Stable Inhibition of Specific Estrogen Receptor α (ERα) Phosphorylation Confers Increased Growth, Migration/Invasion, and Disruption of Estradiol Signaling in MCF-7 Breast Cancer Cells

Abstract

Elevated phosphorylation of estrogen receptor α (ERα) at serines 118 (S118) and 167 (S167) is associated with favorable outcome for tamoxifen adjuvant therapy and may serve as surrogate markers for a functional ERα signaling pathway in breast cancer. It is possible that loss of phosphorylation at S118 and/or S167 could disrupt ERα signaling, resulting in aggressive ERα-independent breast cancer cells. To this end, MCF-7 breast cancer cells were stably transfected with an ERα-specific short hairpin RNA that reduced endogenous ERα. The resulting cell line was stably transfected with wild-type ERα (ER-AB cells), or ERα containing serine to alanine mutation at S118 or S167 (S118A cells and S167A cells, respectively). These stable cell lines expressed approximately equivalent ERα compared with parental MCF-7 cells and were evaluated for growth, morphology, migration/invasion, and ERα-regulated gene expression. S118A cells and S167A cells exhibited increased growth and migration/invasion in vitro. Forward- and side-scatter flow cytometry revealed that S167A cells were smaller in size, and both S118A and S167A cells exhibited less cellular complexity. S118A and S167A cells expressed pancytokeratin and membrane localization of β-catenin and did not express vimentin, indicating retention of epithelial lineage markers. Expression of ERα-target genes and other genes regulated by ERα signaling or involved in breast cancer were markedly altered in both S118A and S167A cells. In summary, attenuated phosphorylation of ERα at S118 and S167 significantly affected cellular physiology and behavior in MCF-7 breast cancer cells, resulting in increased growth, migration/invasion, compromised expression of ERα target genes, and markedly altered gene expression patterns.

Approximately 12% of all women in the United States will develop invasive breast cancer with an estimated 200,000 new cases diagnosed annually, the majority of which will be estrogen receptor-α (ERα) positive (1). Historically, the presence of ERα has been thought to imply competent ERα signaling; thus, ERα-positive tumors were considered prime candidates for ER-targeting drugs such as mixed antagonist tamoxifen and aromatase inhibitors. However, clinical data have shown that approximately half of ERα-positive tumors are resistant to endocrine therapy, and many patients who are initially responsive to endocrine therapy develop resistance even though the tumors remain ERα positive (2, 3). Recent retrospective clinical studies in breast cancer have elucidated a connection between clinical responses to endocrine therapy and histological measurements of ERα phosphorylation levels in archival breast cancer tissue specimens (2). Murphy et al. (4) reported that ERα-positive breast tumors exhibiting elevated ERα phosphorylation at serine 118 (S118) and serine 167 (S167) exhibited a more differentiated phenotype and were clinically less aggressive. Furthermore, elevated phosphorylation at S118 and S167 was associated with better clinical outcome for patients taking adjuvant tamoxifen therapy and was also associated with other positive prognosis markers (5, 6). These data were in agreement with Jiang et al. (7) who reported that relapse-free survival was closely associated with ERα-positive tumors exhibiting ERα S167 phosphorylation. Interestingly, Yamashita et al. (3) reported that tissue samples examined from patients who received various regimens of endocrine treatment (i.e. tamoxifen or aromatase inhibitor agents) showed a correlation between elevated S167 phosphorylation and decreased S118 phosphorylation. Collectively, these studies demonstrated that phosphorylation at certain sites in ERα may serve as surrogate markers for a functional ERα signaling pathway in breast cancer. In this regard, phosphorylation of ERα could identify patients who might be more responsive to endocrine therapy. Laboratory models are needed that can test these hypotheses to evaluate the role of ERα phosphorylation in ERα signaling and endocrine therapy response.

A key mediator of estrogenic effects, ERα is a member of the super family of ligand-activated steroid hormones receptors. Although ERα is activated by ligand binding, ERα function is also regulated in large part by posttranslational modifications, most notably by phosphorylation (8, 9). Functional events such as interaction with chromatin, coactivator recruitment, nuclear localization, receptor stability, and hormone responsiveness may be regulated by individual ERα phosphorylation sites (10–14). Phosphorylation of ERα may occur in both a ligand-dependent and ligand-independent manner and is regulated by numerous growth factor signaling pathways (2, 9, 15, 16). ERα phosphorylation sites have been identified in all functional domains of ERα, although most phosphorylation occurs in the N-terminal AF-1 domain (2, 8, 9, 14).

Many studies have focused on two well-characterized ERα phosphorylation sites, serine 118 and serine 167, which are directly phosphorylated by the MAPK and AKT kinase signaling pathways and indirectly through additional kinase/phosphatase cascades (9, 16). S118 has been shown to be the major site of ERα phosphorylation in response to estradiol stimulation as well as the substrate for many other kinases (9, 13, 16). Unlike many ERα phosphorylation sites, S167 is not contained within a Ser-Pro consensus motif and phosphorylation of S167 by multiple kinases indicates that several signaling pathways regulate this site (9, 16). Despite the plethora of published data regarding ERα phosphorylation, specifically at S118 and S167, the in vivo relevance of phosphorylation of ERα is still unclear.

Previous studies have used transient transfection of ERα phospho-mutants to study the effects of altered ERα phosphorylation on ERα signaling. Ali et al. (18) reported that ERα-negative COS-1 and HeLa cells transiently expressing ERα with a serine to alanine mutation at S118 (ERα-S118) showed reduced transactivation in response to estrogen stimulation compared with wild-type (WT) ERα. Similarly, Bunone et al. (19) reported that SK-BR-3 cells transiently expressing ERα-S118 exhibited a 30% reduction in hyperphosphorylation in response to estradiol. However, Le Goff et al. reported minimal loss in transcriptional activity of cells transiently transfected with ERα-S118 compared with WT ER (15). Additionally, cells transiently expressing ERα with a serine to alanine mutation at serine S167A (ERα-S167) exhibited reduced transcriptional activity (20). Recent studies from this laboratory showed that mutation of S118 to alanine expressed in ERα-negative HeLa cells reduced expression of genes involved in canonical estrogen response element (ERE) signaling and abrogated expression of genes involved in nongenomic signaling (21). Furthermore, this mutation resulted in altered recruitment of coregulators to the promoters of estrogen-regulated genes.

These studies have provided a wealth of information for the current understanding of the role of phosphorylation in ERα function; however, studies conducted in transient expression systems fail to adequately mimic the in vivo microenvironment of patient tumors, making it difficult to predict the effects of chronic changes in ERα phosphorylation on cellular physiology. These previous studies were also designed within a myriad of cellular genetic backgrounds and with some cell lines that did not express endogenous ERα and therefore evolved without responsiveness to ERα signaling. The present study altered ERα S118 or S167 phosphorylation through stable expression of ERα phosphorylation mutations in MCF-7 breast cancer cells in which endogenous expression of ERα was markedly reduced but not completely eliminated. These cell lines permitted the evaluation of morphology, cellular physiology, and ERα signaling in the context of cells that exhibited attenuated phosphorylation at two sites, a condition that would be more reminiscent of altered ERα phosphorylation in a physiological breast cancer tissue.

Materials and Methods

Cell culture

ERα-positive MCF-7 cells were obtained from American Type Culture Collection (Manassas, VA). Cell lines were maintained in DMEM high glucose (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS; Cellgro, Manassas, VA), 1% penicillin-streptomycin (Invitrogen), 2 mm l-glutamine (Invitrogen), and 1 mm sodium pyruvate (Invitrogen). Cells were maintained at 37 C with 5% CO₂. Cells stimulated with hormone in growth assays were cultured overnight in DMEM supplemented with 10% FBS followed by addition of 17β-estradiol (10⁻⁸ m), tamoxifen (10⁻⁷ m), or ICI 182,780 (10⁻⁸ m). For PCR, cells stimulated with 17β-estradiol were first cultured to 80% confluency in DMEM supplemented with 10% FBS. Cells were then transferred to phenol red-free DMEM (Invitrogen) without serum for 48 h followed by incubation with 10⁻⁸ m 17β-estradiol.

