Selective Deletion of Pten in Pancreatic β Cells Leads to Increased Islet Mass and Resistance to STZ-Induced Diabetes^†

Animals.

Targeted deletion of Pten in β cells was achieved by crossing Pten^loxP/loxP mice (C57/129/J background) (30) with rat insulin promoter-Cre (RIP-Cre) mice (C57 background) (43). F₁ generation compound heterozygous animals were backcrossed with Pten^loxP/loxP mice to produce F₂ generation experimental animals. Animals were genotyped by standard genomic PCR techniques (30). Animals were housed in a temperature-, humidity-, and light-controlled room (12-h light/dark cycle), allowing free access to food and water. All experiments were conducted according to the research guidelines of the UCLA Chancellor's Animal Research Committee.

Glucose, insulin, and lipid determinations.

Blood samples were collected from 3-month-old mice. After mice fasted overnight, blood samples were collected by orbital eye bleeding into lithium chloride-coated plasma collection tubes. Plasma was obtained by centrifugation. Plasma insulin levels were measured by enzyme-linked immunosorbent assay according to the manufacturer's protocols (Alpco, Windham, N.H.). Plasma samples were also analyzed for nonesterified fatty acid (Wako) and triglyceride (Thermo DNA) using the manufactured kits. Glucose levels were determined with a Freestyle glucometer (Therasense) with blood samples from tail vein punctures in mice that fasted for 16 h.

Glucose and insulin tolerance tests.

After overnight fasting, mice 3 months of age were injected intraperitoneally (i.p.) with 30% d-glucose (2 g/kg of body weight; Sigma, St. Louis, Mo.) or insulin (1 U/kg; Lilly) (50). Blood glucose concentrations were measured at indicated time points as previously described (50).

Western electrophoresis.

Protein lysates were collected from isolated control and mutant islets. Western blot analyses were performed with PTEN, phosphor-S6, cyclin D, and p27 antibodies. All membranes were also probed with antibodies for β-actin as a loading control.

Immunohistochemistry.

We performed immunohistochemistry on Zn-formalin-fixed, paraffin-embedded sections, following antigen retrieval. Antibodies for PTEN (26H9) and phospho-AKT (Ser473) antibodies were obtained from Cell Signaling Technology, Beverly, Mass.). Glucose transporter 2 (GLUT2) antibody was purchased from Calbiochem, San Diego, Calif. Antibodies for insulin, glucagons, somatostatin, and pancreatic polypeptides were provided by Zymed. Sections were counterstained with hematoxylin.

Stereology measurement of islet area.

Insulin-stained pancreatic sections were analyzed using a MicroBrightField Stereology-assisted microscope mounted on a remote controlled platform. Islet area and number were determined with five cell clusters considered to be islets. Three slides per pancreas, 120 μm apart, were counted on a total of five animals that were 3 months old.

Determination of cell proliferation and apoptosis.

We evaluated cell proliferation on embryonic day 17.5 (E17.5) mice with bromodeoxyuridine (BrdU) pulse-labeling. Pregnant mothers at 17.5 days after plugging were injected with BrdU (100 μg/g of body weight) and sacrificed 45 min later. Embryos were retrieved, decapitated, and fixed in Zn-formalin. For adult animals, BrdU was given i.p. 30 to 45 min before euthanization. BrdU staining was done using a kit from Roche with a modified protocol for immunofluorescence. Sections were costained with insulin. Apoptosis was determined with P17- and STZ-treated mice with a terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) kit from Roche with a modified protocol for immunofluorescence. The same sections were stained with insulin.

Islet culture and glucose stimulation.

Islets were isolated from 3-month-old male animals by collagenase digestion of the pancreas, followed by purification using a Ficoll gradient. Islets were handpicked twice and cultured in RPMI complete medium before stimulation with glucose. After overnight culture, islets (50) were stimulated with 3 mM (low) and 30 mM (high) glucose in RPMI 1640 for 30 min. The amount of insulin released to the culture medium was measured with an enzyme-linked immunosorbent assay kit from ALPCO. The islets were collected and lysed for DNA analysis. For STZ treatment, cultured islets were treated with 2 mM STZ in the culture medium for 30 min. Islets were then allowed to recover overnight. Insulin production in response to 30 mM glucose was then assayed.

