Role of Protein Phosphatase 1 and Inhibitor of Protein Phosphatase 1 in Nitric Oxide–Dependent Inhibition of the DNA Damage Response in Pancreatic β-Cells

Type 1 (insulin-dependent) diabetes is characterized by islet inflammation and autoimmune-mediated destruction of the insulin-secreting pancreatic β-cells (1). β-Cells respond to proinflammatory cytokines that are released during islet inflammation, such as interleukin-1 (IL-1), interferon γ (IFNγ), and tumor necrosis factor α (TNFα), with expression of inducible nitric oxide synthase (iNOS) and production of nitric oxide (2–4). Nitric oxide has long been implicated as one potential mediator of β-cell damage and destruction during the development of autoimmune diabetes, as it is responsible for the inhibition of glucose-stimulated insulin secretion and mitochondrial oxidative metabolism and the induction of DNA damage and, after prolonged exposure to cytokines, nitric oxide is responsible for β-cell death (2–10). Despite these inhibitory actions, nitric oxide also plays protective roles, including the activation of pathways that promote the recovery of oxidative metabolism, insulin secretion, the repair of damaged DNA, and the attenuation of β-cell apoptosis (11–16). These studies highlight the complex and dichotomous role that nitric oxide plays in controlling the response of β-cells to proinflammatory cytokines.

The DNA damage response (DDR) is a signaling pathway activated after DNA double-strand breaks (DSBs) that functions to promote cell cycle arrest and the repair of damaged DNA (17). If DNA repair fails, or DNA damage is too severe for repair, the DDR initiates a proapoptotic signaling cascade resulting in cell death (18). Phosphorylation of histone H2AX on Ser139 (termed γH2AX when phosphorylated) by DDR kinases such as ataxia telangiectasia mutated (ATM) is a rapid event that is considered to be one of the most sensitive markers of DDR activation and DSB formation (19–21). DDR activation has been implicated as a lethal event in cytokine-induced β-cell death, as prolonged cytokine treatment results in nitric oxide–dependent γH2AX formation and ATM-dependent caspase activation (22).

Despite its genotoxic nature, nitric oxide also inhibits DDR activation (16). The phosphorylation of DDR substrates H2AX, KAP1, and p53 in response to genotoxic agents is attenuated by nitric oxide, despite the presence of DNA damage (16). The ability of nitric oxide to inhibit DDR activation occurs at micromolar levels of nitric oxide and is selective for pancreatic β-cells (16). The net result of DDR inhibition by nitric oxide is the suppression of DNA damage–dependent β-cell apoptosis (16). In this study, the ability of oxidants and radicals in addition to nitric oxide to modulate DDR activity and the mechanisms responsible for the selective inhibition of DDR signaling by nitric oxide in pancreatic β-cells were explored. The kinetics of inhibition and the wide range of DDR substrates affected by nitric oxide suggest that phosphatase activation may be responsible for suppressing DDR signaling in the presence of nitric oxide; however, the cell type selectivity of this inhibition indicates that the phosphatase must be selectively regulated in β-cells. Recently, an inhibitor of protein phosphatase 1 (IPP1) was identified to be selectively expressed in β-cells (23). Protein phosphatase 1 (PP1), the phosphatase controlled by IPP1, is a primary regulator of DDR signaling via the dephosphorylation of transducing kinases (such as ATM and ataxia telangiectasia and Rad3-related protein [ATR]) and DDR substrates (H2AX, KAP1, and p53) (24,25). In this study, we explored the potential role of IPP1 and PP1 as potential targets responsible for mediating the inhibitory actions of nitric oxide on DDR signaling.

