PED/PEA-15 Regulates Glucose-Induced Insulin Secretion by Restraining Potassium Channel Expression in Pancreatic β-Cells

The rip-I/ped/pea-15 chimeric gene was obtained by introducing a 2.3-Kb EcoRI fragment containing the entire human ped/pea-15 cDNA at the EcoRI site in a pB5-RIP-β-globin expression vector (kindly provided by Prof. D. Accili, Columbia University, NY). This chimeric gene was microinjected into fertilized C57BL6/SJL mouse eggs. The general procedures used for microinjection and animal selection have been previously described (18). Six F₀ founders were identified by PCR as in reference 13. Mice were fed ad libitum with a standard diet (Research Diets formulas D12328; Research Diets, New Brunswick, NJ) and kept under a 12-h light/12-h dark cycle. Tissue samples were collected rapidly after mice were killed by pentobarbitone overdose. Tissues were snap frozen in liquid nitrogen and stored at −80°C. All procedures described below were approved by the institutional animal care and utilization committee. Blood glucose levels were measured with glucometers (A. Menarini Diagnostics, Florence, Italy); insulin was measured by radioimmunoassay (RIA) with rat insulin as standard (Insulin Rat RIA kit; Linco Research, St. Louis, MO). Fasting plasma free fatty acids were measured with the Wako NEFA C kit (Wako Chemicals, Richmond, VA), and triglycerides were measured with the Infinity triglyceride reagent (Sigma-Aldrich, St. Louis, MO).

Glucose tolerance tests, insulin tolerance tests, and insulin secretion were measured as previously described (19,20). For analyzing glucose utilization by skeletal muscle, an intravenous injection of 1 μCi of the nonmetabolizable glucose analog 2-[1-³H]deoxy-d-glucose (2-DG) (Amersham Pharmacia Biotech, Piscataway, NJ) and an intraperitoneal injection of insulin (0.75 mU/g body wt) were administered to random fed mice. The specific blood 2-DG clearance was determined with 25-μl blood samples (tail vein) obtained 1, 15, and 30 min after injection as previously reported (21). Skeletal muscles were removed 30 min after the injections. Glucose utilization index was determined by measuring the accumulation of radiolabeled compound (22). The amount of 2-DG-6 phosphate per milligram of protein was then divided by the integral of the concentration ratio of 2-DG to the measured unlabeled glucose. Glucose utilization indexes were expressed as picomoles per milligram of protein per minute.

Generation of PED/PEA-15 antibody has been previously reported (8). Immunoistochemical detection of PED/PEA-15, insulin, and glucagon in pancreases from control and transgenic mice were analyzed as described (13). Analysis of serial consecutive islet sections stained with either insulin or PED/PEA-15 antibodies was used to confirm PED/PEA-15 expression in insulin-immunopositive β-cells. Sections were also stained with hematoxylin and eosin. For morphometry, pancreases were obtained from 6-month-old control and transgenic mice, and immunohistochemical detection of insulin was performed in three sections (2–3 μm) separated by 200 μm as in reference 18.

Islets were isolated from 6-month-old mice by collagenase digestion and subsequent centrifugation on a Histopaque (Sigma-Aldrich) gradient as in reference 23. A total of 20 islets were manually selected and preincubated in Dulbecco’s modified Eagle’s medium (Life Technologies) at 37°C in a 5% CO₂ atmosphere for 24 h. Islets were then further incubated in Krebs-Ringer buffer (120 mmol/l NaCl, 1.2 mmol/l MgSO₄, 5 mmol/l KCl, 10 mmol/l NaHCO₃, 1.3 mmol/l CaCl₂, and 1.2 mmol/l KH₂PO₄) for 30 min and then stimulated at 37°C with various concentrations of either glucose for 1 h, KCl for 30 min, or glyburide (Sigma-Aldrich) for 1 h. Islets were subsequently collected by centrifugation and supernatants assayed for insulin content by RIA. For measuring total pancreatic insulin content, pancreases were solubilized in acid-ethanol solution (75% ethanol, 1.5% HCl) overnight at 4°C and centrifuged for 15 min at 800g. Insulin was measured in the supernatants by RIA.