Fluorescence-activated cell sorter (FACS) analysis for cell size.

Islet cells were dissociated with trypsin and passed through a 70 μm cell strainer (BD Falcon, Bedford, MA) to obtain a single-cell suspension. The trypsinized cells were run on a BD FACSCalibur (BD Immunocytometry Systems, San Jose, CA) to measure the cell size, based on forward scatter signal.

STZ-induced diabetes.

Eight-week-old male mice were injected i.p. with multiple subdiabetogenic doses of streptozotocin (31, 38) at a dose of 50 mg of streptozotocin/kg of body weight daily for 5 consecutive days (Sigma, St. Louis, Mo.) to produce a β-cell injury model. On day 8, a group of three animals were sacrificed following BrdU pulse-labeling. Pancreas sections were costained with antibodies against BrdU and insulin or for TUNEL and insulin. For long-term effect, the remaining six animals were evaluated weekly for the development of diabetes (defined as persistent random blood glucose levels of >300 mg/dl). Animals were sacrificed 2 months later, following BrdU pulse-labeling. For the high-dose STZ experiment, 200-mg/kg STZ was given as one dose, and animals were sacrificed 36 h later for tissue collection.

Statistical analysis.

All data are presented as means ± the standard error of the mean. Statistical calculations were performed with Microsoft Excel analysis tools. Differences between individual groups were analyzed by Student's t test, with two-tailed P values of <0.05 considered statistically significant.

β-cell-specific Pten deletion.

To establish a mouse model carrying Pten deletion in the insulin-producing cells, we crossed Pten conditional knockout mice (Pten^loxP/loxP) (30) with the β-cell-specific Cre transgenic line driven by the rat insulin promoter (43). Of the offsprings generated, we used Pten^loxP/loxP;Cre⁺ mice as mutants and Pten^loxP/loxP; Cre⁻ mice as controls for the experiment. PCR analysis of genomic DNA prepared from Pten^loxp/loxp; RIP-Cre⁺ (mutant) mice demonstrated that Cre-mediated Pten deletion was pancreatic specific, with little leakage in the brain (Fig. 1A), consistent with the previous report (17). Immunofluorescent analysis demonstrates that PTEN was localized in the ducts (Fig. 1B, top left, inset) and the nucleus of the murine islet cells (Fig. 1B, top left), similar to the subcellular localization observed in human islets (42). In the mutant pancreatic islets, PTEN was present in the ducts (Fig, 1B, bottom left, inset) but lost in the majority of the β cells, accompanied by increased phospho-AKT staining (Fig. 1B, bottom left). Loss of Pten and activation of AKT also led to the activation of S6K, as indicated by the increased phosphor-S6 protein levels in the mutant pancreatic islet protein lysate (Fig. 1B, top right). We further evaluated the expression of two cell cycle regulators, p27 and cyclin D; both are potential targets for regulation by the PI3K pathway. We found that the cell cycle inhibitor p27 was diminished significantly as a result of Pten deletion in the pancreas, whereas the level of cyclin D was minimally altered (Fig. 1B, bottom right).

Pten deletion leads to an increase in islet numbers and total islet mass.

We examined the impact of Pten deletion on pancreatic islets and β cells. Development of the endocrine pancreas with insulin-expressing cells started as early as E12.5. Just prior to birth, endocrine cells migrate and aggregate to form mature islets with localized insulin secretion. At this stage, more islet clusters were seen in Pten mutant pancreas than in the control pancreas (Fig. 2A, left). Compared to islet mass in control animals, islet mass in postnatal animals was significantly increased in mutant mice as early as postnatal day 15 (P15) and persisted throughout adulthood (Fig. 2A, middle and right). This increase was likely due to increases in both islet number and islet size (Fig. 2B). We quantified islet number and islet size in 3-month-old male mice and demonstrated a 4.5-fold increase in total islet area (Fig. 2C, left). The increase in islet number was true for both small and large islets. Islets larger than 5,000 μM² were only observed in the mutant pancreas but not in the control pancreas (Fig. 2C, right). The increase in islet area was also observed when islets were isolated (Fig. 2B, right) and scored. In this case, when islets from four mutant or control animals were pooled and counted, there was a 1.5-fold increase in the total number of islets recovered (200 versus 300 islets/sample pooled from four animals) and a 2-fold increase in islet size. The phenotype analysis of the mutant animals suggested that PTEN is a crucial player in regulating islet mass. This may happen during embryonic development when neogenesis dominates the islet mass, as well as during adulthood when most β cells come from self-regeneration.

