| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Promoter Interaction, and Changes in Receptor Phosphorylation
Department of Biochemistry and Cancer Biology, Health Science Campus, University of Toledo, Toledo, Ohio 43614
Address all correspondence and requests for reprints to: Brian G. Rowan, Ph.D., Department of Structural & Cellular Biology, SL49, Tulane University School of Medicine and the Louisiana Cancer Research Consortium, 1430 Tulane Avenue, New Orleans, Louisiana 70112. E-mail: browan{at}tulane.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, F/I also decreased estradiol-dependent reporter gene transcription. Overexpression of PKA or PKA inhibitor (PKI) demonstrated that F/I effects on repression of estradiol action occurred through the PKA pathway. 8CPT-2Me-cAMP, a selective inducer of non-PKA signaling, did not alter ER-dependent transcription. In contrast to F/I effects on reporter genes, F/I exhibited gene-specific effects on endogenous, ER-regulated genes. F/I enhanced estradiol induction of pS2 and cMyc but repressed estradiol induction of cyclin D1 mRNA and protein in MCF7 cells. To explore likely mechanisms by which F/I regulated ER, experiments examined estradiol binding, Hsp90 interaction, promoter recruitment, and ER phosphorylation. F/I decreased estradiol binding and increased Hsp90 association with ER. Chromatin immunoprecipitation revealed that F/I recruited ER to both pS2 and cMyc promoters at earlier times than estradiol, and F/I shifted estradiol recruitment of ER to earlier time points. F/I induced a unique ER phosphorylation profile (increase in serine 305 and decrease in serine 118 phosphorylation) that was distinct from estradiol and estradiol + F/I. Taken together, F/I signaling through PKA selectively regulates estradiol-dependent genes in breast cancer, which is associated with reduced ligand binding and changes in promoter interaction and ER phosphorylation. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In the classical, ligand-dependent activation of ER, estrogen binding increases ER phosphorylation at specific sites (4) that facilitate ER dimerization and direct interaction with estrogen response elements (EREs) in the promoter of estrogen target genes (5). Liganded ER recruits a set of transcriptional coactivators with associated histone acetyltransferase activity that facilitates transcriptional activation. At some promoters, ER regulates transcription not through direct interaction with DNA but through binding to other transcription factors at the promoter including Sp-1 (6, 7), activator protein 1 (AP-1) (8, 9), GATA-1 (10), and nuclear factor-
B (11). In addition to ligand-dependent activation, ER may also be activated in the absence of ligand by cross talk with signal transduction pathways that is associated with changes in receptor and coregulator phosphorylation (12, 13, 14, 15, 16, 17, 18).
Protein kinase A (PKA), a holoenzyme composed of two regulatory subunits and two catalytic subunits, phosphorylates substrate proteins on serine/threonine residues. cAMP-dependent pathways govern many endocrine and neural functions. Elevated intracellular cAMP binds to the regulatory subunit of PKA causing phosphorylation of the catalytic subunit of PKA. The phosphorylated catalytic subunit then translocates to the nucleus and phosphorylates transcription factors such as cAMP response element (CRE)-binding protein (CREB) and CRE modulator (19). In addition to activating PKA signaling, cAMP also activates Epac (exchange protein directly activated by cAMP), a pathway that exhibits guanine nucleotide exchange activity (20, 21).
PKA signaling promotes ligand-independent activation of ER and regulates ligand-dependent ER activation (1, 18, 22, 23, 24). Fujimoto and Katzenellenbogen (23) showed that activation of PKA in MCF7 breast cancer cells alters the agonist/antagonist balance of antiestrogens. PKA phosphorylates ER at serine 236 within the DNA-binding domain, and this phosphorylation inhibits ER dimerization in the absence of estradiol (15). Activation of the PKA signaling protects ER from estrogen-induced degradation and promotes ER protein stability (25). In addition, PKA phosphorylates ER at S305, which is associated with tamoxifen resistance (26) and reduced ER acetylation (27).
Activated PKA inhibits both human breast cancer proliferation (28) and estrogen-regulated cell growth (29) in which ER contributes to cAMP level and activity (28, 29). However, the mechanisms by which PKA inhibits ER-positive breast cancer growth and affects both ligand-dependent and ligand-independent activation of ER in breast cancer remain unclear. Here we demonstrate gene-specific regulation of ER-dependent genes by forskolin/3-isobutyl-1-methylxanthine (F/I) that may underlie F/I suppression of breast cancer proliferation. We further describe the molecular signaling pathways and mechanisms that occur during ligand-independent activation by F/I and F/I suppression of estradiol action in breast cancer cells.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
, bar 4) and ERE2TK-luciferase (Fig. 1B, bar 4) in MCF7 cells. Although the magnitude of the ligand-independent activation by F/I relative to estradiol activation varied among experiments, the ligand-independent activation by F/I was consistently and quantitatively less than estradiol activation. This variability of the ligand-independent activation was likely due to subtle changes in intracellular kinase/phosphatase activity present during different experiments. Remarkably, F/I reduced full reporter activation by estradiol for both the ERE-TATA-luc and the ERE-TK-luc reporters (Fig. 1, A and B, bars 5). 4-hydroxytamoxifen [4-(OH)-Tam] did not activate ERE-TATA-luc or ERE-TK-luc (Fig. 1, A and B, bars 3) and repressed the basal transcriptional activity of these ERE-containing promoters. Basal reporter activity could be the result of residual estrogen present in the charcoal-stripped fetal bovine serum (FBS) and/or the result of ER ligand-independent activation by growth factors present in the medium.
