PTPN2, a Candidate Gene for Type 1 Diabetes, Modulates Pancreatic β-Cell Apoptosis via Regulation of the BH3-Only Protein Bim

Rat islets were isolated by collagenase digestion and handpicked under a stereomicroscope, and β-cells were fluorescence-activated cell sorting (FACS)–purified and cultured as previously described (11,19). The β-cell preparations used in this study contained 96 ± 2% β-cells (n = 12). For cytokine treatment, cells were cultured in the medium without serum. For small interfering (si)RNA transfection, BSA- and antibiotic-free medium was used.

Human islets were isolated from four nondiabetic organ donors (age 70 ± 3 years; BMI 25.7 ± 0.3 kg/m²) in Pisa, Italy, with the approval of the local ethics committee. Islets were isolated by enzymatic digestion and density-gradient purification (21), and cultured in M199 medium containing 5.5 mmol/L glucose. The human islets were shipped to Brussels within 1–5 days of isolation. After overnight recovery in Ham’s F-10 containing 6.1 mmol/L glucose, 10% FBS, 2 mmol/L GlutaMAX, 50 μmol/L 3-isobutyl-1-methylxanthine, 1% BSA, 50 units/mL penicillin, and 50 μg/mL streptomycin, islets were dispersed for treatment and viability assays as previously described (19). The percentage of β-cells, examined in the four dispersed islet preparations by staining with anti-insulin antibody (1:1,000; Sigma-Aldrich, Bornem, Belgium) and donkey anti-mouse IgG rhodamine (1:200; Lucron Bioproducts, De Pinte, Belgium), was 71 ± 2%.

The INS-1E cell line (a gift from Dr. C. Wollheim, Centre Medical Universitaire, Geneva, Switzerland) was cultured in Roswell Park Memorial Institute 1640 GlutaMAX-I (22).

The siRNAs used in the study are described in Supplementary Table 1. The optimal concentration of siRNA used for cell transfection (30 nM) was established previously (19). Cells were transfected using the Lipofectamine 2000 lipid reagent (Invitrogen, Carlsbad, CA) as previously described (23). For dose-response experiments, the following concentrations of the siRNA targeting PTPN2 gene were used: 1, 5, 10, and 30 nM. After transfection, cells were cultured for a 48-h recovery period and subsequently exposed to cytokines. All siRNAs used in the current study have been validated for use in rodent and human cells, leading to 60–80% inhibition of the target genes (7,19,24) (J. Barthson, E.N.G., D.L.E., unpublished data).

The following cytokine concentrations were used on the basis of dose-response experiments previously performed by our group (11) (I.S., F.M., D.L.E. unpublished data): recombinant rat IFNα (specific activity: 1 × 10⁸ units/mg; PBL Biomedical Laboratories, Piscataway, NJ) at 1,000 units/mL and human IFNα (specific activity: 1.8 × 10⁸ units/mg; PeproTech Inc., Rocky Hill, NJ) at 2,000 units/mL, rat IFNβ at 1,000 units/mL (specific activity: 8 × 10⁷ units/mg; PBL Biomedical Laboratories), and IFNγ at 100 units/mL and 1,000 units/mL for rat and human (specific activity: 2 × 10⁷ units/mg; R&D Systems, Abingdon, U.K.). The synthetic dsRNA analog PIC (Invivogen, San Diego, CA) was used at the final concentration of 1 μg/mL. PIC transfection was performed in the same conditions described for siRNA, but using 0.15 μL Lipofectamine 2000 per condition (7,11).

The percentage of living, apoptotic, and necrotic cells was determined using the DNA-binding dyes propidium iodide (PI, 5 μg/mL; Sigma-Aldrich, Bornem, Belgium) and Hoechst dye 33342 (HO, 5 μg/mL; Sigma-Aldrich) as described (7,19,25). Results are expressed as percentages (i.e., number of apoptotic cells/total number of cells × 100). Apoptosis was confirmed by Western blot analysis of cleaved caspases 9 and 3 and cytochrome c release as previously described (26,27).

