Peroxisome Proliferator-Activated Receptor γ Activation Restores Islet Function in Diabetic Mice through Reduction of Endoplasmic Reticulum Stress and Maintenance of Euchromatin Structure

Mouse models.Male db/db mice and lean C57BLKS/J mice 8 weeks of age were obtained from Jackson Laboratories and maintained under protocols approved by the University of Virginia and Indiana University Animal Care and Use Committees. These mice therefore exhibited phenotypic features resembling those of progressive type 2 diabetes in humans. On four separate occasions, 12 db/db mice per group were treated either with pioglitazone, which was administered daily by oral gavage at a dose of 20 mg/kg of body weight in 400 μl of phosphate-buffered saline, or with daily gavage of a vehicle only (total, 48 mice per group). Treatment was initiated when the mice were 8 weeks of age, after the onset of hyperglycemia, and continued for 6 weeks.

For the HFD experiments, male C57BLKS/J mice were fed either a regular chow diet (17% of calories from fat) or a Western HFD containing 42% of calories from fat (Harlan Teklad) for 7 weeks. Mice in the HFD group were gavaged daily with pioglitazone at 20 mg/kg as described above or with vehicle only. All mice were kept in a standard light-dark cycle and had access to a regular chow diet and water ad libitum.

For Pdx1 knockout studies of a genetically altered mouse strain in which Pdx1 expression can be inducibly eliminated by administration of doxycycline (“Tet-off” mice), Pdx1^tTA/+ and Tg^Pdx1 mice were obtained from Jackson Laboratories and bred to homozygosity to produce Pdx1^tTA/tTA; Tg^Pdx1 offspring as described previously (20). When the mice were 8 weeks of age, administration of doxycycline was begun as a single intraperitoneal dose of 100 mg/kg and was then maintained in the drinking water at 0.5 mg/ml for 7 days to inducibly delete the Pdx1 gene as described previously (20).

Islet isolation and cell culture.Pancreatic islets were isolated from mice by collagenase digestion as described previously (15, 59), hand picked, and cultured in phenol-free low-glucose Dulbecco's modified Eagle's medium overnight prior to use. INS-1 (832/13) cells were cultured in RPMI 1640 with 11.1 mM glucose supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 10 mM HEPES, 2 mM l-glutamine, 1 mmol sodium pyruvate, and 50 μmol/ml β-mercaptoethanol. Pioglitazone was used at a concentration of 10 μM for in vitro incubations in a final concentration of 0.1% dimethyl sulfoxide (DMSO). Thapsigargan was dissolved in DMSO and added to INS-1 cell cultures at a final concentration of 1 μM in 0.1% DMSO.

Metabolic testing.Intraperitoneal glucose tolerance tests were performed after an overnight fast and intraperitoneal injection of glucose at 1 to 2 g/kg. Blood was sampled from the tail vein at 0, 10, 20, 30, 60, 90, 120, and 180 min, and glucose was measured using a OneTouch Ultra glucometer (Lifescan). Insulin tolerance testing was performed on random fed mice following an intraperitoneal injection of 0.75 to 3 U of insulin/kg. Blood was sampled from the tail vein at 0, 15, 30, 45, and 60 min, and glucose was measured in the manner described above. Groups were compared using two-way analysis of variance (ANOVA).

When they were 8 weeks of age and at the time of sacrifice, animals were weighed and blood glucose measurements for nonfasting mice were obtained from the tail vein. Blood was obtained at sacrifice by cardiac puncture in selected groups for measurement of total cholesterol, triglycerides, free fatty acids, and serum insulin. Samples were centrifuged at room temperature, and serum was separated and stored at −20°C until assayed. Total cholesterol and triglycerides were analyzed using a VetTest chemistry analyzer (Idexx Laboratories). Serum insulin was measured using a commercial radioimmunoassay (Millipore). Free fatty acids were assayed using an HR series NEFA-HR(2) enzymatic assay kit (Wako Diagnostics). Groups were compared by one-way ANOVA followed by the Tukey-Kramer posttest.

Intracellular Ca²⁺ measurements.Intracellular Ca²⁺ ([Ca²⁺]_i) was measured using the ratiometric Ca²⁺ indicator fura-2 -acetoxymethylester (fura-2 AM) and a modification of previously published methods (24). Islets were maintained in 11 mM glucose at all steps prior to studies of oscillations to prevent transient states in oscillations due to shifts in glucose concentrations prior to or during the experiment (40, 62). Islets were loaded with 3 μM fura-2 AM for 30 to 40 min, washed, transferred to a small-volume chamber (Warner Instruments) mounted on the stage of an Olympus BX51WI fluorescence microscope (Olympus), and perifused with 11 mM glucose Krebs-Ringer HEPES-buffered solution with a peristaltic pump (Gilson) at ∼35°C by the use of an in-line heater (Warner Instruments). Islets were incubated within the chamber under these conditions for 10 to 15 min before the experiment was begun. Excitation light from a xenon burner was supplied to the preparation via a light pipe and filter wheel (Sutter Instrument Company). Images were taken sequentially under conditions of 340-nm and then 380-nm excitation using a Hammamatsu ORCA-ER camera (Hammamatsu) to produce data representing each [Ca²⁺]_i ratio from emitted light at 510 nm.

