Cyclin D1 Antagonizes BRCA1 Repression of Estrogen Receptor α Activity

Germ line mutations in the BRCA1 tumor suppressor gene predispose women to ovarian and breast cancer and are associated with the development of prostate cancer in men (1, 2). In members of high-risk families, BRCA1 mutations increase the risk of breast cancer to 80% to 90% by age 70 years. The BRCA1 gene product, a nuclear polypeptide of 220 kDa, colocalizes in nuclear foci with RAD51, BRCA2, and RAD50 (3), consistent with a role for BRCA1 in DNA damage repair. In support of this model, BRCA1 is hyperphosphorylated in response to ionizing radiation by ATM, ATR, and hCds1/Chk2 (4, 5). In addition, protein complexes that regulate repair of DNA double-strand breaks, including hRAD50-hMre11-NBS1, MSH2, MSH6, and proliferating cell nuclear antigen, functionally interact with BRCA1 (3, 6). Furthermore, BRCA1-deficient cells are hypersensitive to ionizing radiation and exhibit defects in both repair of chromosomal double-strand breaks by homologous recombination and transcription-coupled repair processes (3, 7). The observation that BRCA1 associates with components of the DNA damage repair complexes in a temporally and spatially coordinated manner supports a model in which BRCA1 functions as a scaffold protein organizing DNA damage sensors, thereby coordinating the repair of damaged DNA (8).

Consistent with the genetic evidence that BRCA1 functions as a tumor suppressor, overexpression of wild-type (WT) BRCA1 inhibited colony formation and tumor growth in vivo, and antisense BRCA1 accelerated growth of breast cancer cell lines (9, 10). It has been proposed that BRCA1 may regulate expression and/or function of a select subset of genes and that mutation of BRCA1 may contribute to tumorigenesis by altering expression of these genes. In this context, roles for BRCA1 have been proposed as both a general and a selective transcriptional regulator. Components of the RNA polymerase II complex contain BRCA1 (10) and BRCA1 contains an acidic COOH-terminal activation domain sensitive to cancer-associated mutations (11). Although BRCA1 can associate with the RNA polymerase II, recent studies have shown that BRCA1 modulates transcription of only a subset of genes likely through interaction with specific binding proteins. Thus, BRCA1 has little or no effect on transcription controlled by Jun, Fos, Gal4, or USF (6, 12). In contrast, BRCA1 directly binds a subset of transcription factors, including p53, Myc, and ZBRK1, which may contribute to the regulation of a subset of genes and the cell cycle checkpoint control function of BRCA1.

One model proposed to explain the tissue-restrictive pattern of BRCA1-mediated tumor suppression implicates estrogen function in either initiation or progression of tumorigenesis (13). BRCA1 has been hypothesized to function as a tissue-specific tumor suppressor, repressing estrogen receptor (ER) signaling particularly during the early stages of breast tumorigenesis and dedifferentiation. It was considered unlikely that BRCA1 repression of ERα function contributed to the later stages of breast tumorigenesis when ERα expression was frequently reduced (14). In support of this model, BRCA1 was shown to inhibit the activity of the hormone-bound ERα in cultured breast cancer cell lines (12). In subsequent studies, BRCA1 was shown to physically interact in vitro with the ERα to repress ligand-dependent gene expression in breast cancer epithelial cells (15). BRCA1 also inhibited the basal activity of ERα in fibroblasts (16) and repressed a subset of estrogen-responsive genes in cultured 293 embryonal kidney cells (17), emphasizing the importance of cell type–specific regulators in determining BRCA1 function and the selectivity of BRCA1-dependent gene regulation. Together, these studies evidence the importance of understanding the mechanisms by which BRCA1 regulates ERα activity in breast and prostate epithelial cells.

The cyclin D1 gene encodes the regulatory subunit of a holoenzyme that phosphorylates and inactivates the retinoblastoma protein (pRb) tumor suppressor. Cyclin D1 is overexpressed in 30% to 40% of human breast cancers and is associated with poor prognosis in ERα-positive cases (18, 19). Evidence for a biologically relevant role for cyclin D1 in breast tumorigenesis includes findings that mammary gland–targeted cyclin D1 overexpression induced mammary tumors in transgenic mice (20), cyclin D1 antisense blocked ErbB2-induced mammary tumor growth in vivo (21), and mice deleted of the cyclin D1 gene were resistant to ErbB2- or Ras-induced mammary tumorigenesis (22). Given that BRCA1 function is compromised by mutations or reduced expression in human breast cancer and that cyclin D1 is a candidate mammary oncogene, we examined the possibility that cyclin D1 may antagonize BRCA1. In this report, we show that cyclin D1 and BRCA1 form distinct complexes with ERα in breast cancer cells. Cyclin D1 antagonizes BRCA1 repression of estrogen-responsive genes. Cyclin D1 overcomes BRCA1 repression of ERα-dependent gene transcription and competes with BRCA1 for ERα binding.