For promoter studies, INS-1E cells were plated in 24-well plates and transfected with siRNAs as mentioned above. After 12 h of recovery, cells were cotransfected with pRL-CMV encoding Renilla luciferase (Promega, Madison, WI) and a firefly luciferase promoter-reporter construct containing three IFNγ-activated sequence (GAS) consensus sequences (GAS reporter–LR0075; Panomics, Fremont, CA) or four IFN-stimulated response element (ISRE) consensus sequences (ISRE reporter–LR0040). After 24 h of recovery and 4 h of treatment with IFNα or IFNγ, luciferase activities were measured using the Dual-Luciferase Reporter Assay System (Promega) and corrected for the luciferase activity of the internal control plasmid, pRL-CMV (28).

Immunofluorescence, Western blot analysis, mRNA extraction, and real-time PCR were performed as described (27,29). The antibodies used in Western blot analysis are described in Supplementary Table 2.

Data are presented as means ± SEM. Comparisons were performed by ANOVA followed by Student t test with Bonferroni correction or by paired t test as indicated. A P value <0.05 was considered statistically significant.

INS-1E cells were left untransfected or transfected with a control siRNA (siCTRL) or an siRNA targeting PTPN2 (siPTPN2), previously shown to efficiently inhibit PTPN2 expression (19). Cells were subsequently treated for 24 h with IFNα or IFNβ (Fig. 1A). None of the treatments significantly affected cell viability in nontransfected or siCTRL-transfected cells, whereas PTPN2 inhibition exacerbated apoptotic cell death in all IFN-treated cells (Fig. 1A). Additional experiments using a higher concentration of IFNα and IFNβ (2,000 units/mL) confirmed the proapoptotic activity of siPTPN2 (Supplementary Fig. 1A), and the concentration of 1,000 units/mL was chosen for subsequent experiments in INS-1E cells and primary rat β-cells. Because the effects of IFNα and IFNβ were similar (Fig. 1), most subsequent experiments were performed with IFNα. Similar results were observed in primary rat β-cells (Fig. 1B). Thus, inhibition of PTPN2 by two different siRNAs (siPTPN2 and siPTPN2#1) increased apoptosis after a 48-h treatment with type I or II IFNs. Of note, the necrotic component of cell death was negligible (<2% of cells, data not shown) in all experiments.

We next performed viability experiments using different concentrations of siPTPN2 (1, 5, 10, and 30 nM). PTPN2 expression gradually decreased with increasing concentrations of siPTPN2 (Supplementary Fig. 1B), whereas IFNα- and IFNγ-induced β-cell apoptosis progressively increased (Supplementary Fig. 1C).

During viral infections, β-cells release type I IFNs (7,11) that act in an autocrine/paracrine fashion through the type I IFN receptor (IFNAR). To evaluate the potential role of IFNAR in PIC-induced apoptosis, INS-1E cells were transfected with siRNAs targeting PTPN2, IFNAR1, or PTPN2 and IFNAR1 and subsequently transfected with PIC (Fig. 1C). Inhibition of PTPN2 increased PIC-induced apoptosis. IFNAR silencing decreased PIC-induced cell death, whereas the double knockdown of PTPN2 and IFNAR abolished the exacerbating effect of PTPN2 inhibition on PIC-induced apoptosis (Fig. 1C). These data suggest that PIC-induced β-cell apoptosis in PTPN2-silenced cells is at least partially mediated by type I IFNs that are released by the β-cells themselves in response to the synthetic dsRNA.

Type I IFNs preferentially induce STAT1/2 heterodimers that activate ISREs, whereas type II IFNs favor the activation of STAT1 homodimers that bind to the GAS (30). Inhibition of PTPN2 in INS-1E cells significantly increased IFNα-induced ISRE reporter activity (1.7-fold), whereas IFNγ stimulation led to poor ISRE reporter activation, even after PTPN2 silencing (Fig. 2A). The GAS reporter activity was increased in untransfected and siCTRL-transfected INS-1E cells after IFNγ treatment, and PTPN2 silencing further exacerbated activation of the GAS reporter (twofold induction), whereas IFNα did not modify GAS reporter activity (Fig. 2B). Double transfection of PTPN2- and STAT1-targeting siRNAs potently inhibit both target proteins (19), which was confirmed in the present series of experiments (data not shown). PTPN2 knockdown significantly increased IFNα- and IFNγ-induced apoptosis, whereas double knockdown of PTPN2 and STAT1 prevented the proapoptotic effect of PTPN2 inhibition in INS-1E cells exposed to IFNα or IFNγ (Fig. 2C). These results suggest that STAT1 activation plays a key role in both type I and II IFN–induced β-cell apoptosis in PTPN2-deficient cells.

