Estrogen Negatively Regulates Epidermal Growth Factor (EGF ...

0888-8809/05/$15.00/0 Printed in U.S.A.

Molecular Endocrinology 19(11):2660–2670 Copyright © 2005 by The Endocrine Society doi: 10.1210/me.2004-0439

Estrogen Negatively Regulates Epidermal Growth Factor (EGF)-Mediated Signal Transducer and Activator of Transcription 5 Signaling in Human EGF Family Receptor-Overexpressing Breast Cancer Cells Julie L. Boerner, Matthew A. Gibson, Emily M. Fox, Erika D. Posner, Sarah J. Parsons, Corinne M. Silva, and Margaret A. Shupnik Departments of Microbiology (J.L.B., E.D.P., S.J.P., C.M.S.), Internal Medicine-Endocrinology (M.A.G., E.D.P., C.M.S., M.A.S.), and Pharmacology (E.M.F.), and The Cancer Center of the University of Virginia (S.J.P., C.M.S., M.A.S.), Charlottesville, Virginia 22908 Breast cancer cell growth may be stimulated by 17␤-estradiol (E2) or growth factors like epidermal growth factor (EGF). However, tumors typically depend on only one of these pathways and may overexpress either estrogen receptor (ER) or EGF receptor (EGFR) and related family members. Tumors overexpressing EGFR are more aggressive than those expressing ER. Intracellular mediators of these growth-stimulatory pathways are not completely defined, but one potential common mediator of EGF and E2 signaling is the transcription factor signal transducer and activator of transcription 5 (STAT5). To investigate the role of STAT5 in potential crosstalk between E2 and EGF, MDAMB231 and SKBr3 breast cancer cells, which are ER-negative and overexpress human EGF family receptors, were used. Introduction of ER␣ and treatment with E2 decreased EGF-induced tyrosine phosphorylation of STAT5b, basal and EGFinduced STAT5-mediated transcription, and EGF-

stimulated DNA synthesis in these cells. Suppressive effects of E2-⌭R␣ were specific for STAT5, as EGF stimulation of MAPK was unaffected. Deletion/mutation analysis of ER␣ demonstrated that the DNA-binding domain was insufficient, and that the ligand-binding domain was required for these responses. ER␣ transcriptional activity was not necessary for suppression of STAT5 activity. Overexpression of c-Src did not prevent suppression of STAT5 activity by E2 and ER␣. However, ER␣ did prevent basal increases in STAT5 activity with overexpressed c-Src. In the context of human EGF receptor family overexpression, E2-ER opposes EGF signaling by regulating STAT5 activity. STAT5 may be a crucial point of signaling for both E2 and growth factors in breast cancer cells, allowing targeted therapy for many types of breast tumors. (Molecular Endocrinology 19: 2660–2670, 2005)

T

ligand-dependent transcription factors by binding specific E2 responsive elements (EREs) found in responsive gene promoters (5), as well as by associating with other transcription factors tethered to DNA (6). ER proteins act as dimers, with dimerization interfaces in their DNA- and ligand-binding regions. Each ER also contains two activation functions that cooperate for full transcriptional activity, including an N-terminal ligand-independent activation function 1 (AF-1) and a C-terminal ligand-dependent domain. After E2 binding, association of coactivators regulates ER transcriptional activation by recruiting additional proteins that modify chromatin structure or aid in association with the transcriptional complex. E2-bound ER can also participate in nongenomic intracellular signaling, by association with intracellular signaling complexes (reviewed in Ref. 7). All of these mechanisms may play a role in ER-dependent breast cancer. Effective treatment of ER-negative breast cancers has been more elusive and is the subject of intense

HE GROWTH OF 40–70% of human breast cancers is dependent on the steroid hormone 17␤estradiol (E2) (1, 2). E2 dependence allows treatment of the cancer with antiestrogen or, more recently, aromatase inhibitor therapy (1, 3), by preventing E2 actions through its cognate receptor [estrogen receptor (ER)]. There are two ER subtypes (␣ and ␤), with ER␣ more highly expressed in breast cancer (4). ERs act as First Published Online June 23, 2005 Abbreviations: AF-1, Activation function 1; BrdU, bromodeoxyuridine; CMV, cytomegalovirus; DBD, DNA-binding domain; E2, 17␤-estradiol; EGF, epidermal growth factor; EGFR, EGF receptor; ER, estrogen receptor; ERE, E2 responsive element; HER, human EGF receptor; LBD, ligand-binding domain; LHRR, lactogenic hormone response region; PI3K, phosphatidylinositol 3-kinase; PRL, prolactin; STAT, signal transducer and activator of transcription. Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.

2660

Boerner et al. • STAT5 Inhibition by ER␣

investigation. ER-negative tumors frequently overexpress members of the human epidermal growth factor (EGF) family (HER family) of receptors, such as EGF receptor (EGFR)/HER1 and HER2/neu (8–10). HER2 forms heterodimers with EGFR, and responds to EGF binding to this heterodimer. The EGFR family proteins are receptor tyrosine kinases that, when bound to ligand through the extracellular ligand binding domain, result in receptor dimerization, kinase activation, and transphosphorylation on C-terminal regulatory tyrosines (11, 12). These phosphorylated tyrosines provide docking sites for many adapter molecules to activate the MAPK, phosphatidylinositol 3-kinase (PI3K), and signal transducer and activator of transcription (STAT) signaling pathways. Another tyrosine kinase, c-Src, also is overexpressed in many breast cancers, both ER-positive and ER-negative, and in fact is observed as a molecule co-overexpressed with the EGFR in 20% of all breast cancers (13). Cell or tumors that co-overexpress both ER and HER family receptors exhibit altered responses to ER-based therapies. Patients with tumors that overexpress HER family members typically have a poorer prognosis and do not respond as well to antiestrogen therapies, even if tumors express ER (14–16). In MCF-7 breast cancer cells with induced resistance to tamoxifen, an alternative basal growthregulatory pathway was established through increased levels of EGFR and HER2, increased basal phosphorylation of EGFR-HER2 and EGFR-HER3 dimers, and basal activation of an EGFR/HER2/ MAPK signaling cascade (17). Similar pathways are activated in breast cancer cells resistant to the pure ICI antiestrogen faslodex (18). Human breast tumors with acquired resistance to tamoxifen often have increased expression of c-erbB2 and/or EGFR, suggesting that cell signaling context will influence therapeutic responses (16). It is unclear whether steroids and growth factors act together to help cells reach a critical threshold level of common critical signaling molecules or whether there are both common and complementary pathways involved. Could steroid receptors also serve as an internal restraint to signaling through cytoplasmic pathways, requiring two signals (steroid and growth factors) for maximal proliferation? In this case, the loss of steroid receptors, along with increased cellular expression of other signaling molecules, would contribute to the flow of cytoplasmic signals through growth factorstimulated pathways. Reintroduction of ER might then act as a brake to inhibit flow through the common pathway. Cellular stoichiometry between the various receptors, and between groups of signaling molecules, may then be critically important, and overall cell proliferation would thus be determined by cell context and the endogenous hormonal milieu. There is increasing evidence that common downstream mediators are required for the pathways that regulate E2- and EGF-induced growth of breast can-

