17-β Estradiol Protects ARPE-19 Cells from Oxidative Stress through Estrogen Receptor-β

Abstract

Purpose.: To elucidate the mechanism of 17-β estradiol (17β-E₂)–mediated protection of retinal pigment epithelium (RPE) from oxidative stress.

Methods.: Cultured ARPE-19 cells were subjected to oxidative stress with t-butyl hydroxide or hydrogen peroxide in the presence or absence of 17β-E₂. Reactive oxygen species (ROS) were measured using H₂DCFDA fluorescence. Apoptosis was evaluated by cell-death ELISA kit and Hoechst-3486 staining. Mitochondrial membrane potential was measured using the JC-1 assay. Cellular localization of estrogen receptor (ER) was evaluated by confocal microscopy. Gene expression and protein expression was quantified using qRT-PCR and western blotting. Superoxide dismutase and ATP levels were measured using commercial kits.

Results.: ARPE-19 cells expressed significant amounts of ERα and ERβ. Pretreatment with 17β-E2 protected ARPE-19 cells from oxidative stress and apoptosis. 17β-E₂ reduced the ROS levels and mitochondrial depolarization. The 17β-E₂–mediated cytoprotection was inhibited by ER antagonists ICI (ERα and ERβ) and THC (ERβ) but not by tamoxifen (ERα). Knockdown of ERβ expression by siRNA abolished the protective effects of 17β-E₂. Further, qRT-PCR analysis revealed that 17β-E₂ pretreatment upregulated the expression of ERβ and phase II cellular antioxidant genes.

Conclusions.: These results indicate that 17β-E₂ protects ARPE-19 cells from oxidative stress through an ERβ-dependent mechanism. 17β-E₂–mediated cytoprotection occurred through the preservation of mitochondrial function, reduction of ROS production, and induction of cellular antioxidant genes.

Age related macular degeneration (AMD) is the leading cause of blindness among the elderly population in the Western world. ^1,2 The number of patients with AMD is expected to increase from 1.75 million to 3 million in the next decade. ³ Epidemiologic studies suggest a strong association between estrogen deficiency, early menopause, and the development of AMD in women. ^4–7  

The pathogenesis of AMD involves apoptosis of the retinal pigment epithelium (RPE) followed by death of the underlying photoreceptors. ^8,9 Because of its unique location and function, the RPE provides an ideal environment for the accumulation of reactive oxygen intermediates (ROIs), ¹⁰ which, in turn, leads to mitochondrial dysfunction and RPE death. ¹¹ Although antioxidants such as β-carotene, zinc, and vitamins C and E have been shown to slow the progression of this disease, ^12,13 there is no effective treatment for advanced atrophic AMD. 

Several in vitro and in vivo studies have shown that 17β-estradiol (17β-E₂) protects RPE from oxidative stress. ^14–16 Microarray analysis of RPE exposed to hydrogen peroxide (H₂O₂) showed that 17β-E₂ pretreatment induced the upregulation of apoptosis-related protein and protected the RPE from degeneration. ¹⁵ Estrogen deficiency in high-fat fed C57BL/6 mice increased the severity of sub-RPE deposits. ¹⁶ Estrogens have been shown to protect a variety of tissues, including brain, breast, myocardium, and lens epithelium, from oxidative damage by targeting the mitochondrial respiratory chain. ¹⁷ Despite these compelling observations, the mechanism by which the estrogens protect cells from oxidative stress is not completely understood. Some studies suggest an ER-mediated mechanism, ^18,19 whereas others suggest that estrogens act as general antioxidants. ²⁰  

In this study, we set out to investigate the mechanism of 17β-E₂–mediated cytoprotection in RPE cells and to identify the ER subtype that mediates these protective effects. We used the well-characterized model of H₂O₂-induced oxidative stress in ARPE-19 cells as the in vitro model system. This study provides the first definitive evidence that activation of ERβ protects RPE from oxidative stress, which is hypothesized to play a role in pathogenesis of AMD. 

Materials and Methods

Reagents

Dulbecco's modified Eagle's medium (DMEM/ F12) was obtained from Hyclone (Logan, UT). 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). Probe (MitoTracker; M7512) and 2′, 7′-dichlorodihydrofluorescein (H₂DCFDA) were obtained from Molecular Probes (Eugene, OR). H₂O₂ and t-butyl hydroxide (t-BH) were obtained from Fisher Scientific (Pittsburgh, PA); 17α-E₂ and 17β-E₂ were obtained from Sigma Aldrich (St. Louis, MO). Tamoxifen, ICI-182780 (ICI), and tetrahydrochrysene (THC) were purchased from Tocris (Ellisville, MO). Charcoal stripped/dextran-treated fetal bovine serum (csFBS) was obtained from Atlanta Biologicals (Lawrenceville, GA). 

