Transrepression of Estrogen Receptor β Signaling by Nuclear Factor-κB in Ovarian Granulosa Cells

0888-8809/04/$15.00/0 Printed in U.S.A.

Molecular Endocrinology 18(8):1919–1928 Copyright © 2004 by The Endocrine Society doi: 10.1210/me.2004-0021

Transrepression of Estrogen Receptor ␤ Signaling by Nuclear Factor-␬B in Ovarian Granulosa Cells SIMON CHU, YOSHIHIRO NISHI, TOSHIHIKO YANASE, HAJIME NAWATA,

AND

PETER J. FULLER

Prince Henry’s Institute of Medical Research and the Department of Medicine (S.C., P.J.F.), Monash University, Clayton, Victoria 3168, Australia; and Department of Medicine and Bioregulatory Science (Y.N., T.Y., H.N., P.J.F.), Graduate School of Medical Sciences, Kyushu University, Fukuoka 8128582, Japan Estrogen receptor (ER) ␤ is the predominant ER in granulosa cells of the ovary. ER␤ is expressed at high levels in granulosa cell tumors (GCT) and in the human GCT-derived cell lines, COV434 and KGN. To gain insight into ER␤ function in granulosa cells and in GCT, we have used the COV434 and KGN cell lines. Although the cells bind estradiol (E2), transcriptional activation of a transfected estrogen-responsive reporter vector construct (ERE2-luc) by E2 was not observed. Transactivation was also not observed with cotransfected ER␣ or ␤. This transcriptional resistance is specific to steroid receptor transactivation; reporter plasmids that are activated by the transcription factors activator protein 1 (AP-1) and nuclear factor ␬B (NF-␬B) demonstrate both constitutive and inducible transactivation. AP-1 and NF-␬B are known

to cause transrepression of both ER␣- and glucocorticoid receptor-mediated transcription. We therefore examined the possibility that activation of these pathways was responsible for the lack of a response to estrogen by using inhibitors of AP-1 or NF-␬B. The AP-1 inhibitors alone had no effect, whereas inhibition of NF-␬B signaling allowed a 3- to 4-fold E2mediated induction of ERE2-luc. This response was both ligand and ER dependent. Repression of ER␤ signaling by NF-␬B has not previously been reported. Recent evidence suggests that ER␤ may function to promote differentiation. The inhibition of ER␤ in combination with the antiapoptotic properties of NF-␬B may therefore contribute to pathogenesis of GCT. (Molecular Endocrinology 18: 1919–1928, 2004)

E

ER␤ is predominantly and abundantly expressed in granulosa cell tumors (GCT) of the ovary (6). By contrast, ER␣ shows moderate expression across a range of ovarian tumors and indeed, normal ovary (6). Although the significance of ER␤ expression in GCT is unclear, approximately 70% of GCT are known to synthesize estrogen, indicating that ER␤ may contribute to proliferation and/or differentiation. Several signaling pathways are important in mediating or modifying the proliferative response of granulosa cells to FSH stimulation, the primary mediator of granulosa cell proliferation during folliculogenesis (7). Understanding the role of these signaling pathways, including those influenced by estrogen, in human granulosa cells and in GCT requires a suitable in vitro model. Although several animal granulosa cell-derived cell lines have been reported, only six human granulosa cell lines have been established (8). Of these, two have been reported to express the FSH receptor (8, 9). The KGN cell line was established from a recurrent GCT and has detectable aromatase activity that can be further stimulated by FSH (8). The GCT-derived cell line, COV434 (10), has been reported to synthesize 17␤-estradiol (E2) in response to FSH, indicating the presence of the FSH receptor (9). These cell lines may therefore be useful in characterizing the signaling pathways activated by FSH stimulation and also in elucidating the molecular pathogenesis of GCT.

STROGEN MEDIATES IMPORTANT physiological responses in numerous target tissues including the breast, uterus, brain, bone, and ovary. The effects of estrogen are mediated through two distinct receptors, estrogen receptor (ER) ␣ (1) and ER␤ (2). Estrogens enhance FSH-induced granulosa cell proliferation; however, evidence in support of a role for estrogen in human folliculogenesis has until now been limited (3). The major site of ER␤ expression in the female reproductive tract is the granulosa cell. ER␤ null mice exhibit a predominantly ovarian phenotype with partial arrest of follicular development (4). The exact role of ER␤ in the ovary is unclear, although it has been hypothesized that this receptor may mediate differentiation (5). We have previously determined the patterns of expression of the two ERs in a panel of ovarian tumors. Abbreviations: AP-1, Activator protein 1; CHO, Chinese hamster ovary; CREB, cAMP response element binding protein; DES, diethylstilbestrol; E2, estradiol; ER, estrogen receptor; ERE, estrogen response element; ERE2-luc, estrogen-responsive reporter vector construct; FCS, fetal calf serum; GCT, granulosa cell tumors; GR, glucocorticoid receptor; 3H-E2, tritiated E2; I␬B␣, inhibitor ␬B␣; I␬K, inhibitor ␬B kinase; NF-␬B, nuclear factor ␬B; pTAL, enhancer-less reporter; SRC, steroid receptor coactivator; SRE, serum response element. Molecular Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the endocrine community.

1919 Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

1920 Mol Endocrinol, August 2004, 18(8):1919–1928

We have examined the potential utility of these cell lines as models of ER␤ action. In this report, we demonstrate that although both cell lines express ER␤ at the mRNA and protein level, the endogenous receptor is unable to transactivate an estrogen reporter construct. Further investigation reveals that the transcription factor nuclear factor ␬B (NF-␬B) contributes to transrepression of ER␤-mediated transcription.