Paired images were recorded every 5 s for 15 min. After determinations of [Ca²⁺]_i in 11 mM glucose for 15 min, islets were incubated in 3 mM glucose for 15 min and then results were recorded to determine the islet response to 28 mM glucose stimulation. The glucose-stimulated [Ca²⁺]_i response (GSCa) was defined as the difference between ratio measurements (340/380 nm fluorescence) in 28 mM versus 3 mM glucose. Data were analyzed with IP Lab software, version 4.0 (Scanalytics). The periods and amplitudes of [Ca²⁺]_i oscillations were determined using the CLUSTER8 pulse detection algorithm (at settings of peak size = 2, nadir size = 2, minimum = 0, and t score = 2) or by direct visual inspection (41). Groups were compared using one-way ANOVA followed by the Tukey-Kramer posttest.

Measurement of glucose-stimulated insulin secretion and quantitation of total pancreatic insulin content.Approximately 50 islets per group were incubated in a 12-well dish for 1 h at 37°C in Krebs-Ringer HEPES-buffered solution containing 3 mM glucose. Islets were then transferred to Krebs-Ringer HEPES at 3 mM glucose for an additional hour, after which the supernatant was collected for insulin measurement using a two-site immunospecific enzyme-linked immunosorbent assay (ELISA) (Alpco Diagnostics or Crystal Chem). The same islets were subsequently transferred to Krebs-Ringer HEPES-buffered solution containing 28 mM glucose, and insulin release into the medium was measured by ELISA after 1 h of incubation. Total pancreatic insulin content was measured by acid extraction as previously described (1).

Quantitative ChIP assay.Approximately 100 islets were fixed in 1% formaldehyde for 15 min, sonicated to shear DNA to obtain fragments in the range of 800 to 2,000 bp, and subjected to chromatin immunoprecipitation (ChIP) analysis as detailed previously (4, 5) using antibodies against acetylated H3, acetylated H4, and H3-dimethyl-Lys4 (Millipore, Billerica, MA). ChIP assays were performed for at least three independent islet isolation experiments. For each isolation, samples were quantitated in triplicate by SYBR green I-based real-time PCR using forward and reverse primer sequences for the mouse Ins1/2 and Glut2 gene promoters and cycling parameters described previously (5). The forward and reverse primers used to amplify the Ins1/2 promoter (bp −126 to −296 relative to the transcriptional start site) were 5′-TCAGCCAAAGATGAAGAAGGTCTC-3′] and 5′-TCCAAACACTTGCCTGGTGC-3′]. The forward and reverse primers used to amplify the Glut2 promoter (bp −523 to −738 relative to the transcriptional start site) were 5′-ATCTGGCTCCGCACTCTCATCTTG-3′ and 5′-CCCTGTGACTTTTCTGTGTCTTAGG-3′.

Real-time RT-PCR.Total RNA (5 μg) from islets was reverse transcribed at 37°C for 1 h using 15 μg of random hexamers, 0.5 mM deoxynucleoside triphosphates, 5× first-strand buffer, 0.01 mM dithiothreitol, and 200 units of Moloney murine leukemia virus reverse transcriptase (Invitrogen) in a final reaction volume of 20 μl. Real-time reverse transcription-PCR (RT-PCR) was performed for the Pdx1, NeuroD1, Nkx6.1, Pax6, Gck, Kcnj11, Glut2, Ins1/2, IAPP, Irs1, Irs2, LPL, and β-actin genes by the use of TaqMan primer/probe combinations (Applied Biosystems) and cycling parameters described previously (23). The amplified product for each RT-PCR was subcloned into pCR2.1 T/A cloning vector (Invitrogen) and sequenced to confirm the identity of the amplified product. Primers used to amplify components of the ER stress pathway, and sarcoplasmic/ER Ca²⁺ ATPase (SERCA) genes were used as previously described (28, 31). The cycle threshold (C_T) methodology (5) was used to calculate relative quantities of mRNA products from each sample; all samples were corrected for total input RNA by normalizing C_T values to the C_T value of the β-actin message.

Immunohistochemistry and immunocytochemistry.Pancreata from at least three mice per treatment group were fixed by cardiac perfusion with 4% paraformaldehyde, paraffin embedded, and sectioned at 5-μm intervals. INS-1 cells were fixed on coverslips with Z-fix. Immunohistochemical analysis of insulin (rabbit anti-human insulin [Santa Cruz] [1:500]), Pdx1 (rabbit anti-mouse Pdx1 [Millipore] [1:2,000]), and Set7/9 (rabbit anti-human Set7/9 [Millipore] [1:800]) was performed as previously described (11). Immunofluorescence experiments were performed for insulin and glucagon (rabbit anti-human insulin and glucagon; Santa Cruz) (1:500), Pdx1 (mouse anti-human Pdx1; R&D Systems) (1:2,000), Set7/9 (mouse anti-human Set7/9; Abcam) (1:50), and nuclei (Hoechst 33342; Molecular Probes). Secondary antibodies were goat anti-rabbit immunoglobulin G conjugated to Alexa Fluor 555 (1:200 dilution) and donkey anti-mouse antibody conjugated to Alexa Fluor 488 (1:50 dilution) (Molecular Probes). Images were acquired using a Zeiss (Thornwood, NY) LSM 510 confocal microscope.