Mol Endocrinol, November 2005, 19(11):2660–2670 2661

cer cells. However, the common pathways linking these two mitogens have not been completely defined, and multiple points of interaction may occur. One potential common target is the signal transducer and activator of transcription (STAT) family of seven related proteins. STATs were originally identified in the interferon-signaling pathway and have been shown to be activated by a single tyrosine phosphorylation in their C-terminal domain, which leads to STAT dimerization (20). Once dimerized, STATs translocate to the nucleus where they bind DNA to regulate gene transcription (21). STATs are now known to be activated by a plethora of cytokines, growth factors, and hormones, including EGF (22, 23). In addition to activation by tyrosine kinase receptors, many of the STATs are activated by Srcfamily kinases, either directly or through crosstalk with growth factor or cytokine receptors (reviewed in Ref. 22). We reported previously that in EGFR/c-Srcoverexpressing cells, EGFR is phosphorylated at tyrosine 845 (Y845), and EGF stimulation of STAT5b, but not STAT3, is dependent on this phosphorylation. Furthermore, not only does Y845F inhibit EGFstimulated DNA synthesis, but so does dominantnegative STAT5b (21, 23). Steroid receptors, including progesterone and estrogen receptors, have been shown to functionally interact with STAT5 (24–29). For example, progesterone stimulation of breast cancer cells primed the cells for EGF-induced cellular signaling, including increased phosphorylation of MAPK and STAT5 and elevated protein levels of cyclin D1 (27, 29). ER activates STAT5b in aortic epithelial cells, whereas a dominant-negative STAT5 inhibits ER transcriptional activation in ER-positive breast cancer cells (25, 28, 30). Whereas these studies imply a positive relationship between ER and STAT5, ER can also suppress STAT5 activity in some contexts. For example, E2-ER suppresses prolactin (PRL)-stimulated STAT5 transcriptional activity in transfected human embryonic kidney (HEK)293 cells and in HEPG2 hepatocyte cells (26, 31). To gain insight into the relationship between ER and growth factor receptors and signaling pathways emanating from these receptors, we used the human breast cancer cell lines, MDA-MB231 and SKBr3, as models. MDA-MB231 and SKBr3 cells are ER negative and overexpress either EGFR (MDA-MB231) or EGFR and HER2 (SKBr3). Using these models, we found that expression of ER␣ negatively impacted EGF signaling. Treatment with E2 resulted in a decrease in STAT5 transcriptional activity (basally or in response to EGF), decreased STAT5b tyrosine phosphorylation, and decreased EGF-induced DNA synthesis. These inhibitory events required the C-terminal region of ER␣. Such findings suggest that the interference with STAT signaling may represent a viable, alternative target for ablating breast cancer growth.

2662 Mol Endocrinol, November 2005, 19(11):2660–2670

RESULTS E2 Decreases STAT5 Transcriptional Activity To investigate the potential interaction between the ER and STAT signaling pathways, we first examined STAT5 transcriptional activation in MDA-MB231 cells transiently transfected with plasmids encoding ER␣ and luciferase reporter constructs containing either an E2 response element (pGL2ERE) or a STAT5 response element [lactogenic hormone response region (LHRR)] and analyzed for luciferase activity. Figure 1A shows that E2 stimulates transcription of the ERE reporter and is completely inhibited by the pure ER antagonist


ICI 182,780. In contrast, E2 inhibited the basal transcriptional response of the STAT5 response element construct LHRR. The inhibitory effect on LHRR is mediated by ER, as it was completely reversed by treatment with the pure ER antagonist. An approximate 50% inhibition of LHRR promoter activity occurred regardless of the amount of ER␣ transfected (data not shown). Figure 1B demonstrates that, as previously observed, EGF activated STAT5 transcriptional activity. Treatment with E2 not only decreased the basal LHRR activity, but also suppressed EGF stimulation of LHRR transcription to basal levels. EGF had no effect on E2 induction of the ERE reporter. ER␣-E2 Suppresses STAT5b Tyrosine Phosphorylation and Transcriptional Stimulation by EGF

Fig. 1. E2 Specifically Suppresses STAT Responses in MDA-MB231 and SKBr3 Cells Transfected with ER␣ MDA-MB231 cells were transfected with an expression vector for ER␣ (1 ␮g/well CMV-ER␣) and either an EREcontaining luciferase reporter (pGl2-ERE; 1 ␮g/well) or a STAT5 response-element luciferase reporter (LHRR; 1 ␮g/ well). A, Transfected cells were treated with either ethanol vehicle [control (Con)], 10 nM E2, or 10 nM E2 plus 1 ␮m ICI 182,780 (E2 ⫹ ICI) for 24 h before collection and assay. B, Cells were transfected in DMEM without serum, and then incubated in serum-free media with no treatment (Con), E2 (10 nM), EGF (100 ng/ml), or E2 plus EGF for 24 h before collection. In all panels, data shown are normalized luciferase activity and are the mean ⫾ SEM for three experiments with triplicate samples per group. *, P ⬍ 0.05 vs. controls.

To investigate the role of the ER␣ in EGF signaling to the STAT5 pathway, we used an MDA-MB231 cell line stably expressing ER␣ (ERwt). Because expression of ER in cells in which the receptor is not normally expressed could lead to down-regulation of EGFR protein levels (32), we first investigated whether there was a decrease in EGFR protein in the ER␣-expressing cells. Figure 2A demonstrates that neither EGFR nor c-Src, which may mediate EGF signaling to STAT5 (22), is decreased in the cells expressing ER␣ compared with the vector control cells. In MDA-MB231 cells STAT5b is the predominant STAT5 form expressed (Silva, C. M., A. M. Weaver, and E. M. Branch, unpublished data), and STAT5b protein levels also were independent of ER␣ expression (Fig. 2A). Thus, any changes in STAT5b signaling cannot be due to gross changes in STAT5b expression. Because STAT5b must be tyrosine phosphorylated to mediate transcriptional activity, we next examined whether the inhibition of STAT5b transcriptional activity by E2 is mediated through alterations in tyrosine phosphorylation. Stable MDA-MB231 cell lines expressing the ER␣ or vector (pcDNA3) were treated with E2, EGF, or both compounds. In cells expressing the vector only, STAT5b was tyrosine phosphorylated in response to EGF, and this phosphorylation was not affected by E2 (Fig. 2B). However, in cells stably expressing ER␣, EGF-induced tyrosine phosphorylation of STAT5b was not detected in the presence or absence of E2 (Fig. 2B). These data suggest that one mechanism through which ER␣ regulates STAT5b transcriptional activity is by modulating its tyrosine phosphorylation. Given the effect of ER␣ alone on STAT5b tyrosine phosphorylation, we next sought to determine whether ER␣ could affect EGF-mediated STAT5 transcriptional activity, with or without E2. Either MDA-MB231 (Fig. 3A) or SKBr3 (Fig. 3B) cells were transiently transfected with ER␣ or vector alone and stimulated with EGF, E2, or both compounds, and LHRR-luciferase activity was measured. In MDA-MB231 cells, the presence of ER␣ alone did not eliminate EGF-stimulated