Cell culture. Human embryonic kidney 293T, DU-145, and MCF-7 cells were obtained from American Type Culture Collection (Manassas, VA) and maintained in DMEM supplemented with 10% FCS, 2 mmol/L glutamine, 100 units/mL penicillin, and 100 μg/mL streptomycin in humidified atmosphere containing 5% CO₂ at 37°C. Cells were seeded onto tissue culture dishes containing phenol red–free DMEM supplemented with 5% charcoal/dextran–treated fetal bovine serum and cultured for 48 hours before all experimental treatments with hormone.

Plasmids, transfections, and reporter assays. The expression vectors pCMV-cyclin D1 pCMV-cyclin D1-KE, CMV-cyclin D1-GH, pCMV-cyclin E, pCMV-cyclin A, Rous sarcoma virus (RSV)-Renilla luciferase reporter, and CMV-Luc were described previously (23). The human cyclin D1 mutants were derived by PCR-directed amplification using sequence-specific primers and cloned into pRC/CMV. The WT BRCA1 expression plasmid was generated by cloning the BRCA1 cDNA into the mammalian expression vector pcDNA3 (Invitrogen, Carlsbad, CA). The ERα expression constructions in pCI-Neo (Promega, Madison, WI) and glutathione S-transferase (GST) fusion proteins were derived from pHEGO. The reporter plasmid estrogen response element (ERE)₂-Luc contains two copies of the vitellogenin A2 estrogen-responsive enhancer (12). DNA transfection and luciferase assays were done as described previously (12). Cells were transfected using LipofectAMINE Plus and the medium was changed after 3 hours. Luciferase activity was determined after 48 hours. A β-gal reporter plasmid was included to control for transfection efficiency. The fold effect was determined by comparison with the effect of the empty expression vector cassette (pcDNA3).

Western blot and Northern blot. The antibodies used in Western blot analysis were BRCA1 (MS13, a gift from Dr. R. Scully, Harvard Institute of Medicine, Boston, MA), BRCA1 (C-20), ERα (H-184), and pS2 (C-20; Santa Cruz Biotechnology, Santa Cruz, CA). Proteins were visualized by the enhanced chemiluminescence system (Amersham Pharmacia Biotech, Piscataway, NJ). The abundance of immunoreactive protein was quantified using a computing densitometer (ImageQuant version 1.11, Molecular Dynamics, Sunnyvale, CA). Northern blot analysis was done using a full-length pS2 cDNA clone from American Type Culture Collection (cDNA collection).

Immunostaining. The primary antibodies were Ab-3 to cyclin D1, H-184 to ERα, and monoclonal antibodies MS110, AP16, and MS13 (generous gifts from Dr. R. Scully) to BRCA1. The secondary antibodies were goat anti-mouse IgG conjugated to Alexa Fluor 488 (green) and goat anti-rabbit IgG conjugated to Alexa Fluor 568 (red; Molecular Probes, Eugene, OR). MCF-7 cells were cultured on glass coverslip or glass bottom microwells (MatTek Corp., Ashland, MA). The cells were fixed for 8 minutes at 37°C with freshly prepared 3.7% paraformaldehyde solution with 0.18% Triton X-100. The fixation was followed by two washes with PBS. Specimens were blocked with 1% bovine serum albumin in PBS at 37°C for 30 minutes. Cells were incubated with primary antibody diluted with blocking solution at room temperature for 60 minutes, rinsed, and then incubated with secondary antibody conjugated with Alexa Fluor using the same conditions as the primary antibody binding reaction. For confocal microscopy, cells were fixed and permeabilized as described previously (24). Monoclonal BRCA1 antibodies were used at 1:30 dilution of the hybridoma culture supernatant. Images were collected using an Olympus (Melville, NY) IX-70 laser scanning microscope with objective lenses (×60 with oil). Images were collected through the specimens every 2 μm in the vertical plane and overlaid to generated focus composite images and stored as TIFF files. Figures were assembled from the TIFF files using Adobe Photoshop software.

Protein-protein interaction assays in vitro and in cultured cells. In vitro protein-protein interactions were done as described (23). In vitro [³⁵S]methionine-labeled proteins were prepared by coupled transcription-translation with a TNT-coupled reticulocyte lysate kit (Promega) using plasmid DNA (1.0 μg) in a total of 50 μL. GST fusion proteins were prepared from Escherichia coli cells (23). In vitro–translated protein (15 μL of In vitro–translated products) was added to 5 μg GST fusion protein of GST as control in 225 μL binding buffer [50 mmol/L Tris-HCl, 120 mmol/L NaCl, 1 mmol/L DTT, 0.5% NP40, 1 mmol/L EDTA, 2 μg/mL leupeptin, 2 μg/mL aprotinin, 1 mmol/L phenylmethylsulfonyl fluoride (PMSF), 2 μg/mL pepstatin] and rotated for 2 hours at 4°C. Glutathione-Sepharose bead slurry (50 μL) was added and the mixture was rotated for a further 30 minutes at 4°C. Beads were washed five times with binding buffer (1 mL), and binding buffer (30 μL) was added after the final wash. Sepharose beads were washed five times with lysis buffer and boiled in SDS sample buffer, and released proteins were resolved by SDS-PAGE. The gel was fixed, dried, and subjected to autoradiography.

