Regulation of Estrogenic Effects by Beclin 1 in Breast Cancer Cells

Beclin 1 is an essential gene in autophagy, a cell survival pathway that enables use of long-lived proteins as a source of amino acids under conditions of nutritional deprivation (1–4). In autophagy, Beclin 1 interacts with class III phosphatidylinositol-3-OH kinase (vps34) during autophagosomal membrane engulfing of damaged cytoplasmic organelles and long-lived proteins. Autophagy can facilitate cell survival, delaying apoptotic death (5, 6). Alternatively, autophagy can facilitate a form of cell death called autophagic or type II programmed cell death, characterized by autophagic vacuoles in the cytoplasm (7). Autophagy and Beclin 1 are important in the balance of breast cancer cell growth and death (8, 9). The function of Beclin 1 is, in part, defined by its interaction with the antiapoptotic gene products, Bcl-2 and Bcl-xL (8, 9). Beclin 1 is also a tumor suppressor gene because one allele of Beclin 1 is lost in subsets of breast, prostate, and other tumors (10). Overexpression of Beclin 1 in MCF-7 breast cancer cells reduced tumorigenicity in nude mice (11). These results raise questions on molecular pathways by which Beclin 1 modifies estrogenic function in breast cancer.

Estrogens function through the receptors, estrogen receptor (ER)α and ERβ, the ligand-activated transcription factors controlling the growth of ER-positive breast tumors and target tissues (12, 13). ERα mediates the proliferative stimulus of E₂ on ER-positive breast cancer cells, whereas ERβ suppresses cell proliferation (14). Estrogenic ligands alter the conformation of ERs such that ERs recruit coregulatory proteins to the promoter/enhancer sites of responsive genes to facilitate transcription (12, 13, 15). ER-recognition of estrogen response element and associated changes in chromatin stimulate a cell- and tissue-specific network of genes, enabling E₂ to exert multiple functions. In addition, estrogenic action includes nongenomic mechanism(s), involving membrane ERs and activation of kinase cascades in association with G protein–coupled receptors or growth factor receptors (15–17).

Cell signaling through the ERα is pivotal to the regulation of breast cancer cell growth so that the ERα antagonist, tamoxifen, is the first targeted breast cancer therapeutic agent (18). 4-Hydroxytamoxifen is an active metabolite of tamoxifen, with higher affinity for ERα than tamoxifen (19). Although tamoxifen is initially effective against ERα-positive tumors, resistance develops in most women (20). The agonistic activity of tamoxifen on uterine cells is also a cause for concern (21). Raloxifene is an antiestrogen with mixed agonist/antagonist activity, in a tissue-specific manner (22, 23). Raloxifene is an agonist capable of enhancing bone density, whereas it is an antagonist on breast epithelial cells and the reproductive system. Understanding the basis of agonist activity of antiestrogens is an important aspect of overcoming antiestrogen resistance.

For a subset of ERα-positive tumors, cell growth is totally dependent on ERα (12, 13). Functions of ERα as a transcription factor and its ability to bind to accessory proteins and coregulators are critical to the control of cancer cell growth (12, 13). Because Beclin 1 inhibited E₂-dependent MCF-7 tumor growth in immunodeficient mice (10), an interaction between ERα signaling and Beclin 1 function seemed possible. We, therefore, examined the consequences of Beclin 1 overexpression on E₂-induced cell growth and cell signaling. We found that E₂-induced cell signaling and gene expression are modified by Beclin 1 and that an interaction of ERα and Beclin 1 might be involved in the inhibition of estrogenic cell signaling.