Generation of MCF-7 cell lines with stable expression of ERα phosphorylation mutants

MCF-7 cells exhibiting reduced levels of endogenous ERα protein were generated by stable transfection of a short hairpin RNA (shRNA) directed against the 3′-untranslated region (3′-UTR) of human ERα. Sequences encoding the ERα 3′-UTR-directed shRNA (forward 5′-GCGCCGCCTACGAATTTAACGCCGCGGCCGC-3′ and reverse 5′-GCGGCCGCGGCGTTAAATTCGTAGGCGGCGC-3′) were cloned into the Kpn1/Not1 sites of pcDNA3.1 and designated pcDNA3.1-3′-shUTRERαC1. Oligos encoding for shRNA were purchased from Integrated DNA Technologies (Coralville, IA). MCF-7 cells grown to 70% confluency in a 10-mm² dish were transfected with 3 μg pcDNA3.1-3′-shUTRERαC1 and 1 μg pGFP²-C2 (BioSignal Packard, Montréal, Québec, Canada) using Fugene6 transfection reagent (Roche, Indianapolis, IN). Cells were pooled 48 h after transfection, separated, and collected as single positive cells by flow cytometry based on green fluorescent protein (GFP) expression. GFP-expressing single cell isolates were further cultured under G418 selection (1 μg/ml) until confluency in a six-well plate. Microscopic evidence and Western blot analysis revealed no continued expression of ERα-GFP in the surviving resultant colonies. Cells were then assayed for reduced levels of ERα compared with parental MCF-7 cells by Western blot analysis. Two clones, 3′-shUTR 16 and 3′-shUTR 33, demonstrated shRNA-mediated suppression of endogenous ERα with at least 50% reduced ERα protein level as measured by Western blot analysis. These cells were used for subsequent experiments and designated as MCF-7 Low (MCF-7L).

To generate MCF-7 ERα-AB, S118A and S167A stable transfectants, MCF-7L cells were stably transfected with pcDNA3.1 expression plasmids containing either wild-type ERα (pcDNA3.1-ERα-FLAG), ERα mutated at serine 118 to alanine (pcDNA3.1-ERα-FLAG-S118A), or ERα mutated at serine 167 to alanine (pcDNA3.1-ERα-FLAG-S167A), respectively. The FLAG epitope was expressed in frame with the N terminus of ERα for each ERα expression cassette. Briefly, MCF-7L cells were cultured to 80–90% confluency and transfected with 10 μg of either pcDNA3.1-ERα-FLAG, pcDNA3.1-ERα-FLAG-S118A or pcDNA3.1-ERα-FLAG-S167A using FugeneHD transfection reagent (Roche) at a 3:1 DNA to FugeneHD ratio. After 48 h, transfectants were selected by culture in 1 μg/mL G418. Colonies from single cell clones were collected using glass cylinders and cultured to confluency under continuous G418 selection in 6-well plates. Surviving colonies were then subjected to Western blotting with FLAG antibody to screen for the presence of the FLAG epitope on exogenous ERα, and by Western blotting with ERα antibody to identify clones with total ERα level that was comparable to ERα in parental MCF-7 cells.

Immunoprecipitation (IP) and Western blot analysis

To confirm that MCF-7L cells showed a reduction in ERα, total protein was extracted from cells and subjected to Western blotting with an antibody directed against ERα. To confirm that ER-AB cells, S118A cells, and S167A cells were stably transfected with exogenous FLAG-tagged WT ERα, S118A ERα, or S167A ERα, respectively. Protein extracts were first subjected to IP with anti-FLAG antibody followed by Western blotting with an antibody directed against ERα. To assess the relative level of ERα phosphorylation, cell extracts were first subjected to IP using phosphospecific antibodies directed against S118 or S167 followed by Western blotting with an antibody directed against ERα. Protein samples used for IP and Western blot analysis were prepared from cells plated in 100-mm culture dishes in DMEM supplemented with 10% FBS and cultured until 80% confluent. Cells that were to be stimulated with 17β-estradiol were then transferred to phenol-red-free DMEM without serum for 48 h. Cells were then incubated with vehicle (mock stimulated) or 10⁻⁸ m 17β-estradiol for 30 min at 37 C. Whole-cell lysates were prepared using a low-salt lysis buffer [125 mm NaCl, 1m Tris (pH 8), 0.5 m EDTA (pH 8), 10% nonyl phenoxypolyethoxylethanol (Nonidet P-40)] and supplemented with phosphatase inhibitor (1×) and 1× protease inhibitor cocktail (Pierce, Rockford, IL). Collected lysates were incubated on ice for 30 min to ensure efficient lysis of cells. Lysates were centrifuged at 10,000 × g for 5 min, and the protein concentration of the resulting supernatant was determined using a bicinchoninic acid-based colorimetric assay (Bio-Rad, Hercules, CA). Protein samples were prepared for IP using the ReliaBlot IP system according to manufacturer's instructions (Bethyl Laboratories, Montgomery, TX). Briefly, lysates were diluted to 2 mg/ml and 500 μl combined with 10 μg antibody directed against either phospho-ERα S118 (Bethyl Laboratories) or phospho-ERα S167 (Bethyl Laboratories) or 5 μg mouse monoclonal antibody directed against the FLAG epitope (Sigma Chemical Co., St. Louis, MO). The protein-antibody mixture was combined with 10 μl protein A magnetic beads (New England Biolabs, Ipswich, MA) and incubated overnight with rotation at 4 C. After washes with lysis buffer, 1× sample buffer (0.25 m Tris, 10% sodium dodecyl sulfate, 10% glycerol, 1m dithiothreitol, and bromophenol blue) was added to the magnetic beads and boiled for 5 min at 95 C to denature proteins. Samples were loaded onto precast 10% Tris-HCl gels and electrophoresed at 100 V for 90 min. Proteins were then transferred to a nitrocellulose membrane for 60 min at 100 V. Nonspecific binding was blocked by incubating membranes in ReliaBlot Block (Bethyl Laboratories) for 60 min. Membranes were then incubated overnight in primary antibody directed against pan-specific ERα (1:1000; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) diluted in ReliaBlot Block at room temperature (RT). Membranes containing samples prepared from crude protein extracts were incubated with primary antibodies directed against ERα or α-tubulin (1:1000; Millipore, Billerica, MA) diluted in ReliaBlot Block at RT. Membranes were washed in Tris-buffered saline with 1% Tween 20 followed by incubation with horseradish peroxidase-conjugated secondary antibodies diluted in ReliaBlot Block (1:20,000) for 60 min at RT. Specific immunoreactive proteins were detected using West Pico Super Signal (Pierce) chemiluminescence detection reagent. Images were obtained using the Versa Doc imaging system (Bio-Rad Life Sciences, Hercules, CA) and analyzed using Quantity One 1-D image analysis software (Bio-Rad Life Sciences).

Growth analysis

Growth was assessed using the cell proliferation ELISA, bromodeoxyuridine (BrdU) assay kit (Roche) and reduction of 3-(4-5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT; Invitrogen). To measure BrdU incorporation, cells were cultured in DMEM supplemented with 10% FBS at a density of 8.0 × 10³ per well using 96-well culture plates in a 37 C incubator with humidified 5% CO₂. Cells were incubated for 2 h with BrdU using the kit-supplied 1× BrdU labeling agent (Roche). Medium containing BrdU was removed, and cells were fixed and DNA denatured in one step using kit-supplied solution, Fix Denat (Roche), for 30 min. Cells were then incubated with a peroxidase-conjugated secondary antibody directed against BrdU, anti-BrdU (Roche), for 1 h. Cells were washed in kit-supplied 1× washing solution. Cells were incubated with kit-supplied substrate 3,3′,5,5′-tetramethylbenzidine for 1 h at RT. The resulting reaction was read at 405 nm using a scanning multi-well spectrophotometer. To measure MTT reduction, cells were cultured overnight in DMEM supplemented with 10% FBS at a density of 2.0 × 10⁴ in 24=well flat-bottom culture plates in a 37 C incubator with humidified 5% CO₂. Cells were stimulated with 17β-estradiol, tamoxifen, or ICI 182, 780. At 24 and 48 h after stimulation, 125 μl MTT reagent (5 mg/ml) was added to the culture medium, and cells were incubated for 1 h at RT. MTT reagent was removed and replaced with 300 μl dimethylsulfoxide (Sigma), and plates were read at 595 nm. To determine doubling time, cells were cultured in a T25 culture flask at a density of 2.5 × 10⁵ cells per flask. Cells were trypsinized every 48 h and counted using the trypan blue exclusion method. Apoptosis was measured using a well-established annexin-V/propidium (PI) apoptosis staining, followed by flow cytometry. Cells were cultured in a T25 culture flask at a density of 5 × 10³ in DMEM supplemented with 10% FBS and allowed to attach overnight. Cells were then treated with 17β-estradiol, tamoxifen, or ICI 182, 780 for 24 h and stained with annexin-V/PI according to manufacturer's instructions (Sigma).

