ABSTRACT
Redox regulation of transcription factors has recently been demonstrated for AP-1, NF-[kappa]B, Sp-1 and glucocorticoid receptor in vitro and in vivo. The redox state in estrogen-dependent cells possibly influences the function of estrogen receptor (ER), and the regulation of the function of ER is essential for understanding of growth and differentiation of these cells, as well as promotion and progression of estrogen-associated cancer. In this paper, we first analyzed the effects of redox state on transcriptional activity of ER in terms of pS2 mRNA expression and transfection of ERE-CAT plasmid in human breast cancer cells. Addition of H2O2 at low concentrations lowered levels of pS2 mRNA and also down-regulated ERE-CAT activity, which was recovered by transfection of thioredoxin (TRX) expression vector. Next, the transfection of antisense TRX plasmid diminished ERE-CAT activity, and the activity was recovered by co-transfected sense TRX. Furthermore, specific DNA binding activity of recombinant ER was inhibited by sulfhydryl-modifying reagents and restored by the addition of recombinant TRX protein in electrophoretic mobility shift assay. These results in vitro and in vivo revealed that the transcription activity of ER is strongly influenced by its redox state, which is reversibly modulated by endogenous redox effector protein, TRX.
Estrogen and its receptor play an important role in regulating growth and differentiation of the normal epithelium as well as in carcinogenesis of steroid hormone-dependent cancers. In breast cancer, the expression of the estrogen receptor (ER) is closely associated with the cancer biology, especially the development of tumor; i.e., presence or absence of ER is associated with growth of breast cancer cells: breast carcinomas which lack ER expression often reveal more aggressive phenotypes (1 -3 ). Moreover, expression of ER in tumor tissues is a good predictor of prognosis in endocrine treatment (4 ), which aims to inhibit the mitogenic stimulus produced by the ER bound to estrogen.
ER belongs to a family of nuclear receptors that have steroid and thyroid hormones as known ligands, and it regulates the transcription of various genes as a transcription factor upon binding to estrogen response elements (ERE) upstream of the target genes (5 -7 ). The transcriptional regulation of the ER gene is one of the major factors in mediating the estrogen effects on cells. We recently characterized the expression of ER mRNA in human primary breast cancers, and found that transcriptional regulation from a specific distal promoter is responsible for enhanced expression of ER related to breast carcinogenesis (8 ). Besides transcriptional regulation, post-transcriptional events are important in the regulation of ER activity, e. g., phosphorylation of ER protein at specific positions has been shown to influence the function of ER as a transcription factor (9 -11 ). The post-transcriptional regulation of ER by means other than phosphorylation is also an integral factor in understanding its function in cells.
Thioredoxin (TRX) stimulates the growth of various normal and cancer cell lines (12 -15 ), and expression of TRX increases in primary lung cancer, colon cancer, cervical neoplastic squamous cells and hepatocellular carcinoma (16 -18 ). Very recently, it has been reported that the transfection of dominant negative mutant TRX reversed a transformed phenotype of ER-positive human breast cancer cell line, MCF-7, suggesting that endogenous TRX plays an important role in malignant development of breast cancer (19 ). However, the endogenous molecular targets of TRX in breast carcinogenesis have not been identified yet.
Human TRX was originally discovered as a T cell growth factor secreted from HTLV-I-infected T cell (20 ). TRX is a 12 kDa ubiquitous protein that controls the redox state of various target proteins by means of thiol-disulfide exchanges as a constituent of cellular antioxidant defence systems (21 ,22 ). The redox-active sulfhydryls in TRX are located at a highly-conserved active-site sequence -Trp-Cys-Gly-Pro-Cys- (22 ). The pathway for the reduction of a protein disulfide by TRX entails nucleophilic attack by one of the active-site sulfhydryls, formation of a protein-protein disulfide, and subsequent intramolecular displacement of the reduced target proteins with concomitant formation of oxidized TRX (22 ). Recently, involvement of TRX in transcriptional processes of several genes has been suggested: transient transfection or exogenous application of TRX resulted in inhibition of NF-[kappa]B and induction of AP-1 activity (23 ,24 ), and the glucocorticoid receptor (GR) in isolated rat cytosol was stabilized as a reduced and ligand-binding form when TRX and TRX-dependent thiol-disulfide exchange system was added (25 ). Furthermore, we have recently reported that cellular glucocorticoid responsiveness is coordinately modulated by redox state and TRX levels in vivo (26 ). Thus, redox state in cells may be another important mechanism for post-transcriptional regulation of ER function, thereby playing a critical role in regulating the transcription activity of ER. This is the case, for example, in NF-[kappa]B (27 ), AP-1 (28 ), Sp-1 (29 ) and GR (26 ), since the ER has zinc fingers in its DNA binding domain, as Sp-1 and GR do, and since most cysteine residues converged in the DNA binding domain (nine of 13 cysteines). Here, we report for the first time that the cellular content of TRX determines the transcriptional activity of ER, indicating that the function of ER is highly sensitive to the cellular redox state.