Cell culture. MCF-7 cell line was obtained from the American Type Culture Collection. We obtained MCF7.beclin cells expressing Beclin 1 from a tetracycline-repressible promoter (pTRE/flag-beclin 1; ref. 11), and a control cell line containing empty vector (MCF7.control) from Dr. Beth Levine (University of Texas Southwestern Medical Center, Dallas, TX). DMEM, phenol red–free DMEM, fetal bovine serum (FBS), and anti–β-actin antibody were from Sigma Co. Antibiotics, trypsin, and other additives for cell culture medium were purchased from Invitrogen. Anti-Beclin 1, antiphospho-Akt, anti-Akt, anti-pERK, and anti–phospho-ERK antibodies were purchased from Cell Signaling Technology. Anti-ERα antibody was from Santa Cruz Biotechnology. Secondary antibodies conjugated with Alexa Fluor 633 or Alexa Fluor 488 were from Invitrogen.

MCF-7 cells were maintained in DMEM, supplemented with 100 units/mL penicillin, 100 μg/mL streptomycin, 40 μg/mL gentamicin, 2 μg/mL insulin, 0.5 mmol/L sodium pyruvate, 50 mmol/L nonessential amino acids, 2 mmol/L l-glutamine, and 10% FBS. MCF-7.control and MCF7.beclin 1 cells were maintained in DMEM supplemented with 10% FBS, 40 μg/mL gentamicin, 5 mg/mL G418, 2 μg/mL insulin, 200 μg/mL hygromycin, and 2 μg/mL tetracycline. Before each experiment, cells were grown for 3 d in phenol red–free DMEM containing serum treated with dextran-coated charcoal (DCC) to remove serum-derived estrogenic compounds (16). MCF-7.beclin cells were grown for 3 d in phenol red–free DMEM without tetracycline, before plating under phenol red–free conditions.

[³H]-Thymidine incorporation. Cells (0.5 × 10⁶) were seeded in 6-well culture plates in phenol red–free DMEM supplemented with DCC treated serum and additives. After 24 h of plating, cells were treated with 4 nmol/L E₂. DNA synthesis was measured by adding 4 μCi/mL of [³H]-thymidine, 1 h before harvest (24). Cells were treated with 5% trichloroacetic acid followed by 1 N NaOH, and 1 N HCl. The radioactive thymidine in cellular DNA was quantified by liquid scintillation counting.

Cell proliferation. MCF-7.beclin cells were seeded at a density of 5 × 10⁴ cells per well in 24-well plates, in the presence of tetracycline in phenol red–free medium. A parallel study was set up with MCF-7.beclin cells without tetracycline. Cells were dosed 24 h after plating, with 4 nmol/L E₂, and redosed after 48 h. After appropriate treatment periods, live cells were counted using the trypan blue exclusion, using a hemocytometer.

CellTiter Glo assay. This assay measures cell viability by level of ATP content (25). Cells (5 × 10³ per well) were plated in 96-well plates. E₂/antiestrogens were added 24 h after plating. Cells were redosed with medium change at 48 h after the first dose. After 96 h, cells were treated with 0.1 mL of CellTiter Glo (Promega) reagent, incubated for 5 min at 22°C, and luminescence recorded using a GloMax Luminometer (Promega).