For an assessment of β-cell proliferation, a total of 60 random fields from two pancreata per group were stained for Ki67, insulin, and diamidino-2-phenylindole (DAPI) and imaged with a 20× lens objective. The number of cells costained with Ki67 and insulin was quantified and divided by the total number of insulin-positive cells per islet to obtain a proliferative index value (32). For assessment of relative cytoplasmic-to-nuclear distribution of Set7/9, INS-1 cells stained for Set7/9 and DAPI (to visualize nuclei) were visualized using a Zeiss Z1 inverted microscope equipped with an Apotome optical sectioning unit, and cytoplasmic and nuclear intensities were quantitated using Zeiss Axiovision software.

Immunoblot analysis.Aliquots of whole-islet or INS1 cell extract (5 μg) were prepared as described previously (23) and resolved by electrophoresis on a 4-to-20% sodium dodecyl sulfate-polyacrylamide gel followed by immunoblot analysis using polyclonal rabbit anti-Pdx1, monoclonal mouse anti-Set7/9, anti-glyceraldehyde-3-phosphate dehydrogenase (anti-GAPDH), CHOP, and cleaved caspase 3 antibodies. Immunoblots were incubated with near-infrared-fluorophore-labeled secondary antibodies before fluorometric scanning was performed with an Odyssey imaging system (Licor).

Effect of pioglitazone on metabolic homeostasis in db/db mice.To study the effect of PPAR-γ activation on islet function, we began with a mouse model in which db/db mice were treated orally with pioglitazone. db/db mice on the C57BLKS/J background harbor a mutation of the leptin receptor and exhibit progressive obesity, insulin resistance, and islet dysfunction with age (22). At the start of the study, 8-week-old db/db mice weighed approximately 33.5 g (58% more than age-matched lean C57BLKS/J counterparts), consistent with the obese phenotype of these mice. The mice already exhibited a diabetic phenotype, as suggested by their average nonfasting blood glucose level of 268 mg/dl. At the completion of the 6-week study, vehicle-treated db/db mice exhibited an approximately 30% weight gain (to about 44 g), whereas pioglitazone-treated mice exhibited a significantly greater 53% weight gain (to about 51 g) (Fig. 1A). Notwithstanding the greater weight gain, pioglitazone-treated mice displayed 10% lower total cholesterol, 30% lower triglyceride, and 40% lower free fatty acid levels compared to vehicle-treated mice. These values, with the exception of total cholesterol, were similar to those seen with their lean C57BLKS/J counterparts (Fig. 1B to D).

• Open in new tab
• Download powerpoint

FIG. 1.

Pioglitazone treatment results in greater weight gain but improved serum lipids in db/db mice. Male C57BLKS/J-db/db mice were treated with either vehicle (db/db) or pioglitazone (Pio-db/db) for 6 weeks and compared to age- and sex-matched lean C57BLKS/J mice. Following the treatment period, mice were evaluated for body weight (A), serum total cholesterol (B), serum triglycerides (C), and serum free fatty acids (D). Data represent the means ± standard errors of the results obtained with 12 mice per group. *, significantly (P < 0.05) different compared to C57BLKS/J mouse results; #, significantly (P < 0.05) different compared to db/db mouse results.

We next assessed glucose homeostasis in control and treated db/db animals. As shown in Fig. 2A, pioglitazone-treated mice displayed significantly lower random blood glucose values (194 ± 20 mg/dl) compared to vehicle-treated mice (318 ± 20 mg/dl). Consistent with this improvement in nonfasting glucose levels, intraperitoneal glucose tolerance testing demonstrated that pioglitazone-treated mice exhibited enhanced glucose clearance compared to the controls, although the treatment did not completely normalize the results compared to those seen with the lean C57BLKS/J counterparts (Fig. 2B). Remarkably, the average nonfasting insulin levels for pioglitazone-treated mice were 90% higher than those for control animals (38 versus 20 ng/ml), suggesting that the improved nonfasting glucose levels and glucose tolerance in pioglitazone-treated animals could be at least partly explained by an enhanced islet insulin secretory capacity (Fig. 2C). This conclusion was further supported by the results of insulin tolerance testing, which showed that treated and untreated mice were equally insulin resistant after 6 weeks of pioglitazone treatment (Fig. 2D).

• Open in new tab
• Download powerpoint

FIG. 2.

Pioglitazone treatment improves glycemic control and insulin levels in db/db mice. Male db/db mice were treated with either vehicle (db/db) or pioglitazone (Pio-db/db) for 6 weeks and compared to age- and sex-matched lean C57BLKS/J mice. (A) Results of random blood glucose tests at the end of the 6-week treatment period; (B) results of intraperitoneal glucose tolerance tests at the end of the 6-week treatment period. Pio-db/db animals exhibited significantly improved glucose tolerance compared to db/db animals (P < 0.001 for the comparison by two-way ANOVA). (C) Random serum insulin levels at the end of the 6-week treatment period; (D) results of insulin tolerance tests. No differences in insulin tolerance were observed between the db/db and Pio-db/db groups, whereas both groups had significantly impaired insulin tolerance compared to the C57BLKS/J mouse group (P < 0.001 for the comparison by two-way ANOVA). Data represent the means ± standard errors of the results obtained with at least 12 animals per group. *, significantly (P < 0.05) different compared to C57BLKS/J mice; #, significantly (P < 0.05) different compared to db/db mice.