Immunoprecipitation/Western blot analysis was done in HEK 293T cells. Transfected cells were treated for 24 hours with 10 nmol/L 17β-estradiol (E₂) or vehicle as control. Cells were rinsed with PBS, harvested by scraping, pelleted, and lysed in buffer [50 mmol/L HEPES (pH 7.2), 150 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1 mmol/L DTT, 0.1% Tween 20, 0.1 mmol/L PMSF, 2.5 μg/mL leupeptin, 0.1 mmol/L sodium orthovanadate]. Extracts were cleared by centrifugation and further precleared by rocking at 4°C with washed protein A-Sepharose (Roche Molecular Biochemicals, Indianapolis, IN). Precleared extract (500 μg) was immunoprecipitated with 1 μg cyclin D1 antibody (DCS-11, NeoMarkers Lab Vision Corp., Fremont, CA), BRCA1 antibody or equivalent amounts of appropriate control IgG (Santa Cruz Biotechnology), and protein A agarose (50 μL) for 6 hours to overnight at 4°C. Agarose beads were washed five times with lysis buffer and boiled in SDS sample buffer, and released proteins were resolved by 10% SDS-PAGE. The gel was transferred to nitrocellulose, and Western blotting was done.

Chromatin immunoprecipitation assay. Chromatin immunoprecipitation (ChIP) analysis was done following a protocol provided by Upstate Biotechnology (Charlottesville, VA) under modified conditions. MCF-7 cells (1 × 10⁶) were grown in DMEM with 10% charcoal/dextran–stripped serum for 3 days. Cells were treated with E₂ for the time points as indicated in the figure legends (45 minutes, 1 hour, and 12 hours). On E₂ (100 nmol/L) stimulation, the cells were cross-linked by adding 1.1% formaldehyde buffer containing 100 mmol/L NaCl, 1 mmol/L EDTA-Na (pH 8.0), 0.5 mmol/L EGTA-Na, and Tris-HCl (pH 8.0) directly to culture medium for 10 minutes at 37°C. The medium was aspirated, cells were washed thrice using ice-cold PBS containing 10 mmol/L DTT and protease inhibitors and lysed by warm 1% SDS lysis buffer and incubated for 10 minutes on ice. The cell lysates were sonicated to shear DNA to lengths between 200 and 1,000 bp and the samples were diluted to 10-fold in ChIP dilution buffer. To reduce nonspecific background, the cell pellet suspension was precleared with 60 μL salmon sperm DNA/protein A agarose-50% slurry (Upstate Biotechnology) for 2 hours at 4°C with agitation. Chromatin solutions were precipitated overnight at 4°C using 6 μg anti-ERα (H-184, Santa Cruz Biotechnology) with rotation. For a negative control, rabbit IgG was immunoprecipitated by incubating the supernatant fraction for 1 hour at 4°C with rotation. Salmon sperm DNA/protein A agarose-50% slurry (60 μL) was added for 2 hours at 4°C with rotation to collect the antibody/histone complex and washed extensively following the manufacturer's protocol. Input and immunoprecipitated chromatin were incubated at 65°C overnight to reverse cross-links. After proteinase K digestion for 1 hour, DNA was extracted using a Qiagen (Valencia, CA) spin column kit. Precipitated DNAs were analyzed by PCR. In separate experiments, MCF-7 cells were transiently transfected with plasmid DNA encoding cyclin D1, BRCA1, or empty vector and ChIP analysis was conducted for the endogenous pS2 promoter (25 cycles of PCR). The following primers were used for PCR of the pS2 promoter: sense 5′-GGCCATCTCTCACTATGAATCACTTCTGC-3′ and antisense 5′-GCAGAAGTGATTCATAGTGAGAGATGGCC-3′.