Nitric oxide donor (Z)-1-[N-(3-aminopropyl)-N-(3-ammoniopropyl)amino]diazen-1-ium-1,2-diolate (DPTA/NO), SIN-1, and actinomycin D were purchased from Cayman Chemical (Ann Arbor, MI). DPTA/NO was dissolved in 10 mmol/L NaOH prior to use. H₂O₂, camptothecin, menadione, cycloheximide, and tunicamycin were purchased from Sigma-Aldrich (St. Louis, MO). Thapsigargin was purchased from Enzo (Farmingdale, NY). KU-55933 was purchased from EMD Millipore (Billerica, MA). Calyculin A was purchased from Cell Signaling Technology (Beverly, MA). Antibodies used and their sources are as follows: mouse anti–phosphorylated (p-)H2AX (γH2AX, Ser139) and rabbit anti-glucagon (EMD Millipore); rabbit anti–p-KAP1 (Ser824) and rabbit anti-IPP1 (Abcam, Cambridge, MA); mouse anti-GAPDH (Thermo Fisher Scientific, Waltham, MA); mouse antip53, rabbit anti–p-Akt (Ser473), mouse anti-H3, and rabbit anti–p-eIF2α (Ser51; Cell Signaling Technology); guinea pig anti-insulin (DakoCytomation, Carpinteria, CA); mouse anti-PP1 (Santa Cruz Biotechnology, Dallas, TX); and horseradish peroxidase–conjugated donkey anti-rabbit and horseradish peroxidase–conjugated donkey anti-mouse (Jackson ImmunoResearch Laboratories, West Grove, PA).

Islets were isolated from male Sprague-Dawley rats by collagenase digestion and cultured as previously described (26). The insulinoma cell line INS 832/13 was obtained from Chris Newgard (Duke University, Durham, NC). Human embryonic kidney (HEK)293 and αTC3 cells were obtained from American Type Culture Collection. MIN6 insulinoma cells were obtained from Washington University Tissue Culture Support Center. Mouse embryonic fibroblasts (MEFs) were obtained from Fumihiko Urano (Washington University, St. Louis, MO). EndoC-βH1 cells were obtained from Raphael Scharfmann (Paris Descartes University, Paris, France) (27). All cells were cultured as previously described (16) and detached from culture plates using 0.05% trypsin in 0.53 mmol/L EDTA and plated at a density of 250,000 cells/mL (MEF and HEK293) or 500,000 cells/mL (INS 832/13, MIN6, αTC3, and EndoC-βH1) unless otherwise specified. RPMI 1640 medium, DMEM, minimum essential medium α, l-glutamine, sodium pyruvate, HEPES, penicillin, streptomycin, β-mercaptoethanol, and trypsin (0.05% in 0.53 mmol/L EDTA) were purchased from Thermo Fisher Scientific.

SDS-PAGE electrophoresis and Western blot analysis were performed as previously described (15,28). Dilutions of primary antibodies were as follows: γH2AX (1:5,000), GAPDH (1:10,000), p-KAP1 (1:2,000), p-eIF2α (1:1,000), IPP1 (1:10,000), H3 (1:2,000), p-Akt (1:2,000), PP1 (1:100), and p53 (1:1,000). Detection was by chemiluminescence (29).

All transfections were performed using Lipofectamine 2000 (Thermo Fisher Scientific) per manufacturer’s instructions. Control-A small interfering (si)RNA and siRNA targeting IPP1 (used at 100 nmol/L) were purchased from Integrated DNA technologies (Coralville, IA). Pan-PP1 siRNA was purchased from Santa Cruz Biotechnology. After 24 h incubation, transfection media was removed and fresh media was added to the wells. INS 832/13 cells were allowed to incubate in fresh media overnight prior to further treatment. IPP1 plasmid was obtained from Addgene (category no. 31331 [30]) and cloned into a pcDNA3.1 vector. Plasmids transfections were allowed to proceed for 24 h prior to experimentation.

Islet cells were fixed using 4% paraformaldehyde for 15 min, permeabilized with 0.1% Triton X-100 for 2 min on ice, and then blocked using 1% BSA in PBS for 30 min. Primary antibodies to γH2AX, IPP1, insulin, and glucagon were used at 1:500, 5 µg/mL, 1:100, and 1:100, respectively. Alexa Fluor 488–conjugated donkey anti-rabbit, Alexa Fluor 488–conjugated donkey anti-guinea pig, Cy3-conjugated donkey anti-mouse, and Cy3-conjugated donkey anti-rabbit secondary antibodies were used at 1:400. Hoechst 33342 was used as a nuclear stain (2 µg/mL in PBS for 10 min). Images were captured using a Nikon Eclipse 90i confocal microscope. γH2AX-positive cells were quantified using ImageJ (National Institutes of Health). Cell death was assessed using the SYTOX Green nucleic acid stain as previously described (16).