Colony formation assay

To assess the colony formation abilities of S118A cells and S167A cells, cells were cultured in DMEM supplemented with 10% FBS in triplicate at a density of 5.0 × 10³ in 12-well plastic plates. Cells were cultured for 7 d after which culture medium was removed and cells washed three times with 1 ml PBS. Colonies were stained with 0.005% crystal violet (Invitrogen) in 100% methanol for 30 min. Stained colonies were then gently flushed with distilled H₂O for 15 min or until background was clear. Images of colonies were obtained using the Gel Doc imaging system (Bio-Rad Life Sciences), and colonies were counted using the Quantity One 1-D image analysis software (Bio-Rad Life Sciences).

Immunocytochemistry

Cells were cultured as previously described and then plated at a density of 2.0 × 10⁴ cells per well on an eight-chambered slide (Fisher Scientific, Pittsburgh, PA) and allowed to adhere overnight. Basic morphology of cells was evaluated using Diff-Quik (Siemens, Deerfield, IL) staining according to manufacturer's instructions. For cells that were stained using a fluorescent dye-labeled secondary antibody, culture medium was removed, and cells were washed briefly in ice-cold PBS. Cells were then fixed in 4% formalin for 15 min at RT. Cells were washed with PBS and incubated in blocking buffer (10% goat serum, 0.2% Nonidet P-40) in PBS for 30 min. Cells were washed again and incubated in primary antibody diluted in blocking buffer for 30 min. Primary antibodies directed against β-catenin (1:200; Cell Signaling Technology, Danvers, MA), vimentin (1:250; Fisher Scientific) and pancytokeratin (PanCyto, 1:250; AbCam, Cambridge, MA) were used. Cells were washed and incubated in the appropriate fluorescently labeled secondary antibody for 30 min in the dark. Cell nuclei were fluorescently labeled with 1 mm Hoechst dye. Cells were washed and mounted using Prolong Gold antifade reagent (Invitrogen), and slides were allowed to cure overnight in the dark. Images were obtained using a fluorescent Nikon microscope and microphotographed by Sensicam qe high-performance camera using NIS Elements software.

Migration and invasion assay

Invasion and migration of S118A cells and S167A cells were assessed using BD BioCoat Matrigel invasion chambers and control inserts (BD Biosciences, San Jose, CA) according to manufacturer's instructions. Briefly, Matrigel invasion chambers were hydrated using serum-free DMEM for 2 h at 37 C. Using DMEM supplemented with 10% FBS as a chemoattractant, 2.5 × 10⁴ cells were added to Matrigel invasion chambers and control inserts. Cells were treated with 17β-estradiol, tamoxifen, or ICI 182, 780 and incubated for 16 h in a 37 C incubator with humidified 5% CO₂. Cells that had not invaded or migrated through chambers were removed using gentle scrubbing with cotton-tipped applicators. Membranes were dissected out of chambers, and the remaining attached cells were stained using Diff-Quik. Cells were visualized using a Nikon DXM12007 digital camera attached to a Nikon Eclipse TE200 motorized inverted microscope (Nikon, Melville, NY) driven by Act-1 software (Nikon).

Flow cytometry

Qualitative evaluation of the relative size and complexity of the stably transfected cell lines was carried out using flow cytometric assessment. Cells were cultured in DMEM supplemented with 10% FBS until 80% confluent. Cells were washed in PBS and harvested by trypsinization (0.25% trypsin). Flow cytometric analysis was performed using a fluorescence-activated cell sorter (FACS) flow cytometer (BD Biosciences) using 488-nm laser excitation and fluorescence emission at 530 nm and over 575 nm. Forward- and side-scatter measurements were made using linear amplification. Approximately 10,000 events (cells) were evaluated for each sample.

RNA extraction and quantitative real-time RT-PCR

Cells were treated with 17β-estradiol or vehicle for 2, 6, 12, 18, or 24 h. Extraction of total RNA was performed using an RNeasy Mini Kit (QIAGEN, Valencia, CA). Briefly, cells were lysed directly by aspirating cell culture media and adding 600 μl of the kit-supplied lysis buffer to each well. Lysis buffer was supplemented with β-mercaptoethanol (10 μl/ ml) to inactivate ribonucleases (RNase). Lysate was mixed thoroughly by repeated pipetting. To ensure binding of isolated RNA to the spin column membrane, 600 μl 70% ethanol in RNase-free water was added to the lysate and transferred to an RNeasy spin column. Spin columns were centrifuged for 15 sec at 8,000 × g. Column-bound total RNA was washed using buffers supplied in the kit. Total RNA was eluted using 14 μl RNase-free water. Concentration and purity of extracted RNA were determined using a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies Inc., Wilmington, DE). Transcript abundance was measured using i-Script One-Step PCR Kit for probes (Bio-Rad). Previously designed and validated primers and probes were used to detect transcript abundance of target genes. A final concentration of 900 nm for pS2 forward (5′-CGTGAAAGACAGAATTGTGGTTTT-3′) and reverse (5′-CGTCGAAACAGCAGCCCTTA-3′) primers and 250 nm for pS2 probe (5′/56-FAM/TGTCACGCCCTCCCAGTGTGC A/3BHQ_1/-3′), 50 nm for cMyc forward (5′-CGTCTCCACACATCAGCACAA-3′) and reverse (5′-TCTTGGCAGCAGGATAGTCCTT-3′) primers and 25 nm for cMyc probe (5′/56-FAM/ACGCAGCGCCTCCCTCCACTC/3BHQ_1/-3′), and 22.5 nm for cyclin D1 forward (5′-TGGGTCTGTGCATTTCTGGTT-3′) and reverse (5′-GCTGGAAACATGCCGGTTAC-3′) primers and 6.25 nm for cyclin D1 probe (5′/56-FAM/CGGCGCTTCCCAGCACCA A/3BHQ_1/-3′) was used. Briefly, 25 ng RNA was combined with 12.5 μl 2× RT-PCR mix for probes, 1 μl of the appropriate forward and reverse primer, 1 μl of appropriate probe, 0.5 μl of iScript reverse transcriptase and nuclease water for a final volume of 25 μl per reaction. Conditions for cDNA synthesis were 50 C for 10 min, followed by inactivation of reverse transcriptase at 95 C for 5 min; real-time PCR was cycled 45 times and performed at 95 C for 15 min with data collection performed at 60 C for 1 min. All quantitative RT-PCR reactions were performed in an iQ5 multicolor, real-time PCR detection system (Bio-Rad). Cycle threshold (Ct) values of the target gene were normalized (ΔCt) using GAPDH (forward, 5′-CAACGGATTTGGTGGTATTGG-3′; reverse, 5′-GCAACAATATCCACTTTACCAGAGTT-3′; and probe, 5′/56-FAM/CGCCTGGTCACCAGGGCTGCT/3BHQ_1/-3′).

PCR array

Previously extracted RNA from cells stimulated with 10⁻⁸ m 17β-estradiol was used to hybridize to a human estrogen receptor signaling PCR array (SABiosciences, Frederick, MD) targeting genes related to breast cancer regulation and estrogen receptor-dependent signaling. Briefly, 25 ng RNA was reverse transcribed using RT² first-strand kit (SABiosciences) according to manufacturer's instruction. Nascent cDNA was combined with 2× SABiosciences RT² quantitative PCR (qPCR) Master Mix according to manufacturer's instruction and pipetted onto the supplied gene array PCR plate. Briefly, 102 μl cDNA was combined with 1350 μl 2× RT² SYBR Green Mastermix and nuclease water for a final volume of 2700 μl. Conditions for the gene array hybridization were 95 C for 10 min for activation of HotStart DNA polymerase, followed by real-time PCR that was cycled 40 times at 95 C for 15 sec with data collection performed at 60 C for 1 min.