To clarify the death pathways by which PTPN2 deficiency exacerbates type I and II IFN–induced cell death, we first analyzed genes induced by the endoplasmic reticulum stress response, previously shown to be associated with interleukin (IL)-1β +IFNγ-induced apoptosis of β-cells (31). The mRNA expression of C/EBP homologous protein (CHOP), immunoglobulin heavy chain-binding protein (BIP), or X-box binding protein 1 spliced (XBP-1s) was not induced in siPTPN2-transfected INS-1E cells after 24 h of treatment with IFNα or IFNγ compared with siCTRL (Supplementary Fig. 2). There were also no changes in BIP expression after 2, 4, 8, and 16 h of treatment with IFNα or IFNγ and siPTPN2 exposure (data not shown), making it unlikely that endoplasmic reticulum stress response contributes to IFN-induced apoptosis after PTPN2 knockdown.

We next examined the intrinsic mitochondrial pathway of apoptosis, because IFNγ activates this cell death pathway when associated with the proinflammatory cytokines IL-1β or tumor necrosis factor-α (25,27). Immunofluorescence analysis of INS-1E cells in which PTPN2 was silenced demonstrated diffuse BCL2-associated X protein (BAX) staining in untreated cells (Fig. 3, top), whereas it colocalized with the mitochondrial marker ATP synthase after 16 h of treatment with IFNα (Fig. 3, middle) or IFNγ (Fig. 3, bottom), demonstrating that BAX translocates to the mitochondria in siPTPN2-silenced cells. In line with our initial viability experiments (Fig. 1), fewer apoptotic cells were observed in untreated or IFN-treated siCTRL-transfected cells (data not shown).

Western blot evaluation of mitochondrial and cytoplasmic fractions from siCTRL- and siPTPN2-transfected INS-1E cells demonstrated that IFNα and IFNγ induced mitochondrial cytochrome c release to the cytoplasm in PTPN2-deficient cells but not in siCTRL-transfected cells (Fig. 4A and B). In line with these findings, PTPN2 knockdown also induced caspase 3 activation after 16 and 24 h of IFNα and IFNγ treatment (Fig. 4C and D). Taken together, the data from Figures 3 and 4 confirm the results observed in viability assays (Fig. 1) and indicate that apoptosis occurs mainly through the mitochondria-driven intrinsic pathway of cell death in IFN-treated PTPN2-deficient cells.

The members of the B-cell lymphoma 2 (BCL-2) family, death protein 5 (25), BCL2 binding component 3 (p53 up-regulated modulator of apoptosis [PUMA]) (27), and BCL2-like 11 (Bim) (32), have been shown to contribute to β-cell apoptosis after exposure to cytokines (IL-1β + IFNγ) or metabolic stress. IFNα or IFNγ treatment induced death protein 5 and p53 upregulated modulator of apoptosis expressions over time in the treated cells, but there were no significant differences between siCTRL- and siPTPN2-transfected cells (Supplementary Fig. 3). On the other hand, cytokine-induced Bim mRNA upregulation was exacerbated in PTPN2-deficient INS-1E cells after 16–24 h of IFNα or IFNγ treatment compared with their respective controls (Fig. 5A and B). However, there was not a significant modulation of Bim protein expression in siCTRL and siPTPN2-transfected cells after treatment with IFNα (Fig. 5C) or IFNγ (Fig. 5D).

Bim activity may be controlled by posttranscriptional phosphorylation at multiple serine and threonine residues (33). PTPN2 inhibition increased Bim phosphorylation at residue 65 in untreated INS-1E cells, and this effect was prolonged until 4 h of treatment with IFNα or 8 h after treatment with IFNγ (Fig. 5E). Similar phosphorylation levels were observed in CTRL and PTPN2-deficient cells after 16 h of treatment of IFNα or IFNγ. These results suggest that Bim phosphorylation is increased in PTPN2-silenced cells, which may render these cells more sensitive to the proapoptotic effects of IFNs.

