| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
RECEPTORS |
Department of Biochemistry and Molecular Biology, University of Louisville School of Medicine, Louisville, Kentucky 40292
Address all correspondence and requests for reprints to: Carolyn Klinge, Department of Biochemistry and Molecular Biology, University of Louisville School of Medicine, 319 Abraham Flexner Way, Louisville, Kentucky 40202. E-mail: . carolyn.klinge{at}louisville.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ERß and inhibits E2-induced transcription. We examined how SHP affects tamoxifen’s estrogen agonist activity in endometrial cells. We report that SHP interacts with 4-hydroxytamoxifen (4-OHT) or E2-occupied ER in a temperature-dependent manner in vitro. In transient transfection assays, SHP inhibited 4-OHT-stimulated reporter gene activity from an estrogen response element (ERE) in ER-positive RL95-2 but not in HEC-1A human endometrial carcinoma cells transfected with ER or ERß. SHP inhibited E2-induced transcriptional activity in ER- or ERß-transfected HEC-1A or Chinese hamster ovary-K1 cells. SHP inhibition of E2 activity was greater for ER than ERß from the nonpalindromic ERE in the pS2 gene promoter in Chinese hamster ovary-K1 but not HEC-1A cells. Thus, ER subtype, cell type, and ERE sequence influence SHP repressor activity. An ER mutant lacking activator function-1 showed reduced inhibition by SHP. In glutathione S-transferase pull-down experiments, SHP inhibited ER dimerization, providing a possible mechanism to account for the inhibitory effect of SHP on ER activity. These results identify SHP as novel target for blocking 4-OHT agonist activity in endometrial cells. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
SHP inhibits E2-induced transcriptional activity of ERß more than ER, although there are conflicting results about how the ER ligand affects SHP interaction with ERß (3, 4, 5). In glutathione S-transferase (GST) pull-down experiments, SHP interacts with ER in the presence of E2 but not 4-hydroxytamoxifen (4-OHT) (3, 5). Although the actual mechanisms accounting for estrogen-dependent transactivation by ER and ERß remain to be fully elucidated, interaction of estrogen response element (ERE)-bound, agonist-occupied ER with coactivator proteins (e.g. SRC-1 and components of the TFIID complex) is necessary for enhanced gene expression (6, 7). This suggests that one possible mechanism for the estrogen antagonist activity of SHP could be by blocking ER interaction with coactivators.
Tamoxifen (TAM) is used clinically to prevent breast cancer in high-risk women and block disease recurrence in women with breast cancer (8, 9). Unfortunately, TAM increases the risk of endometrial cancer in women taking the drug (10, 11, 12, 13). TAM and its metabolite 4-OHT have estrogen antagonist activity in breast but agonist activity in uterus and bone (9). The mechanisms for these tissue-specific effects of TAM are not completely understood. The agonist activity of TAM in cultured human endometrial cancer cells is blocked by the "pure steroidal antiestrogen" ICI 164,384 (14), indicating that TAM agonist activity is ER mediated. TAM binds to ER with high affinity in the ligand-binding pocket and prevents the ligand binding domain (LBD) of ER from assuming a conformation commensurate with recruitment of coactivators necessary for transcriptional activation (15). Thus, 4-OHT blocks activator function (AF)-2 function. The agonist activity of TAM is cell specific [e.g. TAM is an agonist in COS-7 cells and in SaOS2 human osteoblast cells (16, 17)] but not in MCF-7 cells (18). Deletion analysis of ER showed that amino acids (aa) 41–64 in the N-terminal AB domain, containing AF-1, are required for the agonist activity of 4-OHT in HEC-1A (19). The agonist activity of TAM is also promoter specific [e.g. TAM is an agonist at AP-1 sites in cell lines of uterine) but not breast origin (20, 21, 22)]. Together these results suggest that cell-specific factors mediate the agonist activity of 4-OHT, perhaps through interaction with AF-1.
We investigated whether SHP blocks the agonist activity of 4-OHT in endometrial cancer cells. We found that SHP inhibits 4-OHT agonist activity in RL95-2 cells. In agreement with earlier reports (17, 23), 4-OHT did not have agonist activity in HEC-1A cells. SHP also inhibited E2-induced ERE-driven reporter activity in RL95-2 and ER-transfected HEC-1A human endometrial carcinoma cells. We also demonstrated that SHP is a more potent inhibitor of ER activity than ERß activity for both EREs. An ER mutant lacking AF-1 showed reduced inhibition by SHP, suggesting a role for ER N-C-terminal interaction in SHP action. Our results indicate that SHP could offer a novel target to ameliorate the agonist activity of TAM in endometrium.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(24) and rat ERß (25, 26, 27) were produced in IPLB-Sf21AE cells. Nuclear extracts of ER and ERß were prepared from IPLB-Sf21AE insect cells infected recombinant baculovirus as described previously (28). The concentration of ER was determined by hydroxyapatite (HAP) assay (29) and is reported as dimer concentration (two moles of E2 per mole of ER).
Reporter plasmids
The sequences of EREc38 and pS2 are: 5'-CCAGGTCAGAGTGACCTGAGCTAAAATAACACATTCAG-3' (30) and 5'-CTTCCCCCTGCAAGGTCAGCGTGGCCACCCCGTGAGCCACT-3' (25), respectively. EREc38, 2EREc38 [i.e. two tandem copies of EREc38 (31) and pS2 (25)] were cloned into the pGL3-pro-luciferase plasmid (Promega Corp., Madison, WI). Plasmids were amplified in Escherichia coli strain DH5 and purified using the Bio-Rad Maxi/Midiprep kit (Bio-Rad Laboratories, Inc., Hercules, CA).
