Deletion of Pten in Pancreatic β-Cells Protects Against Deficient β-Cell Mass and Function in Mouse Models of Type 2 Diabetes

Pten^fl/fl mice (exons 4 and 5 of Pten flanked by loxP sites by homologous recombination) were mated with RIPcre mice (Cre transgene under the control of the rat insulin two promoter TgN[ins2-cre]25Mgn from Jackson Laboratories). RIPcre⁺ Pten^+/fl mice were intercrossed to generate RIPcre⁺ Pten^+/+, RIPcre⁺ Pten^+/fl, and RIPcre⁺ Pten^fl/fl mice. RIPcre⁺ Pten^+/+ mice were used as controls. RIPcre⁺ Pten^fl/fl Lepr^db/db mice were generated by breeding RIPcre⁺ Pten^+/fl mice with Lepr^+/db mice (Jackson Laboratories) to obtain RIPcre⁺ Pten^+/fl Lepr^+/db mice. RIPcre⁺ Pten^+/fl Lepr^+/db mice were then intercrossed to generate RIPcre⁺ Pten^+/+ Lepr^+/+ (wild-type), RIPcre⁺ Pten^+/+ Lepr^db/db (db/db), and RIPcre⁺ Pten^fl/fl Lepr^db/db (−/−; db/db) mice. Only male mice were used for experiments and only littermates were used as controls. Genotypes for Cre and Pten gene were determined with PCR using ear clip DNA as described previously (19,23). All mice were maintained on a mixed 129J-C57BL/6 background and housed in a pathogen-free facility on a 12-h light-dark cycle and fed ad libitum with standard irradiated rodent chow (5% fat; Harlan Tecklad, Indianapolis, IN) in accordance with the Ontario Cancer Institute Animal Care Facility protocol, with no restrictions on the animals' activities.

HFD (Dyets lard Surmwit mouse diet DYET# 182084; Fyets Inc., Bethlehem, PA) for RIPcre⁺ Pten^fl/fl and RIPcre⁺ Pten^+/+ mice was started for mice at 2 months of age and continued for 7 months.

All overnight fasts were carried out between 5:00 p.m. and 10:00 a.m. All blood glucose, glucose tolerance tests, insulin tolerance tests, and glucose-stimulated insulin secretion tests were performed on overnight-fasted animals as previously described (23).

Detailed perifusion protocol was described in a previous publication (29). In brief, 60 islets were placed in a perifusion chamber at 37°C with a capacity of 1.3 ml, and perifused with Kreb-Ringer bicarbonate buffer at 1 ml/min. Islets were equilibrated with Kreb-Ringer bicarbonate HEPES buffer (2.8 mmol/l glucose) for 30 min. They were then stimulated with 2.8 mmol/l glucose for 10 min, followed by a 40-min incubation with 16.7 mmol/l glucose. Fractional insulin content was determined with radioimmunoassay kit (Linco Research, St. Louis, MO). At the end of each perifusion, islets were collected and lysed with acid ethanol for assessment of insulin content. First phase secretion was 10 to 25 min, and second phase secretion was 25 to 50 min. Results were presented as insulin secreted normalized to 60 islets and to total insulin content.

Pancreas was fixed in 4% paraformaldehyde in 0.1M PBS (pH 7.4) as previously described (25). For immunohistochemical and immunofluorescent staining antibodies against PTEN (NeoMarker, Fremont, CA), Akt (Cell Signaling Technology, Beverly, MA), p-Akt (Ser473) (Cell Signaling Technology, Beverly, MA), insulin (DAKO), glucagon (NovoCastra Laboratories), laminin (Sigma), β-catenin (BD Transduction Laboratories), GLUT2 (Chemicon, Temecula, CA), PDX-1 (Chemicon, Temecula, CA), and Ki67 (DAKO) were used. Total islet area and total pancreas area were determined on insulin immunohistochemically stained sections with Image-Pro Plus software (Media Cybernetics, Silver Spring, MD) and expressed as total islet area per total pancreas area. Proportions of different islet sizes were measured on H&E-stained pancreas sections: small (<10 cells), medium (10–200 cells), and large (>200 cells). β-Cell size was determined with insulin/DAPI costained pancreas sections. Insulin-positive area was determined with Image-Pro Plus software (Media Cybernetics, Silver Spring, MD) and divided by the number of DAPI-positive nuclei within the insulin-stained areas.