Cell Culture

Primary human RPE cells were isolated from donor eyes provided by the Mid-South Eye Bank (Memphis, TN) using procedures previously described. ²¹ ARPE-19 cells were obtained from the American Type Culture Collection (Manassas, VA). All research herein described adhered to the tenets of the Declaration of Helsinki. All studies and procedures used de-identified tissues and were approved by the University of Tennessee Institutional Review Board. Cells were cultured in DMEM/F12 containing 10% csFBS (complete medium) and were plated in densities mentioned for each assay. After overnight incubation at 37°C, cells were washed and then treated with either vehicle or estradiol for 2 hours in phenol red-free DMEM/F12 + 5% csFBS before exposure to appropriate concentration of t-BH or H₂O₂. In assays with ER antagonists, the antagonists were added 30 minutes before the addition of 17β-E₂. 

Antibodies

Antibodies and dyes used for Western blot analysis and confocal microscopy are listed in Table 1. 

Table 1.

 

List of Antibodies and Dyes Used for Western Blot Analysis and Immunohistochemistry

Antibody/Dye	Source	Concentration Used
Mouse anti-ERα	Cell Signaling Technology (Danvers, MA)	WB, 1:1000
Rabbit anti-ERβ	Upstate (Temecula, CA)	WB, 1:350
Rabbit anti–ERβ (H-150)	Santa Cruz Biotechnology Inc. (Santa Cruz, CA)	IHC, 1:100
Mouse anti–β-actin	Upstate	WB, 1:5000
Mouse anti–RNA Pol II	Upstate	WB, 1:500
Mouse anti–VDAC	ABCAM (Cambridge, MA)	WB, 1:500
Hoechst 33342	Molecular Probes (Eugene, OR)	1 μM
MitoTracker (M7512)	Molecular Probes	100 nM
ToPro-3	Molecular Probes	300 nM
Anti–rabbit/mouse-HRP	Cell Signaling Technology	WB, 1:1000–1:5000
Anti–rabbit-Alexa488	Molecular Probes	IHC, 1:300

Fluorescence Detection of Intracellular Reactive Oxygen Species

ARPE-19 cells (1 × 10⁵ cells/well) in 24-well plates were loaded with 10 μM H₂DCFDA. After 30 minutes of incubation, cells were washed and then incubated with either vehicle or 17β-E₂. Cells were then treated with 150 μM t-BH for 1 hour. Intracellular ROS production was measured using a bioplate reader (Victor³ V; Perkin Elmer, Norwalk, CT; excitation, 485 nm; emission, 535 nm). 

Measurement of Necrosis and Apoptosis

ARPE-19 cells (5 × 10⁴ cells/well) were incubated with either vehicle or 17β-E₂ before exposure to 500 μM H₂O₂. After 24 hours of incubation, cells were trypsinized and the percentage of necrotic cells (trypan blue positive) was determined. Apoptosis was quantified with a commercial kit (Cell Death Detection ELISA Kit; Roche Diagnostics, Mannheim, Germany) in accordance with the manufacturer's instructions. Absorbance was measured using a bioplate reader (Victor³ V [Perkin Elmer]; absorbance, 405 nm). To evaluate the effect of H₂O₂ stress on cellular morphology, cells were stained with Hoechst 33342 dye at 24 hours, and photomicrographs were taken under a microscope (Axiophot; Carl Zeiss, Oberkochen, Germany). 

Measurement of Mitochondrial Potential Using JC-1 Assay

ARPE-19 cells (2.5 × 10⁵ cells/well) in six-well plates were incubated with either vehicle or 17β-E₂ before exposure to 500 μM H₂O₂ and then were incubated for an additional 4 hours at 37°C. The cells were then washed and incubated in the dark with 10 μg/mL JC-1 dye at 37°C for 15 minutes. Mitochondrial membrane potential was measured using a bioplate reader (Victor³ V [Perkin Elmer]; green fluorescence, 535 nm; red fluorescence, 595 nm). 

Measurement of ATP Levels

ARPE-19 cells were treated with vehicle or 17β-E₂ and then exposed to 500 μM H₂O₂ for 4 hours. Cells were washed with PBS, trypsinized, and pelleted at 250g for 5 minutes. Approximately 5 × 10⁴ cells were lysed with 1 mL HBSS containing 0.5% TCA and 4 mM EDTA. The lysate was then centrifuged at 3000g for 10 minutes, and the supernatants were taken for analysis. The pH of samples was adjusted to approximately 7.75 using 25 mM Tris (pH 8.0). An aliquot (25 μL) of the sample was loaded on a 96-well plate to measure intracellular ATP levels using an assay kit (ENLITEN ATP Assay kit; Promega, Madison, WI) in accordance with the manufacturer's instructions. 

