PTPN2, a Candidate Gene for Type 1 Diabetes, Modulates Interferon-γ–Induced Pancreatic β-Cell Apoptosis

Male Wistar rats (Charles River Laboratories, Brussels, Belgium) were housed and used according to the guidelines of the Belgian Regulations for Animal Care. Islets were isolated by collagenase digestion and handpicked under a stereomicroscope. β-Cells were purified by autofluorescence-activated cell sorting (FACSAria; BD Bioscience, San Jose, CA) (28,29). The preparations contained 90.4 ± 3.2% β-cells (n = 4). β-Cells were cultured for 2 days in Ham's F-10 medium containing 10 mmol/l glucose, 2 mmol/l glutaMAX, 50 μmol/l 3-isobutyl-1-methylxanthine, 5% fetal bovine serum (FBS), 0.5% charcoal-absorbed BSA (Boehringer, Indianapolis, IN), 50 units/ml penicillin, and 50 μg/ml streptomycin (29,30). During cytokine exposure, cells were cultured in the same medium but without serum.

Human islets were isolated from 11 nondiabetic organ donors (age 63 ± 5 years; BMI 25.1 ± 0.8 kg/m²) in Pisa, Italy, with the approval of the local ethics committee. Islets were isolated by enzymatic digestion and density-gradient purification (31) and cultured in M199 medium containing 5.5 mmol/l glucose. Their functional status was determined using glucose-stimulated insulin release and was 2.8 ± 0.4 (expressed as stimulation index). 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 exposed to cytokines in the same medium without FBS for 2 days. The percentage of β-cells, examined in the 11 dispersed islet preparations by staining with anti-insulin antibody (1:1,000; Sigma, Bornem, Belgium) and donkey anti-mouse IgG rhodamine (1:200; Lucron Bioproducts, De Pinte, Belgium), was 49 ± 5%.

The rat insulin-producing INS-1E cell line (a gift from Dr. C. Wollheim, Centre Medical Universitaire, Geneva, Switzerland) was cultured as previously described (32,33).

The siRNAs used in this study are listed in Supplementary Table A1 (available in an online-only appendix at http://diabetes.diabetesjournals.org/cgi/content/full/db08-1510/DC1). The best settings for the transfection of siRNAs in INS-1E cells, primary fluorescence-activated cell–sorted rat β-cells, and dispersed human islets were established by using a fluorescein isothiocyanate (FITC)-conjugated siRNA (siGLO Green Transfection Indicator, Thermo Scientific). Different transfection methods were tested, including electroporation, and several lipid reagents. Among those, DharmaFECT 1 (Thermo Scientific) was selected for its high transfection efficiency, namely 95 ± 2.7% in INS-1E cells and 80.7 ± 5.6% in primary rat β-cells. No cell toxicity was observed with this transfection method (% cell death: 9.4 ± 1.7% after transfection vs. 8.4 ± 2.2% in control for INS-1E cells and 9.8 ± 1.3% after transfection vs. 7.4 ± 1.7% in control for primary rat β-cells). Cells were cultured in antibiotic-free medium for at least 24 h before transfection. For transfection, the siRNA and the DharmaFECT were diluted separately in OptiMEM medium and incubated at room temperature for 5 min. Lipid-siRNA complexes were then formed at room temperature for 20 min in a proportion of 0.75, 1.25, and 1.75 μl DharmaFECT to 150 nmol/l of siRNA for INS-1E cells, primary β-cells, and dispersed human islets, respectively. The complexes were diluted five times in antibiotic-free medium and added to the cells at a final concentration of 30 nmol/l siRNA (except when indicated otherwise) for overnight transfection. The concentration of 30 nmol/l was selected after dose-response studies (data not shown). Afterward, cells were cultured for a 48-h recovery period and subsequently exposed to cytokines.

The following cytokine concentrations were used, based on previous dose-response experiments (10,24,33): recombinant human IL-1β (specific activity 1.8 × 10⁷units/mg; a gift from C.W. Reinolds, National Cancer Institute, Bethesda, MD) at 10, 50, or 100 units/ml as indicated; recombinant murine TNF-α (specific activity: 2 × 10⁸units/mg; Innogenetics, Gent, Belgium) at 1,000 units/ml; and recombinant rat and human IFN-γ (specific activity: 2 × 10⁷units/mg; R&D Systems, Abingdon, U.K.) at 100 and 1,000 units/ml for rat cells and human islets, respectively. When cells were treated with cytokines, culture supernatants were collected for nitrite determination (nitrite is a stable product of nitric oxide [NO] oxidation) at OD^{540 nm} using the Griess method (34). Palmitate (sodium salt, Sigma) was dissolved in 90% ethanol, heated to 60°C, and used at a final concentration of 0.5 mmol/l in medium containing 1% BSA (35). The SERCA blocker CPA (Sigma) was dissolved in DMSO and used at a concentration of 25 μmol/l.

