Ligand-dependent interaction of estrogen receptor-alpha with members of the forkhead transcription factor family.

Estrogen acting through the estrogen receptor (ER) is able to regulate cell growth and differentiation of a variety of normal tissues and hormone-responsive tumors. Ligand-activated ER binds DNA and transactivates the promoters of estrogen target genes. In addition, ligand-activated ER can interact with other factors to alter the physiology and growth of cells. Using a yeast two-hybrid screen, we have identified an interaction between ER alpha and the proapoptotic forkhead transcription factor FKHR. The ER alpha-FKHR interaction depends on beta-estradiol and is reduced significantly in the absence of hormone or the presence of Tamoxifen. A glutathione S-transferase pull-down assay was used to confirm the interaction and localized two interaction sites, one in the forkhead domain and a second in the carboxyl terminus. The FKHR interaction was specific to ER alpha and was not detected with other ligand-activated steroid receptors. The related family members, FKHRL1 and AFX, also bound to ER alpha in the presence of beta-estradiol. FKHR augmented ER alpha transactivation through an estrogen response element. Conversely, ER alpha repressed FKHR-mediated transactivation through an insulin response sequence, and cell cycle arrest induced by FKHRL1 in MCF7 cells was abrogated by estradiol. These results suggest a novel mechanism of estrogen action that involves regulation of the proapoptotic forkhead transcription factors.

Estrogen and related steroid ligands play a critical role in the normal development and function of numerous cell types. Estrogens induce physiologic effects through an interaction with nuclear steroid receptors. Two human estrogen receptors have been identified, ER␣ 1 and ER␤ (1)(2)(3)(4). The ERs are members of the steroid-thyroid-retinoic acid superfamily of transcription factors (5). In the classic model of steroid hormone action, ␤-estradiol induces homodimerization of ER, which is able to bind specific regulatory sequences in the promoters of ER target genes called estrogen response elements (EREs) (6). It is through this classic model of steroid hormone action that ERs alter the expression of a set of target genes. Several target genes for ER␣ in hormone-responsive breast tumors have been described including progesterone receptor (PR) (7), pS2 (8), TGF-␣ (9), cathepsin D (10), HSP27 (11), and GREB1 (12). These genes are directly activated by ER␣, and the induction of gene expression depends on the ability for ER␣ to bind to the promoters of each target gene.
Increasingly, it has been reported that estrogen acting through ERs can have profound effects on cell physiology through mechanisms independent of DNA binding. One mechanism that has been proposed is through the ability of ER␣ to regulate the activity of other nuclear transcription factors by mechanisms involving direct protein-protein interactions. In many cases the interactions between ER␣ and other nuclear factors have been shown to be ligand-dependent. One example of this alternate mechanism of gene regulation is the effect of ER␣ on the expression of AP1-regulated genes (13). ER␣ and ER␤ have been shown to interact with AP1 but with differential ligand activation. In experiments in HeLa cells, ER␣ stimulated an AP1 reporter plasmid in the presence of estrogen or antiestrogens. However, ER␤ demonstrated activation of AP1mediated transcription only in the presence of antiestrogens. At the functional level, the classic model of steroid action and this alternate mechanism of ER action are indistinguishable; in both cases estradiol induces alterations in the pattern of gene expression. Through interaction with other nuclear factors, ERs are able to modulate the expression of genes normally controlled by factors that are not thought to be influenced by steroid hormones. More recently there has been an important finding that ERs can regulate the cell cycle and apoptosis through a small fraction of receptors that may be present in the cell membrane (14). In this model of regulation, the ligandactivated ER␣, ER␤, or androgen receptor (AR) was able to attenuate apoptosis through the activation of the Src/Shc/ERK pathway.