Human mammary tumor cells, ZR-75-1, were cultured in RPMI1640 medium supplemented with 10% fetal calf serum, 100 nM 17[beta]-estradiol, 2 mM l-glutamine, and 2 µg/ml of gentamicin at 37oC in a humidified atmosphere of 5% CO2 in air.
Diamide, dithiothreitol (DTT), N-ethylmaleimide (NEM), and 17[beta]-estradiol were purchased from Sigma Co. (St. Louis, MO); other chemicals from Wako Pure Chemical (Osaka, Japan). Recombinant TRX (rTRX) was produced by the method described previously and kindly provided by Ajinomoto Co. Inc., Basic Research Laboratory, Kawasaki, Japan (30 ). Anti-human ER antibody G-20 was obtained from Santa Cruz Biotech. (Nashanic Station, NJ). All enzymes were purchased from TaKaRa Shuzo (Tokyo, Japan).
The bacterial expression plasmid of GST-hER was constructed from pGEX2T expression plasmid (Pharmacia) and wild type of human ER cDNA kindly supplied by Dr B. S. Katzenellenbogen (University of Illinois, Urbana, IL) (31 ). BamHI and EcoRI sites were designed adjacent to the initiation and termination codons of ER cDNA, respectively, using PCR. The amplified cDNA by PCR was subcloned into the BamHI-EcoRI site of pGEX2T. The sequence of cDNA clone in pGEX2T was confirmed by sequencing using ALF Auto Read Sequencing Kit (Pharmacia). The expression plasmid for TRX, pcDSR[alpha]ADF, and antisense TRX expression plasmid, pASADF, were prepared as described previously (26 ). Antisense TRX expression plasmid pcASADF has been ascertained to reduce the cellular thioredoxin protein levels (26 ). The estrogen-responsive reporter construct pERE-CAT was kindly supplied by Dr H. Oshima. In brief, the oligonucleotides containing ERE tandem repeats (non-coding strand, 5'-GATCCAGGTCAGGATGACCTAGCTACGGATCCAGGTCAGGATGACCTAGCTACGGATC-3') were inserted into a pBLCAT2 CAT expression vector (32 ).
Total RNAs were prepared from cultured cells (1-5 × 106 cells) according to the method of Chomczynski and Sacchi (33 ). Semi-quantitative RT-PCR was carried out using the GeneAmp RNA PCR Kit (Takara Shuzo, Tokyo) as previously reported (34 ). Oligonucleotides used in PCR amplification were as follows: PS1, CCAGACAGAGACGTGTACAG; PS2, TGGGACTAATCACCGTGCTG for pS2 (35 ); hER9, GGTACTGGCCAATCTTTCTC; hER10, AACGCGCAGGTC TACGGTCA for ER (8 ); GAP1, ACATCGCTCAGACACCATGG and GAP2, GTAGTTGAGGTCAATGAAGGG for GAPDH (36 ). The primers were designed to sandwich one intron for specific detection of mRNA. Total RNA (1 µg) was reverse transcribed to synthesize cDNA using random hexamers at 42oC and then subjected to PCR amplification with 0.4 µg each of specific primers and 3 µCi of [[alpha]-32P]dCTP (3000 Ci/mmol) in 50 µl mixtures consisting of 10 mM Tris-HCl pH 8.3, 50 mM KCl, 1.5 mM MgCl2, 0.01% gelatin, and 0.2 mM dNTPs (dATP, dTTP, dGTP, dCTP). PCR comprised 26 cycles for GAPDH and pS2, and 30 cycles for ER, with denaturing at 95oC for 1 min, annealing at 62oC for 1 min, and extension at 72oC for 1 min in each cycle using a GeneAmpTM PCR System 9600 (Perkin Elmer Cetus). The PCR products were then subjected to 5% polyacrylamide gel-electrophoresis, and the radioactivity was evaluated using the Fuji BAS2000 (Fuji Film, Tokyo). A linear relation between radiolabeled PCR products and RNA amounts was ascertained in this study along with the confirmation of reproducibility by a method described in a previous report (8 ,34 ).