|
Overexpression of PKA, but not PKI, significantly repressed estradiol activation of ERE-TK-luc (Fig. 2A). In these experiments, tamoxifen antagonist action was assessed by the ability of tamoxifen to repress estradiol activation of the reporter gene (Fig. 2A, bar 2 vs. bar 4). Overexpression of PKA completely repressed estradiol activation of the reporter (Fig. 2A, bar 2 vs. bar 5) and enhanced the tamoxifen repression of estradiol activation (Fig. 2A, bar 4 vs. bar 6). In contrast, overexpression of PKI had only modest effects on estradiol activation (Fig. 2A, bar 2 vs. bar 7) and did not alter tamoxifen repression of estradiol activation (Fig. 2A, bar 4 vs. bar 8). Results similar to those described in Fig. 2A were found with the ERE-TATA-luc reporter (data not shown).
|
). Examination of another tamoxifen-sensitive breast cancer cell line (T47D) revealed that overexpression of PKA in T47D cells decreased estradiol-dependent reporter activation (Fig. 2C, bar 2 vs. bar 5), whereas overexpression of PKI actually increased estradiol-dependent reporter activation (Fig. 2C, bar 2 vs. bar 6). Similar to MCF7 cells, 8CPT-2Me-cAMP had no significant effect on estradiol-dependent reporter activity in T47D cells (data not shown).
MCF7-LCC2 cells are resistant to tamoxifen antagonism of estradiol reporter activation (Fig. 2D, bar 2 vs. bar 4). Interestingly, overexpression of PKA repressed both estradiol reporter activation and estradiol + tamoxifen reporter activation in MCF7-LCC2 cells (Fig. 2D, bars 2 and 4 vs. bars 5 and 6). In contrast, overexpression of PKI had no significant effect on estradiol alone (Fig. 2D, bar 2 vs. bar 7) and only partially blocked estradiol + tamoxifen reporter activation in MCF7-LCC2 cells (Fig. 2D, bar 4 vs. bar 8). Incubation of cells with 8CPT-2ME-cAMP exhibited effects similar to that found with overexpression of PKI in MCF7-LCC2 cells (data not shown). To confirm the functional activity of the PKA and PKI expression vectors and of F/I treatment, MCF7 cells were cotransfected with the cAMP and PKA-activated CRE-luciferase (CRE-luc) reporter. Overexpression of PKA (Fig. 2E) or incubation of cells with F/I (Fig. 2F) resulted in significant activation of the CRE-luc reporter. In contrast, PKI reduced basal CRE-luc reporter activity (Fig. 2E). Incubation of cells with 8CPT-2Me-cAMP showed no significant effect on reporter activity (Fig. 2F). Nonspecific effects of F/I on general transcription were assessed by transfection of MCF7 cells with the constitutive Rous sarcoma virus (RSV)-Luc reporter. F/I, at the concentrations used in this study, had no effect on RSV-Luc activation (data not shown). The results presented above are consistent with F/I regulation of ER-dependent reporter gene activity occurring through selective activation of PKA, but not Epac signaling.
Gene-Selective Effects of F/I on ER-Dependent Transcription
The above data indicated that elevation of cAMP induced ligand-independent activation of ER and also suppressed full estradiol activation of ER-dependent reporter genes. To more comprehensively assess F/I effects on ER-regulated gene expression, four well-characterized ER-regulated genes in MCF7 cells were measured using real-time RT-PCR. F/I exhibited distinct differences in regulation of cMyc, pS2, progesterone receptor (PR), and cyclin D1 mRNA, and this regulation was markedly different from F/I effects on the ER-dependent reporter genes. F/I alone induced the expression of pS2, PR, and cMyc mRNA in MCF7 cells (Fig. 3, panels A, B, and C, respectively). However, in stark contrast to F/I effects on the ERE-TATA-luc and ERE-TK-luc reporters, F/I in combination with estradiol actually elevated mRNA for pS2 and cMyc above that of estradiol alone (Fig. 3, A and C, 
vs. 
). The relative effects of F/I or F/I + estradiol on cMyc mRNA corresponded to similar changes observed with cMyc protein levels (Fig. 3D). Although all treatments (F/I, F/I + estradiol, and estradiol) resulted in a rapid induction of cyclin D1 mRNA at 2 h, only estradiol treatment resulted in induction of cyclin D1 protein (Fig. 3F, lane 2) as previously reported (31, 32, 33). However, the rapid and transient induction of cyclin D1 mRNA at 2 h by F/I and F/I + estradiol was not sufficient to induce cyclin D1 protein. Notably, only estradiol treatment resulted in a reproducible and significant induction of cyclin D1 mRNA at 18 h treatment. It is likely that the sustained induction of cyclin D1 mRNA by estradiol at later time points was responsible for the increased cyclin D1 protein. F/I induction of ER target gene mRNA was ER-dependent because both 4-hydroxytamoxifen (Fig. 3G) and ICI 182,780 (data not shown) significantly reduced F/I induction of cMyc mRNA. F/I induction of PR, cyclin D1, and pS2 mRNA was also blocked by 4-hydroxytamoxifen and ICI 182,780 (data not shown).
|
transcriptional activity. Reporter gene assays were used in conjunction with endogenous gene expression to assess overall effects on gene expression.