qPCR. MCF-7.beclin-1 or MCF-7.control cells (1 × 10⁶) were seeded in 60-mm culture dishes. After 24 h, cells were treated with E₂ (4 nmol/L) for 0, 1, or 2 h. RNA was isolated using Trizol reagent (Invitrogen), and 2 μg RNA was reverse transcribed using the first strand cDNA synthesis kit (Fermentas, Inc.) with random hexamers as primers. The expression of Tff1 (pS2), c-Myc, c-fos, Erg-1, Nur77, and Gapdh genes was determined by real-time PCR using the SYBR Green PCR Master Mix (Bio-Rad) with the following primers: 5′-CAATGGCCACCATGGAGAAC-3′ and 5′-AACGGTGTCGTCGAAACAGC-3′ for Tff1 (188 bp); 5′-CTCCTCACAGCCCACTGGTC-3′ and 5′-CTTGGCAGCAGGATAGTCCTTC-3′ for c-Myc (101 bp); 5′-CGGGCTTCAACGCAGACTA-3′ and 5′-GGTCCGTGCAGAAGTCCTG-3′ for c-Fos (147 bp); 5′-ACCTGACCGCAGAGTCTTTTC-3′ and 5′-GCCAGTATAGGTGATGGGGG-3′ for Erg-1 (110 bp); 5′-CGCACAGTGCAGAAAAACG-3′ and 5′-TGTCTGTTCGGACAACTTCCTT-3′ for Nur77 (145 bp); and 5′-CATGAGAAGTATGACAACAGCCT-3′ and 5′-AGTCCTTCCACGATACCAAAGT-3′ for Gapdh (113 bp). A final volume of 25 μL was used for qPCR in an IQ5 thermocycler (Bio-Rad). Amplification conditions were 95°C for 3 min, followed by 40 cycles of 95°C for 15 s and annealing for 30 s (57°C for Tff1 and 55°C for other genes). qPCR products were normalized relative to that of Gapdh to correct for differences in template input. Results are expressed as fold differences in expression of the indicated gene relative to that of Gapdh. Standard curves were generated for every target using six 4-fold serial dilutions.

Confocal microscopy. Cells were plated in Labtek 6-well slide chamber and dosed after 24 h (16). Cells were fixed in 4% paraformaldehyde, blocked in normal goat serum (5%) in PBS, followed by incubation with anti-ERα antibody (mouse) and anti-Beclin 1 antibody (rabbit) in 2.5% goat serum in PBS. After washing, cells were incubated with appropriate secondary antibodies. Alexa Fluor 633 conjugated anti-mouse IgG was used for ERα (deep red) and Alexa Fluor 488 conjugated anti-rabbit igG (green) was used for Beclin 1. Experiments were repeated using Alexa Fluor 488–conjugated anti-mouse IgG for detecting ERα (green) and Alexa Fluor 633 conjugated anti-rabbit IgG for detecting Beclin 1 (red). Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; 1 nmol/L). Images were recorded using a Zeiss 510 Laser scanning microscope with a ×60 objective at identical intensity settings for all treatment groups. No fluorescence was detected when cells were treated with fluorescence labeled secondary antibody alone.

Immunoprecipitation. Cells (1.5 × 10⁶ per 60-mm dish) were plated in dishes and allowed to attach for 24 h before treatments. After the specified treatments, cells were harvested and cell pellet lysed in buffer containing 50 mmol/L Tris.HCl (pH 7.4), 150 mmol/L NaCl, 1% Triton, 5 mmol/L EDTA, 25 mmol/L sodium fluoride, 25 mmol/L sodium PPi, 2 mmol/L sodium vanadate, 5% glycerol, 1 mmol/L phenylmethylsulfonyl fluoride and 1× protease inhibitor cocktail (Calbiochem; ref. 26). Cell lysate (100 μg in 100 μL lysis buffer) was precleared with Protein A/G agarose and incubated with 10 μL of anti-ERα (mouse monoclonal) antibody for 16 h at 4°C. Reaction mixture was incubated with protein A/G-Sepharose for 1 h at 4°C. Sepharose beads were washed thrice with cold lysis buffer, extracted using Laemmli buffer, and loaded on 12% SDS polyacrylamide gel for immunoblot analysis.

