17β-Estradiol Protects against Oxidative Stress-induced Cell Death through the Glutathione/Glutaredoxin-dependent Redox Regulation of Akt in Myocardiac H9c2 Cells*

The GSH/glutaredoxin (GRX) system is involved in the redox regulation of certain enzyme activities, and this system protects cells from H2O2-induced apoptosis by regulating the redox state of Akt (Murata, H., Ihara, Y., Nakamura, H., Yodoi, J., Sumikawa, K., and Kondo, T. (2003) J. Biol. Chem. 278, 50226–50233). Estrogens, such as 17β-estradiol (E2), play an important role in development, growth, and differentiation and appear to have protective effects on oxidative stress mediated by estrogen receptor α (ERα). However, the role of the ERβ-mediated pathway in this cytoprotection and the involvement of E2 in the redox regulation are not well understood. In the present study, we demonstrated that E2 protected cardiac H9c2 cells, expressing ERβ from H2O2-induced apoptosis concomitant with an increase in the activity of Akt. E2 induced the expression of glutaredoxin (GRX) as well as γ-glutamylcysteine synthetase, a rate-limiting enzyme for the synthesis of GSH. Inhibitors for both γ-glutamylcysteine synthetase and GRX and ICI182,780, a specific inhibitor of ERs, abolished the protective effect of E2 on cell survival as well as the activity of Akt, suggesting that ERβ is involved in the cytoprotection and redox regulation by E2. Transcription of the GRX gene was enhanced by E2. The promoter activity of GRX was up-regulated by an ERβ-dependent element. These results suggest that the GRX/GSH system is involved in the cytoprotective and genomic effects of E2 on the redox state of Akt, a pathway that is mediated, at least in part, by ERβ. This mechanism may also play an antiapoptotic role in cancer cells during carcinogenesis or chemotherapy.

and weakened antioxidative defenses can damage macromolecules such as DNA, lipids, and proteins.
Estrogens play an important role in development, growth, and the differentiation of both female and male secondary sex characteristics (3). Protective effects of estrogen, such as 17␤-estradiol (E 2 ), 3 on oxidative stress have been indicated (4). E 2 regulates longevity signals to enhance resistance to oxidative stress in mice. Inhibitory effects of E 2 on atherosclerosis are mediated by COX-2-derived prostacyclin (5). E 2 induces production of antioxidative enzymes, such as superoxide dismutase (6), ␥-glutamylcysteine synthetase (␥-GCS), and glutathione S-transferase (7). The effects of E 2 are mediated mostly through ER␣, which functions as a ligand-induced transcription factor and belongs to the nuclear receptor superfamily (8). ER␣ binds to a variety of ligands and displays tissue-specific effects through estrogen-response element (ERE). When estrogen-responsive genes do not contain EREs, ER␣ can up-regulate gene expression through AP-1 and Sp1 sites (9). Another ER, ER␤, is expressed in cells targeted by E 2 , including cardiomyocytes (10). However, the role of ER␤ in protection against oxidative stress has not been well studied.
Protein thiols act as redox-sensitive switches and are believed to be a key element in maintaining the cellular redox balance. The redox state of protein thiols is regulated by oxidative stress and redox signaling and important to cellular functions. To maintain the cellular thiol-disulfide redox status, living cells possess two major systems, the thioredoxin (TRX)/TRX reductase system and the glutathione (␥-glutamylcysteinyl glycine, GSH)/glutaredoxin (GRX) system (11). GSH is synthesized in two sequential enzymatic reactions that are each catalyzed by a ratelimiting enzyme, ␥-GCS, and GSH synthetase (12). GRX, also known as thioltransferase, was first discovered as a GSH-dependent hydrogen donor for ribonucleotide reductase in Escherichia coli mutants lacking TRX (13). Oxidized GRX is recycled to the reduced form by GSH with the formation of glutathione disulfide and regeneration of GSH by coupling with NADPH and glutathione disulfide reductase (14). GRX functions via a disulfide exchange reaction by utilizing the active site, Cys-Pro-Tyr-Cys, which specifically and efficiently catalyzes the reduction of protein-SSG mixed disulfide (15). GRX also partially shares its function as a redox sensor with TRX (16,17). Recently, we have found that GRX protects against oxidative stress-induced cell death from apoptosis by regulating the redox state of Akt (18).