F/I Rapidly Recruits ER to pS2 and cMyc Promoters and Alters Estradiol-ER Cyclic Recruitment
Estradiol-activated ER can bind directly to ERE sequences in promoters or may interact with DNA indirectly through protein-protein interactions with other transcription factors (e.g. AP-1 and SP1 proteins) (4). ER is recruited to promoters in a cyclic fashion involving successive rounds of recruitment and dissociation (34, 35, 36). The ER recognition sequences in the cMyc and pS2 promoters have been well characterized. ER is recruited to the ERE region of the pS2 promoter (37) (Fig. 4A) and ERE and AP-1 regions of the cMyc promoter (38) (Fig. 4B). To measure the effect of activated cAMP on ER association with cMyc and pS2 promoters, quantitative chromatin immunoprecipitation (ChIP) assays were performed as previously described by our laboratory (13). Incubation of MCF7 cells with F/I for 45 min significantly increased the association of ER with ERE sequences in the pS2 promoter (Fig. 4C, closed bar) and with the ERE/AP1 sequences in the cMyc promoter (Fig. 4C, open bar). In contrast, treatment with estradiol or F/I + estradiol did not recruit ER to these promoters at this early (45 min) time point (Fig. 4C). However, at longer incubation periods (2 h), incubation with estradiol, but not F/I or F/I + estradiol, significantly increased ER association with the pS2 promoter (Fig. 4D, closed bar) and with the cMyc promoter (Fig. 4D, open bars). To further examine time-dependent ER recruitment to the cMyc promoter and to determine the relative differences between F/I and estradiol, time course recruitment of ER to the cMyc gene was assessed. ER regulates cMyc transcription through both ERE and AP-1 sequences that are in close proximity to one another. Consequently, ChIP assays for the cMyc promoter assessed combined ER recruitment to ERE and AP1 sequences. ER recruitment to the cMyc promoter was significantly increased by F/I at 1 h, F/I + estradiol at 1.5 h, and estradiol at 2 h (Fig. 4E, 
, , and , respectively). A similar trend was observed for the ERE sequence of the pS2 promoter (data not shown).
|
promoter recruitment in MCF7 cells. F/I, but not estradiol, induces a rapid recruitment of ER to pS2 and cMyc promoters. Promoter recruitment by estradiol is delayed (2 h), but this delay is shifted to earlier time points (1.5 h) by the addition of F/I to estradiol treatments.
F/I Reduces Estradiol Binding to ER and Is Accompanied by Increased Association of ER with Hsp90
To examine possible mechanisms by which PKA represses estradiol-dependent transcription of selected genes, the effect of F/I on estradiol binding in intact cells was examined in MCF7 (Fig. 5A), MCF7-LCC2 (Fig. 5B), and T47D cells (Fig. 5C). F/I significantly reduced estradiol binding by 45% in all cell lines. Because ER is expressed approximately 4-fold higher than ERß in MCF7 cells (Fig. 5D), the reduced estradiol binding by F/I was due predominantly to effects on ER. ERß protein expression in MCF7 cells was low compared with ER (Fig. 5D) as previously reported (39), and detection by Western blot using crude cellular extract was weak. To confirm expression of ERß in MCF-7 cells, ERß was immunoprecipitated from cells followed by detection by Western blot (Fig. 5E). Notably, decreased ligand binding by F/I was not the result of reduced ER protein because F/I as well as estradiol did not significantly alter ER protein levels at 2 h incubation and earlier (refer to Fig. 7B). Interestingly, F/I had no effect on estradiol binding to purified ER (Fig. 5F), indicating that other cellular processes, and not direct F/I interaction with ER, were required for reduced ligand binding. Although reduced ligand binding may contribute to F/I antagonism of estradiol activation of cyclin D1 and the ERE-TATA-luc and ERE-TK-luc reporters, the reduced ligand binding by F/I was not sufficient to suppress estradiol induction of cMyc, pS2, and PR.
|
|
. F/I significantly increased Hsp90 association with ER in MCF7 cells (Fig. 6A). As a control we assessed ligand binding and Hsp90 interaction in HeLa cells, an ER-negative cell line in which ER is activated by F/I (Fig. 6B) as previously reported (16). In HeLa cells transfected with ER expression vector, F/I had no effect on estradiol binding (Fig. 6C). In parallel, F/I did not alter the Hsp90 interaction with ER in HeLa cells (Fig. 6D). These data suggest that increased association of Hsp90 with ER may be related to the selective attenuation of estradiol binding by F/I in MCF7 cells and may further lead to selective gene activation by F/I (currently under investigation).
|
is phosphorylated on eight residues (Fig. 7A), and receptor phosphorylation is known to regulate multiple receptor functions including nuclear translocation, ligand binding, ER dimerization, and transcriptional activation (3, 41). To identify F/I-induced changes in ER phosphorylation that may contribute to altered gene expression, phospho-specific antibodies to each ER phosphorylation site were developed (42) and used to profile ER phosphorylation in MCF7 cells. The major effect of F/I on ER phosphorylation was the induction of S305 phosphorylation, reduction of S118 phosphorylation, and suppression of the estradiol-induced S118 phosphorylation (Fig. 7, B–D). Other effects of F/I and estradiol on ER phosphorylation were less pronounced although consistent: estradiol induced phosphorylation at sites S104, S106, S118, S167, S236, S305, T311, and Y537 (Fig. 7B, lane 2); F/I suppressed estradiol induction of S118, S236, and Y537 (Fig. 7B, lane 2 vs. lane 4). F/I repression of estradiol-induced phosphorylation likely contributed to F/I-regulated decreases in ligand binding and estradiol-dependent gene activation. The effects of F/I and F/I + estradiol on S167 phosphorylation were inconsistent (data not shown). F/I and estradiol effects of ER phosphorylation were not the result of changes in ER protein level. Although previous reports showed that estradiol significantly decreased ER protein in MCF7 cells by 4 h and inhibition of PKA by H-89 decreased ligand-free ER protein in MCF7 cells (43), at incubation periods used in the present study (2 h and earlier) estradiol and F/I did not alter ER protein levels (Figs. 5–7).