Cell culture and transient transfection
RL95-2, HEC-1A, and Chinese hamster ovary (CHO)-K1 cells were purchased from American Type Culture Collection (Manassas, VA). CHO-K1 cells were maintained in Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 10% newborn calf serum. RL95-2 cells were maintained in MEM with 5 µg/ml insulin. HEC-1A cells were maintained in McCoy’s medium. RL95-2 and HEC-1A cell maintenance media were supplemented with 10% FCS and 1% penicillin (pen)/streptomycin (strep). All cell culture media and reagents were purchased from Life Technologies, Inc. (Grand Island, NY). For transfection, 2.5 x 105 (HEC-1A and RL95-2) or 1.5 x 105 (CHO-K1) cells were plated in each well of a 24-well plate in IMDM (without phenol red, -) supplemented with 10% charcoal-stripped calf serum (CCS) and 1% pen/strep. After 24 h, the cells were transfected using TransFast (Promega Corp.). HEC-1A and RL95-2 cells were cotransfected (per well) with 400 ng reporter and 100 ng pCMV-ß-gal (CLONTECH Laboratories, Palo Alto, CA) per well. HEC-1A were cotransfected with 5 ng pCMV-ER/well. CHO-K1 cells were cotransfected (per well) with 250 ng reporter, 50 ng pCMV-ß-gal, and 5 ng pCMV-rhER (recombinant human ER), pCMV-rhERß, or pSG5-hER-
1–44 kindly provided by Drs. Benita Katzenellenbogen (32), Eckardt Treuter (3), and Farzad Pakdel (33), respectively. Late-passage (P15–18) RL95-2 were cotransfected with 5 ng pCMV-rh ER or pSG5-hER-1–44 plus 250 ng reporter and 50 ng pCMV-ß-gal. The pGL3-EREc38, pGL3–2EREc38 (two tandem copies of EREc38), and pGL3-pS2 luciferase reporter plasmids have been described (25, 31, 34).
Some cells were cotransfected with SHP in the pCDM8 expression vector [pCDM8-mSHP (mouse SHP)], kindly provided by Dr. David D. Moore, Baylor College of Medicine (2, 5). SHP amounts transfected are indicated in the figure legends. In every experiment conducted here, the luciferase activity determined in the presence of a given amount of transfected SHP and each ligand (e.g. E2 or 4-OHT) was divided by the luciferase activity determined in the presence of the identical amount of SHP in cells treated with ethanol (EtOH). In this way, any SHP-induced alteration in basal activity was eliminated from our assessment of SHP’s effects on E2 or 4-OHT activity. The EREc38 and pS2 ERE-luciferase reporters have been described (25, 34). Some cells were cotransfected with the coactivators pSCT-SRA2 and/or pSG5.HA-protein arginine methyltransferase 1 (PRMT1), which were graciously provided by Dr. Bert W. O’Malley (35) and Dr. Michael R. Stallcup (36), respectively. As with SHP, any effects of steroid receptor coactivator (SRA) and PRMT1 on basal (EtOH) activity were used to normalize the data for cells treated with ligand. One hour after transfection, 20% CCS, 2% pen/strep IMDM was added to bring the final concentration to 10% CCS, 1% pen/strep per well. After 24 h, the cells were treated with vehicle, EtOH, or the ligand(s) indicated in the figures. Treatments were performed in triplicate within each experiment. After 30 h of treatment, the cells were lysed and the cleared extract was assayed for luciferase and ß-galactosidase (ß-gal) activities as described (31). Luciferase activity was normalized using ß-gal activity and was expressed as relative light units as a ratio of activity detected in the EtOH controls. The effect of SHP on E2-stimulated reporter activity was normalized by the effect of SHP on basal luciferase activity for each concentration of SHP transfected. Statistical evaluations of data were performed using t test in Microsoft Corp. Excel and ANOVA in GraphPad Software, Inc. (San Diego, CA) Prism. The 50% inhibitory concentration (IC50) values were calculated in GraphPad Software, Inc. Prism.
Middle-passage (P12) RL95-2 cells were stably transfected with pCDNA3-myc-ER or pCDNA3 using LipofectAMINE reagent (Life Technologies, Inc.). Twenty-four hours after transfection, the cells were rinsed with PBS and MEM containing 5 µg/ml insulin plus 1 mg/ml neomycin. Cells were grown continually in 1 mg/ml neomycin for 2 wk to select for stably transfected cells (37) and were rinsed daily with PBS to remove detached cells. Cells growing in neomycin were pooled and transiently transfected as described above except that 5 ng of pRL-CMV (Promega Corp.) were cotransfected with EREc38-luciferase and pCD8-mSHP and cell extracts were assayed using dual luciferase reporter assay (Promega Corp.).
GST pull-down assays
The plasmid directing the expression of a GST fusion protein for mouse SHP was kindly provided by Dr. David D. Moore [Baylor College of Medicine (2, 5)]. The plasmid directing the expression of a GST fusion protein for human ER LBD was kindly provided by Dr. Janet E. Mertz, University of Wisconsin (38). GST-fusion proteins and GST expressed from pGEX-2TK were purified from E. coli BL-21 cells according to protocols (Pharmacia, Piscataway, NJ). The concentrations of the glutathione (GSH)-Sepharose-purified proteins were determined by DC assay (Bio-Rad Laboratories, Inc.). Protein purity was monitored by using Coomassie Blue staining and by Western blot with an anti-GST-antibody (Pharmacia). GST pull-down assays were performed using identical amounts (in micrograms or moles) of purified GST-fusion proteins and baculovirus-expressed recombinant human ER as described (25, 31, 39). Monoclonal anti-ER antibody (Ab10) was purchased from NeoMarkers (Fremont, CA). GST pull-down competition assays were performed as described (40) except that ER was visualized by Western blotting with Ab10. Data were quantitated from scanned films using Un-Scan-It software (Silk Scientific, Inc., Orem, UT).
