Background
Ovarian cancer (OVCA) is the most lethal gynecological malignancy and fifth leading cause of cancer-related death in women [
1]. The fallopian tube epithelium (referred to as oviduct epithelium in all species except primates) is one of the likely progenitor cell types for the most common and deadly OVCA histotype, high-grade serous cancer (HGSC) with the alternative cellular source being the ovarian surface epithelium (OSE) [
2]. Morphological, immunological, and gene expression analysis of HGSC tumors, suggest a close relationship to fallopian tube epithelium rather than OSE [
3,
4]. Precursor lesions, termed the p53 signature, have been identified in the fimbriated end of the fallopian tube [
5]. In cases of HGSC, these p53 signatures in the fallopian tube had the same mutation as their corresponding ovarian carcinoma [
5]. Additional precursor lesions termed serous tubal intraepithelial carcinomas (STICs) have been identified in the fallopian tubes (but not ovaries) of women undergoing risk-reducing salpingo-oophorectomy, particularly in women with
BRCA1/2 mutations [
6]. Furthermore, mouse models of HGSC have been produced by inducing BRCA, PTEN, and p53 mutation in the oviductal epithelium [
7].
Risk factors for non-heritable OVCA include null parity, infertility, the number of lifetime ovulations and the use of estrogen only-hormone replacement therapy [
8]. Estrogen is a steroid hormone that functions in multiple tissues in the body, including the fallopian tube epithelium. While the reproductive role of estrogen in the fallopian tube is to facilitate movement and maturation of eggs, sperm and fertilized embryos between the ovary and uterus [
9], the function in terms of tumor initiation and progression is not clear. Estrogen signals in the cell through three main receptors. Estrogen receptor α (ERα) and ERβ are ligand activated transcription factors [
10]. G protein coupled receptor (GPER) is a membrane bound factor that signals through a non-genomic mechanism [
11]. Given the recent findings suggesting that OVCA is increased by hormone replacement therapy containing estrogen and the fallopian tube may be the source of HGSC, the estrogen receptor targets in the fallopian tube should be defined [
12].
ERα and a gene target of estrogen signaling, progesterone receptor (PR), are prognostic biomarkers in OVCA [
13]. ERα is expressed in 80 % of HGSC, but PR is expressed in only 31 % [
13]. Successful treatment of OVCA with selective estrogen receptor modulators (SERM) therapy has been limited [
14]. SERMs are ER ligands that function as either agonists or antagonists in a cell type specific manner [
15]. Given that HGSC may arise from the fallopian tube, understanding the response of normal fallopian tube epithelium to estrogen and SERMS is important for understanding the implications of SERM therapy on OVCA risk.
Murine oviduct epithelial (MOE) cells were utilized to investigate estrogen signaling in a putative HGSC precursor. MOE cells are estrogen responsive and the SERMs 4-hydroxytamoxifen (4OHT), raloxifene (RAL) and desmethylarzoxifene (DMA) antagonize 17-βestradiol (E2) in this cell type. The MOE specific transcriptional targets of estrogen signaling were determined by RNAseq. Finally, the receptor status for ERα, ERβ and PR was determined in a panel of HGSC cell lines. Our results highlight the need to consider the E2 response of putative progenitor cell populations of HGSC to investigate estrogen’s role in initiation and progression of OVCA. The occurrence and survival rates for OVCA have not improved in over 40 years emphasizing the necessity for better understanding and effectively treating or preventing this deadly disease. This study demonstrates that estrogen receptor activates unique oviduct-specific targets that may provide new targets for preventing or treating fallopian tube derived tumors.
Methods
Cell culture
MOE and MOSE cells (passages 7-25) were cultured as previously described [
16,
17]. For experiments investigating E
2 and SERM response, cells were cultured at least 48 h in “stripped media” consisting of phenol red free α-modified Eagle’s medium (Life Technologies, Carlsbad, CA) supplemented with 10 % v/v charcoal stripped fetal bovine serum (FBS) (Life Technologies) [
18], 1 mg/mL gentamycin (Mediatech, Manassas, VA), 2 mM L-glutamine (Life Technologies), 100 U/mL penicillin and 50 μg/mL streptomycin (Roche, Indianapolis, IN) prior to treatments. KURAMOCHI (passages 15-25), OVSAHO (passages 45-60) and OVKATE (passages 45-60) cell lines were purchased from the JCRB Cell bank and maintained in RPMI1640 (Mediatech) supplemented with 10 % FBS,100 U/mL penicillin and 50 μg/mL streptomycin. OVSAHO were cultured for 48 hours (h) prior to treatment and treated in phenol-red free RPMI1640 (Life Technologies) supplemented with 10 % charcoal stripped FBS, 100 U/mL penicillin and 50 μg/mL streptomycin. OVCAR4 (passages 15-30) were acquired from the National Cancer Institute Division of Cancer Treatment and Diagnosis Tumor Repository and maintained in the same media as KURAMOCHI plus 1 % ι-glutamine. SKOV3 (passages 10-30) and OVCAR3 (passages 3-30) cell lines were acquired from ATCC (Manassas, VA). SKOV3 were cultured in McCoy’s 5A supplemented with 2.3 g/L sodium carbonate, 10 % FBS, 100 U/mL penicillin and 50 μg/mL streptomycin. OVCAR3 were maintained in minimum essential media supplemented with 20 % FBS, 1 % ι-glutamine, 1 % non-essential amino acids, 1 % sodium pyruvate, 100 U/mL penicillin and 50 μg/mL streptomycin. Cell lines were authenticated by STR analysis at DDC Medical (murine) or UIC DNA Services Facility (human).