F/I resulted in a distinct ER phosphorylation profile (decreased S118 and increased S305 phosphorylation) that was markedly different from vehicle or estradiol-induced phosphorylation. S118 is a MAPK consensus site that is phosphorylated after estradiol treatment and via cross talk with ERK1/ERK2 signaling (3, 17). PKA phosphorylates ER at S305 (26). Further experimentation investigated the involvement of S305 and S118 phosphorylation in F/I action. Because F/I decreased basal S118 phosphorylation, the effect of F/I on ERK1/ERK2 signaling was examined by Western blot analysis using phospho-specific antibodies to activated ERK1/ERK2. Activation/inhibition of ERK1/ERK2 by PKA is cell line and tissue dependent (33, 44, 45, 46). Overexpression of PKA in MCF-7 cells decreased basal phosphorylation of ERK1/ERK2 (Fig. 8A, lane 2). In addition, incubation of MCF-7 cells with F/I resulted in rapid decrease in ERK1/ERK2 phosphorylation (Fig. 8A, lanes 4–7). These data suggest that in MCF-7 breast cancer cells, PKA repressed S118 phosphorylation via reduced ERK1/ERK2 activation.
|
(13, 48, 49) and therefore F/I suppression of S118 phosphorylation likely contributes to F/I effect on reducing estradiol activation of selected genes. However, ligand-independent activation of ER by F/I would not likely be explained by suppression of S118 phosphorylation. To investigate the involvement of S305 in F/I action, ER wild type or ERS305A mutant was overexpressed in MCF7 cells, and ERE-TK-luc reporter activity was measured. F/I induced ligand independent activation of wild-type ER and decreased estradiol-dependent activation of ERE-TK-luc (Fig. 8B). With overexpression of ERS305A, there was reduced ligand-independent activation by F/I compared with overexpression of wild-type ER, and F/I did not significantly reduce estradiol activation of the reporter. This suggests that S305 contributes to both F/I ligand-independent activation of ER and to the F/I-dependent attenuation of estradiol activation of the reporter.
In summary, F/I, estradiol, and estradiol + F/I resulted in three distinct ER phosphorylation profiles in MCF7 cells. These profiles likely represent three different ER conformations that are important for differences in time-dependent recruitment of ER to pS2 and cMyc promoters and define the gene-selective action of F/I and estradiol.
F/I Suppresses MCF7 Cell Growth
MCF7 cells are dependent upon estradiol for proliferation (30), and forskolin exhibits an antiproliferative effect in MCF7 cells through activation of PKA signaling (29). The effect of F/I on estradiol-dependent growth of several ER-positive breast cancer cell lines was measured at different times. At 72 h estradiol significantly induced growth of MCF7 (Fig. 9A, ) and T47D cells (Fig. 9B, ) but did not promote growth of tamoxifen-resistant MCF7-LCC2 cells (data not shown) as previously reported (50). F/I alone (10 µM) decreased growth of MCF7 (Fig. 9A, ) and T47D cells (Fig. 9B, ) but had no effect on MCF7-LCC2 cells (data not shown). Similar results were observed at lower concentrations of F/I (1 and 5 µM) for MCF-7 cells (data not shown). Remarkably, F/I significantly inhibited the estradiol-dependent growth of MCF7 and T47D cells (
in Fig. 9, A and B, respectively).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
(18), although activated PKA also inhibited breast cancer cell growth (28, 52) and estrogen-induced cell growth (29). The present study investigated F/I-activated signaling pathways leading to regulation of ER transcriptional activity in breast cancer cells. F/I induced ligand-independent activation of ER but decreased estradiol activation of ER-dependent reporter genes specifically through PKA signaling. In contrast to general effects on reporter genes, F/I had gene-selective effects on endogenous ER-regulated transcription in which F/I induced PR, cMyc, and pS2 but repressed cyclin D1 expression. Associated with the gene-selective effects of F/I was decreased estradiol binding, increased ER association with Hsp90, rapid recruitment of ER to cMyc and pS2 promoters, and a unique ER phosphorylation profile that was distinct from basal and estrogen-induced phosphorylation. F/I and activation of PKA suppressed breast cancer cell growth and inhibited estradiol-dependent cell growth in accordance with a previous study showing that inhibition of PKA signaling promoted estradiol-dependent cell growth (29). Possibly, the inhibition of cyclin D1 by F/I and PKA is involved in PKA suppression of estradiol-dependent cell growth.
Elevation of cAMP by F/I resulted in ER ligand-independent activation in MCF7, T47D, and MCF7-LCC2 breast cancer cell lines but, unexpectedly, F/I decreased ER activation by estradiol for selected genes. Dissecting the cAMP effect on estradiol-dependent reporter activity revealed that the PKA signaling pathway attenuated estradiol-dependent reporter activity in all cell lines and also reversed tamoxifen resistance in MCF7-LCC2 cells. Unlike PKA, activating Epac signaling (20) by 8CPT-2Me-cAMP treatment had no significant effect on ER activation and only partially reversed tamoxifen resistance in MCF7-LCC2 cells. To our knowledge, this is the first report that separates the Epac pathway from effects on ER signaling in breast cancer cells after elevation of intracellular cAMP.
Previous studies demonstrated conflicting results for PKA effects on ER-dependent reporters. Inhibition of PKA by H-89 significantly stimulated ligand-independent activation of ER and enhanced estradiol-dependent transcriptional activity in MCF7 cells stably transfected with an ERE-luc reporter (43). In contrast, Katzenellenbogen and co-workers (23) reported that overexpression of PKA or treatment with cholera toxin (CT)/3-isobutyl-1-methylxanthine (I) induced estradiol-dependent activity of (ERE)2-TATA-CAT but not ERE-TK-CAT. In addition, PKA or CT/I treatment altered tamoxifen antagonism but had no significant effect on ligand-independent activation of ER in MCF7 cells. Data presented here using endogenous ER-regulated genes regulated by different promoters provide evidence that the conflicting data with reporter genes are likely related to the gene-specific effects of F/I.