Fig. 2. ER␣ Inhibits EGF-Induced STAT5b Tyrosine Phosphorylation without Altering Expression of Key Signaling Molecules A, MDA-MB231 cells stably expressing ER␣ or vector controls (pcDNA3) were lysed, and proteins were separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted for EGFR, c-Src, STAT5, and ER␣, as described in Materials and Methods. B, MDA-MB231 clones that stably express empty vector (pcDNA3) or those that express wild type ER␣ (ERwt) were treated for 15 min with media [control (Con)], 1 ⫻ 10⫺8 M E2, 100 ng/ml EGF, or both. After treatment, detergent lysates were prepared and incubated with anti-STAT5b-specific antibody. Immunoprecipitates were analyzed by Western blotting with antiphosphotyrosine (top) or antiSTAT5b (bottom). The data are representative of at least three experiments. pTyr, Phosphotyrosine.

LHRR-luciferase activity although the fold-stimulation was somewhat diminished (compare white bars in Fig. 3A). However, treatment with E2 completely prevented EGF-induced LHRR-luciferase activity in ER␣-ex-

pressing cells (black bars, Fig. 3A). In SKBr3 cells, similar responses were observed. EGF stimulation of STAT5-mediated transcriptional activity was significantly reduced by the expression of ER␣ (white bars, Fig. 3B), and was further suppressed in the presence of E2 plus ER␣. Thus, ER␣ suppressed the STAT5b transcriptional response to EGF, but this was further reduced in the presence of E2. E2 Does Not Suppress EGF-Induced MAPK Activation

Fig. 3. Comparison of ER ␣ and ER␣-E2 Inhibition of EGFStimulated LHRR Activity MDA-MB231 (panel A) or SKBr3 cells (panel B) were transfected as in Materials and Methods with an expression vector for ER␣ (1 ␮g/well CMV-ER␣) and a STAT5 response-element luciferase reporter (LHRR; 1 ␮g/well). Cells were transfected in DMEM without serum, and then incubated in serumless media with no treatment [control (Con)], E2 (10 nM), EGF (100 ng/ml), or E2 plus EGF for 24 h before collection. In all panels, data shown are normalized luciferase activity and are the mean ⫾ SEM for three experiments with triplicate samples per group. *, P ⬍ 0.05 vs. controls.

To determine whether E2 affects growth factor signaling pathways in addition to STAT5, we examined the MAPK pathway. MAPK is phosphorylated and activated in response to growth factors including EGF and results in stimulation of transcription factors such as ELK-1 (33, 34). We measured transcriptional activation of the GAL4-luciferase construct by GAL4-ELK after stimulation by EGF, with and without ER␣ and E2. Figure 4 demonstrates that in MDAMB231 cells containing vector alone, EGF stimulated ELK transcriptional activity, alone or in combination with E2. In cells expressing ER␣, EGF stimulated ELK transcriptional activity to the same extent as in vector-containing cells, and this was not suppressed by E2. In fact, treatment of cells with E2 alone showed a slight stimulation of ELK transcription. In agreement with the ELK-mediated transcriptional response, EGF also stimulated MAPK phosphorylation in either vector- or ER␣-containing cells, and E2 did not abrogate this effect (Fig. 4B). In cells expressing ER␣, no significant changes in basal or EGF-stimulated levels of either MAPK (3-fold stimulation by EGF) or Akt phosphorylation (3-fold stimulation by EGF; data not shown) were observed. These data provide evidence that the E2 inhibition of STAT5 transcriptional activation (Figs. 1 and 3) is not due to a general inhibition of EGF-mediated signaling and transcriptional responses.



Fig. 4. E2 Does Not Suppress ELK Transcriptional Activation or MAPK Phosphorylation by EGF A, Transient transfection analysis was used to study MAPK-mediated Elk-1 activation cells by cotransfecting MDA-MB231 cells with GAL4-Elk-1 and GAL4-E1B-luciferase vectors. Cells were serum starved and transfected as in Materials and Methods. Twenty-four hours after transfection, cells were stimulated with vehicle, 10 nM E2, 100 ng/ml EGF, or both for 8 h. Luciferase activity for each lysate was determined. The data are presented as mean ⫾ SEM of three experiments. *, P ⬍ 0.05 compared with untreated control. B, Stable MDA-MB231 cells expressing either vector or ER␣ were serum starved for 24 h and treated with 100 ng/ml EGF, 10 nM E2, or both for 15 min. Lysates were collected and separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted for total MAPK (MAPK) and phosphorylated MAPK (pMAPK). A representative immunoblot is shown, and numbers below the blot are densitometric values for the mean ⫾ SEM for four blots. Con, Control; ERwt, wild-type ER; Ave., average.

E2 Suppression of STAT5 Activity Is Independent of the Kinase Activities of Src, MAPK, and Akt One possible mechanism for the inhibition of STAT5 activity by E2 and ER␣ is changes in activities of kinases important to both E2 and growth factor-stimulated pathways. Two ways to address this possibility is to overexpress or inhibit a given kinase to determine whether the inhibitory effect of E2 is relieved or enhanced. One of the most important of these kinases is c-Src, which stimulates STAT activity and is a critical signaling partner with ER and EGF. Figure 5A shows that increasing amounts of c-Src in SKBr3 cells stimulates STAT5 activity, as measured by LHRR luciferase promoter activity. In cells expressing ER, overexpression of c-Src did not change the ability of E2 to decrease LHRR activity. However, ER did prevent basal increases in STAT5 activity observed with overexpressed c-Src, suggesting that titration of c-Src by ER could play some role in suppression of basal STAT activity. In addition to c-Src, MAPK and Akt modulate both E2- and growth factor-signaling pathways (reviewed in Ref. 7). We used pharmacological inhibitors of c-Src (PP2), MAPK (U0126), and PI3K (LY294002) to determine whether inhibiting these kinases would alter the inhibitory effect of ER-E2 signaling on STAT5 activity. Figure 5B demonstrates that although basal STAT5 activity, as measured by LHRR reporter activity, was modified by PP2 and U0126, these responses were

consistent in both vector and ER-containing cells, and none of the tested inhibitors reversed the inhibitory effect of E2 treatment on STAT5 activity. The C-Terminal Region of ER Is Required to Suppress STAT5 Transcription To determine how E2 may be regulating STAT5-mediated transcription through the ER, several ER␣ mutants were tested for their ability to inhibit LHHR activity. Figure 6A shows the mutant ER constructs used in this study, and Fig. 6, B and C, depicts the effects of each of these ER mutants on STAT5 transcriptional activity in MDA-MB231 and SKBr3 cells. Loss of the ability of ER to dimerize (L507R), loss of the helix 12 activity (E542K), mutation of the DNA-binding region (DBDmut), or mutation of the dimerization domain in the DBD (S236E) all result in at least a 75% loss in E2-stimulated transcription on an ERE (Refs. 35–38, and our unpublished data). The truncation mutant (AF1DBD) binds DNA but cannot respond to E2. Mutations in the dimerization (L507R) and coactivator binding (E542K) regions, or truncation of the ER, leaving only the AF-1 and the DNA-binding domains (AF1DBD) also resulted in the inability of E2 to inhibit STAT5b transcriptional activity. In contrast, mutation of the dimerization site in the DBD (S236A), mutation of the growth factor-MAPK-sensitive phosphorylation site in the N terminus (S118A), or mutation of the tyrosine reported to be phosphorylated by c-Src



