Protection of RPE Cells from Oxidative Injury by 15-Deoxy-Δ^12,14-Prostaglandin J₂ by Augmenting GSH and Activating MAPK

Abstract

purpose. The goal of this study was to identify the mechanisms by which 15-deoxy-Δ^12,14-prostaglandin J₂ (dPGJ₂) protects RPE cells from oxidative injury.

methods. Cell viability was determined by MTT assay. Protein expression and activation of signaling molecules were detected by Western blot. Reduced glutathione (GSH) was determined by a colorimetric assay kit. PPARγ expression was knockdown by small interfering (si)RNA technique.

results. dPGJ₂ protected ARPE19 cells from oxidative injury, whereas the synthetic PPARγ agonists AGN195037 and rosiglitazone had no effect. PPARγ knockdown also did not affect dPGJ₂’s protective activity. dPGJ₂ upregulated GSH synthesis via induction of glutamylcysteine ligase. GSH depletion sensitized cells to oxidative stress and completely reversed the protective effect of dPGJ₂. dPGJ₂ activated ERK, JNK, and p38; GSH induction by dPGJ₂ depended partially on JNK and p38. In addition, dPGJ₂ significantly extended hydrogen peroxide–induced activation of JNK and p38, but not of Akt. Inhibition of MEK, JNK, and p38 abolished dPGJ₂’s protection of ARPE19 cells from oxidative injury, whereas inhibiting PI3K/Akt pathway failed to affect dPGJ₂’s protective effect. Heme oxygenase-1 was strongly induced by dPGJ₂ but was not associated with protection.

conclusions. Independent of its PPARγ activity, dPGJ₂ protected cells from oxidative stress by elevating GSH and enhancing MAPK activation. Thus, dPGJ₂ may delay the development of dry-type age-related macular degeneration.

Age-related macular degeneration (AMD) is an idiopathic retinal degenerative disease that predominates in the Western world as a cause of irreversible, profound vision loss in people older than 60 years. ^{1 2 3 4} Atrophic AMD is a multifactorial syndrome with genetic and environmental contributions. The pathologic changes leading to the loss and/or dysfunction of macular retinal pigment epithelial (RPE) cells are believed to be central to the etiology of this disease. Among them, lifelong chronic cumulative oxidative stress constitutes a significant risk factor, ^{5 6 7 8} since the retinal environment favors the generation of reactive oxygen species. Thus, it is of considerable interest in combating oxidative injury to RPE cells to understand factors that modulate the viability of RPE cells in an oxidative stress environment. 

Oxidative stress triggers cell death through the effects of oxidants on signal transduction pathways, ^{9 10} including the activation of sphingomyelinase, ¹¹ caspases, ¹² and cathepsin D. ¹³ One of the major mechanisms by which cells protect themselves against oxidative stress is upregulation of a wide range of antioxidant genes. Among such genes, the rate-limiting enzyme of GSH synthesis glutamate cysteine ligase (GCL) and heme oxygenase (HO)-1 have attracted great interest as modifiers of susceptibility to oxidative stress. GCL is composed of a catalytic (GCLc; molecular weight [MW], ∼73,000) and a modulatory (GCLm; MW ∼30,000) subunit encoded by two different genes. ^{14 15} GCLm is enzymically inactive but plays an important regulatory function by lowering the K _m of GCL for glutamate and raising the K _i for GSH. ^{15 16} HO-1, a stress-inducible enzyme catalyzing the degradation of heme into equimolar biliverdin, carbon monoxide, and free iron, ¹⁷ functions as a cytoprotective defense mechanism against oxidative insults via the antioxidant activity of biliverdin and its metabolite, bilirubin. ¹⁸  

The cyclopentenone 15-deoxy-Δ^12,14-prostaglandin J₂ (dPGJ₂) is a natural ligand for peroxisome proliferator-activated receptor (PPAR)-γ, a nuclear receptor and transcription factor implicated in lipid homeostasis, ^{19 20} inflammation, ^{21 22} and malignancy. ^{23 24} dPGJ₂-activated PPARγ forms a heterodimer with the retinoid X receptor, binds to the PPAR response element, and activates target gene transcription. ²⁵ In addition to the PPARγ pathway, dPGJ₂ can initiate synthesis of genes such as those encoding glutathione S-transferase and HO-1 through the antioxidant response element. ^{26 27} However, dPGJ₂ may also exert its various effects through covalent interaction with other intracellular targets. At least three identified candidates are thought to be possible mediators of the PPARγ-independent actions of dPGJ₂: induction of cyclooxygenase-2 expression, ²⁸ the nuclear factor-κB system, ²⁹ and activation of the extracellular signal-regulated kinase (ERK) pathway. ³⁰  

