Background
Ovarian cancer is a lethal tumor in women worldwide, with the highest mortality rate among gynecological malignancies, and its 5-year survival rate is 30–40% depending on tumor stage [
1,
2]. Although many improvements in surgical techniques and adjuvant therapies have been made, the survival rate of ovarian cancer patients has shown little improvement. The poor prognosis of this malignancy is largely due to late detection, chemoresistance, and a lack of targeted therapies for advanced and recurrent cases [
2]. Evidence from epidemiological and scientific studies has shown that hormones play important roles in ovarian tumorigenesis and progression [
3]. Androgen receptor (AR) is located on the X chromosome and is expressed in a diverse range of normal and cancer tissues. AR binds to its native ligand, 5α-dihydrotestosterone (DHT), with strong affinity in the nucleus. AR dimers bind to androgen response elements (AREs) in the promoter region of target genes [
4]. AR has biological actions in both health and disease and participates in both normal physiological processes and pathological conditions. Several studies have shown that AR is involved in the progression of numerous malignancies, including prostate, bladder, liver, kidney and lung cancers [
5,
6]. In recent years, the relationship between ovarian cancer and the AR signaling axis has become a popular topic of research because polycystic ovary syndrome and obesity are associated with a high risk of ovarian cance [
7,
8]. Meanwhile, emerging evidence has indicated that AR is frequently expressed in various ovarian cancer subtypes, especially epithelial ovarian cancer, and its high expression is associated with a poor prognosis [
9‐
13]. The AR signaling axis promotes the proliferation, migration and invasion of ovarian cancer cells in vitro and in vivo and interacts with many key signaling components, including the IL-6/IL-8 and EGFR pathways [
10,
14]. Nevertheless, the mechanisms by which androgen and AR influence cancer cell growth are complex. Additionally, AR may stimulate cancer development and progression possibly by expanding the population of cancer stem cells (CSCs). CSCs are specific population of cancer cells that are responsible for tumor initiation, drug resistance and metastasis. However, the impact of AR has not been widely pursued in ovarian cancer, and to date, the precise roles of AR and CSCs in ovarian cancer are not fully understood.
There is good evidence to support the view that most human tumors harbor CSCs and that these CSCs possess the biological characteristics of normal stem cells. CSCs also possess high self-renewal, extensive proliferative and strong tumorigenic capacities, and they are the initiating cells in tumor progression, recurrence and chemoresistance [
15,
16]. CSCs are able to persist in tumors and survive under nutrient starvation conditions [
17]. Over the last few decades, a number of studies have identified CSCs in human ovarian cancer cells. The development and progression of ovarian cancer is fueled and sustained by these undifferentiated CSCs. Indeed, ovarian CSCs may be responsible for tumor growth, peritoneal metastasis, chemoresistance and relapse. Because CSCs can potentially arise from oncogenic reprogramming and the dynamic nature of cancer cells, identification of stem cell-related and self-renewal molecules is expected to lead to novel therapeutic targets for cancer [
18].
In terms of CSC markers, Nanog has been identified as a molecule that maintains CSC pluripotency and self-renewal capability [
19]. Nanog, Oct4 and Sox2 are considered pluripotent genes and stem cell markers. Nanog is a nuclear transcription factor that plays a crucial role in pluripotent cells by maintaining their embryonic stem-like properties and in cancer cells by promoting carcinogenesis and reprogramming regulation [
20]. Many studies have indicated that Nanog is expressed in a variety of cancers, and its expression is correlated with poor survival [
21]. Therefore, Nanog can promotes CSC properties and characteristics, and such potentially pluripotent “stem-like” cells may have an impact on tumorigenesis and progression [
22]. Furthermore, Nanog is highly expressed in ovarian cancer tissues. Nanog overexpression in high-grade serous ovarian cancer is significantly associated with increased chemoresistance and poor survival. Although Nanog might function as a stem cell-associated gene involved in ovarian cancer tumorigenesis and prognosis [
23,
24], the regulation of Nanog in ovarian cancer is not well understood.