Transient transfection was performed as described previously (26 ). Briefly, cells were plated on plastic culture dishes to 30-50% confluency, and medium was replaced with RPMI1640 medium. Plasmid cocktail was mixed with Lipofectin reagent (GIBCO Laboratories) and added to the culture. Total amount of the plasmids was kept constant by adding carrier plasmid, pGEM7Z (Promega). After 24 h of incubation, the medium was replaced with fresh RPMI1640 medium supplemented with 5% FCS, and the cells were further cultured in the presence of 100 nM estradiol for 24 h; CAT activity was determined essentially as described previously (37 ) by the standard method using [14C]chloramphenicol as a substrate. The activities were evaluated by autoradiography with a Fuji Bio-Image Analyzer BAS2000 (Fuji film, Tokyo). The transfection efficiency was normalized by luciferase expression using pGL2-control vector (Promega).
Subsequent to bacterial expression of the protein using expression plasmid of GST-hER, the recombinant GST-fused ER (rER) was isolated according to the manufacturer's recommendation using glutathione-Sepharose column chromatography. Purified rER was dialyzed against 100 mM Tris-HCl buffer, pH 8.0, containing 0.5 mM EDTA, 2 mM DTT, 0.1 mM PMSF, and 10 µM ZnSO4. The partially purified rER was checked by SDS polyacrylamide gel-electrophoresis and western blotting using anti-ER antibody.
EMSA for rER was carried out as described before (37 ). Briefly, the ERE probe was end-labeled with [[alpha]-32P]dCTP using the Klenow fragment of DNA polymerase I. The sequences of the oligonucleotides encompassing ERE are: 5'-GGTTTGGCAAGGGTCACAATGACCTCAACA-3' (upper strand sequence) and 5'-GGTGTTGAGGTCATTGTGACCCTTGCCAAA-3' (lower strand sequence). Recombinant ER (50 ng protein per reaction) was incubated with 1.2 ng of ERE probe labeled with [32P]dCTP using Klenow fragment in a 10 µl reaction mixture containing 10 mM Tris-HCl pH 7.5, 50 mM NaCl, 0.5 mM EDTA, 1 mM MgCl2, 2 mM DTT, 4% glycerol and 500 ng poly (dI-dC) (Pharmacia LKB) for 30 min on ice. Radioinert competitor DNA, rTRX, or anti-ER antibody was added when indicated. The reaction mixture was then loaded onto a 5% non-denaturing polyacrylamide gel containing 0.5* TBE. The gels were run at 200 V for 2 h and dried. Results were visualized using the Fuji Bio-Image Analyzer BAS2000.
The human pS2 gene, which has a cis-acting ERE, was initially characterized as a gene whose expression is specifically controlled by estrogen in breast cancer cell lines (38 ,39 ). Thus, pS2 is considered to be one of the representative target genes of ER in mammary cells. We first examined the effect of oxidative stress on mRNA expression of pS2 by adding H2O2 to an ER-positive breast cancer cell line, ZR-75-1. As shown in Figure 1 , the abundant expression of pS2 mRNA in normal culture conditions was dose-dependently reduced by the addition of H2O2. On the other hand, expression of GAPDH mRNA and viability of cells remained unaffected under oxidative stress.
The effect of TRX on ER-mediated transcription was also confirmed by an in vivo complementation assay using the TRX expression plasmid and an antisense TRX expression plasmid. As indicated in Figure 4 , the transfection of antisense TRX plasmid dose-dependently suppressed the ERE-CAT expression (lanes 1-4), while the cotransfection of TRX expression plasmid canceled the negative effects of the antisense TRX and restored ER-mediated transcriptional activation in a dose-dependent manner (lanes 5-8). Thus, it is strongly indicated that cellular TRX levels are one of the determinants in regulation of the transcription activity of ER, and also that TRX acts as an endogenous auxiliary factor for the ER transactivation function.
Since most cysteine residues of ER converge in the DNA binding domain, we next asked whether the DNA binding activity of ER is regulated by redox state. For this experiment we used EMSA with purified ER protein. The binding of recombinant ER protein (rER) to the oligonucleotide probe encompassing a perfect palindromic vitellogenin ERE was observed as shown in Figure 5 A, and this binding dramatically decreased with the addition of radioinert ERE oligonucleotide but not with the addition of unspecific competitors, AP-1 and Sp-1 oligonucleotides. The binding complex also decreased with the addition of anti-ER antibody (Fig. 5 B, lane 8). These observations indicate the specific binding of rER to ERE oligonucleotide, and in Figure 5 A and B the broad minor band in the lower position is presumably the complex of proteolysed rER. After incubation of protein samples with a thiol specific oxidizing reagent diamide, DNA binding of rER drastically decreased (Fig. 5 B, lanes 2-4), a decrease which was restored by the presence of a thiol reducing agent, DTT (lane 5). N-ethylmaleimide (NEM), a thiol alkylating reagent, also eliminated the DNA binding of rER (lanes 6 and 7). These results strongly suggest that the redox state of cysteine residues of ER is important for its DNA binding activity to ERE.