dPGJ₂ can protect cells from oxidative injury in various cell systems examined at submicromolar concentrations. ^{31 32} However, the cellular mechanisms of the protection afforded are still unclear. The cytoprotective effect of dPGJ₂ is apparently not shared by synthetic PPARγ agonists. ^{31 32} dPGJ₂ has been demonstrated to protect PC12 cells from nitrosative insult by inducing the increase of GSH levels in a GCL-dependent manner. ³² HO-1 induction by dPGJ₂ is essential for its anti-inflammatory effect in macrophages ³³ and its cytoprotective effect in neurons against oxidative stress. ³⁴ In our preliminary study, dPGJ₂ was found to be potent in protecting RPE cells from oxidative stress. In this study in which we used the siRNA technique and pharmacological inhibitors, we found that dPGJ₂, independent of its PPARγ activity, protected RPE cells from oxidative injury by raising intracellular GSH levels and extending hydrogen peroxide-induced activation of JNK and p38. 

Materials and Methods

DMEM/F-12 medium with sodium bicarbonate, sodium pyruvate, pyridoxine, and fetal bovine serum were purchased from (Invitrogen-Gibco, Grand Island, NY); the GSH assay kit (GSH-400) from OXIS International (Portland, OR); the MTT cell viability kit from Roche Molecular Biochemicals (Indianapolis, IN); U0126, SP600125, SB202190, and wortmannin from Calbiochem-Novabiochem Corp. (San Diego, CA); l-buthionine-sulfoximine, meta-phosphoric acid, and monoclonal anti-β-actin antibody from Sigma-Aldrich (St. Louis, MO); tin protoporphyrin IX (SnPP) from Frontier Scientific (Logan, UT); dPGJ₂ from Cayman Chemicals (Ann Arber, MI); phospho-Akt (Ser473) antibody, anti-Akt antibody, active anti-ERK, anti-JNK, and general anti-ERK, -p38, and -JNK from Cell Signaling Technology (Beverly, MA); active anti-p38 from Promega (Madison, WI); horseradish peroxidase-conjugated secondary antibodies and anti-human PPARγ antibody from Santa Cruz Biotechnology (Santa Cruz, CA); enhanced chemiluminescence reagents from GE Healthcare (Piscataway, NJ); monoclonal anti-HO-1 and anti-HSP27 antibodies from Stressgen (Victoria, BC, Canada); anti-glutamylcysteine ligase catalytic subunit antibody from Laboratory Vision (Fremont, CA); and PPARγ and GAPDH LUX primer sets from Invitrogen (Carlsbad, CA). 

Cell Culture

The human retinal pigment epithelial cell line ARPE19 was obtained from the American Type Culture Collection (Manassas, VA) and cultured in DMEM/Ham’s F12 supplemented with 10% FCS, penicillin (100 U/mL), and streptomycin sulfate (100 μg/mL). The cells were grown at 37°C in a humidified 5% CO₂ condition and split when approximately 90% confluence was reached. They were obtained at passage 20 and used at passages 21 to 30. 

Cell Viability Assay

ARPE-19 (1 × 10⁴/well) cells were seeded in 96-well flat-bottomed microculture plates for 24 hours, starved in medium with 0.1% FBS for 24 hours, and then treated with various concentrations of hydrogen peroxide in serum-free medium for the indicated time points. Untreated control cells were handled in a similar fashion without hydrogen peroxide. The number of viable cells was then determined by the addition of MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide) for 4 hours according to the manufacturer’s instructions (Roche Molecular Biochemicals). 

siRNA Transfection and PPARγ mRNA Quantitation

Stealth small interfering (si)RNAs targeting position 524 of the PPARγ coding sequence were obtained from Invitrogen (Carlsbad, CA). The target sequence (sense strand) was 5′-GCUUAUCUAUGACAGAUGUGAUCUU-3′. Cells grown in 10-cm plates were transfected overnight with 25 nM siRNA duplexes (Lipofectamine 2000 reagent; Invitrogen). After transfection, the medium was replaced, and cells were maintained in medium containing fetal bovine serum for 48 hours. The effect of siRNA on PPARγ mRNA levels was assessed by RT-qPCR. Briefly, cells transfected with either PPARγ or control siRNA were harvested at different time points after transfection, and total RNA was isolated (RNeasy kit; Qiagen, Valencia, CA). qRT-PCR was performed with a kit (SuperScript III Platinum Two-Step qRT-PCR; Invitrogen). PPARγ mRNA was measured by real-time PCR (TaqMan; Applied Biosystems, Foster City, CA). 

Exposure of RPE Cells to dPGJ₂

ARPE19 cells were subcultured into tissue culture dishes. For the cell viability assay, 24 hours after seeding, the cells were starved in culture medium containing 0.1% FBS and 1 μM dPGJ₂ for 24 hours followed by hydrogen peroxide stimulation in serum-free medium without dPGJ₂. For the immunoblot and GSH assays, 3 days after reaching confluence, the cells were starved in culture medium containing 0.1% FBS and 1 μm dPGJ₂ for 1 day and then were exposed to hydrogen peroxide at 37°C in serum-free medium without dPGJ₂. 