Western blot analysis. Cells (1.5 × 10⁶ per 60-mm dish) were plated in dishes and allowed to attach for 24 h before treatments. After the specified treatments and time periods, cells were harvested and cell pellet was lysed in buffer (16). Proteins (25 μg) were separated on a 12% SDS-polyacrylamide gel, transferred to polyvinylidene difluoride Polyscreen membrane and incubated with 1:200 to 1:1,000 dilution of the primary antibody. Protein bands were visualized using horseradish peroxidase–conjugated secondary antibody (Santa Cruz Biotechnology) and SuperSignal West Peco Chemiluminescent Substrate (Pierce). Membranes were stripped in stripping buffer [62.5 mmol/L Tris.HCl (pH 6.8), 2% SDS, 100 mmol/L (fresh) β-mercaptoethanol] and washed thrice in washing buffer [20 mmol/L Tris.HCl (pH 7.6), 150 mmol/L NaCl, and 0.05% Tween 20]. A second primary antibody was then added and the membrane reprobed as needed. To verify equal protein loading, membranes were stripped and reblotted with anti–β-actin antibody. Lightly exposed Kodak XAR Biomax films were scanned using an Epson B4 Scanner and band intensities quantified using the NIH Image J 1.34S program. Fold changes in the intensity of the protein signals reported are the mean of the results from three experiments.

Chromatin immunoprecipitation assay. Chromatin immunoprecipitation (ChIP) assay was performed as described previously (27) with minor modifications. Cells (8 × 10⁶) in 10-cm dishes were washed once with PBS and cross-linked with 1.5% formaldehyde at 37°C for 10 min. After washing with PBS, cells were collected in 1.6 mL of lysis buffer [0.5% SDS, 5.6 mmol/L EDTA, 33 mmol/L Tris.HCl (pH 8.1), 0.5% Triton X-100, 84 mmol/L NaCl] and incubated on ice for 30 min. Cell lysate was sonicated using Sonicator 3000 (Misonix). Samples were diluted 5-fold with dilution buffer [0.01% SDS, 1.2 mmol/L EDTA, 16.7 mmol/L Tris.HCl (pH 8.1), 1.1% Triton X-100, 167 mmol/L NaCl] and then precleared with salmon sperm DNA/Protein A agarose for 2 h at 4°C. Anti-ERα antibody (HC-20, 5 μg) or rabbit IgG (Santa Cruz) was used to immunoprecipitation from 200 μg of protein. Immunoprecipitated DNA was amplified by PCR using AccuPrime TaqDNA polymerase and visualized with ethidium bromide staining. The following primers from pS2 promoter used were as follows: 5′-CTAGACGGAATGGGCTTC ATG-3′ (forward) and 5′-TCCTCCAACCTG ACCTTAATCC-3′ (reverse).

Statistical analysis. All experiments were repeated at least thrice. Statistical difference between control and treatment groups was determined by one-way ANOVA followed by Dunnet's posttest using SigmaStat statistical program. A P value of <0.05 was considered to be statistically significant.

We tested the effect of E₂ on MCF-7.beclin cells that expressed Beclin 1 from a tetracycline-repressible promoter. MCF-7 cells transfected with control plasmid (MCF-7.control) were also treated with E₂ in the same manner. Cells were plated in 6-well plates, and E₂ was added 24 hours later at 0.1 or 4 nmol/L concentrations. At the time of dosing, 62% ± 3% of the cells were in the G₁ phase, as determined by flow cytometry with propidium iodide staining. [³H]-thymidine incorporation assay was conducted 24 and 48 hours after E₂ treatment. Figure 1A and B shows a comparison of E₂-induced DNA synthesis in MCF-7.beclin and MCF-7.control cells. MCF-7.control cells showed maximal DNA synthesis in the presence of 4 nmol/L E₂ after 48 hours of treatment. The level of DNA synthesis in E₂-treated cells was ∼2.5-fold higher than that of untreated cells at the same time point. In contrast, Beclin 1–transfected cells showed only 75% increase in DNA synthesis at the 48-hour time point in the presence of 4 nmol/L E₂. At the 24-hour time point, MCF-7.control cells showed a 2-fold increase in DNA synthesis due to E₂ treatment. However, only 35% increase in DNA synthesis was observed in MCF-7.beclin cells. To verify the level of Beclin 1 protein expression in MCF-7.beclin cells, Western immunoblots were conducted using Beclin 1 antibody. Our result (Fig. 1C) showed that MCF-7.beclin cells had a 5-fold higher level of Beclin 1 compared with MCF-7.control cells, as expected from its overexpression due to transfection.