Statistical analysis was performed using one-way ANOVA with a Tukey-Kramer post hoc test. The minimum level of significance was P < 0.05. Statistical results are further defined in the figure legends.

Camptothecin stimulates the phosphorylation of DDR substrates KAP1 and H2AX in multiple cell types (Fig. 1). In each insulinoma cell line—INS 832/13 (rat), EndoC-βH1 (human), and MIN6 (mouse)—DDR signaling is impaired by the nitric oxide donor DPTA/NO in a concentration-dependent manner (Fig. 1A–C). DPTA/NO is not effective at inhibiting camptothecin-induced DDR signaling in non-β-cell lines such as MEFs and HEK293 cells (Fig. 1D and E) or mouse α-cells (αTC3) (Fig. 1F). Further, nitric oxide induces DNA strand breaks (31,32) and alone stimulates KAP1 phosphorylation and γH2AX formation in MEFs, HEK293 cells, and αTC3 cells, but not in INS 832/13, EndoC-βH1, or MIN6 cells (Fig. 1). Much like treatment with donors, nitric oxide produced endogenously in INS 832/13 after an 18-h incubation with IL-1 + IFNγ attenuates camptothecin-induced γH2AX and KAP1 phosphorylation (Fig. 2A–C). The inhibitory actions of endogenously produced nitric oxide are attenuated by the NOS inhibitor N^G-monomethyl-l-arginine (l-NMMA) (Fig. 2A–C). Consistent with these findings, we have shown that nitric oxide impairs DDR signaling in cytokine-treated rat islets (16). For determination of whether this inhibition is restricted to β-cells and not other islet endocrine cells such as α-cells, dispersed rat islet cells were pretreated for 18 h with IL-1 and IFNγ to stimulate nitric oxide production and then exposed to H₂O₂ for 30 min to induce DNA damage. The presence of γH2AX in insulin- and glucagon-containing cells was determined by immunofluorescence (Fig. 2D and E). Cytokine pretreatment reduces H₂O₂-induced γH2AX formation in insulin-positive β-cells by >50%, while it has no effect on γH2AX formation in glucagon-positive cells (Fig. 2F). Since β-cells are the islet cellular source of iNOS after cytokine treatment (33) these findings provide additional evidence that the endogenous production of nitric oxide selectively impairs DDR signaling in β-cells.

Reactive oxygen species (ROS) and nitrogen species (RNS) activate distinct signaling pathways in β-cells (34,35). For determination of whether other forms of ROS or RNS suppress DDR activation, INS 832/13 cells were treated for 1 h with camptothecin in the presence or absence of DPTA/NO, SIN-1 (peroxynitrite donor), menadione (superoxide generator), or H₂O₂ (Fig. 3). Nitric oxide (DPTA/NO) inhibits (Fig. 3A and E) while peroxynitrite (SIN-1) and superoxide (menadione) do not modify camptothecin-induced KAP1 phosphorylation and γH2AX formation (Fig. 3), Alone, DPTA/NO and peroxynitrite fail to stimulate DDR activation. Menadione stimulates DDR activation only at the concentration of 60 μmol/L (Fig. 3C) (34). H₂O₂ also stimulates DDR activation yet does not impair camptothecin-induced DDR activation at low concentrations (Fig. 3C–E). At high concentrations of 200–400 μmol/L, H₂O₂ decreases camptothecin-induced γH2AX formation and KAP1 phosphorylation, although at these concentrations it is also highly toxic to INS 832/13 cells (34,36). Thus, DDR inhibition at these high concentrations is likely a consequence of cell death (Fig. 3D and E). These data suggest that the DDR is selectively inhibited by nitric oxide and not by other ROS/RNS.