Understanding the mechanisms that regulate cell physiology in response to hormone is an important step toward developing new therapies for diseases of hormone-responsive tissues. We have identified additional proteins that interact with ligandactivated ER␣ using a yeast two-hybrid screen. One of these was a ligand-dependent interaction between ER␣ and the forkhead family transcription factor FKHR. This interaction was specific for ligand-activated ER␣; no significant interaction could be detected between FKHR and AR, glucocorticoid receptor (GR), progesterone receptor (PR), or vitamin D receptor (VDR). A ligand-specific interaction between ER␣ and the re-lated proteins FKHRL1 and AFX was also confirmed. The interaction between ER␣ and FKHR augmented transactivation through EREs but repressed the ability of FKHR to transactivate through an insulin-response sequence (IRS). Expression of FKHRL1 induced alterations in the cell cycle in MCF7 cells that were reversed with estrogen. These findings establish an important link between ER␣ and transcriptional regulation of the proapoptotic family of forkhead transcription factors and provide a novel mechanism of estrogen action in hormone-responsive cells.

EXPERIMENTAL PROCEDURES
Cell Lines-The cell lines COS-1 and MCF7 were obtained from ATCC and maintained as described previously (15).
Two-hybrid Screen-The full-length ER␣ cDNA was amplified using the oligos 5Ј-ERNde (5Ј-GGGGAATTCCATATGACCATGACCCTCCA-CACCAAAGCATCAGGG-3Ј) and 3Ј-ERBam (3Ј-GCCAGGGGATCCTC-AGACTGTGGCAGGGAAACCCTC-3Ј) and was cloned into the NdeI and BamHI site of pGBKT7 (CLONTECH). This vector encoding the ER-Gal4 DNA binding domain fusion was co-transfected with the human mammary gland MATCHMAKER cDNA library (CLONTECH) into AH109 yeast. The yeast were selected following the instructions of the manufacturer except the plates were treated with water (carrier), ␤-estradiol (final concentration on the plate 100 nM), or Tamoxifen (final concentration 1 M).
DNA Constructs-The human FKHR cDNA (pFB-12A2) (16) was provided by Dr. Karen Arden (The Ludwig Institute for Cancer Research, La Jolla, CA), and the human cDNA clones for FKHRL1 and the triple mutant FKHRL1-TM were provided by Dr. Michael Greenberg (Harvard University) (17). The plasmids pCMVgal50 and p5GE1b-luc were obtained from Dr. Kent Wilcox (Medical College of Wisconsin). Fragments of FKHR were amplified by PCR using the primer pair FKHR111 (5Ј-CGGTCTAGAGAATTCAATTCGTCATAATCTGTCCC-3) and FKHR110 (5Ј-GCCGGTACCTCAGCCTGACACCCAGCTATGTG) or the pair FKHR109 (5Ј-CGGTCTAGAGGTTAACGGGCGTCCCCT-GC) and FKHR110. The PCR products were cloned into pCR3.1 and subcloned into the expression vector pCMVgal50.
The expression plasmid pFKHR-MT was constructed by PCR amplification of the coding region of pFB-12A2 with the primers FKHR5Ј-Bam (5Ј-GGGGGATCCGCCACCATGGCCGAGGCGCCTCA-GGTGGTGGAG) and FKHR3Ј-Xho (5Ј-GGCCTCGAGGCCTGACACC-CAGCTATGTGTCGT) containing a BamHI site at the 5Ј end and an XhoI site at the 3Ј end to allow for in-frame fusion of the FKHR open reading frame with the Myc epitope sequence of pCMV-Tag5A (Stratagene, La Jolla, CA). The PCR product was cleaved with BamHI and XhoI and ligated into pCMV-Tag5A. For cell cycle analysis, the FKHRL1 and FKHRL1-TM cDNAs were subcloned into the green fluorescent protein (GFP) expression plasmid pCMS-EGFP.