The percentage of viable, apoptotic, and necrotic cells was determined after a 15-min incubation with the DNA-binding dyes propidium iodide (PI, 5 μg/ml; Sigma) and Hoechst 33342 (HO, 5 μg/ml; Sigma) (10,29,36). A minimum of 500 cells was counted in each experimental condition. Viability was evaluated by two independent observers, one of them being unaware of sample identity. The agreement between findings obtained by the two observers was >90%. Results are expressed as percent apoptosis, calculated as number of apoptotic cells/total number of cells × 100. In some experiments, apoptosis was confirmed using the Cell Death Detection ELISAplus kit (Roche Diagnostics, Vilvoorde, Belgium), which detects cytoplasmic fragmented DNA.

Poly(A)⁺mRNA was isolated from INS-1E cells, rat primary β-cells, and human islets using the Dynabeads mRNA DIRECT kit (Invitrogen) and reverse transcribed as previously described (10,29,37). The real-time PCR amplification reaction was done as described (29,37), using SYBR Green and compared with a standard curve (38). Expression values were corrected for the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH). IL-1β alone does not modify GAPDH expression, whereas exposure to IL-1β + IFN-γ reduces its expression according to cell death in INS-1E cells (Supplementary Fig. A1 and (7,10)). The primers used in this study are detailed in Supplementary Table A2 (online appendix).

Cells were washed with cold PBS and lysed with either Laemmli buffer or phospho lysis buffer (the compositions of the buffers are provided in Supplementary Table A3). Lysates were then resolved by 8–10% SDS-PAGE and transferred to a nitrocellulose membrane. The antibodies used in this study are listed in Supplementary Table A4. Immunoreactive bands were revealed using the SuperSignal West Femto chemiluminescent substrate (Thermo Scientific), detected using a LAS-3000 charge-coupled device camera and quantified with the Aida Analysis software (Fujifilm).

INS-1E cells were seeded onto glass coverslips and treated as indicated. Cells were washed with cold PBS, fixed for 10 min in 100% methanol at −20°C, washed three times with PBS, quenched 5 min in PBS/0.1% sodium borohydride, washed twice with PBS, and incubated for 5 min in PBS containing 0.2% Triton X-100 (PBST). After a 1-h blocking with 5% normal goat serum in PBST, cells were incubated overnight with PTPN2 antibodies (0.7 μg/ml) in PBST/0.5% normal goat serum (Supplementary Table A4). Cells were washed three times with PBST, and FITC-conjugated goat anti-mouse antibodies (Lucron Bioproducts) were applied for 2 h at 1:200 in PBST. After two washes with PBST, nuclei were counterstained with Hoechst 33342 (HO) for 5 min and washed three times with PBS. Coverslips were mounted in mounting medium (Dakocytomation), and immunofluorescence was visualized on a Zeiss microscope.

Data are presented as means ± SE. Comparisons were performed by two-tailed paired Student's t test or by ANOVA followed by Student's t test with Bonferroni correction. A P value <0.05 was considered statistically significant.

We first evaluated the expression of PTPN2 mRNA (both TC45 and TC48 isoforms) in primary fluorescence-activated cell–sorted rat β-cells, human islets, and rat insulin–producing INS-1E cells and tested whether cytokine treatment would affect its expression. PTPN2 mRNA was expressed in untreated primary rat β-cells and human islets cells, and IL-1β + IFN-γ exposure upregulated its expression by 9.6-fold (rat β-cells) and 2.2-fold (human islets) after 24 and 48 h, respectively (Fig. 1,A and B). A time course of IL-1β + IFN-γ treatment in INS-1E cells indicated that PTPN2 mRNA was already induced after 2 h, increasing progressively up to 24 h (Fig. 1,C). We then confirmed by Western blot that PTPN2 was upregulated by two- and threefold, respectively, in human islets and INS-1E cells after IL-1β + IFN-γ treatment (Fig. 1,D and E andFig. 3,B and C). In accordance with previous reports (13), we observed TC45 (arrow,Figs. 1,D and 3 B) to be the major PTPN2 isoform expressed in β-cells, with TC48 being poorly expressed (Fig. 1,D and 3 B). In contrast, the expression of PTPN22 (another phosphatase associated with type 1 diabetes risk; (39)) was only present in samples from rat spleen and lymph nodes (used as positive controls), whereas no or marginal PTPN22 expression was observed in primary β-cells, human islets, and INS-1E cells. The expression of PTPN22 was not modified by cytokines (Supplementary Fig. A2 in the online appendix).