Pioglitazone enhances islet function.To determine whether pioglitazone treatment enhanced islet function, we first performed immunohistochemistry using pancreatic sections from control and pioglitazone-treated mice. As shown in Fig. 3A to C, it is apparent from hematoxylin and eosin staining that the islets of control- and pioglitazone-treated db/db mice were substantially larger than those of lean C57BLKS/J mice, consistent with the adaptive β-cell hyperplasia of insulin resistance (42). Given the increased islet size in treated mice compared to untreated controls, we next assessed proliferation by staining pancreatic sections with Ki67, insulin, and DAPI. There were no differences in Ki67 staining results observed between islets from treated and untreated animals (data not shown). Moreover, an extensive analysis of cell cycle genes by PCR array profiling revealed no notable differences between control and pioglitazone-treated islets in the expression of genes controlling the G₁-to-S transition (data not shown). These data suggested to us that the larger islet size in pioglitazone-treated animals was not secondary to enhanced islet replication but instead may have resulted from protection from islet cell dropout.

• Open in new tab
• Download powerpoint

FIG. 3.

Islet architecture and insulin staining. Pancreata from male db/db mice, treated with either vehicle (db/db) or pioglitazone (Pio-db/db) for 6 weeks, were fixed and stained and compared to pancreata from age- and sex-matched lean C57BLKS/J mice. (A, B, and C) Hematoxylin and eosin staining of pancreatic sections from C57BLKS/J, db/db, and Pio-db/db mice; (D, E, and F) immunofluorescence staining of islets from C57BLKS/J, db/db, and Pio-db/db mice for insulin (red) and glucagon (green). Nuclei were counterstained with Hoechst dye (blue). The figure shows representative islets from three pancreata analyzed per group of mice.

Next, we examined islet insulin production and secretory capacity. Immunofluorescence staining demonstrated that the islets of control db/db mice exhibited noticeably lower staining intensity for insulin than those of lean C57BLKS/J mice, suggestive of β-cell degranulation (Fig. 3D and E). Pioglitazone treatment, however, restored islet insulin-staining intensity (Fig. 3F). Of note, whereas glucagon-producing α cells occurred in a normal pattern at the islet periphery in lean mice, this architecture appeared disrupted in control-treated db/db mice, with α cells frequently occupying central portions of the islets; pioglitazone treatment did not appear to restore this normal architecture to islets (Fig. 3D to F). Total pancreatic insulin content was measured for three pancreata per group following acid extraction. Mean pancreatic insulin content was significantly higher in the C57BLKS/J group compared to the db/db group (11,335 ng ± 1,064 ng versus 4,854 ± 475 ng [P < 0.05]). Pioglitazone restored total pancreatic insulin content to the levels observed in the C57BLKS/J group. Mean pancreatic insulin content in pioglitazone-treated db/db mice was likewise significantly increased compared to that seen with untreated db/db mice (11,941 ng ± 1,506 ng versus 4,854 ± 475 ng [P < 0.05]). These data suggested that PPAR-γ activation enhances islet insulin production.

To directly assess islet function, we next isolated islets from control and pioglitazone-treated db/db mice and compared their function to that of islets isolated from lean C57BLKS/J mice. Figure 4 shows results from GSCa studies of isolated islets. GSCa is a measure of islet glucose sensitivity that captures the dynamics of the biphasic response, which is similar, but not identical, to those of glucose-stimulated insulin secretion (17). The GSCa, as measured by the change in the fura-2 AM fluorescence ratio after glucose stimulation, was reduced in control db/db islets such that the glucose stimulatory index was only 26% of that observed in islets from lean C57BLKS/J mice. Also, the initial reduction in [Ca²⁺]_i following exposure to high glucose levels was blunted or absent in vehicle-treated db/db mice. This phenomenon has been attributed to the activity of SERCAs (47). Pioglitazone treatment of db/db mice restored both the initial fall in [Ca²⁺]_i and relative islet GSCa responsiveness, with the GSCa index increasing to about 80% of that seen with lean C57BLKS/J islets (Fig. 4A to B). As shown in Fig. 4C, average islet insulin content was significantly lower in vehicle-treated db/db mice compared to that seen with the background strain. Insulin content was restored after pioglitazone treatment. Glucose-stimulated insulin secretion paralleled the GSCa studies for all three groups, with islets from untreated mice secreting significantly less insulin than islets from pioglitazone-treated db/db mice and islets from the background strain. Remarkably, pioglitazone treatment restored glucose-stimulated insulin secretion to the levels observed in the normoglycemic C57BLKS/J mice (Fig. 4D).

• Open in new tab
• Download powerpoint

FIG. 4.

Pioglitazone treatment improves islet function in db/db mice. Islets from male db/db mice, treated with either vehicle (db/db) or pioglitazone (Pio-db/db) for 6 weeks, were compared to islets from age-matched lean C57BLKS/J mice. (A) Results of GSCa studies of isolated islets. The panel shows the continuous fura-2 AM fluorescence ratio (340/380 nm) as glucose in the incubation chamber was increased from 3 mM to 28 mM. Data represent the means ± standard errors of the results obtained with at least 30 islets from 12 different animals per group. (B) Data from panel A were used to calculate a GSCa index, which represents the fura-2 AM fluorescence ratio at 28 mM glucose divided by the ratio at 3 mM glucose. (C) The insulin content of islets used for static release assays was measured by ELISA after acid extraction. (D) Islets were incubated in 3 and 28 mM glucose for 1 h, and insulin secretion into the supernatant was measured by ELISA. Data represent the means of the results of at least three independent experiments performed using 50 islets per group. *, significantly (P < 0.05) different compared to C57BLKS/J mice; #, significantly (P < 0.05) different compared to db/db mice.

