GTx-822, an ERβ-Selective Agonist, Protects Retinal Pigment Epithelium (ARPE-19) from Oxidative Stress by Activating MAPK and PI3-K Pathways

Abstract

Purpose.: The goal of this study was to determine whether an estrogen receptor-β (ERβ)-selective agonist (GTx-822; GTx, Inc., Memphis, TN) could prevent hydrogen peroxide (H₂O₂)-induced oxidative stress in ARPE-19 cells and to elucidate the molecular pathways involved in this protection.

Methods.: The selectivity of GTx-822 for ERβ was determined by receptor-binding assay (RBA) and transactivation assay. Cultured ARPE-19 cells were subjected to oxidative stress with t-butyl hydroxide (t-BH) or hydrogen peroxide (H₂O₂) in the presence and absence of GTx-822. Reactive oxygen species (ROS) was measured by using H₂DCFDA fluorescence. Apoptosis was evaluated by cell death ELISA. Mitochondrial membrane potential was measured with the JC-1 assay. Gene expression and protein expression and activation were quantitated with qRT-PCR and Western blot analysis. Phospho-protein arrays elucidated the activation of protein kinases.

Results.: The RBA and transactivation assay revealed that GTx-822 is an ERβ-selective agonist (K _i = 0.53 nM). GTx-822 prevented oxidative stress in ARPE-19 cells. It preserved mitochondrial function and prevented cellular apoptosis. Pretreatment with GTx-822 increased ERβ gene and protein expression during oxidative stress. Upregulation of the phase II antioxidant genes GPx-2 and HO-1 was also seen in an ERβ-dependent mechanism. GTx-822 pretreatment induced phosphorylation of ERK1/2, PI3-K, and Bad.

Conclusions.: This is the first report to show that GTx-822, an ERβ agonist, can protect ARPE-19 cells from the cellular apoptosis induced by oxidative stress. GTx-822 mediated cytoprotection was mediated through induction of both genomic and nongenomic pathways. The results of this study open new avenues for the use of a selective ERβ agonist in treatment of ocular diseases like AMD where oxidative stress plays a major role in disease pathogenesis.

Mitochondrial dysfunction due to oxidative stress is thought to be a central factor in the pathogenesis of several ocular diseases, such as age-related macular degeneration (AMD), diabetic retinopathy, and glaucoma. ¹ AMD is a common cause of central vision loss in elderly people in the developed world ^2,3 and is characterized by slow degeneration of the central retina. Oxidative damage is likely to be higher in cells that have a high metabolic rate, such as retinal pigmented epithelial (RPE) cells. ⁴ Almost all the oxygen in aerobic cells is consumed by the mitochondria, and approximately 1% to 2% is converted to superoxide that is then converted to H₂O₂ by the mitochondrial respiratory chain. ⁵ This continuous accumulation of free radicals leads to oxidative damage and senescence of the cells. The mitochondrial theory of aging proposes that oxidative stress generates free radical species in cells that then target the mitochondria, especially the mitochondrial DNA, and cause cell death. ^6,7 Of note, mitochondria isolated from the liver and brain of female rats produce less than 50% of H₂O₂ than that isolated from the same organs in male rats, and the difference is lost on ovariectomy. ⁸ Supplementation of estrogen prevented the buildup of H₂O₂ in these rats, indicating a role for estrogen in modulating oxidative stress. ⁹  

Epidemiologic studies suggest that the risk for AMD is increased in women undergoing menopause and that women on hormone-replacement therapy have a reduced risk of having the disease. ^10,11 Several in vitro studies have shown that estradiol (E₂) prevents oxidative stress–induced apoptosis of retinal neurons, ¹² RPE cells, ¹³ and human lens epithelial cells (HLECs). ¹⁴ Despite the promising results, the use of 17β-E₂ to treat oxidative stress in the retina has been limited because of severe side effects that include increased risk of cardiovascular disease and uterine cancer. ^15,16 Estradiol binds with high affinity to estrogen receptor (ER)-α and -β, which are expressed in multiple tissues, thus causing pleiotropic effects. Most of the estrogenic responses in tissues are associated with the activation of ERα, whereas ERβ appears to act as a negative regulator of ERα. ¹⁷ We demonstrated in a prior study that 17β-E₂ protects RPE cells from oxidative stress in an ERβ-dependent manner, suggesting that ERβ-selective agonists could be beneficial for prevention of oxidative stress. ¹³ In this article, we explore the mechanism of RPE protection by one such agonist, GTx-822 (GTx, Inc., Memphis, TN), which has a selectivity for ERβ over that for ERα. This study is the first to demonstrate that an ERβ-selective agonist can protect RPE from oxidative stress and cell death. Pharmacologic antagonism of ERβ reversed the protective effects of GTx-822. Further, ERβ–GTx-822 interaction mediated cytoprotection through the activation of cell survival protein kinases: PI3-K, Akt, ERK1/2, and RSK. Activation of these pathways phosphorylated Bad (a proapoptosis protein), thereby preventing its translocation to the mitochondria and thus inhibiting cell death. 