Western blotting
Cells were plated at a density of 10-30 × 104 cells per well in a 6-well dish in stripped media. Twenty-four hours post plating cells were washed with PBS and the media was replenished. After another 24 h, cells were washed with PBS and treated with DMSO (0.1 %) or compounds for indicated times.
Following treatment, cells were washed with PBS then harvested on ice in RIPA buffer (50 mM Tris, pH 7.6, 150 mM NaCl, 1 % v/v Triton X-100, 0.1 % w/v sodium dodecyl sulfate) and frozen at -80 °C. Lysates were centrifuged at 14,000 rpm for 5 min and clarified samples were quantified by BCA (Pierce, Rockford, IL), separated by SDS PAGE (8 %) at 100 V for 2.5 h then transferred for 2 h at 25 V to nitrocellulose (GE Healthcare Bio-Sciences, Pittsburgh, PA). Blots were blocked in 5 % milk TBS-T (Tris Buffered Saline-Tween 20) for 1 h at room temperature (RT) followed by overnight incubation at 4 °C in primary antibodies: Anti-ERα (1:300, MC-20 or 1:200 HC-20, Santa Cruz, Dallas, TX), Anti-PR (1:500 H-190, Santa Cruz), Anti-actin (1:1000, Sigma-Aldrich). Following 3 washes in TBS-T, blots were incubated in Anti-rabbit horseradish peroxidase-conjugated secondary (1:1000, Cell Signaling, Cambridge, MA) for 30 min at RT. After secondary, blots were washed 3X in TBS-T then bands were imaged using SuperSignal West Femto Chemiluminescent Substrate (Pierce) on a FluorChem E Imager (ProteinSimple, Santa Clara, CA). Densitometry was performed using Protein Simple software.
qPCR
Cells were treated as described for Western blots, then harvested in 500 μL of Trizol Reagent (Life Technologies) and frozen at -80 until processed according to manufacturer’s protocol. Purified RNA was treated with DNAseI (New England Biolabs, Ipswich, MA) for 10 min, followed by heat inactivation. 1 μg of purified RNA was reverse transcribed using Revertaid Reverse Transcriptase (Thermo Fisher Scientific, Waltham, MA) in the presence of Ribolock RNAse Inhibitor (Thermo Fisher Scientific). Expression was monitored using Faststart Universal Sybr Green (Rox) (Roche) using protocol: 10 min at 94 °C, 40 cycles of 10 sec at 94 °C, 30 sec at 60 °C followed by a melt curve. Primers listed in Additional file
1: Table S1.
Immunofluorescence
MOE cells were cultured in charcoal stripped media for 72 h and then plated (50,000 per well) onto Millicell EZ slides (Millipore, Billerica, MA). The next day the cells were fixed with 4 % paraformaldehyde and probed with primary antibody against estrogen receptor α (1:200, MC-20, Santa Cruz) with 10 % goat serum. Slides washed and probed with secondary antibody (AlexaFluro 594 A11037, Life Technologies) before mounting with DAPI containing mounting media (H-1500, Vector Laboratories, Burlingame, CA). Images were captured with a Nikon Eclipse E600 microscope at 40x.
Proliferation assay
Cells were cultured in stripped media for 72 h followed by passaging and plating of 1,000 cells/well in a 96 well plate. Twenty-four hours post plating, cells were treated with solvent or compounds for 72 h then fixed with 20 % w/v Trichloroacetic acid for 24 h at 4 °C and processed for sulforhodamine B colorimetric assay as described [
19]. Absorbance at 505 nm was measured using a BioTek Synergy MX microplate reader (BioTek, Winooski, VT).