Cell Extraction and Immunoblot Assays

ARPE19 cells (1 × 10⁶/10-cm dish or 5 × 10⁵/6-cm dish) were seeded for 3 days. Before stimulation with hydrogen peroxide, the cells were serum starved for 24 hours in medium with 0.1% FBS. After treatment, the cells were washed twice with cold PBS containing 2 mM NaF and 2 mM vanadate and lysed in modified RIPA lysis buffer (150 mM NaCl, 1% Triton X-100, 1 mM Na₃VO₄, 2 mM phenylmethylsulfonyl fluoride, and 50 mM Tris [pH 7.4] with complete protease inhibitor cocktail [SC-29130; Santa Cruz Biotechnology]). Lysates were clarified by centrifugation at 16,000g for 15 minutes at 4°C. Total cell lysates were resolved by SDS-polyacrylamide gels, transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA), and detected with appropriate primary antibodies. The blots were subsequently incubated with secondary antibodies conjugated to horseradish peroxidase, and images were developed using the enhanced chemiluminescence system (GE Healthcare). 

GSH Measurement

The GSH content was measured by using a commercial kit according to the manufacturer’s protocol (GSH-400; OXIS International), with 4-chloro-1-methyl-7-trifluoromethyl-quinolinum methylsulfate. Cells were harvested in 500 μL meta-phosphoric working solution. After centrifugation, 200 μL of supernatant was mixed with 700 μL S3 solution. R1 solution (50 μL) was then added, followed by gentle vortex mixing. After the addition of 50 μL of R2 solution (30% NaOH), the mixtures were incubated at 25 ± 3°C for 10 minutes. The absorbance was read at 400 nm, and the results presented as x-fold increase over the control level, which was arbitrarily defined as 1. 

Statistical Analysis

Statistical significance was determined by paired two-tailed Student’s t-test. P < 0.05 was considered significant for all experiments. The values are presented as the mean ± SEM. 

Results

Protection of ARPE19 Cells from Oxidative Injury by dPGJ₂, Independent of Its PPARγ Activity

Initial experiments were performed to assess the rate of cell viability loss on exposure of the cells to oxidative stress. Starved cells were treated with various doses of hydrogen peroxide for 5 hours, and cell viability was determined with MTT conversion. ARPE19 cells were resistant to low concentrations of hydrogen peroxide, but rapidly lost viability with an increase in hydrogen peroxide concentration. Exposure of the cells to 1 mM hydrogen peroxide for 5 hours reduced viable cells to approximately 20% of control. Pretreatment with 1 μM dPGJ₂ for 24 hours alone did not alter ARPE19 cell viability, but rescued the cells from oxidative injury. The viability was increased from ∼20% to ∼80% after 1 mM hydrogen peroxide treatment (Fig. 1A) . A further increase in dPGJ₂ concentration failed to protect more (data not shown). 

dPGJ₂ is an endogenous PPARγ agonist. We, therefore, examined whether dPGJ₂ works through PPARγ activation. ARPE19 cells were pretreated with 1 μM of synthetic PPARγ agonists AGN195037 and rosiglitazone, and cell viability after oxidative stress was determined by MTT dye conversion. Unlike dPGJ₂, both AGN195037 and rosiglitazone were unable to protect cells from hydrogen peroxide–induced injury (Fig. 1A) , suggesting that dPGJ₂ may work independent of its PPARγ activity. To confirm this possibility, the siRNA technique was used to knockdown PPARγ mRNA. siRNA (25 nM) was required for near complete inhibition of PPARγ mRNA and a single-dose transfection suppressed PPARγ mRNA expression for 1 week (Fig. 1Band data not shown). Immunoblot and densitometry analysis revealed that siRNA gene knockdown correlated with a threefold reduction in PPARγ protein level (Fig. 1C) . However, PPARγ knockdown did not reduce dPGJ₂’s protection of RPE cells from oxidative injury (Fig. 1D) . 

dPGJ₂ Protection of ARPE19 Cells against Oxidative Stress via Augmentation of Intracellular GSH