We next examined E₂-induced changes in the expression of early response genes in MCF-7.control and MCF-7.beclin cells. As shown in Fig. 1D, there were significant changes in the pattern of expression of E₂-induced genes between MCF-7.beclin cells and MCF-7.conrol cells. As expected from the E₂ sensitivity of MCF-7.control cells, there was 2- to 3.5-fold higher levels of transcripts of Tff1, c-myc, c-fos, and Nur77 in these cells, after 2 hours of E₂ treatment, compared with those of untreated cells. In contrast, only c-myc mRNA level showed a moderate increase at 2 hours after E₂ treatment in MCF-7.beclin cells, whereas c-fos, Erg-1, and Nur77 mRNAs decreased, compared with those of untreated cells. All of the tested mRNAs showed lower levels in E₂-treated MCF-7.beclin cells, compared with E₂-treated MCF-7.control cells. These results indicate that the reduction of E₂-induced DNA synthesis in MCF-7.beclin cells is associated with altered expression of early-response genes.

E₂ is known to induce Akt phosphorylation as a part of nongenomic mechanism of action of ERα (16, 28, 29). We, therefore, examined whether E₂ response is modified in MCF-7.beclin cells. Cells were treated with 4 nmol/L E₂ for 5, 10, 20, and 30 minutes. Cells were harvested and phospho-Akt levels determined by Western blots. We found a decrease in Akt phosphorylation in MCF-7.beclin cells treated with E₂, compared with untreated cells (Fig. 2). In contrast, in MCF-7 cells, treatment with E₂ resulted in a 3-fold increase in Akt phosphorylation, 5 minutes after its addition (Fig. 2). We also determined the levels of total Akt and phospho-ERK1/2 in MCF-7.beclin and control MCF-7 cells. Although there was no change in the level of total Akt, p-ERK1/2 showed an increase, 10 minutes after E₂ treatment, in MCF-7.beclin and wild-type cells. The level of β-actin was comparable in all lanes. These results indicate that whereas E₂-induced Akt phosphorylation is down-regulated in the presence of Beclin 1 overexpression, increase in ERK1/2 phosphorylation remained intact.

We next examined the long-term effects of E₂, raloxifene and 4-hydroxytamoxifen, on MCF-7.beclin and MCF-7.control cells using CellTiter Glo assay. Cells were treated with E₂ or antiestrogens for 48 or 96 hours. Assay at 96 hours provided maximal responses to E₂ and antiestrogens. Our results (Fig. 3) show that the presence of E₂ increased the luminescence of MCF-7.control cells by 2-fold, whereas there was only a 33% increase in luminescence in MCF-7.beclin cells. Raloxifene and 4-hydroxytamoxifen inhibited E₂-induced growth of MCF-7.control cells at 500 nmol/L. In MCF-7.beclin cells, raloxifene had no significant effect even at 1,000 nmol/L. With 4-hydroxytamoxifen, E₂-induced proliferative response of MCF-7.beclin cells was not affected at 500 nmol/L 4-hydroxytamoxifen, but it was inhibited at 1,000 nmol/L. However, the basal proliferation of MCF-7.beclin cells or MCF-7.control cells in the absence of E₂ was not affected by these antiestrogens.

As an additional control for the effects of E₂ and antiestrogens in the absence of Beclin 1 overexpression, we studied MCF-7.beclin cells in the presence of tetracycline. E₂ response of cells containing tetracycline was similar to that of MCF-7 cells and antiestrogens suppressed the proliferative effect. However, when cells were treated with 4-hydroxytamoxifen and raloxifene in the absence of E₂, there was a 30% to 40% increase in proliferation, indicating that MCF-7.control and MCF-7.beclin cells containing tetracycline have some differences in antiestrogen responses.