Measurement of Superoxide Dismutase Activity

ARPE-19 cells (4 × 10⁶ cells in 15-cm plates) were incubated with 17β-E₂ or vehicle and then treated with 1 mM H₂O₂ for 24 hours. After the incubation period, mitochondrial and cytosolic fractions were extracted using a mitochondria isolation kit (Pierce, Rockford, IL) in accordance with the manufacturer's instructions. Superoxide dismutase (SOD) activity in the fractions was determined using a SOD assay kit (Assay Designs, Ann Arbor, MI) in accordance with the manufacturer's instructions. 

Real-Time Quantitative PCR

ARPE-19 cells (2.5 × 10⁵ cells in 10-cm plates) were treated with 17β-E₂, followed by 1 mM H₂O₂ for 4 hours. After the incubation period, cells were washed with PBS and RNA extracted using reagent (Trizol; Invitrogen, Carlsbad, CA). Two micrograms of RNA was reverse transcribed (High Capacity cDNA RT Kit; Applied Biosystems, Foster City, CA). Reactions without template served as negative controls. Samples were amplified using ABI probes for genes (ESR1 [ERα], ESR2 [ERβ], GPX1, GPX2, HMOX1 [HO-1], HMOX2 [HO-2]) and master mix (TaqMan Fast Universal PCR Master Mix, No AmpErase UNG on a 7900HT Fast Real-Time PCR System; Applied Biosystems, Foster City, CA). PCR reactions for each gene were performed in triplicate. Quantification was performed by the ΔΔCt method using 18S RNA as loading control. 

Western Blot Analysis

ARPE-19 cell lysates were obtained at 4, 8, 12, and 24 hours after exposure to 1 mM H₂O₂. The lysates were centrifuged at 3000g for 5 minutes, and the supernatants were resolved on a 5% to 20% SDS-PAGE gradient gel (Bio-Rad, Hercules, CA). The nuclear/mitochondrial/cytosolic-enriched fractions were separated in accordance with the manufacturer's instructions in the mitochondria isolation kit (Pierce, Rockford, IL). Proteins were transferred to nitrocellulose membranes, and immunoblotting was performed with antibodies to mouse anti-ERα, rabbit anti-ERβ, and mouse anti-β-actin. Immunodetection was performed with secondary antibody conjugated to HRP and an enhanced chemiluminescence kit (Amersham, Piscataway, NJ). Densitometry was performed using image editing software (Photoshop CS4; Adobe Systems Incorporated, San Jose, CA). 

Immunohistochemistry and Confocal Microscopy

ARPE-19 cells (3 × 10⁴ cells/well) on a four-well chamber slide (LabTek, Scotts Valley, CA) were treated with vehicle or 17β-E₂ and then exposed to 500 μM H₂O₂ for 4 hours. The cells were washed with PBS and incubated with 100 nM probe (MitoTracker; Molecular Probes) at 37°C for 20 minutes. The cells were then fixed with 4% paraformaldehyde for 15 minutes and blocked with buffer (10% normal goat serum, 5% BSA, 0.02% Triton X-100 in PBS) for 60 minutes. The cells were incubated in rabbit anti-ERβ (H-150) (1:200) overnight at 4°C. After the incubation period, the slides were washed with PBS and incubated in secondary antibody (goat anti rabbit Alexa-488; 1:300) and nuclear stain (ToPro-3, 1:300; Molecular Probes, CA) for 60 minutes at room temperature. Slides were mounted with mounting medium (Vectashield; Vector Laboratories, Burlingame, CA). 

Small Interfering RNA Knockdown of ERβ

To knockdown ERβ gene expression from ARPE-19 cells, we used Accell smart pool small interfering RNA (siRNA; mixture of 4 siRNA duplexes; Dharmacon, Inc., Chicago, IL) against ERβ (cat. no. E-003402–00-0010) or ERα (cat. no. E-00341–00-0002) or GAPDH or nontargeting siRNA. Cells were transfected with 1 μM siRNA in siRNA delivery medium (Accell; Dharmacon, Inc.) for 72 hours at 37°C. After the incubation period, ER protein expression was determined in cell lysates by Western blot analysis. Cells in siRNA-containing medium were then used in the JC-1 assay to measure mitochondrial protection. 

Results

Western Blot Analysis for Estrogen Receptor Expression

We first compared the expression of ERα and ERβ in primary hRPE with that observed in ARPE-19 cells. The primary RPE cells were obtained from both male (hRPE-8, hRPE-12, hRPE-14, hRPE-15) and female (hRPE-13) patients (age range, 41–72 years). Western blot analysis of the cell lysates showed that both hRPE cells and ARPE-19 cells express high levels of ERα and ERβ (Fig. 1). ERβ expression levels were similar in all cell lines examined (Fig. 1B). ERα expression was more variable with hRPE 13 from a female donor (∼43 years) with the most ERα expression (Fig. 1C). Because there was no significant difference in ERβ expression between the hRPE and ARPE-19 cells, all our subsequent experiments were carried out using the well-characterized ARPE-19 cells. 