Islets and hypothalami were isolated and protein lysates were obtained as previously described (23). Antibodies against actin (Santa Cruz Biotechnology, Santa Cruz, CA), Akt, p-Akt, FoxO-1 (Cell Signaling Technology, Beverly, MA), p-FoxO-1 (Cell Signaling Technology, Beverly, MA), GAPDH (Cell Signaling Technology, Beverly, MA), GLUT-2, mTOR (Santa Cruz Biotechnology, Santa Cruz, CA), p-mTOR (Cell Signaling Technology, Beverly, MA), p-p53 (R&D systems), p53 (Santa Cruz Biotechnology, Santa Cruz, CA), PDX-1, and PTEN were used (antibody sources for Akt, p-Akt, GLUT-2, PDX-1, and PTEN refer to immunohistochemistry and immunofluorescence).

Isolated islets are incubated in RPMI-1640 (10% FBS) at 37°C overnight, then irradiated with 30Gy of gamma irradiation the next day. Islets were harvested after overnight in culture. RNA was extracted with RNeasy Plus Mini Kit (Qiagen, Valencia, CA) according to protocol provided. Complimentary DNA was synthesized according to protocol published elsewhere (30). PCR was monitored in real time using the ABI Prism 7900HT Real-Time PCR system (Applied Biosysterm). Experiments were performed in triplicate for each sample. Primer sequences for Mdm2, Bax, and p21 are available upon request.

Data are presented as mean ± SEM and were analyzed by one-sample t test, independent-sample t test, and one-way ANOVA with the post hoc Tukey least significant difference test, when appropriate. All data were analyzed using the statistical software package SPSS (version 16.0) for Macintosh.

PTEN transcript and protein levels were measured in islets of mice with type 2 diabetes. In both HFD-induced and db/db mice, PTEN transcripts in islets were significantly increased as accessed by quantitative real-time PCR (Fig. 1,A). PTEN protein expression was also increased as demonstrated by immunohistochemistry of pancreatic sections and Western blotting of isolated islet lysates (Fig. 1,B and C). The increase in islet PTEN expression in both of these type 2 diabetes models was accompanied by attenuated PI3K signaling, as demonstrated by the reduction of p-Akt, p-mTOR, and p-FoxO-1 expression (Fig. 1 C).

We have previously shown that RIPcre⁺ Pten^fl/fl mice exhibit an increase in β-cell mass and function under basal conditions (23). To investigate whether these positive attributes of PTEN deletion in pancreatic β-cells conferred protection against type 2 diabetes, we fed these mice a prolonged HFD for 7 months. Efficient PTEN deletion in β-cells, along with partial deletion in the hypothalamus, persisted in the RIPcre⁺ Pten^fl/fl mice on prolonged HFD (Fig. 2,A and B). Despite their increased weight gain upon HFD feeding, RIPcre⁺ Pten^fl/fl mice remained remarkably euglycemic throughout the duration of prolonged HFD, in contrast to the gradual increase in blood glucose levels in control littermates (Fig. 3,A and B). They also exhibited improved glucose tolerance (Fig. 3,C). The attenuation of insulin secretion in response to glucose is a characteristic β-cell defect in type 2 diabetes (1,–3). Indeed, this attenuation was observed in both the first and second phases of insulin secretion after in vivo glucose challenge in wild-type mice after a prolonged HFD. In contrast, insulin secretion in response to glucose was preserved in RIPcre⁺ Pten^fl/fl mice (Fig. 3 D).

During HFD-induced diabetes, development of peripheral insulin resistance leads to a compensatory increase in β-cell mass to meet the increasing demands for insulin. This compensatory proliferation was observed in RIPcre⁺ Pten^+/+ islets on HFD (Fig. 4,A and B). RIPcre⁺ Pten^fl/fl islets showed an already increased β-cell mass under chow-fed conditions, and we observed no further increase in β-cell mass in the mice on HFD. The increased β-cell mass was due to both an increase in proliferation and β-cell size in RIPcre⁺ Pten^fl/fl mice under both chow and HFD conditions, which likely reflects the direct effects of PTEN deletion in their β-cells (Fig. 4,C, E, and F). Furthermore, age-matched chow- and HFD-fed RIPcre⁺ Pten^fl/fl mice showed similarly increased proportion of large islets (Fig. 4,D). Despite the increased proliferation and cellular growth in RIPcre⁺ Pten^fl/fl islets, their architectures were maintained (Fig. 5 A).