We also examined E₂-induced response of MCF-7.beclin cells by trypan blue exclusion, using a hemocytometer (Fig. 3B). Experiments were conducted in the presence and absence of tetracycline. After 48 hours, E₂ treatment induced only 20% to 30% increase in cell number, regardless of the presence of tetracycline. By 4 days, E₂ treatment of MCF-7.beclin cells showed a 2.5-fold increase in cell number in the presence of tetracycline. However, there was only ∼50% increase in cell number in MCF-7.beclin cells in the absence of tetracycline. This result validates our findings on the influence of Beclin 1 on E₂ response, detected by CellTiter Glo assay.

ERα is known to redistribute within the cell after the addition of E₂ (16, 30). Therefore, we examined whether ERα and Beclin 1 interact and redistribute in the presence of E₂ by confocal microscopy (Fig. 4). In untreated cells, Beclin 1 (green) was found distributed in both cytoplasm and occasionally in the nucleus. In contrast, ERα (red) was mainly found in the nucleus. Addition of E₂ caused a movement of Beclin 1 to the perinuclear area. Addition of E₂ caused a redistribution of ERα throughout the cytoplasm and the nuclei in the case of mitotic cells (Fig. 4A) and toward cytoplasm in the case of clusters of cells (Fig. 4B). Addition of E₂ enhanced colocalization of ERα and Beclin 1, indicated by the yellow color in the merged photographs. Redistribution of Beclin 1 and its accumulation in the perinuclear area was illustrated by the increased colocalization of Beclin 1 with Golgin-97 (Fig. 4C) in E₂-treated cells. These results indicate that localization and function of Beclin 1 and ERα are being modified by E₂.

We next examined the effect of raloxifene and 4-hydroxytamoxifen on the localization of Beclin 1 and ERα in MCF-7.beclin cells. Figure 5 shows a representative confocal microscopic study of MCF7.beclin cells with and without raloxifene treatment. In these experiments, green represents ERα and red represents Beclin 1. Although a different set of secondary antibodies were used, accumulation of ERα in the nucleus and distribution of Beclin 1 throughout the cell remained similar to previous studies (11, 16). However, the intensity of Beclin 1 signal was much stronger for Alexa Fluor 633–conjugated (red) antibody (Fig. 5), compared with Alexa Fluor 488–conjugated (green) antibody (Fig. 4). Addition of raloxifene caused a reorganization of ERα or Beclin 1 so that these proteins were found throughout cytoplasm and nucleus, including membrane projections (Fig. 5). Merge pictures showed association of ERα and Beclin 1 with a strong yellow color, indicating colocalization of ERα and Beclin 1 (Fig. 5) in the presence of raloxifene. Colocalization of Beclin 1 and ERα was also observed in the presence of 4-hydroxytamoxifen (data not shown).

Subsequently, we examined the interaction of ERα and Beclin 1 by coimmunoprecipitation. Our results (Fig. 6A) showed that Beclin 1 was immunoprecipitated in association with ERα. Interestingly, addition of E₂ caused a 1.5- to 2-fold increase in the level of Beclin 1 associated with ERα, whereas the presence of raloxifene caused a 6-fold increase. The presence of 4-hydroxytamoxifen increased the level of Beclin 1 associated with ERα by 2.5-fold. Taken together, results of immunoprecipitation and subsequent immunoblotting confirmed the association of Beclin 1 with ERα, as observed in confocal microscopic studies. Although coimmunoprecipitation of ERα and Beclin 1 was found in untreated cells as well as in E₂- and raloxifene-treated cells, maximal level of coimmunoprecipitated proteins were found in raloxifene-treated cells.

To examine whether band intensities in the immunoprecipiation study has contributions from changes in the level of ERα and Beclin 1, we conducted direct Western blot analysis. Representative blots from 24-hour treatment are shown in Fig. 6B. Beclin 1 level did not undergo any major change due to treatment with E₂ or antiestrogens. ERα level decreased (10–25%) in the presence of E₂ and slightly increased in the presence of antiestrogens (1.5- to 2-fold) after normalizing to β-actin level. These changes in ERα are consistent with reports on the effects of E₂ and antiestrogens on ERα stability (16, 28).

