Resveratrol, a polyphenolic compound found in grapes and wine, is an agonist for the estrogen receptor

Cell Culture.

MCF-7 cells, subclone WS8 (estrogen receptor-positive), MDA-MB-231 cells, subclone 10A (estrogen receptor-negative), and T47D cells, subclone A18 (estrogen receptor-positive) are derived from human breast adenocarcinomas and were provided by V. Craig Jordan (Northwestern University Medical School, Chicago). MCF-7 and MDA-MB-231 cells were grown in MEM supplemented with nonessential amino acids, 10 mM Hepes, and 5% calf serum. T47D cells were grown in RPMI 1640 medium supplemented with nonessential amino acids, 10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin sulfate. BG-1 (human ovarian adenocarcinoma, estrogen receptor-positive) cells, provided by Jeff Boyd, University of Pennsylvania Medical Center, Philadelphia, were grown in DMEM/Ham’s F-12 medium with 10% fetal bovine serum. Estrogen-depleted media contained no phenol red and were supplemented with sera extracted three times with dextran-coated charcoal.

Estrogen Receptor Binding Assay.

Cytosol was prepared from MCF-7 cells grown for 4 days in estrogen-depleted medium. Cells were incubated for 30 min in 10 volumes of ice-cold hypotonic TESH buffer (10 mM Tris⋅HCl/1.5 mM EDTA/1 mM DTT, pH 7.4) (11) containing 1 mM phenylmethylsulfonyl fluoride and 10 μg/ml leupeptin and lysed by 10 passages through a 25-gauge hypodermic needle. After ultracentrifugation (1 h at 100,000 × g), the cytosol (supernatant) fraction was adjusted to 10% glycerol and stored at −70°C.

Aliquots containing 12 μg of protein were incubated with various concentrations of 16α-[¹²⁵I]-iodo-3,17β-estradiol (New England Nuclear; 2200 Ci/mmol) and resveratrol or estradiol in TEGDMo buffer (40 mM Tris⋅HCl, pH 7.4/1 mM EDTA/10% glycerol/10 mM DTT/10 mM Na₂MoO₄, pH 7.4) overnight at 4°C in a final volume of 100 μl (12). Bound and unbound ¹²⁵I-estradiol were separated with dextran-coated charcoal (11) and measured in a γ counter. Nonspecific binding was determined in the presence of a 500-fold excess unlabeled estradiol or DES and was subtracted from total binding.

Plasmids.

The plasmid tk109-luc was constructed by inserting the −109 to +52 fragment of the herpes simplex virus thymidine kinase (tk) gene into pA₃luc (13, 14). Plasmids ERE-tk109-luc and ERE2-tk109-luc were constructed by cloning single or double copies of the Xenopus vitellogenin A₂ estrogen response element (ERE) (15) into the HindIII site of tk109-luc. ERE-tk81-luc contains a single ERE upstream of the minimal thymidine kinase promoter (−81 to +52) in pA₃luc. CMV-luc (16) and FOS-luc (17), which contain the cytomegalovirus and c-fos promoters, respectively, were used as estrogen-nonresponsive controls. The wild-type human estrogen receptor expression vector pSG5-HEGO was provided by Pierre Chambon (Université Louis Pasteur, Strasbourg, France) (18), and the pSG5 control plasmid was purchased from Stratagene.

Transfection and Luciferase Assays.

Cells were grown in estrogen-depleted media and transfected by using liposomes of dioleyl phosphatidylethanolamine and dimethyldioctadecylammonium bromide (Sigma) (19, 20). MCF-7 and MDA-MB-231 cells were transfected in 6-well plates by using 7.5 μg/well and 3 μg/well of the two lipid components, respectively, and 1 μg/well of reporter gene. BG-1 cells were transfected in 12-well plates by using 4.5 μg/well and 7.5 μg/well of the lipids and 2.5 μg/well of reporter gene. Cells were incubated with liposome-DNA complexes in serum-free, estrogen-depleted media for 6–7 h and then transferred to treatment media that contained estradiol, resveratrol, or estrogen antagonists added as stock solutions in absolute ethanol. Ethanol was added to control media to produce the same final solvent concentration (typically 0.1%) in all wells. Resveratrol, 17β-estradiol, DES, and tamoxifen were purchased from Sigma. ICI 182780 was provided by Alan Wakeling (ICI Pharmaceuticals, Macclesfield, England) and Craig Jordan (Northwestern University Medical School, Chicago). Luciferase activity (21) was determined approximately 22 h after transfection, by using an AutoLumat LB953 luminometer (EG & G, Salem, MA).

Reverse Transcription (RT)-PCR Assays for Progesterone Receptor mRNA Expression.

MCF-7 cells were grown in estrogen-depleted medium for 5 days, then treated for 24 h with ethanol (control), estradiol (0.01 or 1 nM), or resveratrol (3, 10, or 30 μM). RNA was isolated by using the RNeasy Mini Kit (Qiagen, Chatsworth, CA). Total RNA (3 μg) was subjected to RT by incubation at room temperature for 10 min followed by incubation at 42°C for 15 min by using conditions described previously (22). PCR was performed by using specific primers for the progesterone receptor gene (23) and for an internal control, glyceraldehyde-3-phosphate dehydrogenase (22). PCR reactions (50 μl) included 1 μl of the RT reaction product and 50 pmol of sense and antisense primers for the receptor and dehydrogenase genes. After 25 cycles, a 20-μl aliquot of each reaction was subjected to polyacrylamide (6%) gel electrophoresis and quantitated as described (22).

Estrogen-Dependent Cell Proliferation.