To assess whether Bim indeed contributes to type I and II IFN–induced β-cell death in PTPN2-inhibited cells, we silenced PTPN2 and Bim in a double-knockdown approach. As previously shown (Fig. 1), inhibition of PTPN2 significantly exacerbated INS-1E cell apoptosis after treatment with IFNα, IFNβ, or IFNγ (Fig. 6A). This effect was counteracted by Bim knockdown, which protected INS-1E cells against IFNα-, IFNβ-, and IFNγ-induced apoptosis in PTPN2-silenced cells by 58, 67, and 78%, respectively (Fig. 6A). These observations were confirmed in FACS-purified rat primary β-cells (Fig. 6B), in which Bim inhibition again reversed the exacerbating effect of PTPN2 knockdown, with a 59 and 63% reduction of cell apoptosis after IFNα and IFNγ treatment, respectively. Similar observations were made in dispersed human islet cells (Fig. 6C). Thus, PTPN2 knockdown exacerbated IFN-induced cell death and Bim knockdown reversed the deleterious effect of PTPN2 knockdown, with a reduction of cell death nearly to basal levels (Fig. 6C).

The protective effects of Bim knockdown were confirmed by Western blot analysis of caspases 9 and 3 activation in INS-1E cells. As shown in Figure 6D–F, PTPN2 knockdown exacerbated caspases 9 and 3 activation in IFNα-, IFNβ-, and IFNγ-treated INS-1E cells, whereas the concomitant inhibition of PTPN2 and Bim abrogated this deleterious effect (Fig. 6E and F). As a whole, these results suggest that Bim plays a central role in the exacerbation of IFN-induced β-cell apoptosis in PTPN2-deficient cells.

To assess whether the effect of PTPN2 on Bim phosphorylation was mediated by Jun NH₂-terminal kinase 1 (JNK1), we performed a double-knockdown approach, targeting PTPN2 and JNK1 and evaluating Bim phosphorylation by Western blot in INS-1E cells. As previously shown (Fig. 5E), PTPN2 knockdown increased Bim phosphorylation mainly in the nontreated condition (Fig. 7A and B), although this effect was also evident after 2 h of treatment with IFNα (Fig. 7A) or IFNγ (Fig. 7B). After 24 h of treatment with IFNα (Fig. 7A) or IFNγ (Fig. 7B), the phosphorylation level observed in CTRL and PTPN2-deficient cells was similar. After JNK1 knockdown or double knockdown of PTPN2 and JNK1, Bim phosphorylation was decreased in both the nontreated condition (Fig. 7A and B) and after 2 or 24 h of IFNα (Fig. 7A) or IFNγ (Fig. 7B) treatment. These results indicate that the protein kinase JNK1 is in part responsible for the hyperphosphorylation of Bim at serine 65 (Ser65) in PTPN2-silenced β-cells.

We next silenced PTPN2 and JNK1 in a double-knockdown approach to test whether the observed decrease in Bim phosphorylation inhibited IFN-induced apoptosis in PTPN2-silenced cells. As shown in Figure 7C, JNK1 knockdown counteracted the increase in cell death in PTPN2-silenced INS-1E cells after 24 h of treatment with IFNα or IFNγ. These results were confirmed in FACS-purified rat primary β-cells, in which double knockdown of PTPN2 and JNK1 partially prevented the increase in IFN-induced cell death in the presence of PTPN2 inhibition (Fig. 7D).

The protective effects of JNK1 knockdown were confirmed by Western blot analysis of cleaved caspase 3 in INS-1E cells (Fig. 7A and B): PTPN2 inhibition increased cleaved caspase 3 in IFNα- and IFNγ-treated cells, whereas the double knockdown of PTPN2 and JNK1 revoked this effect.