Figure 1.

 

ERα and ERβ receptor protein expression in human primary RPE and ARPE-19 cells. Twenty micrograms of protein from RPE cell lysates were resolved by SDS-PAGE and probed with antibodies to ER-α (H-184) and ER-β. Cell lysates from the breast cancer cell line MCF-7 were used as positive controls (A). Relative ERβ (B) and ERα (C) protein expression was determined as a ratio of ER (α/β) and β-actin. *P < 0.05 versus ARPE-19 cells.

Oxidative Stress Assay Conditions

We performed dose-response studies with membrane-permeable oxidants H₂O₂ and t-BH to determine the LC₅₀ for each oxidant. Treatment with H₂O₂ (concentrations up to 1 mM) induced ROS formation, but the magnitude of ROS accumulation was lower than that seen with t-BH (Figs. 2A, 2B). The LC₅₀ was found to be dependent on the confluence of the cultures. The LC₅₀ for subconfluent monolayers was approximately 500 μM H₂O₂ and approximately 150 μM t-BH (Figs. 2A, 2B), whereas in confluent monolayers the LC₅₀ was 1 mM for H₂O₂ and 300 μM for t-BH, respectively (data not shown). Data presented in this study are from H₂O₂-induced stress experiments unless specifically stated. 

Figure 2.

 

17β-E₂ prevents ROS production. Intracellular ROS was measured by H₂DCFDA fluorescence. Dose-response data for intracellular ROS accumulation by H₂O₂ (A) and t-BH (B) are shown. (C) Cells loaded with H₂DCFDA were treated with vehicle or 1 μM 17β-E₂ for 2 hours, followed by a bolus of 150 μM t-BH. Relative fluorescence of samples versus t-BH-only treatment is shown. For all experiments, n = 3; mean ± SEM; ***P < 0.001.

Intracellular Reactive Oxygen Species Accumulation

Treatment of ARPE cells with t-BH (150 μM) increased the intracellular accumulation of ROS (Fig. 2C) compared with untreated cells. 17β-E₂ attenuated the ROS accumulation in a concentration-dependent manner with maximal effect seen at 1 μM concentration (data not shown). This concentration of 17β-E₂ was used for all subsequent experiments. At 1 μM concentration, 17β-E₂ limited ROS accumulation in the cells by 53% compared with t-BH-only controls (Fig. 2C). Pretreatment with ER antagonist ICI reversed the protective effects of 17β-E₂. 

Oxidative Stress–Induced Cell Death

To evaluate the effects of oxidative stress on cell growth, we monitored ARPE-19 cells for changes in morphology, apoptosis, and necrosis 24 hours after exposure to 500 μM H₂O₂. After 24-hour exposure, a significant number of cells showed cell rounding and detached from the plate (Fig. 3A). Nuclear staining (Hoechst-33342) of the cells revealed pyknotic nuclei (Fig. 3A, arrowheads) indicating condensed chromatin (bright punctate labeling) and apoptosis. A significant number of cells (∼25%) underwent necrosis (Table 2). Pretreatment with 17β-E₂ reduced the percentage of necrotic cells to 11.98% ± 4.3% (Table 2). ARPE-19 cells exposed to peroxide had a fivefold increase in apoptosis relative to untreated cells (Fig. 3B). Pretreatment with 17β-E₂ before H₂O₂ exposure reduced the percentage of apoptotic cells to 41% ± 6.79% (Fig. 3B). Further, the morphology of the 17β-E₂–pretreated cells was similar to that of the untreated controls (Fig. 3A) indicating cytoprotection. This protective effect was abolished on pretreatment with ER antagonist ICI, suggesting an ER-dependent mechanism for the action of 17β-E₂ (Fig. 3B). 

Figure 3.

 

17β-E₂ prevents oxidative stress–induced apoptosis of ARPE-19 cells. (A) ARPE-19 cells were treated with 500 μM H₂O₂ only and H₂O₂ + 1 μM 17β-E₂. Nuclei were stained with Hoechst 33342 dye. Representative bright-field (5×) and fluorescent (40×) micrographs at 24 hours after oxidant treatment are shown. White arrowheads: apoptotic cells with brightly stained fragmented nuclei. (B) ARPE-19 cells were treated with 500 μM H₂O₂ only, H₂O₂ + 1 μM 17β-E₂, H₂O₂ + 1 μM ICI, and H₂O₂ + 1 μM 17β-E₂ + ICI. Twenty fours later, 1 × 10⁴ trypan blue–negative cells were used to measure the extent of nucleosomal fragmentation and apoptosis using ELISA. The percentage apoptosis of each sample versus H₂O₂ treatment is depicted. For all experiments, n = 3; mean ± SEM; ***P < 0.001 versus oxidant-only treated cell.