BRCA1 repression of estrogen receptor α transcriptional activity is selectively reversed by cyclin D1. We examined the possibility that cyclin D1 may regulate the ability of BRCA1 to inhibit hormone-bound ERα activity (12). We first assessed the specificity of BRCA1-mediated regulation of E₂-dependent ERα gene function. ERα-positive T47D breast cancer cells were transfected with either BRCA1 or control vector and treated with E₂ or vehicle for 24 hours. BRCA1 protein levels were increased 3- to 4-fold in transfected cells compared with vector (ref. 25; data not shown), a change within the physiologic range that occurs during cell cycle transition. ERα activity measured as E₂-stimulated activation of the estrogen-responsive reporter ERE-TK-Luc was inhibited by expression of BRCA1 (Fig. 1A). Transcriptional repression was selective, as BRCA1 did not inhibit either the activity of E2F-Luc or Sp1-Luc reporters or the transcriptional activity of Gal4-c-Jun or Gal4-E2F-1 (ref. 12; data not shown). Cyclin D1 did not affect activity of the liganded ERα but substantially reversed the BRCA1-mediated repression of ERα, increasing ERE-TK-Luc activity >200-fold (Fig. 1A, note log scale). In contrast, expression of other G₁- and S-phase regulatory cyclins (cyclin A, E, D2, or D3), from the same CMV expression plasmid, did not affect the BRCA1 inhibition of liganded ERα activity (Fig. 1A). Cyclin D1 at the same vector doses did not affect the activity of several different viral luciferase reporter constructions (RSV and CMV) in either the presence or the absence of E₂ in the same cells (data not shown), indicating the effect was specific for the ERE reporter gene. We assessed mutants of cyclin D1, defective in their ability to bind either cyclin-dependent kinase (cdk) 4/cdk6 (cyclin D1-KE), pRb protein (cyclin D1-GH), or ERα coactivator SRC-1 (cyclin D1 L254/255A; ref. 26). These mutants, which were expressed equally to WT cyclin D1 (27), conveyed a similar rescue function to WT cyclin D1 (Fig. 1B). We had the same observation in ERα-deficient DU-145 prostate cancer cells (Fig. 1C and D).

Cyclin D1 overcomes BRCA1-mediated estrogen receptor α activation function-2 activity. We next assessed whether the regulation of ERα activity by cyclin D1 in the presence of BRCA1 required the DNA-binding domain of ERα. The COOH-terminal activation function (AF)-2 domain of the ERα linked to the heterologous Gal4 DNA-binding domain was induced by E₂ in the human ERα-positive MCF-7 breast cancer cells and this activity was inhibited by >90% on expression of BRCA1 (Fig. 2A, note log scale). Cyclin D1 reversed BRCA1 repression of Gal4-ER AF2, suggesting that BRCA1 repression of ERα activity is not dependent on the ERα DNA-binding function. ERα contains two transactivation domains regulating ligand-independent and ligand-dependent gene transcription. AF1 consists of A and B domains, which are in the NH₂ terminus of ERα. We examined whether AF1 contributes to BRCA1 repression of ERα activation. Figure 2B shows that ERα amino acids 1 to 282 (AF1) activated ERE-TK-Luc reporter in the absence of ligand and overexpression of BRCA1 could not repress AF1-mediated ERα activation. Taken together, these studies suggest BRCA1 repression of ERα is independent of the AF1 and DNA-binding domain. Further, the ligand-binding domain is sufficient for BRCA1 repression and cyclin D1 enhances ERα function through either AF1 domain (28, 29) or AF2 by antagonizing BRCA1 repression.

Cyclin D1 rescues BRCA1 repression of 17β-estradiol–induced gene expression. The pS2 gene is induced by estrogen in breast cancer cells (30). To determine whether BRCA1 regulated pS2 expression, T47D cells were transfected with BRCA1 and pS2, gene expression was measured by Northern blot analysis. The 4-fold induction of pS2 gene expression by E₂ was abrogated by BRCA1 expression (Fig. 2C). Western blot analysis showed that cyclin D1 reversed the BRCA1-mediated repression of pS2 abundance in T47D cells (Fig. 2D).

Cyclin D1 and BRCA1 physically associate with estrogen receptor α through similar domains. To investigate the mechanisms by which cyclin D1 antagonized BRCA1 repression of ERα activity, we examined the possibility that cyclin D1 may complex with ERα in vivo under physiologic conditions. Immunoprecipitation with the cyclin D1–specific antibody showed that cyclin D1 associated specifically with ERα in MCF-7 cells (Fig. 3A), consistent with previous in vitro studies (28). T47D cells were subjected to immunoprecipitation with a BRCA1-specific antibody and Western blotting for ERα was done (Fig. 3B). ERα was coprecipitated with BRCA1, whereas control IgG did not precipitate ERα from an equivalent amount of cell lysate. We next determined the domain of ERα required for binding to cyclin D1. A series of mutant ERα expression plasmids were assessed in cultured 293T cells for expression by ERα Western blotting (Fig. 3C). Equal amounts of protein extract from the transfected cells were then subjected to immunoprecipitation with the cyclin D1 antibody and the immunoprecipitated product was subjected to ERα Western blotting. The experiments were conducted on at least three separate occasions with similar results. ERα was identified in the cyclin D1 immunoprecipitation. The abundance of ERα in the cyclin D1 immunoprecipitate was not reproducibly significantly changed by the addition of ligand (Fig. 3D, cyclin D1 immunoprecipitation). The ERα mutants with sequential deletion of the COOH terminus from N595 to N378 were efficiently coimmunoprecipitated with cyclin D1 by the cyclin D1 antibody. However, the ERα mutant 1 to 282 was not coimmunoprecipitated by the cyclin D1 immunoprecipitation, although this mutant was readily detected by direct Western blot analysis (Fig. 3D). These studies show that cyclin D1 binding to the ERα in cultured cells requires ERα residues 282 to 378.