We next conducted ChIP assay to elucidate the mechanistic differences in the action of E₂ and antiestrogens in MCF-7 and MCF-7.beclin cells. A comparison of the effects of E₂, 4-hydroxytamoxifen, and raloxifene on MCF-7 cells and MCF-7.beclin cells is presented in Fig. 6C. ChIP assay showed comparable E₂ responses on ERα occupation of pS2 promoter in both MCF-7 and MCF-7.beclin cells. Thus, the partial inhibition of the growth of MCF-7.beclin cells is not associated with ERα binding to pS2 promoter. In the presence of 4-hydroxytamoxifen and raloxifene, ERα occupation on pS2 promoter was lower in MCF-7.beclin cells than that in MCF-7 cells. This result argues against a role for Beclin 1 as a repressor that binds to both ERα and DNA. Thus, the lack of antiestrogen response in the presence of Beclin 1 might involve the interaction between ERα and Beclin 1, physically sequestering ERα away from promoter sites.

In this study, we examined the effect of Beclin 1 on E₂ function and found that Beclin 1–overexpressing cells had reduced proliferative response to E₂. qPCR analysis of gene expression showed that several E₂-responsive genes were down-regulated in cells expressing Beclin 1. E₂ caused a decrease in Akt phosphorylation in Beclin 1–transfected cells. Growth inhibitory responses of antiestrogens, raloxifene, and 4-hydroxytamoxifen were also limited by Beclin 1 transfection. In contrast, E₂-induced growth of MCF-7.control cells was totally inhibited by antiestrogens. The presence of E₂ caused a reorganization and accumulation of Beclin 1 in the perinucelar areas of cells. Importantly, we found a colocalization of ERα and Beclin 1 by confocal microscopy in MCF-7.beclin cells. An interaction between Beclin 1 and ERα was also evident from the results of coimmunoprecipitation, followed by Western immunoblotting. ChIP assay results suggest that interaction of ERα and Beclin 1 in the presence of antiestrogens leads to decreased promoter occupancy. Our results show that Beclin 1 can down-regulate estrogen-induced signaling events and cell growth in ERα-positive breast cancer cells. Our results further suggest that Beclin 1 may alter sensitivity of antiestrogens to breast cancer cells.

There have been extensive efforts to understand tamoxifen resistance in breast cancer cells (31). Although a primary mechanism of estrogen insensitivity involves the loss of ERα, tamoxifen resistance often occurs in the presence of ERα. Because ERβ has a down-regulatory effect on many estrogenic responses, it is the growth stimulatory responses of ERα that are generally targeted by tamoxifen (20, 32). A basic mechanism for acquired tamoxifen resistance is that ERα becomes insensitive, as the essential growth regulatory circuits are taken over by growth factor receptors, including members of the HER-2 family of receptors, or G protein–coupled receptors (20, 33). Our results show that overexpression of Beclin 1 provides a context under which ERα does not transmit cell signaling responses such as Akt phosphorylation. Down-regulation of E₂-responsive genes, such as c-fos, Erg-1, and Nur77, is indicative of an altered molecular environment for estrogenic function. ChIP assay results showed decreased receptor occupancy of pS2 promoter due to increased Beclin 1 expression, especially in the presence antiestrogens.