Table 2.

 

Quantification of Necrosis by Trypan Blue Method

Treatment	Trypan Blue–Positive Cells (%)	SD
Cells + vehicle	5.33	4.37
500 μM H2O2 + vehicle	24.90	0.80
500 μM H2O2 + 17β-E2	11.98	4.30

 Percentages of ARPE-19 cells undergoing necrosis after 500 μM H₂O₂ treatment. Subconfluent monolayers of ARPE-19 cells were pretreated with vehicle or 1 μM 17β-E2 and were treated with 500 μM H₂O₂ for 24 hours. Cells were trypsinized, and cell suspension was stained with trypan blue. Cells that took up trypan blue were counted using a hemocytometer and were expressed as a percentage of total number of cells. Each experiment was performed at least three times in triplicate, and the values are expressed as ±SD.

Oxidative Stress Affects Mitochondrial Function

Mitochondrial Membrane Potential.

H₂O₂ caused a rapid loss of mitochondrial membrane potential as seen by the increase in the green/red fluorescence ratio (Figs. 4A, 4B). Pretreatment with 17β-E₂ reduced mitochondrial depolarization in a dose-dependent manner, with the maximum (∼50%) reduction seen at 1 μM concentration compared with H₂O₂-only treated cells (Figs. 4A, 4B). This protective effect of 17β-E₂ was also prevented by ICI (Fig. 4B). 

Figure 4.

 

17β-E2 protects mitochondria in H₂O₂-treated ARPE-19 cells. (A) Dose-response data. Mitochondrial membrane potential was measured using JC-1 dye. ARPE-19 cells in serum-free medium were pretreated with vehicle or increasing concentrations of 17β-E₂ and then exposed to 500 μM H₂O₂ for 4 hours. (B) 17β-E₂ protects mitochondria in an ER-dependent manner. ARPE-19 cells in serum-free medium were pretreated with vehicle or 1 μM 17β-E₂, 1 μM ICI, or 1 μM 17β-E₂ + ICI and were exposed to 500 μM H₂O₂ for 4 hours and then stained with JC-1 dye. Mitochondrial membrane potential was calculated as a ratio of green to red fluorescence and is expressed as a percentage of H₂O₂-only treated cells. For all experiments, n = 3; mean ± SEM; **P < 0.01 and ***P < 0.001 versus oxidant-only treated cell.

Intracellular ATP Levels.

H₂O₂ depleted intracellular ATP levels by 80% compared with that of the control untreated cells (Fig. 5A). Pretreatment with 17β-E₂ limited ATP loss to just 30% compared with the controls. 

Figure 5.

 

(A) Cellular ATP levels were measured using a luciferin-luciferase–based assay. Cells were exposed to H₂O₂ with or without 17β-E₂ pretreatment, and ATP levels were measured at 4 hours after oxidative stress. The total amount of ATP in each sample was determined as RLU U/mol ATP. Results are expressed as a percentage of ATP in untreated ARPE-19 cells (control). (B) SOD activity was determined using a colorimetric assay. Mitochondrial and cytosolic fractions were isolated by differential centrifugation of cell lysates at 24 hours, and the levels of SOD in each fraction was estimated using a commercial kit. SOD levels are expressed as a percentage of enzyme activity in untreated ARPE-19 cells (control). For all experiments, n = 3; mean ± SEM; ***P < 0.001 versus oxidant-only treated cell.

Superoxide Dismutase Levels in Cells.

H₂O₂ stress reduced cytosolic (63% ± 2.2%) and mitochondrial SOD (45.10% ± 2.45%) activity compared with control untreated cells (Fig. 5B). Pretreatment with 17β-E₂ partially prevented decreases in mitochondrial (76.17% ± 2.98%) and cytosolic SOD (71.60% ± 2.34%) levels compared with H₂O₂-treated cells. 

Role of ERβ in 17β-E₂–Mediated Protection: Mechanisms

Effect of Non-ER Binding Ligand.

To determine whether ER binding is required for mitochondrial protection from H₂O₂ stress, we used 17α-E₂ (a structural analog of 17β-E₂, which binds ER with 100-fold less potency). Pretreatment with 17α-E₂ protected only a small percentage of cells (17.38% ± 3.07%) from membrane depolarization compared to 17β-E₂ treatment (48.5% ± 3.07%: Fig. 6A). 

Figure 6.