We next examined endogenous Ca²⁺ oscillations in islets. Endogenous Ca²⁺ oscillations are closely correlated to oscillatory insulin secretion from islets, and their periodicity and amplitude have been suggested to reflect overall islet health (14, 44). Panels A to C of Fig. 5 show representative examples of endogenous Ca²⁺ oscillations under conditions of 11 mM glucose for each of the three groups studied. Islets from vehicle-treated db/db mice showed marked impairments in endogenous oscillations, with only about 38% of islets demonstrating oscillatory behavior compared to approximately 98% of islets from lean C57BLKS/J mice demonstrating endogenous oscillations. Pioglitazone treatment significantly improved the percentage of oscillating islets compared to the results seen with vehicle-treated controls, with 86% of islets exhibiting oscillations (Fig. 5D). Pioglitazone treatment also significantly increased the amplitude of oscillations compared to vehicle-treated control results (Fig. 5E and F), suggesting a possible increase in insulin release during each oscillatory peak.

• Open in new tab
• Download powerpoint

FIG. 5.

Pioglitazone treatment improves islet calcium oscillations. Islets from male db/db mice, treated with either vehicle (db/db) or pioglitazone (Pio-db/db) for 6 weeks, were harvested and compared to islets from age-matched lean C57BLKS/J mice. (A to C) Three representative calcium oscillations at an 11 mM glucose concentration from islets isolated from C57BLKS/J mice (A), db/db mice (B), and Pio-db/db mice (C). (D) Percentages of islets exhibiting endogenous calcium oscillations; (E) amplitude of calcium oscillations; (F) period of oscillations. Data in panels D to F represent results obtained from analysis of at least 30 islets from 12 different animals per group. *, statistically (P < 0.05) different compared to C57BLKS/J mice; #, statistically (P < 0.05) different compared to db/db mice.

Pioglitazone preserves islet function in the HFD feeding model.As leptin signaling has previously been shown to play an independent role in islet function in isolation (34), we sought to extend our observations to a second mouse model. To create a model of glucose intolerance, we subjected C57BLKS/J mice to a diet containing 42% calories from fat for 4 weeks, consistent with a Western style of dietary intake. Mice fed an HFD were gavaged daily with pioglitazone or vehicle control, which began with diet initiation. This model enabled us to observe the islet-specific effects of pioglitazone treatment in a model with a phenotype of lesser severity and without a genetic defect in leptin signaling. As shown in Fig. 6A, mice fed an HFD gained significantly more weight than mice fed regular chow. There was no difference in weight between the two HFD-fed groups. Notably, whereas control HFD-fed mice exhibited significantly worse glucose tolerance than chow-fed animals, pioglitazone-treated HFD mice were indistinguishable in this regard from chow-fed mice (Fig. 6B). Insulin drawn during the glucose tolerance test represented in Fig. 6C revealed higher baseline (fasting) levels for both HFD-fed groups compared to the chow group, consistent with relative insulin resistance. Importantly, however, control HFD-fed mice had a blunted insulin response to glucose challenge compared to pioglitazone-treated HFD-fed mice, suggesting that PPAR-γ activation by pioglitazone appeared to preserve islet function in this mouse model of glucose intolerance. To directly determine the functionality of islets, we next isolated islets from all three groups of mice and subjected them to glucose-stimulated insulin secretion studies in vitro (Fig. 6D). Whereas islets from control HFD-fed mice demonstrated a significantly blunted insulin secretory response to glucose compared to the chow-fed controls, islets from pioglitazone-treated mice showed a complete restoration of insulin secretion. These results suggest that PPAR-γ activation leads to improved islet function independently of leptin signaling.

• Open in new tab
• Download powerpoint

FIG. 6.

Pioglitazone treatment improves islet function in mice fed an HFD. Male C57BLKS/J mice were fed an HFD (42% of calories from fat) and gavaged with either vehicle (HFD) or pioglitazone (HFD+Pio) for 4 weeks. Results were compared to those obtained with age- and sex-matched C57BLKS/J mice fed a regular chow diet (17% of calories from fat [Chow]). (A) Body weight at end of study; (B) results of intraperitoneal glucose tolerance tests at the end of study. Pio + HFD animals exhibited significantly improved glucose tolerance compared to HFD animals (P < 0.05 for the comparison by two-way ANOVA). (C) Results of analysis of serum insulin levels at times 0 and 30 min during the glucose tolerance test represented in panel B. (D) Islets were incubated in 3 and 28 mM glucose for 1 h, and insulin secretion into the supernatant was measured by ELISA. Data represent the means of the results of at least three independent experiments performed using 50 islets per group. The results shown in panels C and D were analyzed by one-way ANOVA. #, significantly (P < 0.05) different compared to Chow mice; *, significantly (P < 0.05) different compared to HFD mice.

Pioglitazone treatment enhances expression of genes necessary for glucose-stimulated insulin secretion and calcium mobilization.To explain the improvements in islet function seen with pioglitazone treatment, we asked how the expression of genes important for islet function was affected. Data from quantitative real-time RT-PCR using total islet RNA from C57BLKS/J mice and vehicle- and pioglitazone-treated db/db mice are shown in Fig. 7. Compared to islets from the background strain, islets from diabetic db/db mice had significant reductions in the expression of several genes involved in β-cell development, growth, and function, including NeuroD1, Nkx6.1, Glut2, Ins1/2, Kcnj11, Irs1, and Lpl (Fig. 7A to C). Pioglitazone was effective in significantly restoring or improving the expression of Glut2, Ins1/2, Kcnj11, Nkx6.1, NeuroD1, Pdx1, and Irs1 within the islet (Fig. 7A to C).

