ER-α36, a Variant of ER-α, Promotes Tamoxifen Agonist Action in Endometrial Cancer Cells via the MAPK/ERK and PI3K/Akt Pathways

Background Recently, a novel variant of ER-α, ER-α36 was identified and cloned. ER-α36 lacks intrinsic transcription activity and mainly mediates nongenomic estrogen signaling. Here, we studied the role of nongenomic estrogen signaling pathways mediated by ER-α36 in tamoxifen resistance and agonist action. Methodology The cellular localization of ER-α36 was examined by immunofluorescence in MCF-7 cells and Hec1A cells. MCF-7 breast cancer cells, MCF-7 cells expressing recombinant ER-α36 (MCF-7/ER36), Hec1A endometrial cancer cells and Hec1A cells with siRNA knockdown of ER-α36 (Hec1A/RNAiER36) were treated with 17β-estradial (E2) and tamoxifen (TAM) in the absence and presence of kinase inhibitor U0126 and LY294002. We examined phosphorylation of signaling molecules and the expression of c-Myc by immunoblotting, and tumor cell growth by MTT assay. Conclusions ER variant ER-α36 enhances TAM agonist activity through activation of the membrane-initiated signaling pathways in endometrial cancer, and that ER-α36 is involved in de novo and acquired TAM resistance in breast cancer.


Introduction
Tamoxifen is a selective estrogen receptor modulator (SERM) with mixed agonist/antagonist activities that has been used widely as an effective treatment of all stages of estrogen receptor (ER)positive breast cancer [1]. Tamoxifen suppresses the recurrence of breast cancer and reduces the incidence of contralateral breast cancer by 49% [2]. Tamoxifen has also been used as a chemopreventive agent in women who have high risk for breast cancer [3]. It is believed that tamoxifen acts as an antagonist by competing with estrogens for the ligand binding domain of ER, thereby inhibiting ER-mediated mitogenic estrogen signaling [4]. However, the major obstacle to tamoxifen usage is tamoxifen resistance, which occurs de novo or can be acquired after its use [5]. In addition, tamoxifen usage increases the incidence of endometrial cancer in postmenopausal women with long-term treatment [6]. The molecular mechanisms underlying both de novo and acquired tamoxifen resistance and its agonist action in endometrial tissue are poorly understood.
ER belongs to the steroid hormone family of the nuclear receptor superfamily. It is prevailingly considered that ER acts as a transcription factor that is mainly localized in the cell nucleus [7]. However, accumulating evidence has demonstrated that ER also exists on the plasma membrane and participates in rapid estrogen signaling. It has been reported that ER is modified by posttranslational palmitoylation in the ligand-binding domain that may contribute to its membrane localization [8]. Association of ER and caveolin-1 also was shown to facilitate ER localization on the plasma membrane [9]. Caveolin-1 is a structural protein of caveolae and serves as a scaffold protein to recruit signaling molecules such as growth factor receptors, G proteins, Src family tyrosine kinases and the PI3K [10]. It was postulated that estrogen may rapidly activate different signaling pathways, including MAPK/ERK, phospholipase C, PI3K/Akt and G proteincoupled receptor-activated pathways in the caveolae [11].
Recently, we identified and cloned a novel variant of ER-a with a molecular weight of 36 kDa that was named as ER-a36 [12]. The original 66 kDa ER-a was named ER-a66 [13]. ER-a36 transcript is initiated from a promoter located in the first intron of the ER-a66 gene and is generated from two alternative splicing events. ER-a36 protein thus lacks ligand-dependent andindependent transactivation domain (AF-1 and AF-2), but it retains DNA-binding domain and partial dimerization domain and ligand-binding domains [12]. ER-a36 possesses a unique 27 amino acid domain at the C-terminal that replaces the last 138 amino acids encoded by exons 7 and 8 of ER-a66 gene. Our previous report showed that 17b-estradiol and SERMs such as tamoxifen could induce activation of the MAPK/ERK pathway and stimulate cell proliferation through membrane-associated ER-a36 [14]. We thus hypothesized that ER-a36 may be associated with the agonist activity of tamoxifen. In the present report, we studied the ER-a36 function in ER-positive MCF-7 breast cancer cells and Hec1A endometrial cancer cells, and investigated the contribution of the MAPK/ERK and PI3K/Akt pathways mediated by ER-a36 to the agonist action of tamoxifen in endometrial cancer.