 

ERβ is essential for cytoprotection. (A) JC-1 assay was carried out with or without pretreatment with 17α-E2, tamoxifen, and THC before exposure to vehicle or 17β-E2 and 500 μM H₂O₂. Mitochondrial membrane potential was measured at 4 hours after oxidative stress. Green to red fluorescence ratio is depicted as a percentage of H₂O₂-only treated cells. n = 3; mean ± SEM; ***P < 0.001 versus oxidant-only treated cell. (B) siRNA suppression of ERβ. Representative Western blot of cell lysates with antibody against ERβ in ARPE-19 cells transfected with either mock-transfected (lane 1), siRNA against GAPDH (lane 2), or scrambled nontarget (lane 3) siRNA or siRNA duplexes to ERα (lane 4) and ERβ (lane 5). Western blot against β-actin served as loading control. (C) Mitochondrial protection after ERβ knockdown. ARPE-19 cells transfected with siRNA (as described) were pretreated with vehicle or 17β-E2 and then exposed to H₂O₂ for 4 hours, and mitochondrial membrane potential was measured by JC1 assay. Green to red fluorescence ratio is depicted as a percentage of H₂O₂-only treated cells. n = 3; mean ± SEM; ***P < 0.001 versus vehicle-only control cells.

Effect of ER Inhibitors.

We performed the JC-1 assay in the presence of ER inhibitors ICI (pure α and β antagonists), tamoxifen (ERα partial agonist/antagonist) and THC (pure ERβ antagonist). The antagonists had no effect on loss of mitochondrial membrane potential mediated by H₂O₂ (data not shown). Preincubation of cells with ICI or THC before 17β-E₂ treatment completely abrogated the protective effects of 17β-E₂ (Fig. 6A). In contrast, tamoxifen did not reverse the protective effect of 17β-E₂ (Fig. 6A) but did prevent the moderate protection mediated by 17α-E₂. 

Effect of ERβ Knockdown.

To evaluate the role of ERβ in mitochondrial protection, we performed siRNA-mediated silencing of gene expression. Western blot analysis 72 hours after transfection with the siRNA duplexes specific to ERβ revealed a knockdown of ERβ expression (Fig. 6B, lane 5). No change in ERβ expression was seen when cells were transfected with nontarget siRNA (Fig. 6B, lanes 2–4). When siRNA-transfected cells were subjected to H₂O₂ stress and JC-1 assay was performed, 17β-E₂ prevented the loss of mitochondrial potential induced by H₂O₂ in cells transfected with scrambled or ERα siRNA or GAPDH siRNA (Fig. 6C) but not in ERβ siRNA-transfected cells. These results indicate that ERβ is required for protecting ARPE-19 cells from oxidative stress. 

ERβ Localization in Mitochondria during Oxidative Stress.

To determine ERβ localization in cells, we performed immunohistochemistry and confocal microscopy at 4 hours after oxidative stress. When cells were treated with vehicle alone, ERβ localized to the nucleus, mitochondria (punctate staining colocalized with MitoTracker) and cytoplasm (diffuse staining; Figs. 7A, 7A′). When cells were treated with 17β-E₂ without oxidative stress, ERβ localization was similar to that of control (Figs. 7B, 7B′). On exposure to H₂O₂, the mitochondria collapsed and ERβ was seen to colocalize within the collapsed mitochondria (intense yellow tubular structures in the perinuclear region; Figs. 7C, 7C′, arrowheads). In cells pretreated with 17β-E₂ followed by H₂O₂ treatment, there was a significant increase in ERβ staining in mitochondria (Figs. 7D, 7D′) compared with H₂O₂-only treated cells. In addition, we performed subcellular fractionation of the cells at 4 hours after oxidative stress and analyzed ERβ protein levels in cytosol, mitochondria, and nucleus. Our results show that oxidative stress reduced mitochondrial and nuclear pools of ERβ whereas the cytosolic pool was increased (Fig. 7E). Pretreatment with 17β-E₂ restored the nuclear and mitochondrial ERβ pools (Fig. 7E). 

Figure 7.

 

Localization of ERβ in ARPE-19 cells. Cells were treated with vehicle, 1 μM 17β-E₂ only, 500 μM H₂O₂, or 1 μM 17β-E₂ + 500 μM H₂O₂ and processed for immunohistochemistry. To distinguish cellular compartments, mitochondria were labeled with probe (MitoTracker; red) and the nuclei were stained with nuclear stain (ToPro; blue). ERβ protein expression was detected using polyclonal rabbit anti–ERβ (H-150) antibody and counterstained with anti–rabbit Alexa-488 antibody (green). Images were acquired with a confocal microscope using multitrack configuration. All images were acquired on the same day using the same parameters and were uniformly processed. Representative images at 40× magnification from cells treated with vehicle (A, A′), 1 μM 17β-E₂ only (B, B′), 500 μM H₂O₂ (C, C′), and 500 μM H₂O₂ + 1 μM 17β-E2 (D, D′) are shown in all three separate channels. Right: overlay of all three channels to show colocalization. Arrows: normal punctate colocalization of ERβ in the mitochondria located in the perinuclear region. Arrowheads: ERβ colocalization along the collapsed mitochondria. A′–D′ represent 2.5× digital magnification of the overlay images. (E) Western blot of ERβ: cytosolic, mitochondrial, and nuclear-enriched fractions were obtained at 4 hours after oxidative stress and probed for the presence of ERβ. β-Actin (cytosolic marker), VDAC (mitochondrial marker), and RNA polymerase II (nuclear marker).