The inhibition of DDR signaling by nitric oxide, but not other forms of ROS and RNS, suggests that the pathways responsible for DDR inhibition are activated selectively by nitric oxide. Through the inhibition of the sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA) channel, nitric oxide has been shown to stimulate the unfolded protein response (UPR) (12,37,38), while other forms of ROS and RNS fail to stimulate UPR activation in β-cells (12,34,35). Also, UPR activation in response to nitric oxide inversely correlates with DDR activation, as DDR signaling is suppressed under conditions in which nitric oxide stimulates eukaryotic translation initiation factor (eIF)2α phosphorylation (16). We have shown that nitric oxide and activators of endoplasmic reticulum (ER) stress cause adaptive changes in β-cells that limit extracellular signal transduction (39,40). Specifically, IL-1–induced NF-κB activation and MAPK phosphorylation and IFNγ-induced STAT-1 phosphorylation are attenuated in β-cells exposed to ER stressors and nitric oxide (39,40). Based on these associations, we tested the hypothesis that nitric oxide may inhibit DDR activation through an ER stress–induced pathway. INS 832/13 cells were treated for 2 h with ER stress–inducing agent tunicamycin or thapsigargin in the presence or absence of camptothecin, and UPR activation was confirmed by examining eIF2α phosphorylation (Fig. 4). As expected, tunicamycin, thapsigargin and nitric oxide all stimulate eIF2α phosphorylation, while camptothecin does not (Fig. 4). While all three ER stress inducers activate the UPR, it is only nitric oxide that inhibits camptothecin-induced KAP1 phosphorylation and γH2AX formation (Fig. 4). Further, INS 832/13 cells treated with DPTA/NO are protected from thapsigargin-induced cell death, suggesting that nitric oxide–induced cell death may not be mediated by ER stress pathways (Fig. 4C). These findings disassociate UPR activation from the ability of nitric oxide to inhibit DDR signaling.

Phosphorylation of eIF2α prevents the exchange of guanosine diphosphate for guanosine triphosphate necessary for the progression of the translational initiation complex, and as such, this posttranslational modification effectively limits translation (41). As nitric oxide–dependent inhibition of the DDR correlates with phosphorylation of eIF2α and, presumably, the inhibition of translation, we explored whether inhibition of DDR signaling required new gene expression or protein synthesis. To testing of this hypothesis, INS 832/13 cells were pretreated with the transcriptional inhibitor actinomycin D and then treated with camptothecin in the presence or absence of DPTA/NO for 2 h. Actinomycin D inhibits transcription via intercalation into DNA and stalling of RNA polymerase (42), and like camptothecin, actinomycin D alone induces DDR activation as evidenced by KAP1 phosphorylation and γH2AX formation (43) (Fig. 4D). Nitric oxide inhibits DDR activation in response to camptothecin and actinomycin D (Fig. 4D and E), indicating that inhibition of DDR signaling by nitric oxide occurs independently of new gene transcription. Further, the protein synthesis inhibitor cycloheximide (CHX) does not prevent the inhibitory actions of nitric oxide on camptothecin-induced phosphorylation of KAP1 and H2AX (Fig. 4F and G). As a control, we show that CHX prevents p53 accumulation in camptothecin-treated INS 832/13 cells (consistent with the inhibition of protein synthesis) (44). These data indicate that nitric oxide suppresses DDR signaling by a mechanism that is independent of both new gene transcription and protein synthesis.

The temporal effects of nitric oxide and the ATM kinase inhibitor KU-55933 on phosphorylation of KAP1 and H2AX were examined by adding DPTA/NO and KU-55933 to INS 832/13 cells pretreated for 1 h with camptothecin, followed by determination of DDR activation. In a time-dependent manner, DPTA/NO decreases the phosphorylation of KAP1 and H2AX, with more than a 60% decrease after 2 h of incubation. Nitric oxide is responsible for the loss in phosphorylation of KAP1 and H2AX, as phosphorylation levels are not decreased after 3 h incubation with camptothecin alone (Fig. 5A and C). In contrast, camptothecin-induced KAP1 phosphorylation is rapidly lost in response to KU-55933 (maximal at 30 min) (Fig. 5B and D), while ATM inhibition does not decrease camptothecin-induced γH2AX formation (Fig. 5B and D). Importantly, the phosphorylation of KAP1 on Ser824 is ATM selective (45), while H2AX can be phosphorylated on Ser139 by multiple kinases (e.g., ATR and DNA-dependent protein kinase [DNA-PK]) (20,46,47). These findings indicate that the inhibitory actions of nitric oxide on the DDR are not limited to ATM-dependent phosphorylation events but, rather, extend to additional DDR kinases such as ATR and DNA-PK.