Materials and Methods

Reagents

Phenol red-free Dulbecco's modified Eagle's medium/Ham's F12 (DMEM/ F12), Hanks' balanced salt solution (HBSS) plus calcium and magnesium, phosphate-buffered saline (PBS containing Ca²⁺ and Mg²⁺), 1 M Tris (pH 8.0), and 0.5 M EDTA were obtained from Mediatech (Manassas, VA). 2′,7′-Dichlorodihydrofluorescein (H₂DCFDA) was obtained from Molecular Probes (Eugene, OR). Hydrogen peroxide (H₂O₂) and t-butyl hydroxide (t-BH) were obtained from Fisher Scientific (Pittsburgh, PA). Tetrahydrochrysene (THC) and ICI-182780 (ICI) were purchased from Tocris (Ellisville, MO). Charcoal-stripped, dextran-treated fetal bovine serum (csFBS) was obtained from Atlanta Biologicals (Norcross, GA). Trypsin and 0.25% EDTA was obtained from Invitrogen Corp. (Carlsbad, CA). 

Cell Culture

HEK-293 and ARPE-19 cells were obtained from American Type Culture Collection (ATCC; Manassas, VA). The cells were cultured in DMEM-F12 supplemented with 10% FBS. They were plated in the densities mentioned in the description of each assay and incubated overnight. After the incubation period, the cells were washed and then treated with the appropriate concentration of drugs in phenol red-free DMEM/F12 medium, containing 5% csFBS. 

Transactivation Assay

Rat ERα and ERβ sequences were amplified from rat ovarian cDNA and cloned into a pCR3.1 (Invitrogen) plasmid vector backbone. The transactivation potential of GTx-822 was determine with a luciferin-luciferase–based assay. 

Radioligand Binding Assay

A competitive receptor binding assay was performed with [³H]E₂ and bacterial lysates expressing rat ERα or -β ligand-binding domain (LBD). For competitive binding, increasing concentrations of compounds (10⁻¹¹–10⁻⁵ M) were incubated with [³H]E₂ (1–4 nM) and ER LBD, and the assay was performed. The concentration of compound that reduced the specific binding of [³H]E₂ by 50% (IC₅₀) was determined by computer-fitting the data (Sigma Plot; Systat Software, Inc., San Jose, CA) and by performing nonlinear regression with the four-parameter logistic curve. The equilibrium binding constant (K _i) of each compound was then calculated by: K _i = K _d × IC₅₀/(K _d + L), where K _d is the equilibrium dissociation constant of [³H]E₂, and L is the concentration of [³H]E₂. 

Fluorescent Detection of Intracellular Reactive Oxygen Species (ROS)

The assay was performed as published in our earlier study. ¹³ Briefly, ARPE-19 cells (1 × 10⁵ cells/well) in 24-well plates were preincubated with 10 μM H₂DCFDA for 30 minutes before incubation with either vehicle or compounds for 2 hours. The cells were then treated with 150 μM of t-BH for 1 hour, and intracellular ROS was measured. 

Measurement of Mitochondrial Potential with the JC-1 Assay

The assay was performed as described in our study. ¹³ Briefly, ARPE-19 cells (2.5 × 10⁵ cells/well) in six-well plates were incubated with either vehicle or compounds for 2 hours before exposure to 500 μM H₂O₂. After an incubation period of 4 hours at 37°C, the cells were treated in the dark with 10 μg/mL JC-1 dye at 37°C for 15 minutes before measurement of the mitochondrial membrane potential. 

Measurement of Apoptosis

The assay was performed as detailed in our prior work. ¹³ ARPE-19 cells (5 × 10⁴ cells/well) were incubated in 24-well plates, with either vehicle or respective compounds for 2 hours before exposure to 500 μM H₂O₂. After 24 hours' incubation, cellular apoptosis was quantified (Cell Death Detection ELISA Kit; Roche Diagnostics, Mannheim, Germany), according to the manufacturer's instructions. 

Measurement of ATP Levels

A detailed description of the assay has been reported. ¹³ Briefly, ARPE-19 cells were treated with vehicle or GTx-822 for 2 hours and then exposed to 500 μM H₂O₂ for 4 hours. The cells (5 × 10⁴) were lysed with 1 mL HBSS containing 0.5% TCA and 4 mM EDTA. The lysate was then centrifuged at 16,000g for 10 minutes, and the supernatants were collected for analysis. The pH of the samples was adjusted to ∼7.75 using 25 mM Tris (pH 8.0). Twenty-five microliters of the sample was loaded on a 96-well plate to measure intracellular ATP levels (Enliten ATP Assay kit; Promega, Madison, WI), according to the manufacturer's instructions. 

Real-Time Quantitative PCR

ARPE-19 cells (2.5 × 10⁵ cells in 10-cm plates) were treated with the indicated compounds followed by 500 μM H₂O₂ for 4 hours. RNA extraction and real-time PCR using gene expression assays (TaqMan, ABI) for each gene (ESR1 [ERα], ESR2 [ERβ], GPX1, GPX2, HMOX1 [HO-1], HMOX2 [HO-2], SOD1, SOD2, SOD3) was performed in duplicate in three separate experiments, according to our published method. ¹³ Quantification was performed by the ^ΔΔCt method with 18S RNA as the loading control. 