Fig. 5. E2 Suppression of LHRR-Luciferase Activity Is Independent of the Kinase Activities of Src, MAPK, and Akt A, SKBr3 cells were transfected with the STAT5 response-element luciferase reporter (LHRR; 1 ␮g/well) and either vector alone or ER␣ with increasing amounts of expression vector for c-Src as shown. Transfected cells were treated with either ethanol vehicle [control (Con)] or 10 nM E2 for 24 h before collection and assay for luciferase activity. B, SKBr3 cells were transfected with LHRR expression vector and treated with ethanol (Con) or 10 nM E2 in the presence of various inhibitors (10 ␮M PP2,100 nM U0126, or 10 ␮M LY294002) for 24 h before collection for luciferase assay. In all panels, data shown are normalized luciferase activity, and are the mean ⫾ SEM for three experiments with triplicate samples per group. *, P ⬍ 0.05 vs. controls. LY, 294002.

(Y537F) all had no effect on the ability of ER to regulate STAT5 transcriptional activity. The DNA binding mutant (DBDmut) had only partial effects on suppression, suggesting that the DNA-binding region contributes to, but is not sufficient for, ER effects on STAT5mediated transcription (Fig. 5C). In MDA-MB231 cells, there were equivocal results with the DBDmut (data not shown); this may be because the extent of the response in MDA-MB231 cells was smaller than in SKBr3 cells, and it was more difficult to verify a partial response in that context. Taken together, these data suggest that the C-terminal region of the ER and perhaps the DBD are required for abrogation of STAT5 transcriptional activation. E2 Decreases EGF-Induced Bromodeoxyuridine (BrdU) Incorporation In addition to modulation of transcriptional activity, STAT5 recently has been shown to regulate EGFinduced proliferation. Expression of a dominantnegative form of STAT5, which inhibits both STAT5a and STAT5b, in SKBr3 breast cancer cells led to an inhibition of EGF-induced DNA synthesis (21). To determine whether E2 modulated EGF-induced DNA

synthesis, MDA-MB231 cells stably expressing ER␣ were serum starved and pulsed with BrdU in the presence of EGF, E2, or both. In the presence of EGF, both the vector and ER-expressing cells responded to EGF by stimulating BrdU incorporation (Fig. 7A). E2 alone had no effect in either cell type. However, when E2 and EGF were added together in ER-expressing cells, BrdU incorporation remained at basal levels and was no longer stimulated by EGF. To determine whether the inhibition of BrdU incorporation was dependent on the same regions of ER as those found for inhibition of STAT5 transcriptional activity, MDA-MB231 cells transiently expressing the wild-type ER, S118A-ER, or L507R-ER were again pulsed with BrdU in the presence of EGF, E2, or both. In the presence of E2, both wild-type ER and S118A-ER, which suppressed STAT5-mediated transcription, suppressed BrdU incorporation stimulated by EGF (Fig. 7B). In contrast, the L507R-ER, which lost the ability to suppress STAT5-mediated transcription after E2 treatment, also was unable to decrease EGF-stimulated DNA synthesis in the presence of E2. Thus, the ability of ER mutants to suppress STAT5-mediated transcription correlated with effects on EGF-stimulated DNA synthesis.



Fig. 6. E2 Suppression of LHRR-Luciferase Activity Requires a Functional ER␣ C-Terminal Domain A, ER␣ functional domains, including the hormone-dependent (AF-2) and hormone-independent (AF-1) activation functions, DBD, LBD, and activation helix (Helix 12) are shown. The location of mutations in ER␣ mutants tested for effects on E2 suppression of LHRR-luciferase are also depicted. MDA-MB231 (panel B) or SKBr3 cells (panel C) were transfected with 2 ␮g LHRR-luciferase and 1 ␮g of expression constructs for either wild-type ER␣ (ER␣) or the indicated ER␣ mutations. Transfected cells were treated for 24 h with 10 nM E2 before collection and assay. Normalized luciferase levels were expressed relative to expression of the wild-type ER␣ (set at 1.0) and are the mean ⫾ SEM for four to five experiments with triplicate samples per group. *, P ⬍ 0.05 comparing control and E2-treated samples for each ER␣ construct pair. Con, Control; WT ER, wild-type ER.

DISCUSSION Studies described in this report provide evidence for crosstalk between ER-growth factor receptor signaling pathways that is mediated by STAT5. Our results show that E2, acting through ER, suppresses basal STAT5-mediated transcription as well as suppressing EGF-stimulated STAT5b tyrosine phosphorylation, STAT5-mediated transcription, and EGF-induced DNA synthesis. The suppressive effects of E2-ER were not specific to one cell system, as similar results were observed in both the EGFR-overexpressing MDAMB231 cells and the EGFR/HER2 overexpressing SKBr3 breast cancer line. Inhibitory effects of E2-ER on the STAT pathway were not due to a general inhibition of EGF signaling, as EGF-stimulated MAPK phosphorylation and MAPK-mediated transcription through ELK-1 were not affected. In addition, there were no changes in basal activity of MAPK or Akt in cells stably expressing ER, compared with vector controls. Inhibition of MAPK or Akt activity did not alter the ability of ER and E2 to suppress the STAT5 pathway. In MDA-MB231 cells, ER alone suppressed EGFstimulated phosphorylation of STAT5b, but did not abolish EGF-stimulated transcription. This may be due to the relative sensitivity of the two assays (protein

phosphorylation vs. transcription) or the differences in time frame required to observe phosphorylation (minutes) vs. transcriptional (hours) responses. In SKBr3 cells, some suppression of both basal and EGF-stimulated LHHR-promoter activity by ER alone was observed. This may be due to the relative ease of observing significant suppression with the more robust transcriptional response in SKBr3 cells, or because signaling to the STAT5 pathway is subtly different between the two cell types. A recent report demonstrating that STAT5a and ER␣ can be immunoprecipitated in the absence or presence of E2 (39) supports the idea that some responses could be mediated by the presence of ER alone. The character of the ER-STAT response, whether it is positive or negative, may be dependent on several criteria including the specific end point measured, and the presence or absence of other transcription factors and membrane receptor signals. The nature of the response may also be reflected in the domains of the ER required to mediate the biological effects. For example, E2-ER suppresses PRL-induced STAT5 transcriptional activity in HEK293 cells or clonal hepatocyte cells (26, 31), but amplifies PRL effects on STAT5 in mammary epithelial cells (24). Both responses require the DNA binding domain of ER, and direct inter-