Wound healing assay
Cells were cultured for 72 h as described in the proliferation assay, followed by passaging and plating of 6×104 per well in a 24 well plate. Twenty-four hours post plating a pipette tip induced a scratch across the monolayer of cells. Images of the wound were taken at time zero and 24 h post scratch. The percent closure of the wound was calculated as the area of the scratch at 24 h post scratch divided by the area of the scratch at time zero using ImageJ (National Institutes of Health).
Luciferase assay
SKOV3 cells were trypsinized and plated in 24-well plate (3.5 × 10
4 cells/ well) in stripped media. Incubation of cells with pERE-luciferase plasmid (100 ng/well) [
20], RSV-β-galactosidase (100 ng/well, [
21], and TransIT LT1 transfection reagent (1 μL per well, Mirus Bio, Madison, WI) was performed overnight in fresh media then treated for 24 h. Luciferase production and β-galactosidase activity (for transfection normalization) were measured as described previously [
21].
RNAseq library construction and sequencing
Cells were treated as described in qPCR assay, followed by RNA isolation using Qiagen Qiashredder column, on column DNAse treatment and Qiagen RNAeasy spin columns (Qiagen, Valencia, CA). Library construction and sequencing were performed at the Genomics Core facility at the University of Chicago. RNA quality and quantity were determined with the Agilent Bioanalyzer 2100, with RNA integrity numbers (RIN) of 10 and quantities of 100 ng or more per sample. Samples were enriched for mRNA using oligo-dT columns. Directional 50 bp single-end mRNA libraries were prepared using Illumina TruSeq mRNA Sample Preparation Kits per manufacturer’s instructions. Briefly, polyadenylated mRNAs were captured from total RNA using oligo-dT selection. Next, samples were converted to cDNA by reverse transcription, and each sample was ligated to Illumina sequencing adapters containing unique barcode sequences. Barcoded samples were then amplified by PCR and the resulting cDNA libraries quantified using qPCR. Finally, equimolar concentrations of each cDNA library were pooled and sequenced on the Illumina HiSeq2500.
Transcriptome analysis
The quality of DNA reads, in fastq format, was evaluated using FastQC. Adapters were removed and reads of poor quality filtered. The data was processed largely following the procedure described in [
22]. Briefly, reads were aligned to the
Mus musculus genome (mm10) using TopHat (v2.0.8b). Subsequently, aligned reads, in conjunction with a gene annotation file for mm10 obtained from the UCSC website, were used to determine the expression of known genes using Cufflinks (v2.1.1). Individual transcript files generated by Cufflinks for each sample were merged into a single gene annotation file, which was then used to perform a differential expression analysis with the Cufflinks routine, cuffdiff. Differential expression was determined by cuffdiff using the procedure described in Trapnell et al [
22], using an FDR cutoff value of 0.05. Results of the differential expression analysis were processed with cummeRbund. Differentially expressed genes were separated into upregulated and downregulated lists. A pathway analysis was performed on both gene lists using GeneCoDis [
23‐
25] to identify pathways enriched with genes that were upregulated and downregulated.
Statistical analysis
Data shown are represented as the mean of at least three experiments, with errors bars representing the standard error. Statistical analysis was conducted with GraphPad Prism (GraphPad, La Jolla, CA) using one-way ANOVA with a Tukey’s post hoc test.
Discussion
The epithelium of the fallopian tube is considered to be one of the origins of HGSC and estrogen replacement therapy impacts the risk of OVCA. This study characterized the agonist/antagonist function of SERMs in a normal murine model of fallopian tube epithelium, identified the genes regulated by estrogen signaling in this cell type to shed light on the subset of genes regulated by E2 in this cell type, and determined the E2 responsiveness of a panel of HGSC cell lines.
SERMS antagonize E2 in MOE cells
Tamoxifen is a common treatment for receptor positive breast cancer, as well as given prophylactically for women at high risk for developing breast cancer [
31]. Given the risk associated with Tamoxifen for the uterus, the response of Tamoxifen and other SERMs in the fallopian tube epithelium is highly important for clarifying the risks of using these drugs in women with intact fallopian tubes. Responses such as proliferation or migration has not been well documented in fallopian tube epithelium as compared to the response in the breast or uterine tissue, traditional hormone responsive tumor types [
20]. Reports from our lab indicate that estrogen did not proliferate oviductal epithelium in FVB strains, or in organotypic cultures of the mouse, baboon, or human [
20,
40]. Long term (at least 4 years) Tamoxifen treatment of women with breast cancer has been reported to increase tubal dysplasia [
41]. The anti-estrogen effect reported here on isolated epithelial cell lines suggests the dysplastic effect of tamoxifen may not be due to proliferation, but rather the length of exposure, or a non transcriptional effects, such as DNA damage [
42].