Results
ER-a36 Is Expressed on the Plasma Membrane in MCF-7 and Hec1A Cells ER-a36 is a novel variant of ER-a66 generated by alternative promoter usage and alternative splicing [12]. To examine ER-a36 expression in MCF-7 cells and Hec1A cells, Western blotting analysis was performed using ER-a36 specific antibody against the unique 20 amino acids at the C-terminal of ER-a36. ER-a36 is expressed in both cell lines (Fig. 1A, left). However, Western blot analysis failed to detect ER-a66 expression in Hec1A cells (Fig. 1A, right), consistent with that Hec1A is an ER-negative cancer cell line [15]. To examine the cellular localization of ER-a36, immunofluorescence assay was performed. In both cell lines, immunofluorescence staining revealed an intense plasma membrane distribution pattern (Fig. 1B). Caveolae are invaginated microstructures on the plasma membrane in which caveolin-1 serves as a scaffold protein to form the signaling complex. As shown in Fig. 1C, caveolin-1 was primarily expressed on the cell surface (red). Merged images of ER-a36 and caveolin-1 showed substantial co-localization signals (yellow) on the plasma membrane.
Next, we analyzed ligand-induced ER-a36 expression. Hec1A cell lines were treated with tamoxifen for different time points and ER-a36 expression was assessed by Western blotting analysis, revealing that ER-a36 expression was increased in tamoxifen treated cells (Fig. 1D).

ER-a36 Mediates Estrogen-and Tamoxifen-Stimulated ERK Activation
To probe the mechanism underlying the agonist effect of tamoxifen in endometrial cancer cells, we decided to examine the function of ER-a36 in tamoxifen treated Hec1A cells. We first examined the phosphorylation levels of MAPK/ERK, a serine-threonine kinase involved in cell proliferation [16]. As shown in Fig. 2A and Fig. 2B, E2 or tamoxifen treatments result in rapid phosphorylation of ERK1/2. Reprobing the membrane with a total ERK1/2 antibody indicated that the total ERK1/2 content was not changed, suggesting that the increased ERK1/2 phosphorylation was not caused by increased ERK1/2 expression.
To test the involvement of ER-a36 in the activites of E2 and tamoxifen observed in Hec1A cells that lack ER-a66 expression, we decided to knock down ER-a36 expression with the siRNA approach. We established stable cell lines that express shRNA expression vector against ER-a36 (Hec1A/RNAi cells) and examined ER-a36 expression ( Fig. 2C and 2D). As shown in  , E2 and tamoxifen failed to stimulate phosphorylation of the ERK1/2 in Hec1A cells with ER-a36 knocked down, suggesting that ER-a36 is the receptor that mediates the activities of estrogen and tamoxifen.
However, tamoxifen induced phosphorylation of the ERK1/2 in MCF-7/ER36 cells (Fig. 3D). MEK specific inhibitor U0126 effectively inhibited the ERK1/2 activation stimulated by E2 and tamoxifen (Fig. 3E). Therefore, these results indicated that ER-a36 mediates the Ras/MEK/ERK pathway induced by both estrogen and tamoxifen and suggested that ER-a36 may be involved in tamoxifen resistance and even promote agonist action of tamoxifen.