PTPN2 was mainly located in the nucleus under unstimulated conditions (Fig. 2, left panels). However, a 15-min treatment with IFN-γ alone or in combination with IL-1β or TNF-α resulted in redistribution of the protein between the cytoplasm and the nucleus (Fig. 2, middle and right panels).

We next examined the role of PTPN2 in β-cells. To this end, INS-1E cells were left untransfected, transfected with an irrelevant control siRNA (siCtrl), or with an siRNA targeting both PTPN2 isoforms (siPTPN2). Cells were then left untreated, or treated for 24 h with IL-1β, IFN-γ, TNF-α, or with the combination of IL-1β + IFN-γ or TNF-α + IFN-γ. PTPN2 mRNA expression was significantly upregulated after IL-1β and IL-1β + IFN-γ treatment in untransfected and siCtrl-transfected INS-1E cells (Fig. 3,A). The expression of PTPN2 proteins was increased by all cytokines tested, alone or in combination, in both untransfected and siCtrl-transfected cells (Fig. 3,B and C), suggesting that PTPN2 expression is regulated at both transcriptional and post-transcriptional levels. Transfection with PTPN2 siRNA potently inhibited basal and cytokine-induced PTPN2 expression at the mRNA and protein level (Fig. 3,A–C). We then evaluated whether siRNA-mediated PTPN2 inhibition affects cytokine-induced apoptosis in INS-1E cells. As expected, a 24-h treatment with IL-1β + IFN-γ or TNF-α + IFN-γ induced apoptosis in untransfected and siCtrl-transfected cells, whereas each individual cytokine did not (Fig. 4,A; of note, the necrotic component was <2% under all experimental conditions tested; data not shown). The inhibition of PTPN2 exacerbated β-cell apoptosis after exposure to both IL-1β + IFN-γ and TNF-α + IFN-γ treatments. Importantly, PTPN2 inhibition also rendered IFN-γ treatment toxic to the cells, whereas this cytokine alone failed to induce apoptosis in both control counterparts (Fig. 4,A). These results were confirmed using a second method to evaluate apoptosis, namely quantification of cytoplasmic fragmented DNA (Fig. 4,B), and also with a second siRNA targeting PTPN2 (see below). The increased apoptosis in cytokine-treated PTPN2-deficient INS-1E was not accompanied by a more severe decrease in Ins1 and Ins2 mRNA contents in these cells (Supplementary Fig. A3) and could not be explained by a higher release of NO, since nitrite was similarly produced by untransfected, siCtrl-, and siPTPN2-transfected cells after IL-1β, IL-1β + IFN-γ, and TNF-α + IFN-γ treatments (Fig. 4,C). Blocking PTPN2 using siRNA also augmented IFN-γ–, IL-1β + IFN-γ–, and TNF-α + IFN-γ–induced apoptosis in primary rat β-cells (Fig. 4,D). Similar observations were made in dispersed human islet cells, in which siRNA-mediated PTPN2 inhibition exacerbated IL-1β + IFN-γ–induced cell death by 30–38% (Fig. 4 E and F). This was again independent of NO production (data not shown). This effect seems to be specific for cytokines, since PTPN2 inhibition did not increase β-cell apoptosis in response to palmitate, high glucose, or the SERCA blocker CPA (Supplementary Fig. A4).

Taking into account that STAT1 is a substrate of PTPN2 in other cell types (13) and that it is an important mediator of cytokine-induced β-cell apoptosis (18), we next examined the effect of PTPN2 inhibition on the kinetics and magnitude of IFN-γ–induced STAT1 phosphorylation. STAT1 phosphorylation was highly induced after 15 min of IFN-γ treatment in both untransfected and siCtrl-transfected controls, slowly decreasing between 2 and 24 h (Fig. 5,A and B). However, IFN-γ–induced STAT1 phosphorylation was markedly enhanced in cells lacking PTPN2, reaching a peak at 1–2 h and slowly decreasing afterward (Fig. 5,A and B). We confirmed that the increased STAT1 phosphorylation in PTPN2-deficient cells was not due to an augmentation of total STAT1 content in these cells (Fig. 5,A). Comparable results were observed for STAT3, another target of PTPN2 phosphatase activity, with clearly increased STAT3 phosphorylation in PTPN2-deficient cells (Fig. 5 A and C). These data demonstrate that the phosphatase PTPN2 is a major modulator of IFN-γ–induced STAT1 and STAT3 activity in β-cells. We also evaluated the p42/44 MAPK (ERK), EGFR, and IRβ activation pathways, previously described as PTPN2 targets in other cell types (14,16,40). Neither ERK nor EGFR signaling pathways were affected after PTPN2 inhibition in INS-1E cells (Supplementary Fig. A5). On the other hand, there was an increase in IRβ phosphorylation after 30 min of cytokine treatment in PTPN2-inhibited INS-1E cells that lasted until 14 h (Supplementary Fig. A5).

