Diabetes Volume 64, July 2015 2489

William J. Stanley,1,2 Sara A. Litwak,1 Hong Sheng Quah,1,2 Sih Min Tan,3 Thomas W.H. Kay,1,2 Tony Tiganis,4 Judy B. de Haan,3 Helen E. Thomas,1,2 and Esteban N. Gurzov1,2

Inactivation of Protein Tyrosine Phosphatases Enhances Interferon Signaling in Pancreatic Islets

Diabetes 2015;64:2489–2496 | DOI: 10.2337/db14-1575

Type 1 diabetes (T1D) is the result of an autoimmune that phosphorylate STAT1. These then dimerize, translo- assault against the -producing pancreatic b-cells, cate to the nucleus, and bind g-activated sites of a diverse where chronic local inflammation (insulitis) leads to array of . Blocking activation of this transcriptional b-cell destruction. T cells and macrophages infiltrate into pathway protects islets from immune destruction (3–5). islets early in T1D pathogenesis. These immune cells se- These data indicate that excessive activation of JAK/STAT crete cytokines that lead to the production of reactive signaling in islets during the inflammatory process con- oxygen species (ROS) and T-cell invasion and activation. tributes to b-cell dysfunction and death. Cytokine signal- Cytokine-signaling pathways are very tightly regulated by ing results in the upregulation and release of factors by STUDIES ISLET protein tyrosine phosphatases (PTPs) to prevent exces- b-cells, such as chemokines, that attract immune cells and sive activation. Here, we demonstrate that pancreata amplify the inflammatory process. from NOD mice with islet infiltration have enhanced Protein tyrosine phosphatases (PTPs) are a large oxidation/inactivation of PTPs and STAT1 signaling com- superfamily of that dephosphorylate tyrosine pared with NOD mice that do not have insulitis. Inactiva- tion of PTPs with sodium orthovanadate in human and phosphorylated proteins to oppose the actions of protein rodent islets and b-cells leads to increased activation tyrosine kinases (6,7). PTPs play an important role in the of interferon signaling and chemokine production medi- development of both forms of diabetes (8). The architec- ated by STAT1 phosphorylation. Furthermore, this exac- ture and low thiol pKa of the Cys residue in the active erbated STAT1 activation–induced cell death in islets was site of PTPs renders these proteins highly susceptible to prevented by overexpression of the suppressor of cyto- oxidation by (ROS) (6,8). ROS- kine signaling-1 or inactivation of the BH3-only protein mediated oxidation of the PTP active site Cys inhibits PTP Bim. Together our data provide a mechanism by which activity and prevents substrate binding. Recent studies PTP inactivation induces signaling in pancreatic islets have established that PTP oxidation occurs in vivo under that results in increased expression of inflammatory physiological and pathological conditions such as inflam- genes and exacerbated insulitis. mation (6,8,9). The total serum antioxidant levels of patients with prediabetes and patients with T1D are reduced compared Type 1 diabetes (T1D) is caused by progressive loss of with age-matched controls (10,11). Furthermore, ROS pancreatic b-cells due to an autoimmune assault. Local and oxidative stress have been linked to b-cell cytotoxicity inflammation and proinflammatory cytokines, particularly and are believed to be involved in pathology (12). The interferons (IFNs), play an important role in b-cell loss mechanism, however, remains unclear. Here, we show (1,2). IFN-g signal transduction involves activation of that oxidative stress in pancreatic islets during insulitis the tyrosine kinases Janus kinase (JAK) 1 and JAK2 results in the enhanced oxidation of PTPs and that

1St Vincent’s Institute of Medical Research, Melbourne, Victoria, Australia Received 14 October 2014 and accepted 22 February 2015. 2 ’ Department of Medicine, St Vincent s Hospital, The University of Melbourne, This article contains Supplementary Data online at http://diabetes Melbourne, Victoria, Australia .diabetesjournals.org/lookup/suppl/doi:10.2337/db14-1575/-/DC1. 3Diabetic Complications Division, Baker IDI Heart and Diabetes Institute, Mel- © 2015 by the American Diabetes Association. Readers may use this article as bourne, Victoria, Australia long as the work is properly cited, the use is educational and not for profit, and 4Department of Biochemistry and Molecular Biology, Monash University, Mel- the work is not altered. bourne, Victoria, Australia Corresponding author: Esteban N. Gurzov, [email protected]. 2490 PTP Oxidation Promotes Inflammation Diabetes Volume 64, July 2015