The current study shows that inhibition of the T1D candidate gene PTPN2 sensitizes pancreatic β-cells to apoptosis induced by both type I and II IFNs. We further demonstrate that blocking IFNAR abolishes the exacerbation of apoptosis induced by the viral dsRNA analog PIC in PTPN2-silenced cells. dsRNA is produced in the cytosol of infected cells as a by-product of viral replication. This suggests that type I IFNs released by β-cells in response to a viral insult play an important role in cell death. This is in line with previous observations suggesting that intracellular dsRNA induces type I IFN release by β-cells, which contributes to cell death (7,11).

Silencing of PTPN2 increases ISRE and GAS reporter activity after IFN treatment, and double knockdown of PTPN2 and STAT1 reverses the proapoptotic effect of PTPN2 inhibition. PTPN2 silencing induces BAX translocation to the mitochondria, cytochrome c release to the cytosol, and caspase 3 activation after exposure to type I and II IFNs, characterizing activation of the intrinsic mitochondrial pathway of cell death. This process might be secondary to hyperphosphorylation of Bim in PTPN2-silenced cells via modulation of the protein kinase JNK1. In line with this hypothesis, both Bim and JNK1 knockdown protects β-cells against IFN-induced apoptosis in PTPN2-silenced cells. The key role of Bim was confirmed in human islet cells. This suggests that local IFN production may interact with a genetic factor (PTPN2) to induce aberrant proapoptotic activity of the BH3-only protein Bim, resulting in increased β-cell apoptosis.

Viruses are one of the putative environmental factors associated with T1D triggering in genetically predisposed individuals; several epidemiologic studies in humans and in animal models support this association (34). Thus, it was recently reported that enterovirus-positive blood samples are more frequent among subjects with T1D than among healthy individuals (35), and histologic studies of pancreas from deceased patients with T1D indicate a higher prevalence of Coxsackievirus B4 than in pancreas from control individuals (9,36). Type I IFNs are important mediators of the immune response against viral infections, and several studies demonstrated expression of type I IFNs in the islets and blood of individuals with T1D (37,38) and their role in the initiation or the acceleration of the autoimmune process in NOD mice (39). The present results indicate a potential link between a candidate gene for T1D (PTPN2) and the effects of type I IFNs in the progressive β-cell loss initiated by viral infections in genetically susceptible individuals. Thus, we suggest that development of T1D in some individuals may require a combination of infection by a potential diabetogenic virus (e.g., Coxsackievirus B4) (9), a vigorous and probably excessive local production of IFNs and other chemokines/cytokines, and the presence of particular polymorphisms in a candidate gene (i.e., PTPN2) that exacerbates IFN signaling (present data).

Type I and II IFN–induced β-cell apoptosis in PTPN2-silenced cells is accompanied by the exacerbation of the JAK-STAT signaling pathway (present data). On cytokine treatment, PTPN2 inhibition increases the activity of reporter vectors containing ISRE or GAS consensus sequences, and concomitant knockdown of STAT1 and PTPN2 significantly decreases the exacerbation of IFN-induced apoptosis. Because PTPN2 is a well-known negative regulator of STAT1 activation (40), our data suggest a key role for STAT1 in type I and II IFN–induced β-cell apoptosis. Aberrant activation of the JAK-STAT pathway is associated with both T1D and other autoimmune and inflammatory diseases (41,42). For example, STAT1 phosphorylation is increased in the intestinal mucosa of patients with celiac disease, suggesting that persistent STAT1 activation contributes to maintaining and expanding the local inflammatory response (41).

As mentioned above, the present results demonstrate that PTPN2 knockdown contributes to type I and II IFN–induced β-cell apoptosis via activation of the BH3-only protein Bim and consequent activation of the intrinsic mitochondrial pathway of cell death. This pathway leads to BAX translocation to the mitochondrial membrane, mitochondrial permeabilization, cytochrome c release to the cytosol, activation of caspase 9, subsequent activation of the effector caspase 3, and cell apoptosis. Mitochondrial integrity is controlled by interactions between pro- and antiapoptotic members of the B-cell lymphoma 2 (Bcl-2) protein family (20). Bim is a Bcl-2 member that mediates apoptosis by activating the proapoptotic Bcl-2 members Bax and Bak (43), and it contributes to β-cell apoptosis induced by chronic exposure to high glucose and the Fas-FasL system (32). Bim-dependent apoptosis is regulated by both modulation of its expression and phosphorylation at specific serine or threonine residues (33). We presently show that PTPN2 inhibition increases Bim phosphorylation at Ser65 via JNK1 activation, and that Bim inhibition abrogates IFN-induced cell death in PTPN2-deficient cells. Our results are in line with previous reports showing that MAPK-p38- or JNK-induced phosphorylation of Bim at Ser65 increases its proapoptotic activity (44–46). It is interesting that PTPN2 silencing in β-cells (present data) induces hyperphosphorylation of Bim at residue Ser65 even in untreated cells, which is not sufficient to decrease cell viability. However, when this is combined with exposure to IFNs, cell death is triggered. This suggests that phosphorylation at Bim Ser65 is an apoptosis-sensitizing event in β-cells, leading to apoptosis only in the context of a wider proapoptotic environment prompted by local production of proinflammatory cytokines or viral infection.