To evaluate the role of STAT1 in the exacerbation of cytokine-induced β-cell apoptosis observed after PTPN2 inhibition, we additionally interfered with STAT1 in a double knockdown approach. We first confirmed by Western blot that both PTPN2 and STAT1 siRNAs adequately inhibited their respective targets without affecting the expression of the other protein and also that the double transfection of PTPN2 and STAT1 siRNAs potently inhibited both target proteins (Fig. 6,A). As previously shown (Fig. 4), IL-1β + IFN-γ or TNF-α + IFN-γ induced apoptosis to a similar degree in both control conditions, and PTPN2 inhibition exacerbated cytokine-induced apoptosis, also rendering treatment with IFN-γ alone toxic to the cells (Fig. 6,B). Inhibition of STAT1 protected the cells against IL-1β + IFN-γ- and TNF-α + IFN-γ–induced apoptosis (by 56 and 67%, respectively). STAT1 knockdown abrogated the proapoptotic effect of PTPN2 inhibition in cells exposed to IFN-γ or to combinations of cytokines (Fig. 6 B). These results were confirmed using a second siRNA targeting PTPN2 expression (Supplementary Fig. A6) and suggest that increased STAT1 activity contributes to the aggravation of cytokine-induced apoptosis in PTPN2-deficient β-cells.

We presently show that the phosphatase PTPN2 is expressed in human islets and rat primary and immortal β-cells and that its expression is regulated by the pro-inflammatory cytokines IL-1β, IFN-γ, and TNF-α in these cells. PTPN2 was first identified in humans as a T-cell PTP (41) with two spliced variants, namely TC45 and TC48 (13,42). In rat β-cells and human islets, the TC45 isoform is by far the most prevalent (Fig. 1,D and 3 B), and it is able to shuttle between the nucleus and the cytoplasm after cytokine exposure (Fig. 2). This change in subcellular location of PTPN2 has been shown previously in other cell types (14,16) and allows the phosphatase to target numerous cellular substrates, including cytoplasmic JAK tyrosine kinases, EGFR, the IRβ and nuclear STAT1, STAT3, STAT5, and STAT6 (43,44). Because of its modulatory role in a wide variety of signaling pathways, perturbations in PTPN2 expression are associated with diverse pathogenic processes, including ABC-like diffuse large B-cell lymphomas (45), inflammation-associated tumorigenesis (44), and several autoimmune disorders (see below).

PTPN2 inhibition in β-cells results in an early and more intense STAT1 and STAT3 activation after IFN-γ treatment, suggesting that this phosphatase plays an important role in the dephosphorylation and consequent inactivation of these transcription factors in β-cells. This early regulation of IFN-γ–induced STAT1 activation seems to be critical for the subsequent triggering of β-cell apoptosis, since PTPN2-deficient β-cells show an aggravation of cytokine-induced apoptosis, which is reverted by the combined inhibition of PTPN2 and STAT1. Importantly, PTPN2 inhibition unmasked the proapoptotic effect of IFN-γ alone in both INS-1E and primary β-cells, while it failed to augment palmitate- or CPA-induced β-cell death. Taken together, these observations indicate two interesting aspects that require additional investigation:1) The crucial negative feedback of PTPN2 on the STAT1 signaling pathway suggests that PTPN2 is part of the “defense mechanisms” triggered by β-cells in response to cytokines. It remains to be clarified why this and other defense mechanisms (24) are not sufficient to protect β-cells against apoptosis. 2) IFN-γ alone induces STAT1 activation, but this only leads to apoptosis when PTPN2 activation is prevented. This suggests that STAT1 must cross a “nuclear activation threshold” before becoming proapoptotic to β-cells. It remains to be clarified whether this threshold depends on the intensity/length and/or periodicity of STAT1 activation, as previously suggested for nuclear factor κB (33).

The exacerbation of apoptosis in β-cells with decreased PTPN2 expression is independent of augmented NO production. This contrasts with observations reporting that PTPN2-deficient macrophages produce higher amounts of nitric oxide under inflammatory conditions (40,46), highlighting the differential regulation of inflammation-associated genes in cells from different backgrounds.