RESULTS COV434 and KGN Cells Predominantly Express ER␤ mRNA and Protein Because GCTs predominantly express ER␤ (6), we sought to determine whether the COV434 and KGN cell lines also express this ER isoform using an RTPCR assay (6) designed to detect both ER␣ and ER␤. ER␤ is the predominant isoform in GCT samples as well as in both COV434 and KGN cells (Fig. 1A). No ER␣ mRNA expression was observed in the cell lines. The levels of ER␤ in the cell lines, although lower than that seen in GCT, are more than in normal ovary. Western blot analysis with an ER␤-specific antibody confirmed that both COV434 and KGN cells express ER␤ at the protein level, whereas no ER␣ protein was detectable (Fig. 1B). In contrast, ER␣ but not ER␤ was detected in the ER␣-positive breast cancer cell line

Fig. 1. Expression of Estrogen Receptors in Granulosa Cell Tumors, Normal Ovary, and GCT-Derived Cell Lines COV434 and KGN A, Southern blot analysis of RT-PCR products from individual normal ovary (numbers 1–9), GCT (numbers 10–18), COV434 and KGN samples. amplified using universal ER primers and probed with ER␣- and ER␤-specific 32P-labeled oligonucleotide probes (6). The PCR was performed for 25 cycles. Negative controls in which the reverse transcriptase was omitted (⫺) are also included. B, Western blot analysis of ER␣ and ER␤ in COV434, KGN, HEK293, and T47D cells. Cytosolic (C) and nuclear (N) proteins were prepared as described in Materials and Methods. Twenty-five micrograms of protein were subjected to SDS-PAGE, followed by transfer to polyvinylidene difluoride membrane. The membrane was incubated with monoclonal human ER␣ antibody or polyclonal human ER␤ antibody followed by the secondary horseradish peroxidase-conjugated antibody. Antibody binding was visualized by chemiluminescence.

Chu et al. • Transrepression of ER␤ Signaling

T47D. This contrasts with the study of Garcia Pedrero et al. (11) in which ER␤ was identified in T47D cells; the reasons for this difference are not clear. HEK293 cells are ER mRNA negative, and neither ER␣ or ER␤ were detected by Western blotting. ER␤ in COV434 and KGN Cells Is Functional with Respect to Binding but Is Unable to Transactivate an Estrogen-Responsive Reporter Construct The ability of ER␤ expressed in the COV434 and KGN cell lines to bind E2 was assessed. Cytosolic extracts from COV434 and KGN cells were incubated with increasing concentrations of tritiated E2 (3H-17␤-E2) in the presence or absence of a 200-fold molar excess of diethylstilbestrol (DES). Scatchard analysis was used to determine the binding affinity of the ER. As shown in Fig. 2A, both cell lines exhibited high-affinity E2 binding with equilibrium dissociation constants of 1.6 ⫾ 0.4 ⫻ 10⫺9 M and 3.5 ⫾ 0.7 ⫻ 10⫺9 M for COV434 and KGN cells, respectively. These results demonstrate that ER␤ in both of these GCT-derived cell lines is indeed functional with respect to binding. To test whether the receptor in these cells is functional with respect to transactivation at a canonical estrogen response element (ERE), the cells were transfected with the ERE2-luc reporter plasmid. As shown in Fig. 2B, treatment of COV434 and KGN cells with 10 nM E2 had no effect on ERE2-luc activity as compared with vehicle-treated cells. The potent synthetic estrogen, DES also showed a similar lack of response. No significant difference was observed in luciferase activity between cells treated with vehicle and cells incubated with the ER antagonist ICI 182780, excluding the possibility that agonist activity was present in the system or that the receptor was constitutively active. This observed lack of response was not due to the ERE2-luc reporter used because three other estrogen-responsive reporters also displayed a similar lack of response (data not shown). Conversely, ER␣-positive T47D cells transfected with the ERE2-luc reporter and treated in parallel with E2 exhibited an appropriate response (data not shown). These observations suggested that the ER␤ present in these cells, although able to bind ligand, may be transcriptionally inactive. Exogenous ER␣ and ER␤ in COV434 and KGN Cannot Transactivate ERE2-Luc The response to exogenous ER was examined in the cell lines. When either an ER␣ or ER␤ expression vector was cotransfected with the ERE2-luc reporter in Chinese hamster ovary (CHO) cells (Fig. 3) or HEK293 cells (data not shown), a 3-fold induction of luciferase activity after treatment with 10 nM E2 was observed. However, when either expression vector was transiently transfected into COV434 or KGN cells, no transactivation of ERE2-luc in response to

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

Chu et al. • Transrepression of ER␤ Signaling

Mol Endocrinol, August 2004, 18(8):1919–1928

1921

Fig. 2. Functional Analysis of ER␤ in COV434 and KGN Cells A, Binding of [3H]-17␤-E2 in the presence or absence of a 200-fold excess of E2 for 4 h at 4 C. Unbound radioligand was removed as described, and specific binding was determined by subtracting nonspecific binding from total counts. The Scatchard plot analysis is shown yielding a binding affinity (Kd) of 1.6 ⫾ 0.4 nM and 3.5 ⫾ 0.7 nM for ER binding in COV434 and KGN cells, respectively. B, Transactivation assay measuring ER-mediated transcription in COV434 and KGN cells. COV434 and KGN cells were transiently transfected with the reporter construct ERE2-luc. Twenty-four hours post transfection, cells were treated with either vehicle (ethanol); 10 nM E2; 10 nM DES; or 10 nM E2 plus 10 nM ICI 182,780 and incubated at 37 C for 18 h. Firefly and Renilla luciferase activity was measured. The fold induction of luciferase activity (firefly luciferase/Renilla luciferase) is expressed relative to the baseline activity detected in those cells treated with vehicle alone. Each data point represents the mean ⫾ SEM of at least five determinations run in triplicate.

Fig. 3. Transactivation Assay Measuring Exogenous ER-Mediated Transcription in COV434 and KGN Cells CHO, COV434, and KGN cells were transiently transfected with either an ER␣ or ER␤ expression vector together with a reporter construct ERE2-luc. Twenty-four hours post transfection, cells were treated with either vehicle or 10 nM E2 and incubated at 37 C for 18 h. Firefly and Renilla luciferase activity was measured. The fold induction of luciferase activity (firefly luciferase/Renilla luciferase) is expressed relative to the baseline activity detected in those cells treated with vehicle alone. Each data point represents the mean ⫾ SEM of at least five determinations run in triplicate. *, P ⬍ 0.001 compared with vehicle alone group. **, P ⬍ 0.05 compared with vehicle alone group.