Through antioxidant response elements, dPGJ₂ can induce the cytoprotective mediator GSH, the most abundant intracellular antioxidant in a variety of cell systems. To verify this in RPE cells, the intracellular GSH content was determined by a colorimetric assay kit after exposure of ARPE19 cells to various concentrations of dPGJ₂ for 24 hours, or 1 μM dPGJ₂ for the indicated time points. Incubation with dPGJ₂ triggered a concentration-dependent increase in GSH content in ARPE19 cells (Fig. 2A) . This increase in GSH was detected 24 hours after treatment with 0.25 μM dPGJ₂, the lowest concentration examined, and GSH content reached a peak with 2 μM dPGJ₂ exposure—an approximately threefold increase over the control. dPGJ₂-induced increase in GSH content was also time dependent (Fig. 2B) . With 3 hours of initial exposure to 1 μM PGJ₂, GSH content was unchanged. An increase in GSH content was observed at 6 hours and reached the maximum level after 18 hours of incubation, which is approximately a 2.5-fold elevation over the control level. l-Buthionine-sulfoximine (BSO), a specific inhibitor of the GSH synthesis rate-limiting enzyme glutamylcysteine ligase (GCL), depleted the GSH stores to 30% of control, and this BSO-dependent inhibition of GSH could not be overcome by dPGJ₂ (Fig. 2C) . The GSH level was also substantially depleted (50% reduction) after 3 hours of hydrogen peroxide treatment (Fig. 2C) . Preincubation of ARPE19 cells with dPGJ₂ under oxidative stress conditions restored the GSH level to 150% of the control. To correlate the upregulation of GSH level with dPGJ₂’s protection in ARPE19 cells, we performed MTT assays, with or without the presence of BSO. Treatment with 20 μM BSO did not affect cell viability in the absence of hydrogen peroxide, but sensitized cells to 0.5 mM hydrogen peroxide–induced injury and completely reversed dPGJ₂’s protective action against oxidative injury as illustrated in Figure 2D . These data suggest that dPGJ₂ elevates intracellular GSH concentration, thereby eliminating reactive oxygen species and exerting its cytoprotective effect. 

The principal mechanism used to increase the rate of GSH synthesis is the increased synthesis of the rate-limiting enzyme GCL. ^{35 36} We examined the content of the GCL subunits to test whether the observed increase in GSH concentration on dPGJ₂ exposure resulted from an increased capacity for de novo synthesis. ARPE19 cells were treated with various concentrations of dPGJ₂ for various lengths of time, as indicated in Figure 3A , and relative changes in the content of the GCL catalytic subunit were determined with an anti-GCLc antibody. In the ARPE19 cells, immunoblot analysis revealed that dPGJ₂ caused a time- and dose-dependent increase in the GCLc protein level. A significant increase in the GCLc level was observed 6 hours after treatment with 1 μM dPGJ₂. 

JNK and p38-Dependent Induction of Intracellular GSH by dPGJ₂

The mitogen-activated protein kinase (MAPK) pathway, including ERK, JNK, and p38, is a ubiquitous signal-transductionsystem that responds to a variety of cellular stimuli. We therefore addressed whether dPGJ₂ could modulate MAPKs, thereby regulating GSH levels. Immunoblot analyses with anti-active ERK, JNK, and p38 antibodies showed that the contents of phosphorylated ERK, JNK, and p38 increased in a dose-dependent manner with exposure to dPGJ₂, demonstrating that dPGJ₂ activated ERK, JNK, and p38 in RPE cells (Fig. 3B) . dPGJ₂ preferentially activated p44 over p42 ERK and only activated p46 JNK. Phospho-specific anti-JNK and anti-p38 recognized doublets of phospho-p46 JNK and phospho-p38 in ARPE19 cells; however, only one of them was recognized by general anti-JNK and anti-p38 antibodies. The other could not be confirmed. 

Specific inhibitors for each of the three major MAPK pathways were used to evaluate whether any of the MAPK pathways is involved in the dPGJ₂-induced increase in GSH synthesis. ARPE19 cells were pretreated with the selective inhibitors for MEK, JNK, and p38 before exposure to dPGJ₂, and the resultant GSH levels were determined (Fig. 3C) . The GSH level was significantly reduced in the presence of 20 μM JNK or p38 inhibitor, whereas inhibition of ERK and PI3K showed no effect on dPGJ₂-induced GSH synthesis (Fig. 3Cand data not shown). 

dPGJ₂ Enhancement of Hydrogen Peroxide–Induced Activation of MAPKs but Not of Akt

Cells can adapt to environmental changes by reacting quickly to extracellular stimuli via MAPK pathways. To address the potential roles of MAPKs in mediating dPGJ₂’s protection of RPE cells from oxidative injury, activation of members of the MAPK family, ERKs 1/2, JNK, and p38, was assessed. Both ERK-1 and -2 were activated by hydrogen peroxide, which peaked at 15 to 30 minutes and decreased below the basal level after 60 minutes (Fig. 4A) . With respect to JNK, a similar activation pattern was observed (Fig. 4B) . However, p46 JNK was predominantly activated, whereas p54 JNK was weakly activated in response to hydrogen peroxide. Assessment of p38 using activated p38 antibody demonstrated that hydrogen peroxide stimulated a moderate and transient increase in phospho-p38 (Fig. 4C) , clearly demonstrating that all members of the MAPK family are activated in response to hydrogen peroxide. Preincubation with dPGJ₂ led to a higher level of activation of p44 ERK, but did not change ERK’s response to hydrogen peroxide (Fig. 4A) . However, dPGJ₂ significantly prolonged hydrogen peroxide–induced activation of JNK and p38—in particular JNK (Figs. 4B 4C) . In an intriguing finding, dPGJ₂ priming led to a strong activation of p54 JNK by hydrogen peroxide at later time points (Fig. 4B) . 