Akt/protein kinase B is a pleckstrin homology domain-containing serine/threonine kinase and a critical component of an intracellular signaling pathway that exerts effects on survival and apoptosis (19). Akt has been found to be responsive to extracellular signaling factors, oxidative and osmotic stress, irradiation, and ischemic stress (20). Akt can phosphorylate Bad, caspase-9, and forkhead-related transcription factors, leading to an inhibition of apoptosis (21). The unphosphorylated form of Akt is virtually inactive, and phosphorylation at Thr 308 and Ser 473 stimulates its activity. Inactivation of Akt also occurs via dephosphorylation of the two phosphorylation sites by protein phosphatase 2A (PP2A) (22,23). The activation of Akt contributes to the survival of H 2 O 2 -treated cells (24).
It has been reported that the function of ER-mediated transcriptional activity is regulated by redox (25). However, the precise mechanisms of redox regulation in the E 2 -mediated signal pathways have not been clarified. Here we describe a mechanism for the antiapoptotic effect of E 2 through the regulation of the redox state of Akt under oxidative stress. Treatment of cardiac H9c2 cells with E 2 for 18 h protected against H 2 O 2induced apoptosis. E 2 induced the expression of GRX and ␥-GCS, at least in part, through ER␤-mediated regulation. Elevated GSH and GRX levels potentiated the redox of Akt on the exposure of cells to H 2 O 2 .
Cell Culture-H9c2 cells, a clonal line derived from embryonic rat heart, and human breast cancer SK-BR-3 (SKB3) cells, and MDA-MB-231 (MDA) cells, were obtained from the American Type Culture Collection (CRL-1446). Human breast cancer MCF7 cells were from The Cell Resource Center for Biomedical Research Institute of Development, Aging, and Cancer, Tohoku University (Sendai, Japan). H9c2 cells were routinely maintained in Dulbecco's modified Eagle's medium, or MDA and SKB3 cells were maintained in RPMI1640 medium. The cells were supplied with 10% fetal calf serum in a humidified atmosphere of 95% air and 5% CO 2 at 37°C (23).
Cell Viability-Cell viability was determined by a MTT assay as described (26). Briefly, cells (1500 -5000) were placed in 100 l of medium per well in 96-well plates. Four hours after treatment with various concentrations of H 2 O 2 , the cells were incubated for 4 h at 37°C with 3-(4,5dimethylthiazol-e-yl)-2,5-diphenyltetrazolium bromide (652 g/ml) and lysed with 100 l of 20% SDS, 50% N,N-dimethylformamide (pH 4.7) in each well. After an overnight incubation at 37°C, the absorbance at 570 nm was measured. Wells without cells served as blanks.
Nuclear Condensation-For the histochemical analysis, cells were maintained in a four-well Lab Tec Chamber (Nalge Nunc International, Naperville, IL). After treatment with H 2 O 2 , cells were treated with 10 M Hoechst 33342 for 30 min to estimate the extent of nuclear condensation. They were then washed again with PBS. Fluorescence intensity was examined using an Axioskop2 fluorescence microscope (Carl Zeiss, Jena, Germany), and the findings were analyzed using a charge-coupled device camera (Axio-Cam) and AxioVision software.
Morphological Staining-The immunohistochemical analysis to examine the expression of ER␣ and ER␤ was performed as described previously (27). Briefly, cells were fixed with 4% paraformaldehyde in PBS and then preincubated with blocking solution for 1 h at room temperature. For ER␤ 10% normal goat serum and 1% bovine serum albumin in PBS and for ER␣ 500 g/ml of normal goat IgG and 1% bovine serum albumin in PBS were used, respectively. Next, the samples were incubated with the primary antibodies overnight and washed three times with 0.075% Brij 35 in PBS. Then samples were reacted with HRP-goat anti-mouse IgG or HRP-goat anti-rabbit IgG for 1 h at room temperature and washed three times with 0.075% Brij 35 in PBS. HRP sites were visualized with H 2 O 2 and DAB solution or H 2 O 2 and DAB in the presence of nickel and cobalt ions. As a negative control, normal rabbit IgG and normal mouse IgG were used instead of the primary antibodies. The results of immunohistochemistry for ERs were graded as positive or negative, compared with the staining with IgG or serum of a normal rabbit or mouse.
