Stimulatory Cross-talk between NFAT3 and Estrogen Receptor in Breast Cancer Cells*

Estrogen receptors (ERα and ERβ) are ligand-regulated transcription factors that play critical roles in the development and progression of breast cancer by regulating target genes involved in cellular proliferation. The transcriptional activity of ERα and ERβ is known to be modulated by cofactor proteins. We used a yeast two-hybrid system and identified NFAT3 as a novel ERβ-binding protein. NFAT3 interacted with ERα and ERβ both in vitro and in mammalian cells in a ligand-independent fashion. NFAT3 bound specifically to the ERβ region containing the activation function-1 domain, a ligand-independent transactivation domain. Overexpression of NFAT3 enhanced both ERα and ERβ transcriptional activities in a ligand-independent manner and up-regulated downstream estrogen-responsive genes including pS2 and cathepsin D. Reduction of endogenous NFAT3 with NFAT3 small interfering RNA or overexpression of NFAT3 deletion mutants that lack the ER-binding sites reduced the NFAT3 coactivation of ERα and ERβ. NFAT3 increased binding of ERα to the estrogen-responsive element and was recruited to endogenous estrogen-responsive promoters. NFAT3 was expressed differentially in many breast cancer cell lines and overexpressed in a subset of breast cancer patients. Knockdown of endogenous NFAT3 reduced the growth of human breast cancer ZR75-1 cells in a ligand-independent manner. Taken together, these results suggest that NFAT3 may play important roles in ER signaling and represent a novel target for breast cancer therapy.

Estrogen receptors (ER␣ and ER␤) 3 are members of the steroid hormone superfamily of nuclear receptors that act as ligand-activated transcription factors (1)(2)(3). Both of the two receptors regulate gene transcription either by binding directly to estrogen-responsive elements (ERE) located within the promoter regions of target genes or interacting with other transcription factors such as AP1 and SP1 (4,5). ER␣ and ER␤ share structural similarity characterized by several functional domains. Two distinct activation function (AF) domains contribute to the transcriptional activity of the two receptors. The first activation function AF-1, a ligand-independent transactivation domain, is located at the N terminus, whereas the second, ligand-dependent activation function, AF-2, is located at the C terminus, overlapping the ligandbinding domain. The AF-1 activity of ER␤ is weak, compared with that of ER␣, whereas their AF-2 activities are similar (6). In most cases, the AF-1 and AF-2 domains interact functionally to enhance transcription in a cooperative manner. The DNA-binding domain (DBD) of the two receptors is centrally located. ER␣ and ER␤ possess similar binding affinities for estrogen and their cognate DNA binding site, which is probably caused by the high degree of sequence homology they share in their ligand and DNA binding domains (7). The ligand binding domain shows 58% homology between ER␣ and ER␤. The DNA binding domain is identical between the two receptors except for three amino acids. However, the N terminus containing the AF-1 domain of ER␤ has only 28% homology with that of ER␣.
ERs mediate the effects of estrogen on the development and progression of breast cancer (8 -10). ER␣ has served as an important diagnostic and therapeutic target for prevention and treatment. Activation of ERs is responsible for many biological processes, including cell proliferation, differentiation, motility, and apoptosis (11)(12)(13). ERs exert these functions by regulating genes and signaling pathways involved in cell fate. Regulation of gene expression by the ERs requires the coordinate activity of ligand binding, phosphorylation, and cofactor interactions, with particular combinations probably resulting in the tissue-specific responses elicited by the receptors (14 -16). Receptor phosphorylation can be induced by mitogen-activated protein kinase pathways (17)(18)(19)(20). A growing list of cofactors that regulate ER␣ includes coactivators, such as members of the SRC-1 family (21), p300/CREB-binding protein (22), p68 (23), PBP (24), PRMT2 (25), and ARNT (26), and corepressors, such as N-CoR (27), SMRT (28), MTA1 (29), and Smad4 (30). Most of the identified cofactors are coactivators. Since ER␤ was discovered more recently, very few cofactors for ER␤ have been reported. There have been some cofactors shared between ER␣ and ER␤ (20,31). The observation that certain coactivators such as SRC-3/AIB1/RAC3/ACTR/p/ CIP (32), cyclin D1 (33), and PBP (24) are overexpressed in some breast cancers underscores the importance of nuclear receptor coactivators in transcriptional activation and also points to their possible role in neoplastic transformation. However, the intracellular signaling pathways modulating ER transcriptional activity are not fully elucidated.
In this study, we have identified and characterized a novel ER-interacting protein, NFAT3. NFAT3 is a member of the family of NFAT transcription factors (34 -36). We show that NFAT3 interacts with ER␣ and ER␤ in vitro and in vivo and enhances ER␣-and ER␤-mediated transcriptional activity in a ligand-independent manner. NFAT3 not only is incorporated in the ER␣-ERE complex but also increases binding of ER␣ to the ERE sequence. NFAT3 significantly increases the transcription of various downstream genes of estrogen signaling. We further demonstrate that NFAT3 is frequently up-regulated in a subset of human breast tumors, and knockdown of endogenous NFAT3 with small interfering RNA (siRNA) significantly inhibits breast cancer cell growth.