Fig. 7. Inhibition of EGF-Induced DNA Synthesis by E2 Requires the C-Terminal Domain Upper panel, MDA-MB231 cells stably expressing pcDNA vector or wt-ER␣ were serum starved in phenol red-free media for 18 h followed by incubation with BrdU and the indicated ligands (EGF at 100 ng/ml; E2 at 10⫺8 M) for 8 h. Cells were then fixed and stained for BrdU using anti-BrdU Texas Red and assayed by immunofluorescence microscopy. Lower panel, MDA-MB231 cells were transiently cotransfected with wild-type ER, S118A-ER, or L507R-ER and green fluorescent protein (ratio of 10:1) and serum starved in phenol red-free media for 18 h followed by stimulation with BrdU and the indicated ligands as above for 8 h. Cell were fixed and stained for BrdU using anti-BrdU Texas Red. Green fluorescent protein-positive cells were scored for BrdU incorporation by immunofluorescence microscopy. The data represent the mean ⫾ SEM of three to four experiments. *, P ⬍ 0.05 as compared with control; #, P ⬍ 0.05 as compared with EGF-stimulated cells. Con, Control; ERWT, wild-type ER.

actions between ER and STAT5b in vitro occurs via the ER DBD (24, 26). Presumably, these are nuclear events. In contrast, in endothelial cells, treatment with E2 alone stimulates STAT activation and ␤-casein promoter activity by a process requiring the ligand-binding domain (LBD) but not the DBD of the ER (25). Within the LBD, dimerization but not coactivator binding is required. It is postulated that this process occurs primarily via nongenomic signaling through the MAPK, c-Src, and PI3K pathways (25). To our knowledge, these studies are the first to examine crosstalk of ER and the EGF signaling pathway via STAT5. The activity of ER mutants tested was consistent for both suppression of STAT5 transcriptional activity and suppression of EGF-stimulated DNA synthesis. The ability of ER/E2 signaling to inhibit STAT5b tyrosine phosphorylation by EGF within 15 min of treatment suggests a role of rapid E2 cytoplasmic signaling events. In our studies, inhibitory effects of E2-ER on STAT5 signaling required the LBD of ER


and its ability to dimerize. Other investigators have found that nongenomic signaling through ER to STATs (25) and other signaling molecules (40) also requires the LBD and dimerization of this domain. We did not observe altered MAPK or Akt activity in MDA-MB231 cells stably expressing ER vs. vector alone, unlike long-term E2-deprived breast cancer cells (41). Further, inhibition of MAPK and PI3K by pharmacological inhibitors did not alter E2-ER suppression of STAT signaling. The c-Src inhibitor PP2 alone suppressed STAT5 signaling, and increasing c-Src expression did not stimulate STAT5 activity when ER was coexpressed, suggesting some titration of c-Src by ER may occur. Inhibition of c-Src completely prevents stimulated STAT5 phosphorylation by EGF (data not shown), and it was not possible to assess the effects of E2-ER under these conditions. Other investigators observed that expression of a DBD mutation of ER abolished E2 effects on PRLstimulated transcription on STAT5 elements, or prevented ER-STAT binding (26, 41). Using the same DNA-binding mutant, we observed only a partial amelioration of the suppressive response. Our data also suggest that the DBD alone (AF1DBD) is insufficient to mediate the suppressive effects of ER on a STAT5 response element. We suggest that several domains of the ER may be responsible for inhibiting STAT5 transcriptional activity, phosphorylation, and EGFstimulated proliferation. The role of multiple ER␣ domains may reflect a crosstalk of ER␣- and STAT5signaling pathways in the nucleus (genomic) as well as at the membrane/cytoplasm. The studies we present here provide evidence for crosstalk between the ER- and EGFR-signaling pathways that impinge on the STAT5 pathway. Our studies demonstrate that the suppression of STAT5b phosphorylation and STAT5-mediated transcription by E2-ER can have biological consequences, in that E2-ER suppressed EGF-stimulated DNA synthesis in MDA-MB231 cells. These data are in agreement with previous work demonstrating that in MDA-MB231 cells engineered to overexpress ER, E2 inhibits cell proliferation in the presence of fetal calf serum, which contains several growth factors (43). Similar growth inhibition is observed in several ER-negative cell lines with introduced ER and subsequent E2 treatment (43, 44). Although the mediator of the growth-inhibitory response to E2-ER was not defined in these earlier studies, our data suggest that the STAT5 could play a role, with reintroduction of ER acting as a brake to inhibit signaling flow from growth factors through the common STAT5 pathway. Modulation of the STAT5 pathway may provide a unique opportunity for targeted therapy in both ERpositive (E2-dependent) and ER-negative (growth factor-dependent) breast cancers. STATs have been targeted for therapeutic regimens for several cancers, including breast, head and neck, and prostate cancer (51), and animal models using dominant-negative forms of STAT5 have been shown to prevent tumor



growth in xenograft studies (28). Although STAT5 is involved in many signaling pathways, including responses to cytokines and growth factors, knockdown of STAT5a/b is not embryonic lethal but does have severe defects in mammary gland development and some growth deficiencies (52, 53). Therapeutic targeting of the STAT5 pathway after development would likely have more impact in breast cells compared with other tissues, and cancer cells may be more susceptible to suppression of increased STAT5 activation from steroids or growth factors.