GST pull-down experiments were next conducted using in vitro–translated, ³⁵S-labeled ERα incubated with equal amounts of GST-cyclin D1 fusion protein. Approximately 14% of the ERα input bound to GST-cyclin D1 (Fig. 3E). Deletion of the ERα residues 282 to 378 again abrogated binding of the ERα to cyclin D1.

Cyclin D1 competes with BRCA1 for estrogen receptor α binding. To investigate the mechanisms by which cyclin D1 rescued BRCA1 repression of liganded ERα activity, we determined the domains of the ERα required for binding to cyclin D1. As shown in Fig. 4A, five fragments that cover the full length of ERα were fused with GST. These GST fusion proteins were assessed for their binding capacity to cyclin D1 and BRCA1 (Fig. 4B and C). Equal amounts of GST fusion protein were incubated with the in vitro–translated, ³⁵S-labeled cyclin D1. The ERα 338 to 379 associated efficiently with cyclin D1. GST pull-down experiments were next conducted using ³⁵S-labeled BRCA1 amino acids 1 to 302, which was identified previously as the minimal region required for binding to ERα (15). The same domain of ERα (338-379) required for binding to cyclin D1 was also sufficient for binding to BRCA1 (Fig. 4C). Requirement of the same region of ERα for cyclin D1 or BRCA1 binding raised the possibility that cyclin D1 and BRCA1 may compete for binding to ERα. We examined this possibility by conducting ERα pull-down experiments with either cyclin D1 (Fig. 4D) or BRCA1 (Fig. 4E). We found that preincubation with increasing amounts of BRCA1 decreased cyclin D1 binding to ERα. Furthermore, preincubation with cyclin D1 reduced BRCA1 association with ERα (Fig. 4E). These results show that cyclin D1 has the ability to inhibit BRCA1 association with ERα, providing one potential mechanism for antagonism.

17β-Estradiol facilitates formation of an active estrogen receptor α complex in vivo. As cyclin D1 and BRCA1 were both capable of binding ERα, we reasoned that distinct complexes may form between these proteins in vivo serving to physically segregate their interactions and functional activities. BRCA1 localizes to discrete nuclear dots during S phase in MCF-7 cells (10). BRCA1 functions as a cell cycle checkpoint protein through its BRCT domain and BRCA1 associates with RNA polymerase II holoenzyme (10), suggesting that BRCA1 might be directly involved in transcriptional regulation. We examined the possibility that E₂ may alter the BRCA1/ERα association using immunofluorescent staining to view the dynamic chances of BRCA1 in response to E₂. Cells were grown in the absence of E₂ for 48 hours followed by E₂ treatment. Monoclonal antibody against BRCA1 (MS110, AP16, and MS13) and polyclonal ERα antibody (H-184) were used in this study. In the absence of E₂, most cells expressed ERα and BRCA1 to a comparable level (Fig. 5A, green for BRCA1 and red for ERα). Colocalization of BRCA1 and ERα (yellow) was observed in nuclear dots. E₂ treatment for 1 hour reduced BRCA1 and ERα colocalization (Fig. 5B). Approximately 50% of the cells maintained relatively strong expression of ERα (red) with reduced colocalization of BRCA1 in the nucleus. In ∼50% of the remaining cells, BRCA1 dots remained (Fig. 5B, green), but costaining of ERα was substantially reduced. Thus, the addition of ligand reduces ERα/BRCA1 colocalized dots. These studies do not necessarily reflect activity of the ERα within the local chromatin structure of an ERE; therefore, ChIP assays were conducted.

Cyclin D1 and BRCA1 form functionally distinct complexes with estrogen receptor α in the context of an estrogen-responsive gene promoter. It has been reported that ERα and p160 coactivators are recruited to the promoters of endogenous estrogen-responsive target genes contributing to transcriptional activation of ERα (31). Given the evidence that BRCA1 repressed ERα transactivation and estrogen-induced pS2 gene expression, we examined the recruitment of BRCA1 to the endogenous estrogen-responsive target gene pS2 in the context of its local chromatin structure using ChIP assays. The estrogen-responsive MCF-7 cell line was grown in the absence of estrogen for 3 days and then treated with either E₂ or vehicle control for 1 hour. ChIP assays were done to determine the endogenous transcriptional complexes present on the pS2 promoter. E₂ increased ERα recruitment to the pS2 gene promoter (Fig. 6). BRCA1 occupancy on the pS2 gene promoter was also observed and E₂ significantly decreased BRCA1 recruitment to the pS2 gene promoter (Fig. 6A). To examine the possibility that cyclin D1 may antagonize BRCA1 function at an estrogen-responsive promoter through inhibiting BRCA1 recruitment, we expressed cyclin D1 in MCF-7 cells and assessed BRCA1 occupancy by ChIP assay (Fig. 6B). Cyclin D1 expression reduced the amount of BRCA1 recruited to the pS2 promoter in both basal and E₂-induced states. BRCA1 transfection increased BRCA1 occupancy at the promoter ∼3-fold. Cyclin D1 expression reduced the effect of BRCA1 expression on BRCA1 occupancy at the pS2 promoter.