Western blot
For Western blotting, 2 x 105 CHO-K1 cells/well were plated in a 12-well plate and were transfected with 30 ng ER or ERß mammalian expression vector plus 1.5 µg pCD-M8 (negative control) or 1.5 µg pCD-M8-SHP. This gives the same ratio of ER:SHP as that used in 24 wells used in transient transfection assays in which 250 ng of SHP inhibited the activity of 5 ng ER. Whole-cell extracts (WCEs) were prepared in RIPA buffer (50 mM Tris-HCl, pH 7.4; 1% Nonidet P-40; 0.25% Na-deoxycholate; 150 mM NaCl; 1 mM EDTA; 1 mM PMSF; 1 µg/ml of each: aprotinin, leupeptin, pepstatin; 1 mM Na3VO4; 1 mM NaF). Then 100 µg of WCE were separated on 10% polyacrylamide SDS gels and electroblotted onto a polyvinylidene difluoride (PVDF) membrane (NEN Life Science Products, Boston, MA). The transfer was monitored by prestained MultiMark (Invitrogen, San Diego, CA). Following the transfer, the blots were probed with ER monoclonal antibody Ab10 or ERß polyclonal antiserum PA1-311 (Affinity BioReagents, Inc., Golden, CO). Immunodetection employed Renaissance enhanced luminol reagent (NEN Life Science Products) as described (41). Data were quantitated from scanned films using Un-Scan-It software (Silk Scientific, Inc.; Ref. 41).
[3H]E2 binding assay
Specific [3H]E2 binding to ER in WCEs of RL95-2 cells (100 µg protein) using the HAP assay (29). Binding reactions were performed in triplicate for both total and nonspecific binding (42).
Bacterial expression of SHP
Mouse SHP (mSHP) was expressed from the pT7lac-mSHP plasmid provided by Dr. David Moore (2, 5, 43) and bacterial lysates were prepared from E. coli using CellLytic B-II bacterial cell lysis/extraction reagent (Sigma) per the manufacturer’s instructions. Protein purity was estimated by Coomassie staining of a 10% SDS-PAGE gel.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
and ERß activity in CHO-K1 cells- and ERß-stimulated reporter gene activity from three tandem copies of the Xenopus vitellogenin A2 ERE in U2-OS cells (5) and from 2xERE-tk-luc and 3xERE-tk-luc (i.e. two and three tandem copies of the Xenopus vitellogenin A2 ERE) in transfected 293 cells (3). No one has evaluated how SHP affects E2-stimulated reporter activity from a single ERE or from naturally, imperfect (nonpalindromic) EREs, constructs that more closely resemble the EREs found in most ER-regulated genes. Here we tested the effect of SHP on E2 stimulated reporter gene activity from a single palindromic ERE and the naturally occurring, imperfectly palindromic ERE from the estrogen-regulated pS2 gene (Fig. 1
). For these experiments, cotransfection assays were performed in CHO-K1 cells because this cell line requires exogenous ER or ERß to activate ERE-driven reporter gene expression. Thus, the transcriptional response of each ER isoform in isolation can be evaluated separately with each ERE.
|
than ERß from the palindromic ERE called EREc38 in transiently transfected CHO-K1 cells (Fig. 1A
). Cotransfection with an SHP expression vector resulted in inhibition of E2-induced reporter activity from EREc38 in a concentration-dependent manner with both ER and ERß; however, statistical significance for ERß was reached only at 500 ng SHP (Fig. 1A). Thus, SHP had a greater inhibitory effect on E2-induced activity for ER vs. ERß. SHP had no effect on basal activity from either the pGL3-pro-luciferase vector or the ERE-luciferase reporter (data not shown). These data are consistent with the lack of effect of SHP on basal activity from an MMTV-luciferase reporter in transiently transfected US-OS cells (5). In contrast, SHP inhibited basal luciferase reporter activity from five tandem HNF-4 binding sites in transiently transfected HepG2 cells (46) and from three tandem EREs in U2-OS cells transiently transfected with ER or ERß (5).
Consistent with our previous results (25, 41, 47), the pS2 ERE showed lower induction by E2 with either ER or ERß, compared with EREc38 (compare Fig. 1, A and B). ER and ERß bind the pS2 ERE with reduced affinity, compared with EREc38 (48). In general, SHP showed less inhibition of E2-induced activity by either ER or ERß with the pS2 ERE (Fig. 1B), with the exception that 100 and 250 ng SHP inhibited ERß activity more for pS2 ERE than EREc38. ERß was inhibited less by SHP than ER at the pS2 ERE. These results indicated that the sequence of the ERE impacts the repressive activity of SHP with E2-activated ER and ERß in CHO-K1 cells.
