Inactivation of Protein Tyrosine Phosphatases Enhances Interferon Signaling in Pancreatic Islets

Type 1 diabetes (T1D) is caused by progressive loss of pancreatic β-cells due to an autoimmune assault. Local inflammation and proinflammatory cytokines, particularly interferons (IFNs), play an important role in β-cell loss (1,2). IFN-γ signal transduction involves activation of the tyrosine kinases Janus kinase (JAK) 1 and JAK2 that phosphorylate STAT1. These then dimerize, translocate to the nucleus, and bind γ-activated sites of a diverse array of genes. Blocking activation of this transcriptional pathway protects islets from immune destruction (3–5). These data indicate that excessive activation of JAK/STAT signaling in islets during the inflammatory process contributes to β-cell dysfunction and death. Cytokine signaling results in the upregulation and release of factors by β-cells, such as chemokines, that attract immune cells and amplify the inflammatory process.

Protein tyrosine phosphatases (PTPs) are a large superfamily of enzymes that dephosphorylate tyrosine phosphorylated proteins to oppose the actions of protein tyrosine kinases (6,7). PTPs play an important role in the development of both forms of diabetes (8). The architecture and low thiol pKa of the Cys residue in the active site of PTPs renders these proteins highly susceptible to oxidation by reactive oxygen species (ROS) (6,8). ROS-mediated oxidation of the PTP active site Cys inhibits PTP activity and prevents substrate binding. Recent studies have established that PTP oxidation occurs in vivo under physiological and pathological conditions such as inflammation (6,8,9).

The total serum antioxidant levels of patients with prediabetes and patients with T1D are reduced compared with age-matched controls (10,11). Furthermore, ROS and oxidative stress have been linked to β-cell cytotoxicity and are believed to be involved in pathology (12). The mechanism, however, remains unclear. Here, we show that oxidative stress in pancreatic islets during insulitis results in the enhanced oxidation of PTPs and that inactivation of PTPs in human and rodent islets and β-cells leads to increased INF signaling, chemokine production, and cell death.

Mice were maintained at St Vincent’s Institute, and experiments were approved by the institutional animal ethics committee. The list of mice is provided in Supplementary Table 1.

Mouse pancreata were frozen in optimal cutting temperature embedding medium (Sakura Finetek, Tokyo, Japan) or fixed with formalin and embedded in paraffin. Sections were incubated with mouse anti-phosphorated (p)STAT1 (BD Biosciences, San Jose, CA) and guinea pig anti-insulin (DAKO, Glostrup, Denmark) or rabbit anti-nitrotyrosine (Millipore, Billerica, MA) primary antibodies. The sections were then incubated with secondary Alexa Fluor 488 anti-mouse and Alexa Fluor 568 anti-guinea pig (both from Molecular Probes, Life Technologies, Carlsbad, CA) and counterstained with DAPI for immunofluorescence or secondary biotinylated anti-rabbit (Vector Laboratories, Burlingame, CA) for immunohistochemistry.

Human pancreata were obtained, with informed consent from next-of-kin, from heart-beating, brain-dead donors by the Australian Islet Transplant Consortium, and approved by the human ethics committees of the hospitals involved and the Australian Red Cross. Human islets were isolated as described previously (13). Islets were used from donors without diabetes who were an average age of 64 ± 3 years. Mouse islets were isolated as described previously (14). The insulin-producing MIN6 and NIT-1 cell lines were cultured in DMEM (Life Technologies) supplemented with 10% fetal calf serum. Cytokine concentrations were selected based on previous time course and dose-response studies (15).

Analyses were performed with the ddCT method using β-actin as internal controls. Probes are provided in Supplementary Table 2.

