Background
Ovarian cancer is the fifth most common cause of death because of cancer in women and is the leading cause of death from gynaecological malignancy in the developed world [
1]. Due to missing screening methods and its aggressive behaviour, a vast number is diagnosed at an advanced stage [
2]. Steroid hormones have an influence on ovarian cancer cells [
3] and it has been shown that 40–60% of ovarian cancers express estrogen receptor (ER) α [
4,
5]. In advanced stages the selective estrogen receptor modulator tamoxifen is used in patients as a well-tolerated and also effective treatment [
6‐
8]. Moreover, use of peri- and postmenopausal hormone therapy has been shown to increase ovarian cancer risk [
9]. One extra ovarian cancer case per 1000 users can be observed in women who use hormone therapy for 5 years after the age of 50 years [
9].
Investigating the underlying mechanisms, it is inevitable to consider the two ER types, ERα and β. So far, little is known about the molecular mechanisms of ERβ function in ovaries and ovarian cancers. However, it has been shown that both receptor types exert different biological functions [
10,
11]. Given that ERβ is able to counteract ERα signaling in some settings, loss of ERβ is thought to enhance ERα-mediated proliferation of hormone-dependent cancer cells [
12]. Moreover, the influence of ERb signaling on apoptosis pathways has been shown [
13].
Comparing normal ovarian tissue with epithelial ovarian cancers, a loss of ERβ expression and a decrease in ERβ/ERα ratio can be observed [
14‐
16]. Furthermore, in metastases of ovarian cancers a complete loss of ERβ was observed, whereas in the corresponding primary tumors low expression levels were still measurable [
15]. A positive correlation of ERβ expression with survival has been shown in ovarian cancer patients as well as animal models [
17,
18].
In vitro studies on other hormone-dependent tumors as breast and prostate cancers revealed a tumor suppressive role of ERβ [
10,
19]. Fewer reports suggest that this receptor plays a similar role in ovarian cancer. Recently, we investigated the effect of ERβ overexpression on the SK-OV-3 ovarian cancer cells. Particularly overexpression of ERβ1 inhibited growth and motility of these cells and induced apoptosis. In addition, we observed specific changes in gene expression. Interestingly, the antitumoral effects of ERβ were independent of estradiol and functional ERα. However, we were able to show an increased transcription of cyclin-dependent kinase inhibitor 1, a decrease in cyclin A2 transcripts and an up-regulation of fibulin 1c [
20].
In another study, proliferation of ERα expressing BG − 1 ovarian cancer cells decreased after reintroduction of ERβ expression [
17]. An increased expression of ERβ was associated with a decreased number of cells in S phase, whereas more cells were found in the G2/M phase. Also the cell cycle regulators cyclin D1 and A2 were affected by ERβ expression. When ERβ was reintroduced, total retinoblastoma (Rb), phosphorylated Rb and phospho-AKT content decreased. A part of the antiproliferative effect of ERβ was explained by the strong inhibition of ERα activity and expression by ERβ [
17,
21]. To examine the role of ERβ in a more physiological model of ovarian carcinogenesis, Bossard et al. orthotopically transplanted ERβ expressing ovarian cancer cells in ovaries of Nude mice, which reduced both tumor growth and the presence of tumor cells in sites of metastasis, and led to improved survival [
17].
The suggested role of ERβ as tumor suppressor and the observed decrease of expression in ovarian cancer cells raise the question, whether ERβ expression in these cells might be high enough to make this receptor a potential target in ovarian cancer therapy. Thus, we investigated the effect of ERβ agonists on proliferation and gene expression of two ovarian cancer cell lines.