RNAseq identifies targets of E2 signaling in MOE cells
Many E
2 regulated genes in other normal and cancer cell types were also regulated in MOE cells including
Pgr,
Greb1,
Csf2, and
Dhrs9 [
43,
44], while other E
2 regulated genes in MCF7 cells were not regulated (
Nrip1) [
35] or even expressed (
Tff1) in MOE cells [
45]. Interestingly, in contrast to MCF7 [
45], the majority of differentially expressed genes were upregulated in MOE cells at the time point (24 h) probed. Table
1 and
2 lists the most highly regulated genes in response to E
2 in MOE cells, some of which are relatively uncharacterized. BPI fold containing family C gene (
Bpifc), involved in lipid binding [
46], demonstrated the largest induction in response to E
2. One gene specifically upregulated by 4OHT was the RNA component of mitochondrial processing endonuclease (
Rmrp), reported to be upregulated in 4OHT resistant as compared to 4OHT sensitive breast cancer cells [
47]. The biological processes enriched in MOE cells included regulation of proliferation and apoptosis, response to hormone stimulus and calcium ion homeostasis. The RNAseq analysis identified genes responsible for both positive and negative regulation of proliferation, which may reconcile the lack of significant proliferative increase in the FVB MOE background. In the fallopian tube, calcium is required for sperm capacitation [
48] and cilia beating [
49], therefore E
2 may regulate calcium levels in the fallopian tube as part of reproductive biology.
Comparison of the MOE E
2 responsive genes and The Cancer Genome Atlas of ovarian cancer tumors identified a number of genes with significance in OVCA in common between the two groups [
27,
50,
51]. For example, E
2 increased expression of the cyclin dependent kinase 2 (
Ccnd2) gene, which is altered (via amplification or mRNA upregulation) in 12 % of ovarian tumors, and has been shown to be upregulated in some OVCAs [
52].
Jak2, and
Kit were also upregulated in response to E
2 and altered in 12 % and 8 % of ovarian tumors, respectively. The
St3gal1 gene, encoding a glycosyltransferase, was upregulated by E
2 and altered (mostly amplified or showing mRNA upregulation) in 30 % of ovarian tumors. The significance of ST3Gal1 in OVCA is unknown, but has been linked to colon cancer [
53]. Cyclin dependent kinase 1 (
Ccnd1) was downregulated in response to E
2 and altered in 8 % of ovarian tumors. Interestingly, CCND1 is overexpressed in cisplatin resistant testicular cancer and OVCA [
54]. The overlap of genes regulated by E
2 in MOE cells and alteration in OVCA provides a number of potential new targets for further investigation of E
2 regulation in OVCA.
Need for better E2 responsive models of HGSC
The most frequently used estrogen responsive OVCA cell lines are not ideal models of HGSC including SKOV3 and the NIEHs BG1 cells [
38,
39]. Two other estrogen responsive OVCA cell lines, PEO1 and PEO4, have recently been reported as HGSC [
55]. These cell lines proliferate in response to E
2 in culture and xenografts and E
2 increases risk of distant metastases [
32,
33]. The HGSC cell lines investigated in this study express much less ER and PR receptors compared to the estrogen responsive MCF7. Nevertheless, the likely HGSC cell line OVSAHO proliferates in response to E
2, but not SERMs. Further validation of other/more estrogen responsive HGSC cell lines is desperately needed to aid in understanding the role of estrogen in OVCA and whether ER expressing HGSC would respond to anti-estrogen therapy. By studying how the cancers and different progenitors, such as the oviductal or OSE, respond to estrogen may aid in the use of SERMs in tumors that express ER or help to uncover if long-term use of Tamoxifen could enhance dysplastic lesions in the fallopian tube.
Conclusion
This study shows that the fallopian tube epithelia respond to E2 stimulation by regulating expression of a tissue-specific set of target genes. All SERMs tested inhibited E2-stimulated responses, showing SERMs are antagonistic action in the fallopian tube. Thus direct effects of E2 on the fallopian tube epithelium may play a role in the development of ovarian cancer.
Acknowledgements
We are grateful to Barbara Vanderhyden for FVB MTEC and MOSE cell lines, Ronny Drapkin for the FTSEC cell line, Greg Thatcher for the SERMs and Pavel Petukov for the HDAC inhibitor. This work was supported by the Northwestern University NGS Core Facility.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (
http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Authors’ contributions
GMH and JEB conceived of experiments. GMH, MD, and DAD performed experiments. MJS analyzed RNAseq data. Finally, GMH, MD, and JEB prepared manuscript. All authors approved the final manuscript.