In order to assess the direct effects of PTEN deletion specifically in β-cells, we measured insulin secretion during perifusion assay on isolated islets ex vivo. Under basal condition, both RIPcre⁺ Pten^+/+ and RIPcre⁺ Pten^fl/fl islets demonstrated similar insulin release after glucose stimulation (Fig. 4,G and H). However, after HFD feeding, RIPcre⁺ Pten^fl/fl mice maintained robust insulin secretion in response to glucose in contrast to islets of HFD-fed RIPcre⁺ Pten^+/+ mice, which showed attenuation in both phases of insulin secretion (Fig. 4 I and J). These data suggest that diabetes protection in HFD-fed RIPcre⁺ Pten^fl/fl mice are likely due to both preserved β-cell function and mass.

To determine the mechanisms underlying the enhanced β-cell growth and maintained function after HFD feeding in islets of RIPcre⁺ Pten^fl/fl mice, we assessed for alterations in the PI3K signaling pathway. HFD has been shown to attenuate phosphorylation of Akt in β-cells (7). However this decline was not observed in PTEN-deficient islets, which indicates persistent activation of the PI3K pathway even after prolonged exposure to HFD (Fig. 5,A and B). FoxO-1 and mTOR, both downstream effectors of PI3K signaling, showed increased phosphorylation in HFD-fed RIPcre⁺ Pten^fl/fl islets (Fig. 5,B). The increased p-mTOR was in keeping with the increased β-cell size in RIPcre⁺ Pten^fl/fl mice (Fig. 4,E and F). The expression of GLUT-2, a marker of β-cell differentiation and a rate-limiting component in glucose sensing, has also been shown to decline with the progression of type 2 diabetes (31,32). GLUT2 remained high in HFD-fed RIPcre⁺ Pten^fl/fl islets to a similar degree as the islets of chow-fed RIPcre⁺ Pten^+/+ mice, in contrast with the decline of GLUT2 in islets of HFD-fed RIPcre⁺ Pten^+/+ mice (Fig. 5,A and B). Furthermore, PDX-1, a downstream transcriptional target of insulin signaling important in β-cell differentiation and growth (33), was also increased in RIPcre⁺ Pten^fl/fl islets, which may also contribute to the preserved β-cell mass and function in HFD-fed RIPcre⁺ Pten^fl/fl mice (Fig. 5 A and B). However, RIPcre⁺ Pten^fl/fl mice also exhibited increased peripheral insulin sensitivity because of partial neuronal deletion of PTEN (L.W. and M.W., unpublished data). To circumvent the potential confounding effects of enhanced insulin sensitivity, we next examined RIPcre⁺ Pten^fl/fl mice in a db/db model of type 2 diabetes which did develop insulin resistance as described below.

Both RIPcre⁺ Pten^fl/fl Lepr^db/db mice and RIPcre⁺ Pten^+/+ Lepr^db/db littermates exhibited similar degrees of weight gain and insulin resistance (Fig. 6,A and D). However, despite severe insulin resistance in RIPcre⁺ Pten^fl/fl Lepr^db/db mice, they continued to remain remarkably euglycemic and showed normal glucose tolerance (Fig. 6,B and C). Furthermore, in vivo GSIS experiments showed robust insulin secretion in response to glucose stimulation in RIPcre⁺ Pten^fl/fl Lepr^db/db mice (Fig. 6,E). Interestingly, islets of RIPcre⁺ Pten^fl/fl Lepr^db/db mice demonstrated similar degrees of hypertrophy and proliferation as those of RIPcre⁺ Pten^+/+ Lepr^db/db islets and did not show a further increase in β-cell mass compared with db/db controls (Fig. 7,B and C). Islets of RIPcre⁺ Pten^fl/fl Lepr^db/db mice showed no signs of disorganized architecture (Fig. 7 A).

To assess the direct effects of PTEN deletion on β-cell function in the absence of leptin signaling, we examined islet function ex vivo by perifusion. RIPcre⁺ Pten^+/+ Lepr^db/db islets showed diminished insulin secretion, whereas robust first phase insulin secretion was observed in RIPcre⁺ Pten^fl/fl Lepr^db/db islets after glucose stimulation, further confirming the preserved glucose-responsive insulin secretory function in PTEN-deficient β-cells in db/db mice (Fig. 6 F and G). Thus, in contrast to the HFD-fed RIPcre⁺ Pten^fl/fl mice, where the combination of increased β-cell mass and function likely contributed to the protective effect against diabetes, increased β-cell function and not mass in RIPcre⁺ Pten^fl/fl Lepr^db/db mice was likely responsible for their diabetes protection.