Immunoblot Analysis-Cultured cells were harvested and lysed for 20 min at 4°C in lysis buffer as described previously (17). The supernatants obtained by centrifugation of the lysates at 8000 ϫ g for 15 min were used in subsequent experiments. Protein concentrations were determined using a BCA assay kit (Pierce). Protein samples were electrophoresed on 10, 12.5, or 15% SDS-polyacrylamide gels under reducing conditions, except for thiol-modified protein samples. The proteins in the gels were transferred onto nitrocellulose membranes. The membranes were blocked in Tris-buffered saline (10 mM Tris-HCl (pH 7.5) and 0.15 M NaCl; TBS) containing 0.05% Tween 20 (v/v) (TBST) and 5% (w/v) nonfat dry milk and then reacted with primary antibodies in TBST containing 3% (w/v) bovine serum albumin overnight with constant agitation at 4°C. After several washes with TBST, the membranes were incubated with horseradish peroxidase-conjugated anti-IgG antibodies. Proteins in the membranes were then visualized using the enhanced chemiluminescence detection kit (Amersham Biosciences) according to the manufacturer's instructions.
Akt Activity Assay-Akt activity was assayed using an Akt assay kit (Cell Signaling Technology) according to the manufacturer's directions with GSK3␣/␤ fusion protein (GSK3␣/␤) as a substrate. Phosphorylation of GSK3␣/␤ was assessed by immunoblot analysis using a specific antibody. Briefly, Akt was immunoprecipitated from cell lysates using the anti-Akt antibody, and then the immunoprecipitates were incubated at 30°C for 30 min in an assay mixture containing GSK3␣/␤. Phosphorylated proteins were separated by 12.5% SDS-PAGE and then transferred to nitrocellulose membranes to detect phosphorylated GSK3␣/␤ using an anti-phosphorylated GSK3␣/␤ antibody.
Protein Phosphatase Assay-PP2A activity was assayed spectrophotometrically using a Ser/Thr phosphatase assay kit 1 (Upstate Biotechnology, Inc.) according to the manufacturer's instructions. The phosphopeptide RKpTIRR and p-nitrophenyl phosphate were used as phosphatase substrates.
Determination of Redox States-The redox states of proteins were assessed by modifying free thiol with AMS (28). Briefly, after incubation with or without H 2 O 2 , cell lysates or proteins were treated with trichloroacetic acid at a final concentration of 7.5% to denature and precipitate the proteins as well as to avoid any subsequent redox reactions. The protein precipitates were collected by centrifugation at 12,000 ϫ g for 10 min at 4°C. The pellets were rinsed in acetone and centrifuged twice and then dissolved in a buffer containing 50 mM Tris-HCl (pH 7.4), 1% SDS, and 15 mM AMS. Proteins were then separated by 10% SDS-PAGE without using any reducing agents and blotted to nitrocellulose membranes. Proteins in the membranes were visualized by immunoblotting as described above.
Northern Blot Analysis-A 764-bp DNA fragment (bp 865-1628) of fulllength ␥-GCS heavy subunit cDNA was obtained by digestion with PstI (29). The probes were radiolabeled with 32 P using a random primer labeling kit (Takara Biomedicals, Shiga, Japan). The isolation of cytoplasmic RNA and Northern blotting were essentially performed as described by Sambrook et al. (30). Isolated RNAs (30 g) were electrophoresed on a 1% agarose gel containing 0.6 M formaldehyde, transferred to a nylon membrane, and then hybridized with 32 P-labeled probes. Autoradiographed membranes were analyzed using a BAS5000 bioimage analyzer (Fuji Photo Film). A specific system for the amplification of mRNA was also used: an mRNA-selective PCR kit (Takara-Biomedicals; distributed by BioWhittaker, Europe). It had a total volume of 25 l, comprising 12.5 l of 2ϫ buffer II, 5 mM MgCl 2 , 0.5 M PE1 and PE2, 1 mM each dNTP, 0.4 units of RNAsin (40 units/l), 0.5 units of Rtase XL (5 units/l), and 0.5 units of optimized Taq. As material, 1 g of total RNA extracted from the cells was used. The 330-bp oligonucleotides for GRX (rat GRX sequence, accession number AF167981) were obtained using as a sense primer 5Ј-GCA TGG CTC AGG AGT TTG TGA ACT