MATERIALS AND METHODS Transient Transfection, ER Constructs, and Luciferase Assays MDA-MB231 cells were maintained in Cellgro DMEM (Mediatech/Fisher, Herndon, VA) with 10% fetal bovine serum (GIBCO/Invitrogen, Grand Island, NY) and 100 U/ml penicillin and 100 ␮l/ml streptomycin (GIBCO/Invitrogen). For transfection studies, 231 cells were plated in six-well plates (Corning) at a density of 2.5 ⫻ 105 cells per well in phenol red-free DMEM (Mediatech/Fisher) containing 5% charcoal-stripped newborn calf serum, 100 U/ml penicillin, 100 ␮l/ml streptomycin, and 2 mM L-glutamine (GIBCO/Invitrogen). The next day, cells were transfected with Lipofectin reagent (Invitrogen, Carlsbad, CA) for 5 h according to the manufacturer’s instructions, using 20 ␮l of Lipofectin per well and indicated DNA constructs. After this incubation, each well was washed once with PBS, media were replaced, and cells were incubated for 18 h. Cells were then treated with fresh media containing the treatments as indicated for an additional 24 h. Either ethanol or 10 mM acetic acid with 1% BSA was used as vehicle control for E2 and EGF, respectively. After treatment, the cells were washed once with PBS and collected in 250 ␮l of 1⫻ Cell Culture Lysis Reagent (Promega Corp., Madison, WI). Lysates were then assayed for luciferase activity using a Turner TD-20e luminometer (Molecular Dynamics, Inc., Sunnyvale, CA), and for ␤-galactosidase activity. ␤-Galactosidase activity from each sample was measured using a ␤-Gal Assay Kit (Invitrogen), in a ELx800 Microplate Reader (Bio-Tek Instruments, Inc., Winooski, VT). Luciferase activity from each sample was normalized either to protein levels determined by the Bio-Rad Protein Assay (Bio-Rad Laboratories, Inc., Richmond, CA) or to ␤-galactosidase activity. Each treatment was performed in triplicate, and each experiment was repeated at least three times. SKBr3 cells were maintained in the same media as MDA-MB231 cells and were also plated at a density of 2.50 ⫻ 105 cells per well in six-well plates for transfection. Transfection studies and analysis were performed identically as for MDA-MB231 cells, except that cells were transfected the day after plating with 2.5 ␮l Lipofectamine 2000 reagent (Invitrogen) per ␮g of DNA, according to the instructions of the manufacturer, and incubated overnight. After incubation the cells were treated for 24 h with fresh media containing appropriate treatments. Reporter and Expression Constructs To assess ER activity on a canonical E2 response element, 1 ␮g of the pGL3–2ERE reporter, containing two consensus EREs upstream of a PRL TATA box and the firefly luciferase gene, was added to each well of a six-well plate already seeded with cells (45). To determine transcriptional activity of STAT5, two copies of the LHRR from the ␤-casein gene were linked upstream to luciferase and used as previously de-

scribed (46). All human ER␣ constructs were cloned into the EcoR1 site of the pcDNA 3.1 expression vector containing a cytomegalovirus (CMV) promoter (Invitrogen). Wild-type CMV-hER␣, except where noted, was added at 1 ␮g/well. All mutant constructs were created by QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA) of CMVhER␣, and include: AF-1/DBD, containing only the N-terminal, DNA binding domain and nuclear localization signal (47), the DNA-binding and dimerization mutant S236E (35), the N-terminal ras-dependent phosphorylation mutant, S118A (36), the dimerization mutant L507R (37), the helix 12 coactivator-binding mutant E542K (49), the tyrosine phosphorylation mutant Y537F (38), and the DNA-binding mutant (E203G,G204S,A207V) (49), referred to in this manuscript as DBDmut. All mutants, added at 1 ␮g/well, were expressed efficiently and in equal amounts in transfected cells and were tested for appropriate activity on the pGL2-ERE promoter. To determine whether overexpression of c-Src restored Stat5 activity, a CMV-c-Src expression construct was transfected into SKBr3 cells with the expression vector alone (PCDNA3) (0.02–0.5 ␮g) as a negative control. In studies to assess the role of other intracellular signaling pathways on ER regulation of STAT5, the MAPK kinase inhibitor U0126 (10 ␮M), the PI3K inhibitor LY294002 (50 ␮M), or the c-Src inhibitor PP2 (10 ␮M) was added to transfected cells 1 h before E2 treatment. PP2 was obtained from Calbiochem (La Jolla, CA), and all other inhibitors were from Sigma Chemical Co. (St. Louis, MO). To study MAPK-mediated Elk-1 activation, 231 cells were transiently transfected with both 2 ng of GAL4-Elk-1 and 0.1 ␮g of GAL4-E1B-Luc reporter vectors encoding the GAL-Elk fusion protein and the GAL-luciferase reporter gene, respectively, as previously described (34). Transfections were performed as above and 24 h after transfection, cells were treated with 10 nM E2, 100 ng/ml EGF, or both for 8 h. Luciferase activity was measured for each lysate, and samples were normalized by ␤-galactosidase cotransfected with above vectors. All transfections were conducted at least three times, and the data are presented as mean ⫾ SEM. Western Blotting MDA-MB231 clones expressing vector (pcDNA3) or wild-type ER (ERwt) were kindly provided by Dr. Theresa Guise (Department of Internal Medicine and Division of Endocrinology, University of Virginia). They were maintained in DMEM/10% fetal calf serum with G418 and passaged twice weekly. Cells were treated in 10-cm dishes with either media alone, 1 ⫻ 10⫺8 M, or 100 ng/ml EGF for 15 min at 37 C. After treatment, media was removed, cells were washed twice in 1⫻ Dulbecco’s PBS, and lysed in RIPA buffer (0.15 M NaCl, 1% Triton X-100, 1% deoxycholate, 5 mM EDTA, 50 mM Tris, pH 7.4) containing 1⫻ protease inhibitor cocktail (Set I; Calbiochem, La Jolla, CA) and 10 mM sodium orthovanadate. Lysates were centrifuged at 15,000 rpm at 4 C to remove cellular debris. For immunoprecipitations, lysates were then incubated overnight at 4 C with STAT5b-specific polyclonal antibody (50). Immunoprecipitates were captured on protein A agarose beads and removed by boiling for 5 min in 1⫻ Laemmli buffer. Samples were analyzed by denaturing PAGE (7.5% acrylamide) and then electrophoretically transferred to nitrocellulose. Nitrocellulose blots were probed with either antiSTAT5b or antiphosphotyrosine (py99, Santa Cruz Biotechnology Inc., Santa Cruz, CA) antibodies. Horseradish peroxidase-conjugated antirabbit or antimouse secondary antibodies were used followed by enhanced chemiluminescence (Amersham Pharmacia Biotech, Arlington Heights, IL) to detect binding of primary antibodies. For immunoblotting, 100 ␮g lysate was separated by SDS-PAGE and transferred unto nitrocellulose as described above. Membranes were blotted with antibodies specific for EGFR (Cell Signaling Technologies, Waltham, MA), Src (2–17), STAT5b, ER␣ (Santa Cruz), phospho-MAPK (Sigma), MAPK (B3B9, gift



from M. Weber), pAkt (Cell Signaling Technologies, Beverly, MA), or Akt (Cell Signaling Technologies). 7. BrdU Incorporation Assays Cells were plated on coverslips and grown overnight. Transient transfections were performed using FuGENE6 (Roche Clinical Laboratories, Indianapolis, IN) and wild-type ER, S118A-ER, or L504R-ER plasmid DNA in the presence of 1/10⫻ green fluorescent protein DNA as a marker of transfection. Twenty four hours later cells were placed in serumfree media for an additional 24 h. Cells were then stimulated with EGF (100 ng/ml) and/or E2 (10⫺8 M) in the presence of BrdU for 8 h. Cells on coverslips were fixed by incubation with 4% paraformaldehyde for 20 min, and the nuclei were permeabilized with 2 M HCl for 1 h at 37 C, followed by neutralization with 1 M Borate Buffer. Cells were blocked in 20% goat serum for 1 h at 37 C and then incubated with anti-BrdU-Texas Red (Molecular Probes, Eugene, OR). After 1 h incubation at room temperature, cells were washed three times with PBS and mounted onto glass slides using Vectashield mounting medium (Vector Laboratories, Inc., Burlingame, CA) and Cytoseal (Richard-Allan Scientific, Kalamazoo, MI). Green fluorescent protein-positive cells were scored for BrdU incorporation into newly synthesized DNA, whereas untransfected cells in the same slide were used as internal controls. Cells (50–100) were counted for each treatment/experiment, with each experiment being repeated at least three times.