ERβ-Induced Gene Expression

Pretreatment with 17β-E₂ before H₂O₂ stress increased ERβ gene expression by 4.3-fold compared with controls. Pretreatment with ICI before exposure to 17β-E₂ prevented the upregulation of ERβ gene expression (Fig. 8A). Western blot analysis for ERβ showed that protein levels increased by 8 hours after stress and were reduced to baseline by 24 hours (Fig. 8B). Treatment with 17β-E₂ without oxidative stress had no effect on the protein levels of ERβ (Fig. 8C). On the other hand, ERα gene expression was downregulated with H₂O₂ stress, and 17β-E₂ or ICI pretreatment had no beneficial effects on gene expression (Fig. 8A). 

Figure 8.

 

Analysis of gene and protein expression in ARPE-19 cells. (A) Gene expression was determined using qRT-PCR. Cell lysates from the different treatment paradigms were analyzed for changes in expression of phase II genes and ERα and ERβ. Data are represented as fold change relative to untreated (control) cells (baseline expression). Data from 4-hour time point alone is shown. (B) ERβ protein expression during oxidative stress. Representative Western blot analysis shows the kinetics of ERβ protein levels at 4, 8, 12, and 24 hours. Lane 1: untreated cells; lane 2: vehicle + 500 μM H₂O₂; lane 3: 500 μM H₂O₂ + 1 μM 17β-E2. Relative ERβ expression (ERβ/β-actin ratio) was determined using imaging editing software. (C) Effect of 1 μM 17β-E2-only treatment on ERβ protein expression. n = 3; mean ± SEM; **P < 0.01.

H₂O₂ stress downregulated gene expression of GPx2 (−4.25-fold) and HO-1 (−0.52-fold) compared with vehicle-treated cells (Fig. 8A). When the cells were pretreated with 17β-E₂ before oxidative stress, there was an upregulation of HO-1 expression (2.5-fold), whereas GPx2 level was brought back to that of vehicle-treated cells (Fig. 8A). Induction of the phase II genes was completely abolished on treatment with the ER antagonist ICI. No change in the expression of HO-2 (constitutive form of HO) or GPx3 was observed with any treatments. 

Discussion

In these studies we showed that 17β-E₂ at pharmacologic concentrations prevented the accumulation of ROS, collapse of mitochondrial potential, loss of cellular ATP, and apoptosis of the RPE cells caused by oxidative stress. We also showed that the protection of mitochondria by 17β-E₂ is ER dependent because the ER antagonist (ICI 182170) blocks the beneficial effects. ICI is a pure antagonist of ER and leads to the degradation of ER receptors. ²² Therefore, we infer that the lack of protective effect in the presence of ICI is due to the lack of ER in the cells after treatment. Additionally 17α-E₂, a weak ERα ligand, provided only marginal protection compared with 17β-E_2, indicating a requirement for tight binding with the receptor. These data indicate that 17β-E₂ protects mitochondria in ARPE-19 cells through a specific receptor-mediated mechanism. Several reports have shown that estrogens protect mitochondria in human lens epithelial cells (HLECs) from H₂O₂ death. ^20,23,24 The molecular mechanism of protection in HLECs was a rapid nongenomic-type response and was ER dependent. 

Estrogens mediate their physiological effects through two subtypes of receptors, ERα and ERβ. These are nuclear receptors that, on ligand binding, translocate to the nucleus and modulate gene expression. ²⁵ Although few studies suggest that ERβ acts as a negative regulator of ERα, ^26,27 the role of ERβ in cellular physiology is not clearly understood. Βoth ER subtypes are expressed in the male and female retina. ^28,29 Similar to these studies, our RNA and protein expression results confirmed that ERα and ERβ are expressed in the RPE. To identify which of the ER subtypes is required for protecting RPE from oxidative stress, we used ER subtype-specific antagonists and targeted knockdown of ER gene expression. Our results show that antagonists that block ERβ (ICI or THC) completely abrogated the beneficial effects of 17β-E₂. Preincubation with tamoxifen (ERα antagonist) reversed the effect of 17α-E2 (ERα agonist) but did not block 17β-E₂–mediated cytoprotection. Barkhem et al. ³⁰ showed that 17α-E2 has partial agonism in 293/hERβ cells and full agonist activity in 293/hERα cells. These results indicate that the mild protection seen with 17α-E2 is probably an ERα-mediated effect. Further, siRNA knockdown of ERβ gene expression, but not ERα gene expression, resulted in the loss of mitochondrial potential even in the presence of 17β-E₂. These data indicate that 17β-E₂ mediates its protection by engaging ERβ but not ERα. This is the first report to elucidate the role of ERβ in protecting mitochondria in RPE cells from oxidative damage. Previously, it had been demonstrated that ERβ plays an important role in regulating extracellular matrix turnover during oxidant injury in RPE both in vitro and in vivo. ^14,31 These observations indicate that ERβ can protect RPE from oxidative stress through multiple mechanisms. 