GCA AGA TTC AG-3Ј and, as an antisense primer, 5Ј-CCT TTC ATA ACT GCA GAG CTC CAA TCT GCT TCA GC-3Ј. The 410-bp oligonucleotides for ␤-actin (rat ␤-actin sequence, accession number BC063166) were obtained using 5Ј-GAG CTA TGA GCT GCC TGA CG-3Ј and 5Ј-AGC ATT TGC GGT GCA CGA TG-3Ј. The 410-bp oligonucleotides for ␤-actin (human ␤-actin sequence, accession number NM_001101.2) were obtained using 5Ј-GAG CTA CGA GCT GCC TGA CG-3Ј and 5Ј-AGC ATT TGC GGT GCA CGA TG-3Ј. The 325-␤ oligonucleotides for ER␣ (rat ER␣ sequence, accession number AY280663.1) and 280-bp oligonucleotides for ER␤ (rat ER␤ sequence, accession number U57439) were obtained using 5Ј-GCT CTT GAC AAA CCC A-3Ј and 5Ј-GCG GCG TTG AAC TCG TAG-3Ј and 5Ј-GGC TGA GGA AAG CAC CTG TC-3Ј and 5Ј-GCG GCG TTG AAC TCG TAG-3Ј, respectively. Similarly, the sense and antisense primers for the oligonucleotides for human ER␣ (accession number NM_000125.2) were 5Ј-GAC TAT GCT TCA GGC TAC C-5Ј and 5Ј-GGT TCC TGT CCA AGA GCA AG-3Ј, whereas those for human ER␤ (accession number NM_001437.1) were 5Ј-GTG GTC CAT CGC CAG TTA TC-3Ј and 5Ј-GCA CTT CTC TGT CTC CGC AC-3Ј, respectively. The RT-amplification was carried out as follows: 30 min at 50°C for the reverse transcription, denaturation for 5 min at 95°C, and then a succession of 23 (35 for ERs) cycles as follows: 30 s at 95°C, 40 s at 65°C (55°C for ERs), and 90 s at 72°C. Amplification took place after 5 min at 72°C.
Determination of Cellular Glutathione Levels-Total GSH and GSSG levels were measured using a total glutathione quantification kit (Dojindo Molecular Technologies, Inc.) according to the manufacturer's directions. Briefly, 5,5Ј-dithiobis-(2-nitrobenzoic acid) and GSH react to generate 2-nitro-5-thiobenzoic acid. The concentration of GSH in the sample solution was determined by measuring absorbance at 412 nm. For quantification of GSSG, cell lysates were treated with 2-vinylpyridine and triethanolamine to block the sulfhydryl residue of GSH. GSSG in the sample solution was reduced to GSH using a reducing mixture containing GSSG reductase and NADPH, and the levels of GSSG were determined photometrically as for GSH.
3-Phosphoinositide-dependent Protein Kinase-1 Activity-3-Phosphoinositide-dependent protein kinase-1 (PDK1) activity was estimated using an assay kit according to the manufacturer's instructions (Upstate Biotechnology). Briefly, recombinant human active PDK1 (Upstate Biotechnology) protein was incubated first with inhibitors for 30 min at 30°C and then with inactive glucocorticoid-inducible kinase 1 (SGK1) for 30 min at 30°C. Next, the PDK1-dependent SGK1 kinase activity was estimated by incubating the reaction mixture with glycogen synthase kinase 3 (GSK3) peptide as a substrate for 10 min at 30°C in the presence of [␥-32 P]ATP.
Generation of Luciferase Reporter Constructs-A 2.0-kb fragment of the human GRX gene promoter (Ϫ2023 to Ϫ22) (31) was amplified by PCR using Pfu turbo DNA polymerase (Stratagene). The primers used were as follows: a forward primer (5Ј-GGA CTG AGT GAG AGG CAG ACA ATA GTC TCC-3Ј) and a reverse primer (5Ј-CGG GAA GAA TCC TCA GTT GCA GGT ATT GCT TGG-3Ј). The PCR product was subcloned into pUC18 to obtain pUC18-pro-GRX. PUC18-pro-GRX was digested with HindIII, and the resulting fragment containing the promoter region from Ϫ2023 to Ϫ22 was inserted into the HindIII site of the reporter vector pGL3-Basic (Stratogene) to give pGL3-pro-GRX. To generate a deleted version of the luciferase reporter construct (pGL3-pro-GRX-del), pGL3-pro-GRX was digested with KpnI and PvuII (Takara Biomedicals). Site-directed mutagenesis for luciferase vectors was performed with pGL3-pro-GRX (Ϫ2023 to Ϫ22) as a template by using a QuikChange site-directed mutagenesis kit (Stratagene). The oligonucleotides used were as follows: electrophoretic response element (EpRE)-like 1 forward (5Ј-GCT CCC CCT CCG GGA CTC AGA ATC TGG-3Ј) and EpRE-like 1 reverse (5Ј-CCA GAT TCT GAG TCC CGG AGG GGG AGC-3Ј). The nucleotide sequence was confirmed by sequencing with an ALFexpress II system (Amersham Biosciences).