Although F/I attenuated estradiol activation of reporter genes, F/I demonstrated a selective activation of endogenous genes containing complex promoters. F/I had no effect on estradiol-dependent transcription of PR and exhibited an additive effect on estradiol-dependent transcription of cMyc and pS2 mRNA. Similar results with pS2 were reported by El Tanani and Green (53) in which elevation of cAMP by CT/I exhibited an additive effect on estradiol-dependent transcription of the pS2 and LIV-1 genes in MCF7 cells. In the present study, F/I induced cyclin D1 mRNA at early time points similar to results with estradiol treatment. However, at later time points F/I repressed estradiol induction of cyclin D1 mRNA with the end result being suppression of estradiol-induced cyclin D1 protein. The differences in the ER-regulated promoter sequences in each of these genes may provide clues to the F/I-selective repression of cyclin D1 but not cMyc, PR, or pS2. (The role of specific ERE sequences in ER promoter interaction is reviewed in Ref. 54 .) Estradiol regulates cMyc transcription through ERE and AP-1 sequences (38), PR transcription through ERE, AP-1, and SP1 sequences (55, 56), and pS2 transcription through ERE, SP1, and SP3 sequences (37, 57). In contrast, although cyclin D1 contains a partial and imperfect ERE (58), ER regulates cyclin D1 predominantly through a cAMP response element (CRE) in MCF7 cells (31). In addition, Nadella and Kirschner (59) provided clear genetic evidence that cyclin D1 expression was regulated by PKA signaling. ATF-2 binding to the CRE is crucial for activation of the cyclin D1 promoter, and it is possible that F/I treatment disrupts this interaction. Although Sabbah et al. (31) reported that forskolin had no effect on cyclin D1 promoter activity in MCF7 cells, the effect of forskolin or F/I on ATF-2 binding to the CRE was not investigated. It was reported that PKA inhibits the cyclin D1 promoter by phosphorylation of CREB that binds to the CRE (32). CREB may compete with ATF-2 for binding to the CRE and prevent cyclin D1 activation by ER.
There are likely both direct and indirect effects of F/I on ER. F/I induction of ER target genes was ER dependent, i.e. both 4-hydroxytamoxifen or ICI, 182,780 were able to reverse F/I induction of ER target genes. F/I reduced ER ligand binding and increased association of ER with Hsp90. Some of these effects of F/I on ER function are likely related to direct effects of F/I on altering ER phosphorylation at serines 305 and 118. However, these data do not exclude the possibility that F/I regulates ER function through indirect mechanisms.
This is the first report to demonstrate ligand-independent recruitment of ER to pS2 and cMyc promoters induced by F/I. Recruitment of ER by F/I was rapid and, similar to estradiol, occurred in a cyclic fashion. However, a clear correlation between F/I recruitment of ER to promoters and levels of mRNA induction could not be made. It will be of interest to profile the proteins associated with ER at each promoter during both ligand-independent activation and ligand-dependent activation to determine whether different protein complexes contribute to selective gene activation by F/I and PKA signaling (currently under investigation).
It is widely accepted that ligand binding is regulated by both Hsp90 association and nuclear receptor phosphorylation (4, 41, 60, 61, 62). In the absence of ligand, Hsp90 interacts with the ligand-binding domain of ER (63) and dissociates upon ligand binding, leading to tight association of ER with the nuclear compartment (60). In the present study, F/I increased Hsp90 association with ER, which was accompanied with a decrease in ligand binding. As shown in HeLa and MCF7 cells, the F/I effect on ligand binding was correlated with Hsp90 association with ER. This suggests that F/I induces conformational changes in ER and/or Hsp90 that decrease estradiol binding and increase Hsp90 association (the effect of estradiol or F/I + estradiol on Hsp90 association with ER could not be examined because estradiol results in tight association of ER with nuclear structures, thereby preventing extraction for coimmunoprecipitation assays). Remarkably, decreased estradiol binding and increased association of ER with Hsp90 by F/I did not prevent ER recruitment to promoters, suggesting that an ER-Hsp90 complex may be present at the promoter as previously suggested for the GR-Hsp90 complex (64).
F/I, estradiol, and F/I + estradiol resulted in three distinct ER phosphorylation profiles. In general, ligand-independent activity of ER results from changes in AF-1 (A/B) domain phosphorylation (4), S118 phosphorylation, a MAPK site that is enhanced by estradiol (2), is required for full ER transcriptional activity by estradiol (13, 17). S118 phosphorylation was dramatically reduced by F/I and F/I + estradiol compared with vehicle. It is likely that activation of PKA by F/I in MCF7 cells blocks MAPK phosphorylation. In this context, PKA suppression of MAPK could also reduce steroid receptor coactivator 1 regulation of ER (46) because steroid receptor coactivator 1 contains several MAPK consensus sites that are required for full coactivator activity (63, 65). S305 is contained within a PKA consensus sequence located at the interface of the hinge (D) domain and the ligand binding (E/F) domain and within the nuclear localization region (63). A previous study showed that forskolin results in phosphorylation of S305 in U2OS osteosarcoma cells (26) and Cos-1 cells transfected with wild-type ER (42). The present study demonstrates the involvement of S305 in the ligand-independent activity as well as in repressing estradiol-dependent transcription of ERE-luciferase-containing promoter. The dual effect of PKA on ER phosphorylation in MCF7 cells (induction of S305 and suppression of S118 phosphorylation) may explain both the ligand-independent activity (induction of S305) and suppression of estradiol-dependent activity (suppression of S118), at least for the simple promoters present in the reporter genes.
Interestingly, F/I did not induce phosphorylation of S236, a PKA consensus site (15), but instead modestly inhibited estradiol-dependent phosphorylation of S236 as well as Y537. S236 phosphorylation prevents ER dimerization in the absence of estradiol (15). Y537 phosphorylation regulates ligand binding and transcriptional activity (60, 61). Similar to suppression of S118 phosphorylation, F/I inhibition of S236 and Y537 would be consistent with the reduced estradiol binding and reduced ligand-dependent activation observed here. A similar effect of forskolin was reported by Blok et al. (66) where forskolin blocked androgen binding to its cognate receptor by dephosphorylating androgen receptor. Remarkably, F/I in combination with estradiol resulted in a unique and distinct phosphorylation profile compared with estradiol or F/I alone (phosphorylation of S104, S106, S305, and T311). S104 and S106 can be phosphorylated by estradiol binding or by cross talk with cyclin A/cyclin-dependent kinase 2 signaling (67), and these sites potentiate estradiol action and induce ligand-independent activation of ER (67). T311 is phosphorylated by a P38/SAPK2a complex and regulates ER nuclear localization (68).