We next determined the effects of PTEN deletion on PI3K signaling in β-cells in the db/db model. RIPcre⁺ Pten^+/+ Lepr^db/db islets showed reduced levels of p-Akt, whereas PTEN deletion completely rescued this defect (Fig. 7,A and D). Phosphorylation of mTOR and FoxO-1 also remained high in RIPcre⁺ Pten^fl/fl Lepr^db/db islets, consistent with constitutively active PI3K cascade (Fig. 7,D). In keeping with its normal insulin secretion in response to glucose, GLUT2 was highly expressed in islets of RIPcre⁺ Pten^fl/fl Lepr^db/db mice in contrast to the near-complete loss of GLUT2 expression in islets of Lepr^db/db mice (Fig. 7,A and D). PDX-1 was also downregulated in RIPcre⁺ Pten^+/+ Lepr^db/db islets, whereas the RIPcre⁺ Pten^fl/fl Lepr^db/db islets continued to show intense expression of PDX-1 and nuclear localization of the protein to a similar degree as the wild-type islets on chow diet (Fig. 7 A and D).

Given PTEN′s role as a tumor suppressor in various cancer-prone tissues (16,–18), we assessed for changes suggestive of cellular transformation and response to DNA-damaging stimuli such as gamma irradiation. Despite the significant increase in islet mass in both HFD-fed RIPcre⁺ Pten^fl/fl and RIPcre⁺ Pten^fl/fl Lepr^db/db islets, intact islet integrity was observed even in the 9-month-old aged mice fed either HFD or on db/db background. Intact laminin staining in the basement membrane as well as localization of β-catenin to the plasma membrane both demonstrated absence of deregulated growth (Fig. 8,A). Furthermore, induction of DNA damage with gamma irradiation showed a similar degree of DNA-repair and apoptosis response in both RIPcre⁺ Pten^+/+ and RIPcre⁺ Pten^fl/fl islets, including phosphorylation of p53 and cleavage of MDM2 (Fig. 8,C). The transcript levels of p53 target genes, including Bax, Mdm2, and p21 also showed similar changes in response to gamma irradiation between RIPcre⁺ Pten^+/+ and RIPcre⁺ Pten^fl/fl islets (Fig. 8,B). Interestingly RIPcre⁺ Pten^fl/fl islets exhibited significantly reduced expression of proapoptotic gene Bax under basal condition, which is consistent with their constitutively active PI3K signaling (Fig. 8 B). Thus, PTEN deletion in β-cells protects against β-cell dysfunction in both HFD and db/db models of type 2 diabetes with no finding suggestive of deregulated growth. Furthermore, PTEN-deleted β-cells still maintained intact response to DNA-damaging stimuli, which further suggests that they are not more prone to tumor formation.

In pre-diabetic individuals, insulin resistance in the classic metabolic tissues including the liver, muscle, and fat is present; however, glucose homeostasis can be maintained as long as pancreatic β-cells are able to increase insulin production to compensate for increased insulin demands (7,34,35). Accumulating evidence shows that a defect in insulin-PI3K signaling in pancreatic β-cells may contribute to the development of β-cell dysfunction, leading to the onset of type 2 diabetes. Indeed, mice with genetic ablation of insulin/IGF-1 receptors specifically in the β-cells demonstrate reduced β-cell function (10,12,13). Therefore, a new paradigm of type 2 diabetes pathogenesis suggests that diminished insulin responsiveness specifically in the pancreatic β-cells plays a central role in disease development.

Interestingly, PTEN, a critical negative regulator of the PI3K signaling pathway, has been shown to be upregulated in β-cells that have complete absence of insulin/IGF-1 signaling (25), leading to a hypothesis that PTEN may play a causal role in the pathogenesis of β-cell dysfunction. Here we have shown that when type 2 diabetes was induced, either by HFD or global leptin signaling deficiency, increased PTEN expression was observed in islets along with diminished PI3K signaling, suggesting that β-cell PTEN may indeed play a critical role in the development of type 2 diabetes. We and others have already shown that tissue-specific increase in PI3K signaling in pancreatic β-cells resulting from deletion of PTEN leads to increased β-cell mass and function under basal conditions (23,24). However, the biologic outcome from continued deletion of PTEN in β-cells under metabolically stressed conditions was elusive. In this report, we show that after prolonged HFD, RIPcre⁺ Pten^fl/fl mice continued to exhibit increased islet mass and were protected against the loss of glucose-stimulation insulin secretion. The enhanced β-cell function was demonstrated not only in vivo, but also in vitro, which supports the direct effect of PTEN in β-cells. However, RIPcre⁺ Pten^fl/fl mice exhibited enhanced insulin sensitivity, which may have masked the full potential effects of PTEN deletion in pancreatic β-cells against metabolic stress in the HFD model.