Measurement of SOD Activity

SOD activity was determined according to our published protocol. ¹³ Briefly, ARPE-19 cells (4 × 10⁶ cells in 15-cm plates) were treated with vehicle or GTx-822 and then incubated with 1 mM H₂O₂ for 24 hours. After the incubation period, mitochondrial and cytosolic fractions were extracted with a mitochondria isolation kit (Pierce, Rockford, IL), according to the manufacturer's instructions. SOD activity in the fractions was determined with a kit (Assay Designs, Ann Arbor, MI), according to the manufacturer's instructions. 

Western Blot Analysis

ARPE-19 cell lysates were obtained at 4, 8, 12, and 24 hours after exposure to vehicle or compound, with or without 500 μM H₂O₂. Immunoblot analysis and detection were performed with antibodies (Table 1), using published methods. ¹³ Phospho-MAPK array was performed on cell lysates (250 μg protein) obtained at 4 hours after oxidative stress, according to the manufacturer's instructions (R&D Systems, Minneapolis, MN). Densitometry was then performed (Photoshop CS4; Adobe Systems Inc., San Jose, CA). 

Table 1.

 

List of Antibodies Used for Western Blot Analysis

Antibody	Source	Concentration Used
Rabbit anti-ERβ	Upstate (Temecula, CA)	WB - 1:350
Rabbit anti-phospho ERK1/2	Cell Signaling Technology (Danvers, MA)	WB: 1:1000
Rabbit anti-ERK1/2 (Total)	Cell Signaling Technology	WB: 1:1000
Rabbit anti-phospho Akt (Thr308)	Cell Signaling Technology	WB: 1:1000
Rabbit anti-phospho Akt (Thr473)	Cell Signaling Technology	WB: 1:1000
Mouse anti-Akt (total)	Cell Signaling Technology	WB: 1:1000
Rabbit anti-phospho Bad (Serl12)	Cell Signaling Technology	WB: 1:500
Rabbit anti-phospho Bad (Serl36)	Abcam (Cambridge, MA)	WB: 1:500
Mouse anti-β-actin	Upstate (Temecula, CA)	WB 1:5000

Results

We recently demonstrated that estradiol (17β-E₂) protects ARPE-19 cells from hydrogen peroxide–induced stress through engagement of ERβ. ¹³ In the present study, we tested the effectiveness of the nonsteroidal ERβ-selective agonist GTx-822 (Fig. 1) in protecting RPE mitochondria from oxidative stress. 

Figure 1.

 

Chemical structure of GTx-822.

Binding Affinity and Transactivation Potential of GTx-822

Competitive receptor binding assays to estimate the relative binding affinity of GTx-822 and E₂ to ER were performed. GTx-822 bound both ER subtypes, but showed a higher selectivity for ERβ than for ERα (Table 2). GTx-822 bound ERβ with affinity (K _i = 0.53 nM) similar to that of 17β-E₂ (K _i = 0.39 nM), but bound ERα with significantly less affinity than 17β-E₂. RBA, calculated as the ratio of the K _i values for the compound of interest to 17β-E₂, indicated that GTx-822 bound more preferentially to ERβ (∼77-fold) than to ERα. In vitro transactivation assays in HEK-293 cells showed that GTx-822 activated the transcription of ERβ with eightfold higher potency than that observed with ERα. Also, GTx-822 did not antagonize E₂-mediated transactivation of either ERα or -β (Table 2). GTx-822 did not induce maximum transactivation of ERα or -β (ERα E _max = 78.1%, ERβ E _max = 58.3%), indicating that the compound was a partial agonist of ERβ. 

Table 2.

 

RBA and Transcriptional Activation of ER by GTx-822

Compounds	Receptor Binding					Transactivation
	K i ERα (nM)	K i ERβ (nM)	RBA ERα	RBA ERβ	Ratio RBA α:β	Agonist ERα (nM)	Agonist ERβ (nM)	Antag. ERα (nM)	Antag. ERβ (nM)
Estradiol	0.29	0.39	1	1	1	0.016	0.121	—	—
GTX-822	30.01	0.53	103.48	1.36	76.60	326.8	40.5	>10,000	>3,000
E max						78.1%	58.3%