inactivation of PTPs in human and rodent islets and to sulfonic acid (PTP-SO3H) using pervanadate (Sigma- b-cells leads to increased INF signaling, chemokine pro- Aldrich). Then, “hyperoxidized” PTPs were detected by duction, and cell death. immunoblotting with the PTPox antibody (R&D Systems, Minneapolis, MN). Immunoprecipitation was performed RESEARCH DESIGN AND METHODS using the PTPox antibody (R&D Systems) and radioimmu- Mice noprecipitation assay buffer. Antibody/protein complexes Mice were maintained at St Vincent’s Institute, and were collected with Protein G PLUS-Agarose (Santa Cruz experiments were approved by the institutional animal Biotechnology, Santa Cruz, CA), washed, and then boiled ethics committee. The list of mice is provided in Supple- in sample buffer (Santa Cruz Biotechnology) to remove mentary Table 1. the antibody/protein complex. Equal amounts of proteins Immunohistochemistry and Immunofluorescence were resolved by 10% SDS-PAGE and immunoblotted Mouse pancreata were frozen in optimal cutting temper- with the antibodies indicated in Supplementary Table 3. ature embedding medium (Sakura Finetek, Tokyo, Japan) All blots shown are representative for two to four inde- or fixed with formalin and embedded in paraffin. Sections pendent experiments. were incubated with mouse anti-phosphorated (p)STAT1 Cell Viability (BD Biosciences, San Jose, CA) and guinea pig anti-insulin Human islet preparations were assessed for viability by (DAKO, Glostrup, Denmark) or rabbit anti-nitrotyrosine simultaneously staining monodispersed islet cells with 5 (Millipore, Billerica, MA) primary antibodies. The sections mg/mL 7-aminoactinomycin D (7-AAD) for islet viability, were then incubated with secondary Alexa Fluor 488 anti- 10 mmol/L Newport green (NG) for b-cells, and 1 mmol/L mouse and Alexa Fluor 568 anti-guinea pig (both from tetramethylrhodamine-ethyl-ester (TMRE) for apoptosis Molecular Probes, Life Technologies, Carlsbad, CA) and (Molecular Probes, Invitrogen, Grand Island, NY), and counterstained with DAPI for immunofluorescence or analyzing by flow cytometry following the method of Ichii secondary biotinylated anti-rabbit (Vector Laboratories, et al. (16). Mouse islets were dispersed into single cells Burlingame, CA) for immunohistochemistry. with trypsin. DNA fragmentation was analyzed by staining with propidium iodide, as previously described (17). The Cell Culture and Treatments percentage cell death of MIN6 cells was determined in at Human pancreata were obtained, with informed consent least 600 cells per experimental condition by inverted from next-of-kin, from heart-beating, brain-dead donors by fluorescence microscopy after staining with the DNA the Australian Islet Transplant Consortium, and approved dyes Hoechst-33342 (10 mg/mL) and propidium iodide by the human ethics committees of the hospitals involved (5 mg/mL). and the Australian Red Cross. Human islets were isolated as described previously (13). Islets were used from donors Statistical Analysis without diabetes who were an average age of 64 6 3years. Data are means 6 SEM of three to five independent Mouse islets were isolated as described previously (14). The experiments. Comparisons between groups were made by insulin-producing MIN6 and NIT-1 cell lines were cultured paired t test or by ANOVA, followed by t test with the in DMEM (Life Technologies) supplemented with 10% fetal Bonferroni correction. calf serum. Cytokine concentrations were selected based on previous time course and dose-response studies (15). RESULTS PTPs Are Oxidized in Pancreata From NOD Mice Real-Time PCR To directly evaluate the role of insulitis-induced PTP Analyses were performed with the ddCT method using oxidation, we used nontransgenic NOD mice and trans- b -actin as internal controls. Probes are provided in Sup- genic NOD mice expressing proinsulin (PI) in MHC class II