T47D cells were estrogen-depleted for 5 days and seeded into 96-well plates at 5000 cells/well. Treatment media (100 μl/well) were added on the following day and replaced at 48-h intervals until the end of the experiment. Cell density was measured via the tetrazolium reduction assay (Promega) (24) after 0, 2, 4, 6, and 8 days of culture. The absorbance (490 nM) of the formazan product was measured directly in the 96-well plates with an EL 312e microplate reader (Biotek Instruments, Luton, U.K.).

Resveratrol Binds to the Human Estrogen Receptor.

The structure of resveratrol is compared with DES and estradiol in Fig. 1A. The similarities are particularly striking for resveratrol and DES, which share the stilbene backbone. The phenolic A ring that is characteristic of steroidal estrogens is present in all three compounds. These structural similarities prompted us to assess whether resveratrol might interact with the estrogen receptor. Competition binding studies were performed by using extracts of MCF-7 cells, which contain large amounts of estrogen receptor. As shown in Fig. 1B, resveratrol inhibited the binding of ¹²⁵I-estradiol. The degree of inhibition depended on the concentrations of both resveratrol and the labeled ligand. At 0.1 nM ¹²⁵I-estradiol, the IC₅₀ for resveratrol was ≈10 μM, indicating that it is a relatively weak ligand for the receptor. Nevertheless, this concentration is similar to those at which resveratrol exerts other biological actions (8). At higher concentrations, inhibition by resveratrol appeared to plateau. This may be because of the presence of multiple estradiol- or resveratrol-binding proteins in the MCF-7 extracts.

Resveratrol Functions as an Agonist for Estrogen Receptor-Mediated Transcription.

The ability of resveratrol to bind to the estrogen receptor raised the possibility that it might function as an agonist or antagonist. In the presence of an agonist, the estrogen receptor initiates transcriptional activation by binding to specific EREs in the promoters of target genes (15). The actions of resveratrol were tested initially by using ERE-tk109-luc, a reporter gene that contains a single copy of an ERE upstream of the thymidine kinase promoter (Fig. 2A). In MCF-7 cells, resveratrol produced dose-dependent transcriptional activation with half-maximal induction at 5–10 μM, corresponding to the concentration required for inhibition of estradiol binding. Although considerably less potent than estradiol, resveratrol produced a maximal level of induction that was 2- to 3-fold greater than that achieved by using maximal doses of estradiol (0.1 nM). This superagonism by resveratrol was highly reproducible and was not observed with DES, which produced the same level of activation as estradiol at saturating (1 nM) concentrations (Fig. 2B). The superagonism by resveratrol also was observed with the ERE-tk81-luc reporter gene.

Transfected MCF-7 cells were treated with combinations of resveratrol and estradiol to determine whether their actions were additive, synergistic, or antagonistic. Suboptimal doses of estradiol and resveratrol were additive (Fig. 2C), but maximal activation by resveratrol was not increased further by estradiol. These results are consistent with both compounds activating the same receptor.

Several additional types of control experiments were performed to confirm that the effects of resveratrol are estrogen receptor-mediated (Fig. 3). Deletion of the ERE from ERE-tk109-luc eliminated transcriptional activation by resveratrol (and estradiol) at each dose tested (Fig. 3A). This result, in combination with experiments using ERE-tk81-luc, confirms that the response of the reporter gene to resveratrol is not caused by estrogen receptor-independent activation of the basal promoter. Resveratrol did not activate other control reporter plasmids such as CMV-luc and FOS-luc, which are not estrogen-responsive (data not shown). Two estrogen antagonists, tamoxifen and ICI 182780, inhibited reporter gene activation by resveratrol (Fig. 3B). An estrogen receptor-negative cell line, MDA-MB-231, was used to confirm that estrogen receptor was required for resveratrol action (Fig. 3C). In the absence of cotransfected estrogen receptor, resveratrol was inactive whereas coexpression of the estrogen receptor with the ERE2-tk109-luc reporter gene conferred responsiveness to resveratrol. Thus, resveratrol action requires an ERE in the reporter gene and the participation of estrogen receptor, and it is inhibited by specific estrogen antagonists.

Cell-Specific Effects of Resveratrol.

The effects of resveratrol on estrogen receptor-mediated transcriptional activation of ERE2-tk109-luc were compared in human MCF-7 breast cancer cells and human BG-1 ovarian carcinoma cells, each of which contain endogenous estrogen receptors (Fig. 4). As with the single-ERE reporters, resveratrol was superagonistic in MCF-7 cells (56-fold induction vs. 24-fold by estradiol). In contrast, in BG-1 cells, transcriptional activation by resveratrol (8-fold) was somewhat less than that by estradiol (12-fold).

Resveratrol Stimulates Expression of Endogenous Estrogen-Regulated Genes.

The effects of resveratrol and estradiol on expression of progesterone receptor mRNA were analyzed in MCF-7 cells to determine whether resveratrol stimulates expression of natural estrogen-regulated genes as well as transfected reporter genes. As shown in Fig. 5A, resveratrol (10–30 μM) was as effective as a maximal (1 nM) dose of estradiol in activating progesterone receptor gene expression. Resveratrol and estradiol also induced expression of the pS2 gene in MCF-7 cells, and both of these estrogen-responsive genes were stimulated in T47D cells (data not shown).

Resveratrol Stimulates the Proliferation of Estrogen-Dependent Breast Cancer Cells.

T47D cells, an estrogen-dependent breast cancer cell line, were used to compare the effects of resveratrol and estradiol on cell growth. As shown in Fig. 5B, resveratrol (10 μM) induced T47D proliferation to the same extent as estradiol (0.1 nM). Cells grew equally well in 0.1 or 1 nM estradiol (data not shown), indicating that this rate of growth represents maximal estrogenic stimulation. The effect of resveratrol, like that of estradiol, was blocked by the estrogen antagonist, ICI 182780.