Data analysis

All data are expressed as mean ± sd. All statistical analyses were performed using GraphPad Prism version 5 (GraphPad, Inc., San Diego, CA). P values were calculated using ANOVA followed by the Tukey's multiple-comparison test. P < 0.05 was considered significant.

Results

Generation of ERα phosphorylation mutant cell lines with attenuated phosphorylation at serine 118 and serine 167

Preliminary data showed that ERα-negative HeLa cells when transiently transfected with ERα mutated at serine 167 to alanine exhibited increased growth relative to HeLa cells transiently transfected with wild-type ERα (Supplemental Fig. 1, published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org). To determine whether these effects would be reproduced in a cell line that expressed endogenous ERα with a functional estradiol signaling pathway, ERα-positive MCF-7 breast cancer cells (MCF-7) were stably transfected with shRNA that targeted the 3′-UTR of ERα to reduce endogenous ERα. Endogenous ERα expression in MCF-7 cells was reduced by more than 50%. These cells were designated as MCF-7-Low (designated MCF-7L cells). The shRNA used to reduce endogenously expressed ERα did not affect exogenous plasmid-derived ERα that lacked the 3′-UTR of endogenous ERα. Reduced expression of ERα in the MCF-7L cells was confirmed by Western blots using crude extracts from the cell lines (Fig. 1A, lanes 1 and 2). The MCF-7L cells were stably transfected with ERα expression plasmids for wild-type ERα (designated ER-add back or ER-AB cells), point mutation in serine 118 to alanine (designated S118A cells), or point mutation in serine 167 to alanine (designated S167A cells). Mutation of S118 or S167 to alanine did not significantly alter the total ERα protein levels in comparison with MCF-7 parental cells (Fig. 1A, lanes 3–5). A FLAG epitope was engineered in frame with the N terminus of ERα to distinguish the transfected ERα from the endogenous ERα. Only cell clones showing a positive Western blot signal at the ERα molecular mass (∼67 kDA) after IP with FLAG antibody were used in subsequent experiments (Fig. 1B). It was noted that in the S118A cells and the S167A cells, some of the ERα protein detected by Western blot could be the endogenous ERα that did not contain mutations at S118 and S167 and was therefore competent to be phosphorylated at these sites. The remaining ERα resulted from the stable expression of WT ERα, S118A, or S167A expression plasmids.

Fig. 1.

Generation and validation of MCF-7L, ER-AB, S118A, and S167A cells. A, MCF-7 parental cells were stably transfected with shRNA directed against the 3′-UTR of endogenous ERα. MCF-7L cells were then stably transfected with ERα expression plasmids containing wild-type ERα-FLAG (ER-AB), ERα-FLAG mutated at serine 118 to alanine (S118A), or ERα-FLAG mutated at serine 167 to alanine (S167A). Expression of ERα was confirmed in crude cellular extracts by Western blotting (WB) with an antibody directed against pan-ERα (ERα). α-Tubulin was used as a loading control. B, To confirm stable transfection, extracts of MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells were subjected to IP with an antibody directed against the FLAG epitope tag followed by Western blotting with antibody against pan-ERα (ERα). C, To evaluate ERα phosphorylation, cells were cultured to 80% confluency in DMEM supplemented with 10% FBS and then transferred to serum-free phenol-red-free DMEM followed by incubation with or without 10⁻⁸ m 17β-estradiol for 30 min. Cell extracts were prepared for IP using antibodies directed against ERα phosphorylated at S118 (p118) or ERα phosphorylated at S167 (p167) followed by Western blotting with an antibody directed against pan-ERα (ERα). Data are representative from one of at least three independent experiments.

Previous reports have shown that S118 and S167 were rapidly phosphorylated in response to estradiol (9). To evaluate levels of ERα phosphorylation in the S118A cells or S167A cells, cells were incubated with or without estradiol for 30 min, and ERα was immunoprecipitated with phosphospecific antibodies to S118 or S167. Western blotting was then performed with an antibody directed against total ERα. Because of the low reactivity of phosphospecific antibodies for ERα, Western blots using crude cellular extracts did not result in strong enough specific signal above background (data not shown) to visualize distinct bands on the blots. We found that the approach of using IP with the phosphospecific antibody followed by Western blot with total ERα antibody resulted in a robust and reproducible signal by Western blot analysis (22). In agreement with previously published studies (15, 21), MCF-7 parental cells incubated with estradiol exhibited increased phosphorylation at S118 and S167 (Fig. 1C, lanes 1 and 2). ER-AB cells exhibited similar estradiol-induced phosphorylation of S167 as MCF-7 parental cells; however, estradiol-induced phosphorylation of S118 was compromised in ER-AB cells (Fig. 1C, lanes 3 and 4). As expected, basal and estradiol-stimulated phosphorylation of S118 was reduced in S118A cells. Interestingly, basal phosphorylation of S167 was modestly increased in the S118A cells, and there was a suppression of S167 phosphorylation after estradiol treatment (Fig. 1C, lanes 5 and 6). Unexpectedly, in the S167A cells, basal S167 phosphorylation was modestly increased compared with parental MCF-7 cells, suggesting that the endogenous ERα may have been hyperphosphorylated at S167. Estradiol treatment resulted in marked reduction in S167 phosphorylation in the S167A cells. Basal and estradiol-stimulated phosphorylation of S118 in the S167A cells were similar to parental MCF-7 cells (Fig. 1C, lanes 7 and 8). These data indicated that ER-AB cells may have exhibited a disruption in estradiol-induced S118 phosphorylation. S118A cells showed disruption in estradiol-induced S118 and S167 phosphorylation, and S167A cells showed a disruption of estradiol-induced S167 phosphorylation. These data suggested a possible link or dependency between the phosphorylation at these two sites.

Attenuated phosphorylation of ERα at S118 and S167 resulted in increased cell growth

In comparison with MCF-7 parental cells, MCF-7L cells, S118A cells, and S167A cells exhibited increased growth as assessed by MTT assay, whereas growth of ER-AB cells was not significantly different from MCF-7 parental cells (Fig. 2A). Synthesis of DNA as measured by BrdU incorporation showed that only MCF-7L cells had significantly higher incorporation of BrdU; ER-AB cells, S118A cells, and S167A cells were not significantly different from MCF-7 parental cells (Fig. 2B). To further evaluate the effect of attenuated phosphorylation at S118 and S167 on cell growth, cell doubling time was measured. There was no significant difference in cell number for any of the cell lines at d 2 and 4 (Fig. 2C). At d 6, S167A cells and MCF-7L cells were significantly higher in number in comparison with MCF-7 parental cells. There was no significant difference in cell number when comparing S118A cells and ER-AB cells with MCF-7 parental cells. A further assessment of growth was performed using colony-formation assays. In comparison with MCF-7 parental cells, there was no difference in colony formation for MCF-7L cells (Fig. 2D). ER-AB cells formed significantly fewer colonies than MCF-7 parental cells, whereas S118A cells and S167A cells formed significantly more colonies compared with MCF-7 parental cells. Taken together, these data demonstrated increased growth and/or colony formation of the S118A cells and S167A cells as well as the MCF-7L cells in comparison with MCF-7 parental cells and ER-AB cells.

Fig. 2.

Growth assessment of S118A cells and S167A cells. A, Growth analysis of MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells by MTT assay. Cells were cultured in DMEM supplemented with 10% FBS for 48 h. Bars represent mean ± sd. **, P ≤ 0.05 compared with MCF-7 parental cells. B, Proliferation of cells was measured by incorporation of BrdU. Data are reported as mean ± sd of three independent experiments. **, P ≤ 0.05 compared with MCF-7 parental cells. C, MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells were cultured in DMEM supplemented with 10% FBS at a density of 2.5 × 10⁵ per culture flask. Cell number was determined after 2, 4, or 6 d in culture. Data presented represent mean ± sd of three independent experiments. **, P ≤ 0.05 compared with MCF-7 parental cells. D, Colony formation units (CFU) was determined after 7 d culture in DMEM supplemented with 10% FBS. Colonies were stained with crystal violet and counted. Data are representative of three or more independent experiments yielding similar results. **, P ≤ 0.05 compared with MCF-7 parental cells.