EXPERIMENTAL PROCEDURES
Plasmids-pERE-LUC (estrogen-responsive element-containing luciferase reporter), pcDNA3-ER␣, and pCMX-SRC-1 have been described previously (37). To construct pcDNA3-ER␤, full-length human ER␤ cDNA was obtained by standard PCR amplification from a mammary two-hybrid cDNA library (Clontech). The amplified ER␤ cDNA was cloned into pcDNA3. For the yeast two-hybrid assay, the bait plasmid pAS2-ER␤-(1-167) was generated by inserting a PCR-amplified cDNA fragment containing the AF-1 region of ER␤ into pAS2-1. pAS2-ER␤-(131-324) and pAS2-ER␤-(286 -530) were constructed by inserting into the pAS2-1 vector the ER␤ cDNA fragments containing the DNA-binding domain and AF-2 domain of ER␤, respectively. pEGFP-NFAT3 was a generous gift from Dr. Toren Finkel (National Institutes of Health, Bethesda, MD) (38). The FLAG-tagged NFAT3 expression plasmid was cloned into a pcDNA3 vector linked with FLAG at the amino terminus by PCR using pGFP-NFAT3 as the template. Deletion mutants of NFAT3 were constructed by inserting PCR-generated fragments from the corresponding NFAT3 cDNAs into the pcDNA3-FLAG vector. The lac-tagged ER␣ and ER␤ expression plasmids were constructed by inserting corresponding PCR-generated fragments into a pRC-CMV (Invitrogen) vector linked with amino acids 1-355 of Escherichia coli lac repressor at the amino terminus. Plasmids encoding GST fusion proteins were prepared by amplification of each sequence by standard PCR methods, and the resulting fragments were inserted in frame into pGEX-KG (Amersham Biosciences). All plasmids were verified by restriction enzyme analysis and DNA sequencing.
Yeast Two-hybrid Screening-The Matchmaker two-hybrid system kit (Clontech) was used for detecting specific proteins interacting with the ER␤ bait protein as described by manufacturer. Briefly, the bait plasmid pAS2-ER␤-(1-167) and a human mammary cDNA prey library (Clontech) were sequentially transformed into the Saccharomyces cerevisiae strain CG1945. Transformants were plated on synthetic medium lacking tryptophan, leucine, and histidine but containing 1 mM 3-aminotriazole. Approximately 0.6 million transformants were screened. The candidate clones were rescued from the yeast cells and reintroduced back to the same yeast strain to verify the interaction between the candidates and the ER␤ bait. The specificity of the interaction was determined by comparing the interactions between the candidates and various bait constructs. The unrelated prey plasmid pACT2-lamin and the empty vector pACT2 were examined as negative controls.
GST Pull-down Assay-The GST alone and GST fusion proteins were expressed in bacteria and bound to glutathione-Sepharose beads according to the manufacturer's instructions (Amersham Pharmacia). The expression plasmid for the ER␣, ER␤, or NFAT3 was used for in vitro transcription and translation in the TNT system (Promega). The 35 S-labeled ER␣, ER␤, or NFAT3 was incubated with ϳ1 g of GST fusion protein bound to glutathione-Sepharose beads in 500 l of binding buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 0.3 mM dithiothreitol, 0.1% Nonidet P-40, and protease inhibitor tablets from Roche Applied Science) at 4°C. The beads were precipitated, washed four times with binding buffer, eluted by boiling in SDS sample buffer, and analyzed by SDS-PAGE. After electrophoresis, radiolabeled proteins were visualized by autoradiography.
Coimmunoprecipitation-For transfection-based coimmunoprecipitation assays, 293T cells were transfected with the indicated plasmids using Lipofectamine 2000 (Invitrogen), lysed in 500 l of lysis buffer (50 mM Tris at pH 8.0, 500 mM NaCl, 0.5% Nonidet P-40, 1 mM dithiothreitol, and protease inhibitor tablets from Roche Applied Science), and immunoprecipitated with anti-FLAG-agarose beads (Sigma) overnight at 4°C. The beads were washed four times with the lysis buffer and eluted in SDS sample buffer. The eluted proteins were separated by SDS-PAGE, followed by Western blotting with anti-Lac (Stratagene) or anti-FLAG (Sigma) antibody according to the standard procedures.
For detecting interaction of endogenous ER␣ with NFAT3, human breast cancer ZR75-1 cells were lysed in 500 l of lysis buffer and immunoprecipitated with anti-ER␣ or control serum (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). After extensive washing with the lysis buffer, the immunoprecipitates were resolved by SDS-PAGE, followed by Western blot analysis using anti-NFAT3 (Santa Cruz Biotechnology).
Luciferase Assay-ZR75-1, MCF-7, and MDA-MB-435 cells were maintained in DMEM (Invitrogen) supplemented with 10% fetal bovine serum. For transfection, cells were seeded in 12-well plates containing phenol red-free DMEM (Invitrogen) supplemented with 5% charcoalstripped fetal bovine serum (Hyclone). The cells were transfected using Lipofectamine 2000 (Invitrogen) with 0.2 g of ERE-LUC, 50 ng of ER␣ or ER␤ expression vector, 250 ng to 1.0 g of the expression vector for NFAT3, and 0.1 g of ␤-galactosidase reporter as an internal control. After treatment with 10 nM 17␤-estradiol (E2), 100 nM 4-hydroxytamoxifen (4-OHT), or 100 nM ICI 182,780 for 24 h, the cells were harvested. Cell extracts were analyzed for luciferase and ␤-galactosidase activities as described previously (39). All experiments were repeated at least three times with similar results.