Luciferase Activity Assay-Each vector was introduced into H9c2 cells by using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. After 12 h of transfection, cells were harvested for 24 h and then treated with E 2 (100 nM) or left untreated for 18 h. Then luciferase activity was assayed with cellular extracts by using a dual luciferase reporter assay system (Promega).
Electrophoretic Mobility Shift Assay-The electrophoretic mobility shift assay for the GC box and EpRE-like 1 element was performed as described (32). Briefly, oligonucleotides were annealed to doublestranded oligonucleotides and then labeled with [␥-32 P]ATP using T4 polynucleotide kinase. Oligonucleotides specific for the GC box and EpRE-like 1 element were prepared according to the nucleotide sequence of the human GRX promoter region. Oligonucleotides used were as follows: EpRE-like 1 element, 5Ј-CCC TCC GTG ACT CAG AAT CTG GCT TTC-3Ј; mutated EpRE-like 1 element, 5Ј-CCC TCC GGG ACT GTA AGC ACT TTA TGC TTC-3Ј. Binding reactions were performed in 15 l of reaction mixture (25 mM Tris, pH 7.0, 6.25 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM dithiothreitol, 50 mM KCl, and 10% glycerol) containing 10 g of nuclear extract and 25 ng of labeled oligonucleotide. For the supershift assay, specific antibodies were added to the reaction mixture during the binding reaction for 30 min.
Statistical Analysis-Data were presented as means Ϯ S.D. Differences were examined by using Student's t test. A value of p Ͻ 0.05 was considered significant.

RESULTS
Expression of ERs-The expression of ERs in H9c2 cells was estimated immunohistochemically and genetically. Fig. 1 shows the results of the immunohistochemical analysis. Unlike MCF7 cells, which are known to express both ER␣ (Fig. 1A) and ER␤ (Fig. 1B), H9c2 cells expressed ER␤ (Fig. 1F) but not ER␣ (Fig. 1E). Fig. 1I shows the results of the RT-PCR analysis. ER␤ mRNA but not ER␣ mRNA was detected in H9c2 cells. On the other hand, both ER mRNAs were detected in MCF7 cells.
Cytoprotective Effect of E 2 on Oxidative Stress-We tested the cytoprotective effect of E 2 on oxidative stress-induced apoptosis in H9c2 cells. Hydrogen peroxide induces apoptosis or early mitochondrial dys-function in cardiac H9c2 cells (32,33). Since 10% fetal calf serum, required for maintaining cultured cells, reduces oxidative stress modification of cells, in order to observe the effect of E 2 on H 2 O 2 -induced oxidative stress, the concentration of fetal calf serum in the medium was changed from 10 to 0.5% in the experiments that followed. As shown in   (Fig. 2B). The increase in cell viability caused by E 2 observed in 18 h continued until 24 h and then declined until 36 h (Fig. 2C). Morphologically, H 2 O 2 -induced DNA condensation was observed (Fig. 3, A versus E). E 2 protected against DNA condensation (Fig. 3, B versus F). ICI182,780, an ER antagonist, abolished the pro-tective effect of E 2 (Fig. 3, C versus G). PPT (0.5 mM), a specific inhibitor of ER␣, had no apparent influence on the protective effect of E 2 (Fig. 3, D versus H). These results suggest that the protective effect against oxidative stress observed on treatment of the cells with E 2 for 18 h involves transcriptional regulation mediated by ER␤ through a genomic pathway in this cell line. Unless otherwise indicated, subsequent experiments on the effect of E 2 were done by incubating the cells with 100 nM E 2 for 18 h. E 2 Stimulated the Activity of Akt in Response to H 2 O 2 -The Akt cascade is known to mediate the survival function. The Akt signal is involved in both the genomic (34) and the nongenomic pathway of E 2 (35). We tested the involvement of Akt in the cytoprotective effect of E 2 in ER␤-positive H9c2 cells. Phosphorylation of Akt (Ser 473 ) was promoted by H 2 O 2 in 10 min by 1.7-fold, and the control level was reached in 60 min (Fig. 4, A and B). Prior treatment with E 2 for 18 h resulted in a further increase in the H 2 O 2 -induced phosphorylation of Akt in 10 min by 4.1-fold, and the phosphorylation continued until 30 min (Fig. 4, A and B, lanes 6 versus lanes 2 and lanes 7 versus lanes  3, respectively). ICI182,780 abolished the effect of E 2 (Fig. 4, C and D,  lanes 6 versus lanes 4). The H 2 O 2 -induced enhancement of Akt activity estimated using GSK3␣/␤ as a substrate was increased by E 2 (Fig. 4, E and F, lanes 5 versus lanes 2 and lanes 6 versus lanes 3,  respectively), concomitant with the increase in the phosphorylation of Akt. The activity of PDK1, upstream of Akt, was stimulated by H 2 O 2 ; however, E 2 had no apparent effect on the activity of PDK1 (Fig. 4G). The phosphorylation of Akt is regulated by PP2A (18). The activity of PP2A assayed spectrophotometrically using RKpTIRR and p-nitrophenylphosphate as substrates was not affected by H 2 O 2 and E 2 (Fig. 4H). The data suggest that the change in the activity of PP2A  is not involved in the up-regulation of the phosphorylation of Akt by E 2 .