The reporter plasmid 3XIRS-LUC contains three copies of the IRS element linked to the herpes simplex virus thymidine kinase gene promoter driving the expression of the Photinus pyralis luciferase gene and was provided by Drs. Guan and Tang (University of Michigan). The reporter pVit ERE-Luc contains three copies of the ERE from the vitellogenin A1 gene and was provided by Dr. Craig Jordan (Northwestern University).
Transactivation Assays-DNA was introduced into either MCF7 or COS-1 cells by lipid-mediated transfection using Fugene-6 (Roche Molecular Biochemicals) according to manufacturer instructions. For COS-1 cells, the cells were plated at 2.5 ϫ 10 5 cells/well in 6-well plates 24 h prior to transfection and were 80% confluent at the time of transfection. Each transfection contained 250 ng of Gal4 reporter plasmid (p5GE1b-luc), 100 ng of p␤Gal-Control vector, and various amounts of pCMVgal50/FKHR109 -110, pCMVgal50/FKHR111-110, or pCMV-gal50 as indicated. All transfections were performed in triplicate, and the cells were harvested 48 h post-transfection. Cell extracts for luciferase or ␤-galactosidase assays were prepared using a luciferase sssay system (Promega). Luciferase assays were performed with 5 l of cell extract and 100 l of luciferase assay buffer. The enzyme activity was measured for 2 s in a luminometer (Analytical Luminescence Laboratory, San Diego, CA). ␤-galactosidase activity in cell extracts was assayed using the Galacto-Light system (Applied Biosystems, Bedford, MA).
For MCF7 cell transfections, cells were seeded at 5 ϫ 10 5 /well in 6-well plates in growth medium the day before transfection. On the day of transfection the medium was replaced with fresh growth medium. DNA-lipid complexes were formed in serum-free minimum essential medium (MEM) and then placed on the cells. Generally, 1 g each of reporter plasmid and transactivator plasmid and 0.4 g of p␤gal-control plasmid were used per sample. After 18 h of incubation in the presence of the DNA-lipid complexes the transfection mixture was removed, and the cells were washed twice with phosphate-buffered saline. For hormone response assays, phenol red-free MEM containing 10% charcoal dextran-treated fetal bovine serum with or without the appropriate hormone was added as described previously (12). For assays not requiring hormone-controlled conditions the transfection mixture was replaced with MEM plus 10% fetal bovine serum. The cells were replaced at 37°C/5% CO 2 until harvest. The luciferase activity was assayed using 10 l of lysate and 100 l of luciferase assay reagent, and luminescence was measured as described above. ␤-galactosidase activity in 10 l of the same lysates was measured as described above. For each sample the luciferase relative light units were normalized with the ␤-galactosidase relative light units.
Cell Cycle Analysis-MCF7 cells were transfected with the expression plasmids pCMS-EGFP, pCMS-FKHRL1-WT, or pCMS-FKHRL1-TM by electroporation. For each sample, 4 ϫ 10 6 cells were washed with cold MEM and then resuspended at 5 ϫ 10 7 cells/ml in cold MEM. Ten micrograms of DNA was added, and the cells were electroporated in 0.4-cm cuvettes using a Bio-Rad Gene Pulser set at 260V, 960 F, and 200 ⍀. Growth medium was added to bring the volume to 12 ml, and the cells were plated in four 6-cm dishes at 37°C/5% CO 2 . After 18 h, the medium was removed, and the cells were rinsed twice with PBS. The appropriate medium was added to each plate, and the cells were replaced at 37°C for 24 h.
Samples were processed for cell cycle analysis by flow cytometry as described (18). Briefly, the cells were harvested by brief trypsinization and placed on ice. All steps were performed at 4°C. The cells were washed twice in PBS with 1% fetal bovine serum and fixed with 1% formaldehyde in PBS for 1 h. After two additional washes the cells were fixed in 70% ethanol/30% PBS overnight at 4°C. The cells were washed twice before resuspension in PBS ϩ 1% fetal bovine serum, RNase A (100 g/ml), and propidium iodide (40 g/ml). After incubation at 37°C for 1 h, 20,000 cells from each sample were analyzed on a FACStar flow cytometer for GFP and propidium iodide fluorescence. The data were analyzed using Flowjo software (Treestar Software). The GFP-positive single cell population was plotted as a histogram of propidium iodide fluorescence, and the Watson Pragmatic model (19) was used to perform cell cycle analysis.