E2 treatment was observed (Fig. 3). The baseline transcriptional response for ER␣ in untreated cells is increased for reasons that are unclear. This result indicated that repression of ER-mediated transactivation was occurring in the COV434 and KGN cell lines.

Glucocorticoid-Mediated Transactivation Is Also Repressed in COV434 and KGN Cells This repression of transactivation was not limited to ER-mediated signaling. When the glucocorticoid receptor (GR) was transfected into these cells together

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

1922

Mol Endocrinol, August 2004, 18(8):1919–1928

Chu et al. • Transrepression of ER␤ Signaling

with the GRE-luc reporter construct, no induction of the reporter was observed in response to 30 nM dexamethasone (Fig. 4), whereas a 4-fold increase in luciferase activity was observed in CHO cells in response to dexamethasone (Fig. 4). This result demonstrated that the observed lack of transactivation was not restricted to the ER in these cell lines but suggested that it may be more generalized. Transcriptional Repression Is Restricted to the Steroid Receptor-Mediated Transcriptional Response To investigate whether the transcriptional repression seen is restricted to the steroid receptors, the COV434 and KGN cells were transiently transfected with reporter constructs carrying enhancer elements for second messenger pathways. In COV434 and KGN cells, the cAMP response element binding protein (CREB), heat shock, and MAPK reporters could all be induced by the appropriate stimulus (Fig. 5). In both cell lines, the activator protein 1 (AP-1) and NF-␬B signaling pathways were constitutively activated. Under serum-free conditions, transfection with either the AP-1 or NF-␬B reporter alone elicited a 4- to 5-fold induction of luciferase activity (Fig. 5) as compared with cells transfected with the enhancer-less reporter (pTAL). For AP-1, this constitutive activation was not further induced upon treatment with 10% fetal calf serum (FCS), a stimulus for the AP-1 pathway (Fig. 5). In contrast, treatment of cells transfected with the NF-␬B reporter with TNF␣ resulted in a further induction of luciferase in the order of approximately

Fig. 4. Transactivation Assay Measuring GR-Mediated Transcription in COV434 and KGN Cells CHO, COV434, and KGN cells were transiently transfected with a GR expression vector together with the reporter construct GRE-luc. 24 h post transfection, cells were treated with either vehicle or 30 nM dexamethasone (DEX) and incubated at 37 C for 18 h. Firefly and Renilla luciferase activity was measured. The fold induction of luciferase activity (firefly luciferase/Renilla luciferase) is expressed relative to the baseline activity detected in those cells treated with vehicle alone. Each data point represents the mean ⫾ SEM of at least five determinations run in triplicate. *, P ⬍ 0.001 compared with vehicle alone group.

Fig. 5. Transcriptional Repression Is Specific to the Steroid Receptor-Mediated Transcriptional Response COV434 and KGN cells were grown in serum-free media and transiently cotransfected with reporter constructs for other signaling pathways. The reporter plasmids used were as follows: CREB (CRE), heat shock (HSE), SRE, AP-1, and NF-␬B. pTAL-luc contains the same promoter as the reporter plasmids without the response elements. Twenty-four hours post transfection, cells were treated with either vehicle or the appropriate stimulus as shown in the key and incubated at 37 C for 18 h. Firefly and Renilla luciferase activity was measured. The fold induction of luciferase activity (firefly luciferase/Renilla luciferase) is expressed relative to the baseline activity detected in those cells treated with vehicle alone. Each data point represents the mean ⫾ SEM of at least five determinations run in triplicate. *, P ⬍ 0.05 compared with vehicle alone group. **, P ⬍ 0.001 compared with vehicle alone group. ˆ, P ⬍ 0.05 compared with pTAL group. #, P ⬍ 0.001 compared with pTAL group.

30-fold (COV434) and 13-fold (KGN) over that of vector alone (Fig. 5). These results show that COV434 and KGN cells have functional pathways for CREB, heat shock, and MAPK, while also displaying constitutive activation of the AP-1 and NF-␬B pathways. The lack of a transcriptional response to steroid hormone is therefore seen for ER- and GRmediated transactivation. Inhibiting AP-1 Using MAPK Inhibitors Does Not Rescue Inhibition of ER-Mediated Transactivation on the ERE AP-1 transcription factors can cause transrepression of both the GR (12) and ER␣ (13). To determine whether this was the case for ER␤, inhibitors of the

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

Chu et al. • Transrepression of ER␤ Signaling

Mol Endocrinol, August 2004, 18(8):1919–1928

1923

Fig. 6. Effect of MAPK Inhibitors on SRE-, AP-1-, and ER-Mediated Transactivation COV434 and KGN cells were transiently transfected with either the SRE-luc (A), AP-1-luc (B), or the ERE2-luc (C). Twenty-four hours post transfection, cells were treated with either vehicle, the appropriate stimulus or the appropriate stimulus in the presence of one of the MAPK inhibitors as shown in the key and incubated at 37 C for 18 h. Firefly and Renilla luciferase activity was measured. The fold induction of luciferase activity (firefly luciferase/Renilla luciferase) is expressed relative to the baseline activity detected in those cells treated with vehicle alone. Each data point represents the mean ⫾ SEM of at least five determinations run in triplicate wells. *, P ⬍ 0.001 compared with 10% FCS. **, P ⬍ 0.05 compared with the vehicle alone group.