It has been reported that inactive Akt develops a redox-sensitive intramolecular disulfide bond close to its activation loop (18), and recently we found that the redox state of Akt is modulated by H 2 O 2 (19). Fig. 5A shows the redox state of Akt assessed by modifying free thiol with AMS. In control cells, Akt existed mostly in an oxidized form (lane 1). Treatment of cells with H 2 O 2 resulted in a further increase in an oxidized form of Akt (lanes 2 and 3). In the cells treated with E 2 for 18 h, Akt existed more in a reduced form (lane 4). The reduced form of Akt, once decreased by H 2 O 2 for 30 min, was restored again in 60 min (lanes 5 and 6). The data suggested that E 2 maintains Akt in a reduced form under oxidative stress. The redox state of Akt is regulated by the GSH/GRX system, and this system protects cells against H 2 O 2 -induced apoptosis by preventing the association of Akt with PP2A (19). Then we estimated the effect of E 2 on the phosphorylation of Akt in the presence of buthionine sulfoximine (BSO), a specific inhibitor of ␥-GCS, or cadmium, an inhibitor of GRX. ␥-GCS is a rate-limiting enzyme of GSH synthesis. The effect of E 2 on the phosphorylation was abolished both by BSO (Fig.  5B) and by cadmium (Fig. 5C). These results suggest that E 2 increases the levels of GSH/GRX to protect cells against oxidative stress. E 2 Induces the Expression of ␥-GCS and GRX-We tested if E 2 increases the levels of GSH and GRX. E 2 increased the levels of GSH Phosphorylation of Akt was detected by immunoblot analysis using specific antibodies as described under "Materials and Methods." B and D, band intensity was estimated densitometrically, and the phosphorylation rates are expressed as the intensity of phosphorylated Akt relative to total Akt (Akt-p/Akt). E, activity of GSK3␣/␤. The kinase activity of Akt was measured based on the phosphorylation of GSK3␣/␤ as described under "Materials and Methods." F, band intensity was estimated densitometrically, and the phosphorylation rates are expressed as arbitrary units. G, the activity of PDK1. The experimental conditions are the same as in E. H, the activity of PP2A. The activity of PP2A was measured as described under "Materials and Methods." The data are the mean Ϯ S.D. of three independent analyses (B, D, and F). *, p Ͻ 0.05 compared with cells without E 2 at each time point; **, p Ͻ 0.05 compared with cells with H 2 O 2 and E 2 without ICI182,780; ***, p Ͻ 0.05 compared with cells with H 2 O 2 without E 2 at each time point. (Fig. 6A). The level of GSH was 24.8 Ϯ 4.0 nmol/10 6 cells in control cells and 38.5 Ϯ 5.2 nmol/10 6 cells in cells treated with 100 nM E 2 for 18 h. The level of GSSG in the control cells was ϳ2 nmol/10 6 cells and was not changed by E 2 treatment (data not shown). The expression of ␥-GCS was up-regulated by 100 nM E 2 by 1.6-fold in 18 h (Fig.  6, B and C). Similarly, 100 nM E 2 increased the expression of GRX by 1.8-fold in 18 h (Fig. 6, D and E). ICI182,780 abolished the E 2 -dependent up-regulation of GSH synthesis as well as GRX synthesis (Fig. 6, A-E). It is suggested that the redox state of Akt is regulated by an E 2 -dependent enhancement of the GRX/GSH system.