To examine the direct interaction between ER and TRX proteins in vitro, we performed EMSA using rER and recombinant TRX protein (rTRX). After incubation of ER protein samples with diamide, DNA binding of rER to the ERE oligonucleotide probe sharply dropped with increased concentrations of diamide (Fig. 6 , lanes 2-4). Addition of rTRX to the samples dose-dependently squelched the inhibitory effect of diamide on DNA binding of rER (lanes 6-11); DNA binding was enhanced to a level higher than that in control or in the samples restored by co-presence of 10 mM DTT. Since only two purified proteins, rER and rTRX, were used in this assay system, the results indicate that TRX reduces ER by a direct protein-protein interaction.
Understanding the ER-mediated regulation of various genes is indispensable for the biology of breast epithelium cells and also important for breast carcinogenesis. We previously found that the enhanced expression of ER mRNA in human ER-positive breast carcinomas can be ascribed to transcription from a distal upstream promoter of the ER gene, suggesting a tumor-specific transcription mechanism of ER in breast carcinogenesis (8 ). Apart from the transcriptional regulation of the ER, post-transcriptional events such as phosphorylation or redox-dependent modification are also important determinants in the biological function of ER. Here we present evidence that ER-mediated gene expression is coordinately modulated by cellular redox state and TRX levels using breast cancer cells. Moreover, we indicate that the redox regulation of ER is ascribable to thiol-mediated modification of ER by oxidants and TRX.
As we have found in transcriptional regulation of ER in human breast cancer, accumulated evidence suggests that a tumor-specific mechanism is important in post-transcriptional regulation of ER. During tumor promotion and progression of cancer, the cells are stimulated in autocrine or paracrine manners by various growth factors and cytokines, resulting in increased oxidative stress. Oxidative stress induces the expression of TRX (40 -42 ), and enhanced expression of TRX has been observed in various cancers (16 -18 ). Recently, Gallegos et al. reported that a dominant-negative mutant of TRX inhibited tumor cell growth and reversed a transformed phenotype of human breast cancer cells, suggesting that endogenous TRX may play a crucial role in developing breast cancer (19 ). Since ER plays an important role in the proliferation and progression of ER-positive breast cancer cells, our results suggest that ER may be one of the major targets of TRX in breast cancer, and that TRX, at least in part, is involved in cell proliferation and malignant development of breast cancer through redox regulation of ER. Involvement of other proteins such as Ref-1 (43 ) in the redox regulation of ER will also be an important target in future study. Alteration of the redox regulation of ER in human breast carcinomas, compared with that in normal mammary gland, needs to be further investigated, particularly its association with clinical characteristics of breast cancer.
Our results showed that the overexpression of TRX resulted in a significant increase in transcriptional activity of ER, indicating that at least a part of cellular ER is sequestered in a transcriptionally inactive state, probably due to oxidative modification of the thiol residues in the DNA binding domain. All cysteine residues within proteins including transcription factors must not be equally sensitive to redox state in the cells, and ER seems to be one of the most sensitive proteins playing a physiologically important role. We therefore propose that the function of ER is controllable in not only a negative but also a positive direction by cellular redox state or various reducing regents. In the case of estrogen-dependent cancer, treatment with antiestrogen is a standard endocrine therapy. However, while 70-80% of ER-positive cancers respond to endocrine therapy, the rest are resistant. Pharmacological modulation of ER function through redox state may be a useful tool for future treatment of these resistant cases. Moreover, the cellular levels of TRX also appear to influence the sensitivity of several cancer cells to a variety of superoxide-generating anticancer drugs; for example, cisplatin-resistant cancer cells, which often show high levels of TRX, become susceptible to anticancer drugs after antisense-mediated sequestration of TRX (44 ). Thus, TRX can be not only a novel target in anticancer therapy but a useful marker for diagnosis and prognosis, particularly in steroid hormone-dependent cancers.
We thank Drs Benita S. Katzenellenbogen (University of Illinois, Urbana, IL), Hisaji Oshima (NIH, Bethesda, MD) for providing plasmid, and Dr Masakazu Toi (The Tokyo Metropolitan Institute for Medicinal Sciences) for providing cell line. We thank Drs Kei Nakachi, and Hirota Fujiki (Saitama Cancer Center Research Institute) for valuable comments. This study was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan and for a 2nd-Term Comprehensive 10-Years Strategy for Cancer Control from the Ministry of Health and Welfare of Japan.
*To whom correspondence should be addressed. Tel: +81 48 722 1111; Fax: +81 48 722 1739; Email: shayashi@saitama-cc.go.jp
REFERENCES