Total (reversible and irreversible) PTP oxidation was assessed essentially as described previously (9). Briefly, PTP oxidation results in two pools of PTPs: oxidized (PTP-SOH; inactive) and reduced (PTP-S⁻; active). To detect oxidized PTPs in NOD mice, pancreas samples were alkylated with N-ethylmaleimide (NEM, Sigma-Aldrich), rendering active PTPs resistant to further modification, whereas oxidized PTPs remain unaffected. Excess NEM was removed by column filtration, and reversibly oxidized PTPs were reduced with dithiothreitol (Sigma-Aldrich). After a buffer exchange, the reduced PTPs (representing PTPs that initially were reversibly oxidized) were oxidized to sulfonic acid (PTP-SO₃H) using pervanadate (Sigma-Aldrich). Then, “hyperoxidized” PTPs were detected by immunoblotting with the PTPox antibody (R&D Systems, Minneapolis, MN). Immunoprecipitation was performed using the PTPox antibody (R&D Systems) and radioimmunoprecipitation assay buffer. Antibody/protein complexes were collected with Protein G PLUS-Agarose (Santa Cruz Biotechnology, Santa Cruz, CA), washed, and then boiled in sample buffer (Santa Cruz Biotechnology) to remove the antibody/protein complex. Equal amounts of proteins were resolved by 10% SDS-PAGE and immunoblotted with the antibodies indicated in Supplementary Table 3. All blots shown are representative for two to four independent experiments.

Human islet preparations were assessed for viability by simultaneously staining monodispersed islet cells with 5 μg/mL 7-aminoactinomycin D (7-AAD) for islet viability, 10 μmol/L Newport green (NG) for β-cells, and 1 μmol/L tetramethylrhodamine-ethyl-ester (TMRE) for apoptosis (Molecular Probes, Invitrogen, Grand Island, NY), and analyzing by flow cytometry following the method of Ichii et al. (16). Mouse islets were dispersed into single cells with trypsin. DNA fragmentation was analyzed by staining with propidium iodide, as previously described (17). The percentage cell death of MIN6 cells was determined in at least 600 cells per experimental condition by inverted fluorescence microscopy after staining with the DNA dyes Hoechst-33342 (10 µg/mL) and propidium iodide (5 µg/mL).

Data are means ± SEM of three to five independent experiments. Comparisons between groups were made by paired t test or by ANOVA, followed by t test with the Bonferroni correction.

To directly evaluate the role of insulitis-induced PTP oxidation, we used nontransgenic NOD mice and transgenic NOD mice expressing proinsulin (PI) in MHC class II bearing cells (NOD PI mice) (18). NOD PI mice do not develop insulitis due to immune tolerance to PI, but otherwise have a normal, functional immune system. At 12 weeks of age, NOD mice developed insulitis and oxidative stress that correlated with increased nitrotyrosine-positive islets in the pancreas (Fig. 1A–C), as previously reported (19). Insulitis and nitrotyrosine staining were markedly reduced in NOD PI transgenic mice at 12 weeks of age (Fig. 1A–C). Enhanced nitrotyrosine staining is indicative of oxidative stress and results from the increased production of superoxide and nitric oxide and the generation of peroxynitrite. Therefore, we next examined whether the increased ROS promote PTP oxidation and inactivation in the pancreata of NOD mice. To this end, we used an antibody (PTPox) developed against the signature motif of the prototypic PTPN1/PTP1B oxidized to the irreversible sulfonic (–SO₃H) state (9,20). This antibody can detect most classical PTPs when oxidized to the sulfonic state. Pancreata from 12-week-old NOD and NOD PI mice were homogenized in the presence of N-ethylmaleimide to prevent postlysis oxidation and to alkylate all reduced and active PTPs. This was followed by the reduction of reversibly oxidized PTPs and their subsequent hyperoxidation to the −SO₃H state for detection with PTPox by immunoblot analysis (Fig. 1D). In pancreata from NOD mice we detected an increase in the oxidation status of several phosphatases; these included PTPox species with molecular masses of ∼45/48, 50, and 67 kDa. These bands comigrate with the phosphatases PTPN2, PTPN1, and PTPN6, respectively. Importantly, there were no significant differences in the total expression of these PTPs in NOD and NOD PI mice (Fig. 1D). We monitored for the oxidation status of PTPN2 and PTPN6 proteins by immunoprecipitation (Fig. 1E and Supplementary Fig. 1A). We found PTPN2 and PTPN6 in PTPox immunoprecipitates from NOD mice (Fig. 1E), consistent with increased oxidation. We next determined if the oxidation of PTPs in immune-infiltrated pancreata was mediated by the enhanced ROS levels. We found that treating NOD mice with the antioxidants N-acetyl cysteine (NAC) or mito-TEMPO (mTEMPO) in the drinking water attenuated PTP oxidation (Fig. 1F). Moreover, mTEMPO decreased immune infiltration (Fig. 1G).