The Gene Promoter Activity of GRX Is Regulated by E 2 via an EpRElike Element-As reported by Montano et al. (7), the expression of the ␥-GCS heavy (catalytic) subunit is up-regulated by E 2 via an EpRE (5Ј-(G/A)TGACNNNGC(G/A)-3Ј), not by an ERE. To investigate the mechanism of the transcriptional regulation of GRX by E 2 , we used a 2.0-kb genomic fragment containing the promoter region of GRX inserted into a luciferase vector, pGL3 Basic. The promoter region contains no apparent ERE or EpRE. There were two EpRE-like sites (EpRElike 1 (Ϫ1380 to Ϫ1370; GTGACTCAGAA) and EpRE-like 2 (Ϫ347 to Ϫ337; GTGAGTAAGCA)) and Sp1 (Ϫ1217 to Ϫ1208, GCCCCGC-CTC). The luciferase activity of the cells previously treated with E 2 for 18 h was almost lost when the EpRE-like 1 site was deleted or mutated (Fig. 7). Deletion of EpRE-like 2 or Sp1 had no apparent effect on the E 2 -induced up-regulation of the luciferase activity. E 2 Up-regulated the ER␤-EpRE-like 1 Complex Formation-To investigate the importance of the EpRE-like elements in the E 2 -induced expression of GRX, an electrophoretic mobility shift assay was performed with nuclear extracts from the cells treated with E 2 for 18 h using 32 P-labeled oligonucleotides designed for EpRE-like 1. As shown in The addition of the anti-ER␤ antibody caused the ER␤-DNA-binding complex to disappear (lane 12), indicating the involvement of ER␤ as a transcription factor that bound to the EpRE-like 1 site. The EpRE-like 1 of GRX did not bind with Nrf2, Sp1, c-Jun, or c-Fos (lanes 9 -11), different from the EpRE site of the ␥-GCS heavy subunit (7). On the other hand, neither EpRE-like 2 site nor the Sp1 site was stimulated by E 2 (data not shown).  Effects of E 2 on levels of GSH, the ␥-GCS heavy subunit, and GRX were estimated in the presence or absence of ICI182,780, as described under "Materials and Methods." A, cells were treated with 100 nM E 2 for 0 -24 h, and the levels of GSH in the cell lysates were estimated. Cells were incubated with 100 nM E 2 for 6 h for the analysis of the expression of the ␥-GCS heavy subunit by Northern blotting (B) and that of GRX by RT-PCR (D). The expression of ␥-GCS was expressed as relative intensity (percentage of control) (C), and that of GRX was expressed as the intensity of GRX/ ␤-actin (E). Each datum is a mean Ϯ S.D. of three independent analyses. *, p Ͻ 0.05 compared with untreated cells.
Important Role of ER␤ in Other Cells-To further confirm the role of ER␤ in protection against oxidative stress through redox regulation of Akt, we employed human breast cancer cells, SK-BR-3 and MDA-MB-231 cells. As shown in Fig. 9A, an RT-PCR analysis revealed that these cells mainly expressed ER␤ mRNA. A stimulatory effect of E 2 on the activity of Akt was observed in these cells (Fig. 9, B and C). However, ICI182,780 abolished the protective effect of E 2 (Fig. 9, B and C). E 2 induced the expression of GRX (Fig. 9, D and E). The results suggested that the cytoprotective effect of E 2 is mediated through redox regulation of Akt activity in ER␤-expressing cells.

DISCUSSION
ER␤-mediated Cytoprotection against Oxidative Stress-Estrogenic hormones are required for the growth and differentiation of female FIGURE 7. The EpRE-like element is important for the E 2 -dependent induction of the GRX promoter in H9c2 cells. Left, schematic representation of luciferase vector constructs for the human GRX promoter. Each luciferase vector construct was generated as described under "Materials and Methods." Right, luciferase activity of the vector constructs for the human GRX gene promoter in H9c2 cells. The cells were transiently transfected with the GRX promoter-luciferase gene fusion plasmids. After the transfection, luciferase activity was assayed with cellular extracts as described under "Materials and Methods." Each value represents the mean of at least three independent experiments, and the S.D. was always within 10% of the mean. *, p Ͻ 0.05 compared with control. FIGURE 8. The EpRE-like 1 element is responsive to E 2 in electrophoretic mobility shift assays. H9c2 cells were incubated with 100 nM E 2 for 18 h, and the nuclear extracts were prepared as described under "Materials and Methods." 32 P-Labeled oligonucleotides specific to EpRE-like elements 1 and 2 of the GRX gene promoter were prepared and incubated with each nuclear extract and then subjected to a 5% nondenatured PAGE. In lanes 1, 5, and 6, the nuclear extract (NE) was free. In lanes 6, 8, and 14, 32 P-labeled mutant oligonucleotides were used. In lanes 9 -12, ϩAb, specific antibodies were added to the reaction mixture during the binding reaction for the supershift assay. Arrowhead, protein-DNA complex.