S118A cells and S167A cells retained sensitivity to estradiol and antagonists but exhibited reduced apoptosis

In subsequent experiments, the ER-AB cell line, and not MCF-7L, was used as the control for comparison with S118A cells and S167A cells because the ER-AB cells expressed similar levels of total ERα that was capable of being phosphorylated at S118 and S167. In comparison with MCF-7 parental cells, there were no significant differences in growth rates in response to 17β-estradiol, tamoxifen, or ICI 182,780 for ER-AB cells, S118A cells, or S167A cells, indicating that the S118A and S167A cells still retained sensitivity to ligands. Estradiol resulted in modest stimulation of growth for all cell lines (Fig. 3A). All cell lines exhibited comparable growth inhibition by tamoxifen and ICI 182,780 (Fig. 3, B and C). Apoptosis was assessed to determine whether changes in the relative level of apoptosis could contribute to the overall elevated growth potential of S118A cells and S167A cells in the absence of added ligands. Both S118A cells and S167A cells exhibited significantly reduced apoptosis in medium without added ligands (Fig. 3D). There was no significant difference in apoptosis for all cell lines with the addition of estradiol or tamoxifen (Fig. 3, E and F), although ER-AB and S167A cells exhibited reduced apoptosis in response to ICI 182,780 (Fig. 3G). These data suggest that the increased growth of the S118A and S167A cells in the absence of added ligands was due, in part, to reduced apoptosis.

Fig. 3.

Response of S118 cells and S167A cells to estradiol and ER antagonists tamoxifen and ICI 182,780. Effects of estradiol and antiestrogens on growth were evaluated by MTT assay. Cell lines were cultured in DMEM supplemented with 10% FBS and 17β-estradiol (E2, 10⁻⁸ m) (A), tamoxifen (TAM, 10⁻⁷ m) (B), or ICI 182,780 (ICI, 10⁻⁸ m) (C) for 24, 48, 72, or 96 h. To evaluate the rate of apoptosis in S118A cells and S167A cells, cell lines were cultured in DMEM supplemented with 10% FBS at a density of 5.0 × 10³ for 48 h. Cells were incubated with vehicle (Veh) (D), 17β-estradiol (10⁻⁸ m) (E), tamoxifen (10⁻⁷ m) (F), or ICI 182,780 (10⁻⁸ m) (G). Cells were then labeled with annexin-V/PI and analyzed using flow cytometry. Data presented represent three independent experiments yielding similar results. **, P ≤ 0.05 compared with MCF-7 parental cells.

S118A cells and S167A cells retained epithelial lineage cell markers

Light microscopy of S118A cells and S167A cells revealed cells growing in smaller clusters compared with MCF-7 parental cells (Fig. 4). Additionally, cell morphology, as assessed with Diff-Quik staining, revealed that S118A cells and S167A cells exhibited cytoplasmic extensions suggestive of migratory cells (Fig. 4, arrows) that were not evident in parental MCF-7 parental cells or ER-AB cells. These observations coupled with the increased growth observed in S118A cells and S167A cells suggested possible alterations in the epithelial-like nature of these cell lines. To test the relative epithelial and mesenchymal nature of the cells, cell lines were stained with epithelial lineage marker pancytokeratin and mesenchymal lineage marker vimentin and evaluated for the subcellular localization of β-catenin. Both S118A and S167A exhibited positive staining for pancytokeratin, indicating that the cell lines retained this epithelial lineage marker. Additionally, S118A cells and S167A cells showed β-catenin staining in the membrane, indicating that β-catenin was present in adherence junctions but not in the nucleus, another feature of epithelial-lineage cells (23). The mesenchymal marker vimentin was not detected in any of the cell lines. MDA-MB-231 breast cancer cells that exhibit a mesenchymal phenotype (24) were used for comparative purposes for morphology and staining with pancytokeratin, β-catenin, and vimentin. Taken together, the data presented in this study indicated that phosphorylation at S118 and S167 was necessary for a fully functional ERα signaling pathway in MCF-7 breast cancer cells. Furthermore, phosphorylation at S118 and S167 was associated with epithelial-like features of MCF-7 cells. However, the S118A cells and the S167A cells exhibited reduced clustering and cytoplasmic extensions that would indicate possible increased migration and invasive capacity.

Fig. 4.

Basic morphology and assessment of cell lineage markers for S118A cells and S167A cells. MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells were cultured in DMEM supplemented with 10% FBS at 2.0 × 10⁴ cells per well on eight-chambered slides. Cells were stained with Diff-Quik to evaluate basic morphology of the cells. Cells were subjected to immunocytochemistry with antibodies directed against pancytokeratin (PanCyto), β-catenin (βcat), or vimentin. Arrows indicate cytoplasmic extensions.

S118A cells and S167A cells exhibited increased migration and invasion

To investigate the effects of attenuated phosphorylation of ERα at S118 and S167 on migration and invasion potential, cell lines were cultured in Boyden chambers for 16 h using inserts coated without or with Matrigel to measure migration or invasion, respectively, in the presence of 17β-estradiol, tamoxifen, or ICI 182,780. In the absence of added hormones (vehicle), S118A cells, S167A cells, and ER-AB cells exhibited increased migration in comparison with MCF-7 parental cells (Fig. 5A). S118A cells and S167A cells exhibited increased migration when incubated with estradiol, tamoxifen, or ICI 182,780 in comparison with MCF-7 parental cells. Furthermore, S118A cells and S167A cells exhibited increased migration when incubated with estradiol in comparison with ER-AB cells (Fig. 5A). In the absence of added hormones (vehicle), ER-AB cells, S118A cells, and S167A cells exhibited increased invasion in comparison with MCF-7 parental cells (Fig. 5B), and S118A cells and S167A cells exhibited increased invasion in comparison with ER-AB cells (Fig. 5B). ER-AB cells, S118A cells, and S167A cells exhibited increased invasion in response to estradiol, tamoxifen, and ICI 182,780 in comparison with MCF-7 parental cells. Additionally, S167A cells incubated with estradiol and S118A cells incubated with tamoxifen or ICI 182,780 exhibited significantly increased invasion in comparison with ER-AB cells (Fig. 5B). Interestingly, S167A cells incubated with ICI 182,780 exhibited significantly less invasion in comparison with ER-AB cells (Fig. 5B). These data demonstrated the increased migration and invasion of the S118A cells and S167A cells in comparison with MCF-7 parental cells both in the absence and presence of estradiol. Remarkably, the cell lines were partially insensitive to antiestrogens to suppress migration/invasion, even though the cells remained sensitive to tamoxifen and ICI 182,780 growth inhibition (see Fig. 3, A–C). Notably, the ER-AB cell also exhibited increased migration and invasion compared with MCF-7 parental cells, although not to the degree of migration of S167A cells or invasion of S118A cells and S167A cells. It was noted in Fig. 1 that there was disruption in ligand-induced S118 phosphorylation in the ER-AB cells, and this may be related to the partial increase in migration and invasion observed by ER-AB cells in these assays.

Fig. 5.

Migration and invasion of S118A cells and S167A cells in response to estradiol and ER antagonists. A, Migration of MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells was determined after incubation of cells with vehicle (Veh), 17β-estradiol (E2, 10⁻⁸ m), tamoxifen (TAM, 10⁻⁷ m), or ICI 182,780 (ICI, 10⁻⁸ m) in uncoated Boyden chambers using DMEM supplemented with 10% FBS as a chemoattractant. **, P ≤ 0.05 compared with MCF-7 parental cells; #, P ≤ 0.05 compared with ER-AB cells. B, Invasion of MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells was determined by incubating cells with vehicle, 17β-estradiol (10⁻⁸ m), tamoxifen (10⁻⁷ m), or ICI 182,780 (10⁻⁸ m) in Matrigel-coated Boyden chambers using DMEM supplemented with 10% FBS as the chemoattractant. Data are representative of three or more independent experiments yielding similar results. **, P ≤ 0.05 compared with MCF-7 parental cells; #, P ≤ 0.05 compared with ER-AB.