As discussed above, the transcription factor STAT1 has a major role in cytokine-induced β-cell apoptosis in vitro (18) and in vivo (27). It is thus conceivable that a genetically determined modification in the expression and/or the function of PTPN2 may sensitize β-cells to IFN-γ/STAT1–driven proapoptotic signaling, amplifying β-cell loss under inflammatory conditions, and in some cases, contributing to diabetes. The potential effect of the transient hyper-phosphorylation of STAT3 or IRβ in PTPN2-inhibited cells remains to be clarified. STAT3 has been reported to be dispensable for β-cell development and function (47), but it is generally associated with increased survival of tumoral cells (48), whereas the insulin receptor may exert positive feedback on β-cell function and survival (20).

Over the past few years, several new candidate genes have been identified in human type 1 diabetes. Most of these genes are expressed in the immune system, including major histocompatibility complex (e.g., HLA-DRB1), CTLA4, IL-2Rα (CD25), IFIH1 (MDA5), and several protein tyrosine phosphatases, namely PTPN22, PTPN11, and the recently reported PTPN2 (3,49). Modified function or regulation of these genes could contribute to the development of an autoimmune response. Thus, nonsynonymous single-nucleotide polymorphisms (SNPs) in PTPN22 (human LYP) have been associated with type 1 diabetes and other autoimmune processes, probably secondary to defective deletion of autoreactive T-cells during thymic selection (39). The fact that PTPN22 is not expressed in β-cells (present data) reinforces the likelihood that this is solely due to effects on the immune system. PTPN2-null mice present severe abnormalities in the immune system, resulting in a systemic inflammatory disease and widespread tissue infiltration by mononuclear cells (13,46,50). Moreover, the PTPN2 gene is associated with several other autoimmune disorders besides type 1 diabetes, including Crohn's disease, ulcerative colitis, and rheumatoid arthritis (49,51,52). It is currently unknown whether the SNPs identified in the PTPN2 gene on chromosome 18p11 will lead to a gain or a loss of function of the protein. PTPN2 overexpression in other cell types has been shown to induce a p53- and caspase 1–dependent apoptosis (53), whereas suppression of PTPN2 sensitize β-cells to cytokine-induced apoptosis (present data). In experiments on INS-1E cells, we observed that DNA vector–based overexpression of PTPN2 induced 30–50% of apoptosis 36 h after transfection, whereas there was <8% apoptotic cells in control-transfected cells (data not shown). These preliminary results require experimental confirmation in primary rat and human β-cells using adenoviral vectors to overexpress PTPN2, but they suggest that PTPN2 must be tightly regulated to avoid deleterious effects.

Pancreatic β-cells themselves may play an active role in the pathogenesis of type 1 diabetes (21), and at least three of the type 1 diabetes–associated genes are expressed in β-cells, namely insulin, IFIH1 (MDA5) (6,54), and PTPN2 (present study). β-Cells have been shown to participate actively in the recruitment and activation of the immune system by secreting various chemokines and cytokines under inflammatory conditions (18,21,55) and by providing “danger signals” to the immune system during apoptosis, especially in the context of local inflammation (56). We presently demonstrate that changes in PTPN2 function in β-cells sensitize these cells to surrounding proapoptotic inflammatory signals (e.g., cytokines), potentially amplifying β-cell loss and insulitis. This suggests that PTPN2, a type 1 diabetes–associated gene, may modulate β-cell apoptosis in early type 1 diabetes independently of its potential effects on the immune system.

This work was supported by grants from the Fonds National de la Recherche Scientifique (FNRS-FRSM) Belgium, the Communauté Française de Belgique–Actions de Recherche Concertées (ARC), the European Union (STREP Savebeta, contract number 036903; in the Framework Programme 6 of the European Community), and the Belgium Program on Interuniversity Poles of Attraction initiated by the Belgium State (IUAP P6/40). F.M. is the recipient of a post-doctoral fellowship from FNRS, Belgium. M.L.C. is the recipient of a scholarship from CAPES (Brazilian Coordination for the Improvement of Higher Education Personnel).

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

We thank John A. Todd and Dr. Kate Downes, Department of Medical Genetics, University of Cambridge, Cambridge, U.K., for helpful discussions, and Hindrik Mulder for helpful suggestions on siRNA transfection methods. We thank M.A. Neef, G. Vandenbroeck, M. Urbain, J. Schoonheydt, R. Leeman, A.M. Musuaya, and S. Mertens from the Laboratory of Experimental Medicine, ULB, for excellent technical support.