PI3K/Akt pathway is a vital signaling pathway in cell survival. To address the potential roles of PI3K/Akt in RPE cell oxidative stress signaling, we assessed the activation of Akt by using an anti-pS⁴⁷³ Akt antibody. Hydrogen peroxide-induced Ser473 phosphorylation peaked at 15 minutes and thereafter declined below the basal level at 60 minutes (Fig. 4D) . Unlike responses seen with JNK and p38, hydrogen peroxide–triggered Akt activation was not modified by dPGJ₂ treatment (Fig. 4D) . 

Inhibition of MAPKs Abolished dPGJ₂’s Protective Action against Oxidative Injury

To determine the effect of prolonged MAPK activation on dPGJ₂’s protective effect, we evaluated roles of MAPKs in oxidative stress–induced cell killing by applying selective MAPK inhibitors. ARPE19 cells were pretreated with 20 μM of U0126 (MEK), SP600125 (JNK), and SB202190 (p38) for 60 minutes before the cells were exposed to hydrogen peroxide for 5 hours. MAPKs seemed to play no apparent role in hydrogen peroxide–induced cell killing (Fig. 5A) . In comparison, inhibition of ERK, JNK, and p38 completely abrogated dPGJ₂’s protection of ARPE19 cells from oxidative injury (Fig. 5B) , implying that enhanced and/or extended activation of ERK, JNK, and p38 is involved in mediating dPGJ₂’s protective effect. Pretreatment of ARPE19 cells with 100 nM of the PI3K inhibitor wortmannin, which completely blocked Akt activation by hydrogen peroxide, failed to reverse dPGJ₂’s protective action against oxidative injury (Fig. 5Band data not shown). Taken together, these results suggest that dPGJ₂-mediated cell adaptation to oxidative stress is MAPK specific. 

Induction of HO-1 by dPGJ₂

To address the role of HO-1 in dPGJ₂’s protection of ARPE19 cells, HO-1 expression before and after dPGJ₂ treatment was examined by immunoblot analysis with an anti-HO-1 antibody. The HO-1 protein level was low in dPGJ₂-untreated ARPE19 cells. Incubation of ARPE19 cells with 0.25 to 4 μM dPGJ₂ for 18 hours caused an increase in HO-1 protein, which was apparent with 0.25 μM dPGJ₂ and occurred without a decrease at the doses examined (Fig. 6A) . dPGJ₂ (1 μM) almost achieved a maximum induction of HO-1. HO-1 was significantly induced 3 hours after incubation with 1 μM dPGJ₂ and reached a peak after 6 hours. In contrast, HSP27, a low-molecular-mass heat shock protein, was abundant in ARPE19 cells and was not induced by dPGJ₂ treatment, indicating that induction of HO-1 is a dPGJ₂-specific effect. In the presence of 20 μM MAPK or 200 nM PI3K inhibitors, no inhibition of HO-1 induction by dPGJ₂ was observed (Fig. 6B) . 

To link the HO-1 induction by dPGJ₂ with its protective role, cell viability was examined in the presence of the HO-1-specific inhibitor, tin protoporphyrin IX (SnPP). SnPP alone had no toxic effect on cells up to a dose of 10 μM and dPGJ₂’s protection of ARPE19 cells from hydrogen peroxide was not abrogated by SnPP (Fig. 6C) , indicating that the induction HO-1 is not associated with dPGJ₂’s protection of RPE cells from oxidative injury. 

Discussion

Enhancing the ability of RPE cells to protect themselves from oxidative injury may provide a therapeutic opportunity to delay or stop the development of dry-type age-related macular degeneration. We therefore explored how the endogenous PPARγ agonist dPGJ₂ rescued RPE cells from oxidative injury. Two lines of evidence supported that dPGJ₂ exerted its protective effect independent of its PPARγ activity. Synthetic PPARγ agonists, AGN195037 and rosiglitazone, did not protect RPE cells from hydrogen peroxide-induced injury (Fig. 1A) , and PPARγ knockdown had no effect on dPGJ₂’s protective action (Fig. 1D) . 