MIN-6 cells were cultured in Dulbecco’s modified Eagle’s medium containing 25 mmol/l glucose, 50 μmol/l 2-mercaptoethanol, and 10% FCS (Biochrom) at 37°C in a 5% CO₂ atmosphere (24). INS-1 cells were cultured in RPMI-1640 medium (Life Technologies) containing 11 mmol/l glucose, 50 μmol/l 2-mercaptoethanol, and 10% fetal bovine serum (Biochrom) at 37°C in a 5% CO₂ atmosphere (25). Stable transfection of the ped/pea-15 cDNA was performed with the lipofectamine method (13) according to the manufacturer’s instructions. Transient transfection of the PKC-ζ cDNA was also performed with the lipofectamine method. By using PCAGGS-β-Gal as a reporter, transfection efficiency was between 65 and 85%, staining with the chromogenic substrate 5-bromo-4-chloro-3-indolyl b-d-galactopyranoside.

Total cellular RNA was isolated from pancreatic islets and tissue samples by using the RNeasy kit (Qiagen Sciences) according to the manufacturer’s instructions. Northern blot analysis was performed as in reference 18. For real-time RT-PCR analysis, 1 μg islet or cell RNA was reverse transcribed using Superscript II Reverse Transcriptase (Invitrogen, Carlsbad, CA). PCR were analyzed using SYBR Green mix (Invitrogen). Reactions were performed using Platinum SYBR Green qPCR Super-UDG using an iCycler IQ multicolor Real Time PCR Detection System (Biorad, Hercules, CA). All reactions were performed in triplicate, and β-actin was used as an internal standard. Primer sequences used were as follows: PED/PEA-15 forward 5′-TTCCCGCTGTTCCCTTAGG-3′, PED/PEA-15 reverse 5′-TCTGGCTCATCCGCATCC-3′; Kir6.2 forward 5′-TCCACCAGGTAGACATCCC-3′, inwardly rectifying K⁺ channel 6.2 (Kir6.2) reverse 5′-TAGGAGCCAGGTCGTAGAG-3′; sulfonylurea receptor 1 (SUR1) forward 5′-GCATCAACTTGTCTGGTG-3′, SUR1 reverse 5′-ACTGTCCTCTTGTCATCC-3′; forkhead box A2 (Foxa2) forward 5′-CACCTGAGTCCGAGTCTG3′, Foxa2 reverse 5′-CGAGTTCATGTTGGCGTAG-3′; glucose transporter 2 (Glut2) forward 5′-ACAGTCACACCAGCATAC-3′, Glut2 reverse 5′-ACCCACCAAAGAATGAGG3′; glucokinase (GK) forward 5′-GCTGTACGAAAAGATCATTGG-3′, GK reverse 5′-CTCCGTGGAACAGAAGATTC-3′; hexokinase (HK)1 forward 5′-ACGACACCCCAGAGAACATC-3′, HK1 reverse 5′-AGTCTCCGAGGCATTCAGC-3′; HK2 forward 5′-CTCAGGACACCACAGGCTAC-3′, HK2 reverse 5′-TTCACTCACTGGGTCACTAAGG-3′; and β-actin forward 5′-GCGTGACATCAAAGAGAAG-3′, β-actin reverse 5′-ACTGTGTTGGCATAGAGG-3′.

Tissues homogenates and cell lysates were separated by SDS-PAGE and analyzed by Western blot as previously described (13,17,26). Membranes were probed with antibodies to PED/PEA-15 (8), PKC-ζ (Santa Cruz Biotechnology, Santa Cruz, CA), or tubulin (Santa Cruz Biotechnology). For analyzing atypical PKC activity, cells were incubated with Krebs-Ringer buffer for 30 min and then incubated with either 2.8 or 16.7 mmol/l glucose containing medium for 1 h at 37°C in a 5% CO₂ atmosphere. Cells were solubilized in lysis buffer (25 mmol/l Tris-HCl, pH 7.4, 0.5 mmol/l EDTA, 0.5 mmol/l EGTA, 0.05% Triton X-100, 10 mmol/l β-mercaptoethanol, 1 μg/ml leupeptin, and 1 μg/ml aprotinin) for 1 h at 4°C. Lysates were precipitated with a PKC-ζ antibody (17), and PKC activity was assayed using the SignaTECT PKC assay system (Promega, Madison, WI) according to the manufacturer’s instructions. Determination of PKC activity using the Ac-MBP (4–14) or the pseudosubstrate region of PKC-ε (specific for atypical PKC) provided consistent results.