MAPK pathways targeting the ERK (PD98059), p38 (SB203580), and c-Jun N-terminal kinase (SP600125) pathways were used. These three MAPK pathways are important in the AP-1 signal transduction pathway (14). All three MAPK inhibitors effectively inhibited the stimulated response on the serum response element (SRE) (Fig. 6A). PD98059 (10 ␮M) fully inhibited the constitutive response seen for the AP-1 signaling pathway in both cell lines, whereas SB203580 partially inhibited the response in the COV434 cells but not in the KGN cells (Fig. 6B). As shown in Fig. 6C, inhibition of AP-1 activity in the cell lines was unable to rescue ER-mediated transactivation at the ERE2-luc reporter in response to E2. In the presence of PD98059, which fully abrogated the constitutive activity of AP-1, the addition of E2 had no effect on the reporter. This result indicates that constitutively activated AP-1 signaling in the COV434 and KGN cells is not the cause of the lack of ER-mediated transactivation. Inhibiting NF-␬B Restores ER- and GR-Mediated Transactivation NF-␬B is known to interact with the GR and ER␣ (15) to mutually inhibit transactivation in response to ligand. To investigate whether constitutive activation of the NF-␬B signaling pathway may be the cause for the lack of response seen in COV434 and KGN cells, an inhibitor of NF-␬B was used. BAY11-

708 inhibits the phosphorylation of the ␣-subunit of inhibitor ␬B (I␬B␣), preventing nuclear translocation of NF-␬B. This inhibitor suppressed the observed constitutive activation of NF-␬B in COV434 and KGN cells at concentrations of 30 ␮M and 5 ␮M, respectively (Fig. 7A). When COV434 and KGN cells were transfected with the GRE-luc reporter construct, with or without cotransfection of GR, and treated with the inhibitor and 30 nM dexamethasone, a 3- to 4-fold induction of luciferase was observed for both the endogenous and exogenous receptor (Fig. 7B). This indicates that these cells have a functional endogenous GR that is unable to mediate ligand-dependent transactivation due to constitutive NF-␬B activity. To determine whether this could explain the lack of response to E2 in these cells, COV434 and KGN cells were transfected with the ERE2-luc reporter either in the absence or presence of the BAY11-7082 compound and treated with either 10 nM E2, 10 nM DES, or 10 nM E2, and 10 nM ICI 782180 (Fig. 7C). In the absence of BAY11-7082, no induction of the reporter was observed upon ligand treatment. However, in the presence of BAY117082, a 3- to 4-fold induction of luciferase activity was observed upon treatment with either E2 or DES. When ICI 182780 is added in the presence of E2, no induction is observed, arguing that this is receptor

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

1924 Mol Endocrinol, August 2004, 18(8):1919–1928

Chu et al. • Transrepression of ER␤ Signaling

Fig. 7. Effect of BAY11-7082 on NF-␬B-, GR-, and ER-Mediated Transactivation The COV434 and KGN cell lines were transiently transfected with either NF-␬B-luc (A), GRE-luc (B) or ERE-luc (C). Twenty-four hours post transfection, cells were treated with either vehicle or the appropriate stimulus in the presence of BAY11-7082 (20 ␮M, COV434; 5 ␮M, KGN) as shown in the key and incubated at 37 C for 18 h. Firefly and Renilla luciferase activity was measured. The fold induction of luciferase activity (firefly luciferase/Renilla luciferase) is expressed relative to the baseline activity detected in those cells treated with vehicle alone. Each data point represents the mean ⫾ SEM of at least five determinations run in triplicate wells. *, P ⬍ 0.05 compared with vehicle ⫹ BAY11-7082 group. **, P ⬍ 0.01 compared with TNF␣ group. #, P ⬍ 0.05 compared with TNF␣ group. ˆ, P ⬍ 0.001 compared with vehicle alone group.

mediated. Reciprocal transrepression between NF-␬B and hormone receptors has been reported (16). To test whether this was the case in the GCTderived cell lines, the COV434 and KGN cells were transfected with the NF-␬B reporter and treated with varying concentrations of E2 (10 nM and 1 ␮M). E2 treatment had no effect on the activity of the NF-␬B reporter (data not shown).

DISCUSSION Both ER␣ and ER␤ share many functional properties; however, it is evident that ER␤ also has unique functional capabilities. The identification of ER␤-responsive genes remains elusive; most cell lines express both receptors or only ER␣. To identify ER␤-regulated genes, we have made use of the GCT-derived cell lines, COV434 (10) and KGN (8), both of which predominantly express ER␤ mRNA and immunoreactivity. ER␤ in the cell lines binds ligand with an affinity consistent with that reported for ER␤ in normal tissues (17). The significance of the difference in the binding affinity between the two cell lines is not clear. Determination of binding affinities for ER␤ has been performed using whole tissue or expression systems,

whereas binding studies performed in cell lines have been in cells expressing both ER␣ and ER␤. To our surprise, in both cell lines, neither E2 nor DES was able to transactivate four different reporter constructs containing either a single or multiple copies of an ERE. Furthermore, when exogenous ER (␣ or ␤) was cotransfected with the reporter, no response to E2 was observed. We concluded that this lack of response was specific to steroid receptor-mediated transactivation because a failure of GR-mediated transactivation was also observed. One of the key stages in follicular development is the acquisition by the growing follicle of specific molecular pathways that enable differentiation to the preovulatory stage (7). A major component of this phase is the FSH-induced proliferation of granulosa cells between the preantral and the preovulatory stages. We have previously demonstrated that GCT display a profile of expression for FSH-induced genes, which is remarkably similar to that reported for the preovulatory granulosa cells of the normal ovary (18). The status of other signal transduction pathways in GCT, however, has not been investigated. By using a series of reporter constructs for various signal transduction pathways, we observed striking similarities between the COV434 and KGN cell lines. Unlike ste-