Several signaling pathways that have been identified in CSCs appear to be important for maintenance of the CSC phenotype, and the participation of Nanog is a key factor [
25,
26]. Meanwhile, AR has been implicated as a “molecular switch” that functions by coordinating and regulating the expression of the stem cell network. AREs contain the Nanog promoter. Thus, AR might regulate the Nanog pathway to promote the stem-like differentiation, proliferation and migration of some cancer cell types [
27,
28]. In addition, our previous study confirmed that Nanog is associated with androgen/AR and plays an important role in the regulation of stemness in liver cancer cells [
29]. However, these mechanisms remain unclear in ovarian cancer.
Therefore, in this study, we conducted an investigation to determine whether the AR signaling axis regulates Nanog and promotes stem-like differentiation and proliferation in ovarian cancer cells. The Clustered Regularly Interspaced Short Palindromic Repeats/protein 9 (CRISPR/Cas9) system is a simple and efficient genome-editing tool that can be applied to various cell types. We used the CRISPR/Cas9 system to target Nanog and inserted an endogenous green fluorescent protein (GFP) marker. With the CRISPR/Cas9 system, the GFP marker can directly and accurately reflect Nanog expression. Based on CRISPR/Cas9 technology, we constructed a stable and reliable cell marker model that avoids the genomic instability induced by integration of indirect genetic markers via virus vectors and the disadvantages of markers that decay over time. Hence, this approach is an efficient and stable method to examine how the androgen signaling axis regulates Nanog in ovarian cancer (Research model, Additional file
1: Figure S1). We hope to explore new treatment strategies or drug targets for ovarian cancer treatment using this technology.
Cell transfection
Cells were cultured for transfection in 24-well plates. When the cells reached approximately 70–80% confluence, they were transfected with 0.25 μg of PX330-Nanog-gRNA plasmid and/or 0.15 μg of the Nanog-2A-GFP homogeneous arm vector with Effectene transfection reagent (Qiagen, GER). After 72 h, GFP fluorescence was examined, and cells were sorted.
Fluorescence-activated cell sorting
All transfected cells stably expressing GFP were sorted with a FACSAria II cell sorter (BD Biosciences, USA). Individual SKOV3 or A2780 cells with GFP expression were seeded into 96-well plates for expansion, and then labeled GFP (+) cells were verified by Sanger sequencing. Single clones of SKOV3 + 5 or A2780 + 20 GFP (+) cells and GFP (−) cells were cultured and passaged for further studies.
RNA extraction and real-time qPCR analysis
Total RNA was extracted using an Eastep Super RNA extraction kit (Promega, USA), and then, 1 μg of RNA was converted to cDNA (reaction system 10 μl) with an Advantage® RT-for-PCR kit (Takara, JPN). After 2 μl of cDNA was mixed with SYBR Green (Bio-Rad Laboratories Ltd., USA), quantitative real-time qPCR (RT-qPCR) (reaction system 20 μl) was performed using a CFX96™ Real-Time system (Bio-Rad). GAPDH was used as the internal control. Then, we calculated the mRNA transcript abundance relative to that of GAPDH. All experiments were performed in triplicate. RT-qPCR primers are listed in the Additional file
1: Table S1.
Western blotting analysis
For protein extraction, the tissues were first ground, and then protein was extracted with tissue lysis buffer (Thermo Fisher Scientific, USA). Cells were washed with ice-cold PBS and lysed with cell lysis buffer. Whole-cell lysates were collected from 1 × 106 cells. Protein concentration was measured with a BCA protein assay kit (Beyotime, China). Western blotting assays were performed as described previously. Briefly, 50 μg of protein was loaded. The primary antibodies (1:1000) were as follows: anti-Nanog (Cell Signaling Technology, USA), anti-AR (N-20, Santa Cruz Biotechnology, USA), anti-Oct4 (Abcam, UK), anti-Sox2 (Abcam, UK) and anti-GAPDH (Cell Signaling Technology, USA). The secondary antibody was diluted 1:3000. All antibodies were used at the dilution recommended by the manufacturer.
Immunofluorescence analysis
Cells were grown in 24-well plates that were preloaded with glass slides and cultured for 12 h. Then, the cells were fixed with 4% paraformaldehyde for 30 min and permeabilized with 0.3% Triton (Sigma Aldrich, USA) for 10 min. After blocking with 10% BSA (Sigma Aldrich, USA), the slides were incubated with primary antibodies, namely, anti-Nanog (1:200) and anti-AR (1:1000), for 16 h at 4 °C and then with the Alexa Fluor® 568-conjugated secondary antibody (1:1000) (Thermo Fisher Scientific, USA) for 1 h at room temperature. The slides were treated with DAPI (1:1000) for 15 min and mounted with 40% glycerin before examination. Immunofluorescence was detected with a confocal microscope (Carl Zeiss Jena. JER) using the following parameters: objective, 40×; scan mode, 1024 × 1024; and a scan time of 6.25 s.