Data were analyzed with Statview software (Abacus Concepts) by one-factor ANOVA. P values <0.05 were considered statistically significant. The total area under the curve for glucose response during the insulin tolerance test was calculated by the trapezoidal method (13).

To investigate the relevance of PED/PEA-15–induced β-cell dysfunction to glucose tolerance, we used the insulin promoter to generate transgenic mice featuring selective overexpression of PED/PEA-15 in the β-cells (β-tg; Fig. 1A). Three lines of β-tg mice (L5, L8, and L9) were established in which ped/pea-15 mRNA as well as PED/PEA-15 protein showed a 10- to 20-fold increased expression in the β-cells but unchanged expression in other tissues (Fig. 1B and C). Transgenic mice were fertile and generated viable offsprings showing no significant growth alterations compared with their nontransgenic littermates (controls) or other apparent abnormalities. L8 ped/pea-15 transgenic mice, both male and female, exhibited slightly reduced fasting blood glucose levels compared with control mice (males 84 ± 6 vs. 91 ± 7 mg/dl; females 82 ± 8 vs. 92 ± 6 mg/dl; significant at P < 0.05). Similarly reduced fasting blood glucose levels were measured in the other transgenic lines (data not shown). However, glucose loading (2 g/kg) rendered the β-tg mice significantly more hyperglycemic during the following 120 min compared with control mice (Fig. 2A and B). Very similar abnormalities were observed in the L5 and L9 lines (data not shown), indicating that β-cell overexpression of PED/PEA-15 is sufficient to impair glucose tolerance in mice. As shown in Fig. 2, sex had no effect on the impairment in glucose tolerance caused by β-cell overexpression of PED/PEA-15. Aging (3–14 months) also determined no significant changes in PED/PEA-15–induced decreases in glucose tolerance (data not shown).

L8 β-tg mice featured no significant differences in fasting nonesterified free fatty acid and triglyceride blood concentrations compared with their controls (free fatty acids 0.82 ± 0.03 vs. 0.78 ± 0.05 mmol/l; triglycerides 115 ± 7 vs. 111 ± 8 mg/dl). However, fasting insulin levels were 80% increased in the transgenics (0.31 ± 0.07 vs. 0.17 ± 0.03; P < 0.001). The same results were obtained in the L5 and the L9 lines, with no sex- or age-related variation (data not shown). To verify whether β-cell overexpression of PED/PEA-15 is accompanied by reduced insulin sensitivity, we performed insulin tolerance tests. Intraperitoneal injection of insulin (0.75 mU/g) evoked comparable hypoglicemic responses in both β-tg and control mice (Fig. 2C) with no significant difference in glucose areas under the curve in all of the transgenic lines (C 9,180 ± 1,900 mg/dl per 120 min; L5 8,898 ± 1,800 mg/dl per 120 min; L8 8,805 ± 1,600 mg/dl per 120 min; L9 9,121 ± 1,600 mg/dl per 120 min). Furthermore, insulin-dependent glucose uptake in quadriceps and tibialis muscles from β-tg and control mice exhibited no difference (Fig. 2D).

We then investigated how glucose tolerance is reduced in the β-tg and analyzed their islets by immunohistochemistry and morphometry. Islets from β-tg mice immunostained with PED/PEA-15 antibody significantly more intensely than those from control mice (Fig. 3A). Importantly, islets from the β-tg mice were larger and showed a more elongated and irregular shape but feature normal distribution of α- and β-cells compared with control islets (Fig. 3B). Also, quantitative morphometry revealed twofold increased islet and β-cell mass per unit of total pancreatic area in the β-tg mice (P < 0.01) with no significant difference in their pancreas weight (Table 1).

In control mice, a sevenfold increase in insulin secretion was observed 3 min after intraperitoneal glucose injection, and the levels remained higher than baseline values for up to 30 min, indicating second-phase response (Fig. 3C). Based on the area under the curve, the acute first-phase insulin secretory response to glucose was reduced by more than threefold in β-tg mice (P < 0.001). The second-phase response was also significantly impaired in the β-tg compared with control mice (P < 0.001). Again, almost identical decreases in early and late insulin responses were observed in both male and female β-tg mice and in all of the lines produced (data not shown). Thus, in vivo, isolated β-cell overexpression of PED/PEA-15 increases fasting insulinemia but impairs further insulin secretion response to hyperglycemia.