siRNA Experiments-To construct NFAT3 siRNA expression vector, a DNA fragment containing an inverted repeat of the target sequence GTGAGTGAGATCATTGGCC, corresponding to the coding region 2641-2659 relative to the first nucleotide of the start codon, spaced by the 9-nt sequence TTCAAGAGA and a poly(T) stretch as a terminator for RNA polymerase III, was synthesized and cloned under control of the U6 promoter into the BamHI/HindIII sites of pSilencer2.1-U6neo (Ambion). Plasmid pSilencer2.1-U6neo negative control (Ambion) was used as a control vector. This control vector produces universal scramble siRNA that has no significant homology to mouse, rat, or human gene sequences. Transient or stable transfection of the vector-based siRNA into ZR75-1 cells was performed using Lipofectamine 2000 according to the manufacturer's instructions (Invitrogen). For stable transfection, transfected cells were selected in 1200 g/ml G418 (Invitrogen) and screened by Western blot using anti-NFAT3 antibody.
Gel Shift Assay-The gel shift assay was performed as described previously (37). Briefly, in vitro translated protein or the same amount of unprogrammed lysate (Promega) was combined with the 32 P-labeled ERE (5Ј-AGCTCTTTGATCAGGTCACTGTGACCTGACTTT-3Ј) or mutant ERE (EREM; 5Ј-AGCTCTTTGATCAGTACACTGTGAC-CTGACTTT-3Ј) double-stranded oligonucleotide in binding buffer (10 mM Tris, pH 7.5, 50 mM NaCl, 1 mM MgCl 2 , 4% glycerol, 0.5 mM EDTA, 0.5 mM dithiothreitol, and 0.1 g/l poly(dT-dC)) in the presence or absence of E2 (1 M). The binding reaction was incubated at room temperature for 20 min in a total volume of 20 l. Increasing amounts of in vitro translated NFAT3 or the same amount of unprogrammed lysate (Promega) were added to the binding reaction. For antibody supershift analyses, the reactions were incubated with 1 l of anti-ER (Santa Cruz Biotechnology) or anti-NFAT3 (Santa Cruz Biotechnology) antibody for 15 min at room temperature. Protein-DNA complexes were resolved on a 5% native polyacrylamide gel and visualized by autoradiography.
Immunohistochemistry-Formalin-fixed paraffin-embedded sections of 20 paired cancerous and noncancerous breasts (obtained from 304th Hospital, Beijing) were used to determine NFAT3 protein expression. The sections were deparaffinized, rehydrated, and pretreated with 3% H 2 O 2 for 20 min to quench endogenous peroxidase activity. The antibody-binding epitopes of the antigens were retrieved by microwave treatment, and the sections were then preincubated with 10% goat serum to block nonspecific binding. Rabbit anti-human NFAT3 polycolonal antibody (Santa Cruz Biotechnology) diluted 1:200 was used as the primary antibody, and the specimens were incubated with it for 1 h at room temperature, followed by the addition of biotinylated anti-rabbit secondary antibody and streptavidin-horseradish peroxidase (Zymed Laboratories Inc.). 3,3Ј-Diaminobenzidine was used as a chromogen, and hematoxylin was used for counterstaining. For negative controls, normal rabbit IgG (Santa Cruz Biotechnology) or phosphatebuffered saline was substituted for the primary antibody. In addition, preabsorbed antibody with an excess amount of GST-NFAT3 fusion protein or the blocking peptide GSSSLSSWSFFSDASDEAALYA abolished the staining. NFAT3 expression was classified into two categories, depending on the percentage of cells stained and/or the intensity of staining: Ϫ, 0 -10% positive tumor cells; ϩ, Ͼ10% positive tumor cells.
Cell Growth Assay-Stable transfection of ZR75-1 cells with expression vector for NFAT3 siRNA or control siRNA was performed using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Transfected cells were selected in 1200 g/ml G418 (Invitrogen) and screened by Western blot with anti-NFAT3 antibody. G418-resistant control and NFAT3 siRNA-transfected pooled or single cell clones were selected in 3 weeks and amplified for cell growth analysis. Briefly, cells were seeded in a 24-well plate. After overnight incubation, the medium was replaced by phenol red-free DMEM medium supplemented with 5% charcoal-stripped fetal bovine serum and incubated for an additional 24 h. Cells were incubated in the presence or absence of 10 nM E2 for 5 days. The medium was changed at day 3. Cell growth was analyzed by crystal violet assay as described previously (42). Briefly, cells were fixed by 1% glutaraldehyde for 15 min and stained with 0.5% crystal violet for 15 min at room temperature. Plates were washed with distilled water several times and air-dried. The dye was eluted by Sorenson's solution for 30 min at room temperature with constant shaking. A microplate reader was used to read aliquots of eluant at 590 nm.

RESULTS
Identification of NFAT3 as an ER␤-interacting Protein-To identify proteins that could be involved in regulation of ER signaling, we screened a human mammary cDNA library using amino acids 1-167 containing the ER␤ AF1 domain as bait in the yeast two-hybrid system. Several positive clones were identified to interact with the ER␤, one of which carries an in-frame fusion of the carboxyl-terminal region of NFAT3 (residues 613-902) to the GAL4 transactivation domain. The specificity of this interaction was confirmed in a direct two-hybrid binding assay (Fig. 1A). Transformation of yeast cells with NFAT3 in pACT2 vector alone or together with the GAL4 DNA-binding domain (DBD) in pAS2-1 vector or together with an unrelated protein, lamin C, fused to the GAL4 DBD instead of ER␤-(1-167), did not activate the his and lacZ reporter genes.