It is widely believed that ER posttranslational modifications, including phosphorylation, affect chromatin remodeling and ER promoter occupancy (69). Therefore, the differences in ER phosphorylation profiles represented by the three different treatment regimens (estradiol, F/I, or estradiol + F/I) likely result in distinct receptor conformations that impact the dynamic association of ER with ER target promoters. F/I regulated dual phosphorylation events (induction of S305 phosphorylation and repression of S118 phosphorylation) that were correlated with rapid recruitment of ER to pS2 and cMyc promoters. Recently, Kumar and co-workers (70) found that mutation of S118 to alanine (mimic dephosphorylation) and mutation of S305 to glutamic acid (mimic phosphorylation) inhibited estradiol-dependent transcription. This dual mutant is very similar to the ER phosphorylation pattern induced by F/I and reproduces the F/I suppression of estradiol activation of the reporter. F/I + estradiol treatment resulted in three additional phosphorylation events (phosphorylation of S104, S106, and T311) and was associated with a delay in ER promoter recruitment compared with F/I alone, whereas estradiol alone resulted in phosphorylation of eight ER phosphorylation sites and was associated with further delay in promoter recruitment.
F/I alone inhibited growth of ER-positive, tamoxifen-sensitive MCF7 and T47D breast cancer cells but not ER positive, tamoxifen-resistant MCF7-LCC2 cells. Furthermore, F/I blocked estradiol-dependent growth of tamoxifen-sensitive breast cancer cells. Previous studies reported an antiproliferative effect of cAMP in MCF7 cells in which ER contributed to cAMP levels and activity (28, 29), and ER predicted the antiproliferative effect of cAMP (29). However, activated cAMP also suppressed growth of ER-negative MDA-MB-231 breast cancer cells (28), suggesting an ER-independent pathway for cAMP growth suppression. de Cremoux and colleagues (71) characterized tamoxifen-resistant sublines of MCF7 cells and found that both ER and PR mRNA were decreased and ERß mRNA was increased in MCF7-LCC2 cells compared with parental, tamoxifen-sensitive MCF7 cells. In addition, the differences in the bcl-2/bax ratio and overexpression of TGFß in MCF7-LCC2 cells possibly contributed to resistance (72). It is likely that combined changes in ER expression, TGFß expression, and the bcl-2/bax ratio contribute to the inability of F/I to have an antiproliferative effect on MCF7-LCC2 cells.
It is likely that the inhibition of cyclin D1 by F/I contributes to suppression of estradiol-dependent growth of breast cancer cells. Typically, estradiol regulates cell cycle progression from the G0–G1 phase by induction of cMyc and cyclin D1 followed by activation of Cdk4 and Cdc25A enzymatic activity in a time-dependent manner (73). It is possible that the antiproliferative effect of PKA is mediated, in part, by inhibition of cyclin D1, which in turn blocks cells at G0–G1. F/I induction of cMyc alone would not be sufficient for cells to progress through the G1 phase.
A model for estradiol and F/I effects on ER activation in breast cancer cells is presented in Fig. 10. Estradiol treatment dissociated the Hsp90-ER complex and induced all ER phosphorylation sites, resulting in transcription of reporter genes and endogenous genes (cMyc, pS2, PR, and cyclin D1) and breast cancer cell growth (Fig. 10A). In contrast, F/I (Fig. 10B) increased Hsp90-ER association, decreased estradiol binding, induced S305, and repressed S118 phosphorylation. This was accompanied by ligand-independent activation of ER and decrease of estradiol-dependent reporter activity in a PKA-dependent manner. However, the F/I effect on estradiol-dependent transcription was gene specific, resulting in inhibition of cyclin D1 expression and inhibition of breast cancer cell growth. Cross talk between PKA and estradiol signaling (Fig. 10C) resulted in a unique phosphorylation profile and also inhibited estradiol induction of cyclin D1 expression that may inhibit estradiol-dependent growth of breast cancer cells (29).
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
(D12), and actin (C-11) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). ER-P-S104, ER-P-S106, ER-P-S118, ER-P-S167, ER-P-S236, ER-P-S305, ER-P-T311, and ER-P-Y537 antibodies were provided by Bethyl Laboratories (Montgomery, TX). ER (VP-E613) antibody, peroxidase-labeled antimouse and peroxidase-labeled antirabbit were purchased from Vector Laboratories (Burlingame, CA). ERß (14C8) antibody was purchased from GeneTex (San Antonio, TX). Protein A Sepharose beads (CL-4B) were purchased from Amersham Biosciences (Piscataway, NJ). Protein A magnetic beads were purchased from New England Biolabs (Ipswich, MA).
Cell Culture
MCF7 ER-positive tamoxifen-sensitive breast cancer cells and HeLa cells were maintained in phenol red-free DMEM containing 10% FBS, 1% Penicillin/streptomycin, and 2% glutamine. T47D ER-positive tamoxifen-sensitive breast cancer cells were maintained in phenol red-free RPMI media containing 10% FBS, 1% Penicillin/streptomycin, and 5 µg/ml insulin. MCF7-LCC2 ER-positive tamoxifen-resistance breast cancer cells were maintained in phenol red-free incomplete MEM containing 5% stripped FBS (FBS were incubated with dextran-coated charcoal to remove steroids). All cells were incubated at 37 C and 5% CO2.