plementary Table 2. bearing cells (NOD PI mice) (18). NOD PI mice do not PTP Oxidation, Immunoprecipitation, and Western develop insulitis due to immune tolerance to PI, but other- Blotting wise have a normal, functional immune system. At 12 Total (reversible and irreversible) PTP oxidation was weeks of age, NOD mice developed insulitis and oxidative assessed essentially as described previously (9). Briefly, stress that correlated with increased nitrotyrosine-positive PTP oxidation results in two pools of PTPs: oxidized islets in the pancreas (Fig. 1A–C), as previously reported 2 (PTP-SOH; inactive) and reduced (PTP-S ; active). To de- (19). Insulitis and nitrotyrosine staining were markedly tect oxidized PTPs in NOD mice, pancreas samples were reduced in NOD PI transgenic mice at 12 weeks of age alkylated with N-ethylmaleimide (NEM, Sigma-Aldrich), (Fig. 1A–C). Enhanced nitrotyrosine staining is indicative rendering active PTPs resistant to further modification, of oxidative stress and results from the increased produc- whereas oxidized PTPs remain unaffected. Excess NEM tion of superoxide and nitric oxide and the generation of was removed by column filtration, and reversibly oxidized peroxynitrite. Therefore, we next examined whether the PTPs were reduced with dithiothreitol (Sigma-Aldrich). increased ROS promote PTP oxidation and inactivation in After a buffer exchange, the reduced PTPs (representing the pancreata of NOD mice. To this end, we used an anti- PTPs that initially were reversibly oxidized) were oxidized body (PTPox) developed against the signature motif of the diabetes.diabetesjournals.org Stanley and Associates 2491

Figure 1—Oxidative stress and pancreatic PTP oxidation in NOD mice with prediabetes. A: Insulitis development was scored on pancreata collected from female NOD PI and NOD mice at 12 weeks of age. Serial sections (3 mm thick) were prepared at three levels (200 mm apart) and scored in a blinded manner using the following scale: 0 = no infiltrate, 1 = peri-islet infiltrate, 2 = extensive (>50%) peri-islet infiltrate, 3 = intraislet infiltrate, and 4 = extensive intraislet infiltrate (>80%) or total b-cell loss. Greater than 50 islets per pancreas were scored. The percentage of islets per pancreas with each score was calculated. Scores were added to give an overall score for each pancreas, and data are shown as the mean. B: Paraffin sections (3 mm) of pancreas from 12-week-old NOD PI and NOD mice stained with hematoxylin and eosin (H&E). C: Paraffin sections were immunolabeled with the oxidative stress marker nitrotyrosine and counterstained with hematoxylin. D: Immunoblot analysis with PTPox antibody to determine relative PTP oxidation (PTP-SO3H) or for the indicated PTPs in 12-week-old 2492 PTP Oxidation Promotes Inflammation Diabetes Volume 64, July 2015 prototypic PTPN1/PTP1B oxidized to the irreversible sul- IFN-g treatment in controls and in PTP inactivated cells – A B fonic ( SO3H) state (9,20). This antibody can detect most (Fig. 2 and ). The phosphorylation of STAT1 occurred classical PTPs when oxidized to the sulfonic state. Pan- with different kinetics in the b-cell lines but was markedly creata from 12-week-old NOD and NOD PI mice were ho- prolonged in cells in which PTPs were inhibited with mogenized in the presence of N-ethylmaleimide to prevent Na3VO4 (Fig. 2A and B). Comparable results were observed postlysis oxidation and to alkylate all reduced and active in mouse islets treated with Na3VO4 or a second PTP in- PTPs. This was followed by the reduction of reversibly hibitor (PTPXVIII) and in human islets treated with oxidized PTPs and their subsequent hyperoxidation to Na3VO4 (Fig. 2C and D). the 2SO3H state for detection with PTPox by immunoblot During insulitis, locally produced cytokines can both analysis (Fig. 1D). In pancreata from NOD mice we detected contribute to b-cell apoptosis (1) and stimulate the pro- an increase in the oxidation status of several phosphatases; duction of several chemokines through STAT1 activation these included PTPox species with molecular masses of in islets (24). PTP inactivation enhanced the expression of ;45/48, 50, and 67 kDa. These bands comigrate with the the