Both tamoxifen and raloxifene are selective estrogen receptor modulators, dependent on the presence of coactivator/corepressor proteins (34). Interaction of ligands with ERα and ERβ induces subtle conformational changes that amplified by the binding of tissue-specific proteins (35, 36). Raloxifene was first approved for the prevention of osteoporosis because it enhanced bone density (37, 38). After large comparative study of tamoxifen and raloxifene (STAR trial), raloxifene was found to decrease the incidence of breast cancer and was less prone to induce endometrial cancer, compared with tamoxifen (39). Cell culture studies have shown raloxifene resistance (40). The observation that both 4-hydroxytamoxifen and raloxifene showed a lack of sensitivity in MCF-7.beclin cells, indicates that a common factor in antiestrogen resistance might be the functional availability of ERα, when the receptor is complexed with proteins such as Beclin 1. Such an interaction might allow sequestering of Beclin 1, until conditions of nutritional deprivation or stress arise.

Beclin 1–overexpressig cells were first investigated by Liang and colleagues (11). Although these cells had decreased clonogenicity and showed a different morphology from that of wild-type MCF-7 cells, nutritional deprivation was necessary to produce a large number of autophagosomes (11). Although Beclin 1 was localized in the endoplasmic reticulum, mitochondria, perinuclear membrane as well as the nucleus, Beclin 1 mutants deficient in nuclear export were unable to facilitate autophagy (8). Interestingly, ERα also shuttles between nucleus, cytoplasm, and peripheral membrane and interacts with multiple proteins in these locations in regulating E₂-induced expression of hundreds of proteins (29). Thus, the interaction of ERα and Beclin 1 may affect the function of both Beclin 1 and ERα.

Our results on DNA synthesis and CellTiter Glo assays showed that MCF-7.beclin cells had a reduced growth response in the presence of E₂, compared with vector control cells. However, E₂-response was still present, yielding 30% to 75% increase in DNA synthesis at 24 and 48 hours, respectively. Phospho-Akt levels in MCF-7.beclin cells showed a decrease after treatment with E₂. Because E₂-induced phosphorylation of Akt is dependent on active ERα (15), the association of ERα and Beclin 1 may be involved in decreased Akt phosphorylation. In contrast, ERK1/2 was activated by E₂ in Beclin 1–overexpressing cells. ER-dependent nongenomic functions are dependent on many coregulators, including modulator of nongenomic actions of receptor (41) and growth factor receptors (42, 43). E₂ facilitates an interaction between ERα and epidermal growth factor receptor member, ERBB4 (43). Cross-talk also exists between ERα and ERBB2 (HER-2) signaling and insulin-like growth factor pathways, contributing to tamoxifen resistance (42, 44). Different G protein–coupled receptors and G proteins (such as Gαi and Gβγ) may be also involved in E₂ responses (45).

Recent studies suggest that Bcl-2 binding domain of Beclin 1 serves as a point of cross-talk between autophagic and apoptotic pathways (46). Bcl-2 down-regulation by an antisense oligonucleotide provoked autophagic cell death in HL-60 cells (47). Interestingly, crystallographic studies identified Beclin 1 as a Bcl-2 homology domain 3–only (BH3-only) protein (9, 48). This observation would classify Beclin 1 along with other proapoptotic BH3-only proteins, such as Bim, Bid, and Bok, suggesting a proapoptotic function that might be activated under unique tissue/cellular context (49). Whether the interaction of ERα with Beclin 1 modulates proapoptotic function by sequestering essential growth stimulatory proteins needs to be investigated.

In summary, our results reveal an interaction between ERα and Beclin 1 in breast cancer cells. This interaction may modulate the function of ERα and Beclin 1. In the context of ERα function, we found that Beclin 1–transfected cells were less sensitive to E₂-induced growth stimulation and to the growth inhibitory effects of antiestrogens. Thus, a novel function for Beclin 1 might involve down-regulation of the action of ERα, contributing to resistance of breast cancer cells to antiestrogens.

No potential conflicts of interest were disclosed.

Grant support: National Cancer Institute grants CA80163 (T. Thomas and T.J. Thomas) and CA42439 (T. Thomas and T.J. Thomas).

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. Beth Levine (University of Texas Southwestern Medical Center, Dallas, TX) for the MCF-7.Beclin and MCF-7.control cells.