ER-a36 Mediates Estrogen-and Tamoxifen-Stimulated PI3K/Akt Activation
It is well known that the serine/threonine kinase Akt, or protein kinase B, plays an important role in cell proliferation and survival by inhibition of apoptosis [18]. We tested if E2 and tamoxifen treatment also induces activation of the Akt pathway in Hec1A cells. Treatment of E2 and tamoxifen led to rapid phosphorylation of Akt ( Fig. 4A and 4B) whereas both E2 and tamoxifen failed to induce the Akt phosphorylation in Hec1A/RNAi cells (Fig. 4C). Tamoxifen also induced Akt phosphorylation in MCF-7 cells that highly express recombinant ER-a36 (Fig. 4E). Pretreatment with the PI3K inhibitor LY294002 abrogated the Akt phosphorylation stimulated by E2 or tamoxifen in both cell lines ( Fig. 4D and 4F), indicating that ER-a36 mediates tamoxifen induced Akt phosphorylation through the PI3K pathway in these cells. Thus, our data suggested that ER-a36-mediated activaton of the PI3K/Akt pathway may also be involved in resistance and agonist action of tamoxifen.

ER-a36 Is Involved in Regulation of c-Myc Protein Expression in Hec1A Cells
Protooncogene c-Myc has profound mitogenic effects in cancer cells through its ability to promote cell cycle progression [19]. Antisense oligonucleotides to c-Myc can inhibit breast cancer cells proliferation [20]. Tamoxifen inhibits estrogen-induced c-Myc expression in ER-a66-positive breast cancer cells. However, c-Myc plays an important role in tamoxifen agonist action [21]. We measured the expression levels of c-Myc in Hec1A cells treated with E2 or tamoxifen. As shown in Fig. 5A, treatment with E2 or tamoxifen induced c-Myc expression in Hec1A/V cells but not in Hec1A/RNAi ER-a36 cells, which could be effectively abrogated by the MEK inhibitor U0126 (Fig. 5B) and PI3K inhibitor LY294002 (Fig. 5C).

ER-a36 Mediates Tamoxifen-Stimulated Cell Proliferation in Hec1A Cells
To further study the role of ER-a36 in tamoxifen agonist activity in endometrial cancer cells, Hec1A/V and Hec1A/RNAi cells were treated with tamoxifen and their prolifaration was measured with the MTT assay. MTT assay showed that tamoxifen stimulated growth of Hec1A/V cells. However, tamoxifen was able to inhibit the growth of Hec1A/RNAi cells in a dosedependent fashion (Fig. 6A). The cell proliferation induced by tamoxifen was inhibited by the MEK inhibitor U0126 and PI3K inhibitor LY294002 (Fig. 6B), suggesting that both the MAPK/ ERK and PI3K/Akt pathways were involved in E2 and tamoxifen stimulated cell growth in endometrial cancer cells.
We observed that tamoxifen strongly inhibited cell proliferation in the MCF-7/V cells, consistent with previous reports that tamoxifen functions as a potent antagonist in ER-positive breast cancer MCF-7 cells [22]. However, MCF-7/ER36 cells that constitutively express high levels of recombinant ER-a36 exhibited insensitivity to tamoxifen treatment (Fig. 6C). The MEK inhibitor U0126 and PI3K inhibitor LY294002 furthermore inhibited growth of both cell lines (Fig. 6D). These results again suggest that high level of ER-a36 expression may confer resistance to tamoxifen.