RESULTS
Ligand-dependent Interaction of ER␣ and FKHR-The yeast two-hybrid system was used to identify proteins with liganddependent interaction with ER␣. Full-length ER␣ was cloned as a fusion protein with the DNA binding domain of Gal4. A normal human mammary epithelial cell cDNA library cloned to create a fusion protein with the Gal4 activation domain was used in yeast co-transfection. Yeast transformants were screened first on Leu/Trp medium to select for double transformants. To identify proteins that interact with ER␣ in a ligandspecific fashion, yeast colonies were selected on Leu/Trp/His/ Ade medium supplemented with ␤-estradiol, tamoxifen, or no ER␣ ligand. Approximately 300 yeast colonies were identified that demonstrated a ligand-dependent growth phenotype, and four colonies were isolated that demonstrated growth only with ␤-estradiol (data not shown). The plasmids encoding the activation domain fusion proteins were recovered from these four yeast transformants, and the inserts were sequenced. One of these inserts was found to encode FKHR from amino acid 211 through the stop codon, which had been cloned as a fusion protein in-frame with the Gal4 activation domain.
The plasmid encoding FKHR (AD/FKHR) was co-transfected into yeast with the plasmid expressing the ER␣-Gal4 DNA binding domain fusion protein (DNA-BD/ER). As seen in Fig. 1, there was ␤-estradiol-dependent growth on Leu/Trp/His/Ade medium of yeast co-transfected with the fusion plasmids containing ER␣ and FKHR. By contrast, tamoxifen failed to allow these yeast transformants to grow. As controls, a T antigen-Gal4 activation domain fusion (AD/T) was not able to interact with ER␣ as evidenced by the lack of growth on His/Ade medium, whereas DNA-BD/p53 and AD/T co-transfection generated yeast capable of growth independent of ER␣ ligand. These results indicate an estradiol-dependent interaction between ER␣ and FKHR.
A GST pull-down assay was used to confirm the ligand-dependent interaction between ER␣ and FKHR. The full-length FKHR protein was synthesized by in vitro transcription/translation from the cloned cDNA. The labeled FKHR protein was incubated with a panel of GST-fusion proteins encoding a variety of nuclear hormone receptor ligand binding domains. The ligand binding domains of ER␣, AR, GR, PR, and VDR were each tested in parallel in the presence and absence of their respective ligands. As seen in Fig. 2, only the ER␣-GST fusion protein in the presence of ␤-estradiol was able to bind FKHR. There was no significant interaction detected between FKHR and the ligand binding domains of the other nuclear hormone receptors.
Domains of FKHR Involved in ER␣ Binding-To determine the domain of FKHR involved with the interaction with ER␣, subregions of FKHR were expressed individually using in vitro transcription/translation and tested for ligand-dependent interaction with ER␣ using the GST pull-down assay (Fig. 3). The region of FKHR from amino acid 211 to 655 (fragment I) interacted with ER␣-GST, and the interaction was enhanced greatly greatly with ␤-estradiol. The region of FKHR from amino acids 211 to 466 (fragment II) and 211 to 561 (fragment VI) behaved in a similar fashion; both interacted with ER␣, and the binding was enhanced by ␤-estradiol. However, FKHR amino acids 445-655 (fragment III) and 280 -655 (fragment IV) bound to ER␣-GST only in the presence of ␤-estradiol. Neither fragment V (amino acids 280 -446) nor fragment VII (amino acids 280 -561) interacted with ER␣-GST. These results demonstrate that two regions of FKHR interact with ER␣. The carboxyl-terminal region from amino acids 561 to 655 binds tightly to ER␣, and binding completely depends on ␤-estradiol. A second region from amino acids 211 to 280, which includes part of the forkhead domain, interacts weakly with ER␣ in the absence of ligand; however, binding is enhanced significantly in the presence of ␤-estradiol.