The regulation of DDR signaling is controlled, in part, by the actions of protein phosphatases (24,48). Since the inhibitory actions of nitric oxide on DDR activation are selective for β-cells, the possibility that β-cells selectively express a phosphatase or pathway leading to phosphatase activation that suppresses DDR signaling was explored. Jiang et al. (23) identified IPP1 as a protein selectively expressed in β-cells. IPP1 is a negative regulator of PP1, the primary regulatory phosphatase of the DDR with substrates that include ATM, p53, KAP1, and γH2AX (49–52). Using the Genevestigator database (53) to identify phosphatase-related genes that are differentially expressed in β-cells, we identified Ppp1r1a (which encodes IPP1 as a potential regulator of phosphatase activity that is selectively expressed in β-cells). Out of 292 different mouse tissues and cell lines examined, the insulinoma cell lines R7T1 and MIN6 were found to have the highest expression of Ppp1r1a (Fig. 6A). Similar observations were made in rat, as pancreatic islets have been found to contain the second-highest and INS-1 insulinoma cells the ninth-highest levels of Ppp1r1a expression out of a total of 101 different rat tissues and cell lines that have been evaluated (data not shown). At the protein level, IPP1 is highly expressed in insulinoma cell lines INS 832/13, MIN6, and RINm5F and mouse islets compared with the levels expressed in MEFs, RAW264.7 macrophages, MCF-7 cells, HepG2 hepatocytes, HEK293 cells, and the islet α-cell line αTC3 (Fig. 6B). IPP1 expression appears to localize exclusively with insulin-containing β-cells in rat islets (Fig. 6C).

Since the inhibition of DDR signaling by nitric oxide is selective for β-cells, and β-cells selectively express IPP1, an inhibitor of PP1, we hypothesized that nitric oxide may disrupt the inhibition of PP1 by IPP1, allowing PP1 to catalyze the dephosphorylation of DDR substrates. Two approaches were used to test this hypothesis. First, INS 832/13 cells were mock transfected or transfected with siRNA targeting IPP1 and then treated for 2 h with camptothecin in the presence or absence of DPTA/NO, and DDR activation was examined. Camptothecin-induced KAP1 phosphorylation and γH2AX formation are attenuated by DPTA/NO (Fig. 7A–C) in control cells or to levels similar to the levels of inhibition observed in INS832/13 cells deficient in IPP1 (70% knockdown) (Fig. 7D). Camptothecin-induced phosphorylation of KAP1 and H2AX are similarly attenuated by nitric oxide during shorter exposures to camptothecin (15 and 30 min) (Fig. 7E). Further, overexpression of IPP1 in HEK293 cells that do not express this inhibitor of PP1 does not modify the response of these cells to camptothecin or nitric oxide (Fig. 7F and G). Despite its selective expression in β-cells, these findings suggest that IPP1 does not participate in the inhibition of DDR signaling by nitric oxide.