To address whether overexpression of SHP impacted the expression of ER or ERß in transfected CHO-K1 cells, Western blots were performed using WCEs (Fig. 1C). Quantitation of the amount of ER and ERß performed in separate transfection experiments is summarized in Fig. 1D. Cotransfection with SHP appeared to increase the amount of ER and ERß detected. Whereas no change in ER protein was detected in SHP-transfected cells regardless of treatment, E2 or 4-OHT decreased the amount of ERß detected in cells cotransfected with SHP. However, the amount of ERß was not different from that in EtOH control of cells not transfected with SHP. In summary, no reduction in ER or ERß expression was detected with SHP transfection, indicating that the repressive effect of SHP on ER-mediated transactivation is not caused by ER down-regulation.
SHP inhibits E2- and 4-OHT-induced reporter activity in RL95-2 cells
SHP showed less inhibitory activity toward endogenous ER in MCF-7 human breast cancer cells, compared with human 293 embryonic kidney cells transfected with ER or ERß (3). To address the effect of cell type and endogenous vs. exogenous ER expression on SHP’s inhibition of ER activity, transient transfection experiments were performed using the EREc38 luciferase reporter in ER-positive (49), estrogen-responsive RL95-2 human endometrial cancer cells (Fig. 2). The data in Fig. 2A indicate that 100 nM 4-OHT has agonist activity similar to 10 nM E2 in RL95-2 cells. Furthermore, 4-OHT does not inhibit E2-agonist activity in RL95-2 cells. Thus, 4-OHT is an ER agonist in RL95-2 cells. Cotransfection of SHP in RL95-2 cells had no effect on basal activity from EREc38-luciferase reporter (Fig. 2B). Next, we investigated the effect of SHP on E2 and 4-OHT agonist activity with the EREc38 and pS2 ERE luciferase reporters in RL95-2 cells (Fig. 2, C and D). SHP inhibited the agonist activity of both E2 and 4-OHT from both EREs in a concentration-dependent manner. SHP inhibited 4-OHT agonist activity more than E2 activity for EREc38, suppressing luciferase activity below basal expression. This indicates an active repressor activity of SHP with 4-OHT-occupied ER on this ERE. For pS2, SHP inhibition was equivalent for E2 and 4-OHT, suggesting an ERE sequence-dependent alteration in SHP-ER interaction.
|
expression in RL95-2 cells, HAP assays were performed to examine specific [3H]E2 binding (Fig. 2E). The results show that treatment of RL95-2 cells with E2 increased [3H]E2 binding and that cotransfection with SHP decreased [3H]E2 binding to control levels. These data indicate E2 increases ER expression in RL95-2 cells and that SHP may inhibit E2-stimulated endogenous ER expression in RL95-2 cells, a result in agreement with the data showing SHP inhibition of ERE-driven reporter activity in these cells. In late-passage RL95-2 cells transfected with ER, higher specific [3H]E2 binding was detected, in agreement with higher ER expression. Treatment of ER-transfected RL95-2 cells with E2 increased [3H]E2 binding and cotransfection with SHP resulted in a further increase in specific [3H]E2 binding, indicating that SHP does not inhibit the expression of ER in RL95-2 cells.
SHP inhibits E2-induced reporter activity in HEC-1A cells transfected with ER
To examine SHP activity in another human endometrial cell line, HEC-1A cells were selected because they were reported to express ER and exhibit estrogen responsiveness (50, 51). However, similar to other published reports (23, 45), HEC-1A cells displayed little endogenous estrogen responsiveness in the absence of transfected ER (Fig. 3A). When transfected with ER, E2 induced EREc38-driven reporter activity in a concentration-dependent manner in HEC-1A cells (Fig. 3A). In contrast to our findings in RL95-2 cells, 4-OHT, at concentrations from 0.1 nM to10 µM, had no agonist activity in ER-transfected HEC-1A cells and inhibited E2-stimulated reporter activity from EREc38 (Fig. 3A). Thus, 4-OHT is an ER antagonist in HEC-1A cells.
|
inhibited E2-stimulated reporter gene activity in a concentration-dependent manner (Fig. 3A). SHP exhibited similar inhibitory activity against E2-ER stimulated luciferase expression from EREc38 in CHO-K1 and HEC-1A cells (compare Figs. 1A and 3A). As anticipated because 4-OHT showed no ER agonist activity in transiently transfected HEC-1A cells, cotransfection with SHP had no effect on 4-OHT activity (data not shown). Similar to results in CHO-K1 cells, overexpression of SHP did not decrease the amount of ER protein detected in WCEs from transiently transfected HEC-1A cells (data not shown).
SHP inhibits E2-induced reporter activity in HEC-1A cells transfected with ERß
Next we tested the effect of SHP on ERß activity in HEC-1A cells. When transfected with ERß, E2 induced EREc38-driven reporter activity in a concentration-dependent manner (Fig. 3B). 4-OHT had no agonist activity in ERß-transfected HEC-1A cells and inhibited E2-stimulated reporter activity from EREc38. Thus, 4-OHT is an ERß antagonist in HEC-1A cells. In contrast to results in transiently transfected HepG2 cells (44), 4-OHT appeared equally antiestrogenic with ERß and ER in HEC-1A cells. In contrast to results with ER (Fig. 3A), SHP inhibited only E2-stimulated ERß activity at the highest concentration transfected (Fig. 3B). These data indicate that SHP is less effective at inhibiting activated ERß than ER in HEC-1A cells as well as in CHO-K1 cells. Similarly, SHP was previously shown to be less efficient at inhibiting ERß vs. ER activity from three tandem copies of a consensus ERE in U2-OS cells (5). Similar to results in CHO-K1 cells, overexpression of SHP did not decrease the amount of ERß protein detected in WCEs from transiently transfected HEC-1A cells (data not shown).