Interaction of NFAT3 with ER␣ and ER␤ in Vitro-To confirm the interaction between NFAT3 and ER␤, GST pull-down experiments were performed in which in vitro translated [ 35 S]methionine-labeled NFAT3 was incubated with full-length GST-ER␤ in the presence or absence of E2. The binding of NFAT3 to GST-ER␤, but not to GST, was observed both in the absence and in the presence of E2 ( Fig. 2A). E2 did not significantly increase the interaction of NFAT3 with GST-ER␤. Like E2, antiestrogens, 4-hydroxytamoxifen (4-OHT), and ICI 182,780 did not affect the ability of NFAT3 to interact with ER␤. As a control for the

Physical and Functional Interaction between NFAT3 and ER
effectiveness of E2 and antiestrogens in the GST pull-down assay, however, E2 enhanced the binding of SRC-1, an ER␣ coactivator, to ER␣ as previously reported (1). NFAT3 also showed interaction with ER␣ ( Fig. 2B).
Mapping of the ER␤ and NFAT3 Interaction Domains-To delineate the domains in the ER␤ that mediate the protein-protein interaction with NFAT3 in vitro, GST pull-down experiments were performed. Consistent with the yeast two-hybrid results, the GST-ER␤-(1-148) containing the AF1 domain bound specifically to in vitro translated 35 S-labeled NFAT3, but the GST-ER␤-(145-315) containing the DBD and the GST-ER␤-(248 -530) containing the AF2 did not (Fig. 2C). Since the AF1 domain of ER␤ has only about 30% homology with that of ER␣, the GST-ER␣-(1-180) containing the AF1 domain was also used in GST pull-down experiments. Like GST-ER␤-(1-148), containing the AF1 domain, GST-ER␣-(1-180) interacted with in vitro translated NFAT3 (Fig. 2D).
To determine which region of the NFAT3 protein binds to ER␤, a series of mutant GST-NFAT3 fusion proteins were used in GST pulldown experiments. As shown in Fig. 2E, ER␤ bound full-length NFAT3-(1-902) and the three NFAT3 fragments (NFAT3-(1-261), containing the transactivation domain; NFAT3-(261-450), containing the regulatory domain; and NFAT3-(613-902), containing the transactivation domain), whereas NFAT3-(451-612), containing the DNA-binding domain, failed to interact with ER␤. These results suggest that the N-or C-terminal region of NFAT3 is sufficient for the interaction with ER␤.
Interaction of NFAT3 with ER␣ and ER␤ in Vivo-To determine whether NFAT3 interacts with ER␣ and ER␤ in vivo, 293T cells were transfected with FLAG-tagged NFAT3 and lac-ER␣ or lac-ER␤ and cultured both in the presence (Fig. 2, F and G) and absence of 10 nM E2 (data not shown). FLAG-NFAT3 was immunoprecipitated from cell lysates by an anti-FLAG antibody and analyzed for ER␣ or ER␤ binding by Western blotting analysis. The results showed that both ER␣ and ER␤ could be coimmunoprecipitated in a ligand-independent manner in the presence, but not in the absence, of FLAG-NFAT3 (Fig. 2, F and  G; data not shown). The in vivo interaction of NFAT3 with ER␣ and ER␤ was unlikely to be mediated by nucleic acids, since it was not affected by the treatment of ethidium bromide that disrupts DNA-protein interaction (data not shown).
To ascertain the NFAT3-ER␣ interaction in a more physiological context, the endogenous ER␣ protein from human breast cancer ZR75-1 cells was immunoprecipitated with an anti-ER␣ antibody. Subsequent immunoblotting with an anti-NFAT3 antibody indicated that the endogenous NFAT3 was coprecipitated with ER␣ (Fig. 2H). Different ligands, E2, the ER␣-selective ligand propylpyrazoletriol, and the ER␤-selective ligand diarylpropionitrile, did not affect the ability of NFAT3 to interact with ER␣. In the negative control experiment, normal rabbit serum or an irrelevant antibody (anti-GST antibody) did not immunoprecipitate NFAT3 ( Fig. 2H; data not shown).
Ectopic Expression of NFAT3 Increases the Transcriptional Activity of ER␣ and ER␤-Having firmly established that NFAT3 is an ER␣-and ER␤-binding protein, we tested the effect of overexpression of NFAT3 on the transcriptional activity of ER␣ and ER␤. ZR75-1 cells were cotransfected with the ERE-containing reporter ERE-LUC, ER␣, or ER␤ and NFAT3. As shown in Fig. 3A, in the presence of E2, NFAT3 enhanced the transcriptional activity of ER␣ and ER␤ 4.0-and 3.9-fold, respectively, whereas in the absence of E2, NFAT3 enhanced ER␣ and ER␤ transcriptional activity 1.9-and 2.1-fold, respectively, suggesting that the activation of ER␣ and ER␤ by NFAT3 is ligand-independent. The effect of NFAT3 on the transactivation of ER␣ and ER␤ is not restricted to a single cell type, since NFAT3 also enhanced the transactivation of ER␣ and ER␤ in human breast cancer MCF7 cells (Fig. 3B). NFAT3 increased ER␤ transcriptional activation in a dose-dependent manner. The protein levels of ER␣ and ER␤ in MCF7 and ZR75-1 cells containing transiently expressed NFAT3 were similar to those observed in cells without cotransfected NFAT3 (data not shown). Therefore, the enhancement of the transactivation of ER␣ and ER␤ by NFAT3 was not due to a modulation of the protein levels of ER␣ and ER␤.