• Open in new tab
• Download powerpoint

FIG. 7.

Pioglitazone treatment improves the expression of key β-cell genes. Islets from age- and sex-matched male C57BLKS/J and db/db mice, treated with either vehicle (db/db) or pioglitazone (Pio-db/db) for 6 weeks, were harvested and subjected to real-time RT-PCR for analysis of islet growth and function genes (A), glucose-sensing genes (B), growth factor-signaling genes (C), and SERCA genes (D). Data represent the means ± standard errors of the results of at least three different biologic replicate experiments. Data were analyzed by one-way ANOVA. *, statistically (P < 0.05) different compared to C57BLKS/J mice; #, statistically (P < 0.05) different compared to db/db mice.

Because previous data have demonstrated a positive link between IRS1 signaling and the expression of genes involved in intracellular calcium homeostasis (28), we also examined how PPAR-γ activation might affect the expression of the genes encoding the SERCAs. In this regard, db/db mice have also been shown to exhibit reduced SERCA activity and protein expression in islets, a finding that may explain the defects in GSCa and calcium oscillations that we observed (47). As shown in Fig. 7D, expression of genes encoding SERCA2a, SERCA2b, and SERCA3 was decreased 8- to 10-fold in db/db mice compared to the control background strain. Strikingly, pioglitazone treatment resulted in a dramatic improvement in the expression of each of these genes, restoring them to the levels seen in the control strain (Fig. 7D).

Pioglitazone treatment enhances markers of euchromatin at the Ins1/2 and Glut2 genes.In prior studies, we and others had demonstrated that β-cell genes, particularly Ins1/2 and Glut2, appear to be regulated through changes in covalent histone modifications that alter chromatin accessibility (6). Expression of Ins1/2 and Glut2 was increased 15- and 7-fold, respectively, in islets following 6 weeks of daily oral therapy with pioglitazone (Fig. 7). To investigate a link between PPAR-γ activation and a chromatin basis for Ins1/2 and Glut2 activation, we performed ChIP assays using islet extracts from vehicle- and pioglitazone-treated mice. ChIP assays were performed using antibodies to the euchromatin markers acetylated histone H3, acetylated histone H4, and histone H3-dimethyl-Lys4, and their abundance at the proximal Ins1/2 promoter was quantitated using real-time PCR. As shown in Fig. 8A and B, no differences between vehicle control and treated db/db mice in acetylated H4 levels at the proximal Ins1/2 promoters were observed, although a statistically significant twofold increase in acetylated H3 was observed in the treatment group compared to the controls. In prior studies, we demonstrated a striking enrichment of H3-dimethyl-Lys4 at the Ins1/2 and Glut2 genes in β-cell lines and islets and showed this modification to be closely linked to RNA polymerase activation and gene transcription (6, 11). As shown in Fig. 8C, the H3-dimethyl-Lys4 marker diminished progressively at the Ins1/2 gene with progressive diabetes in the db/db mouse. Pioglitazone therapy enhanced H3-dimethyl-Lys4 nearly threefold at both the Ins1/2 and Glut2 genes, suggesting that both genes had been activated via a mechanism enhancing euchromatin.

• Open in new tab
• Download powerpoint

FIG. 8.

Pioglitazone treatment enhances euchromatin markers at the Ins1/2 and Glut2 promoters. Islets from male db/db mice, treated with either vehicle (db/db) or pioglitazone (Pio-db/db) for 6 weeks, were harvested and subjected to ChIP analysis as detailed in Materials and Methods. Real-time PCR was used to quantitate recovery of the proximal Ins1/2 and Glut2 promoters, and results are expressed as percent recovery of the gene fragment relative to input chromatin. (A) ChIP using normal rabbit serum (NRS) or antibody to acetylated H4 (Ac H4), followed by analysis of recovered Ins1/2 promoter fragments; (B) ChIP using NRS or antibody to acetylated H3 (Ac H4) followed by analysis of recovered Ins1/2 promoter fragments; (C) ChIP using NRS or antibody to H3-dimethyl-Lys4, followed by analysis of recovered Ins1/2 promoter fragments; (D) ChIP using NRS or antibody to H3-dimethyl-Lys4, followed by analysis of recovered Glut2 promoter fragments. Results represent the means ± standard errors of the results of three independent ChIP experiments. Results were analyzed using either a t test or one-way ANOVA (C). *, results were significantly (P < 0.05) different for the comparison shown.