SHP inhibits E2-induced reporter activity from the pS2 ERE in ER-transfected HEC-1A cells
To determine whether SHP inhibits E2-induced activity from the nonpalindromic pS2 gene ERE, HEC-1A cells were cotransfected with ER or ERß, a reporter plasmid bearing the pS2 gene ERE and SHP. In contrast to the results detected in CHO-K1 cells (Fig. 1B), E2 treatment resulted in a similar induction of luciferase activity from the pS2 ERE reporter with both ER and ERß in HEC-1A cells (Fig. 3C). SHP inhibited the E2-induced pS2 ERE reporter activity by ER and ERß in a concentration-dependent manner. In contrast to the greater inhibition of E2-induced ER than ERß activity by SHP in CHO-K1 cells (Fig. 1B), ER and ERß appeared to be equally inhibited by 250- and 500-ng SHP cotransfection in HEC-1A cells. These data indicate that cell type influences ER and ERß transactivation and SHP inhibition at the pS2 ERE.
Deletion of AF-1 does not block SHP inhibition of ER activity
The N-terminal AF-1 function of ER has both constitutive and MAPK-induced activity (52). There are conflicting data regarding the N terminus AF-1 in ERß. Some reports show that the N terminus of ERß has no independent AF-1 activity, but N-terminal residues of ERß interact with AF-2 of ERß (44, 53, 54, 56). However, SRC-1 was shown to activate ERE-driven reporter activity by unoccupied ERß in transiently transfected Cos, HeLa, and 293 T cells, indicating that agonist occupation of the ERß LBD is not necessary for SRC-1 coactivator activity (57). Further experiments showed that SRC-1 interacted directly with aa 104–122 in the N terminus of ERß and that phosphorylation of Ser104 by MAPK increased SRC-1-ERß interaction, providing a molecular basis for ligand-independent activation of ERß by MAPK (57). Residues critical for AF-1 are located between aa 39 and 44 in human ER (33). Because ERß has minimal AF-1 activity (44, 53, 54, 56) and shows less inhibition by SHP, compared with ER [Fig. 1 and (5)], we examined the effect of SHP on transcriptional activity from ER-1–44, a mutant lacking the first 44 N-terminal amino acids and that is devoid of AF-1 activity (33). First, in agreement with a previous report (33), we noted at least 50% lower transcriptional activity for ER-1–44, compared with full-length ER (Fig. 4A), indicating a role for AF-1 in mediating E2-induced transcriptional activity from two EREc38 in CHO-K1 cells. The E2-induced transcriptional activity of ER-1–44 was inhibited by coadministration of 4-OHT, and 4-OHT alone showed no agonist activity with ER-1–44 (Fig. 4A). Therefore, the E2-induced activity detected for ER-1–44 is specific to that receptor.
|
-1–44 activity in transiently transfected CHO-K1 cells (Fig. 4A). By plotting the luciferase assay data as percentage of E2-induced activity, a more direct comparison can be made for SHP’s inhibitory effect on ER vs. ER-1–44 (Fig. 4B). These data show that SHP has less inhibitory activity with ER-1–44 than for ER. Next, we compared the effect of SHP inhibition on inhibition of E2-induced reporter activity from EREc38 with ER, ERß, and ER-1–44; the data were again normalized to the E2-induced activity so that a more direct comparison of the extent of SHP inhibition can be ascertained (Fig. 4C). The 50% inhibitory concentration values for the inhibition of E2 activity from EREc38 in CHO-K1 cells by SHP are 345, 1733, and 651 ng for ER, ERß, and ER-1–44, respectively, indicating that SHP had the least inhibitory effect on ERß.
By an unknown mechanism, RL95-2 lose the endogenous estrogen responsiveness with increased passage number (58). The replication of late-passage RL95-2 (i.e. after the 15th passage) is independent of E2, and transient transfection with an ERE reporter construct fails to generate luciferase activity in response to E2 treatment (data not shown). Loss of estrogen responsiveness could result from a loss of ER expression. To test this hypothesis, late-passage (passage 15–18) RL95-2 cells were transfected with ER and the EREc38-luciferase reporter plasmid and treated with E2 or 4-OHT alone or in combination (Fig. 5A). Transfection of ER restored estrogen responsiveness in this assay. However, in contrast to early-passage RL95-2 cells (Fig. 2), 4-OHT had no estrogen agonist activity in the ER-transfected late-passage RL95-2 cells (Fig. 5A), suggesting a loss of other cell-factors necessary for 4-OHT agonist activity. SHP inhibited the E2-induced ER activity in the transfected late-passage RL95-2 cells; however, the inhibition was less than that detected in the earlier-passage RL95-2 cells (compare Figs. 5A and 2C). The decrease in SHP inhibition in late-passage RL95-2 cells may indicate a reduction in the expression of SHP-interacting factors required for inhibition of ER activity.
|
-1–44 showed lower E2-induced activity than ER (Fig. 4A), when transfected into late-passage RL95-2 cells, the E2-induced activity for ER-1–44 was similar to transfected ER activity (Fig. 5A). This indicates that cell-specific factors may obviate N-terminal differences between ER and ER-1–44 in late-passage RL95-2 cells but not in CHO-K1 cells. These data support the report that ER AF-1 and AF-2 contribute equally to E2-stimulated reporter gene expression in CHO-K1 cells (33). In contrast to cells transfected with ER, 4-OHT exhibited agonist activity when ER-1–44 was transfected into late-passage RL95-2 cells. However, 4-OHT agonist activity in ER-1–44 transfected late-passage RL95-2 cells was reduced, compared with that in estrogen-responsive RL95-2 cells (compare Figs. 2A and 5A). These data are the first demonstration that the first 44 aa containing AF-1 (33), located at the N terminus of ER are not essential for 4-OHT agonist activity at an ERE sequence. SHP inhibited E2 and 4-OHT agonist activity in RL95-2 cells transfected with ER-1–44. The molecular mechanism accounting for the disparity in response to 4-OHT between ER and ER-1–44 is unknown but suggest that unknown factors in late-passage RL95-2 cells selectively inhibit 4-OHT agonist activity by interacting with the N terminus of ER.