In the db/db model, RIPcre⁺ Pten^fl/fl Lepr^db/db mice developed obesity with significantly diminished insulin sensitivity to a similar degree as the RIPcre⁺ Pten^+/+ Lepr^db/db littermate controls (Fig. 6 A and D). Yet, RIPcre⁺ Pten^fl/fl Lepr^db/db mice remained completely glucose tolerant and euglycemic throughout the 7-month period. Islets from RIPcre⁺ Pten^fl/fl Lepr^db/db mice demonstrated an increase in islet area to a similar degree as those of RIPcre⁺ Pten^+/+ Lepr^db/db mice. However RIPcre⁺ Pten^fl/fl Lepr^db/db islets remained responsive to glucose-stimulated insulin secretion. This is likely attributed by the PI3K signaling which remained activated in these islets. As such GLUT-2 and PDX-1, which are important in glucose sensing and cell differentiation, are maintained. These data illustrate the importance of the negative regulation that PTEN exerts on β-cells, whereupon PTEN deletion leads to enhanced PI3K signaling and protection against β-cell dysfunction that occurs in type 2 diabetes.

Given the well known tumorigenic effects of PTEN deletion in some tissues, we assessed for evidence of deregulated growth (16,–18). PTEN deficient islets showed intact architecture and demonstrated intact DNA repair response to gamma irradiation as wild-type counterparts. Furthermore, previously we have shown that PTEN deletion in combination with the activation of the cMyc oncogene is still not able to lead to tumor formation (36). However, more extensive analysis and longitudinal observation are required to conclusively demonstrate the full impact of PTEN deletion in pancreatic β-cells on tumorigenesis. It is worth noting that PTEN-deleted β-cells behave differently than β-cells that express constitutively active Akt. Although the RIP-Akt transgenic mice exhibit increased islet mass, these mice have compromised β-cell function (37,38). PTEN-deficient β-cells, in contrast, do not show loss of function even after a prolonged exposure to HFD, despite the continued proliferation and PI3K signaling within these cells. This important distinction between PTEN-deleted and Akt-overexpressing β-cells illustrates the fundamental difference between removing a physiologic brake, which would allow for PI3K signaling to return to physiologic levels, whereas forced expression of Akt may lead to supraphysiologic levels of PI3K signaling that can lead to cell proliferation and dedifferentiation. Further investigations to delineate other important genes or pathways that distinguish between these two modes of enhancing PI3K signaling may unveil clinically safe therapeutic targets for β-cells.

In summary, our findings show that PTEN plays a critical role in the pathogenesis of β-cell dysfunction in type 2 diabetes and highlight the tissue-specific role of PTEN in attenuating PI3K signaling in β-cells. Our results show that PTEN deletion in β-cells leads to an increase in β-cell mass and function as a result of restoring PI3K signaling to a physiologic level in diabetic condition. Indeed, PTEN deletion leads to complete protection against β-cell defects even in the face of severe insulin resistance, providing genetic evidence of the critical negative regulation that PTEN exerts in impeding β-cell function and growth that occurs in type 2 diabetes. Our results highlight opportunities for therapeutic targeting of PTEN in β-cells for treatment or prevention of type 2 diabetes.

This work was supported by grants to M.W. from the Canadian Institutes of Health Research (CIHR) MOP-201188 and the Canadian Diabetes Association, as well as a CIHR grant MOP-64464 to H.Y.G.

L.W. is supported by a Doctoral Research Award from CIHR.

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

L.W. researched data, contributed to discussion, wrote the manuscript, and reviewed/edited the manuscript. Y.L., S.Y.L., K.-T.T.N., S.A.S., and A.S. researched data. T.W.M. and H.G. contributed to discussion. M.W. contributed to discussion and reviewed/edited the manuscript.