Forward- and side-scatter flow cytometry demonstrated altered cell size and cellular complexity of S118A cells and S167A cells

To determine the effect of attenuated phosphorylation of ERα at S118 and S167 on cellular size and complexity, forward-scatter flow cytometry (to approximate cell size) and side-scatter flow cytometry (to approximate cell granularity and complexity, e.g. organelles) was performed. Forward-scatter flow cytometry revealed that S167A cells were relatively smaller than MCF-7 parental cells; there was no significant difference in size between MCF-7 parental cells, ER-AB cells, or S118A cells (Fig. 6A). However, side-scatter flow cytometry revealed that S118A cells and S167A cells both exhibited relatively less cellular complexity compared with MCF-7 parental cells; there was no difference in complexity between MCF-7 parental cells and ER-AB cells (Fig. 6B). These data indicated that attenuated phosphorylation of ERα at S118 and S167 resulted in markedly reduced cellular size and complexity compared with MCF-7 parental cells.

Fig. 6.

Forward- and side-scatter flow cytometry to measure cell size/complexity of S118A cells and S167A cells. Cell lines were cultured in DMEM supplemented with 10% FBS until 80% confluent. Cells were qualitatively analyzed using flow cytometry for relative comparison of size (forward scatter) (A) and complexity (side scatter) (B). Data are representative of at least three independent experiments yielding similar results. Bars represent mean ± sd. **, P ≤ 0.05 compared with MCF-7 parental cells.

Attenuated phosphorylation of ERα at S118 and S167 resulted in altered ERα-regulated gene expression

To determine whether S118A cells or S167A cells exhibited altered ERα transcriptional activity, the expression of three well-characterized ERα-regulated genes, pS2, cyclin D1, and cMyc, were examined. These genes represent three different paradigms for ERα interaction with promoters for regulation of gene transcription. pS2 is an example of direct binding of ERα to a canonical ERE; cyclin D1 represents the paradigm of indirect interaction of ERα with promoters via binding to Jun/ATF2 transcriptional heterodimers at cAMP response elements on DNA; cMyc is another example of indirect interaction of ERα with promoters via binding to Jun/Fos transcriptional heterodimers at AP-1 sites on DNA (21). In MCF-7 parental cells, pS2 mRNA peaked after 6 h incubation with estradiol and returned to baseline expression by 12 h (Fig. 7A). Similar pS2 mRNA levels were observed for ER-AB cells and S118A cells. However, estradiol induction of pS2 mRNA was completely absent in S167A cells. Estradiol induced cyclin D1 expression 5-fold in MCF-7 parental cells that peaked at 12 h (Fig. 7B). ER-AB cells exhibited a similar temporal cyclin D1 expression pattern as MCF-7 parental cells, although peak expression level was slightly reduced. Both S118A cells and S167A cells exhibited a shift toward earlier peak induction and a lower overall fold induction of cyclin D1. In S118A cells and S167A cells, peak cyclin D1 induction was 3-fold and occurred at 6 h with a return to baseline expression at 12 h, the time point where peak cyclin D1 mRNA levels were observed in MCF-7 parental cells and ER-AB cells. MCF-7 parental cells showed variable mRNA induction of cMyc over a 24-h time period (Fig. 7C). Expression of cMyc in ER-AB cells was similar to that of MCF-7 parental cells. In comparison with MCF-7 parental cells, S118A cells exhibited a modest compromise in cMyc expression, and S167A cells showed similar or greater levels of cMyc mRNA after 12 h incubation with estradiol. In summary, the absence of estradiol induction of pS2 in S167A cells suggested that phosphorylation of S167 may be required for transactivation at ERE-driven promoters. Additionally, attenuated phosphorylation of ERα in S118A cells and S167A cells resulted in altered expression of genes that are regulated when ERα interacts with promoters through indirect binding.

Fig. 7.

Estradiol-dependent transcription in S118A cells and S167A cells. MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells were cultured until 80% confluent in DMEM supplemented with 10% FBS. Cells were transferred to serum-free phenol-red-free DMEM for 48 h followed by incubation with or without 10⁻⁸ m 17β-estradiol for 2, 6, 1-, 18, or 24 h, and then total RNA was extracted. Gene expression of pS2 (A), cyclin D1 (B), and cMyc (C) was measured using real-time RT-PCR normalized to GAPDH. Data presented are representative of at least three independent experiments yielding similar results.

To further assess the consequences of attenuated phosphorylation of ERα at S118 and S167, a real-time qPCR array was evaluated for altered expression of genes known to be involved in estrogen-mediated signaling and in breast cancer. Cell lines were incubated with estradiol for 6 h, and gene expression was evaluated. Fold differences in gene expression were expressed relative to gene expression in MCF-7 parental cells using a 2-fold difference in expression as the threshold for reporting gene induction or gene repression. Genes were grouped based on a role in either ERα signaling or general involvement in breast cancer. The qPCR array compared gene expression among the cell lines that were treated with estradiol for 6 h; the qPCR array did not compare estradiol treatment with vehicle. In total, expression of 84 genes was evaluated (Supplemental Fig. 2). In ER-AB cells, four genes showed increased expression and four genes showed reduced expression, with the remaining 76 genes showing no significant changes in expression compared with MCF-7 parental cells. The four genes down-regulated in ER-AB cells were involved in ERα signaling, whereas three of the genes induced were associated with breast cancer and correlated with breast cancer prognosis; the remaining induced gene was involved in ERα signaling. In S118A cells, 22 genes were down-regulated, and there was no gene induction for any of the 84 genes on the qPCR array (Table 1). Of the 22 genes down-regulated in S118A cells, three were involved in ERα signaling with the remaining 19 genes associated with breast cancer. The majority of these 19 genes (12 of 19) were genes correlated with breast cancer prognosis. In addition to the 22 genes that were down-regulated 2-fold or greater in S118A cells, an additional 22 genes were completely undetectable in the assay. A combined total of 44 of 84 genes were either down-regulated or undetectable in S118A cells compared with parental MCF-7 cells. Two of the genes that were undetectable in S118A cells were involved in ERα signaling, and the remaining genes were associated with breast cancer with the majority of these genes (15 of 22 genes) related to the prognosis of breast cancer patients. In S167A cells, three genes were down-regulated and one gene was induced compared with MCF-7 parental cells (Table 1). Two of the three down-regulated genes in S167A cells were associated with breast cancer, whereas the remaining gene was involved in ERα signaling. Expression of nine genes was completely undetectable in S167A cells. A combined total of 12 of 84 genes were either down-regulated or undetectable in S167A cells compared with parental MCF-7 cells. Eight of the undetectable genes in S167A cells were involved in breast cancer, and one gene was involved in ERα signaling. Interestingly, one gene involved in ERα signaling was induced in S167A cells. Taken together, these data demonstrate that attenuated phosphorylation of S118 and S167 markedly comprised estradiol regulation of a panel of ERα-regulated gene expression and breast cancer-related genes in MCF-7 cells with the majority of changes being down-regulation or absence of expression.

Table 1.

ERα-mediated gene expression in MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells

Gene name	Fold change relative to MCF-7 parental cells
	ER-AB	S118A	S167A
B-cell CLL/lymphoma 2	2.73	−2.88	1.79
BCL2-like 2	0.53	ND	0.53
Cyclin A2	1.31	ND	1.16
Cyclin E1	1.19	ND	0.93
Cyclin-dependent kinase inhibitor 1A	4.64	−4.40	−2.62
Cyclin-dependent kinase inhibitor 2A	1.25	ND	1.09
Claudin 7	1.40	−0.35	1.28
Catenin, β1, 88 kDa	1.09	0.48	1.12
Cathepsin B	0.95	ND	1.03
Cytochrome P450, family 19, subfamily A, polypeptide 1	1.63	ND	ND
V-erb-b2 erythroblastic leukemia viral oncogene homolog 2, neuro/glioblastoma-derived oncogene homolog (avian)	0.84	ND	0.76
Estrogen receptor 1	1.45	−0.37	1.14
Fas	0.92	ND	1.15
Fibronectin leucine-rich transmembrane protein 1	1.25	0.38	1.09
GATA-binding protein 3	0.69	ND	0.95
Inhibitor of DNA binding 2, dominant-negative helix-loop-helix protein	2.45	1.01	1.57
IGF-binding protein 2, 36 kDa	0.67	ND	1.02
IL-2 receptor, α	1.31	ND	1.30
IL-6	1.23	0.44	ND
IL-6 receptor	1.06	0.26	0.97
IL-6 signal transducer	0.93	0.42	0.91
Integrin, α6	1.17	0.45	ND
Integrin, β4	1.30	0.41	1.18
Jun proto-oncogene	0.87	ND	0.77
V-kit Hardy-Zuckerman 4 feline sarcoma viral oncogene homolog	1.14	0.42	1.06
Kruppel-like factor 5 (intestinal)	1.09	0.13	1.03
Kallikrein-related peptidase 5	1.14	ND	0.91
Keratin 19	1.19	−0.02	1.30
MAPK kinase 7	0.91	0.35	ND
Antigen identified by monoclonal antibody Ki-67	1.29	ND	1.26
Nuclear transcription factor Y, β	1.42	ND	0.95
Nerve growth factor receptor	1.07	ND	0.93
Nonmetastatic cells 1, protein (NM23A) expressed in	1.08	0.28	0.82
Progesterone receptor	5.24	1.84	3.24
Plasminogen activator, urokinase	1.32	0.10	1.65
Prostaglandin-endoperoxide synthase 2	1.41	ND	ND
Serpin peptidase inhibitor, clade A	1.24	ND	1.16
Serpin peptidase inhibitor, clade B (ovalbumin), member 5	1.09	0.55	ND
Serpin peptidase inhibitor, clade E	1.37	ND	ND
Solute carrier family 7 (amino acid transporter light chain, L system), member 5	0.37	0.10	ND
Small proline-rich protein 1B	1.53	−0.05	1.23
Stanniocalcin 2	0.35	ND	0.40
Trefoil factor 1	0.41	1.05	0.69
TGFα	0.79	ND	0.63
Tyrosine kinase with Ig-like and EGF-like domains 1	0.95	0.31	1.10
TNFα-induced protein 2	0.79	0.41	ND
Tumor protein p53	0.81	0.27	0.76
Vascular endothelial growth factor A	0.38	ND	0.25

MCF-7 parental cells, ER-AB cells, S118A cells, and S167A cells were incubated with estradiol for 6 h. RNA was reverse transcribed and applied to an 84-gene qPCR array containing genes involved in estrogen signaling or in breast cancer. Gene expression was normalized using GAPDH. Gene expression fold change after estradiol treatment is presented relative to gene expression in MCF-7 parental cells (set equal to 1.00). Only genes that exhibited a 2-fold or greater change in expression relative to MCF-7 parental cells are presented in this table (shown in bold). ND, Not detected.

Discussion

The ERα status of breast tumors is used to predict response to endocrine therapy and to gauge overall prognosis in breast cancer patients. However, half of all ERα-positive tumors are resistant to endocrine therapy, indicating the need for more stringent diagnostic and prognostic tools. S118 and S167 are two of the most well-characterized ERα phosphorylation sites and have been shown to correlate with overall positive prognosis, higher survival rates, and increased relapse-free survival as well as being associated with a functional estrogen signaling pathway (5, 6). In the current study, ERα in MCF-7 breast cancer cells was reduced over 50%, and then stable cell lines were created by transfection with WT ERα (ER-AB cells) or with ERα containing a serine to alanine mutation at S118 or S167 to generate S118A cells or S167A cells, respectively. Growth analysis of S118A cells and S167A cells demonstrated increased growth and colony formation in comparison with parental MCF-7 cells. S118A and S167A cells maintained sensitivity to estradiol and antiestrogens similar to that of MCF-7 parental cells but exhibiting reduced rates of apoptosis. Although S118A and S167A exhibited increased migration and invasion in comparison with MCF-7 parental cells, both cell lines retained expression of epithelial cell surface markers. Finally, expression of three known ERα target genes and a qPCR array of 84 genes revealed markedly compromised estradiol-regulated transcription in S118A cells and S167A cells. These findings demonstrate that changes in a single phosphorylation site in ERα had profound changes on MCF-7 breast cancer cell behavior, morphology, and gene expression.

Most published studies investigating ERα phosphorylation have used transient transfection of ERα phosphorylation-site mutants that may not accurately recapitulate endogenous ERα expression in cells and tissues. A search of the literature revealed that studies using ERα phosphorylation mutants transiently expressed in cells containing endogenous ERα were limited to evaluation of acute transcriptional responses. To study ERα phosphorylation in a physiological context that would be more closely related to the in vivo microenvironment of breast cancer, MCF-7 breast cancer cells that expressed endogenous ERα were used. These cells have evolved to be estradiol dependent for growth and to have a functional ERα signaling pathway. To overcome the deficiencies of a transient transfection system or transfection into ERα-negative cells, ERα phosphorylation mutants S118A and S167A were stably expressed in MCF-7 cells in which endogenous ERα was suppressed over 50%. It is important to note that a complete ablation (100%) of ERα was not feasible because MCF-7 cells are dependent on ERα signaling for growth (25), and complete ablation of ERα might not yield viable clones. In the present study, it was likely that some of the expressed ERα in ER-AB, S118A, and S167A cells was derived from endogenous ERα expression that would be fully competent to be phosphorylated at S118 and S167. Hence, in the stable cell lines derived for this study, the phosphorylation at S118 and S167 was attenuated but not completely lost, a feature that is more reminiscent of changes in ERα phosphorylation that might occur in vivo. It may have been possible to culture MCF-7 cells completely devoid of ERα in estrogen-depleted medium, eventually rendering the cells estrogen independent; however, this may have led to estrogen hypersensitivity, which would further complicate the interpretations of ERα phosphorylation in cell physiological responses. Interestingly, the ER-AB cells did not exhibit estradiol-induced phosphorylation of S118, similar to what was observed in the S118 cells. It is unknown why the ER-AB cells that express only WT ER at comparable levels to MCF-7 parental cells did not exhibit estradiol-induced phosphorylation of S118. This represents a potential limitation to the experimental approach of depleting endogenous protein and stably expressing mutations in the same protein.

Because S118 and S167 are major ERα phosphorylation sites in MCF-7 cells, and these sites may serve as surrogate markers for a functional ERα, it was possible that mutating S118 and S167 to the nonphosphorylatable alanine would result in loss of ERα signaling and possibly loss of viability in MCF-7 cells that had evolved to depend on a functional estrogen signaling pathway. In previous reports, ERα-positive MCF-7 cells, ERα-negative COS-1 cells, and ERα-negative MDA-MB-231 cells transiently transfected with ERα-S118A or ERα-S167A did not undergo cell death in the short term, and these single point mutations had only modest effects on cell behavior and physiology (15, 26). In the present study using stable transfection, S118A and S167A cells exhibited increased growth in comparison with MCF-7 parental cells and ER-AB cells. Proliferation of estrogen-dependent breast cancer cells is regulated during the early G1 phase of the cell cycle. Ligand-bound ERα up-regulates cell-cycle-specific cyclins, which ultimately stimulates cells to progress through the G1/S checkpoint of the cell cycle. Although neither S118A cells nor S167A cells exhibited increased BrdU incorporation, both cell lines exhibited reduced rates of apoptosis. These results suggested that although the S118A and S167A cells were not proliferating at a higher rate than the MCF-7 parental cells or the ER-AB cells, the reduced apoptosis likely contributed to the overall increased cell growth. It is plausible that disruption of ERα signaling by attenuated phosphorylation of S118 and S167 resulted in compensation by other growth-promoting pathways such as epidermal growth factor receptor (EGFR) or IGF-I signaling pathways, both of which have been shown to promote cell survival and proliferation (25, 27). Activation of these pathways was shown to result in phosphorylation of not only ERα but also ERα coregulators (27). Treatment of cells with estradiol increased EGFR and IGF-I ligands, resulting in activation of these pathways (27) that, in turn, have been shown to down-regulate ERα via activation of the AKT and MAPK pathways, ultimately relieving cells of estrogen dependency (25).