A summary of SHP’s effect on endogenous ER activity from early-passage (passage 6–12), estrogen-responsive RL95-2 vs. late-passage RL95-2 transfected with ER or ER-1–44 is provided in Fig. 5B. The data are displayed as percentage of E2 or 4-OHT activity. These normalized data show that SHP inhibition is reduced in the late-passage RL95-2 cells.
To examine whether the observed loss of SHP inhibition could be compensated by overexpression of ER, RL95-2 cells were stably transfected with ER and transiently transfected with SHP and an ERE-luciferase reporter (Fig. 5C). As a control, cells were stably transfected with pCDNA-3 and then transiently transfected with ER, SHP, and ERE-luciferase reporter. The data indicate that overexpression of ER blocks SHP inhibition of E2-induced transcriptional activity in late-passage RL95-2 cells. We suggest that it is the direct interaction of SHP with ER that inhibits ER activity, rather than a decrease in ER protein.
Coactivators partially block SHP inhibition
A previous report showed that coexpression of SHP and the coactivator TIF2 resulted in enhanced E2-mediated activation of ER in a mammalian two-hybrid assay (4). To test whether the coactivators SRA or PRMT can compete with SHP for ER and thereby decrease SHP’s inhibition of E2-activation of ER, we performed transient transfection experiments in early-passage (i.e. estrogen-responsive) RL95-2 cells. SRA was selected because it interacts with SRC-1 and activates ER through AF-1 (35) and PRMT is a secondary coactivator that interacts with ER and coactivators including GRIP1 (36). Both SRA and PRMT stimulated E2-induced activity in RL95-2 cells (Fig. 6). Both SRA and PRMT at least partially relieved SHP’s inhibition of E2 activity in these cells. These data indicate that SHP competes with coactivators SRA and PRMT for activation of ER.
|
with SHP LBD construct in the presence of E2, and 4-OHT was shown to decrease but not abolish SHP-GST-ER interaction (5). These results and those of other investigators (4) indicate that SHP interacts with the E2-occupied ER AF-2 region with higher affinity than the unoccupied or 4-OHT-occupied ER LBD. We examined the effect of incubation temperature and ligand on the interaction of baculovirus-expressed, intact recombinant human ER with GST-SHP (Fig. 7). The rationale for performing the GST pull-down experiments at various temperatures is taken from a paper showing that incubation of nuclear corepressor (NCoR) with GST-RAR
at increasing temperatures (i.e. 20, 30, or 37 C) resulted in decreased ligand-independent interaction of NCoR with GST-RAR and increased RAR antagonist (AGN193109)-dependent interaction of NCoR with GST-RAR (59). As a negative control, neither E2- nor 4-OHT-occupied ER interacted with GSH-Sepharose (Fig. 7C and data not shown). Increasing amounts of unoccupied, E2- or 4-OHT-occupied ER interacted with GST-SHP as the incubation temperature was increased from 21 to 30 and 37 C (Fig. 7). ICI 182,780-occupied ER did not interact with GST-SHP, regardless of incubation temperature (data not shown). These results indicate that ER interacts directly with SHP and that incubation temperatures in the physiological range enhance unoccupied and 4-OHT- and E2-occupied ER interaction with SHP in vitro. We noted that E2-ER showed the highest interaction with GST-SHP at 37 C as measured by immunoblotting (Fig. 7).
|
interaction inhibits ER dimerization in vitro both in vivo and in vitro (3, 4, 5), it was of interest to determine whether SHP affects ER dimerization. GST pull-down competition assays (40) were performed to analyze baculovirus-expressed rhER binding to GST-hER LBD in vitro in the presence or absence of bacterial-expressed SHP (Fig. 8). These competition assays were performed in the presence of fixed amounts of E2- or 4-OHT-occupied ER and GST-hER LBD. Increasing amounts (25–100 µl) of unlabeled, bacterial-expressed SHP were added such that SHP was present from 2- to 8-fold excess, as estimated from Coomassie-stained gels (data not shown). ER interacted with the GST-hER LBD and increasing concentrations of SHP inhibited ER-GST-hER LBD binding, reflecting a competition between rhER and SHP for binding the GST-hER LBD. Addition of bacterial lysate of parental vector transformed bacteria had no affect on ER-GST-hER LBD binding (data not shown). These data suggest that SHP binding to the ER LBD can inhibit ER dimerization.
|
binding in vitro affected E2 ligand binding, HAP assays were performed in the presence of a fixed amount of baculovirus-expressed hER, [3H]E2, and bacterially expressed SHP. In these experiments, SHP inhibited E2-ER binding by 67% at a ratio of SHP:ER of approximately 2:1, based on protein concentrations (Fig. 9). These data suggest that SHP interaction with ER can inhibit E2 binding.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
AF-1 (19) and cell- and promoter-specific factors (17). Selectively blocking the agonist activity of TAM in uterus could be beneficial for women taking TAM.