Hormone treatment
DHT and ASC-J9 were used as the drugs for hormone treatments. The drugs concentrations used were based on our previous study [
29]: DHT 10 nM and ASC-J9 5 μM. Cell viability was tested using an MTS Cell Proliferation Assay kit (Promega, USA). Cells were seeded in 96-well plates at 1 × l0
4 cells/well, cultured for 24 h, and then treated with 10 nM DHT (Dr. Ehrenstorfer GmbH, GER) and/or 5 μM dimethylcurcumin (ASC-J9) (MedChem Express, USA). Dimethyl sulfoxide (DMSO) served as the vehicle control. After further incubation for 24 and 48 h, MTT reagent (5 mg/l) was added to each well, and the cells were further incubated for 4 h. The number of viable cells was calculated relative to the appropriate controls. The mean ± SD values from three independent experiments are shown.
For hormone treatments, cultured cells were washed and placed into serum-free medium for 24 h. Then, 10 nM DHT or 5 μM ASC-J9 dissolved in DMSO was added into the medium. The cells were divided into three groups: DHT, DHT + ASC-J9 and DMSO and were harvested after 24 h.
Luciferase reporter gene and lentivirus vector assay
Luciferase activity was detected with a Luciferase Reporter Assay System (Promega, USA) according to the manufacturer’s protocol. The luciferase backbone lentivirus was used for insertion of Nanog promoter regions with different transcriptional start site (TSS) lengths (− 500 bp, − 1000 bp, − 1500 bp). Three days after transfection, the cells were starved in FBS-free medium for 24 h and then treated with vehicle or DHT with or without ASC-J9 for another 24 h. To examine the effect of androgen on the Nanog gene promoter, the activities of different sections of the Nanog gene were tested by stimulating and (or) suppressing androgen signaling. We used constructed pGL3.0 firefly luciferase gene reporters and lentivirus vectors with three different regions of the Nanog gene.
iCELLigence
Cell proliferation in vitro was investigated with iCELLigence software (ACEA Biosciences, USA). We sorted and treated the GFP (+) and GFP (−) cell groups treated with DHT or DHT + ASC-J9. An equal number of GFP (+)/GFP (−) cells (5 × 103 cells) were seeded. GFP (+)/GFP (−) groups were divided into two subgroups: 10 nM DHT and 10 nM DHT plus 5 μM ASC-J9. The compounds were added to each well. Then, the cells were cultured and monitored for 25–37 h. All experiments were performed in triplicate.
Transwell assay
For Transwell assays, 6.5-mm chambers with 8-μm pores (Corning, USA) were placed into 24-well plates and used to assess the migration of SKOV3 + 5 and A2780 + 20 GFP (+) or (−) cells. Cells were grown in Transwell chambers without Matrigel and treated with either DMSO, DHT or DHT + ASC-J9. An equal number of cells (5 × 104 cells in 100 μl of serum-free modified 1640 medium) were seeded into the upper chambers of polycarbonate Transwell filters. A total of 600 μl of modified 1640 medium containing 10% FBS was added to the bottom of the chambers. After incubation for 16 h, migrated cells were fixed with methanol and stained with hematoxylin and eosin. Images of four random fields in each membrane were captured with a microscope. For analysis, the cells number in four fields was calculated at 40× magnification (images shown in Additional file 1: Figure S6).
For GFP (+)/(−) cell identification, SKOV3 + 5 or A2780 + 20 single GFP(+) or (−) cells were sorted via FACS. Then, cells were seeded into 96-well ultralow attachment plates (Corning). The medium comprised DMEM/F12 (Invitrogen, USA), 2% B27 (Prepro Tech, USA), 20 ng/ml EGF (Prepro Tech, USA), 20 ng/ml bFGF, and 4 μg/ml heparin. The cells were treated for 14 days before analysis.