PPAR-γ activation increases nuclear Set7/9 localization.To investigate the underlying mechanism for the observed enhancement of H3-dimethyl-Lys4 at the Ins1/2 and Glut2 genes, we examined the mRNA and protein levels of Set7/9. Set7/9 is a histone H3-Lys4-specific methyltransferase (37, 54) that is enriched in islets, physically interacts with Pdx1, and occupies known Pdx1 target genes such as Ins1/2 and Glut2 (4, 6, 11). RT-PCR and immunoblot analysis of islets revealed that Set7/9 protein levels were increased in response to pioglitazone treatment but with no effect seen on the transcript levels of the gene encoding Set7/9 (Setd7) (Fig. 9A and B). Consistent with the RT-PCR data (Fig. 7A), Pdx1 protein levels were also increased in islets from pioglitazone-treated mice (Fig. 9B). Interestingly, whereas Set7/9 exhibits a predominantly nuclear staining pattern in islets of normal mice (11), it displayed a strikingly more cytoplasmic pattern in islets from vehicle-treated db/db mice (Fig. 9C and G). Pioglitazone treatment of db/db mice, however, appeared to completely restore the nuclear staining pattern of Set7/9 (Fig. 9D and H), thereby suggesting a potential mechanism for the increased H3-dimethyl-Lys4 modifications observed at the Ins1/2 and Glut2 genes in treated animals. Importantly, Pdx1 staining, although reduced in control db/db mice, remained nuclear in both groups (Fig. 9E and H).

• Open in new tab
• Download powerpoint

FIG. 9.

Expression patterns of Pdx-1 and the methyltransferase Set7/9. Islets and pancreata from male db/db mice, treated with either vehicle (db/db) or pioglitazone (Pio-db/db), were harvested and subjected to real-time RT-PCR, immunoblotting, or immunohistochemistry. (A) Results of real-time RT-PCR analysis for Setd7 mRNA by the use of RNA from isolated islets. Data represent the means ± standard errors of the results obtained from two independent islet isolation experiments using six animals in each experiment. (B) Results of immunoblot analysis for Set7/9 (upper panel), Pdx-1 (lower panel), or GAPDH (both panels) obtained using total protein from isolated islets. Data are from pooled-islet experiments performed using six different animals per group. Data from a second islet pool were similar. (C to H) Pancreata from db/db and Pio-db/db animals were peroxidase stained for Set7/9 (C and D) and Pdx-1 (E and F) and counterstained with hematoxylin or stained for both Set7/9 (green) and Pdx-1 (red) and visualized by immunofluorescence (G and H). Nuclei were counterstained using Hoechst dye in (G and H). Data in panels C to H show representative islets from among three pancreata analyzed per group of mice.

PPAR-γ activation directly improves islet function in db/db mice.To characterize the potential direct effect of pioglitazone on islet function, we incubated islets from 8- to 9-week-old diabetic db/db mice with pioglitazone at a concentration of 10 μΜ for 24 h, followed by an assessment of islet function and gene expression. Islet function was assessed via the measurement of GSCa and insulin secretion. Although GSCa results were not as robust in islets treated in vitro as in islets exposed to chronic pioglitazone treatment in vivo (compare Fig. 10A and Fig. 4A), we did observe significant improvements in treated islets compared to untreated islets. Figure 10A and B demonstrate that pioglitazone-treated islets exhibited a significantly reduced basal fura-2 ratio and an improved GSCa index, trends similar to those observed in islets exposed to pioglitazone in vivo. However, static glucose-stimulated insulin release assays failed to reveal differences in insulin secretory function between the two treatment groups, suggesting that the Ca²⁺ imaging data reflect a more sensitive early measure of islet functionality.

• Open in new tab
• Download powerpoint

FIG. 10.

PPAR-γ activation in vitro improves glucose-stimulated calcium response in db/db islets. Islets from 8-week-old male db/db mice were incubated for 24 h with 10 μM pioglitazone (Pio-db/db) or vehicle control (db/db). (A) Results of GSCa studies of isolated islets. The panel shows the continuous fura-2 AM fluorescence ratio (340/380 nm) changes as glucose in the incubation chamber was increased from 3 mM to 28 mM. Data represent the means ± standard errors of the results for at least 10 islets per group. (B) Data from panel A were used to calculate a GSCa index, which represents the fura-2 AM fluorescence ratio under conditions of 28 mM glucose divided by the ratio under conditions of 3 mM glucose. (C) Islets were incubated in 3 and 28 mM glucose for 1 h, and insulin secretion into the supernatant was measured by ELISA. Data represent the means of the results of at least three independent experiments performed using 50 islets per group. Results were analyzed using a t test. *, statistically different (P < 0.05) compared to db/db islets treated with vehicle control.

To characterize what factors might be contributing to the observed improvements in islet Ca²⁺ responses, we first measured the expression of the same key β-cell genes as examined in our studies in vivo (Fig. 7A and B). Surprisingly, we did not see changes in the expression of any of these genes following short-term treatment with pioglitazone (data not shown). Importantly, however, the genes encoding the SERCA proteins were all uniformly activated, in similarity to data observed with animals treated in vivo (Fig. 11A). We hypothesized that the improved GSCa and enhanced SERCA gene expression might have arisen from a reduction in ER stress. Others have previously described an induction of ER stress in islets from db/db mice as a mechanism underlying the pathogenesis of islet failure in these animals (9). Moreover, known inducers of ER stress (e.g., thapsigargin) have been shown to produce defects in GSCa that are similar to those we observed in islets of untreated db/db mice (47). Thus, we considered the possibility that the direct effects of pioglitazone on db/db islet calcium homeostasis that we observed here may have been secondary to attenuation of ER stress. We measured the expression of several genes in the ER stress pathway, including Chop, Bip, and total and spliced Xbp1. We observed a significant reduction in the expression of Chop and spliced Xbp1, together with a nonsignificant decrease in total Xbp1 expression (Fig. 11B). There was no difference in the levels of Bip expression between treated and untreated islets. These data suggest that a direct effect of PPAR-γ activation in islets may be to reduce ER stress.

• Open in new tab
• Download powerpoint

FIG. 11.