The phosphorylation of DDR proteins is a dynamic process that involves a balance between phosphorylation by kinases and dephosphorylation by a number of serine/threonine phosphatases (24,25). As shown in Fig. 5B, inhibition of ATM after camptothecin treatment leads to rapid dephosphorylation of its substrate KAP1 within 30 min. PP1 is a phosphatase known to regulate the phosphorylation status of a number of DDR substrates, including KAP1 and H2AX (25,49,52,54). For determination of whether PP1 participates in the regulation of the DDR by nitric oxide, INS 832/13 cells were treated with camptothecin in the presence or absence of DPTA/NO and the PP1 inhibitor calyculin A (55). In addition to PP1, this phosphatase inhibitor has also been shown to inhibit PP2A, PP4, and PP6 (55). At concentrations of 50 and 100 nmol/L, calyculin A alone increased Akt and KAP1 phosphorylation, and the formation of γH2AX (Fig. 8A), suggesting that PP1 (as well as other phosphatases) is constitutively active and limits DDR signaling under basal conditions. Importantly, calyculin A does not modify the inhibitory actions of nitric oxide on camptothecin-induced KAP1 phosphorylation (lower band) and γH2AX formation in response to camptothecin (Fig. 8A and B). Further, nitric oxide suppresses DDR signaling in response to calyculin A alone (Fig. 8A, last lane). These findings indicate that PP1 activity is not required for the inhibition of DDR signaling in β-cells exposed to nitric oxide. To complement these pharmacological approaches, we transfected INS 832/13 cells with siRNA designed to target all isoforms of rat PP1 or scramble control siRNA (Fig. 8C); the INS 832/13 cells were then treated with DPTA/NO, camptothecin, or camptothecin and DPTA/NO for 1 h. While PP1 knockdown (∼70% knockdown of PP1 protein) increased basal levels of γH2AX, it does not modify the inhibitory actions of nitric oxide on camptothecin-induced γH2AX formation (Fig. 8C–E). These findings suggest that nitric oxide–dependent inhibition of the DDR does not require the activation of these Ser/Thr phosphatases known to regulate DDR signaling.

The DDR responds to DSBs with the activation of signaling pathways that promote cell cycle arrest and initiate DNA repair; however, when DNA damage is too severe for repair, it activates pathways that result in apoptotic death or cell senescence (17). We have shown that DDR signaling is inhibited by the presence of micromolar levels of nitric oxide in a cell type–selective manner (16). The inhibition of this signaling pathway occurs in the presence of DNA damage, indicating that nitric oxide uncouples DSB formation from the activation of DDR signaling (16). The inhibition of DDR signaling by nitric oxide is selective for pancreatic β-cells (Figs. 1 and 2) and appears to protect these cells from DNA damage–dependent apoptosis (16). Nitric oxide fails to suppress DDR signaling in all other cell types that we have examined to date (16) (Figs. 1 and 2). In this study, the mechanisms responsible for nitric oxide–dependent inhibition of the DDR were explored.

While cell signaling in response to ROS and RNS is oxidant dependent in β-cells (34,35), oxidants in addition to nitric oxide do not impair DDR signaling in β-cells (Fig. 3). These findings indicate that pathways selectively activated by nitric oxide are responsible for suppression of DDR signaling in β-cells. Nitric oxide causes ER stress and activation of the UPR in β-cells via the inhibition of SERCA channels (37), while oxidants such as hydrogen peroxide, peroxynitrite, and superoxide do not activate this pathway (34,35). We show that activation of the UPR (evident by phosphorylation of eIF2α) in nitric oxide–treated β-cells correlates with inhibition of the DDR (16) (Fig. 4), and UPR activators have been shown to attenuate DNA damage–induced apoptosis in response to camptothecin and etoposide in hepatocellular carcinoma 3B cells (56). While these findings are similar to the protective actions of nitric oxide in preventing camptothecin-induced β-cell death (16), UPR activation does not appear to participate in the inhibition of DDR signaling by nitric oxide in β-cells. The ER stressors tunicamycin and thapsigargin do not modify camptothecin-induced phosphorylation of KAP1 or the formation of γH2AX in β-cells, and nitric oxide affords protection from ER stress–induced cell death (Fig. 4C), disassociating UPR activation from the regulation of DDR signaling by nitric oxide. Further, inhibitors of transcription and translation have no effect on nitric oxide–mediated inhibition of DDR signaling, indicating that effects of nitric oxide are mediated by modulating signaling cascades and do not require new gene transcription or protein synthesis (Fig. 4D–F).