8.

9.

10.

11.

12.

13.

Acknowledgments We thank Dr. Theresa Guise and Juan Juan Yin for the stable MDA-MB231-ER-expressing cells and Dr. Michael Weber for the MAPK antibody. We thank Kristin Laughlin for technical assistance and Angela Groehler for careful reading of the manuscript.

Received October 29, 2004. Accepted June 15, 2005. Address all correspondence and requests for reprints to: Margaret A. Shupnik, Department of Internal MedicineEndocrinology, University of Virginia, Charlottesville, Virginia 22908. E-mail: [email protected]. This work was supported by National Institutes of Health Grants F32-CA093028-01 (to J.L.B.); R01-DK57082 (to M.A.S.); R01-CA85462 (to C.M.S.); and R01-CA71449 (to S.J.P.); and developmental funds from the University of Virginia Cancer Center.

REFERENCES 1. McGuire WL 1975 Endocrine therapy of breast cancer. Annu Rev Med 26:353–363 2. Nilsson S, Makela S, Treuter E, Tujague M, Thomsen J, Andersson G, Enmark E, Pettersson K, Warner M, Gustafsson JA 2001 Mechanisms of estrogen action. Physiol Rev 81:1535–1565 3. Baum M 2004 Current status of aromatase inhibitors in the management of breast cancer and critique of the NCIC MA-17 Trial. Cancer Control 11:217–221 4. Herynk MH, Fuqua SA 2004 Estrogen receptor mutations in human disease. Endocr Rev 25: 869–898 5. Hall JM, Couse JF, Korach KS 2001 The multifaceted mechanisms of estradiol and estrogen receptor signaling. J Biol Chem 276:36869–36872 6. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M,

14.

15.

16.

17.

18.

19. 20. 21.

22.

Chambon P, Evans RM 1995 The nuclear receptor superfamily: the second decade. Cell 83:835–839 Levin ER 2003 Bidirectional signaling between the estrogen receptor and the epidermal growth factor receptor. Mol Endocrinol 17:309–317 Koenders PG, Beex LV, Geurts-Moespot A, Heuvel JJ, Kienhuis CB, Benraad TJ 1991 Epidermal growth factor receptor-negative tumors are predominantly confined to the subgroup of estradiol receptor-positive human primary breast cancers. Cancer Res 51:4544–4548 Nicholson S, Wright C, Sainsbury JR, Halcrow P, Kelly P, Angus B, Farndon JR, Harris AL 1990 Epidermal growth factor receptor (EGFr) as a marker for poor prognosis in node-negative breast cancer patients: neu and tamoxifen failure. J Steroid Biochem Mol Biol 37:811–814 Nicholson RI, McClelland RA, Gee JM, Manning DL, Cannon P, Robertson JF, Ellis IO, Blamey RW 1994 Epidermal growth factor receptor expression in breast cancer: association with response to endocrine therapy. Breast Cancer Res Treat 29:117–125 Greenfield C, Hiles I, Waterfield MD, Federwisch M, Wollmer A, Blundell TL, McDonald N 1989 Epidermal growth factor binding induces a conformational change in the external domain of its receptor. EMBO J 8:4115–4123 Lax I, Bellot F, Howk R, Ullrich A, Givol D, Schlessinger J 1989 Functional analysis of the ligand binding site of EGF-receptor utilizing chimeric chicken/human receptor molecules. EMBO J 8:421–427 Biscardi JS, Belsches AP, Parsons SJ 1998 Characterization of human epidermal growth factor receptor and c-Src interactions in human breast tumor cells. Mol Carcinog 21:261–272 deFazio A, Chiew YE, Sini RL, Janes PW, Sutherland RL 2000 Expression of c-erbB receptors, heregulin and oestrogen receptor in human breast cell lines. Int J Cancer 87:487–498 Fitzpatrick SL, Brightwell J, Wittliff JL, Barrows GH, Schultz GS 1984 Epidermal growth factor binding by breast tumor biopsies and relationship to estrogen receptor and progestin receptor levels. Cancer Res 44: 3448–3453 Schiff R, Massarweh SA, Shou J, Bharwani L, Mohsin SK, Osborne CK 2004 Cross-talk between estrogen receptor and growth factor pathways as a molecular target for overcoming endocrine resistance. Clinical Cancer Res 10(Suppl):331S–336S Knowlden JM, Hutcheson IR, Jones HE, Madden T, Gee JMW, Harper ME, Barrow D, Wakeling AE, Nicholson RI 2003 Elevated levels of epidermal growth factor receptor/c-erbB2 heterodimers mediate an autocrine growth regulatory pathway in tamoxifen-resistant MCF-7 cells. Endocrinology 144:1032–1044 McClelland RA, Barrow D, Madden T, Dutkowski CM, Pamment J, Knolden J, Gee JM, Nicholson RI 2001 Enhanced epidermal growth factor receptor signaling in MCF7 breast cancer cells after long-term culture in the presence of the pure antiestrogen ICI 182,780 (Faslodex). Endocrinology 142:2776–2788 Ihle JN, Kerr IM 1995 Jaks and Stats in signaling by the cytokine receptor superfamily. Trends Genet 11:69–74 Darnell Jr JE 1997 STATs and gene regulation. Science 277:1630–1635 Kloth MT, Laughlin KK, Biscardi JS, Boerner JL, Parsons SJ, Silva CM 2003 STAT5b, a mediator of synergism between c-Src and the epidermal growth factor receptor. J Biol Chem 278:1671–1679 Silva CM, Boerner JL, Parsons SJ 2003 Interactions of STATs with Src family kinases. In: Sehgal PB, Levy DE, Hirano T, eds. Signal transducers and activators of transcription (STATs): activation and biology. Dordrecht, The Netherlands: Kluwer Academic Publishers; 223–236