To investigate whether ER␣ itself is required for the effect of NFAT3 on the ERE-LUC reporter gene transcription, human breast cancer MDA-MB-435 cells, which lack ER␣, were cotransfected with the ERE-LUC reporter and NFAT3, together with or without ER␣. Cotransfection of NFAT3 and the ERE-LUC reporter in MDA-MB-435 cells did not increase the ERE-LUC reporter transcription, whereas cotransfection of these genes with human ER␣ expression vector led to activation of the ERE-LUC (Fig. 3C). Therefore, NFAT3 acts through ER␣ to increase ERE-LUC reporter transcription.
To examine the effects of antiestrogens on ER␣ transactivation by NFAT3, MCF7 cells were cotransfected with the ERE-LUC reporter, ER␣, and NFAT3 and subsequently treated with antiestrogens, 4-OHT and ICI 182,780 (Fig. 3D). Both ICI 182,780 and 4-OHT completely blocked the effects of NFAT3 on ER␣ transcriptional activity in the presence or absence of E2.
Reduction of Endogenous NFAT3 Decreases the Transcriptional Activity of ER␣ and ER␤-To investigate the role of endogenous NFAT3 in ER␣-and ER␤-mediated transcriptional activation, ZR75-1 cells, which expressed higher levels of NFAT3 (Fig. 7A), were transfected with NFAT3 siRNA. As shown in Fig. 4A, the NFAT3 siRNA  DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 effectively inhibited the expression of NFAT3 protein 48 h after transfection, whereas universal scramble siRNA (control) had no effect. As another control, the NFAT3 siRNA did not reduce the expression of ER␣ protein. Suppression of the normal expression of NFAT3 in ZR75-1 cells by the specific NFAT3 siRNA significantly decreased the transcriptional activity of ER␣ and ER␤ (Fig. 4B). These results further suggest that NFAT3 can enhance the transcriptional activity of ER␣ and ER␤.

Interaction of NFAT3 with ER␣ and ER␤ Is Required for the Maximal Enhancement of Transactivation Function of ER␣ and ER␤-To exam-
ine the possibility that the interaction of NFAT3 with ER␣ and ER␤ is required for the enhancement of the transcriptional activation of ER␣ and ER␤, two NFAT3 mutants, NFAT3-(1-612) and NFAT3-(451-612), were made. In the NFAT3-(1-612) mutant, one of the interaction regions from amino acids 613-902 of NFAT3 was deleted, whereas in the NFAT3-(451-612) mutant, all of the interaction regions were deleted. ZR75-1 cells were co-transfected with the ERE-LUC reporter, ER␣ or ER␤, and FLAG-tagged NFAT3, NFAT3-(1-612), or NFAT3-(451-612). As shown in Fig. 4C, the mutation lacking one of the ER binding sites reduced the coactivation of ER␣ and ER␤ by NFAT3, whereas the mutation lacking all of the NFAT3 binding sites further reduced the enhancement of the transcriptional activity of ER␣ and ER␤ by NFAT3. Notably, FLAG-tagged NFAT3, NFAT3-(1-612), and NFAT3-(451-612) were expressed at comparable levels (Fig. 4D). Taken together, these findings suggest that the NFAT3 action of ER␣ and ER␤ by their binding contributes to the maximal enhancement of the transcriptional activity of ER␣ and ER␤.
NFAT3 Enhances the Expression of Endogenous ER␣ Target Genes-To corroborate the results of the luciferase reporter assay, the effect of NFAT3 on the expression of endogenous ER␣ target genes was examined. The E2-deprived MCF-7 cells stably expressing either the empty vector or NFAT3 were treated with 10 nM E2 for 20 h. Semiquantitative RT-PCR was performed using the primers specific for ER␣-responsive genes, pS2, cathepsin D, and HSP27. As expected, stable transfection of NFAT3 increased expression of NFAT3 mRNA, and E2-dependent increase in expression of the ER␣-responsive genes was observed in the vector-transfected cells (Fig. 5A). Importantly, the transfection of NFAT3 further enhanced the expression of pS2 and cathepsin D, but not that of Hsp27, both in the absence and in the presence of E2 (Fig.  5A). Consistent with the RT-PCR results, the E2 stimulation of pS2 and cathepsin D and their enhancement by NFAT3 was also observed at the protein level by Western blot (Fig. 5B). These data suggest that NFAT3 selectively increases the expression of endogenous ER␣-responsive genes. NFAT3 Increases Binding of ER␣ to ERE Sequence-Since NFAT3 is a transcription factor, we first tested if NFAT3 binds to the consensus ERE, using a gel shift assay. As expected, the 32 P-labeled ERE, but not EREM, bound to in vitro-translated ER␣ in the absence or presence of E2 ( Fig. 6A and data not shown). The binding was specifically inhibited by a 100-fold molar excess of a cold ERE oligonucleotide. The addition of human anti-ER␣ antibody to the reaction caused a supershift. However, NFAT3 did not bind to the ERE (Fig. 6A).