Discussion
Tamoxifen is a SERM that has been widely used to treat advanced ER-positive breast cancer and to prevent breast cancer in high risk pre-and post-menopausal women as a chemopreventive agent [23,24]. However, tamoxifen also has partial estrogenic activity in the uterus that can lead to endometrial hyperplasia [25]. Long-term tamoxifen usage is associated with an increased incidence of endometrial cancer [26]. Here we reported that a novel variant of ER-a, ER-a36, that is highly expressed on the plasma membrane of Hec1A endometrial cancer cells and in the endometrial cancer specimens from patients who had been treated with tamoxifen for at least three years. Both E2 and tamoxifen induced cell proliferation of Hec1A cells presumably through the ER-a36 mediated non-genomic signaling pathways.
A number of hypotheses have been postulated to explain tamoxifen's agonist action in endometrial carcinogenesis. It has been suggested that reactive metabolites of tamoxifen form DNAadducts and generate mutagenicity in the endometrial tissue [27]. It has also been demonstrated that the AF1 domain of ER-a66 as well as cell-and promoter-specific coactivator recruitment are involved in the tamoxifen agonist action [28,29]. The role of tamoxifen in endometrial carcinogenesis may utilize distinct genomic activity [30]. Recently, accumulating evidence suggested that membrane-initiated signaling pathways confer tamoxifen resistance and agonist action through different kinase cascades and distinct second messengers [15].
The MAPK family consists of ERK, JNK and P38. ERK plays an essential role in cell growth and proliferation. JNK and P38 are involved in cell differentiation and apoptosis induced by stress stimuli such as UV light [31], c radiation [32,33], DNA-damaging and chemopreventive drugs [34]. Many oncogenic signaling molecules activate the MAPK/ERK pathway [35]. ERK expression is usually increased and its activity is up-regulated in breast cancer tissues compared to neighboring normal tissues [36]. Furthermore, tamoxifen resistance in vivo is predominantly mediated by non-genomic mechanisms. Genomic estrogen action seems less active [37,38]. In this study, we found that ER-a36 mediates both E2-and tamoxifen-induced activation of the MAPK/ERK pathway and ER-a36 overexpression in tamoxifen sensitive MCF-7 cells reduced sensitivity to tamoxifen. In addition, ER-a36 mediates tamoxifen induced activation of the MAPK/ ERK pathway and also contributes to agonist action of tamoxifen in Hec1A endometrial cancer cells. Endometrial cancer tissues that highly express ER-a36 also displayed high levels of the ERK phosphorylation.
The PI3K/Akt pathway plays an important role in cell growth and survival [39]. Akt is activated by many signaling pathways, such as overexpression of growth factor receptors, [40]. Introduction of a constitutively active Akt into MCF-7 cells could induce tamoxifen resistance by protecting cells from tamoxifen-induced apoptosis [41]. In addition, the Akt activity is dramaticaly increased in tamoxifen-resistant MCF7 cells [18]. In phosphorylated Akt-positive patients, endocrine therapy has worse efficacy than in phosphorylated Akt-negative patients [42]. In this study, we found ER-a36 mediated tamoxifen-stimulated activation of Akt in cells with high levels of ER-a36 expression suggesting that the activation of the PI3K/Akt pathway mediated by ER-a36 contributes to the resistance and agonist action of tamoxifen.
The c-Myc protein is a nuclear transcription factor that plays an essential role in cell growth [43]. Previous studies have demonstrated that MAPK/ERK and PI3K/Akt pathways regulate c-Myc protein expression [44,45,46]. We found both E2 and tamoxifen induced c-Myc expression through ER-a36-mediated activation of ERK and Akt. Incubation of Hec1A cells with MEK inhibitor U0126 and PI3K inhibitor LY294002 blocked E2-and tamoxifen-induced c-Myc expression. Therefore, tamoxifen exerts agonist action through ER-a36-mediated non-genomic pathway.
In summary, we report here that ER-a36 is expressed on the plasma membrane and in the cytoplasm of endometrial carcinoma cells. We further demonstrated that both E2 and tamoxifen promoted proliferation of endometrial cancer cells through ER-a36-mediated activation of the MAPK/ERK and PI3K/Akt pathways and ER-a36 overexpression led to tamoxifen resistance in MCF-7 cells. Our results provide important novel information   to further understand the molecular mechanisms underlying the agonist action of tamoxifen.