The forkhead domain of FKHR is homologous with the two other family members, FKHRL1 and AFX. A BLASTp search of the carboxyl-terminal 95 amino acids of FKHR also demonstrated strong homology between these three proteins in this region. FKHRL1 has 39% identity and 55% homology with FKHR in the carboxyl-terminal 95 amino acids. As seen in Fig.  4, a comparison of the region of FKHR from amino acids 592 to 633 demonstrates 65% identity and 76% homology with FKHRL1 amino acids 601-643. AFX contains a region from amino acids 462 to 501 with 56% identity and 82% homology with this region of FKHR. This finding suggests that FKHRL1 and AFX may also bind to ER␣ in a ligand-dependent fashion. To test this possibility, in vitro synthesized FKHRL1 and AFX were tested for ER␣ binding using the GST pull-down assay. As seen in Fig. 5, both AFX and FKHRL1 bound to ER␣ only in the presence of ␤-estradiol. In addition, a FKHRL1 protein with mutations of the three regulatory phosphorylation sites (FKHRL1-TM) (17) also bound to ER␣ in a ligand-dependent fashion.
The conserved nature of the regions of FKHR that interact with ER␣ suggests that these are important functional domains. In addition, the highly acidic pI of the carboxyl terminus suggests that this region might function as an activation domain. To test this possibility, the region of FKHR from 211 to 655 and 534 to 655 were cloned into a eukaryotic expression vector as fusion proteins with the DNA binding domain of Gal4. The ability of these fusion proteins to transactivate transcription was assayed by co-transfection with a Gal4 luciferase reporter. As seen in Fig. 6, the region of FKHR from 211 to 655 provides trans-activating function with a nearly linear increase of induction with increased expression. The activity from the FKHR fusion protein is approximately one-third of the activity FIG. 1. Yeast two-hybrid assay reveals a ligand-dependent interaction of FKHR with ER␣. AH109 yeast was transfected with the expression plasmids encoding fusion proteins with the Gal4 DNA binding domain or the activation domain as shown. Yeast were selected on Leu/Trp plates to confirm the expression of both plasmids. Transformants were subsequently plated on Leu/Trp/His/Ade medium with ␤-estradiol (E2), Tamoxifen (Tam), or no ER␣ ligand (None) as shown. Note the growth of DNA-BD/ER and AD/FKHR only in the presence of ␤-estradiol.

FIG. 2. FKHR binds specifically to ligand-activated ER␣ and not to other nuclear hormone receptors.
Full-length FKHR synthesized using in vitro transcription/translation was incubated with a panel of GST-receptor ligand binding domain fusion proteins representing AR, ER␣ (ER), GR, PR, and VDR. Incubations were conducted in the presence (ϩ) or absence (Ϫ) of the appropriate ligands followed by SDS-polyacrylamide gel electrophoresis as described under "Experimental Procedures." The location of bound FKHR protein was revealed by autoradiography and is indicated by an arrow. of a VP16-Gal4 fusion (data not shown). The terminal 122 amino acids (534 -655) contained all the transactivation function of the larger insert at every concentration tested (Fig. 6). These data are in agreement with earlier studies that localized the activation domain of FKHR to amino acids 596 -655 (20).
FKHR Augments ER␣ Transactivation at an ERE-Because the carboxyl-terminal region of FKHR demonstrated strong ligand-dependent interaction with ER␣, it seemed reasonable to postulate that the interaction of ER␣ would have an effect on transactivation by these two factors. We were not able to examine the effect of an ER␣ interaction using the Gal4 system because ER␣ alone was able to induce expression from the Gal4 luciferase reporter independent of the FKHR-Gal4 fusion proteins. Therefore, it was necessary to use a different reporter assay and cell system to examine the functional consequences of the FKHR-ER␣ interaction.