The balance of phosphorylation catalyzed by phosphatidylinositol 3-kinase–related kinases (PIKK) ATM, ATR, and DNA-PK (57) and dephosphorylation by protein phosphatases such as PP1 controls the magnitude of DDR signaling (24,48,58). As such, DDR substrate dephosphorylation in the presence of nitric oxide may be due to inhibition of kinase activity, stimulation of phosphatase activity, or a combination of both. Inhibitors of ATM stimulate a rapid decrease in KAP1 phosphorylation (an ATM-selective substrate) without modifying γH2AX formation, while nitric oxide decreases γH2AX formation and KAP1 phosphorylation. These findings indicate that the inhibitory actions of nitric oxide on DDR signaling are not limited to ATM-dependent pathways but, rather, extend to other PIKK-mediated signaling cascades such as ATR and DNA-PK. Consistent with this hypothesis, we have shown that nitric oxide inhibits PIKK signaling in other pathways, including Akt (a substrate of phosphatidylinositol-4,5-bisphosphate 3-kinase [PI3K]) (15) and p70S6K (a substrate of mammalian target of rapamycin [mTOR]) (59).

The discordance in the kinetics of KAP1 dephosphorylation in response to nitric oxide (half-maximal inhibition ∼60 min) when compared with the ATM inhibitor KU-55933 (∼10 min) (Fig. 5) suggest that either nitric oxide is less effective than ATM inhibitors at decreasing kinase activity or that nitric oxide increases phosphatase activity. Further, if the inhibitory actions of nitric oxide extend to PIKK kinases in addition to ATM, and kinase inhibition is the mechanism responsible for DDR inhibition, then it would be necessary for ATR and DNA-PK to be inhibited by nitric oxide with similar kinetics. While possible, this explanation is unlikely given the unique mechanisms that regulate the activation of each of these PIKKs (57). Also, the inhibitory actions of nitric oxide on DDR signaling are selective for pancreatic β-cells, while all cell types express PIKK kinases (60), further suggesting that inhibition of kinase activity is not a likely mechanism of action for nitric oxide. Because of these findings, we explored the possibility that β-cells may selectively express a phosphatase or protein that modifies phosphatase activity that allows for this β-cell–selective response.

Phosphatase regulation of DDR signaling is accomplished primarily through activity of the phosphoprotein phosphatase P (PPP) family, including phosphatases such as PP1, PP2A, PP4, PP5, and PP6 (24,25). These enzymes have broad substrate specificity and dephosphorylate multiple proteins involved in the DDR signaling cascade (24,25). Based on the cell type–selective nature of DDR inhibition by nitric oxide, we hypothesized that β-cells may selectively express a phosphatase or phosphatase-related protein that enables this regulation. Jiang et al. (23) identified IPP1 to be highly enriched in β-cells compared with other cell types and suggested that it could be an effective biomarker for β-cells. We confirm these findings of β-cell–selective expression and show that IPP1 is not expressed in cell types in which nitric oxide fails to inhibit DDR signaling (16) (Fig. 6). IPP1 is a negative regulator of PP1, a phosphatase with substrate specificity that includes DDR pathway members ATM, KAP1, H2AX, and p53 (49–52). Despite this correlation, IPP1 does not appear to participate in the inactivation of DDR signaling in response to nitric oxide. Knockdown using siRNA in INS 832/13 cells does not modify the inhibitory actions of nitric oxide on camptothecin-induced DDR signaling (Fig. 7A–E). Further, overexpression of IPP1 in cells that are not responsive to nitric oxide (HEK293 cells) and lack IPP1 does not sensitize these cells to DDR inhibition by nitric oxide (Fig. 7F and G). Also, members of the PPP family (PP1, PP2A, PP4, PP5, and PP6), which are considered to be the primary regulators of DDR signaling (24), do not appear to participate in the inhibitory actions of nitric oxide on DDR signaling in β-cells. While knockdown of PP1 increases basal levels of γH2AX, consistent with the role of PP1 in the regulation of this protein, it does not prevent the loss of γH2AX that occurs in response to nitric oxide (Fig. 8C–E). Also, nitric oxide suppresses DDR signaling in the presence of calyculin A, an inhibitor of PP1, PP2A, PP4, and PP6 (55).