Recruitment of RNA polymerase II to active promoters by liganded ERα simultaneously signals for the ubiquitination of ERα and degradation of the receptor (26). Susceptibility of ERα to ligand-induced degradation has been linked to phosphorylation of the receptor. Specifically, loss of S118 phosphorylation increased ERα stability primarily through diminished recruitment of ubiquitin ligases (26, 28, 29). Interestingly, S167 phosphorylation did not appear to affect ERα stability (26). Furthermore, ERα-S118A had reduced affinity for ligand but maintained affinity for coactivator interaction, suggesting that these cells were dependent on ligand-independent ERα signaling. Cells transiently expressing ERα-S118A exhibited modest decreases in transactivation despite the increased stability of ERα (26). In the present study using stable transfection, neither the S118A cells nor S167A cells showed significant alteration in overall levels of ERα despite these cell lines exhibiting marked reductions in ERα-regulated gene transcription. Additional studies will be needed to measure protein stability and turnover to ascertain the effect of stable changes in phosphorylation on ERα protein levels.

The presence of cytoplasmic extensions and the increased migration and invasion of S118A and S167A cells suggested these cells may have acquired some features of an epithelial to mesenchymal transition. Gadalla et al. (30) reported that silencing of ERα in MCF-7 cells led to the development of a mesenchymal-like phenotype characterized by loss of epithelial cellular organization. Moreover, ERα-positive ZR-75-1 cells transfected with siRNA directed against ERα grew in a more dispersed pattern and exhibited cytoplasmic extensions (30). In the present study, it was remarkable that the loss of S118 or S167 phosphorylation with the maintenance of total ERα levels equivalent to the parental MCF-7 cells resulted in cell lines that also exhibited cytoplasmic extensions, reduced cell clustering, and increased migration/invasion, reminiscent of the MCF-7 cells with silenced ERα as reported by Gadalla et al. (30). This suggested that reduced phosphorylation at these two sites mimicked the phenotype of loss of ERα protein in terms of acquisition of an altered cell morphology and increased migration/invasion. Although S118A and S167A cells exhibited this morphology and migration/invasion behavior, an examination of cell markers revealed that both cell lines retained expression of pancytokeratin and the membrane localization of β-catenin, both hallmarks of epithelial-lineage cells. In normal, differentiated breast epithelial tissue, β-catenin is localized to the cell membrane and complexes with cadherins in adherens junctions joining the cytoskeleton with epithelial cells (23, 31). Although the exact mechanism of β-catenin in breast cancer progression remains unclear, reduced localization of β-catenin to the membrane and increased accumulation in the cytoplasm and nucleus has been correlated with poor clinical outcome in breast cancer patients (31). Furthermore, loss of association between E-cadherin and β-catenin at the membrane has been associated with dedifferentiated breast tumors as well as increased invasion and migration of cells (32, 33). In normal breast tissue, cytoplasmic accumulation of β-catenin is prevented by rapid ubiquitination followed by proteasomal degradation (23, 31). However, stabilization of β-catenin by WNT ligands prevents this degradation, permitting cytoplasmic accumulation and eventual translocation of β-catenin to the nucleus (23, 31–33). Nuclear β-catenin serves as a convergence point for numerous signaling pathways, most notably the WNT signaling pathway (31, 33). There was no detection of cytosolic or nuclear β-catenin in either S118A or S167A cells. Furthermore, S118A and S167A cells did not express the mesenchymal marker vimentin, suggesting that even though S118A and S167A cells had acquired some mesenchymal-like features, the cells remained epithelial-like. Taken together, the data presented in this study indicated that phosphorylation at S118 and S167 was necessary for a fully active ERα signaling pathway and, by extension, also indicated that these phosphorylation events were required for maintenance of a fully differentiated epithelial phenotype in MCF-7 cells. Loss of these phosphorylation sites resulted in acquisition of a migratory/invasive phenotype along with other changes in cellular morphology (smaller cell size and less cellular complexity) that are hallmarks of mesenchymal cells.

Examination of select ERα-target gene expression revealed altered expression in both S118A and S167A cells. Most notably, estrogen-induced expression of pS2 in S167A cells was not detectable above basal level expression. Interestingly, Valley et al. (26) showed that pS2 expression in cells transiently expressing ERα-S167A was induced in response to estrogen but the fold increase was mainly due to the fact that basal expression of pS2 was severely depressed. Shah and Rowan (34) also showed that ERα-negative HeLa cells transiently transfected with ERα-S167A exhibited reduced estrogen-stimulated association of ERα with the pS2 promoter. Overexpression of pS2 is a hallmark of many malignancies including breast cancer. A previous study correlating pS2 overexpression with ERα positivity showed that tumors overexpressing pS2 had a more positive prognosis and were more likely to respond to endocrine therapy (35). These studies suggested that pS2 expression may also serve as a surrogate marker for a functional ERα signaling pathway, similar to the hypothesis that was tested in the present study with regard to phosphorylation at S118 and S167 as surrogates for functional ERα signaling. In S118A cells, there was no significant alteration of pS2 expression compared with parental MCF-7 or ER-AB cells. Similarly, Duplessis et al. (21) reported that estradiol-stimulated pS2 expression was not significantly different in ERα-negative HeLa cells transiently transfected with either WT ERα or ERα-S118A. In contrast, both S118A cells and S167A cells exhibited markedly altered cyclin D1 expression with a diminished peak accumulation that occurred at 6 h and quickly returned to baseline by 12 h. Parental MCF-7 cells and ER-AB cells exhibited a higher peak cyclin D1 mRNA accumulation that occurred at a later time point (12 h) and was sustained for a longer period of time. Although temporal expression of cMyc was altered in both S118A and S167A mutants, no clear pattern of expression was determined. Both cyclin D1 and cMyc are active in ushering cells past the G1/S checkpoint, ultimately allowing cells to complete the cell cycle. ERα-stimulated expression of cyclin D1 resulted in increased mRNA and protein expression, leading to the formation of cyclin D1-cdk4 complexes that phosphorylate and inactivate the tumor suppressor retinoblastoma, resulting in cell proliferation (36–38). Overexpression of cyclin D1 has been correlated with loss of cell cycle control and shown to mimic estrogen-stimulated proliferation in the absence of estrogen, permitting cells to escape apoptosis (37). Cyclin D1 is a convergence point for estrogen and growth factor cross talk because its expression can be directly regulated by kinases downstream of EGFR, independent of ERα (37). It is plausible that the increased growth exhibited by S118A and S167A cells was correlated to the early peaks in cyclin D1 expression effectively forcing cells through the cell cycle and preventing apoptosis. As with cyclin D1, cMyc is crucial for estrogen-stimulated proliferation, and cMyc expression was shown to be significantly increased in response to estradiol (39). Furthermore, overexpression of cMyc was capable of mimicking estrogen-stimulated proliferation in the absence of estrogen (39). Although no clear pattern of cMyc expression was observed in S118A and S167A cells, expression of cMyc was significantly elevated at 12 h in S167A cells in contrast to S118A cells, which exhibited no significant estrogen-induced expression of cMyc at the time points examined. Interestingly, qPCR array data demonstrated that genes regulated by ERα signaling or genes involved in breast cancer were markedly reduced in S118A cells. A large number of genes were significantly down-regulated or completely undetectable in S118A cells. A similar result was found with S167A cells, although the effects were not as pronounced as with the S118A cells. It is important to note that many of the ERα signaling genes represented on the array are cell cycle regulated. Under normal culture conditions, approximately 60% of cells are in the G1 phase, 25% are in S phase, and the remaining 15% are in G2 phase (17). The use of unsynchronized cells in the present study may have contributed to some of the observed results.

In summary, attenuated phosphorylation of ERα at S118 and S167 significantly affected cellular physiology and behavior of MCF-7 breast cancer cells resulting in increased growth, migration, and invasion; compromised expression of ERα target genes; and markedly altered gene expression patterns. It is possible that S118A and S167A cells were more efficient at by-passing cell cycle check points, ultimately allowing more cells to complete the cell cycle leading to elevated growth. Reduction in cell size (S167A cells) and complexity (S118A and S167A cells) suggested that these cells spend less time in G2. These data support the hypothesis that S118 and S167 phosphorylation may serve as surrogate markers of a functional ERα signaling pathway. It is plausible that attenuated phosphorylation at these sites in breast cancer patients may lead to tumor cells that are more aggressive, ultimately leading to poorer prognosis. These cell lines provide valuable models for studying the effects of long-term attenuated ERα phosphorylation on cell physiology, ERα function, tumor biology, and response to therapeutics in an ERα-positive genetic background.