The functional effects of the ER␣-FKHR interaction on the transcriptional activation of an ERE by ER␣ was examined using MCF7 cells (an ER␣-positive hormone-responsive breast carcinoma cell line) that has been shown to have negligible expression of FKHR (21). The activity of the Vit-ERE-LUC reporter plasmid was assayed in MCF7 cells in which an FKHR expression plasmid was co-transfected in the absence or presence of ER␣ ligands. As seen in Fig. 7A, ␤-estradiol treatment resulted in the expected transactivation of expression from the Vit-ERE-LUC plasmid. Also as expected, Tamoxifen blocked the activity of ␤-estradiol. In the absence of ligand, FKHR had no effect on expression; however, in the presence of ␤-estradiol, FKHR resulted in a reproducible augmentation of ER␣-mediated transactivation. In the presence of FKHR, Tamoxifen again reduced expression to baseline levels. These results indicate that FKHR can augment ER␣ transactivation through ERE sequences in a ligand-dependent fashion.
ER␣ Represses FKHR Transactivation at an IRS-The effect of the ER␣-FKHR interaction was examined for alterations in the ability of FKHR to transactivate through the IRS element from the IGFBP-1 promoter. FKHR-mediated transactivation of the 3XIRS-LUC reporter plasmid was assayed in MCF7 cells co-transfected with an FKHR expression plasmid in the presence or absence of ER␣ ligands. As seen in Fig. 7B, ␤-estradiol had negligible effects on expression from the 3XIRS-LUC reporter, whereas FKHR expression augmented luciferase expression in the absence of ER␣ ligand. In contrast, ␤-estradiol completely repressed FKHR-mediated transactivation. Tamoxifen partially repressed FKHR-mediated expression but had no effect on luciferase expression in the absence of FKHR. These results suggest that the ligand-dependent interaction of ER␣ and FKHR inhibits FKHR-mediated transcriptional activation.
FKHRL1-induced Cell Cycle Alteration Is Abrogated by Estrogen-FKHR and related family members are known to induce cell cycle arrest or apoptosis in tumor cells (17,22,23). The above experiments demonstrated that the interaction between ligand-activated ER␣ and FKHR repressed transactivation by FKHR while augmenting ER␣-mediated transactivation. To examine the physiologic consequences of this interaction, the effect of estrogen treatment on FKHR-induced cell cycle alterations was examined in MCF7 cells. FKHRL1 or the triple mutant FKHRL1-TM was cloned into the CMS-EGFP vector, which constitutively expresses EGFP. The plasmid pCMS-EGFP, pCMS-FKHRL1, or pCSM-FKHRL1-TM was introduced into MCF7 cells, and the cells were cultured in the presence or absence of ␤-estradiol for 24 h. The cells were analyzed by flow cytometry for EGFP expression to identify transfected cells and for DNA content indicative for cell cycle stage or apoptosis. Histograms of DNA content for EGFPpositive cells are shown in Fig. 8A. In the absence of ␤-estradiol, FKHRL1 expression reduced the percentage of cells in G 1 from 46% (pCSM-EGFP-transfected) to 39%. FKHRL1-TM contains three mutations in which all three phosphorylation sites have been mutated to alanine (17). Hence, the triple mutant cannot be inactivated by Akt-mediated phosphorylation. Ex-pression of FKHRL1-TM resulted in a progressive decrease in the percentage of cells in G 1 to 30%. In the presence of ␤-estradiol, transfection with vector alone, FKHRL1, or FKHRL1-TM had no significant effect on cell cycle. As seen in Fig. 8, the percentage of cells in G 1 transfected with FKHRL1 or FKHRL1-TM in the presence of estrogen was 50.2 and 48.5%, respectively. These values were not statistically different from vector-transfected cells. In FKHRL1-TM-transfected cells, there was a statistically significant difference in the percentage of cells in G 1 comparing cells grown in the presence and absence of estrogen (p Ͻ 0.001). These data, particularly when viewed in combination with the alterations in the percentage of cells in G 2 and the sub-G 1 (apoptotic) regions, demonstrate that FKHRL1, and to a greater extent FKHRL1-TM, induce changes in cell cycle in MCF7 cells that are reversed with ␤-estradiol.