This is the first report showing that SHP inhibits gene transcription activated by 4-OHT that displays estrogen agonist activity in RL95-2 human endometrial carcinoma cells expressing endogenous ER. Previously, SHP was shown to inhibit 4-OHT agonist activity with ER in U2-OS cells (5), and the authors noted that this result was unexpected based on the minimal interaction of 4-OHT-occupied GST-ER LBD with [35S]met-SHP in vitro.
In rats, the oviduct and uterus express immunoreactive ER but not ERß (68, 69). At the mRNA level, ER is the predominant ER expressed in human glandular epithelial and stromal endometrial cells, although ERß is also expressed (70, 71). Similarly, at the protein level, ER predominates over ERß in normal and malignant endometrial tissues (72). Immunohistochemical localization of ER and ERß revealed that both ER and ERß are expressed in the nuclei of glandular epithelial and stromal human endometrial cells (73). Luminal epithelial cells expressed immunoreactive ERß (73). Interestingly, levels of ER appear to decline with endometrial tumorigenesis (72, 74), and the resulting increased ERß:ER ratio may play a role in the progression of myometrial invasion (74).
We showed that the temperature of incubation of intact ER with GST-SHP had a profound effect on the amount of ER bound. Incubation temperatures of 30 and 37 C increased ER-SHP interaction for unoccupied and 4-OHT- or E2-occupied ER. E2-occupied ER showed greater interaction with GST-SHP, compared with 4-OHT-occupied or unoccupied ER. These results indicated that ER interacts directly with SHP and that incubation temperatures in the physiological range enhance unoccupied and 4-OHT- and E2-occupied ER interaction with SHP in vitro. Thus, incubation temperatures more similar to those in vivo increase ER-SHP interaction in vitro. In contrast, ICI 182,780-occupied ER did not interact with GST-SHP.
We speculated that under physiological conditions in endometrial cells, SHP interacts with unoccupied and E2- or 4-OHT-occupied ER in a manner dependent on the presence or absence of other cellular factors. Such ER-interacting proteins may, as previously suggested (15), compete with SHP for interaction with ER through the coactivator groove formed in the LBD upon activation of AF-2 by an agonist ligand. Another potential mechanism could be that these factors stabilize the SHP-ER interaction when the coactivator groove is occupied by helix 12 [e.g. when ER is occupied by 4-OHT (75)]. The lack of SHP repression of basal activity of the ERE-driven reporters seen here may reflect the lowered occupancy of EREs by ER in the absence of ligand (76) or may reflect a lack of interaction of SHP with basal transcription factors. Furthermore, in the absence of ligand, ER may be in a chaperonin complex (77, 78, 79) that may block SHP-ER interaction. The importance of cellular factors in mediating SHP’s inhibitory effects was demonstrated by the decrease in SHP’s inhibition of E2-activated ER activity in RL95-2 cells with increasing passage number.
The RL95-2 cell line was derived from a grade 2 moderately differentiated adenosquamous carcinoma of the endometrium from a 65-yr-old woman (80). The cell line was originally characterized as ER positive (4500–5000 ER molecules/cell), glandular epithelial, and nonstromal (80). Karyotypically RL95-2 is trisomic 8 (47,XX,+8) (80). Although some studies reported that RL95-2 cells were not estrogen responsive unless transfected with ER (58, 81, 82), another study showed that early-passage RL95-2 expressed ER and that ER expression is lost with passage (83).
The observed decrease in SHP’s inhibition of E2-activated ER activity in RL95-2 cells with increasing passage number was not because of loss of ER expression in RL95-2 alone because we performed experiments with cotransfected ER using an amount of ER expression vector that gave the maximal estrogenic response as measured by ERE-driven luciferase activity. It is possible that SHP interacts with corepressor complexes (84, 85), a possibility reflected by the suppression of transcription below basal levels in RL95-2 cells (Fig. 2C).
SHP and dosage-sensitive sex-adrenal hypoplasia congenita-critical region on the X chromosome, gene 1 (86) differ from other orphan receptors because, although they contain putative LBDs, they lack DNA-binding regions (2). SHP is expressed in a variety of tissues including uterus (3) and inhibits the activities of RAR, RXR, and TR (2) as well as ER (5) and ERß (3, 4). Mechanisms accounting for the inhibitory effect of SHP include: 1) inhibition of DNA binding by nuclear receptor dimers; 2) recruiting corepressors, although not NCoR (2); and 3) blocking coactivator interaction (4).
A previous report showed that mutations within helix 12 in the LBD, mutations that result in an incomplete AF-2 surface, abolish ER interaction with SHP (4). These results along with the data presented here indicate that SHP interacts with ER through the LBD. Our data showing that coactivators SRA and PRMT at least partially blocked SHP’s ability to inhibit E2 activity in RL95-2 cells indicate that competition between SHP and coactivators for ER account for SHP’s estrogen antagonist activity. The data showing that deletion of the first 44 aa of the N terminus of ER, containing AF-1 (33), decreased SHP’s inhibitory effect, indicate a role for N-C-terminal interactions in ER-SHP interaction.
As reported previously using U2-OS cells transfected with a luciferase reporter containing three tandem EREs (5), we showed that at either a single, palindromic ERE or the imperfect pS2 ERE, ERß was less inhibited by SHP than ER in transfected CHO cells. In contrast, in HEC-1A cells, similar experiments showed both less inhibition by SHP and approximately equal inhibition of ER and ERß by SHP with either EREc38 or the pS2 ERE. These results support a report showing that ER AF-1 and AF-2 are equally active in RL95-2 cells (33). A chimeric ER/ß construct containing the N terminus of ER fused to the DNA binding and LBD of ERß showed reduced activity in HEC-1A cells, compared with the identical construct in CHO cells (53), suggesting the absence of cellular factors involved in transcriptional activation through interaction with the N terminus of ER from HEC-1A cells.