Similar as in intact mice, β-tg mouse islets also exhibited a twofold increased insulin release when incubated in medium containing 2.8 mmol/l glucose (Fig. 4A; P < 0.001). Raising glucose concentration did not further increase the insulin release by these islets, while enhancing that from control mice by 3.9-fold (P < 0.001). Similarly, islets from transgenic mice ubiquitously expressing PED/PEA-15 (13) showed 2.4-fold increased insulin release when exposed to 2.8 mmol/l glucose (P < 0.001) and no further release upon exposure to 16.7 mmol/l glucose (Fig. 4B). At variance with glucose, exposure to the membrane depolarising agent potassium chloride (33 mmol/l) caused a comparable release of insulin by the β-tg and the control islets, suggesting an effect of PED/PEA-15 on earlier events of β-cell response to glucose. Total insulin content in pancreases was 25% increased in β-tg mice (2.9 ± 0.4 ng/mg in β-tg vs. 2.3 ± 0.2 ng/mg in control mice; P < 0.05).

To gain further insight into the mechanisms leading to dysfunction of β-cells overexpressing PED/PEA-15, we have profiled the expression of several genes relevant to the glucose sensory apparatus by real-time RT-PCR analysis of total RNA isolated from islets of β-tg mice (Fig. 4C). The amounts of mRNAs for the Sur1 subunit of the ATP-sensitive K⁺ channel was significantly reduced by 30% in these islets compared with those in control mice (P < 0.001). Consistently, islets from both the β-tg and the transgenic mice featuring ubiquitous overexpression of PED/PEA-15 exhibited no secretory response to the K⁺ channel locking agent glyburide (Fig. 4D). RNAs encoding the Kir6.2 potassium channel subunit and the Sur1/Kir6.2 upstream regulator Foxa2 were also reduced, respectively, by 30 and 35% in the islets from the ped/pea-15 β-cell–specific transgenics (P < 0.01). Decreased expression of Sur1, Kir6.2, and Foxa2 mRNAs was confirmed by Northern blot analysis (Fig. 4E). At variance, the abundance of glucokinase, HK1, HK2, and GLUT2 mRNAs did not differ between transgenic and control mice (Fig. 4C and E).

A ∼35% decrease in the abundance of the Sur1, Kir6.2, and Foxa2 mRNAs has also been observed in the glucose-responsive MIN-6 and INS-1 β-cell lines upon transfection with ped/pea-15 cDNA (Min-6_ped/pea-15, INS-1_ped/pea-15; Fig. 5A; P < 0.001). As in the mouse islets, these changes were accompanied by a significantly greater release of insulin when in low-glucose medium (2.8 mmol/l glucose) with no further insulin secretion upon exposure to 16.7 mmol/l glucose (Fig. 5B; P < 0.001). This raise in glucose concentration caused a 2.5- and 4-fold enhancement in insulin secretion, respectively, by the untransfected INS-1 and MIN-6 cells, however. In both the mouse islets and the cultured β-cell lines, the effect of glucose on insulin release was paralleled by a twofold increase in immunoprecipitated PKC-ζ activity, with no measurable change in its expression (Fig. 5C). Interestingly, however, the effect of glucose was almost completely abolished in cells overexpressing PED/PEA-15 whether the MIN-6 or the β-tg mouse islets. Similar data were obtained in the INS-1 cells (data not shown), further suggesting an important role of atypical PKCs in regulating genes of the insulin secretion pathway. To further analyze this issue, we forced the expression of PKC-ζ in the MIN-6_ped/pea-15 and the INS-1_ped/pea-15 cells. The overexpression largely rescued PKC-ζ activity in the precipitates from glucose-exposed MIN-6_ped/pea-15 cells (Fig. 6A). Importantly, this rescue was accompanied by recovery of the Sur1, Kir6.2, and Foxa2 gene expression (Fig. 6B). In part, the changes in insulin release caused by PED/PEA-15 were also reverted, determining recovery of glucose-induced insulin secretion (Fig. 6C). The same results were obtained using the INS-1 cells (data not shown), indicating that the changes in atypical PKC function caused by PED/PEA-15 overexpression represent upstream abnormalities impairing the function of multiple genes involved in regulating insulin secretion by the β-cell.