dPGJ₂ can also initiate gene transcription of antioxidant and detoxification enzymes via the antioxidant response element (ARE, also called the stress response element or the electrophile response element). ^{27 37} To address whether dPGJ₂ exerted its protective role via induction of antioxidant genes in RPE cells, two cytoprotective enzymes, GCL and HO-1, were examined. GCL is the rate-limiting enzyme of de novo synthesis of the most abundant non-protein thiol antioxidant GSH. GSH levels were first evaluated on dPGJ₂ exposure, by using a GSH assay kit. A two- to threefold increase in GSH levels was detected after dPGJ₂ incubation and at least 3 hours of latent time was observed in the experimental conditions described (Figs. 2A 2B) . Consistent with this, Aoun et al. ³⁸ reported that approximately 5 hours of preincubation with dPGJ₂ is necessary to achieve significant neuroprotection. ³⁸ Exposure of RPE cells to hydrogen peroxide depleted intracellular GSH levels by 50%, and preincubation with dPGJ₂ completely restored GSH levels to 150% of control under oxidative stress (Fig. 2C) . Depletion of GSH through inhibition of the GSH synthesis enzyme GCL by BSO significantly sensitized RPE cells to oxidative stress and completely abolished the protection by dPGJ₂ (Fig. 2D) . The increase in GSH levels was preceded by upregulation of GCLc protein (Fig. 3A) . The 5′ region of GCL harbors the ARE that can be bound and activated by many basic leucine zipper transcription factors including Nrf, Jun, and Fos. ^{39 40 41} MAPKs are important signal transducers activating these transcription factors, and activity of MAPKs can be modified by dPGJ₂ treatment in various cell systems. ^{42 43} In human RPE cells, all MAPKs were activated on dPGJ₂ exposure, with a preference for p44 ERK and p46 JNK (Fig. 3B) . Therefore, the role of the MAPK pathway in mediating GSH response was tested. Inhibition of the ERK signaling pathway with U0126 had no effect on dPGJ₂-dependent induction of GSH. In contrast, inhibition of JNK and p38 MAPK reduced GSH induction by 33% with SP600125 (P = 0.01) and 31% with SB202190 (P = 0.02), respectively (Fig. 3C) . Immunoblot data showed that the protein level of GCLc was not apparently affected by the inhibitors of MAPK (data not shown). However, it is possible that the MAPKs can regulate GSH synthesis via modifying GCLm that lowers the K _m of GCL for glutamate and raises the K _i for GSH. ^{15 16} Unfortunately, we are unable to test this possibility due to the unavailability of an anti-GCLm antibody. 

dPGJ₂ can activate MAPK pathways and raise GSH levels in a JNK/p38-dependent and -independent manner, thereby rendering cells resistant to oxidative injury by detoxifying reactive oxidants. Central among the signal transducers of oxidant-induced death are MAPKs. We thus assessed the roles of the MAPK pathway and dPGJ₂-induced modification of MAPK signaling in oxidative injury. Hydrogen peroxide activated ERK, JNK and p38 MAPK with different kinetics (Figs. 4A 4B 4C) . ERK and JNK were activated, rapidly reaching a peak at 30 minutes and then declining to basal levels (Figs. 4A 4B) . P38 MAPK was initially inactivated, activated, and then rapidly declined to below the basal level (Fig. 4C) . dPGJ₂ significantly extended hydrogen peroxide–induced activation of JNK (Fig. 4B)and p38 MAPK (Fig. 4C) . In particular, dPGJ₂ priming changed the kinetic of p54 JNK activation by hydrogen peroxide, leading to a sustained and strongly enhanced activation at later time points (Fig. 4B) . Modifying activation of MAPKs by dPGJ₂ seemed to be selective, because dPGJ₂ did not activate Akt or modify hydrogen peroxide-induced activation of Akt (Fig. 4D) . Inhibition of ERK, JNK, and p38 with pharmacologic inhibitors neither exaggerated nor rescued hydrogen peroxide–induced cell death (Fig. 5A) . However, dPGJ₂’s protective effect was completely reversed by inhibiting the MAPKs ERK, JNK, and p38 (Fig. 5B) . As a control, inhibition of phosphatidylinositol-3 kinase/Akt pathway did not affect dPGJ₂-dependent protection of RPE cells from oxidative injury. Consistent with our data in Figure 5A , Garg and Chang ³¹ have reported activation of ERK by hydrogen peroxide in RPE cells, but no role of ERK was demonstrated in oxidative injury. In contrast, priming with dPGJ₂ changed the kinetics of MAPK activation in response to hydrogen peroxide and sustained activation of MAPK delivered an antiapoptotic signal in RPE cell oxidative stress signaling. Taken together, these data demonstrate that sustained activation of the MAPK pathway is a key determinant of the cytoprotective effect of dPGJ₂. Consistent with our observations, sustained ERK activation is reported to mediate adaptive cytoprotection in cardiomyocytes during recovery from stimulated ischemia. ⁴⁴ Furthermore, sustained JNK activation by tumor necrosis factor delivers an antiapoptotic signal in NF-κB–null 32D myeloid cells. ⁴⁵  