Cell Culture and Cell Lines
MCF-7 human breast cancer cells were obtained from ATCC (Manassas, VA), and human Hec1A endometrial cancer cells were obtained from Dr. Li-Hui Wei (Peking University People's Hospital, Beijing). Both cell lines were maintained at 37uC with 5% CO 2 in appropriate culture medium. To establish MCF-7 cells expressing recombinant ER-a36, cells were plated at a density of 1610 5 cells per 60-mm dish and transfected 24 hours later with an expression vector driven by the cytomagalovirus (CMV) promoter in the mammalian expression vector pCB6+ as described elsewhere [14], using the Lipofectamine 2000 (Invitrogen, Carlsbad, CA). The expression vector contains the full-length ER-a36 cDNA. The empty expression vector was also transfected into cells to serve as a control. Forty-eight hours after transfection, the cells were re-plated and selected with 500 mg/ml of G418 for two weeks. The medium was changed every three days until colonies appeared. Clones were expanded for further analysis. Clones with high ER-a36 expression were a mixture of more then twenty clones and termed MCF-7/ER-a36. A cell line with pooled clones transfected with the empty expression vector was named MCF-7/V and used as a control.
We also established cell lines from Hec1A cells transfected with an ER-a36 shRNA expression vector (Hec1A/RNAi) and the empty expression vector (Hec1A/V). Briefly, ER-a36 shRNA expression vector pRNAT-U6.1/Neo plasmid containing the shRNA against ER-a36 (GenScript Corp. TX) and the empty expression vector were transfected into Hec1A cells with Lipofectamine 2000 according to the manufacturer's instruction. Forty-eight hours after transfection, the cells were re-plated and selected with G418 (600 mg/ml) for two weeks. Clones were expanded for further analysis.

Immunofluorescence and Confocal Microscopy
The cellular localization of protein was determined by indirect immunofluorescence. Hec1A or MCF-7 cells cultured on sterile glass coverslips were fixed in 4% paraformaldehyde in PBS for 10 min. After being permeabilized with 0.4% Triton X-100 at room temperature for 10 min, cells were blocked in 4% BSAsupplemented PBS for 1 h and incubated overnight at 4uC with anti-ER-a36-specific antibody. After three washes in PBS, the cells were labeled with FITC-conjugated secondary antibody. The DNA dye Hoechst 33258 was used for nuclear staining.
For double staining of ER-a36 and caveolin-1, after ER-a36 staining and wash in PBS, the cells were blocked in 4% BSA-supplemented PBS for 1 h at room temperature. After incubation with anti-caveolin-1-TRITC antibody overnight, the cells were further washed in PBS and stained with Hoechst 33258. Microscopic analyses were performed using a Confocal Laser-Scanning Microscope (Zeiss LSM 510 META, Germany).

MTT Assay
Cell proliferation was analyzed using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay [47]. Briefly, cells were seeded in 96-well plates to a final concentration of 5610 3 /well for MCF-7/V and MCF-7/ER36 cells or 3610 3 /well for Hec1A/V and Hec1A/RNAi cells. MCF-7/V and MCF-7/ ER36 cells were incubated in DMEM medium containing 10% FCS with the indicated treatments. Hec1A/V and Hec1A/RNAi cells were incubated in phenol-red free medium containing 2.5% dextran charcoal-stripped FCS (Biochrom AG, Berlin, Germany) with the indicated treatments. The cells were then incubated with MTT (0.5 mg/ml) for 4 h at 37uC. After removal the medium containing the MTT reagent, 150 ml of DMSO were added to each well. The plates were read at wavelength of 490 nm using a microplate reader (Bioteck Powerwave TM , USA).

Western Blotting Analysis
Cells were grown in phenol-red-free medium with 2.5% dextran charcoal-stripped FCS for 48-72 hours and then switched to medium without serum 12 h before stimulation by the agents indicated. The cells were collected in ice-cold PBS, and the cell extracts were prepared in RIPA buffer with the proteinase inhibitor cocktail from Sigma (St. Louis, MO). Cell lysates were boiled with gel-loading buffer for 5 min at 100uC, resolved on 10% SDS-PAGE, transferred to PVDF membranes, probed with appropriate antibodies and visualized with enhanced chemiluminescence (ECL) detection reagents (Amersham Pharmacia Biotech., Piscataway, NJ).

Statistical Analysis
Statistical analysis was performed using paired-samples t-test, or ANOVA followed by the Student-Newman-Keuls testing to determine differences in means. P,0.05 was considered statistically significant.