To further address this issue, we used islets from ped/pea-15–null mice. These animals have been previously characterized and reported (27) and feature no PED/PEA-15 expression in β-cells (Fig. 7A). Importantly, islets from these animals evidenced a twofold increased PKC-ζ activation following glucose exposure, as compared with islets from control mice (P < 0.01; Fig. 7B). In parallel with this enhanced response, islets from the ped/pea-15–null mice exhibited a significant twofold increase in sensitivity to glucose and glyburide effect on insulin secretion compared with control islets, with unchanged response to potassium chloride (P < 0.001; Fig. 7C). Abundance of the Sur1, Kir6.2, and Foxa2 mRNAs was also increased by 45% in these islets (P < 0.001; Fig. 7D), indicating that these genes represent major effectors through which PED/PEA-15 physiologically controls insulin secretion.

In the present study, we generated transgenic mice overexpressing PED/PEA-15 selectively in β-cells (β-tg mice). We found that PED/PEA-15 overexpression determines islet hyperplasia with a specific increase in β-cell mass. This may result from a direct effect on β-cell survival as PED/PEA-15 exerts a broad antiapoptotic action in a number of different cell types (10,15,27–30), including β-cells (C.M., unpublished observation). Increased β-cell mass is consistent with the increased fasting serum insulin levels detected in the β-tg mice. Indeed, these mice do not feature resistance to insulin action. They do show grossly impaired glucose tolerance and reduced insulin secretion in response to hyperglycemia, however. This latter defect is comparable with that occurring in mice with ubiquitous overexpression (13), leading us to conclude that β-cell overexpression of PED/PEA-15 is sufficient to impair glucose tolerance in mice. Previous studies in isolated cells and in vivo evidenced that chronic insulin resistance may impair β-cell function, ultimately reducing glucose response (2). Secondary impairment of β-cell function may also occur in the transgenic mice overexpressing PED/PEA-15 ubiquitously. However, the findings in the present study evidence that impaired response to glucose is a direct consequence of PED/PEA-15 overexpression in the β-cells. Consistently, islets from β-tg mice persistently show similarly impaired glucose response in culture as well as in the intact animal. Also, the expression of ped/pea-15 cDNA in different β-cell lines inhibits insulin release following increased glucose concentration in the culture medium.

The defective insulin secretion induced by β-cell overexpression of PED/PEA-15 affected both the early and the late phases of insulin response to glucose, raising the possibility that the overexpression of PED/PEA-15 impairs the glucose-sensing machinery. Earlier reports evidenced that the KIR6.2 and SUR1 potassium channel subunits couple glucose metabolism to the electrical activity of β-cell membrane and insulin secretion by setting the resting membrane potential below the activation threshold for voltage-gated Ca²⁺ channels (31). We found that the membrane depolarizing agent potassium chloride causes a similar release of insulin by both transgenic and control mouse islets, indicating that the site of PED/PEA-15 action on glucose-induced secretion is upstream the membrane depolarization step. Indeed, β-cell expression profiling revealed that mRNA abundance of both the Kir6.2 and the Sur1 transcripts are significantly depressed by the overexpression of ped/pea-15. PED/PEA-15 action on the Kir6.2 and Sur1 genes represents a specific abnormality, as other genes encoding key components of the glucose-sensitive insulin secretion machinery were unaffected by PED/PEA-15. The latter include GLUT2, HK1, and HK2. Consistent with the functional consequence of this defect, glyburide showed impaired action on insulin secretion in PED/PEA-15–overexpressing islets.