 Receptor-binding assay: The equilibrium dissociation constant (K _d) was determined by incubating LBD protein lysates with increasing concentrations of [³H]E₂ in buffer A (10 mM Tris [pH 7.4], 1.5 mM disodium EDTA, 0.25 M sucrose, 10 mM sodium molybdate, 1 mM PMSF) with or without excess unlabeled E₂ at 4°C for 18 hours. After the incubation period, the lysates were harvested onto GF/B filters (Perkin Elmer, Inc., Wellesley, MA), and the filters were washed with buffer B (50 mM Tris; pH 7.2) and then dried at room temperature. A scintillation cocktail (Microscint-O; Perkin Elmer) was added to the filter plates, and radioactive counts were performed (TopCount NXT Microplate Scintillation Counter; Perkin Elmer). K _d for ER (ERα, 0.71 nM; ERβ, 1.13 nM) was determined by computer, the data were fitted (Sigma Plot; Systat Software, San Jose, CA), and nonlinear regression analysis was performed with the ligand-binding one-site saturation curve.  Transactivation assay: For this assay, HEK-293 cells (1.25 × 10⁵ cells/well) were transfected with ERE-LUC (firefly luciferase), constitutively expressed CMV-LUC (renilla luciferase), and ERα or -β with transfection reagent (Lipofectamine; Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. Twenty-four hours after transfection, the cells were treated with increasing concentrations of compounds (range, 10⁻¹⁰–10⁻⁵ M) in DMSO, estradiol (range, 10⁻¹³–10⁻⁷ M) in EtOH, or a combination of ERβ ligands and estradiol in volumes that kept the final concentration of vehicle in the media at 0.1%. Forty-eight hours after drug treatment, a reporter assay (Dual-Luciferase Reporter Assay; Promega) was performed with a multilabel plate reader (Victor³ V; Perkin Elmer Inc.) to measure firefly (FF) and renilla (Ren) luminescence. The concentration of compounds that stimulated the activity of the receptor by 50% (EC₅₀) of the maximum response (E _max) was determined by computer-fitting the data (Sigma Plot; Systat Software), and nonlinear regression was performed with the four-parameter logistic curve. The concentration of drug that inhibited 50% of the E _max (IC₅₀) induced by 1 nM E₂ was determined by the same method.

GTx-822 Protection of ARPE-19 Cells from H₂O₂-Induced Oxidative Stress

Similar to our study of 17β-E2, ¹³ we examined the effect of GTx-822 in protecting RPE from the oxidants H₂O₂ and t-BH. 

Effect on ROS Accumulation.

There was a rapid accumulation of ROS (measured by H₂DCFDA fluorescence) after treatment with 150 μM t-BH (Fig. 2A). When the cells were treated with GTx-822 (0.01–3 μM) before exposure to t-BH, GTx-822 dose dependently prevented the accumulation of ROS in the cells (Fig. 2A). ROS accumulation was reduced by 50% at 1 to 3 μM GTx-822 treatment, when compared with that of the control cells (Fig. 2A). When the cells were treated with the ER antagonist ICI (182170) before GTx-822 pretreatment, the protective effect of GTx-822 was abrogated (Fig. 2B), indicating an ER-mediated mechanism of protection. 

Figure 2.

 

GTx-822 protected ARPE-19 cells from oxidative stress by an ERβ-dependent mechanism. (A) ROS accumulation: dose–response relationship of GTx-822. Cells were treated with either vehicle or increasing concentrations of GTx-822 (0.01–3 μM) for 2 hours before treatment with 150 μM t-BH for 1 hour. Intracellular ROS was measured by H₂DCFDA fluorescence. (B) ROS accumulation: ICI reversed the effect of GTx-822. Cells were treated with vehicle, 1 μM GTx-822, 1 μM ICI, or 1 μM GTx-822+1 μM ICI for 2 hours before treatment with 150 μM t-BH or vehicle for 1 hour. Intracellular ROS was measured by H₂DCFDA fluorescence. (C) JC-1 assay: Dose–response relationship of GTx-822. Mitochondrial membrane potential was measured by JC-1 assay. Cells were pretreated with vehicle or increasing concentration of GTx-822 (0.01–3 μM) for 2 hours before treatment with 500 μM H₂O₂ for 4 hours. (D) JC-1 assay: Effect of ER isotype–specific antagonists. ARPE-19 cells were treated for 2 hours with vehicle, 1 μM GTx-822, 1 μM ICI, 1 μM THC, 1 μM tamoxifen, 1 μM GTx-822+1 μM ICI, 1 μM GTx-822+1 μM THC, or 1 μM GTx-822+1 μM tamoxifen and were later exposed to 500 μM H₂O₂ for 4 hours. Mitochondrial potential was calculated as a ratio of green to red fluorescence and is expressed as a percentage of H₂O₂-only treated cells. In all experiments, n = 3, mean ± SEM; **P < 0.01, ***P < 0.001, versus oxidant-only–treated cells.

Effect on Mitochondrial Potential.

The mitochondrial membrane potential was measured as a change in fluorescence of the potentiometric dye JC-1. An increase in the ratio of green to red fluorescence is an indicator of membrane depolarization. ¹⁸ On exposure to H₂O₂, the ARPE-19 cells underwent rapid depolarization, as seen by an increase in the green/red fluorescence ratio (Fig. 2C). When the cells were treated with GTx-822 (0.01–3 μM) before H₂O₂, membrane depolarization was prevented in a dose-dependent manner, with maximum effect observed at 1 μM, at which the green/red fluorescence ratio was similar to that in untreated control cells (Fig. 2C, 2D) and so we used this concentration of GTx-822 for all subsequent assays. Preincubation with ICI (full antagonist of ERα and -β) or THC (ERα agonist and full ERβ antagonist ¹⁹ ) before GTx-822 treatment abrogated the protective effect of GTx-822 on H₂O₂-induced mitochondrial depolarization (Fig. 2D). Tamoxifen (ERα antagonist) did not inhibit GTx-822-mediated mitochondrial protection (Fig. 2D). Pretreatment with the antagonists ICI, THC, or tamoxifen alone did not show any protection against H₂O₂-mediated stress (Fig. 2D). These data show that mitochondrial protection of GTx-822 is mediated through ERβ and requires ligand activation of ERβ. 