Colony formation assays were performed in 24-well plates. Cells were sorted and cultured in DMEM supplemented with 10% FBS for 14 days. In terms of hormone treatment, after seeding for 24 h, the medium was changed to DMEM with low FBS (2%), with or without the indicated reagents (DHT, DHT + ASC-J9 or DMSO). The medium was replaced every two days, and the cells were treated for 14 days. On the final day, the cells were fixed with 4% paraformaldehyde and stained with freshly prepared crystal violet for 20 min. Colony formation was observed under a microscope (Nikon, Japan).
Immunohistochemistry
Immunohistochemical staining was performed using a conventional protocol. In brief, the tumor tissues were fixed in 4% paraformaldehyde immediately after removal from the mice. Then, samples were embedded in paraffin, sectioned, and immunostained. The sections were incubated with anti-AR (Abcam, UK) or anti-Nanog (Abcam, UK) primary antibodies at 4 °C for 16 h and secondary antibodies (DaKo, DEN) at 37 °C for 0.5 h. Finally, sections were counterstained with Mayer’s hematoxylin and mounted. Sections were visualized, and images were captured with an Olympus camera.
Xenograft study
All animal experiments were performed in accordance with the “Guide for the Care and Use of Laboratory Animals” and were approved by the Animal Ethics Committee of the Third Military Medical University. All mice were maintained in pathogen-free conditions. Nanog A2780 + 20 GFP (+) or GFP (−) cells (5 × 104) were suspended in PBS and mixed with Matrigel (volume ratio 1:1) (BD Biosciences, USA). Then, 200 μl of this mixture was injected subcutaneously into the dorsal surface of 5-week-old female nude mice. Five mice with 2 transplant sites each were used for the experiment. All mice were sacrificed after 30 days, and the tumor masses were measured. Moreover, to further validate the stemness of the GFP (+) cells and the effect of the hormone treatment in vivo, we constructed another animal model by injecting 5 × 103 A2780 + 20 cells into the dorsal surface of 12 female nude mice. When the xenografts reached approximately 180 mm3, the animals were divided into two groups: androgen treatment and anti-androgen treatment groups. The (clinically used) drugs were as follows: the androgen was a testosterone undecanoate soft capsule (N.V. Organon, NL), and the anti-androgen was a bicalutamide tablet (AstraZeneca, UK). Intragastric drug administration was performed every three days, and the sizes of the xenografts were measured. The doses of the androgen and anti-androgen (according to the clinical reference dose) were 6 mg/kg and 8.4 mg/kg, respectively. After 9 treatments, the mice were sacrificed, and the tumors were imaged with an in vivo imaging system (PerkinElmer, USA).
Statistical analyses
Each experiment was performed in triplicate. Statistical tests were performed using SPSS 17.0 and GraphPad Prism 5.0 (GraphPad Software). In addition, one-way ANOVA or a t-test was performed with normalization to obtain P-values. The criteria for statistical significance were P < 0.05 (*), P < 0.01 (**), and P < 0.001 (***). Relative analysis of AR and Nanog expression was calculated with Pearson’s correlation coefficient using the formula in SPSS 17.0.
Discussion
In this study, we have demonstrated that the AR signaling axis induces Nanog-mediated stemness properties and tumorgenicity in ovarian cancer. To the best of our knowledge, few studies have addressed the relationship between AR and Nanog in ovarian cancer or the regulation of OCSCs. Several studies have reported that AR expression is higher in ovarian cancer than in normal ovaries [
30]. Hence, we hypothesized that the AR signaling axis is associated with ovarian cancer. Our study began by investigating the expression levels of AR and Nanog in the same cohort of samples via IHC and western blot analysis (Fig.
1). We determined that the AR and Nanog expression levels in ovarian tumors were higher than those in the ovaries (Fig.
1a-
c). AR was expressed strongly in the cytoplasm and nucleus of ovarian tumor cells (Fig.
1d). We wondered whether a direct relationship existed between AR and Nanog in ovarian cancer. Moreover, we examined whether this signaling axis was consequently involved in OCSC regulation.
Therefore, endogenous Nanog labeling in ovarian cancer cell lines was conducted to explore the association and possible mechanisms of the two genes in ovarian cancer. We generated several single clones with the correct GFP marker (A2780 + 20 and SKOV3 + 5) by using the CRISPR/Cas9 insertion system and then compared the GFP (+) and GFP (−) cells under various conditions (Fig.