E₂ induces ERα to the ERE with cyclical kinetics (31). To examine the role of cyclin D1 and BRCA1 in regulating the kinetics of ERα recruitment to the ERE in the context of its local chromatin structure, ChIP assays were conducted. The pS2 promoter was used as a target of ERα. Cells were treated with E₂ for 0, 1, or 12 hours (Fig. 6C). These studies showed that ER is recruited to an ERE in the presence of ligand within 1 hour (Fig. 6C). Cyclin D1 enhanced ligand-induced recruitment of ERα to the ERE. BRCA1 expression reduced ERα recruitment in MCF-7 cells at both 1 and 12 hours of E₂ treatment. The coexpression of cyclin D1 and BRCA1 resulted in the relative increase of ERα recruitment at an ERE compared with the effect of BRCA1 alone (Fig. 6C). Compared with the effect of BRCA1, coexpression of cyclin D1 enhanced ERα recruitment to the ERE after 12 hours of ligand treatment. As cyclin D1 increases in response to E₂ treatment in ERα-positive breast cancer cells after 6 to 12 hours (32), the inhibition of BRCA1 occupancy at an ERE may contribute to the physiologic mechanism by which E₂-induced cyclin D1 expression antagonizes BRCA1 repression of ERα function in the context of the local chromatin structure. The finding that cyclin D1 enhances ERα recruitment to an ERE in the presence of E₂ and reduces the effect of BRCA1 to inhibit ER recruitment to an ERE may contribute to the ability of cyclin D1 to overcome BRCA1 repression at an ERE.

The ChIP assays for BRCA1 confirmed previous findings that BRCA1 occupies an ERE and that E₂ reduces BRCA1 recruitment (ref. 16; Fig. 6C). Herein, BRCA1 expression enhanced BRCA1 recruitment to the ERE. The expression of cyclin D1 reduced BRCA1 recruitment to the ERE at 0 hour but did not significantly alter the abundance of BRCA1 at an ERE at 1 and 12 hours of E₂ treatment.

The ChIP assays of cyclin D1 at the ERE showed that E₂ enhanced cyclin D1 recruitment to an ERE at 1 hour and that forced expression of cyclin D1 enhanced cyclin D1 presence at the ERE. The expression of BRCA1 reduced the ligand-induced recruitment of cyclin D1 to the ERE at 1 hour. Together, these studies suggest that cyclin D1, BRCA1, and ERα are recruited to an ERE in breast epithelial cells and that the presence of E₂ induces dynamic changes in the presence of these proteins with distinct kinetics. E₂ treatment (1 hour) is associated with the recruitment of ERα, increased presence of cyclin D1, and reduced BRCA1. The expression of BRCA1 reduces the recruitment of cyclin D1 at 1 hour and reduces ERα recruitment at 12 hours. These findings are consistent with the model in which BRCA1 reduces ERα occupancy at an ERE. In addition, these studies show that coexpression of cyclin D1 reduces the ability of BRCA1 to inhibit ERα recruitment in the presence of E₂ at 12 hours to an ERE. These findings are consistent with the model in which BRCA1 reduces ERα occupancy at an ERE in the presence of ligand and that cyclin D1 antagonizes this function, particularly at 12 hours.

Mapping functional domains of cyclin D1. The domains of cyclin D1 required for binding to ERα were assessed using expression plasmids encoding additional mutants of cyclin D1 (Fig. 7A). In vitro–translated [³⁵S]cyclin D1 and COOH-terminal truncated mutants were incubated with GST-ERα 282 to 420. Deletion of residues 142 to 178 of cyclin D1 completely abolished binding to ERα (Fig. 7A). The domains of cyclin D1 required for antagonism of BRCA1-mediated repression of ERα activity was next determined using expression plasmids encoding mutants of cyclin D1. The expression of mutant cyclin D1 proteins was at least equal to that of the WT cyclin D1 expression vector (27). Deletion of cyclin D1 residues 142 to 178 abolished functional antagonism of BRCA1 to <1% of WT cyclin D1. None of the NH₂-terminal cyclin D1 mutants provided BRCA1 rescue function in DU-145 cells (Fig. 7B) or T47D cells (data not shown). (In T47D cells, the rescue function of cyclin D1 1 to 142 was <3% of WT cyclin D1.) Thus, the domain of cyclin D1 between 142 and 178, which is predicted to form a helix-loop-helix (HLH) structure (27), is necessary for antagonism of BRCA1 repression of ERα activity.