Effect on Cellular Apoptosis.

When ARPE-19 cells were incubated with 500 μM H₂O₂ for 24 hours, a significant portion (fivefold compared with the control) of the cells underwent nucleosomal fragmentation and apoptosis (Fig. 3A). Pretreatment with GTx-822 prevented H₂O₂-induced apoptosis. This protective effect was reversed when the cells were treated with the ER antagonist ICI before GTx-822 (Fig. 3A). 

Figure 3.

 

(A) Apoptosis assay: ARPE-19 cells were treated for 2 hours with vehicle or 1 μM GTx-822, 1 μM ICI, or 1 μM GTx-822+1 μM ICI before 500 μM H₂O₂ or vehicle was added. Twenty-four hours after H₂O₂ stress, 1 × 10⁴ trypan blue–negative cells were used to measure the extent of nucleosomal fragmentation by ELISA. Percentage apoptosis of each sample versus H₂O₂-only treatment is shown. (B) Cellular ATP content: ATP levels were measured in a luciferin-luciferase–based assay. ARPE-19 cells were treated for 2 hours with vehicle or 1 μM GTx-822 before addition of 500 μM H₂O₂. ATP levels were measured 4 hours after oxidative stress. The total amount of ATP in each sample was determined as relative luminescence units (RLU)/mole of ATP. Results are expressed as a percentage of ATP in normal untreated ARPE-19 cells. In all experiments, n = 3, mean ± SEM; ***P < 0.001 versus oxidant-only–treated cells.

Effect on Intracellular ATP Levels.

Hydrogen peroxide treatment reduced intracellular ATP levels by 80% compared with those in the untreated control cells (Fig. 3B). When the cells were treated with GTx-822 before H₂O₂ stress, the loss of ATP was only 20% compared with that from the vehicle-treated cells. 

GTx-822 Induction of ERβ and Phase II Genes

We evaluated the ability of GTx-822 to modulate the expression of the phase II genes glutathione peroxidase (GPx-2, -3) hemoxygenase (HO-1, -2) and superoxide dismutase (SOD-1, -2, -3). These enzymes act as cellular antioxidants that neutralize the buildup of ROS in cells and protect the cells from oxidative injury. Exposure to H₂O₂ downregulated the gene expression of GPx2 by 3.5-fold, compared with that of untreated cells (Fig. 4A). Pretreatment with GTx-822 prevented the downregulation of GPx2 and also upregulated the levels of HO-1 (∼2 fold). These protective effects were reversed when the GTx-822 treatment was applied in the presence of the antagonist ICI (Fig. 4A). These data indicate that GTx-822 modulation of antioxidant gene expression requires ERβ. No significant change in expression of ERa, GPx3, or HO-2 or the SOD genes was observed after oxidative stress (Fig. 4A) or GTx-822 pretreatment. We have shown in a prior study that H₂O₂ stress reduces mitochondrial, and cytosolic SOD activity (∼42% and ∼61%, respectively), compared with that in control cells. Pretreatment with 17β-E₂ partially restored SOD activity. ¹³ Similar to the protective effect of 17β-E₂, pretreatment with GTx-822 brought back mitochondrial and cytosolic SOD activity to ∼82% and ∼75%, relative to that in the control cells (Fig. 4B). 

Figure 4.

 

GTx-822 increased the expression of ERβ and cellular antioxidant genes. (A) qPCR analysis: ARPE-19 cells were treated with vehicle, 1 μM GTx-822, 1 μM ICI, or 1 μM GTx-822+1 μM ICI for 2 hours and then exposed to 500 μM H₂O₂ or vehicle for 4 hours. RNA was isolated, and quantitative RT-PCR was performed to evaluate the expression of the ERa, ERb, GPx2, GPx3, HO-1, HO-2, SOD1, SOD2, and SOD3 genes. Data are expressed as the mean ± SEM of duplicate amplifications from three experiments; **P < 0.01 versus oxidant-only–treated cells. (B) SOD activity: Mitochondrial and cytosolic fractions were isolated by differential centrifugation of cell lysates at 24 hours, and the SOD activity in each fraction was determined. SOD levels are expressed as a percentage of enzyme activity in untreated ARPE-19 cells (control). In all experiments, n = 3, mean ± SEM; **P < 0.01 versus oxidant only treated cells. (C) Western blot analysis: ARPE-19 cells were treated with vehicle or 1 μM GTx-822 and then exposed to 500 μM H₂O₂ for 4, 8, 12, and 24 hours. Twenty micrograms of protein from cell lysates were resolved by SDS-PAGE and probed with antibody against ERβ. Lane 1: cells only; lane 2: cells+500 μM H₂O₂; and lane 3: cells+500 μM H₂O₂+1 μM GTx-822. The blot is representative of results in three separate experiments. The relative ERβ expression (ERβ/β-actin ratio) was determined by densitometric analysis. *P < 0.05 versus oxidant-only–treated cells.