DISCUSSION
The forkhead or "winged helix" transcription factors have been shown to play important roles in cell differentiation, embryogenesis, and oncogenesis (24,25). FKHR was first identified as a transcription factor that was involved in a translocation with PAX3 in alveolar rhabdomyosarcoma (26 -28). Based on homology in the forkhead domain, FKHR is most closely related to two other human genes, FKHRL1 (16) and AFX (29), and the Caenorhabditis elegans homolog Daf-16 (30). Studies of the functional regulation and physiologic effect of this gene family have confirmed the similarity initially based on their structural homology. The activity of FKHR and the related proteins are all regulated through phosphorylation by Akt (17,(31)(32)(33)(34). In HepG2 cells, FKHR has been reported to transactivate the IGFBP-1 promoter through the IRS (21, 34 -36). Insulin inactivation of FKHR is mediated through phosphorylation at three residues of FKHR (31,37). Phosphorylated FKHR is retained in the cytoplasm, and its exclusion from the nucleus is associated with a loss of target gene expression (17,31,32).
A number of studies have shown that FKHR and the related proteins can regulate apoptosis and cell cycle arrest. Cytokine stimulation in lymphocytes is mediated through phosphatidylinositol 3-kinase, and down-regulation of p27 kip1 is a common pathway leading to cell cycle stimulation. It has been shown in B cells that the inactivation of FKHRL1 by phosphorylation results in decreased p27 kip1 expression and reduced apoptosis (22). Similarly in neuronal tissue, FKHRL1 has been shown to FIG. 7. Effects of the ER␣-FKHR interaction on transactivation by ER␣ and FKHR. A, the Vit-ERE-LUC reporter was transfected into MCF7 cells with (ϩ) or without (Ϫ) pFKHR-MT as described under "Experimental Procedures." Phenol red-free estrogen-free medium containing no ligand (Ϫ), 17-␤-estradiol (E2) at 10 Ϫ8 M, or Tamoxifen (TAM) at 10 Ϫ6 M was added at 24 h as indicated. The cells were harvested at 48 h and assayed for luciferase and ␤-galactosidase activity as described under "Experimental Procedures." B, the 3XIRS-LUC reporter was transfected into MCF7 cells with (ϩ) or without (Ϫ) pFKHR-MT as described for A. Cells were treated with hormone-defined medium as described for A and then harvested at 72 h. Luciferase and ␤-galactosidase activity was assayed as described under "Experimental Procedures." RLU, relative light units.
induce apoptosis through transcriptional mechanisms that induce a number of genes including Fas ligand (17). Akt-mediated phosphorylation of FKHRL1 is induced by survival factors and is associated with retention of FKHRL1 in the cytoplasm. In addition, the PTEN tumor suppressor gene is known to antagonize the activity of phosphatidylinositol 3-kinase (38). In PTEN-deficient cells, FKHR and FKHRL1 are retained in the cytoplasm, whereas restoring PTEN induces the translocation of FKHR to the nucleus and transcriptional activity (23). In cells that demonstrate PTEN-mediated cell cycle arrest or apoptosis, FKHR activation is able to induce similar changes in cell cycle or apoptosis. AFX has also been shown to induce cell cycle arrest through the activation of p27 kip1 (39). Finally, these pathways have also been shown to be functional in breast cells. In MDA-MB-231 breast carcinoma cells, epidermal growth factor treatment activates phosphatidylinositol 3-kinase and Akt and has been shown to induce FKHR phosphorylation and nuclear exclusion (40).