STAT1-regulated chemokines CXCL9, CXCL10, and phosphatases PTPN2, PTPN1, and PTPN6, respectively. Im- CXCL11 in mouse islets (Fig. 2E). fi portantly, there were no signi cant differences in the total Enhanced IFN Signaling Induces b-Cell Death expression of these PTPs in NOD and NOD PI mice (Fig. D We next evaluated whether PTP inhibition affects IFN- 1 ). We monitored for the oxidation status of PTPN2 and induced apoptosis in b-cells and primary islets. MIN6 cells PTPN6 proteins by immunoprecipitation (Fig. 1E and Sup- were treated with a combination of the PTP inhibitor and plementary Fig. 1A). We found PTPN2 and PTPN6 in PTPox IFN-g (Fig. 3A). None of the treatments alone significantly immunoprecipitates from NOD mice (Fig. 1E), consistent affected cell viability, whereas PTP inhibition exacerbated with increased oxidation. We next determined if the oxida- apoptotic cell death in IFN-g–treated cells (Fig. 3A). Impor- tion of PTPs in immune-infiltrated pancreata was mediated tantly, we confirmed this result in human islets and b-cells by the enhanced ROS levels. We found that treating NOD and in mouse primary islets (Fig. 3B–D and Supplementary mice with the antioxidants N-acetyl cysteine (NAC) or mito- Fig. 2A–D). To determine the death pathways by which TEMPO (mTEMPO) in the drinking water attenuated PTP PTP inactivation exacerbates IFN-g–induced cell death, oxidation (Fig. 1F). Moreover, mTEMPO decreased immune we first analyzed the STAT1-signaling pathway, previously infiltration (Fig. 1G). shown to be associated with cytokine-induced apoptosis of STAT1 is a direct substrate of PTPN2, and this PTP b-cells (1). Islets lacking STAT1 expression or overexpress- has been implicated in b-cell function and survival ing the suppressor of cytokine signaling (SOCS)-1 in b-cells (21,22). Given the systemic PTP oxidation evident in were significantly protected against cell death induced by NOD mice, we next asked whether the increased ROS PTP inhibition and IFN-g treatment (Fig. 3D). promote PTP oxidation and inactivation to exacerbate We next examined Bcl-2 modulators of the intrinsic STAT1 signaling. We therefore isolated pancreas from mitochondrial pathway of apoptosis. The focus was on the 12-week-old NOD and NOD PI mice and measured BH3-only proteins Bim and upregulated modulator of pSTAT1 levels by immunofluorescence. In keeping with apoptosis (PUMA) because these molecules have been a previous study (23), we observed increased STAT1 acti- implicated in the mechanism of cytokine-mediated b-cell vation in immune-infiltrated islets from NOD mice (Fig. death (2,22). Islets derived from PUMA-knockout mice 1H and Supplementary Fig. 1B). were not protected from cell death (Fig. 3E). However, Inactivation of PTPs Increases IFN-g Signaling in Islets Bim inactivation significantly reduced cell death induced To study whether PTP inactivation enhances IFN-g signal- by PTP inactivation and IFN-g treatment (Fig. 3E), sug- ing and STAT1 activation in b-cells, we used the reversible gesting a major role for this proapoptotic protein. PTP inhibitor sodium orthovanadate (Na3VO4). MIN6 and Type I IFN has been associated with the development NIT-1 cells were treated with the PTP inhibitor and IFN-g, of autoimmune diabetes in patients with diabetes and in and the effect of PTP inhibition on the kinetics and mag- NOD mice (22). Thus, we examined the effect of PTP nitude of IFN-g–induced STAT1 phosphorylation was eval- inactivation in mouse islets treated with IFN-a. Similar uated. STAT1 phosphorylation was highly induced after to IFN-g, IFN-a induced cell death after PTP inactivation