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

Chu et al. • Transrepression of ER␤ Signaling

roid receptor-mediated signaling, the CREB, heat shock, and MAPK signal transduction pathways were inducible, whereas the AP-1 and NF-␬B signaling pathways were constitutively active. It is known that the AP-1 and NF-␬B families of transcription factors regulate an array of gene networks that are induced in response to growth factors, mitogens, tumor promoters, DNA-damaging agents, or oxygen radicals (15). Interactions between different classes of transcription factors is illustrated by welldocumented interactions of AP-1 and NF-␬B family members with the nuclear receptors (16, 19). The AP-1 family of proteins belongs to a class of transcription factors encoded by protooncogenes that regulate various aspects of cell proliferation (14). AP-1 can be composed typically of either homodimers or heterodimers between members of the Jun and the Fos families. The expression and/ or activation of these factors is regulated by several signaling pathways, in particular the MAPK pathway. The expression pattern of the AP-1 transcription factors has been investigated in rodent ovarian granulosa cells and appears to be able to regulate different functions at specific stages of granulosa cell differentiation and proliferation (20). Components of the AP-1 complex interact with members of the steroid receptor superfamily to cause transrepression of GR-mediated transcription, and antagonism of ER␣-mediated transactivation (16). Interactions between ER␣ and AP-1 signaling are complex. In several cell lines, activation of AP-1 or overexpression of c-Jun and c-Fos cause repression of ER-mediated transactivation (13, 21). We hypothesized that inhibition of the constitutive AP-1 activity evident in both GCT-derived cell lines may restore ER␤-mediated transactivation. In both cell lines, the ERK signaling pathway inhibitor, PD-98059, significantly reduced the AP-1 activity in both cell lines but did not reverse the repression of steroid-mediated transcription. Activation of this signaling pathway does not therefore appear to be the primary mechanism of the repression of ER- or GR-mediated signaling in these cells. The role of the NF-␬B family of proteins in immune, inflammatory, and apoptotic responses is well documented (22). This signaling pathway is activated by growth factors, cytokines, and mitogens to control cell proliferation, differentiation, and morphogenesis (22). FSH stimulates NF-␬B-DNA binding in cultured rat granulosa cells; it is this pathway that mediates the regulation of the X-linked inhibitor of apoptosis by FSH (23). NF-␬B is a heterodimeric transcription factor comprised of the protein subunits: p50/p105 (NF␬B1), p65/RelA, p52/100 (NF-␬B2), c-Rel, and RelB. NF-␬B is localized to the cytoplasm by association with an inhibitory protein I␬B. Activation of NF-␬B involves the phosphorylation of I␬B by inhibitor ␬B kinase (IKK), targeting it for proteosomal degradation by the ubiqiuitin pathway (24). On release from I␬B, NF-␬B translocates to the nucleus and activates the

Mol Endocrinol, August 2004, 18(8):1919–1928

1925

transcription of genes possessing NF-␬B consensus DNA binding sites in their promoters (24). Recent evidence indicates that NF-␬B and the signaling pathways that are involved in its activation are important for tumor development. Constitutive activation of NF-␬B is emerging as a hallmark of various solid tumors, including breast, colon, prostate, and epithelial ovarian carcinomas (25) and is also observed in many tumor cell lines (25). The mode of constitutive NF-␬B activation in tumor types is varied and complex. Activation of the ras pathway in human tumors might be a means to induce constitutive NF-␬B activity and oncogenesis (26, 27). Granulosa cells have been transformed using oncogenic H-ras (28). However the mechanism by which NF-␬B signaling is up-regulated in the COV434 and KGN cell lines is not known. The constitutive activity is blocked by the IKK inhibitor, BAY11-7082, suggesting that activation is occurring before this step in the signaling pathway. As several pathways converge on IKK, further characterization of these cell lines and indeed of GCT is needed to identify the point of activation. Although, as noted above, ras activation may be responsible, we have not found activating mutations in either of the three ras genes in these cell lines (Jamieson, S., and P. J. Fuller, unpublished observation). Whether activation of NF-␬B signaling also occurs in vivo in GCT remains to be determined. We have observed strong similarities in the profiles of gene expression between these cell lines and the GCT (our unpublished observations), suggesting this may be the case. Members of the nuclear receptor superfamily and NF-␬B subunits have been shown to physically interact. The most well-characterized interactions are those of NF-␬B and GR, resulting in a mutual transcriptional antagonism (15). Recently, it has been demonstrated, in an in vivo system, that p65, p50, and I␬〉␣ interact with the GR not only in nucleus but also in the cytosol in absence of ligand (29). This mutual transrepression is thought to play a part in the modulation of inflammation and immunosuppression. Several groups have reported a reciprocal transcription inhibition between agonist-bound ER␣ and activated NF-␬B (19, 30) This cross-talk is cell type dependent and involves direct protein-protein interactions, inhibition of DNA binding, induction of I␬B expression, and coactivator sharing. Although the cross talk between ER␣ and NF-␬B has been demonstrated in a number of cell types in vitro, it should be noted that all these experiments have involved transfected ER in overexpression systems. ER␤, when transfected into HepG2 cells, is able to repress NF-␬B activation in a ligand-dependent manner (31, 32). Ligand-dependent repression of NF-␬B activity by transfected ER␤ has also been reported in the osteoblastoma U2–05 cell line (33). Curiously, in the current study, only repression of ER␤ signaling by NF-␬B was found, not the reverse. The reasons for this lack of reciprocal transrepression are not clear. In contrast to the present study, these previously reported interactions of ER␤