Our studies have demonstrated that FKHR, FKHRL1, and AFX bind to ER␣ in a ligand-dependent fashion. One of the regions of FKHR involved in the interaction corresponds to the transactivation domain in the carboxyl terminus. The interaction of ER␣ with this region of FKHR was strongly dependent on ␤-estradiol. The data from reporter assays indicated that ER␣ repressed FKHR transactivation, and the repression by ER␣ was dependent on ␤-estradiol and partially reversed by Tamoxifen. The reporter assay data were supported by data showing that FKHRL1 and FKHRL1-TM induced alterations of the cell cycle in MCF7 cells that were reversed by estrogen. One interpretation is that under the conditions used here, FKHRL1 is inducing a cell cycle block in G 2 /M that is removed by estrogen. These results are consistent with a recent study showing that FKHR expression reduced the colony formation of MCF7 cells (41).
Estrogen is known to act as a mitogen in hormone-responsive breast tumors, and the interaction between ER␣ and forkhead transcription factors may explain the cell cycle arrest induced by estrogen withdrawal or Tamoxifen treatment. It has also been reported that estrogen replacement in postmenopausal women is protective for Alzheimer's disease (42,43). Because FKHRL1 has been shown to induce cell death by apoptosis in neuronal cells (17), it is plausible that ER␣ interaction with forkhead proteins in neuronal tissue may support cell survival.
The interaction between ER␣ and FKHR augmented ER␣dependent transactivation through EREs. This result was in contrast to a recent report by Zhao et al. (41), who reported that FKHR repressed ER␣ transactivation through EREs. However, those experiments were performed in HepG2 cells that constitutively express high levels of FKHR, which was in contrast to our experiments that were performed in hormone-responsive MCF7 cells that express negligible levels of FKHR. It may be that the differences in the physiologic effects of the ER␣-FKHR interaction observed depend on the cell line and the promoter being studied. Nevertheless, the finding that the ER␣-FKHR interaction augments ER␣ transactivation at an ERE while repressing FKHR-mediated transactivation makes physiologic sense in light the role of estrogen as a mitogen. Our data suggest that in cells in which both factors are expressed, the presence of estrogen stimulates expression of ER␣-responsive target genes while repressing FKHR target genes, resulting in progression of the cell cycle. Estrogen withdrawal (or treatment with antiestrogens) promotes FKHR activation, expression of FKHR target genes, loss of ER␣ target genes, and cell cycle arrest.
A variety of cofactors have been described for ER␣, and the repertoire of cofactors that are expressed will vary from cell to cell and within the same cell type depending on the physiologic conditions. This consideration was also evident in our examination of the effect of ER␣ on FKHR-mediated transactivation through an IRS element. Repression by ER␣ was maximal at 60 -72 h after ␤-estradiol treatment but was negligible at 24 h and intermediate at 48 h (data not shown). This means that the physiologic effect of the ER␣-FKHR interaction in MCF7 cells is likely to be influenced by factors induced by ␤-estradiol and may not be present in cells that do not express the receptor.
In summary, we have demonstrated a ligand-dependent interaction between ER␣ and the FKHR family of forkhead transcription factors. These data provide a novel mechanism of estrogen regulation through the modulation of forkhead transcription factors. This finding describes an important link between cell surface signaling mechanisms that act through the phosphatidylinositol 3-kinase pathway and nuclear hormone receptors. The physiologic consequences of this interaction will likely depend on the cell type and growth state of the cell. However, ER␣ and the forkhead proteins are found in a plethora of cell types, and this interaction is likely to influence either cell cycle arrest or apoptosis in a variety of tissues. The identification of cross-talk between these two signaling mechanisms also provides an important mechanism whereby steroid hormones can influence the action of cell surface receptors and cell surface ligands can influence the actions of steroid hormones.