Methods
Material
The human ovarian cancer cell line OVCAR-3 was obtained from American Type Culture Collection (ATCC #HTB-161, Manassas, USA), and OAW-42 ovarian cancer cells were obtained from Sigma Aldrich (#85073102, St. Louis, USA). The cells were maintained in phenol red-free DMEM culture medium that was obtained from Invitrogen (Karlsruhe, Germany) containing FCS that was purchased from PAA (Pasching, Austria). RNeasy Mini Kit was obtained from Qiagen (Hilden, Germany). Transfectin reagent was obtained from BioRad (Hercules, USA). OptiMEM medium were purchased at Invitrogen (Karlsruhe, Germany). ESR2 and control siRNAs were from Ambion (Life Technologies, USA). Serum Replacement 2 (SR2) cell culture supplement and 17-β estradiol were from Sigma-Aldrich (Deisenhofen, Germany). ERβ agonists ERB-041 and WAY-200070 were from Tocris (Bristol, UK). 5α-androstane-3β, 17β-diol (3β-Adiol) was from Sigma (Deisenhofen, Germany) and Liquiritigenin from Extrasynthese (Lyon, France).
Cell culture, transfection and proliferation assays
OVCAR-3 and OAW-42 cells were maintained in DMEM/F12 medium supplemented with 10% FCS at 37 °C in a humidified atmosphere containing 5% CO2. For transfection, 4 × 105 cells per well of a 6-well dish were seeded in DMEM/F12 containing 10% FCS. The next day, 2 ml fresh culture medium was added to the cells. 5 μl Transfectin reagent (BioRad) and a mix of three ESR2 siRNAs (10 nM each) were used to prepare transfection solution in OptiMEM medium (Invitrogen). The siRNA mix contained three different ESR2-specific Silencer siRNAs (siRNA IDs 145,909, 145,910, 145,911, Ambion), targeting exons 1, 2 and 3 of ESR2 mRNA. As a negative control, Silencer Negative control siRNA #1 (Ambion) was used. Gene knockdown of ESR2 was verified by means of Western blot analysis 72 h after siRNA treatment as described below. For cell proliferation assays, cells cultured in DMEM/F12 supplemented with 10% FBS or serum replacement 2, both containing 0.1 nM E2, were seeded in 96-well plates in triplicates (1000 cell/well). For agonist analyses, ERβ agonists were added in a 10 nM concentration 1day later. The relative numbers of viable cells were measured on days 0, 3, 4, 5, 6 and 7 using the fluorimetric, resazurin-based Cell Titer Blue assay (Promega) according to the manufacturer’s instructions at 560Ex/590Em nm in a Victor3 multilabel counter (PerkinElmer, Germany). Cell growth was expressed as percentage of cells transfected with negative control siRNA. Growth data were statistically analyzed by the Kruskal–Wallis one-way analysis of variance.
Antibodies and Western blot analysis
OAW-42 and OVCAR-3 cells were lysed in RIPA buffer (1% (
v/v) Igepal CA-630, 0.5% (
w/
v) sodium deoxycholate, 0.1% (
w/
v) sodium dodecyl sulphate (SDS) in phosphate-buffered solution (PBS) containing aprotinin and sodium orthovanadate. Aliquots containing 10 μg of protein were resolved by 10% (
w/
v) SDS–polyacrylamide gel electrophoresis, followed by electrotransfer to a PVDF hybond (Amersham, UK) membrane. Immunodetection was carried out using monoclonal ERβ (ESR2) antibody 14C8 (ab288, Abcam, Germany), diluted 1:100 in PBS containing 5% skim milk (
w/
v), ERα (ESR1) antibody 6F11 (ab9269, Abcam, Germany) (1:500), lipocalin-1 (LCN1) antibody STJ96584 by St John’s Laboratory (London, UK) (1:300), Patched 2 (PTCH2) antibody ABIN1673339 (1: 500) by antibodies-online (Aachen, Germany), Mitochondrially Encoded NADH Dehydrogenase 6 (MT-ND6) antibody ABIN311275 (1:1000) by antibodies-online (Aachen, Germany), β-actin (ACTB) antibody (clone AC-74) from Sigma Aldrich (Munich, Germany) followed by horseradish peroxidase conjugated secondary antibody (1:50,000) which was detected using chemiluminescence (ECL) system (Amersham, Buckinghamshire, UK). The Western blot results from three independent protein isolations were densitometrically analyzed using ImageJ [
22] and expressed in percentage of cell treated with a vehicle control.