To investigate whether NFAT3 enhances the binding of ER␣ to ERE, a constant in vitro translated ER␣ and 32 P-labeled ERE were incubated with increasing amounts of in vitro translated NFAT3 or purified GST-NFAT3 fusion protein. As shown in Fig. 6B, NFAT3 increased the ER␣ binding to ERE sequence in a dose-dependent manner. Antibody supershift experiments showed that NFAT3 was present in the ER␣-ERE complexes (Fig. 6B). Thus, these data suggest that the NFAT3 action on transcription may be due to enhanced ER␣ binding to ERE.

. Knockdown of endogenous NFAT3 or the deletion mutants of NFAT3 reduces ER␣ and ER␤ transcriptional activity in ZR75-1 cells. A, Western blotting with various
antibodies showing the specific knockdown effect of the NFAT3 siRNA on the endogenous NFAT3 protein level. Cells were transfected with NFAT3 siRNA or scramble siRNA (control) plasmid. Forty-eight hours after transfection, whole-cell extracts were prepared and probed with anti-NFAT3 (Santa Cruz Biotechnology), ER␣ (Santa Cruz Biotechnology), or GAPDH antibody (Biogenesis). B, luciferase reporter assays in the control and NFAT3 knockdown cells. Cells were cotransfected with 0.2 g of ERE-LUC, 50 ng of the expression plasmid for ER␣ or ER␤, and 1.0 g of NFAT3 siRNA as indicated. Cells were treated and analyzed as described in the legend to Fig. 3A. C, luciferase reporter assays with the NFAT3 deletion mutants. Cells were cotransfected with 0.2 g of ERE-LUC, 50 ng of the expression plasmid for ER␣ or ER␤, and 1.0 g of the expression vector for FLAG-tagged NFAT3, NFAT3-(1-612), or NFAT3-(451-612) as indicated. Cells were treated and analyzed as described in the legend to Fig. 3A. D, Western blotting showing expression of FLAG-tagged NFAT3, NFAT3-(1-612) and NFAT3-(451-612). Cells were transfected as in C. Cell extracts were prepared, and equivalent amounts of each extract were detected with anti-FLAG antibody (Sigma). DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 NFAT3 Is Recruited to Endogenous E2-responsive Promoters-To further assess the role of NFAT3 as a coactivator for ER-mediated transcription, we performed ChIP experiments using pS2 and cathepsin D promoters. ZR75-1 cells were grown in the absence of estrogen for at least 3 days followed by the addition of either no hormone or E2 for various times. Promoter occupancy before and after E2 treatment at the estrogen response elements within the endogenous pS2 and cathepsin D gene promoters by ER␣ and NFAT3 was then detected by ChIP using antibodies against ER␣ or NFAT3 and semiquantitative PCR with primers flanking the estrogen-responsive elements of the pS2 and cathepsin D promoters. As expected, ER␣ displayed a clear time-dependent recruitment to the pS2 and cathepsin D promoters (Fig. 6C). Importantly, NFAT3 also revealed a distinct time-dependent recruitment to the pS2 and cathepsin D promoters. The specificity of factor association within the estrogen-responsive region of the pS2 and cathepsin D promoters was confirmed by ChIP analysis using normal rabbit serum or an irrelevant antibody, anti-FLAG monoclonal antibody. These antibodies failed to immunoprecipitate pS2 and cathepsin D promoter sequences ( Fig. 6C and data not shown). Further specificity of the ChIP analysis was demonstrated by the inability to detect occupancy by ER␣ or NFAT3 of a region ϳ2 kb upstream of the pS2 or cathepsin D promoter (Fig. 6C). Thus, these results reveal the association of NFAT3 with ER␣ at endogenous estrogen-responsive promoters under physiologically relevant conditions in vivo.