NOD PI vs. NOD pancreas samples. E: Immunoprecipitation (IP) of PTPox proteins from NOD pancreas samples and immunoblotting for PTPN6 (67 kDa) or PTPN2 (45/48 kDa). IgG heavy chain (HC) band (50 kDa) is shown. F: Female NOD mice (6 weeks old) were treated with NAC (1 mg/mL; Sigma-Aldrich, St. Louis, MO) for 8 weeks or mTEMPO (1 mmol/L; Enzo Life Sciences, Farmingdale, NY) for 14 weeks in drinking water and pancreata extracted for an assessment of PTP oxidation by SDS-PAGE. G: Insulitis scores on pancreata collected from female NOD mice that were treated with the control or with mTEMPO for 20 weeks (n = 8 per group). H: Cryosections of pancreata from 12-week-old NOD PI or NOD mice stained with antibodies recognizing pSTAT1, insulin, and DAPI (bar = 50 mm). Quantification of pSTAT1 staining after correction for nuclear DAPI in islets is shown. A total of 25 islets from 4 NOD PI mice, 7 islets without infiltration, and 124 islets with infiltration from 6 NOD mice were scored. *P < 0.05, ***P < 0.001. Confocal image of overlapped staining (green: pSTAT1, red: insulin, blue: DAPI) demonstrates pSTAT1/insulin costaining (white arrows). Scale bar is 10 mm. diabetes.diabetesjournals.org Stanley and Associates 2493

Figure 2—IFN-g–induced STAT1 signaling in b-cells and pancreatic islets. Time course of STAT1 protein activation after IFN-g (100 units/mL; BioLegend, San Diego, CA) and PTP inactivation with Na3VO4 (100 mmol/L, Sigma-Aldrich) in MIN6 (A) or NIT-1 (B)cells.Celllysateswere subjected to Western blotting with antibodies detecting pSTAT1, STAT1, or b-actin as loading control. The intensity values for the proteins were corrected by the values of the housekeeping protein b-actin and are shown as arbitrary units (A.U.). Results are the means 6 SEM of three to five independent experiments. *P < 0.05. C: Western blots demonstrate activation of STAT1 in mouse islets treated for 8 h with IFN-g (100 units/mL), Na3VO4 (100 mmol/L), the PTP inhibitor XVIII (PTPXVIII, 1 mmol/L; Millipore), or combination as indicated. The result is representative of two independent experiments. D: Activation of STAT1 in human islets isolated from two organ donors after treatment for 24 h with IFN-g (100 units/mL), Na3VO4 (100 mmol/L), or combination. E: Mouse islets were treated for 8 h with IFN-g,Na3VO4, or combination. Quantitative real-time PCR for STAT1-dependent chemokines and b-actin expression was then performed. Individual chemokine values have been divided by the housekeeping b-actin and presented as fold induction related to the Na3VO4-treated samples (considered as 1). Results are the means 6 SEM of four independent experiments. *P < 0.05, **P < 0.01. 2494 PTP Oxidation Promotes Inflammation Diabetes Volume 64, July 2015

Figure 3—Inactivation of PTPs potentiates cell death induced by IFNs. A: Cell death of MIN6 cells was evaluated by Hoechst-33342 (blue) and propidium iodide (red) 24 h after IFN-g (100 units/mL) and PTP inactivation with Na3VO4 (100 mmol/L) as indicated. Arrows indicate propidium iodide–positive cells. Data shown are means 6 SEM of four independent experiments. **P < 0.01. B: Human islets were cultured for 48 h with IFN-g (100 units/mL), Na3VO4 (100 mmol/L), or combination, and viability was measured by 7-AAD staining. Data shown are means 6 SEM of three independent experiments. *P < 0.05. C: Human b-cells from two organ donors treated for 4 h with IFN-g (100 units/mL), Na3VO4 (100 mmol/L), or combination. Human islets were dispersed into single cells and stained with 7-AAD, NG, and TMRE. After gating out 7-AAD–positive cells, NG-positive cells (b-cells) were then analyzed for their percentages of TMRE-positive cells (viability). D: DNA fragmentation was measured by flow cytometry in islets from wild-type, STAT12/2, or RIP-SOCS1 mice cultured for 24 h in medium containing IFN-g,Na3VO4, or combination. Data shown are means 6 SEM of four independent experiments. *P < 0.05, **P < 0.01. E: DNA fragmentation was measured by flow cytometry after incubation of wild-type C57BL/6, Puma2/2,orBim2/2 islets for 24 h with IFN-g,Na3VO4, or combination. Data shown are means 6 SEM of four independent experiments. *P < 0.05. F: DNA fragmentation was measured by flow cytometry in islets from wild-type or IFNAR12/2 mice cultured for 24 h in medium containing IFN-a (PBL Interferon- Source, Piscataway, NJ), Na3VO4, or combination. Data shown are means 6 SEM of three independent experiments. **P < 0.01. G: Islets from GPx1+/+ and GPx12/2 mice were isolated, treated with IFN-g for 24 h, and processed for immunoblot analysis. H: DNA fragmentation was measured by flow cytometry after incubation of GPx1+/+ and GPx12/2 islets for 24 h with IFN-g (100 mU/mL), NAC (1 mmol/L), or combination. Data shown are means 6 SEM of three independent experiments. *P < 0.05.