GeneChip™ microarray assay
Processing of the RNA samples (two biological replicates from OVCAR-3 and OAW-42 cells treated with E2 (0.1 nM) in combination with ERβ agonists (10 nM) or vehicle controls for 48 h) was performed at the local Affymetrix Service Provider and Genomics Core Facility, “KFB - Centre of Excellence for Fluorescent Bioanalytics” (Regensburg, Germany;
www.kfb-regensburg.de).
Samples were prepared for microarray hybridization as described in the Affymetrix GeneChip® Whole Transcript (WT) Sense Target Labelling Assay manual. Double-stranded cDNA was generated from 300 ng of total RNA. Subsequently, cRNA was synthesized using the WT cDNA Synthesis and Amplification Kit (Affymetrix). cRNA was purified and reverse transcribed into single-stranded (ss) DNA. Subsequently a combination of uracil DNA glycosylase (UDG) and apurinic/apyrimidinic endonuclease 1 (APE 1) was used to fragment ssDNA, which was afterwards labelled with biotin (WT Terminal Labelling Kit, Affymetrix). In a rotating chamber, 2.3 μg DNA were hybridized to the GeneChip Human Gene 1.0 ST Array (Affymetrix) for 16 h at 45 °C. After washing and staining the hybridized arrays in an Affymetrix Washing Station FS450 using preformulated solutions (Hyb, Wash & Stain Kit, Affymetrix), the fluorescent signals were measured with an Affymetrix GeneChip® Scanner 3000-7G.
Microarray data analysis
Summarized probe signals were created by using the RMA algorithm in the Affymetrix GeneChip Expression Console Software and exported into Microsoft Excel. Data was then analysed using Ingenuity IPA Software (Ingenuity Systems, Stanford, USA) and Genomatix Pathway Analysis software (Genomatix, Munich, Germany). Genes with more than 2-fold changed mRNA levels after ERβ knockdown in both biological replicates were considered to be differentially expressed and were included in the analyses.
Discussion
In this study, for the first time we report significant inhibitory effects of ERβ agonists on growth of ovarian cancer cell lines. In turn we demonstrated a significant proliferation increase after siRNA-mediated knockdown of ERβ, corroborating both our agonist findings and the suggested tumor suppressor role of this receptor in ovarian cancer. Though all ERβ agonists inhibited ovarian cancer cell growth, their effect on gene expression partially differed due to their known structural differences.
In ovarian cancer, steroid hormone receptors ERα and β are commonly expressed. Especially in normal ovarian tissue ERβ shows high expression levels, which decrease during carcinogenesis [
3,
14,
15,
23‐
26]. This loss of ERβ could be an important step for the development of ovarian cancer and might even be a general mechanism during tumorigenesis of estrogen-dependent tissues. A number of in vitro studies, including one from our group, support the tumor-suppressive role of ERβ in ovaries [
20,
27‐
33].
The results of our knockdown experiments, clearly suggesting an antiproliferative effect of ERβ in ovarian cancer cells, are in line with previous studies by us and others, reporting growth inhibition after overexpression of ERβ or growth increase after knockdown of this receptor [
17,
20].