Luciferase Assay
MCF7, T47D, MCF7-LCC2, or HeLa cells (0.5 x 106) were plated in each well of six-well plates in medium containing 5% stripped FBS for 48 h. Cells were transfected with ERE2e1b-firefly luciferase, ERE2TK-firefly luciferase, or CRE-firefly luciferase by using Fugene 6 (Roche Diagnostics, Indianapolis, IN) for 24 h. In experiments in which overexpression of PKA, PKI, wild-type ER, or ERS305A was performed, ERS305A (obtained from Dr. Rakesh Kumar, University of Texas M.D. Anderson Cancer Center), pFC-PKA, RSV-PKI (obtained from Dr. Mario Ascoli, University of Iowa), ER-pCR3.1, or empty vector were transfected along with the reporter constructs. Cells were incubated with vehicle, 17ß-estradiol (10–8 M), 4-hydroxytamoxifen (4-(OH)-Tam) (10–7 M), forskolin (10 µM) + IBMX (100 µM) (F/I), or 8CPT-2Me-cAMP for 18 h. Media were replaced with 400 µl of 1x sodium dodecyl sulfate (SDS) lysis buffer and incubated for 1 h at –70 C, and firefly luciferase activity was measured using the Luciferase Assay System Kit from Promega Corp. (Madison, WI). Relative luciferase units were normalized to the total protein content of each sample. Constitutive Renilla luciferase [1-phRL-null reporter vector (Promega)] and pCMV-ß-gal reporters were markedly activated by F/I treatment and/or overexpression of PKA and PKI precluding normalization of ERE2e1b-firefly luciferase and ERE2TK-firefly luciferase assays to transfection efficiency (data not shown).
Real-Time RT-PCR
Real-time RT-PCR was performed as previously described by our laboratory (13, 74). Briefly, MCF7 cells (4 x 106) were plated in 10-cm dishes in phenol red-free DMEM containing 5% stripped FBS for 3 d followed by incubation with vehicle, estradiol (10–8 M), F/I (10 µM/100 µM) (F/I), 4-(OH)-Tam (10–7 M), or combination for 2, 6, 12, 18, or 24 h. Cells were harvested and total RNA was extracted using the Trizol (Invitrogen, Carlsbad, CA)/chloroform method. Total mRNA was adjusted to 200 ng/reaction and reverse transcribed using the TaqMan Reverse Transcription Kit (Applied Biosystems, Foster City, CA). cDNA was measured by real-time RT-PCR as previously described (13, 74, 75). The following primers and probes were used.
Progesterone receptor (PR): forward (FWD), 5'-CTATGCA-GGACATGACAACACAAA-3',
Reverse (REV), 5'-TGCCTCTCGCCTAGTTGATTAAG-3',
Probe-5'/56FAM/CCTGACACCTCCAGTTCTTTGCTGACAAG/3BHQ-1/-3'
Glyceraldehyde-3-phosphate dehydrogenase (GAPDH): FWD, 5'-CAACGGATTTGGTGGTATTGG-3',
REV, 5'-GCAACAATATCCACTTTACCAGAGTT-3',
Probe-5'/56FAM/CGCCTGGTCACCAGGGCTGCT/3BHQ-1/-3'
cMyc: FWD, 5'-CGTCTCCACACATCAGCACAA-3',
REV, 5'-TCTTGGCAGCAGGATAGTCCTT-3',
Probe-5'/56FAM/ACGCAGCGCCTCCCTCCACTC/3BHQ-1/-3'
pS2: FWD, 5'-CGTGAAAGACAGAATTGTGGTTTT-3'
REV, 5'-CGTCGAAACAGCAGCCCTTA-3'
Probe-5'/56FAM/TGTCACGCCCTCCCAGTGTGCA/3BHQ-1/-3'
Cyclin D1: FWD, 5'-TGGGTCTGTGCATTTCTGGTT-3'
REV, 5'-GCTGGAAACATGCCGGTTAC-3'
Probe-5'/56FAM/CGGCGCTTCCCAGCACCAA/3BHQ-1/-3'
The CT value of the gene of interest was normalized to GAPDH and calibrated to vehicle treatment CT value to obtain the fold difference.
Western Blot
MCF7 cells (4 x 106) were plated in 10-cm dishes in phenol red-free DMEM containing 5% charcoal-stripped FBS for 3 d. Fresh medium was replaced before treatment, and cells were incubated with vehicle, estradiol (10–8 M), 4-(OH)-Tam (10–7 M), F/I (10 µM/100 µM) (F/I), or a combination of F/I + ligand for 2 or 24 h. To detect ERK1/ERK2 phosphorylation, cells were either transfected with empty vector or PKA expression vector, or incubated with F/I for 5, 10, 15, or 30 min. Cells were harvested in PBS and cell pellets were lysed with high-salt lysis buffer (0.4 M NaCl; 5 mM sodium fluoride; 10 mM Tris, pH 8; 2 mM EDTA; 2 mM EGTA; 1 mM sodium orthovanadate; 0.1% Triton X-100, 10 mM ß-mercaptoethanol, 5 mM ß-glycerophosphate, protease inhibitor mix; 0.1 mM PMSF) for 10 min on ice, and cellular debris was removed by centrifugation at 20,000 x g for 10 min. SDS loading buffer was added to the total cellular extract, and samples were boiled for 2 min. Proteins were separated by 10% SDS-PAGE analysis and transferred to nitrocellulose membrane. Protein expression and ER phosphorylation were determined by Western blotting with specific antibodies listed above, and expression signals were obtained by enhanced chemiluminescence. Protein expression was normalized to ß-actin levels. ER and ERK1/ERK2 phosphorylation was normalized to the total ER and ERK1/ERK2 protein expression, respectively. Relative quantification of ER and ERß in MCF-7 crude cell extract by Western blot was determined by comparison of signal intensity to a known amount of purified ER and ERß.