STAT1 is a direct substrate of PTPN2, and this PTP has been implicated in β-cell function and survival (21,22). Given the systemic PTP oxidation evident in NOD mice, we next asked whether the increased ROS promote PTP oxidation and inactivation to exacerbate STAT1 signaling. We therefore isolated pancreas from 12-week-old NOD and NOD PI mice and measured pSTAT1 levels by immunofluorescence. In keeping with a previous study (23), we observed increased STAT1 activation in immune-infiltrated islets from NOD mice (Fig. 1H and Supplementary Fig. 1B).

To study whether PTP inactivation enhances IFN-γ signaling and STAT1 activation in β-cells, we used the reversible PTP inhibitor sodium orthovanadate (Na₃VO₄). MIN6 and NIT-1 cells were treated with the PTP inhibitor and IFN-γ, and the effect of PTP inhibition on the kinetics and magnitude of IFN-γ–induced STAT1 phosphorylation was evaluated. STAT1 phosphorylation was highly induced after IFN-γ treatment in controls and in PTP inactivated cells (Fig. 2A and B). The phosphorylation of STAT1 occurred with different kinetics in the β-cell lines but was markedly prolonged in cells in which PTPs were inhibited with Na₃VO₄ (Fig. 2A and B). Comparable results were observed in mouse islets treated with Na₃VO₄ or a second PTP inhibitor (PTPXVIII) and in human islets treated with Na₃VO₄ (Fig. 2C and D).

During insulitis, locally produced cytokines can both contribute to β-cell apoptosis (1) and stimulate the production of several chemokines through STAT1 activation in islets (24). PTP inactivation enhanced the expression of the STAT1-regulated chemokines CXCL9, CXCL10, and CXCL11 in mouse islets (Fig. 2E).

We next evaluated whether PTP inhibition affects IFN-induced apoptosis in β-cells and primary islets. MIN6 cells were treated with a combination of the PTP inhibitor and IFN-γ (Fig. 3A). None of the treatments alone significantly affected cell viability, whereas PTP inhibition exacerbated apoptotic cell death in IFN-γ–treated cells (Fig. 3A). Importantly, we confirmed this result in human islets and β-cells and in mouse primary islets (Fig. 3B–D and Supplementary Fig. 2A–D). To determine the death pathways by which PTP inactivation exacerbates IFN-γ–induced cell death, we first analyzed the STAT1-signaling pathway, previously shown to be associated with cytokine-induced apoptosis of β-cells (1). Islets lacking STAT1 expression or overexpressing the suppressor of cytokine signaling (SOCS)-1 in β-cells were significantly protected against cell death induced by PTP inhibition and IFN-γ treatment (Fig. 3D).

We next examined Bcl-2 modulators of the intrinsic mitochondrial pathway of apoptosis. The focus was on the BH3-only proteins Bim and p53 upregulated modulator of apoptosis (PUMA) because these molecules have been implicated in the mechanism of cytokine-mediated β-cell death (2,22). Islets derived from PUMA-knockout mice were not protected from cell death (Fig. 3E). However, Bim inactivation significantly reduced cell death induced by PTP inactivation and IFN-γ treatment (Fig. 3E), suggesting a major role for this proapoptotic protein.

Type I IFN has been associated with the development of autoimmune diabetes in patients with diabetes and in NOD mice (22). Thus, we examined the effect of PTP inactivation in mouse islets treated with IFN-α. Similar to IFN-γ, IFN-α induced cell death after PTP inactivation (Fig. 3F). The effect was specific because islets deficient in IFN-α receptors (IFNAR1^−/−) were protected from cell death (Fig. 3F).

To directly assess the role of ROS in IFN-γ signaling, we used mouse pancreata deficient for the cytosolic and mitochondrial antioxidant enzyme glutathione peroxidase 1 (GPx1) that converts H₂O₂ to water. GPx1 deficiency results in elevated H₂O₂ levels and oxidative stress and increases pancreatic PTP oxidation (Supplementary Fig. 3) (25). In line with these data, isolated islets from GPx1-knockout mice had enhanced STAT1 activation after IFN-γ treatment (Fig. 3G). Moreover, GPx1-knockout islets were highly sensitive to IFN-γ, and cell death was prevented with antioxidant treatment (Fig. 3H). Taken together, these results suggest that oxidative stress causes the inactivation of PTPs to enhance IFN signaling and the inflammatory response to promote cell death in pancreatic islets.