(Fig. 3F). The effect was specific because islets deficient in To directly assess the role of ROS in IFN-g signaling, 2/2 IFN-a receptors (IFNAR1 ) were protected from cell we used mouse pancreata deficient for the cytosolic and death (Fig. 3F). mitochondrial antioxidant peroxidase diabetes.diabetesjournals.org Stanley and Associates 2495

1 (GPx1) that converts H2O2 to water. GPx1 deficiency a better and more effective strategy to prevent b-cell de- fi results in elevated H2O2 levels and oxidative stress and struction. Understanding which speci c PTP to overex- increases pancreatic PTP oxidation (Supplementary Fig. 3) press will be the subject of future work. (25). In line with these data, isolated islets from GPx1- Although in this study we focused our attention on knockout mice had enhanced STAT1 activation after PTP inactivation in islets and b-cells, it is probable that IFN-g treatment (Fig. 3G). Moreover, GPx1-knockout the oxidation and inactivation of PTPs in immune cells islets were highly sensitive to IFN-g,andcelldeath also contribute to the inflammatory process of T1D. For was prevented with antioxidant treatment (Fig. 3H). example, CD8+ T cells deficient for PTPN2 and cross- Taken together, these resultssuggestthatoxidative primed by b-cell self-antigens escape tolerance and stress causes the inactivation of PTPs to enhance IFN acquire cytotoxic T-cell activity, resulting in b-cell destruc- signaling and the inflammatory response to promote tion in the RIP-mOVA model of autoimmune diabetes cell death in pancreatic islets. (28). Nevertheless, the role of global PTP oxidation/inac- tivation in immune cells during insulitis remains to be DISCUSSION determined. The current study demonstrates for the first time that Dying b-cells may act as a “danger signal” in early T1D PTPs are inactivated upon immune infiltration to the and, together with the local release of proinflammatory pancreas, acting as an upstream event of the IFN- cytokines and chemokines, induce the amplification of the signaling pathway in islets in T1D. This builds on our autoimmune reaction. We have demonstrated that PTP recently published work showing the in vivo inactiva- inactivation plays a key role for b-cell death in the context tion of PTPs by oxidative stress in and insulin of IFN signaling. Our research highlights the potential for resistance (9). oxidative stress and PTP oxidation to drastically alter Pancreatic b-cells express very low levels of antioxidant cellular signaling in pathology (i.e., T1D), which has rele- enzymes and are extremely sensitive to oxidative stress vant implications for the development of effective treat- induced by inflammation (12). Inhibition of NADPH oxi- ments of the disease. dases, such as NOX1, that produce superoxide protects b-cells from the toxic effects of inflammatory cytokines (26). However, the mechanism(s) linking ROS to the pro- Acknowledgments. The authors thank L. Elkerbout, L. Yachou-Wos, motion of inflammation and thereby b-cell death remain(s) S. Fynch, S. Thorburn, and C. Selck (St Vincent’s Institute) for technical assistance unclear. Here, we demonstrated that insulitis results in and Dr. T. Loudovaris and L. Mariana (Australian Islet Transplant Consortium, St PTP oxidation in the pancreas of NOD mice with predia- Vincent’s Institute) for human islets. betes. Furthermore, global inactivation of PTPs in the Funding. This work was supported by a National Health and Medical Re- search Council of Australia project grant (APP1071350) and fellowship (H.E.T.). islets results in increased pSTAT1 and downstream tar- E.N.G. is