GTx-822 pretreatment induced the expression of ERβ mRNA (∼fourfold versus the control cells). To verify ERβ overexpression at the protein level, we performed Western blot analysis for ERβ with cell lysates obtained at various times after oxidative stress. Our data revealed that ERβ protein expression was increased at 8 hours and returned to baseline by 24 hours (Fig. 4C). 

GTx-822 Activation of the ERK and PI3-K Cell Survival Pathways

H₂O₂-mediated cell death in RPE cells involves the activation of stress kinases (p38 and JNK). ²⁰ The activation and phosphorylation of kinases by ligands have been shown to play an important role in recruitment and activation of steroid receptors and their co-activators. ^21,22 To determine whether GTx-822-mediated cytoprotection involves activation or inhibition of kinase pathways, we used a phospho-MAPK array that detects 18 different phosphorylated kinases (Fig. 5A). In cells treated with vehicle or GTx-822, only ERK1/2 phosphorylation was evident. On treatment with H₂O₂, phosphorylation of ERK1/2 and Akt1/2/3 was reduced, whereas the phosphorylation of the death kinase p38α was increased (Fig. 5A). When the cells were treated with GTx-822 and then exposed to H₂O₂ stress, a significant increase in phosphorylation of the serine/threonine kinases Akt1/2/3 and ERK1/2 (versus H₂O₂-only–treated cells) was observed (Fig. 5A). In addition, the activation of p38α observed in H₂O₂ treatment was significantly attenuated in cells pretreated with GTx-822. To confirm these results, we performed Western blot analysis with antibodies specific to phospho-ERK1/2, total ERK1/2, phospho-Akt (Ser473 and Thr308), and total Akt (Fig. 5B). Pretreatment with GTx-822 increased levels of phosphorylated Akt (Ser473 and Thr308 sites) and ERK1/2 (Thr202/Tyr204 and Thr185/Tyr187). No change in the levels of total Akt or ERK1/2 and β-actin was seen. 

Figure 5.

 

GTx-822 activated ERK1/2 and Akt pathways and inhibited the p38α-MAPK pathway. (A) Phospho-protein array: ARPE-19 cells were treated with vehicle or 1 μM GTx-822 and then exposed to either vehicle or 500 μM H₂O₂ for 4 hours. The cells were lysed and 250 μg protein was loaded onto the array. The phospho-protein levels were determined by using the antibody cocktail provided by the manufacturer. (B) Western blot analysis: 10 to 20 μm of protein from the cell lysates were resolved by SDS-PAGE and probed with antibodies against the total proteins (Akt and ERK1/2), phospho-ERK1/2, and phospho-Akt (Ser473 and Thr308). β-Actin was used as the loading control. Lane 1: cells+vehicle; lane 2: cells+500 μM H₂O₂; and lane 3: cells+500 μM H₂O₂+1 μM GTx-822. The blots are representative of three separate experiments. The relative phospho-protein levels (phospho-protein/total protein) were determined by densitometric analysis. **P < 0.01 versus H₂O₂-only–treated cells.

Effect of Protein Kinase Inhibitors on Mitochondrial Membrane Potential

To explore the importance of protein kinase pathway activation, we performed the JC-1 assay in the presence of specific inhibitors of MAPK (U0126), PI3-K-Akt (LY294002), and Ras (FTI277). The PI3-K kinase inhibitor LY294002 and MAPK inhibitor U0126 prevented GTx-822-mediated protection of mitochondrial membrane potential, whereas the Ras inhibitor FTI227 had only a marginal effect (Fig. 6). These data, combined with our results from the MAPK array, indicate that GTx-822-mediated cytoprotection requires the activation of the PI3-K-Akt and ERK/p90^Rsk1 pathways. 

Figure 6.

 

Inhibition of MAPK and PI3-K reversed the protection of GTx-822. ARPE-19 cells in serum-free medium were treated for 2 hours with vehicle or 1 μM GTx-822, 1 μM GTx-822+1 μM U0126, 1 μM GTx-822+1 μM Ly294002, or 1 μM GTx-822+1 μM FTI277 and then exposed to 500 μM H₂O₂ for 4 hours. Mitochondrial membrane potential was measured using the JC-1 assay. Mitochondrial potential was calculated as a ratio of green to red fluorescence and is expressed as a percentage of H₂O₂-only–treated cells. All experiments, n = 3; mean ± SEM; ***P < 0.001 versus H₂O₂-only–treated cells.

GTx-822 Inhibition of the Translocation of Bad

It has been reported that E₂ prevents H₂O₂-induced apoptosis in MCF-7 cells by phosphorylation of Bad (a Bcl-2 family proapoptotic protein) through activation of the Ras/PI-3K/Akt and Ras/ERK/p90^Rsk1 pathways. ²³ In apoptotic or stressed cells, Bad translocates to the mitochondria and heterodimerizes with Bcl2 and Bcl-X_L (the antiapoptosis Bcl-2 family proteins). These events then lead to opening of the mitochondrial membrane pore followed by the release of cytochrome c, leading to apoptosis. In normal cells, Bad is phosphorylated at specific sites (Ser112 and Ser136) and is sequestered in the cytoplasm by 14-3-3 protein. Phosphorylation at Ser112 is mediated by ERK1/2-p90^Rsk1, whereas Ser136 is phosphorylated by Akt. ^24,25 In ARPE-19 cells exposed to H₂O₂, Bad phosphorylation was reduced at both the serine residues (Ser112 and Ser136; Fig. 7), whereas in GTx-822-pretreated cells, Bad was phosphorylated at both residues, indicating the activation of both the PI3K-Akt and ERK1/2-p90^Rsk1 pathways. These data indicate that GTx-822 protects mitochondria by activation of signaling pathways that prevent Bad translocation to the mitochondria. 