The current study demonstrates for the first time that PTPs are inactivated upon immune infiltration to the pancreas, acting as an upstream event of the IFN-signaling pathway in islets in T1D. This builds on our recently published work showing the in vivo inactivation of PTPs by oxidative stress in obesity and insulin resistance (9).

Pancreatic β-cells express very low levels of antioxidant enzymes and are extremely sensitive to oxidative stress induced by inflammation (12). Inhibition of NADPH oxidases, such as NOX1, that produce superoxide protects β-cells from the toxic effects of inflammatory cytokines (26). However, the mechanism(s) linking ROS to the promotion of inflammation and thereby β-cell death remain(s) unclear. Here, we demonstrated that insulitis results in PTP oxidation in the pancreas of NOD mice with prediabetes. Furthermore, global inactivation of PTPs in the islets results in increased pSTAT1 and downstream targets of IFNs and contributes to the β-cell’s own demise. The JAK/STAT pathway modulates immune-mediated β-cell dysfunction and death. In line with our results, deficiency of STAT1 in NOD mice prevents islet inflammation (3) and protects β-cells against immune-mediated destruction induced by multiple low doses of streptozotocin (4). Moreover, overexpression of SOCS-1 in β-cells inhibits IFN signaling and protects NOD mice from insulitis and diabetes (5). Interestingly, one of the PTPs oxidized in the NOD pancreas is PTPN2, the inactivation of which has been shown in previous studies to enhance STAT1 phosphorylation and sensitize β-cells to apoptosis induced by IFNs (8,21,22). Consistent with our data, silencing of the proapoptotic molecule Bim prevents β-cell death induced by knockdown of PTPN2 and IFN treatment (22). In humans, genome-wide association studies have linked PTPN2 polymorphisms with T1D (27). Thus, our results establish an important novel mechanism of inactivation of T1D candidate genes by oxidative stress during insulitis. Indeed, overexpression of specific PTPs (e.g., PTPN2) might be a valuable strategy to protect β-cells against immune infiltration. However, if the phosphatase active site is inactivated by oxidative stress, as we demonstrate in the current study, combined therapies to overexpress PTPs and prevent ROS formation might be a better and more effective strategy to prevent β-cell destruction. Understanding which specific PTP to overexpress will be the subject of future work.

Although in this study we focused our attention on PTP inactivation in islets and β-cells, it is probable that the oxidation and inactivation of PTPs in immune cells also contribute to the inflammatory process of T1D. For example, CD8⁺ T cells deficient for PTPN2 and cross-primed by β-cell self-antigens escape tolerance and acquire cytotoxic T-cell activity, resulting in β-cell destruction in the RIP-mOVA model of autoimmune diabetes (28). Nevertheless, the role of global PTP oxidation/inactivation in immune cells during insulitis remains to be determined.

Dying β-cells may act as a “danger signal” in early T1D and, together with the local release of proinflammatory cytokines and chemokines, induce the amplification of the autoimmune reaction. We have demonstrated that PTP inactivation plays a key role for β-cell death in the context of IFN signaling. Our research highlights the potential for oxidative stress and PTP oxidation to drastically alter cellular signaling in pathology (i.e., T1D), which has relevant implications for the development of effective treatments of the disease.

Acknowledgments. The authors thank L. Elkerbout, L. Yachou-Wos, S. Fynch, S. Thorburn, and C. Selck (St Vincent's Institute) for technical assistance and Dr. T. Loudovaris and L. Mariana (Australian Islet Transplant Consortium, St Vincent’s Institute) for human islets.

Funding. This work was supported by a National Health and Medical Research Council of Australia project grant (APP1071350) and fellowship (H.E.T.). E.N.G. is supported by a JDRF fellowship. The St Vincent’s Institute receives support from the Operational Infrastructure Support Scheme of the Government of Victoria.

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

Author Contributions. W.J.S., S.A.L., H.S.Q., and S.M.T. researched data. T.W.H.K., T.T., J.B.d.H., and H.E.T. contributed to experimental design and discussion and reviewed and edited the manuscript. E.N.G. researched data, contributed to discussion, designed experiments, and reviewed, edited, and wrote the manuscript. E.N.G. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.