supported by a JDRF fellowship. The St Vincent’s Institute receives b ’ gets of IFNs and contributes to the -cell s own demise. support from the Operational Infrastructure Support Scheme of the Government The JAK/STAT pathway modulates immune-mediated of Victoria. b-cell dysfunction and death. In line with our results, Duality of Interest. No potential conflicts of interest relevant to this article deficiency of STAT1 in NOD mice prevents islet inflam- were reported. mation (3) and protects b-cells against immune-mediated Author Contributions. W.J.S., S.A.L., H.S.Q., and S.M.T. researched destruction induced by multiple low doses of streptozo- data. T.W.H.K., T.T., J.B.d.H., and H.E.T. contributed to experimental design tocin (4). Moreover, overexpression of SOCS-1 in b-cells and discussion and reviewed and edited the manuscript. E.N.G. researched inhibits IFN signaling and protects NOD mice from insu- data, contributed to discussion, designed experiments, and reviewed, edited, litis and diabetes (5). Interestingly, one of the PTPs oxi- and wrote the manuscript. E.N.G. is the guarantor of this work and, as such, dized in the NOD pancreas is PTPN2, the inactivation of 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. which has been shown in previous studies to enhance STAT1 phosphorylation and sensitize b-cells to apoptosis References induced by IFNs (8,21,22). Consistent with our data, si- 1. Eizirik DL, Colli ML, Ortis F. The role of inflammation in insulitis and beta- lencing of the proapoptotic molecule Bim prevents b-cell cell loss in type 1 diabetes. Nat Rev Endocrinol 2009;5:219–226 death induced by knockdown of PTPN2 and IFN treat- 2. Gurzov EN, Eizirik DL. Bcl-2 proteins in diabetes: mitochondrial pathways of ment (22). In humans, genome-wide association studies b-cell death and dysfunction. Trends Cell Biol 2011;21:424–431 have linked PTPN2 polymorphisms with T1D (27). Thus, 3. Kim S, Kim HS, Chung KW, et al. Essential role for signal transducer and our results establish an important novel mechanism of activator of transcription-1 in pancreatic beta-cell death and autoimmune type 1 diabetes of nonobese diabetic mice. Diabetes 2007;56:2561–2568 inactivation of T1D candidate genes by oxidative stress fi 4. Gysemans CA, Ladrière L, Callewaert H, et al. Disruption of the gamma- during insulitis. Indeed, overexpression of speci c PTPs interferon signaling pathway at the level of signal transducer and activator of (e.g., PTPN2) might be a valuable strategy to protect transcription-1 prevents immune destruction of beta-cells. Diabetes 2005;54: fi b-cells against immune in ltration. However, if the phos- 2396–2403 phatase active site is inactivated by oxidative stress, as we 5. Chong MM, Chen Y, Darwiche R, et al. Suppressor of cytokine signaling-1 demonstrate in the current study, combined therapies to overexpression protects pancreatic beta cells from CD8+ T cell-mediated auto- overexpress PTPs and prevent ROS formation might be immune destruction. J Immunol 2004;172:5714–5721 2496 PTP Oxidation Promotes Inflammation Diabetes Volume 64, July 2015