Pten deletion leads to increased cell proliferation and decreased cell death under physiological conditions.

During pancreatic development, cell proliferation peaks around late embryonic development and early postnatal life, followed by a wave of apoptosis at postnatal 2 to 3 weeks during pancreatic remodeling (16). Cell proliferation and cell death are limited in the adult endocrine pancreas (16). Deletion of Pten led to a significant increase in β-cell proliferation during embryogenesis, as measured by BrdU pulse-labeling at E17.5 (Fig. 3A). More than 8% of insulin-positive cells were BrdU⁺ in the mutant animals, compared to 3% double-positive cells in the controls. There was also significant proliferative activity adjacent to the insulin-positive cells in the mutant pancreas (Fig. 3A, arrows), which may represent increased progenitor cell activity. Deletion of Pten also led to a significant decrease in apoptotic rate during pancreas remodeling at an early postnatal stage. A twofold decrease in TUNEL-positive cells was observed in the mutant pancreas at P17 (Fig. 3B). Taken together, these data demonstrate that PTEN plays a critical role in regulating β-cell growth and survival.

Pten-null pancreas retains normal islet functions.

To determine whether Pten-null β cells are fully differentiated and functional, we first stained islets with anti-insulin antibody. Immunoreactivity to insulin was detected as early as E12.5 in both the control and mutant pancreases (data not shown). By E15.5, all pancreases analyzed contain insulin-positive cells (Fig. 4A, left). In the postnatal stage when endocrine cells form clustered islets, mutant islets showed immunoreactivity to insulin, although with less intensity than in the control islets in both P15 and 3-month-old mice (Fig. 4A). No significant changes in localization of the non-insulin-producing cells were observed (Fig. 4B), while the number of insulin-producing cells appeared to be increased in mutant islets (Fig. 4B). Insulin-producing cells were localized in the center of the islet, while non-insulin-producing cells were predominantly distributed near the periphery of the islets in both control and mutant pancreas. In some islets, non-insulin-producing cells were disbursed in the islets. This was observed for both control and mutant animals and was unlikely to be due to Pten deletion. The ability of Pten-null β cells to secrete insulin in response to glucose stimulation was further quantified using isolated islets. Pten-null islets were responsive to glucose stimulation and secreted insulin in a manner similar to islets from control animals (Fig. 4C, left). Consistent with this observation, no statistically significant changes in serum insulin levels could be observed in the mutant animals, compared to the controls (Fig. 4C, right).

The mutant islets also displayed normal staining patterns of GLUT2 (Fig. 4D, left). With high-power microscopy (magnification, ×100), Pten-null β cells did appear slightly enlarged, compared to controls (Fig. 4D, left). To determine whether Pten deletion leads to changes in β-cell size, we subjected dissociated islet cells to FACS analysis. Forward scatter signal showed no significant difference in cell size between the control and mutant β cells (Fig. 4D, right).

Pten deletion in pancreatic β cells did not significantly alter the glucose homeostasis of mutant animals.

Similar to our previous studies with hepatocyte- and adipose tissue-specific Pten deletion (28, 50), PTEN loss in β cells led to hypoglycemia (Fig. 5A). This lower serum glucose level may be one explanation for the decreased insulin-staining density of the mutant islets. Nevertheless, mutant and control animals responded similarly to a glucose challenge (Fig. 5B). An insulin tolerance test was also performed (Fig. 5C). Although insulin injection led to a similar fold decrease in plasma glucose levels in control and mutant animals, most of the mutant animals could not complete the test, due to their significantly lower fasting glucose levels, and had to be rescued with glucose injection. No significant changes in serum lipid indexes in control and mutant animals were observed (Fig. 5D).

