animal and human type 2 diabetes (6,,–9). Hyperglycemia causes this effect in animals (7,10), but no mechanistic details are known except for a study that reported accelerated GIP-R degradation in islets cultured at high glucose conditions (24). Figure 4 supports our proposal by showing parallel reductions in islet expression of PPARγ and GIP-R in the hyperglycemic ZF rats. However, Px ZF rats also have a reduction in islet PPARα mRNA expression (13), so the exact mechanism is unknown. A related issue is the lowered β-cell expression for the other incretin hormone receptor (GLP-1) in animals and humans with type 2 diabetes (7,8). However, the molecular mechanism is likely different from the GIP-R depletion based on the study by Xu et al. (7) that performed 96-h glucose infusions in conscious rats and found islet GLP-1R mRNA levels fell 50%, whereas GIP-R mRNA levels modestly increased, plus they noted the same divergence in 2-day high glucose cultured islets. Also, it is likely that the lowered islet incretin receptor expression in type 2 diabetes is multifactorial, with recent interest in disrupted TCF7L2 signaling (8). Thus, our results suggest that β-cell PPARγ regulation of GIP-R expression is a newly identified feature of incretin (patho)-physiology and TZD therapeutics. Also the lowered islet PPARγ expression in Px ZF rats (glycemia 15 mmol/l) may provide a new understanding why TZD therapy is so powerful in pre-diabetes and early type 2 diabetes as opposed to the more modest efficacy with advanced type 2 diabetes (19,25).

In summary, GIP-R is a newly described PPARγ-regulated gene in β-cells. Expression of both genes is increased in rat models with β-cell adaptation to diverse stimuli. Also, TZDs augment GIP potentiation of glucose-induced insulin secretion through this mechanism. In contrast, this expression system is downregulated in diabetic rats. As such, our studies have uncovered a new mechanism for regulatory control of the β-cell GIP-R expression and may have added a new understanding to incretin (patho)-physiology and TZD therapy.

This work was supported by National Institutes of Health grants DK56818 (J.L.L.) and DK68329 (T.L.J.), the Juvenile Diabetes Research Foundation (M.P.), and the Iacocca Foundation (M.P.).

No potential conflicts of interest relevant to this article were reported.

The authors would like to thank Dr. Richard Mortenson and Dr. Sheng Zhong Duan (University of Michigan) for the PANC PPARγ^−/− mouse model, Dr. Timothy Kieffer (University of British Columbia) for the GIP-R antiserum, and Dr. Michael Wolfe (Boston University) for GIP-R promoter constructs.