Previous studies in mice evidenced that ablation of either the Sur1 or the Kir6.2 genes results in a phenotype reminiscent of that characterizing the β-tg mice (32–34). Indeed, both the Sur1^−/− and the Kir6.2^−/− mice feature blunted insulin secretion response to sulfanylureas. The defective insulin response to the sulfanylurea glyburide seems even more pronounced in the β-tg mice as, in these animals, <30% decrease in Sur1 and Kir6.2 gene expression blocks glyburide action. The simultaneous impairment in Sur1 and Kir6.2 gene function may result in an additive effect on glyburide action in the β-tg mice. Similar to the β-tg, fasted Sur1^−/− mice are more hyperinsulinemic than control animals due to persistent activation of voltage-gated calcium channels (33,34). These mice are also significantly more hypoglycemic than controls when fasted. Furthermore, both the Sur1^−/− and the Kir6.2^−/− mice feature impaired glucose tolerance and glucose-induced first- and second-phase insulin secretion. Thus, at least in part, the loss of potassium channels caused by β-cell overexpression of PED/PEA-15 may account for the abnormalities in basal and glucose-stimulated insulin secretion and in glucose tolerance observed in the β-tg mice.

Recent evidence in skeletal muscle and fat indicated that high cellular levels of PED/PEA-15 impair the function of the atypical PKC isoform PKC-ζ through a phospholipase d–mediated mechanism (13,17). Interestingly, the activity of PKC-ζ is also decreased in islets from the β-tg transgenics compared with those from control mice. PKC-ζ overexpression in β-cells transfected with a ped/pea-15 cDNA simultaneously rescues glucose-induced insulin secretion and the expression of the Sur1 and Kir6.2 genes. These data evidence that PKC-ζ activation is sufficient to upregulate these genes, while not proving that the loss of PKC-ζ accompanying PED/PEA-15 overexpression causes Sur1 and Kir6.2 downregulation. Another atypical PKC isoform, PKC-λ, has recently been shown to play a major role in regulating glucose-induced insulin secretion by modulating the expression of a number of pancreatic β-cell genes (35). However, the latter report did not exclude that, as we now show, the activation of PKC-ζ may also upregulate Sur1 and Kir6.2 gene function. Indeed, while clearly distinct and involved in a number of specialized mechanisms, PKC-ζ and -λ also feature structural similarities and many common functions (36). Downregulation of PKC-ζ and/or -λ may cause impaired glucose-induced insulin response in β-tg mouse islets by reducing Sur1 and Kir6.2 gene expression.

Atypical PKCs have been reported to control the function of a number of transcription factors (37,38). Very recent data by Hashimoto et al. (35) evidenced that in PKC-λ–null mice, the abundance of pancreatic β-cells, the expression of the Foxa2 mRNA in addition to that of Sur1 and Kir6.2, is significantly reduced. In addition, in pancreatic β-cells, the expression of the Sur1 and Kir 6.2 genes is dependent on Foxa2 activity (25,39,40). Similar to the PKCλ^−/− mice, we now report that Foxa2 mRNA abundance is also reduced in islets from β-tg transgenics. Thus, the impaired activity of PKC-ζ in β-cells overexpressing PED/PEA-15 may also prevent normal potassium channel generation and glucose-regulated secretion of insulin by depressing Foxa2 expression. Supporting this conclusion, we evidenced that forcing the expression of PKC-ζ in PED/PEA-15–overexpressing β-cells rescued Foxa2 levels, in addition to those of Kir6.2 and SUR1, and glucose-induced insulin secretion. Importantly, islets from ped/pea-15–null mice feature upregulated PKC-ζ function. The enhanced PKC-ζ activity is accompanied by specific increases in the abundance of Kir6.2, Sur1, and Foxa2 mRNAs and augmented insulin secretion in response to glucose. Thus, Foxa2-mediated control of glucose sensitivity represents a physiological function of PED/PEA-15 in the β-cell. Dysregulation of this mechanism may be relevant for type 2 diabetes in humans as well.

This work was supported, in part, by the European Community’s FP6 EUGENE2 (LSHM-CT-2004-512013) grants from the European Foundation for the Study of Diabetes to F.B., the Associazione Italiana per la Ricerca sul Cancro (AIRC) to F.B. and P.F., and the Ministero dell’Università e della Ricerca Scientifica (PRIN to F.B. and P.F and FIRB RBNE0155LB to F.B.) and a grant from Association pour la Récherche contre le Cancer (no. 3500 to H.C.). The financial support of Telethon–Italy is gratefully acknowledged.

The technical help of Salvatore Sequino is also acknowledged.