In this study, evidence is presented to show that new gene expression, protein synthesis, UPR activation, and PP1 activity do not participate in the regulation of DDR activity by nitric oxide. While the precise mechanism remains elusive, the pathways by which nitric oxide regulates DDR activity are selective for pancreatic β-cells, as nitric oxide has failed to attenuate DDR signaling in all other cell types examined to date. Furthermore, the pathways are controlled by nitric oxide and not other forms of ROS/RNS such as superoxide, peroxynitrite, or hydrogen peroxide. It is tempting to speculate that nitric oxide modifies DDR signaling via S-nitrosation and inhibition of the apical DDR kinases ATM, ATR, and DNA-PK; indeed, DNA-PK has been shown to undergo S-nitrosation and inhibition after S-nitrosocysteine treatment and in cells overexpressing nNOS (61,62). Unfortunately, much like direct kinase inhibition, this type of mechanism falls short of explaining the cell type selectivity of the response.

It has long been known that treatment of islets (rodent and human) with proinflammatory cytokines such as IL-1, IFNγ, and TNFα results in the inhibition of function and destruction of islets and β-cells (5,63). Nitric oxide, generated via iNOS in response to proinflammatory cytokines, is the primary mediator of the damaging actions of cytokines on β-cell function and viability (2,4,5). In addition to causing damage, we have shown that when β-cells actively produce nitric oxide at micromolar levels, they are resistant to apoptosis (16), a finding counter to studies suggesting that cytokines direct β-cell apoptosis (63–66). This finding also challenges the hypothesis that cytokines are solely damaging to β-cells and suggests that the response of β-cells to cytokines may be physiological, potentially protecting β-cells from insults that result in the loss of viability. In support of a physiological protective role for β-cell responses to cytokines, islets have the ability to completely recover from the inhibitory actions of a 15-h incubation with IL-1 by simply removing the cytokine from the culture by washing (67). Further, nitric oxide stimulates the expression of GADD45α and the induction of a DNA repair response (13), the activation of the UPR (12,37), induction of the heat shock response (68), and activation of pathways that limit the response of β-cells to extracellular signaling such as cytokine stimulation (39,40). Together with the identification of DDR inhibition, these findings provide experimental evidence suggesting that the response of β-cells to cytokines may be far more complicated than simply the activation of cell death pathways and that this response may include the activation of protective pathways that promote β-cell function and viability. Additional studies are necessary to identify the mechanisms by which nitric oxide selectively protects β-cells from apoptosis associated with DNA damage; however, it is hoped that this information will begin to define, at a mechanistic level, the physiological reasons why β-cells are the only endocrine cell type in islets to produce nitric oxide in response to cytokines and the role of this nitric oxide production in supporting β-cell function and viability.

Acknowledgments. The authors thank Jennifer A. McGraw (Department of Biochemistry, Medical College of Wisconsin, Milwaukee, WI) for technical assistance and Drs. Polly Hansen and Kasia Broniowska (Department of Biochemistry, Medical College of Wisconsin, Milwaukee, WI) for helpful discussions related to this project. A gift to support research was provided by Scott Tilton.

Funding. This work was supported by the following National Institutes of Health grants: National Institute of Diabetes and Digestive and Kidney Diseases grant DK-052194 and Division of Intramural Research, National Institute of Allergy and Infectious Diseases, grant AI-044458 (to J.A.C.). B.J.O. was supported by an American Heart Association predoctoral fellowship (14PRE20380585).

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

Author Contributions. B.J.O., A.N., and J.A.C. were responsible for the conception and design of the research. B.J.O., A.N., S.C.P., and C.T.Y. performed the experiments. B.J.O., A.N., S.C.P., C.T.Y., and J.A.C. analyzed the data and interpreted the results of the experiments. B.J.O. and J.A.C. prepared the figures and drafted the manuscript. B.J.O., A.N., S.C.P., C.T.Y., and J.A.C. edited, revised, and approved the final version of the manuscript. B.J.O. and J.A.C. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.