Because ER and ERß differ substantially in their N termini (87), we examined whether deletion of the first 44 aa of the N terminus of ER, containing AF-1 (33), would decrease the inhibitory effect of SHP. Our results showing a decrease in SHP inhibition of ER-1–44, compared with ER, support a role for the N terminus of ER in mediating responsiveness to SHP’s repressor activity.
Because AF-1 has been reported to mediate TAM agonist activity (88) and because a recent report demonstrated N-C-terminal interaction in ER that impacts both AF-1 and AF-2 functions (33), we speculated that SHP may block AF-1 from interacting with coactivators that mitigate 4-OHT agonist activity. Indeed, our data showing that the AF-1-specific ER coactivator SRA (35) is able to partially restore E2-dependent transcription, fit this hypothesis. Another coactivator that interacts with AF-1 of ER and enhances the transcriptional activity of E2- or 4-OHT-occupied ER is p68 (89). Therefore, it is possible that overexpression of p68 would likewise decrease AF-1 blockade by SHP.
In addition to SHP’s possible impact on AF-1 function, we showed that SHP inhibited ER dimerization in a GST pull-down assay. This offers an additional explanation for the ability of SHP to repress both E2- and 4-OHT-stimulated ER activity because ER dimerization is a requisite for ERE binding and transcriptional activity (90, 91, 92, 93). We also reported that SHP inhibits E2-ER binding in vitro, suggesting that SHP interaction with the ER LBD may interfere with ligand binding.
Our results also showed that the effectiveness of SHP as repressor of E2-ER activity also depends on the ERE sequence in CHO-K1 cells. SHP was more effective at inhibiting E2-induced ER activity from a palindromic ERE than from the imperfect ERE from the pS2 gene promoter in CHO-K1 cells. In contrast, in HEC-1A cells, SHP was equally effective at repressing E2 activity for ER and ERß from EREc38 and the pS2 ERE. These results are further indication that cell-specific factors mitigate SHP’s ER repressor activity.
The estrogen responsiveness of HEC-1A cells appears to differ between laboratories. HEC-1A cells were reported to be ER negative (94), and neither E2 nor TAM affected the proliferation of HEC-1A cells (23). We observed very low endogenous ER activity using an ERE-driven reporter assay of E2-treated, transiently transfected HEC-1A cells. However, in agreement with reports by other investigators (50, 51), HEC-1A cells transfected with ER were estrogen responsive. Our observation that 4-OHT acts as an ER antagonist in HEC-1A cells contradicts reports showing that TAM or 4-OHT act as an ER agonists in transfected HEC-1A cells (45, 50, 51) but agrees with other reports showing no agonist activity of TAM or 4-OHT in HEC-1A cells (17, 23).
SHP is known to have direct repressor activity as well as preventing the interaction of coactivators with DNA-bound nuclear receptors (46), thus offering a molecular explanation for our observations: The concentration and identity of corepressors and coactivators are likely different between CHO-K1 and HEC-1A cells. We and others envision a coregulator assembly exchange process between coactivators and corepressors for ER to activate or repress gene transcription (95).
In summary, our results identify SHP as an inhibitor of 4-OHT agonist activity in RL95-2 human endometrial carcinoma cells that express endogenous ER. We conclude that SHP does not decrease ER expression, but rather it is the direct interaction of SHP with ER that inhibits ER transcriptional activity. In addition, we show that SHP inhibits E2-induced activity from a single palindromic or natural, nonpalindromic ERE from the pS2 gene promoter in CHO-K1 or HEC-1A cells transfected with ER or ERß. We report that SHP displays a greater inhibitory effect on ER than ERß. These results provide the basis for further experiments addressing ways to selectively inhibit 4-OHT agonist activity by SHP in endometrial cells. Selective inhibition of 4-OHT agonist activity in human endometria including cellular proliferation, and transcription of endogenous estrogen target genes would be desirable in women taking tamoxifen to prevent or treat breast cancer who are at increased risk of developing endometrial cancer (11).
|
Acknowledgments |
|---|
and ERß and Timothy L. Ramsey for cloning myc-ER into pCDNA3. We thank Loretta Doan, Jennie E. Lee, and Dr. Padmaja B. Nair Thomas for their assistance in some of the experiments reported here. We thank Dr. Barbara J. Clark for her thoughtful suggestions on this manuscript. |
Footnotes |
|---|
Abbreviations: aa, Amino acids; Ab10, monoclonal anti-ER antibody; AF, activator function; CCS, charcoal-stripped calf serum; CHO, Chinese hamster ovary; CMV, cytomegalovirus; ERE, estrogen response element; EtOH, ethanol; ß-gal, ß-galactosidase; GSH, glutathione; GST, glutathione S-transferase; HAP, hydroxyapatite; IMDM, Iscove’s modified Dulbecco’s medium; LBD, ligand binding domain; mSHP, mouse SHP; NCoR, nuclear corepressor; 4-OHT, 4-hydroxytamoxifen; pen, penicillin; PRMT1, protein arginine methyltransferase 1; PVDF, polyvinylidene difluoride; rhER, recombinant human ER; SHP, short heterodimer partner; SRA, steroid receptor coactivator; strep, streptomycin; TAM, tamoxifen; WCE, whole-cell extract.
Received July 9, 2001.
Accepted for publication November 6, 2001.