Quantitative ChIP Assay
Quantitative ChIP assays were applied as previously described by our laboratory with little modification (13, 76). MCF7 cells (10 x 106) were plated in 15-cm dishes in phenol red-free DMEM containing 5% stripped FBS for 3 d. Medium was replaced before cells were incubated with vehicle, estradiol (10–8 M), F/I (10 µM/100 µM), or a combination of estradiol and F/I for 45 min, 1 h, 1.5 h, and 2 h. Chromatin and proteins were cross-linked with formaldehyde (final concentration, 1%) for 15 min at 37 C. Cells were washed with PBS, harvested, and lysed in SDS lysis buffer (1% SDS; 10 mM EDTA, pH 8.0; 50 mM Tris-HCl, pH 8.0) that contained protease inhibitors and phosphatase inhibitors. Cells were incubated in lysis buffer for 10 min followed by sonication on ice (four times, 10 sec each, with 40-sec intervals). Cellular debris was removed by centrifugation of samples at 20,000 x g rpm for 10 min, and the total cellular extract was diluted 10-fold in ChIP dilution buffer (0.01% SDS; 1.1% Triton X-100; 1.2 mM EDTA, pH 8.0; 16.7 mM Tris-HCl, pH 8.0; 167 mM NaCl) that contained protease inhibitors and phosphatase inhibitors. Ten percent of the diluted samples were frozen at –70 C for use as input. Samples were precleared with 2 µg of normal mouse IgG (Santa Cruz), 3 µg of sheared salmon sperm DNA, and 100 µl of protein A Sepharose beads in PBS (1:1) for 2 h at 2 C. Beads were removed by centrifugation at 1500 x g for 5 min, and samples were incubated with 2 µg of ER (D-12) antibody for 18 h at 4 C with rotation. Negative control samples were incubated with normal mouse IgG. Sheared salmon sperm DNA (3 µg) and Protein A magnetic beads (5 µl) were added, and samples were incubated for 2 h at 4 C with rotation. Beads were separated and washed twice with ChIP washing buffer (0.1% SDS; 1% Triton X-100; 2 mM EDTA, pH 8.0; 20 mM Tris-HCl, pH 8.0; 150 mM NaCl) and once with TE buffer (10 mM Tris-HCl, pH 8.0; and 1 mM EDTA, pH 8.0). Protein-chromatin complexes were eluted and reverse cross-linked by incubating beads with elution buffer (1% SDS, 0.1 mM NaHCO3) at 65 C for 12 h (input samples were also incubated at this step). The DNA was purified with QIAquick PCR purification kit (QIAGEN, Chatsworth, CA) and resuspended in 50 µl of ddH2O. ChIP was measured by applying validated primers and probes for real-time RT-PCR for pS2 and cMyc as previously described (13). The CT value of the sequence of interest was normalized to input samples and calibrated to the vehicle treatment CT value to obtain the fold difference.
Ligand Binding in Intact Cells
MCF7, MCF7-LCC2, T47D cells (0.5 x 106/well) were plated in six-well plates for 3 d in medium containing charcoal-stripped FBS. Medium was replaced and cells were pretreated with vehicle or F/I for 1 h. Cells were incubated with 1.5 nM [2,4,6,7-3H(N)]estradiol, (20 µCi; PerkinElmer, Boston, MA) in the presence or absence of 1.5 µM of estradiol (cold) for 1 h. Cells were washed and incubated with 700 µl of absolute ethanol for 20 min at room temperature. Samples (500 µl) were collected, and radioactivity was measured and normalized to the total protein amount. HeLa cells were plated in 10-cm dishes in phenol red-free DMEM containing stripped FBS for 2 d, cells were transfected with ER (500 µg) for 24 h, harvested, and replated (1 x 106) in six-well plates for 1 d. Ligand binding was performed as described above.
In Vitro Ligand Binding
Purified ER (500 ng) was preincubated with F/I, vehicle, or ICI 182,780 (10–9 M) for 1 h in TESH buffer (10 mM Tris, 1.5 mM EDTA, 1 mM dithiothreitol, pH 7.4) on ice. [2,4,6,7-3H(N)]Estradiol (1 µM) was added with or without 10 mM estradiol (cold), and samples were incubated on ice for 1 h. Dextran-coated charcoal (DCC) slurry (1.0 g charcoal, 0.05 g dextran T-70 in 100 ml of TESH buffer) was added for 30 sec and DCC was separated by immediate centrifugation at 5000 x g for 5 min at 4 C. Supernatant was collected and radioactivity was measured.
Immunoprecipitation of ERß
MCF7 cells were plated in 15-cm dishes in phenol red-free DMEM containing 5% charcoal-stripped FBS for 3 d. Cells were washed with PBS and harvested, and cell pellets were lysed with high-salt lysis buffer for 15 min on ice. Cellular debris was removed by centrifugation at 20,000 x g for 10 min. Protein A Sepharose beads were incubated with ERß antibody for 2 h at room temperature and washed twice with PBS. Total cellular extract was diluted with PBS and incubated with beads for 18 h at 4 C with rotation. Beads were washed twice with PBS, and SDS loading buffer was added to the beads and boiled for 2 min. Immunoprecipitated ERß was detected by Western blot as described above.
Coimmunoprecipitation of ER and Hsp90
MCF7 and HeLa cells were plated in 10-cm dishes in phenol red-free DMEM containing 5% stripped FBS for 2 d. HeLa cells were transfected with human ER (500 µg) for 24 h using Fugene 6. Cells were incubated with vehicle or F/I for 2 h and harvested. Cell pellets were incubated with HEMG buffer [10 mM HEPES, 3 mM EDTA, 20 mM sodium molybdate, 5% glycerol (pH 7.4), containing protease and phosphatase inhibitors] on ice for 10 min and ruptured using a Dounce homogenizer (60 strokes). Debris was removed by centrifugation at 20,000 x g for 30 min at 4 C. ER (D-12) antibody (2 µg) was incubated with protein A Sepharose beads (in PBS) for 3 h at room temperature. Beads were washed three times in PBS and added to the total cellular extract. Samples were incubated at 4 C for 18 h with rotation. Beads were washed with coimmunoprecipitation buffer (10 mM Tris-HCl; 3 mM
, Significant difference compared with estradiol treatment. P < 0.05 was considered significant. Data are representative of results from at least three separate experiments. RLU, Relative luciferase units.