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

1926

Chu et al. • Transrepression of ER␤ Signaling

Mol Endocrinol, August 2004, 18(8):1919–1928

with NF-␬B all involve exogenous, and hence overexpressed, transcription factors. An exception to this is an in vivo study of inhibition of the inflammatory response in the liver by genistein, an ER agonist with modest selectivity toward ER␤ (34). Other nuclear receptors that have been demonstrated to physically interact with NF-␬B in similar experimental systems include the mineralocorticoid, androgen, and progesterone receptors (19, 30). In the COV434 and KGN cell lines, inactivation of NF-␬B signaling by inhibition of I␬B␣ phosphorylation was able to rescue ER␤-mediated transactivation at the ERE, leading to a 3- to 4-fold induction in response to ligand. This observation suggests that ER␤ and the NF-␬B transcription factors may interact in a similar manner to that observed for other steroid receptors. The mechanism of this observed transrepression of ER␤mediated transactivation by NF-␬B still needs to be investigated. A level of modulation may be provided by the steroid receptor coactivator, SRC-3 (35). IKK is found to complex with SRC-3 and, at least in vitro, it phosphorylates SRC-3. This would suggest that NF-␬B signaling might enhance ER-mediated signaling; however, SRC-3 levels are relatively low in GCT (Hussein-Fikret, S., and P. J. Fuller, unpublished observation). A direct proteinprotein interaction is the most likely explanation as has been demonstrated for other steroid receptors. Whether this interaction involves the p65 or p50 subunits of NF-␬B remains to be determined. The significance of constitutive activation of NF-␬B with consequent transrepression of ER␤ in the pathogenesis of GCT remains speculative. There are several lines of evidence to suggest that the role of ER␤ in granulosa cells may be antiproliferative (5, 36). It has been reported that ER␤ acts as a modulator of ER␣ and may have a role in suppressing ER␣-mediated proliferation (37). Thus activation of NF-␬B signaling in GCT may provide a survival advantage not only through its antiapoptotic and proproliferative effects but also by its repression of ER␤ signaling. GCTs exhibit a poor response to endocrine therapy. The lack of a response to estrogen and/ or antiestrogen may be due to activation of the NF-␬B pathway in GCT. In this study, we report an interaction between endogenous ER␤ and the NF-␬B transcription factor. We show that NF-␬B contributes to transrepression of ER␤ transactivation. Although NF-␬B transrepression of other steroid hormone receptors and indeed transrepression of NF-␬B signaling by ER␤ have been previously reported, this is the first report to our knowledge that demonstrates transrepression of ER␤ signaling by NF-␬B.

MATERIALS AND METHODS Cell Lines Cell lines used in these studies: granulosa cell carcinoma (COV434 and KGN) (8, 10), breast carcinoma (T47D), human

epithelial kidney cell line (HEK293), and CHO cells. The cells were maintained in DMEM/F12 supplemented with 10% FCS at 37 C in a 95% air/5% CO2 humidified incubator. Cells were maintained in serum-free DMEM/F12 without phenol red for experiments measuring E2-mediated transactivation. 17␤-E2, DES, and dexamethasone were purchased from Sigma (St. Louis, MO). Chemical inhibitors used were: ICI 182780 (Tocris, Bristol, Newcastle-upon-Tyne, UK); BAY11-7082 (Calbiochem, La Jolla, CA); and PD98059, SB203580, and SP600125 (all from Alexis Biochemicals, San Diego, CA). Isolation of RNA from Tissue Specimens Ovarian GCT (n ⫽ 9) and normal ovarian tissue (n ⫽ 9) were obtained in a study of serum inhibin levels in ovarian tumors (38). Most tissues have been examined in previous studies for inhibin subunit gene expression (39), ER␤ gene expression (6), and FSH-regulated gene expression (18). The RNA from the tissue and the two GCT-derived cell lines was extracted using the guanidine thiocyanate/caesium chloride method as described previously (39). This study protocol was approved by the Research and Ethics Committee of Monash Medical Centre, and all women gave written informed consent for the studies. RT-PCR Assay Expression of ER␣ and ER␤ was determined by RT-PCR using gene-specific primers and probes combined with Southern blot analysis as previously described (6). Antibodies and Western Blot Analysis Cytosolic and nuclear extracts from COV434, KGN, T47D, and HEK293 cells were prepared as described elsewhere (40). Western blot analysis involved standard procedures with an enhanced chemiluminescence detection kit (Amersham Biosciences, Buckinghamshire, UK). The antibodies used were an anti-ER␣ mouse monoclonal antibody (1:100; Novacastra, Newcastle-upon-Tyne, UK) and an anti-ER␤ affinitypurified polyclonal antibody (1 ␮g/ml; Zymed, South San Francisco, CA). Ligand Binding Assays Ligand binding assays were performed using cytosolic extracts (40). Extracts were incubated at 37 C for 1 h with increasing concentrations of [3H]-17␤-E2 (0.26, 0.64, 1.6, 4, 10, 25 nM). Nonspecific binding was assessed by adding a 200-fold excess of nonradioactive E2. Unbound steroid was removed using dextran-coated charcoal. Radioactivity was measured in a Packard 2500 TR liquid scintillation counter (Packard, Meriden, CT). Scatchard analysis (41) was used to determine the binding affinity for the ligand. Each concentration was assayed in triplicate. Plasmids Reporter vectors used in these studies: ERE2-luc; the Mercury Pathway Profiling System vectors (CLONTECH, Palo Alto, CA) for CREB (CRE-luc), AP-1-luc, GR (GRE-luc), NF␬B-luc, heat shock (HSE-luc), and serum (SRE-luc) response elements; the enhancer-less reporter (pTAL-luc). The ER␤ expression vector was constructed for the purposes of this study by subcloning the full-length human ER␤ into the pcDNA3.1 vector (Invitrogen Life Technologies, Carlsbad, CA). The expression vectors pCMV5-hER␣ and RSV-hGR have been described previously (42, 43). The pRL-tk plasmid containing the Renilla luciferase gene (Promega, Madison, WI) was used as a transfection efficiency control.

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

Chu et al. • Transrepression of ER␤ Signaling

Mol Endocrinol, August 2004, 18(8):1919–1928

Transfections and Luciferase Assays Cells were transfected by using Superfect reagent (QIAGEN, Valencia, CA) according to the manufacturer’s instructions. Reporter vectors were transfected at concentrations of 1 ␮g/well, and the expression vectors pCMV5-hER␣, pcDNAER␤, and RSV-hGR were transfected at 5 ng/well. Cells were treated post transfection with the specific stimulus or vehicle for 18 h. Firefly luciferase activity was measured by using the luciferase reporter gene assay (Roche, Mannheim, Germany); Renilla luciferase was measured using the Renilla Luciferase Assay system (Promega). Firefly luciferase activity was first normalized to the level of Renilla luciferase activity. Results were then calculated as fold induction relative to either expression of vehicle-treated group (for ERE2-luc), or the control empty expression vector (for Mercury Pathway Profiling System vectors).