Figure 7.

 

GTx-822 protected ARPE-19 cells from oxidative stress by activating Bad. Western blot analysis of ARPE-19 cells treated with vehicle or 1 μM GTx-822 for 2 hours and exposed to 500 μM H₂O₂ for 4 hours. Twenty micrograms of cell lysate was resolved by SDS-PAGE and probed with antibodies against phospho-Bad (Ser112 and Ser136). β-Actin was used as the loading control. Lane 1: cells+vehicle; lane 2: cells+500 μM H₂O₂; lane 3: cells+500 μM H₂O₂+1 μM GTx-822. The blots are representative of three separate experiments. The relative phospho-protein levels (phospho-Bad/β-actin) were determined by densitometric analysis. **P < 0.01 versus H₂O₂-only–treated cells.

Discussion

Estrogens prevent cellular apoptosis through several mechanisms, including acting as nonspecific antioxidants, modulation of gene expression, ²⁶ and activation of nongenomic signaling pathways. ^23,27,28 These mechanisms are dependent on the cell type and pathophysiological condition of the cell. Recently, we showed that 17β-E₂, but not 17α-E₂, protects ARPE-19 cells from oxidative stress mediated by H₂O₂, ¹³ indicating the requirement for strong ER-ligand interaction for mediating these effects. We also showed that the protection of mitochondria in ARPE-19 cells from oxidative stress by estrogen was mediated through ERβ and not ERα. ¹³ In this study, we tested a nonsteroidal ERβ-selective agonist GTx-822 for its ability to protect ARPE-19 cells from oxidative stress. 

Our data showed that GTx-822 pretreatment prevented H₂O₂-induced accumulation of ROS, loss of mitochondrial potential and ATP, thereby preventing apoptosis and death of ARPE-19 cells (Figs. 2, 3). Further, we showed that the ERβ antagonist THC reversed the effects mediated by GTx-822, whereas the ERα antagonist tamoxifen did not prevent GTx-822-mediated protection (Fig. 2). These observations with GTx-822 are similar to our findings with 17β-E2 ¹³ and confirm that ERβ, not ERα, activation is necessary to protect ARPE-19 cells from oxidative stress. This is the first study to show the effectiveness of an ERβ-selective agonist for protecting cells from oxidative stress. 

Classic estrogen receptor signaling (genomic signaling) involves the ligand bound receptor translocating to the nucleus, where it binds to estrogen-responsive elements (EREs) of genes and activates or represses gene expression. ²⁹ An interesting finding in our study was that pretreatment with GTx-822 during oxidative stress upregulated the levels of ERβ gene and protein. To our knowledge, this is the first report to show that an ERβ ligand can upregulate the levels of its cognate receptor. A comparative bioinformatics search revealed two putative EREs in the proximal promoter of the human ERβ sequence (Table 3). An earlier study using genome-wide analysis of the ERα and RNAPolII binding sites revealed ER binding sites in both the ESR1 (ERα) and ESR2 (ERβ) genes, ³¹ supporting the potential for estrogen self-regulation of both ER subtypes. Our gene expression studies also showed that GTx-822 upregulated the expression of the antioxidant genes GPx2 and HO-1 (Fig. 4). Phase II proteins such as GPx2 and HO-1, neutralize the reactive oxygen species and protect the cells from oxidative injury. GPx2 and HO-1 gene proximal promoters lacked putative EREs (data not shown), suggesting that GTx-822 modulates the expression of these target genes through indirect or nonclassic mechanisms. 

Table 3.

 

EREs in the ERβ Promoter Sequence

Gene Name	mRNA Accession	PWM Score	Position (Relative to TSS)	Strand	Chromosome	ERE Sequence
ESR2	NM_001473	0.952	3805 Upstream	(−)	14	TGGTCAAGGTGTCCC
		0.902	1332 Upstream	(−)	14	TGGTCAGGCTGGTCT

 The bioinformatics search of the human ERβ promoter sequence was performed as described by Jin et al. ³⁰

In the past decade, findings in several studies have suggested that membrane-bound estrogen receptors elicit rapid nongenomic signaling cascades that modulate cellular physiology in mammary gland, bone, brain, and cardiovascular system. ³² These rapid effects of estrogens are mediated by the activation of signaling cascades such as MAPK, Src kinase, and PI3-K. ^{33 –35} H₂O₂ stress in RPE cells has been shown to activate cell death pathways, such as p38 and JNK. ²⁰ Our study showed that the ERβ ligand GTx-822 activated the cell survival pathways PI3-K-Akt and ERK1/2-p90^Rsk1 while inhibiting the cellular death pathway involving p38-MAPK. The activation of these cascades by GTx-822 was observed only after oxidative stress and not under basal conditions, indicating a specific protective response during stress (Fig. 5). Estrogen-mediated nongenomic signaling involving MAPK has been described in HLECs subjected to the lethal effects of the oxidant H₂O₂, ³⁶ indicating a role for the nongenomic signaling in ocular tissues. Both ERα and -β are known to activate the PI3-K-Akt pathway. However, a physical interaction between the liganded receptor and p55 subunit of PI3-K has been shown only for ERα. ³⁴ Immunoprecipitation of spermatozoa has shown that ERβ co-immunoprecipitates with Akt, ³⁷ indicating a direct interaction between ERβ and Akt. We were unable to detect a direct interaction between ERβ and Akt in immunoprecipitation assays in our laboratory (data not shown). 