6. Ostman A, Frijhoff J, Sandin A, Böhmer FD. Regulation of protein tyrosine 18. French MB, Allison J, Cram DS, et al. Transgenic expression of mouse phosphatases by reversible oxidation. J Biochem 2011;150:345–356 proinsulin II prevents diabetes in nonobese diabetic mice. Diabetes 1997;46: 7. Tiganis T, Bennett AM. Protein tyrosine phosphatase function: the substrate 34–39 perspective. Biochem J 2007;402:1–15 19. Reddy S, Bradley J. Immunohistochemical demonstration of nitrotyrosine, 8. Gurzov EN, Stanley WJ, Brodnicki TC, Thomas HE. Protein tyrosine phos- a biomarker of oxidative stress, in islet cells of the NOD mouse. Ann N Y Acad Sci phatases: molecular switches in metabolism and diabetes. Trends Endocrinol 2004;1037:199–202 Metab 2015;26:30–39 20. Persson C, Kappert K, Engström U, Ostman A, Sjöblom T. An antibody- 9. Gurzov EN, Tran M, Fernandez-Rojo MA, et al. Hepatic oxidative stress based method for monitoring in vivo oxidation of protein tyrosine phosphatases. promotes insulin-STAT-5 signaling and obesity by inactivating protein tyrosine Methods 2005;35:37–43 phosphatase N2. Cell Metab 2014;20:85–102 21. Moore F, Colli ML, Cnop M, et al. PTPN2, a candidate gene for type 1 di- 10. Maxwell SR, Thomason H, Sandler D, et al. Antioxidant status in patients abetes, modulates interferon-gamma-induced pancreatic beta-cell apoptosis. with uncomplicated insulin-dependent and non-insulin-dependent diabetes Diabetes 2009;58:1283–1291 mellitus. Eur J Clin Invest 1997;27:484–490 22. Santin I, Moore F, Colli ML, et al. PTPN2, a candidate gene for type 1 di- 11. Rocic B, Vucic M, Knezevic-Cuca J, et al. Total plasma antioxidants in first- abetes, modulates pancreatic b-cell apoptosis via regulation of the BH3-only degree relatives of patients with insulin-dependent diabetes. Exp Clin Endocrinol protein Bim. Diabetes 2011;60:3279–3288 Diabetes 1997;105:213–217 23. Suk K, Kim S, Kim YH, et al. IFN-gamma/TNF-alpha synergism as the final 12. Lenzen S. Oxidative stress: the vulnerable beta-cell. Biochem Soc Trans effector in autoimmune diabetes: a key role for STAT1/IFN regulatory factor-1 2008;36:343–347 pathway in pancreatic beta cell death. J Immunol 2001;166:4481–4489 13. Campbell PD, Weinberg A, Chee J, et al. Expression of pro- and anti- 24. Moore F, Naamane N, Colli ML, et al. STAT1 is a master regulator of apoptotic molecules of the Bcl-2 family in human islets postisolation. Cell pancreatic beta-cell apoptosis and islet inflammation. J Biol Chem 2011;286: Transplant 2012;21:49–60 929–941 14. McKenzie MD, Dudek NL, Mariana L, et al. Perforin and Fas induced by 25. Merry TL, Tran M, Stathopoulos M, et al. High-fat-fed obese glutathione IFNgamma and TNFalpha mediate beta cell death by OT-I CTL. Int Immunol 2006; peroxidase 1-deficient mice exhibit defective insulin secretion but protection from 18:837–846 hepatic steatosis and liver damage. Antiox Redox Signal 2014;20:2114–2129 15. Gurzov EN, Ortis F, Cunha DA, et al. Signaling by IL-1beta+IFN-gamma and 26. Weaver JR, Holman TR, Imai Y, et al. Integration of pro-inflammatory cy- ER stress converge on DP5/Hrk activation: a novel mechanism for pancreatic tokines, 12-lipoxygenase and NOX-1 in pancreatic islet beta cell dysfunction. Mol beta-cell apoptosis. Cell Death Differ 2009;16:1539–1550 Cell Endocrinol 2012;358:88–95 16. Ichii H, Inverardi L, Pileggi A, et al. A novel method for the assessment of 27. Barrett JC, Clayton DG, Concannon P, et al.; Type 1 Diabetes Genetics cellular composition and beta-cell viability in human islet preparations. Am J Consortium. Genome-wide association study and meta-analysis find that over 40 Transplant 2005;5:1635–1645 loci affect risk of type 1 diabetes. Nat Genet 2009;41:703–707 17. McKenzie MD, Jamieson E, Jansen ES, et al. Glucose induces pancreatic 28. Wiede F, Ziegler A, Zehn D, Tiganis T. PTPN2 restrains CD8⁺ T cell re- islet cell apoptosis that requires the BH3-only proteins Bim and Puma and multi- sponses after antigen cross-presentation for the maintenance of peripheral tol- BH domain protein Bax. Diabetes 2010;59:644–652 erance in mice. J Autoimmun 2014;53:105–114