The current studies provide the first evidence that cyclin D1 can directly regulate a transcriptional function of BRCA1. Herein, BRCA1 inhibited E₂-induced endogenous gene expression at levels of BRCA1 that are within the physiologic range. BRCA1 and ERα colocalized in nuclear dots. The addition of E₂ reduced BRCA1/ERα colocalization. In the context of the local chromatin structure by ChIP assay, BRCA1 binding to the estrogen-responsive pS2 gene was reduced by E₂ or cyclin D1 overexpression. As cyclin D1 abundance is induced by diverse mitogenic and oncogenic signals (21, 33, 34), the antagonism by cyclin D1 of BRCA1 function at the level of the liganded ERα may contribute to signal transduction cross-talk between oncogenic signals and BRCA1 function.

The direct association between BRCA1 and ERα as shown herein provides a physical means by which BRCA1 may be tethered to select transcriptional regulatory loci. In the current studies, BRCA1 and ERα were physically associated using in vitro GST pull-down and coimmunoprecipitation assays (Figs. 3 and 4). The direct binding of BRCA1 to ERα may serve to select BRCA1 genomic targets as shown herein for the pS2 gene. BRCA1 interacts with general transcriptional regulators, including p300/CBP (35), corepressors (CtIP/CtBP and HDAC complexes; refs. 36, 37), and the BRG1 subunit of the SWI/SNF complex (35, 38), yet regulates expression of a relatively specific subset of transcription factors and target genes. BRCA1 directly binds a subset of transcription factors, including p53, Myc, and ZBRK1, without affecting transcription controlled by Jun, Fos, Gal4, or USF (6, 12). The selective regulation of a subset of genes by BRCA1 likely contributes to the diverse role of BRCA1 including cell cycle checkpoint control function (39, 40). The ability of BRCA1 to repress ERα signaling is not limited to one or a few estrogen-responsive genes, such as pS2. Previously, we showed that in MCF-7 breast cancer cells overexpression of BRCA1 blocks the E₂-inducible expression of ∼90% of the E₂-responsive transcriptosome. Thus, as determined by DNA microarray analysis and confirmed by semiquantitative and quantitative reverse transcription-PCR analysis, exogenous BRCA1 blocked the ability of E₂ to alter the expression of the vast majority of E₂-responsive genes in MCF-7 cells (41).

The current findings extend previous studies in which E₂ reduced BRCA1 binding to an ERE (16) by firstly demonstrating the kinetics of this process, secondly by demonstrating the effect of cyclin D1 expression on ERα recruitment, and thirdly by characterizing the effect of cyclin D1 to inhibit BRCA1 recruitment in the basal state and after 12 hours of E₂ treatment. The ChIP studies herein show that cyclin D1 antagonizes BRCA1 inhibition of ERα occupancy at an ERE in the context of its local chromatin structure. This model is compatible with the identification of a ternary complex between cyclin D1/ERα and cyclin D1 by immunoprecipitation herein and suggests cyclin D1 regulates function of proteins in the complex at the local promoter level. The differences between ChIP and immunoprecipitation studies typically relate to the relatively small fraction of the target protein (in this case, the ERα) that is recruited to the promoter in the context of its local chromatin structure. Frequently, proteins identified in immunoprecipitation/Western analysis are not the same as those present within the transcriptional complex; alternatively, the dynamics of the entire pool of interacting proteins (identified by immunoprecipitation/Western blotting) is not the same as the dynamics of the specific pools of proteins present at the relevant promoter sites.

It is important to note that BRCA1 levels that antagonized ERα signaling herein are consistent with changes in BRCA1 levels that occur during physiologic and pathologic perturbations. BRCA1 levels change 8-fold during cell cycle transition and BRCA1 expression is high in proliferating mammary epithelial cells undergoing differentiation during puberty and pregnancy (42). Thus, BRCA1 is highly expressed in specific windows of time that may be most pertinent to the BRCA1 tumor suppressor function. The mechanism by which cyclin D1 antagonized BRCA1 function is quite distinct from our previous studies of p300 function (25). p300 exhibits a “rescue” function that is quite distinct from its ERα coactivator function. BRCA1 down-regulates p300 and exogenous p300 rescues BRCA1 repression of ERα. The difference between the rescue and coactivator function was shown by the finding that the CH3 domain of p300 was both necessary and sufficient to rescue repression of ERα activity by BRCA1, whereas its coactivator function requires other domains (e.g., histone acetyl transferase domain and SRC-1-binding domain) that are not required for rescue. Assessment of several other ERα coactivators (SRC-1 and PCAF) failed to rescue BRCA1 repression. Clearly, the coactivation and rescue functions are separable structurally and functionally.