Treatment with pharmacologic inhibitors of either the PI3-K-Akt or ERK1/2 pathway abrogated the protective effects of GTx-822, indicating the importance of both pathways for mitochondrial protection. In MCF-7 cells exposed to TNFα or H₂O₂, E₂ was shown to reduce cellular apoptosis by activation of signaling pathways that lead to phosphorylation and deactivation of the proapoptotic protein Bad. ²³ Phosphorylated Bad is a substrate for sequestration by 14-3-3 proteins in the cytoplasm. In apoptotic cells, unphosphorylated Bad translocates to the mitochondria and dimerizes with the antiapoptotic protein Bcl-X_L and inactivates it. These events then lead to depolarization of mitochondrial potential, release of cytochrome c, and apoptosis of the cells. ²⁵ In our study, Bad phosphorylation was reduced at the Ser112 and Ser136 residues in ARPE-19 exposed to oxidative stress (Fig. 7). Pretreatment with GTx-822 increased Bad phosphorylation at both the Ser112 and the Ser136 residues, indicating a mechanism of mitochondrial protection similar to that described for MCF-7 cells. ²³ It was surprising to note that the use of either the PI3-K inhibitor LY294002 or the MEK1/2 inhibitor U0126 completely abrogated the protective effects (Fig. 6). Since the activation of the signaling events by GTx-822 was seen only in the presence of H₂O₂ (which downregulates both pathways) we were unable to isolate the effect of the inhibitors U0126 or LY294002 separately. This result could explain the all-or-nothing effect on mitochondrial potential observed in our experiments with the inhibitors. Alternatively, there could be cross-talk between the PI3K-Akt and ERK1/2 signaling pathways. ³⁸  

Based on our results, we propose a two-part model for GTx-822-mediated protection of ARPE-19 cells (Fig. 8). This model involves a nongenomic rapid phase during which the ligand-ERβ complex at the plasma membrane activates signaling through PI3-K and/or ERK1/2 kinases, which leads to sequestration of Bad, thereby protecting the mitochondria from oxidative damage. After this early phase, we envision a second phase of regulation for long-term protection, involving the activation of transcriptional events that modulate expression of antioxidant genes (HO-1 and GPx2) and/or ERβ. Akt phosphorylation has been shown to modulate the transcription of HO-1 ³⁹ and the GSH/GRX redox system in HepG2 and H9c2 cells by alternate mechanisms, such as Nrf2 translocation and binding to EpRE-like elements. ²⁸ Phosphorylation of the Akt ⁴⁰ or MAPK pathway ⁴¹ has been shown to modulate ERβ-mediated transcription through nonclassic mechanisms. In the absence of ERβ-binding sites in the promoter regions of GPx2 and HO-1, we hypothesize that GTx-822 induces the expression of these antioxidant genes through the nonclassic pathways explained earlier. Taken together, the data presented in this article reveal that GTx-822 protects ARPE-19 cells from oxidative stress via nongenomic and genomic ERβ-dependent mechanisms. These results reveal the potential for ERβ agonists in the treatment of diseases involving oxidative stress, such as AMD, as an essential component for their pathogenesis and progression. 

Figure 8.

 

Proposed mechanism for GTx-822-mediated protection from oxidative stress. (1) Immediate response (nongenomic signaling). GTx-822, in an ERβ-dependent mechanism, inhibits p38α MAPK and activates Akt and ERK1/2. Activated Akt and ERK1/2 phosphorylate Bad at Ser136 and Ser112, respectively. Phosphorylation of Bad leads to sequestering by 14-3-3 protein, thus inhibiting translocation of Bad to the mitochondria and preventing loss of mitochondrial membrane potential. (2) Prolonged response (genomic). Ligand-bound ERβ can activate transcription through direct interaction with promoters containing EREs in the nucleus or indirectly through phosphorylation of Akt. Phospho-Akt can then activate transcription through NF-κB/AP-1 interactions.

Footnotes

 Supported by GTx, Inc.

Footnotes

 Disclosure: A. Giddabasappa, GTx, Inc. (E, I, F); M.N. Bauler, GTx, Inc. (E, I, F); C.M. Barrett, GTx, Inc. (E, I, F); C.C. Coss, GTx, Inc. (E, I, F); Z. Wu, GTx, Inc. (E, I, F); D.D. Miller GTx, Inc. (E, I, F); J.T. Dalton, GTx, Inc. (E, I, F); J.R. Eswaraka, GTx, Inc. (E, I, F)