5. 6. 7. 8.

9.

Statistics All experiments were carried out more than three times with triplicate plates per point. All values represent the mean ⫾ SEM. A Student’s t test was used for statistical evaluation. P ⬍ 0.05 was considered to indicate statistical significance.

Acknowledgments We thank Ms. Sue Panckridge for helping with the preparation of the manuscript, Dr. Michael Clarkson for critically reviewing this manuscript, our clinical colleagues for assistance with the collection of the tissues, Professor Frank H. de Jong (Erasmus Medical Center, Rotterdam, The Netherlands) for generously providing the COV434 cell line, Dr. Guck Ooi and Dr. Jean-Francois Ethier from our institute for providing cell lines, Dr. Benita S. Katzenellenbogen (Dept. of Molecular and Integrative Physiology, The University of Illinois, Urbana, IL) for providing the pCMV5-hER␣ expression vector, and Dr. Sylvia Lim-Tio (Box Hill Diabetes and Endocrine Centre, Box Hill Hospital, Victoria, Australia) for providing the ERE2-luc reporter construct.

Received January 16, 2004. Accepted May 14, 2004. Address all correspondence and requests for reprints to: Professor Peter J. Fuller, Prince Henry’s Institute of Medical Research, P.O. Box 5152, Clayton, Victoria 3168, Australia. E-mail: [email protected]. This work was supported by the National Health and Medical Research Council of Australia. S.C. is the recipient of the National Health and Medical Research Council of Australia (NHMRC) Dora Lush Postgraduate Scholarship. P.J.F. is the recipient of a Senior Principal Research Fellowship from the NHMRC.

10.

11.

12. 13.

14. 15.

16. 17. 18.

19.

20.

REFERENCES 1. Mangelsdorf DJ, Thummel C, Beato M, Herrlich P, Schutz G, Umesono K, Blumberg B, Kastner P, Mark M, Chambon P, Evans RM 1995 The nuclear receptor superfamily: the second decade. Cell 83:835–839 2. Enmark E, Gustafsson JA 1999 Oestrogen receptors—an overview. J Intern Med 246:133–138 3. Drummond AE, Findlay JK 1999 The role of estrogen in folliculogenesis. Mol Cell Endocrinol 151:57–64 4. Krege JH, Hodgin JB, Couse JF, Enmark E, Warner M, Mahler JF, Sar M, Korach KS, Gustafsson JA, Smithies O 1998 Generation and reproductive phenotypes of mice

21.

22. 23.

1927

lacking estrogen receptor ␤. Proc Natl Acad Sci USA 95:15677–15682 Britt KL, Findlay JK 2002 Estrogen actions in the ovary revisited. J Endocrinol 175:269–276 Chu S, Mamers P, Burger HG, Fuller PJ 2000 Estrogen receptor isoform gene expression in ovarian stromal and epithelial tumors. J Clin Endocrinol Metab 85:1200–1205 Richards JS 1994 Hormonal control of gene expression in the ovary. Endocr Rev 15:725–751 Nishi Y, Yanase T, Mu Y, Oba K, Ichino I, Saito M, Nomura M, Mukasa C, Okabe T, Goto K, Takayanagi R, Kashimura Y, Haji M, Nawata H 2001 Establishment and characterization of a steroidogenic human granulosa-like tumor cell line, KGN, that expresses functional folliclestimulating hormone receptor. Endocrinology 142: 437–445 Zhang H, Vollmer M, De Geyter M, Litzistorf Y, Ladewig A, Durrenberger M, Guggenheim R, Miny P, Holzgreve W, De Geyter C 2000 Characterization of an immortalized human granulosa cell line (COV434). Mol Hum Reprod 6:146–153 van den Berg-Bakker CA, Hagemeijer A, FrankenPostma EM, Smit VT, Kuppen PJ, van Ravenswaay Claasen HH, Cornelisse CJ, Schrier PI 1993 Establishment and characterization of 7 ovarian carcinoma cell lines and one granulosa tumor cell line: growth features and cytogenetics. Int J Cancer 53:613–620 Garcia Pedrero JM, Del Rio B, Martinez-Campa C, Muramatsu M, Lazo PS, Ramos S 2002 Calmodulin is a selective modulator of estrogen receptors. Mol Endocrinol 16:947–960 Cabral AL, Hays AN, Housley PR, Brentani MM, Martins VR 2001 Repression of glucocorticoid receptor gene transcription by c-Jun. Mol Cell Endocrinol 175:67–79 Doucas V, Spyrou G, Yaniv M 1991 Unregulated expression of c-Jun or c-Fos proteins but not Jun D inhibits oestrogen receptor activity in human breast cancer derived cells. EMBO J 10:2237–2245 Shaulian E, Karin M 2002 AP-1 as a regulator of cell life and death. Nat Cell Biol 4:E131–E136 De Bosscher K, Vanden Berghe W, Haegeman G 2003 The interplay between the glucocorticoid receptor and nuclear factor-␬B or activator protein-1: molecular mechanisms for gene repression. Endocr Rev 24: 488–522 Pfahl M 1993 Nuclear receptor/AP-1 interaction. Endocr Rev 14:651–658 Hanstein B, Liu H, Yancisin MC, Brown M 1999 Functional analysis of a novel estrogen receptor-␤ isoform. Mol Endocrinol 13:129–137 Chu S, Rushdi S, Zumpe ET, Mamers P, Healy DL, Jobling T, Burger HG, Fuller PJ 2002 FSH-regulated gene expression profiles in ovarian tumours and normal ovaries. Mol Hum Reprod 8:426–433 McKay LI, Cidlowski JA 1999 Molecular control of immune/inflammatory responses: interactions between nuclear factor-␬B and steroid receptor-signaling pathways. Endocr Rev 20:435–459 Sharma SC, Richards JS 2000 Regulation of AP1 (Jun/ Fos) factor expression and activation in ovarian granulosa cells. Relation of JunD and Fra2 to terminal differentiation. J Biol Chem 275:33718–33728 Tzukerman M, Zhang XK, Pfahl M 1991 Inhibition of estrogen receptor activity by the tumor promoter 12-Otetradeconylphorbol-13-acetate: a molecular analysis. Mol Endocrinol 5:1983–1992 Baldwin Jr AS 1996 The NF-␬B and I␬B proteins: new discoveries and insights. Annu Rev Immunol 14:649–683 Wang Y, Chan S, Tsang BK 2002 Involvement of inhibitory nuclear factor-␬B (NF␬B)-independent NF␬B activation in the gonadotropic regulation of X-linked inhibitor of apoptosis expression during ovarian follicular development in vitro. Endocrinology 143:2732–2740