Pten deletion protects animals from STZ-induced β-cell injury.

To determine the pathophysiological significance of Pten deletion in insulin-producing cells, we subjected control and mutant animals to low-dose STZ treatment. One week after initial STZ treatment, islets from the control pancreas started to atrophy, a sign of injury, while no obvious phenotype was observed on mutant islets (Fig. 6C, D, I, and J). Two months following initial STZ treatment, no visible islets could be found in the control pancreas under low-power microscopy, while isolated, irregular, atrophied islets could be found under high-power magnification (Fig. 6E and K). In contrast, a significant number of healthy islets could be found in the mutant pancreas (Fig. 6F and L). These morphologically healthy islets remained insulin positive (Fig. 6P and R). Lymphocyte infiltration was clearly visible in both control and mutant pancreases (Fig. 6E, F, I, J, K, and M). Two months after STZ treatment, small β-cell clusters with little or no lymphocyte infiltration could be observed in the mutant but not control pancreas (Fig. 6L, inset). These clusters may be islets newly formed after injury. Immunostaining to insulin was diminished in all animals treated with STZ but appeared to be retained more in the mutants (Fig. 6R).

Mutant pancreas exhibited increased BrdU-positive cells (5%) in comparison to control pancreas (1.5%) 7 days after STZ treatment (Fig. 7A). In control animals, BrdU-positive cells had weaker nuclear PTEN staining than the BrdU-negative cells (Fig. 7B, top, inset). Similar to the proliferative activity seen during the pancreatic development (Fig. 3), significantly increased proliferative activity peripheral to the islets was also observed in the mutant pancreas (Fig. 7A, bottom left). The Majority of these BrdU-positive/insulin-negative-staining cells stained negative for PTEN (Fig. 7B, bottom, inset). Cell death could be detected in the control pancreas in response to STZ treatment, while apoptotic cells were rare in mutant pancreas (Fig. 7C).

As a result of β-cell damage, blood glucose levels increased significantly in the control animals from approximately 120 mg/dl to 400 mg/dl, while the blood glucose levels of mutant animals did not change significantly (from 106 mg/dl to 120 mg/dl) (Fig. 7D, left). Three weeks after STZ treatment, 80% of the control animals developed hyperglycemia (random glucose level, >300 mg/dl). No mutant animals (n = 6) developed hyperglycemia during the entire experimental period of 2 months (Fig. 7D, right). Furthermore, the glucose levels of mutant animals remained at 120 mg/dl for the remainder of the 2 months of the experimental period. This normal glucose level in Pten mutant mice was consistent with the healthy islet morphology observed in the mutant pancreas and suggests that PTEN loss could efficiently protect β cells from STZ-induced injury and maintain glucose homeostasis.

Pten deletion in the pancreas protected islets from STZ-induced oxidative stress.

To dissect the mechanism of Pten deletion-induced cytoprotection of β cells, we subjected isolated control and mutant islets to STZ treatment in vitro to avoid the complication of in vivo immune response. Immediately after a 30-min treatment with STZ (2 mM), both control and mutant islets lost responsiveness to glucose stimulation (data not shown). After overnight recovery was permitted, control islets were mostly destroyed, while mutant islets still retained their islet morphology (Fig. 8A), suggesting that PTEN loss protects islets from oxidative stress induced by STZ. To further test whether Pten deletion will protect islets from high-dose STZ (200 mg/kg) treatment in vivo, we examined the mutant pancreas 36 h after STZ treatment, when control animals started to show significant damage (Fig. 8B, top). Deletion of Pten did not protect the animals from the high-dose STZ treatment, as indicated by TUNEL-positive staining (Fig. 8B, bottom). Both control and mutant islets had diminished insulin staining after STZ treatment (data not shown). Thus, although Pten deletion can effectively protect the islet from low-dose STZ treatment, it cannot prevent cell death caused by higher doses of oxidative stress.