2). CRISPR/Cas9 technology was used to construct a stable GFP marker for subsequent experiments that could avoid the disadvantages of other methods, such as GFP degradation. We also confirmed that AR and Nanog were co-localized in ovarian cancer cells (Fig.
3). Nanog activity was promoted by AR, which was consistent with the results of a study by Kregel et al. [
27]. Meanwhile, crystallographic data have shown that AR contains an additional interface that stabilizes the AR dimer/ARE complex. In contrast, the dimerization strengths of other steroid receptors would not be sufficient to retain stable binding to selective AREs [
31]. We interpreted these data as strong evidence confirming this pathway. In this study, using luciferase assays, we showed that the AR signaling axis induced Nanog promoter activity in ovarian cancer cells (Fig.
4). Ovarian cancer cells treated with different hormones were used to confirm the effect of the AR signaling axis. Moreover, to explore the possible regulation of AR and Nanog in ovarian cancer cells, the proliferation and migration abilities of single-clone Nanog GFP (+) cells with or without hormone treatments were examined (Figs.
5-
6). DHT distinctly increased GFP (+) cell proliferation and migration. Under DHT + ASC-J9 treatment, no differences were evident between the Nanog GFP (+) and GFP (−) cells. Additionally, our previous study suggested that GFP (−) cells treated with DHT could undergo dedifferentiation via transformation from non-CSCs to stem-like cells [
29]. For this reason, DHT not only promoted the migratory ability of the GFP (+) cells but also affected GFP (−) cells in this experiment. We also examined the Oct4 and Sox2 protein expression levels. Interestingly, Oct4 and Sox2 levels increased along with Nanog expression when cells were treated with DHT (Fig.
7). In addition, DHT clearly enhanced ovarian cancer cell sphere and colony formation compared with the vehicle control and DHT/ASC-J9 treatment (Fig.
8). Thus, our in vitro studies showed that androgen induces Nanog-mediated stemness properties and tumorgenicity in ovarian cancer cells directly through AR. Our in vivo results suggest that Nanog induces stemness properties and promotes ovarian tumorigenicity by interacting with AR (Fig.
8). This study demonstrates that the AR signaling axis interaction with Nanog induces and maintains OCSC stemness by activating the Nanog promoter. We presume that an interaction effect is present between Nanog and AR and suggest that the AR-Nanog signaling pathway participates in OCSC regulation. Even though ovarian cancer has a complex network and microenvironment, this phenomenon may mean that increased AR activity is related to a high risk of this disease. However, the evidence is not sufficient to demonstrate the mechanism of the AR-Nanog signaling axis in patients. Nevertheless, we suggest that the AR-Nanog signaling axis participates in OCSC regulation and may be a prognostic bio-marker.
We present compelling evidence to show that Nanog plays a vital role in malignant diseases and is correlated with the clonogenic growth, tumorigenicity and invasiveness of cancer cells [
11,
32,
33]. Hence, Nanog appears to function as a cooperating or potentiating pro-tumorigenic molecule in the appropriate context [
34]. The highlight of our study is our cell model, which was successfully built with CRISPR/Cas9 technology, and the Nanog GFP marker can be used to monitor and study authentic CSCs or pluripotency. The 2A-tdTomato sequence can reportedly be inserted by homogenous recombination to replace the stop codon of the porcine gene. Thus, fluorescence can accurately show activation of the endogenous gene through CRISPR/Cas9 [
35]. These results indicate that the knock-in reporter system can be used to efficiently monitor the pluripotency status of cells. Using the CRISPR/Cas9 knock-in reporter system, we can monitor and investigate the complicated functions and regulation of genes more directly. Since the conditions of ovarian cancer cells are complex, other cell lines and networks require further elaboration.
Previous studies have not demonstrated efficient effects when anti-androgen is used to treat patients, but we should still consider the effects and design more animal studies to investigate this process [
36]. In the future, we hope to identify new anti-androgen approaches to eliminate OCSCs or to inhibit cancer cell growth in patients with high AR expression.
Acknowledgments
Special thanks to professor TC Li in the Department of Obstetrics & Gynaecology, Faculty of Medicine, Prince of Wales Hospital, The Chinese University of Hong Kong, for writing assistance. The authors would also like to thank Dr. Dev Sooranna, Imperial College London, for editing the manuscript.