23. Tice DA, Biscardi JS, Nickles AL, Parsons SJ 1999 Mechanism of biological synergy between cellular Src and epidermal growth factor receptor. Proc Natl Acad Sci USA 96:1415–1420 24. Bjornstrom L, Kilic E, Norman M, Parker MG, Sjoberg M 2001 Cross-talk between Stat5b and estrogen receptor-␣ and -␤ in mammary epithelial cells. J Mol Endocrinol 27:93–106 25. Bjornstrom L, Sjoberg M 2002 Signal transducers and activators of transcription as downstream targets of nongenomic estrogen receptor actions. Mol Endocrinol 16: 2202–2214 26. Faulds MH, Pettersson K, Gustafsson JA, Haldosen LA 2001 Cross-talk between ERs and signal transducer and activator of transcription 5 is E2 dependent and involves two functionally separate mechanisms. Mol Endocrinol 15:1929–1940 27. Lange CA, Richer JK, Shen T, Horwitz KB 1998 Convergence of progesterone and epidermal growth factor signaling in breast cancer. Potentiation of mitogen-activated protein kinase pathways. J Biol Chem 273: 31308–31316 28. Yamashita H, Iwase H, Toyama T, Fujii Y 2003 Naturally occurring dominant-negative Stat5 suppresses transcriptional activity of estrogen receptors and induces apoptosis in T47D breast cancer cells. Oncogene 22: 1638–1652 29. Richer JK, Lange CA, Manning NG, Owen G, Powell R, Horwitz KB 1998 Convergence of progesterone with growth factor and cytokine signaling in breast cancer. Progesterone receptors regulate signal transducers and activators of transcription expression and activity. J Biol Chem 273:31317–31326 30. Yamashita H, Nishio M, Fujii Y, Iwase H 2004 Dominantnegative Stat5 inhibits growth and induces apoptosis in T47D-derived tumors in nude mice. Cancer Sci 95: 662–665 31. Cao J, Wood M, Liu Y, Hoffman T, Hyde J, Park-Sarge OK, Vore M 2004 Estradiol represses prolactin-induced expression of Na⫹/taurocholate cotransporting polypeptide in liver cells through estrogen receptor-␣ and signal transducers and activators of transcription 5a. Endocrinology 145:1739–1749 32. Sheikh MS, Shao ZM, Chen JC, Li XS, Hussain A, Fontana JA 1994 Expression of estrogen receptors in estrogen receptor-negative human breast carcinoma cells: modulation of epidermal growth factor-receptor (EGF-R) and transforming growth factor ␣ (TGF ␣) gene expression. J Cell Biochem 54:289–298 33. Duan R, Xie W, Burghardt RC, Safe S 2001 Estrogen receptor-mediated activation of the serum response element in MCF-7 cells through MAPK-dependent phosphorylation of Elk-1. J Biol Chem 276:11590–11598 34. Song RX, McPherson RA, Adam L, Bao Y, Shupnik M, Kumar R, Santen RJ 2002 Linkage of rapid estrogen action to MAPK activation by ER␣-Shc association and Shc pathway activation. Mol Endocrinol 16:116–127 35. Chen D, Pace PE, Coombes RC, Ali S 1999 Phosphorylation of human estrogen receptor ␣ by protein kinase A regulates dimerization. Mol Cell Biol 19:1002–1015 36. Kato S, Endoh H, Masuhiro Y, Kitamoto T, Uchiyama S, Sasaki H, Masushige S, Gotoh Y, Nishida E, Kawashima H, Metzger D, Chambon P 1995 Activation of the estrogen receptor through phosphorylation by mitogen-activated protein kinase. Science 270:1491–1494


37. Schodin DJ, Zhuang Y, Shapiro DJ, Katzenellenbogen BS 1995 Analysis of mechanisms that determine dominant negative estrogen receptor effectiveness. J Biol Chem 270:31163–31171 38. Feng W, Ribeiro RC, Wagner RL, Nguyen H, Apriletti JW, Fletterick RJ, Baxter JD, Kushner PJ, West BL 1998 Hormone-dependent coactivator binding to a hydrophobic cleft on nuclear receptors. Science 280:1747–1749 39. Wang Y, Cheng CH 2004 ER␣ and STAT5a cross-talk: interaction through C-terminal portions of the proteins decreases STAT5a phosphorylation, nuclear translocation and DNA-binding. FEBS Lett 572:238–244 40. Razandi M, Pedram A, Merchenthaler I, Greene GL, Levin ER 2004 Plasma membrane estrogen receptors exist and functions as dimers. Mol Endocrinol 18:2854–2865 41. Santen RJ, Song RX, Zhang Z, Yue W, Kumar R 2004 Adaptive hypersensitivity to estrogen: mechanism for sequential responses to hormonal therapy in breast cancer. Clin Cancer Res 10(Suppl):337S–345S 42. Bjornstrom L, Sjoberg M 2002 Mutations in the estrogen receptor DNA-binding domain discriminate between the classical mechanism of action and cross-talk with Stat5b and activating protein 1 (AP-1). J Biol Chem 277: 48479–48483 43. Lazennec G, Bresson D, Lucas A, Chauveau C, Vignon F 2001 ER ␤ inhibits proliferation and invasion of breast cancer cells. Endocrinology 142:4120–4130 44. Kushner PJ, Hort E, Shine J, Baxter JD, Greene GL 1990 Construction of cell lines that express high levels of the human estrogen receptor and are killed by estrogens. Mol Endocrinol 4:1465–1473 45. Schreihofer DA, Resnick EM, Lin VY, Shupnik MA 2001 Ligand-independent activation of pituitary ER: dependence on PKA-stimulated pathways. Endocrinology 142: 3361–3368 46. Silva CM, Lu H, Day RN 1996 Characterization and cloning of STAT5 from IM-9 cells and its activation by growth hormone. Mol Endocrinol 10:508–518 47. Jazaeri O, Shupnik MA, Jazaeri AA, Rice LW 1999 Expression of estrogen receptor ␣ mRNA and protein variants in human endometrial carcinoma. Gynecol Oncol 74:38–47 48. Weis KE, Ekena K, Thomas JA, Lazennec G, Katzenellenbogen BS 1996 Constitutively active human estrogen receptors containing amino acid substitutions for tyrosine 537 in the receptor protein. Mol Endocrinol 10: 1388–1398 49. Mader S, Kumar V, de Verneuil H, Chambon P 1989 Three amino acids of the oestrogen receptor are essential to its ability to distinguish an oestrogen from a glucocorticoid-responsive element. Nature 338:271–274 50. Kloth MT, Catling AD, Silva CM 2002 Novel activation of STAT5b in response to epidermal growth factor. J Biol Chem 277:8693–8701 51. Yu H, Jove R 2004 The STATs of cancer—new molecular targets come of age. Nat Rev Cancer 4:97–105 52. Levy DE 1999 Physiological significance of STAT proteins: investigations through gene disruption in vivo. Cell Mol Life Sci 55:1559–1567 53. Teglund S, McKay C, Schuetz E, van Deursen J, Stravopodis D, Wang D, Brown M, Bodner S, Grosveld G, Ihle JN 1998 Stat5a and Stat5b proteins have essential and nonessential, or redundant, roles in cytokine response. Cell 93:841–850

Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.