HO-1 can defend against oxidative insults through the antioxidant activity of biliverdin and its metabolite, bilirubin, ¹⁸ and HO-1 induction by dPGJ₂ is reported to be dependent on p38 MAPK in macrophages. ³³ HO-1 protein was markedly induced in ARPE19 cells on dPGJ₂ exposure (Fig. 6A)and this induction appeared not to depend on activation of the MAPK or phosphatidylinositol-3 kinase pathway (Fig. 6B) . Inhibition of HO-1 activity by SnPP, however, failed to abrogate dPGJ₂’s protective effect (Fig. 6C) , clearly indicating that induction of HO-1 did not correlate with dPGJ₂’s defensive action against oxidative injury. In comparison, combination of SnPP with dPGJ₂ appeared to protect cells better than did dPGJ₂ alone (Fig. 6C) . SnPP itself was unlikely to function as an antioxidant, since SnPP did not protect cells from oxidative injury (Fig. 6C) . Alternatively, dPGJ₂-treated cells expressed a high level of HO-1 that could trigger a Fenton reaction in the presence of hydrogen peroxide, generating more toxic hydroxyl radicals. Therefore, inhibition of HO-1 activity by SnPP could make cells more resistant to hydrogen peroxide. In agreement with this hypothesis, HO-1 overexpression has been reported to enhance hydrogen peroxide’s cytotoxicity but to protect cells from tert-butyl hydroperoxide. ⁴⁶  

In summary, we confirmed dPGJ₂’s cytoprotective role in RPE cells and revealed that dPGJ₂ mediated its effect through the GCL–GSH pathway, as observed in other cell systems, independent of its PPARγ activity. We further supported the claim that its dPGJ₂’s PPARγ activity is not essential for dPGJ₂’s protective action by showing no effect of siRNA knockdown of PPARγ on cell viability. A more intriguing finding was that dPGJ₂ priming led to sustained activation of MAPK, in particular JNK and p38, on hydrogen peroxide treatment, and this prolonged MAPK activation was essential for dPGJ₂’s cytoprotection. These findings may suggest pharmacological means to delay or stop the development of dry-type AMD. 

 Supported by Allergan, Inc.

 Submitted for publication March 23, 2006; revised May 25 and July 14, 2006; accepted September 19, 2006.

 Disclosure: S. Qin, Allergan, Inc. (E, F); A.P. McLaughlin, Allergan, Inc. (E, F); G.W. De Vries, Allergan, Inc. (E, F)

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

dPGJ₂’s cytoprotective effect was independent of its PPARγ activity. (A) The synthetic PPARγ agonists AGN195037 and rosiglitazone failed to protect RPE cells from oxidative injury. ARPE19 cells (1 × 10⁴/well, 96-well plate) were seeded for 1 day and starved for 1 day in medium containing 0.1% FBS. During starvation, cells were incubated with 1 μM dPGJ₂, 1 μM AGN195037, or 1 μM rosiglitazone and then stimulated in serum-free medium with hydrogen peroxide for 5 hours. Cell viability was calculated based on the ability of cells to convert MTT. Data are the mean ± SEM of results in three independent experiments. **P < 0.01 versus hydrogen peroxide–treated groups. (B) Time course of PPARγ mRNA knockdown by siRNA. ARPE19 cells were transfected with 25 nM PPARγ siRNA. Total RNA was isolated at various time points after transfection (2–8 days). The PPARγ mRNA level was determined by RT-qPCR and expressed as a percentage of the control. (C) Effect of PPARγ siRNA on protein expression. ARPE19 cells were transfected with 25 nM control or PPARγ siRNA. After 3 days, cell lysates were prepared, and equal amounts of protein were separated on SDS-polyacrylamide gels. PPARγ protein was detected using anti-PPARγ antibody. The intensity of bands was determined by densitometry, and PPARγ protein expression was expressed as percentage of the control. (D) Dispensability of PPARγ for dPGJ₂’s protection. ARPE19 cells were transfected with 25 nM PPARγ siRNA and replaced in 96-well plate 3 days after transfection. Cell viability was assayed as described in Figure 1A . Data are the mean ± SEM of three independent experiments. **P < 0.01 and ***P < 0.001 versus hydrogen peroxide–treated groups.

Figure 2.