Physical and Functional Interaction between NFAT3 and ER
NFAT3 Is Overexpressed in Breast Cancer Patients-To investigate the pathological role of NFAT3 in breast cancer, Western blot analysis was performed with an anti-NFAT3 antibody. Differential expression of NFAT3 protein was detected in the breast cancer cell lines tested ( To determine the expression of NFAT3 in breast cancer patients, immunohistochemistry was performed with the anti-NFAT3 antibody (Fig. 7, B and C). Specimens from 20 breast cancer cases were used. NFAT3 was undetectable in noncancerous breast tissues adjacent to tumor in 17 of 20 (85%) cases (Fig. 7B, left), whereas negative staining for NFAT3 was observed in only 8 of 20 (40%) breast cancers. Low level staining for NFAT3 was observed in 3 of 20 (15%) cases of noncancerous breast tissues adjacent to tumor. NFAT3 was detected at an intermediate or high level in 12 of 20 (60%) breast tumors. NFAT3 staining was predominantly localized to the tumor cytoplasm of epithelial cells (Fig.  7B, right). No stromal staining was observed. Specificity of staining was FIGURE 5. Exogenous expression of NFAT3 increases expression of ER target genes. MCF-7 cells stably transfected with NFAT3 or empty vector were cultured in phenol red-free DMEM and treated with control (0.1% ethanol) vehicle or 10 nM E2 for 20 h. A, the cells were harvested, and total RNA was extracted. Semiquantitative RT-PCR was performed as described under "Experimental Procedures." ␤-Actin serves as an internal control. B, the cells were harvested and lysed in radioimmune precipitation buffer, and conditioned medium was concentrated using a 3-kDa membrane. The concentrate was used for Western blot analysis of the expression of pS2 and cathepsin D, and the whole cell lysate was used for Western blot analysis of the expression of NFAT3 and GAPDH. GAPDH served as a loading control. FIGURE 6. NFAT3 forms a complex with ER␣ on DNA. A, NFAT3 did not bind to the ERE. Gel shift assay was performed using 32 P-labeled ERE probe and in vitro translated ER␣ or NFAT3 as indicated, in the presence of 1 M E2. For competition experiments, a 100-fold molar excess of unlabeled ERE was incubated with the radioactive probe. The 32 P-labeled mutant EREM probe was used as a negative control. Supershifts were performed using specific anti-ER␣ antibodies (Santa Cruz Biotechnology). The protein-DNA mixtures were fractionated on a nondenaturing acrylamide gel. The gel was dried and subjected to autoradiography. B, NFAT3 increased protein-DNA complex formation. Increasing amounts of in vitro translated NFAT3 were mixed with in vitro translated ER␣ and 32 P-labeled ERE in the presence of 1 M E2. For supershift experiments, anti-NFAT3 antibodies (Santa Cruz Biotechnology) were included in the binding reactions as indicated. C, recruitment of ER␣ and NFAT3 to estrogen-responsive promoters. ZR75-1 cells, cultured in the absence of estrogen, were treated without (time 0) and with 10 nM E2 for 30 and 60 min. Soluble chromatin was prepared and subjected to immunoprecipitation by using normal serum (negative control) or antibodies for ER␣ or NFAT3. Immunoprecipitated DNA was PCR-amplified with primers that annealed to the proximal region of pS2 and cathepsin D promoters or the region ϳ2 kb upstream of the pS2 or cathepsin D promoter.
confirmed by use of the purified GST-NFAT3 fusion protein or the blocking peptide that completely suppressed staining. Of 20 breast cancer patients, 17 were ER␣-positive, out of which 12 are NFAT3-positive. None of the ER␣-negative breast cancer patients showed NFAT3 expression. However, many ER␣-negative breast cancer cell lines, such as MDA-MB-231, MDA-MB-435, and MDA-MB-453 cells, were shown to express NFAT3 (Fig. 7A). Thus, more clinical samples are needed to determine the correlation between NFAT3 and ER␣. Taken together, these data suggest that NFAT3 is overexpressed in a subset of breast cancer patients.
Reduction of Endogenous NFAT3 Inhibits Breast Cancer Cell Growth -Since NFAT3 is overexpressed in a subset of breast cancer patients, the effect of knockdown of endogenous NFAT3 on growth of ZR75-1 cells was examined. NFAT3 siRNA-expressing single clones or pooled clones were obtained. As shown in Fig. 8A, a representative single clone with NFAT3 siRNA showed a much lower level of expression of NFAT3 than that with control siRNA. In the presence or absence of estrogen, the NFAT3 siRNA-expressing clone displayed a reduced rate of proliferation in comparison with the control siRNA clone (Fig. 8B). Similar results were obtained with pooled clones or other NFAT3 siRNA-expressing single clones. The results indicate that reduction of endogenous NFAT3 inhibits breast cancer cell growth in a ligand-independent manner.

DISCUSSION
In this study, we have identified the transcription factor NFAT3 as a new ER-interacting partner. The physical interaction has been validated by a number of in vitro and in vivo assays, including yeast two-hybrid, in vitro GST pull-down, and in vivo coimmunoprecipitation. Overexpression of NFAT3 enhances the transcriptional activity of ER␣ and ER␤ and the expression of some estrogen-responsive genes. In sharp contrast, knockdown of endogenous NFAT3 reduces the ER␣ and ER␤ transactivation, suggesting that NFAT3 plays an important role in ER signaling. Moreover, the lack of the ER binding sites in NFAT3 reduces or abolishes the ER␣ and ER␤ transactivation by NFAT3. NFAT3 is recruited to endogenous estrogen-responsive promoters. Since NFAT3 does not bind to ERE sequence, these data suggest that NFAT3 is a novel coactivator for ER.
To date, a number of coactivators for ER and other hormone receptors have been identified and characterized. The number of ER␣ coactivators is much more than that of ER␤ coactivators. There are some coactivators shared between ER␣ and ER␤ (20,31). For instance, SRC-1, the p160 steroid receptor coactivator family member, interacts both with ER␣ and ER␤ and enhances their transcriptional activity. Most of the ER coactivators identified have been shown to bind to the AF-2 region of ER. They include the three members of the p160 family (SRC-1/NcoA1 (20), SRC-2/GRIP1/TIF2 (43), and SRC-3/AIB1/RAC3/ ACTR/p/CIP (32)), p300/CREB-binding protein (22), and TIF1␣ (44). These coactivators increase ER transcriptional activity in a ligand-dependent manner. Recently, a few coactivators have been shown to interact with the AF-1 region, a ligand-independent transactivation domain, of ER. p68, an RNA helicase, interacts with ER␣ AF-1 but not with ER␣ AF-2 and stimulated ER␣-dependent reporter gene expression (23). Like p68, NFAT3 binds to the AF-1 but not the AF-2 in both the presence and absence of estrogen. The specific interaction between NFAT3 and the AF-1 may explain why NFAT3 regulates ER transcriptional activity in a ligand-independent manner. ER␣ and ER␤ differ mostly in the N-terminal AF-1 domain and to a lesser extent in the AF-2 domain. The AF-1 domains of ER␣ and ER␤ only show about 30% homology. However, the DNA-binding domains of ER␣ and ER␤ are identical except for three amino acids. The fact that NFAT3 increases both ER␣ and ER␤ transcriptional activity through interaction with the AF-1 suggests the importance of the conserved amino acid residues in the AF-1 domains of ER␣ and ER␤.