reproductive tissues, contribute to male fertility, and play a role in maintaining cardiovascular, skeletal, and neural cell functions (9). Estrogen has been widely used to regulate fertility, relieve postmenopausal symptoms, and decrease the incidence and recurrence of mammary tumors. The ERs were the first members of the nuclear receptor family to be identified. ER␣ has been well characterized and plays a major role in E 2 -mediated genomic actions in both reproductive and nonreproductive tissues. ER␣-mediated cytoprotection against oxidative stress-induced cell damage has been reported in neurological cells (33,34) and breast cancer cells (35). On the other hand, the role of ER␤ is not well understood. A report using microarray analyses showed that most of the genes regulated by ER␤ are distinct from those regulated by ER␣ in response to E 2 and selective estrogen receptor modulators (36). ER␤ regulates plasminogen activator inhibitor-1 in endothelial cells, and a clinical evaluation of ER␤ was suggested as a prognostic or predictive factor of drug resistance in breast cancer (37). These results suggest a significant role for ER␤ in the regulation of cellular function, although the function of ER␤ and its precise mechanism are still unclear (3). Thus, this is the first report aimed at the significant role of ER␤-mediated signals of E 2 in redox regulation in response to oxidative stress.

Involvement of Akt in the Cytoprotection of E2 Mediated by ER␤-
The importance of Akt has been suggested in the cytoprotective effect of E 2 against oxidative stress. This effect of E 2 was rapid and nongenomic in neurological cells (38), vascular endothelial cells (39), and ovarian cancer cells (40). On the other hand, Stoica et al. (41) reported that ER␣-mediated signals up-regulated the expression of Akt in ER␣-positive MCF-7 breast cancer cells. They also demonstrated that Akt-mediated signals up-regulated the expression of ER␣ in these cells, suggesting that Akt plays a central role in the growth and survival of breast cancer cells; however, the mechanism by which Akt is activated by E 2 was not fully characterized.
In the present study, we were interested in the possible involvement of Akt signals in the ER␤-mediated anti-apoptotic effect against oxidative stress. We employed H9c2 cells that apparently express only ER␤ (Fig. 1). We found that 1) H 2 O 2 -induced apoptosis was prevented when the cells were incubated with E 2 for over 18 h; 2) the anti-oxidative effect of E 2 was mediated by a genomic pathway through ER␤; and 3) E 2 retained the level of phosphorylated Akt in response to H 2 O 2 via the GSH/GRX system.
Role of the GSH/GRX System in ER␤-mediated Akt Signals-We reported previously a role for the GRX/GSH system in the regulation of Akt phosphorylation (19). Akt is a Ser/Thr protein kinase with antiapoptotic and oncogenic activities. Akt is activated through a growth factor receptor-mediated activation of the phosphatidylinositol 3-kinase pathway (21). The unphosphorylated form of Akt is virtually inactive, and phosphorylation at Thr 308 and Ser 473 stimulates its activity. Inactivation of Akt also occurs via dephosphorylation of the two phosphorylation sites by PP2A (23,24). The activation of Akt contributes to the survival of H 2 O 2 -treated cells (25). H 2 O 2 induces oxidation of Akt at Cys 297 and Cys 311 , and the oxidized form of Akt can be dephosphorylated by PP2A (19). PP2A is a major Ser/Thr phosphatase implicated in the regulation of many cellular processes, including the regulation of different signal transduction pathways, cell cycle progression, DNA replication, gene transcription, and protein translation (42). Yasukawa et al. have reported that Akt is also inactivated by S-nitrosylation at Cys 224 in NO donor-treated cells (43). Furthermore, we recently reported that the phosphorylation of Akt is down-regulated by cytoplasmic calcium (32).  Fig. 1. B, phosphorylation of Akt. The effect of E 2 on the phosphorylation of Akt under oxidative stress was estimated by immunoblot analysis using specific antibodies as described under "Materials and Methods." C, band intensity was estimated densitometrically, and the phosphorylation rates are expressed as the relative intensity of phosphorylated Akt to total Akt (Akt-p/Akt). The data are the mean Ϯ S.D. of three independent analyses (B and D). D, gene expression of GRX. The effect of E 2 on levels of the GRX was estimated as described under "Materials and Methods." E, band intensity was estimated densitometrically and expressed as the intensity of GRX/␤-actin. Each datum is a mean Ϯ S.D. of three independent analyses. *, p Ͻ 0.05 compared with cells with H 2 O 2 without E 2 ; **, p Ͻ 0.05 compared with cells with E 2 without ICI182,780.