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009

1928 Mol Endocrinol, August 2004, 18(8):1919–1928

24. Karin M 1999 How NF-␬B is activated: the role of the I␬B kinase (IKK) complex. Oncogene 18:6867–6874 25. Karin M, Cao Y, Greten FR, Li ZW 2002 NF-␬B in cancer: from innocent bystander to major culprit. Nat Rev Cancer 2:301–310 26. Finco TS, Westwick JK, Norris JL, Beg AA, Der CJ, Baldwin Jr AS 1997 Oncogenic Ha-Ras-induced signaling activates NF-␬B transcriptional activity, which is required for cellular transformation. J Biol Chem 272: 24113–24116 27. Wang W, Abbruzzese JL, Evans DB, Larry L, Cleary KR, Chiao PJ 1999 The nuclear factor-␬B RelA transcription factor is constitutively activated in human pancreatic adenocarcinoma cells. Clin Cancer Res 5:119–127 28. Amsterdam A, Zauberman A, Meir G, Pinhasi-Kimhi O, Suh BS, Oren M 1988 Cotransfection of granulosa cells with simian virus 40 and Ha-RAS oncogene generates stable lines capable of induced steroidogenesis. Proc Natl Acad Sci USA 85:7582–7586 29. Widen C, Gustafsson JA, Wikstrom AC 2003 Cytosolic glucocorticoid receptor interaction with nuclear factor-␬B proteins in rat liver cells. Biochem J 373:211–220 30. McKay LI, Cidlowski JA 1998 Cross-talk between nuclear factor-␬B and the steroid hormone receptors: mechanisms of mutual antagonism. Mol Endocrinol 12:45–56 31. Bhat RA, Harnish DC, Stevis PE, Lyttle CR, Komm BS 1998 A novel human estrogen receptor ␤: identification and functional analysis of additional N-terminal amino acids. J Steroid Biochem Mol Biol 67:233–240 32. Tyree CM, Zou A, Allegretto EA 2002 17␤-Estradiol inhibits cytokine induction of the human E-selectin promoter. J Steroid Biochem Mol Biol 80:291–297 33. Quaedackers ME, Van Den Brink CE, Wissink S, Schreurs RH, Gustafsson Jk JA, Van Der Saag PT, Van Der Burg BB 2001 4-Hydroxytamoxifen trans-represses nuclear factor-␬B activity in human osteoblastic U2-OS cells through estrogen receptor (ER) ␣, and not through ER ␤. Endocrinology 142:1156–1166

Chu et al. • Transrepression of ER␤ Signaling

34. Evans MJ, Eckert A, Lai K, Adelman SJ, Harnish DC 2001 Reciprocal antagonism between estrogen receptor and NF-␬B activity in vivo. Circ Res 89:823–830 35. Smith CL, O’Malley BW 2004 Coregulator function: a key to understanding tissue specificity of selective receptor modulators. Endocr Rev 25:45–71 36. Shim GJ, Wang L, Andersson S, Nagy N, Kis LL, Zhang Q, Makela S, Warner M, Gustafsson JA 2003 Disruption of the estrogen receptor ␤ gene in mice causes myeloproliferative disease resembling chronic myeloid leukemia with lymphoid blast crisis. Proc Natl Acad Sci USA 100:6694–6699 37. Weihua Z, Saji S, Makinen S, Cheng G, Jensen EV, Warner M, Gustafsson JA 2000 Estrogen receptor (ER) ␤, a modulator of ER␣ in the uterus. Proc Natl Acad Sci USA 97:5936–5941 38. Healy DL, Burger HG, Mamers P, Jobling T, Bangah M, Quinn M, Grant P, Day AJ, Rome R, Campbell JJ 1993 Elevated serum inhibin concentrations in postmenopausal women with ovarian tumors. N Engl J Med 329: 1539–1542 39. Fuller PJ, Chu S, Jobling T, Mamers P, Healy DL, Burger HG 1999 Inhibin subunit gene expression in ovarian cancer. Gynecol Oncol 73:273–279 40. Stein B, Rahmsdorf HJ, Steffen A, Litfin M, Herrlich P 1989 UV-induced DNA damage is an intermediate step in UV-induced expression of human immunodeficiency virus type 1, collagenase, c-fos, and metallothionein. Mol Cell Biol 9:5169–5181 41. Scatchard G 1949 The attractions of protein for small molecules and ions. Ann NY Acad Sci 51:660–672 42. Sun J, Meyers MJ, Fink BE, Rajendran R, Katzenellenbogen JA, Katzenellenbogen BS 1999 Novel ligands that function as selective estrogens or antiestrogens for estrogen receptor-␣ or estrogen receptor-␤. Endocrinology 140:800–804 43. Keightley MC, Curtis AJ, Chu S, Fuller PJ 1998 Structural determinants of cortisol resistance in the guinea pig glucocorticoid receptor. Endocrinology 139:2479–2485

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

Downloaded from mend.endojournals.org at Monash University Library on September 2, 2009