It has been reported that the NFAT family of transcription factors plays important roles in the development of several organ systems, FIGURE 7. Expression of NFAT3 protein in human breast cancer cell lines and breast cancer specimens. A, Western blot from the selected cell lines was performed using an anti-NFAT3 antibody (Santa Cruz Biotechnology). GAPDH was used as an internal control. B, representative immunohistochemical staining of NFAT3 expression in human breast cancer tissues. A case of breast carcinoma showed increased staining of cancerous cells (right) compared with negative staining of adjacent noncancerous cells (left) using anti-NFAT3 (Santa Cruz Biotechnology). Cells with red-brown staining are considered as positive. C, summary of the results from the immunohistochemistry assay.  DECEMBER 30, 2005 • VOLUME 280 • NUMBER 52 including the immune system, vasculature, nervous system, and heart (45)(46)(47). Four closely related members of the NFAT family, NFAT1/ NFATc2/NFATp, NFAT2/NFATc1/NFATc, NFAT3/NFATc4, and NFAT4/NFATc3/NFATx, have been identified and characterized, each with distinct temporally and spatially regulated expression patterns. All of the NFAT members contain a highly conserved N-terminal regulatory NFAT homology region and a C-terminal Rel homology region for DNA binding. The regions of the NFAT members located outside the DNA-binding and regulatory domains have relatively little sequence conservation. There are two transactivation domains in NFAT3, one located N-terminal to the regulatory domain and the other located C-terminal to the DNA-binding domain. Although an NFAT-like factor, named NFAT5/NFATz/TONEBP (48), was discovered, it differs in structure significantly from above four NFAT proteins. Therefore, it is a matter of dispute whether NFAT5 should be included within the NFAT family. Our study showed that two separate regions, one containing the N-terminal transactivation and regulatory domains and the other containing the C-terminal transactivation domain of NFAT3, interact with ER␤, whereas the region containing the DNA-binding domain of NFAT3 does not. The presence of two interacting regions may allow an efficient recruitment of NFAT3 to the ER transcription complex and therefore promote ER-mediated gene transcription. On the other hand, NFAT3 is a transcription factor that undergoes shuttling between the cytoplasm and nucleus. Under basal, unstimulated conditions, NFAT3 resides in the cytoplasm. Stimuli that elicit calcium mobilization cause the rapid dephosphorylation of NFAT3 by the Ca 2ϩ /calmodulin-dependent phosphatase calcineurin and its translocation to the nucleus. Since the NFAT3 protein is limited in quantity inside the nucleus under certain conditions, the presence of two binding regions may allow effective competition for ER once NFAT3 is translocated to the nucleus. How NFAT3 undergoes shuttling between the cytoplasm and nucleus in breast cancer cells remains to be elucidated.

Physical and Functional Interaction between NFAT3 and ER
Recently, we and others have identified several ligand-independent coactivators for ER, such as XBP-1 (37) and cyclin D1 (33). XBP-1 enhances ER␣ transcriptional activity possibly by regulation of large scale chromatin unfolding (49). Cyclin D1 is the first coactivator to have been demonstrated to increase ER␣ transactivation in a ligand-independent manner. Cyclin D1 enhances binding of both liganded as well as unliganded ER␣ to ERE. Like cyclin D1, NFAT3 increases the transcriptional activity of ER in a ligand-independent manner, possibly due to enhanced ER binding to ERE. Since a variety of other mechanisms, including local chromatin remodeling, enzymatic modification of histone tails, modulation of the preinitiation complex via interactions with RNA polymerase II and general transcription factors, increase of ER homodimerization, and effects on RNA processing, have been reported to be responsible for activation of ER by coactivators, we cannot exclude other possible mechanisms of NFAT3 coactivation of ER.
NFAT factors are expressed in multiple cell types and at different developmental stages, where they contribute to diverse functions (34 -36). NFAT1, NFAT2, and NFAT4 are predominantly expressed in immune cells and participate in the activation of T and B cells. NFAT3 mRNA has been shown to be primarily expressed in nonimmune tissues, including heart and brain. Consistent with the NFAT3 expression pattern, NFAT3 activities have been implicated in cardiac hypertrophy, learning and memory, and adipocyte differentiation. We identified a new role for NFAT3 in breast tissues. NFAT3 protein was expressed in many breast cancer cell lines and breast cancer patients. Importantly, NFAT3 protein is elevated in breast tumors compared with matched noncancerous breast tissues, suggesting that NFAT3 may play an important role in breast cancer development and progression. In sup-port of this, we further demonstrated that reduction of endogenous NFAT3 significantly inhibited breast cancer cell growth. Therefore, NFAT3 may be a useful target for breast cancer therapy.