ERβ was shown to localize to the mitochondria of HLECS and protect the cells from oxidative stress. ²⁴ In RPE cells, unliganded ERβ was distributed in the cytoplasm, nuclei, and mitochondria, similar to the localization of the wild-type receptor in HLECs. ²⁴ During oxidative stress, the receptor predominantly colocalized with the collapsed mitochondria in the perinuclear region. In cells pretreated with 17β-E_2, mitochondrial membrane potential was preserved and there was increased ERβ receptor colocalization with the mitochondria as early as 4 hours. At this time point, there was a fourfold increase in transcript levels of ERβ, but protein levels were increased only 8 hours after exposure. These data, along with our subcellular fractionation results (Fig. 7E), indicate that at the earlier time points, 17β-E₂ may stabilize the mitochondrial ERβ levels or induce translocation of ERβ to the mitochondria. Although the exact mechanism by which mitochondria-associated ERβ protects RPE cells from oxidative stress is unknown, we speculate that ERβ and possibly estrogen bound to the receptor associate with the mitochondrial permeability transition pore and stabilize the mitochondrial membrane, similar to that suggested for HLECs. ²⁴  

Estrogens, estrogen metabolites, and phytoestrogens have been reported to induce phase II detoxification enzymes in myocardial cells, ³² IMR-32 cells, ³³ and HepG2 cells. ³⁴ Phase II proteins such as HO-1 and GPx neutralize the reactive oxygen species and protect the cells from oxidative injury. These proteins provide cells with the ability to mount a prolonged and sustained defense against the deleterious effect of oxidants such as superoxide and H₂O₂. In our study, 17β-E₂ upregulated HO-1 and GPx2 gene expression in H₂O₂-stressed cells, and this induction was inhibited by ER antagonist ICI, indicating a receptor-mediated regulation of gene expression. Previous studies have shown both ER-dependent and -independent mechanisms for HO-1 expression. In H9c2 myocardial cells ³² and HepG2 cells, ³⁴ the expression of phase II genes was ER dependent, but in IMR-32 cells, ³³ the ER was not required for phase II gene expression. 

Apart from the effect on phase II genes, we also show that 17β-E₂ modulates intracellular levels of SOD. SOD is an enzyme that is involved in the cellular defense against oxidative stress by detoxification of ROS. SOD has three isoforms, cytosolic Cu/Zn SOD (Cu/ZnSOD), mitochondrial manganese SOD (MnSOD), and extracellular SOD (ecSOD). Reduction in MnSOD levels is associated with disruption of mitochondrial function, accelerated DNA strand breakage, and precocious neuronal degeneration. ³⁵ In our study, H₂O₂ stress reduced the levels of the mitochondrial MnSOD by more than 50%, indicating that this could be a contributing factor to mitochondrial dysfunction. 17β-E₂ treatment significantly alleviated the loss of MnSOD and of Cu/Zn SOD in H₂O₂-treated cells, indicating an additional role for estrogens in maintaining cellular antioxidant function. 

In summary, our data suggest that activation of ERβ protects RPE cells from oxidative damage in two phases, a quick or immediate phase that involves rapid translocation of ligand-receptor complex to mitochondria to protect the mitochondrial pore and function and a prolonged but more sustained phase that involves upregulation of ERβ and phase II antioxidant genes such as HO-1 and GPx2. Our results suggest that the effects of 17β-E₂ are ERβ specific and open avenues for using selective ERβ ligands for reducing oxidative stress, an important factor in the pathogenesis of AMD. 

Footnotes

 Supported by GTx Inc.

Footnotes

 Disclosure: A. Giddabasappa, GTx Inc. (F, I, E); M. Bauler, GTx Inc. (F, I, E); M. Yepuru, GTx Inc. (F, I, E); J. Dalton, GTx Inc. (F, I, E); J. Eswaraka, GTx Inc. (F, I, E); E. Chaum, GTx Inc. (F)