The domain of cyclin D1 required to antagonize BRCA1 repression at an ERE is distinct from the previously described cdk-, pRb-, or SRC-1-binding residue, requiring a domain between residues 141 and 178, which was required for ERα binding, and is predicted to form a HLH-like structure (27). The domain of cyclin D1 required for reversal of BRCA1 repression of the ligand-bound ERα correlated with the domain of cyclin D1 required for ERα binding. Cyclin D1 inhibited BRCA1 association and ERα in vitro and reduced BRCA1 recruitment to endogenous estrogen-responsive gene using ChIP assays. These studies suggest the physical association of cyclin D1 with ERα may exclude BRCA1 association with ERα and thereby contribute to activation at an ERE. BRCA1 binds strongly to an ERα region (amino acids 338-379) and much more weakly to a second site (amino acids 420-595). Cyclin D1 interacts with the ERα at the major site of interaction with BRCA1, consistent with a model in which cyclin D1 may exclude physical access of BRCA1 to the ERα. Alternatively, the interaction between cyclin D1 and ERα may alter the conformation of BRCA1 bound to the ERα and/or change binding of other BRCA1-associated proteins (Fig. 8). These two models remain nonmutually exclusive and compatible with the ChIP data, indicating reduced association of BRCA1 at an ERE in the presence of coexpressed cyclin D1 after 12 hours of E₂ treatment. As distinct clusters of genes are induced by E₂ with distinct temporal kinetics by genome-wide analysis, the identification of the genes regulated by E₂ at 12 hours and repressed by BRCA1 may be of considerable interest.

It is important to consider the current findings in the context of recent epidemiologic data. On one hand, immunohistochemical and microarray analyses of human breast tumors harboring BRCA1 mutations are usually ER negative (43–45). Further, BRCA1 mutant breast cancers have a basal epithelial phenotype unlike ER-positive tumors (46). In contrast, prophylactic oophorectomy significantly reduces the risk of breast cancer in BRCA1 carriers (50% overall and as high as 75% if done at an early age; refs. 47, 48). This is prima facie evidence that ovarian hormones contribute to breast cancer risk in BRCA1 carriers. Secondly, in the general female population, early first pregnancy (which is associated with very high circulating levels of E₂ and progesterone) reduces the risk for breast cancer. However, in female carriers of BRCA1 mutations, early first pregnancy increases breast cancer risk, providing further evidence that the hormonal milieu influences the development of breast cancer in these women (49). Thirdly, studies by Zheng et al. (16) established that inactivation of BRCA1 (by gene deletion or antisense) causes ligand-independent activation of ER. Finally, although BRCA1 mutant cancers are usually ER negative, sporadic breast cancers are usually ER positive. Multiple studies indicate that 30% to 40% of sporadic breast cancers show absent or reduced expression of BRCA1 (50–52) and 46% of sporadic breast cancers show loss of one BRCA1 allele (i.e., these cancers may be haploinsufficient for BRCA1; ref. 53). Together, these findings suggest that BRCA1 regulation of ERα function may be important to both BRCA1 carriers and sporadic breast cancers, which are known to exhibit a high incidence of ER positivity.

Cyclin D1, which is overexpressed in human breast cancer, is both necessary and sufficient for mammary tumorigenesis in transgenic mice (20, 22). The BRCA1 protein is WT in the majority of human breast cancers, and inhibition of BRCA1 activity by cyclin D1 would be predicted to provide both a growth advantage imparted by cyclin D1 overexpression and an important activation through enhancement of liganded ERα activity. Cyclin D1 overexpression is a common early event during mammary tumorigenesis and the loss of ERα expression in BRCA1-associated tumors is likely to represent a relatively late event (14). Like cyclin D1 overexpression, BRCA1 mutations found in human breast cancer fail to repress liganded ERα activity (12). The prevailing view in cancer cell biology is that two general classes of cancer genes exist (54). The first class of genes regulates genomic stability and includes antimutators and DNA repair proteins. The second class of genes is cell cycle control genes involved in cellular proliferation and tumor growth, including cyclins, cdk inhibitors, and pRb. BRCA1 plays an essential role in the maintenance of genomic integrity during DNA damage repair (6, 55) functioning as a sensor of DNA damage by physically interacting with and thereby serving to recruit components of the DNA repair apparatus. The finding in the current studies that the product of the cell cycle control gene cyclin D1 physically interacts with BRCA1 and antagonizes BRCA1 repression function provides evidence that BRCA1 functions as a tumor suppressor by physically interacting with both classes of cancer-promoting genes.

Grant support: Susan Komen Breast Cancer Foundation, BCTR0504227 (C. Wang) Breast Cancer Alliance, Inc., grants R01CA70896, RO1CA75503, RO1CA86072, and RO1CA86071 (R.G. Pestell), RO3AG20337 (C. Albanese), RO1CA18119 (B.S. Katzenellenbogen), RO1ES109169, RO1CA82599, and RO1CA80000; and U.S. Army Breast Cancer Research Program grant DAMD17-99-1-89254 (E.M. Rosen).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. P. Chambon for the ERα constructs.