PPAR-γ activation in vitro upregulates SERCA genes and reduces ER stress in db/db islets. Islets from 9- to 10-week-old db/db mice, treated with either vehicle (db/db) or 10 μM pioglitazone (db/db + Pio) for 24 h, were subjected to real-time RT-PCR for analysis of SERCA genes (A) and ER stress markers (B). Data represent the means ± standard errors of the results of five biologic replicate experiments. Results were analyzed by one-way ANOVA. *, significantly (P < 0.05) different compared to db/db mouse group results.

PPAR-γ activation reduces ER stress and maintains Set7/9 nuclear occupancy.To investigate further the relationship between PPAR-γ activation, ER stress, and Set7/9 nuclear occupancy, we studied a model of β-cell ER stress in vitro through the use of INS-1 (832/13) cells, a well-characterized glucose-responsive β-cell line (19). INS-1 cells were treated with 1 μM thapsigargin for 6 h to induce a state of ER stress similar to that observed in db/db mice (47). As shown in Fig. 12, treatment with thapsigargin led to a reduction in Pdx1 protein levels and increases in CHOP and cleaved caspase 3 levels, whereas coincubation with pioglitazone led to recovery of Pdx1 levels and decreases in CHOP and cleaved caspase 3 levels. These data demonstrate a direct effect of PPAR-γ activation in the reduction of ER stress and preservation of Pdx1 levels in β cells.

• Open in new tab
• Download powerpoint

FIG. 12.

PPAR-γ activation in vitro reduces ER stress in INS-1 (832/13) β cells. (A) INS-1 β cells were incubated with vehicle (DMSO), 1 μM thapsigargin (Thap), or 1 μM thapsigargin plus 10 μM pioglitazone (Thap + Pio) for 6 h and then subjected to immunoblot analysis for the proteins indicated. (B to D) Quantitation of the immunoblots shown in panel A. Data represent the means ± standard errors of the results of three independent biological replicate experiments. Data were analyzed by one-way ANOVA, and statistical significance (P < 0.05) was determined for the data in panel D for the comparison of Thap versus DMSO and of Thap + Pio versus Thap.

Next, we assessed the nuclear occupancy of Set7/9 in INS-1 cells treated with thapsigargin and/or pioglitazone. Panels A and B of Fig. 13 show that Set7/9 occupies a slightly more predominant nuclear distribution in INS-1 cells (see quantitation in Fig. 13G), whereas treatment with thapsigargin led to strongly cytoplasmic redistribution (see Fig. 13C, D, and G) similar to that observed in db/db mice. Concurrent treatment with thapsigargin and pioglitazone caused restoration of Set7/9 nuclear occupancy, a finding again similar to that observed with pioglitazone-treated db/db mice (Fig. 13E, F, and G). Thus, the results in Fig. 13 suggest that the islet chromatin and functional phenotype we observed in pioglitazone-treated db/db mice may have resulted from a direct action of PPAR-γ in reducing ER stress in the islet.

• Open in new tab
• Download powerpoint

FIG. 13.

PPAR-γ activation in vitro promotes nuclear recovery of Set7/9 in INS-1 (832/13) β cells. INS-1 β cells were incubated with vehicle (DMSO) (A and B), 1 μM thapsigargin (Thap) (C and D), or 1 μM thapsigargin plus 10 μM pioglitazone (Thap + Pio) (E and F) for 6 h and were then subjected to fixation and immunostaining for Set7/9 or DAPI (to visualize nuclei), as indicated at the top. (G) The relative ratios of nuclear intensity to cytoplasmic intensity for Set7/9 were calculated for a minimum of 10 cells per set of conditions. Data in panel G were analyzed by one-way ANOVA, and statistical significance (P < 0.05) was obtained for the comparision of Thap versus DMSO and of Thap+Pio versus Thap.

Finally, we asked how PPAR-γ activation could lead to alterations in Set7/9 localization. In prior studies, we showed that Set7/9 is recruited to target genes such as Ins1/2 and Glut2 via physical interaction with Pdx1 (6, 11). We hypothesized therefore that this interaction may lead to the “chaperoning” of Set7/9 into the nucleus and that states that lead to reduced Pdx1 levels (such as ER stress) may reduce Set7/9 nuclear transport. To test this hypothesis directly, we utilized Tet-off mice (see reference 20). As shown in Fig. 14, control mice exhibited normal nuclear localization of both Pdx1 and Set7/9 in pancreatic islets. However, administration of doxycycline for 7 days led to nearly complete elimination of Pdx1 and to both reduction and redistribution of Set7/9 into the cytoplasm. These data therefore suggest strongly that recovery of Pdx1 levels upon PPAR-γ activation (either by direct activation of the Pdx1 gene or indirectly through the reduction of ER stress) may therefore allow recovery of Set7/9 levels in the nucleus.

• Open in new tab
• Download powerpoint

FIG. 14.

Inducible deletion of Pdx1 in pancreatic islets causes a nuclear-to-cytoplasmic shift in Set7/9. Pdx1^tTA/tTA; Tg^Pdx1 mice (20) were treated with vehicle for 7 days as described in Materials and Methods, and pancreata were harvested and fixed for immunostaining for Pdx1, Set7/9, and DAPI, as indicated. (A to C) Pancreatic section from a representative control animal treated with vehicle (−DOXY); (D to F) pancreatic section from a representative animal treated with doxycycline (+DOXY).