SNPs in the PTPN2 gene are implicated in susceptibility to several autoimmune disorders and inflammatory diseases, including T1D, Crohn’s disease, and celiac disease (1,47,48). One of the SNPs (rs2542151) in the PTPN2 gene is associated with an earlier onset of T1D, marking a genetic difference between patients with early and late onset of the disease (49). The intronic risk allele (rs1893217-C) in PTPN2 correlates with decreased IL-2 receptor signaling in CD4⁺ T cells, as well as with decreased PTPN2 expression in CD4⁺ CD45RO T cells (15). Our present data demonstrate that inhibition of PTPN2 activity in pancreatic β-cells increases IFN-induced cell death, suggesting that SNPs leading to a decreased expression of this gene may sensitize β-cells to apoptosis after a triggering event (e.g., type I IFNs produced by β-cells in response to a viral infection). In line with this possibility, PTPN2 knockdown in a model of Crohn’s disease leads to increased claudin-2 (pore-forming protein) expression, further increasing IFNγ-induced intestinal epithelium permeability, which is an important feature of Crohn’s disease (17). Moreover, PTPN2 knockout mice have abnormal production of IFNγ, tumor necrosis factor-α, and other cytokines, resulting in a systemic inflammatory disease (50). Further studies are now required to define how allelic differences in humans affect PTPN2 activity and function, and the impact of these differences in immune responses and β-cell apoptosis.

The current study suggests that environmental factors (e.g., viral infections and the resulting IFN production) may interact with a genetic predisposition factor (e.g., reduced PTPN2 activity) to induce aberrant proapoptotic activity of the BH3-only protein Bim, resulting in increased β-cell apoptosis via the intrinsic apoptotic pathway. Moreover, it indicates that PTPN2 modulates the apoptotic activity of Bim via regulation of the protein kinase JNK1. These findings provide the first indication of a direct interaction between a candidate gene for T1D and the activation of a specific downstream proapoptotic pathway in pancreatic β-cells. Further studies of the molecular interactions between predisposition genes and environmental triggers for autoimmune diseases should clarify the early triggering of autoimmunity and indicate novel avenues for their prevention.

This work was supported by grants from the Communauté Francaise de Belgique–Actions de Recherche Concertées, Fonds National de la Recherche Scientifique, Belgium, the Belgium Program on Interuniversity Poles of Attraction initiated by the Belgium State (IUAP P6/40) and the European Union (project NAIMIT, in the Framework Programme 7 of the European Community). I.S. is the recipient of a postdoctoral fellowship from the Education Department of the Basque Country. F.M. is the recipient of a postdoctoral fellowship from Fonds National de la Recherche Scientifique, Belgium. E.N.G. is supported by an EMBO long-term fellowship.

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

I.S. and F.M. contributed to the original idea and the design of the experiments, researched data, contributed to discussion, and wrote, revised, and edited the manuscript. M.L.C. and E.N.G. researched data, contributed to discussion, and revised and edited the manuscript. L.M. and P.M. researched data. D.L.E. contributed to the original idea and the design of the experiments, contributed to discussion, and wrote, revised, and edited the manuscript.

The authors thank G. Vandenbroeck, R. Makhnas, A.M. Musuaya, S. Mertens, M.A. Neef, and M. Pangerl from the Laboratory of Experimental Medicine, Université Libre de Bruxelles, for excellent technical support.