In our study we addressed the question, whether expression of ERβ in ovarian cancer cells still might be high enough to make this receptor a potential target in ovarian cancer therapy. Thus, we investigated how ovarian cancer cells responded to treatment with ERβ agonists, which have been reported to bind preferentially to this receptor, but only to a much smaller extent to ERα. 3β-Adiol (5α-androstane-3β, 17β-diol) is a dihydrotestosterone metabolite which does not bind androgen receptors. However, it efficiently binds ERβ [
34] and acts as a physiological ERβ-activator in different tissues [
35,
36]. ERB-041 and WAY-200070 are highly specific synthetic ERβ agonists [
37,
38]. ERB-041 is known to display a more than 200-fold selectivity for ERβ than for ERα (EC
50 ERβ = 2 nM), WAY-200070 still has a 68-fold higher selectivity for ERβ than for ERα (EC
50 ERβ = 2 nM [
39]). Liquiritigenin is a plant-derived flavonoid from licorice root, which acts as a highly selective agonist of ERβ (EC
50 ERβ = 36.5 nM [
40]). Recently, we have shown that Liquiritigenin and 3β-Adiol inhibit proliferation of different breast cancer cell lines. However, proliferation of ERα-positive breast cancer cell lines was not affected by the agonists WAY200070 and ERB-041 [
41,
42]. We decided to use a 10 nM concentration of the agonists only, because the EC50 values for ERβ binding of all drugs are in the low nanomolar range, and possible ERβ-unspecific effects of higher drug concentrations on proliferation e.g. via ERα activation thus could be ruled out. Though all agonists affected proliferation regardless of the serum supplement used, our observation that agonist effects in the presence of 10% FCS were higher on OVCAR-3, but lower in OAW-42 cells compared to defined growth-factor free serum replacement might be explained by the different mutation status of these cell lines. OAW-42 cells derive from ascites from a serous ovarian cancer, they obtain mutations of
BRCA1 and
PIK3CA, but not of
p53 [
43]. OVCAR-3 cells were attained from ascites of a patient with high-grade serous ovarian cancer (G3) and exhibit a mutation of
p53 [
43]. Thus, proliferation of OVCAR-3 cells, which is elevated due to mutated p53 and is further increased by growth factors, might be more sensitive to growth inhibition by ERβ agonists [
44].
The transcriptome analyses of both cell lines we performed after treatment with ERβ agonists ERB-041, Liquiritigenin and WAY-200070 revealed possible molecular mechanisms underlying the observed antiproliferative effects. In our study we observed down-regulation of
PTCH2 in OAW-42 cells both on the mRNA and protein level after treatment with ERβ agonist WAY200070.
PTCH2 gene encodes a transmembrane receptor and is part of the hedgehog signaling pathway, which is known to play an important role in the development of several malignancies [
45‐
49]. High expression of
PTCH2 was associated with a poorer survival in patients with bladder cancer [
47]. Recently, Worley et al. showed a significant overexpression of
PTCH2 in ovarian clear cell carcinoma and associated endometriosis [
50]. Given that knockdown of PTCH2 was reported to exert significant growth inhibition in a clear cell cancer cell line, this gene might be in part responsible for the observed growth inhibitory effects of this ERβ agonist [
50].
Pathway analysis suggested that the observed effects of ERβ agonists are mediated by β-catenin (CTNNB1) and amyloid β precursor protein (APP), which have been reported to form a complex [
51]. Expression of
APP and
CTNNB1 previously has been reported to be inducible by estrogens [
52,
53].
CTNNB1 activity has been reported to be inhibited by
ESR2 and is known to affect expression of
EpCAM and
PTCH2, which could explain the link between ERβ agonists and decreased expression of
PTCH2 and
EpCAM we observed in OAW-42 cells [
54‐
56]. The fact that estrogen-inducible
APP has been reported to increase expression of
ND6 and
PTCH2 provides a putative molecular mechanism between
ESR2 knockdown and the observed downregulation of
ND6 and
PTCH2 [
57,
58].
Our observation of
LCN1 downregulation particularly by ERB-041 in both cell lines could be explained by the fact that E2 has been reported to regulate
LCN1 gene expression [
59,
60]. The role of this transporter of small lipophilic ligands in cancer is unclear. However, it remains to be investigated whether
LCN1 might exert tumor-promoting functions like its family member
LCN2 known to induce epithelial to mesenchymal transition and to promote breast cancer invasion in an ERα-dependent manner [
61,
62].