 

Protection of ARPE19 cells from oxidative injury by augmenting intracellular GSH levels. (A, B) Dose and time-dependent induction of GSH by dPGJ₂. ARPE19 cells (5 × 10⁵/6-cm dish) were seeded for 3 days, followed by starvation in medium containing 0.1% FBS. The cells were incubated with various doses of dPGJ₂ for 24 hours (A) or with 1 μM dPGJ₂ at the indicated time points (B). GSH levels were assayed with a GSH assay kit. (C) Restoration of hydrogen peroxide-depleted GSH by dPGJ₂. Confluent cells were incubated with 1 μM dPGJ₂ or 20 μM BSO for 16 hours, followed by incubation with or without 1 mM hydrogen peroxide for 3 hours before cell harvest for the GSH assay. (D) Blockade of dPGJ₂’s protective action by GSH depletion. ARPE19 cells (1 × 10⁴/well, 96-well plate) were seeded for 1 day and starved for 1 day in medium containing 0.1% FBS. During starvation, the cells were incubated with 20 μM BSO and/or 1 μM dPGJ₂ and then stimulated in serum-free medium with hydrogen peroxide for 5 hours. Cell viability was calculated based on the ability of cells to convert MTT. Data shown are the mean ± SEM of results in three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 versus control (A–C) and versus the hydrogen peroxide–treated group (D).

Figure 3.

 

JNK- and p38-dependent induction of GSH by dPGJ₂. (A) Induction of GCLc by dPGJ₂. Confluent ARPE19 cells were incubated with various doses of dPGJ₂ for 18 hours or with 1 μM dPGJ₂ for various lengths of time. GCLc protein was detected with anti-GCLc antibody, and equal loading was verified by an anti-β-actin antibody. (B) Activation of ERK, JNK, and p38 by dPGJ₂. Confluent ARPE19 cells were incubated with 1 μM dPGJ₂ for 18 hours. Activation of ERK, JNK, and p38 was detected with their respective anti-active antibodies. Equal protein loading was demonstrated by immunoblot of the striped membranes with the respective general antibody. (C) JNK- and p38-dependent induction of GSH by dPGJ₂. Confluent ARPE19 cells were preincubated with 20 μM inhibitors of MEK, JNK, and p38 for 60 minutes followed by 1 μM dPGJ₂ treatment for 16 hours. The GSH level was determined with a colorimetric assay kit GSH-400. Data are the mean ± SEM of results in three independent experiments. *P < 0.05 versus the dPGJ₂-treated group.

Figure 4.

 

Activation of ERK, JNK, p38, and Akt by hydrogen peroxide and effects of dPGJ₂ on hydrogen peroxide-induced activation of MAPK and Akt in ARPE19 cells. ARPE19 cells (10 × 10⁵/10-cm dish) were seeded for 3 days, starved for 1 day in medium containing 0.1% FBS with or without the presence of 1 μM dPGJ₂, and then stimulated with 1 mM hydrogen peroxide for the indicated times without dPGJ₂. Activation of ERK (A), JNK (B), p38 (C), and Akt (D) was detected with their respective anti-active antibodies. An equal protein load was demonstrated by immunoblot of the striped membranes with the respective general antibody. Band densities were analyzed on computer and are presented as the mean ± SEM (n= 3) of the x-fold increase over the respective control (no dPGJ₂ priming and no hydrogen peroxide treatment) that were arbitrarily defined as 1. *P < 0.05 versus dPGJ₂ primed without hydrogen peroxide treatment.

Figure 5.

 

Abrogation of dPGJ₂’s protection by MAPK inhibitors but not by a PI3K inhibitor. ARPE19 cells (1 × 10⁴/well, 96-well plate) were seeded for 1 day, starved for 1 day in medium containing 0.1% FBS, either preincubated with 20 μM MAPK inhibitor for 60 minutes and then stimulated with hydrogen peroxide (A) or preincubated with 1 μM dPGJ₂ for 24 hours followed by 1 hour incubation with MAPK inhibitor or PI3K inhibitor, and then exposed to hydrogen peroxide in serum-free medium for 5 hours (B). Cell viability was calculated based on the ability of cells to convert MTT. Data are the mean ± SEM of results in three independent experiments. **P < 0.01 versus the hydrogen peroxide–treated group.

Figure 6.

 

Dispensability of HO-1 activity for dPGJ₂’s protective action. (A) Dose- and time-dependent induction of HO-1 by dPGJ₂. Confluent ARPE19 cells in 6-cm dishes were treated with various doses of dPGJ₂ for 18 hours or with 1 μM dPGJ₂ for the indicated time points. Cell lysate was separated on 4% to12% SDS-polyacrylamide gels, and protein expression was detected by immunoblot with respective antibody. (B) The dispensability of MAPK and AKT in HO-1 induction by dPGJ₂. Confluent ARPE19 cells were preincubated with the inhibitors of MEK, JNK, p38, or PI3K for 60 minutes followed by 1 μM dPGJ₂ treatment for 16 hours. HO-1 protein level was detected by Western blot with anti-HO-1 antibody. (C) No requirement of HO-1 activity for dPGJ₂’s protective effect. Cells were incubated with 1 μM dPGJ₂ for 12 hours followed by SnPP/dPGJ₂ for 12 hours and then were stimulated in serum-free medium with hydrogen peroxide for 5 hours. Data shown is the mean ± SEM of three independent experiments. **P < 0.01 and ***P < 0.001 versus hydrogen peroxide–treated groups.