Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com

ScienceDirect

www.e l sev i e r . com / l oca te / s c r

Stem Cell Research (2014) 13, 24–35

Ligand-independent androgen receptorspromote ovarian teratocarcinoma cellgrowth by stimulating self-renewal of cancerstem/progenitor cells

Wei-Min Chunga,b,1, Wei-Chun Changb,1, Lumin Chena,b, Tze-Yi Linb,Liang-Chi Chenb, Yao-Ching Hunga,b,⁎, Wen-Lung Maa,b,⁎,2

a Sex Hormone Research Center, Graduate Institution of Clinical Medical Science, School of Medicine,China Medical University, Taichung 404, Taiwanb Sex Hormone Research Center, Department of Obstetrics and Gynecology, and Department of Pathology,China Medical University Hospital, Taichung 404, Taiwan

Received 4 October 2013; received in revised form 30 March 2014; accepted 7 April 2014Available online 15 April 2014

Abstract Background: Ovarian teratocarcinoma (OVTC) arises from germ cells and contains a high percentage of cancerstem/progenitor cells (CSPCs), which promote cancer development through their ability to self-renew. Androgen and androgenreceptor (androgen/AR) signaling has been reported to participate in cancer stemness in some types of cancer; however, thisphenomenon has never been studied in OVTC.Methods: Ovarian teratocarcinoma cell line PA1 was manipulated to overexpress or knockdown AR by lentiviral deliversystem. After analyzing of AR expression in PA1 cells, cell growth assay was assessed at every given time point. In order todetermine ligand effect on AR actions, luciferase assay was performed to evaluate endogenous and exogenous AR function inPA1 cells. CD133 stem cell marker antibody was used to identify CSPCs in PA1 cells, and AR expression level in enriched CSPCswas determined. To assess AR effects on CD133+ population progression, stem cell functional assays (side population, sphereformation assay, CD133 expression) were used to analyze role of AR in PA1 CSPCs. In tissue specimen, immunohistochemistrystaining was used to carry out AR and CD133 staining in normal and tumor tissue.Results: We examined androgen/AR signaling in OVTC PA1 cells, a CSPCs-rich cell line, and found that AR, but not androgen,promoted cell growth. We also examined the effects of AR on CSPCs characteristics and found that AR expression was moreabundant in CD133+ cells, a well-defined ovarian cancer stem/progenitor marker, than in CD133− populations. Moreover,results of the sphere formation assay revealed that AR expression was required to maintain CSPCs populations. Interestingly,

Abbreviations: CSPCs, cancer stem/progenitor cells; OVTC, ovarian teratocarcinoma; AR, androgen receptor; DHT, 5α-dihydrotestosterone;DMEM, Dulbecco's modified Eagle medium; Ct, threshold value; hrEGF, human recombinant epidermal growth factor; hrbFGF, human recombinantbasic fibroblast growth factor.⁎ Corresponding authors at: Sex Hormone Research Center, Graduate Institution of Clinical Medical Science, School of Medicine, China Medical

University, Taichung 404, Taiwan.E-mail addresses: ych6375@gmail.com (Y.-C. Hung), maverick@mail.cmu.edu.tw (W.-L. Ma).

1 These authors contribute equally to this work.2 Dr. Wen-Lung Ma is the indicated author of communication.

http://dx.doi.org/10.1016/j.scr.2014.04.0031873-5061/© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/3.0/).

25Ligand-independent role AR in the development of OVTC

this AR-governed self-renewal capacity of CSPCs was only observed in CD133+ cells. In addition, we found that AR-mediatedCSPCs enrichment was accompanied by down-regulation of p53 and p16. Finally, co-expression of AR and CD133 was moreabundant in OVTC lesions than in normal ovarian tissue.Conclusion The results of this study suggest that AR itself might play a ligand-independent role in the development of OVTC.

© 2014 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/3.0/).

Introduction

Ovarian tumors are classified according to the assumed celltype of origin, namely surface epithelium (ovarian carcinoma),stroma (ovarian adenoma), salpin (fallopian tube cancer), andgerm cells (ovarian teratoma) (Zhang et al., 2008; Koshy et al.,2005; Wasim et al., 2009). Ovarian teratoma, a mixed germcell tumor, comprises approximately 20% of ovarian neoplasms(Sviracevic et al., 2011; Oliveira et al., 2004; Moniaga andRandall, 2011). Teratocarcinoma (OVTC), a very aggressiveovarian tumor, accounts for approximately 10–15% of germcell tumor (Koshy et al., 2005; Sviracevic et al., 2011; Oliveiraet al., 2004; Moniaga and Randall, 2011); however, thepathogenesis of this type of tumor is poorly understood. It iswell known that sex steroid hormones such as androgens,estrogens, and progesterone play pivotal roles in reproductivefunction throughout reproductive ages (Walters et al., 2008;Ahmad and Kumar, 2011); however, little is known about theirimpact on OVTC development and progression.

Cancer stem/progenitor cells (CSPCs) are thought of as aconfounding factor for carcinogenesis and cancer progressionbecause of their capacity for unlimited self-renewal. Studieshave shown that several glycoproteins, namely CD133 (Baba etal., 2009; Chen et al., 2014), CD117 (Luo et al., 2011), CD24(Gao et al., 2010), and CD44 (Meng et al., 2012), as well as thetranscription factors OCT-4 (Jiao et al., 2013) and Nanog (Leeet al., 2012), are markers of CSPCs in ovarian tissue. PA1, anOVTC cell line, is an excellent cell line for studying cancerstem cell characteristics because it abundantly expresses theCSPCs marker CD133 (Skubitz et al., 2013; Guo et al., 2011;Chung et al., 2013). In addition to using biomarkers todistinguish CSPCs, analysis of sphere-forming capacity is alsoa reliable method for detecting and characterizing CSPCspopulations (Chung et al., 2013; Williams et al., 2010; Szoteket al., 2006; Fong and Kakar, 2010).

Androgens play important roles in both male and femalereproductive organs (Chang et al., 2014; Fauser et al., 2011;Matsumoto et al., 2008). The androgens act through theandrogen receptor (AR), a transcription factor that belongs tothe nuclear receptor superfamily (Ma et al., 2014). AR exertsphysiological and pathological functions in organisms bytranslocating to the nucleus upon binding to androgens,where it binds to specific DNA sequences (Wang et al., 2005,2009; Ma et al., 2012a). However, non-classical androgen/ARaction has been documented in a variety of pathophysiologicalconditions (Simoncini and Genazzani, 2003; Baron et al., 2004;Singh et al., 2006; Bonaccorsi et al., 2006). There are twotypes of non-classical androgen/AR actions: ligand-mediatedtransient androgen/AR actions and ligand-independent ARactions (Fujimoto et al., 2001; Heinlein and Chang, 2002;Kousteni et al., 2001; Losel et al., 2003). Ligand-mediated

transient androgen/AR actions are stimulated by its binding toandrogen, while ligand-independent AR actions are mediatedby growth factors (Culig et al., 1994), or protein kinases (Lyonset al., 2008; Culig, 2004). However, both types of non-classical actions are the result of the translocation of AR to thenucleus where it acts as a DNA-binding transcription factorthat regulates target gene expression. In this study, weexamined whether androgen/AR signaling is involved in thedevelopment of OVTC in vitro and in humans.

Materials and methods

Ovarian teratocarcinoma patient data from a singlecohort study

Specimens of ovarian teratocarcinoma (n = 7) analyzed inthis study were obtained from patients who had received thediagnosis during the period 1987–2013 at the China MedicalUniversity (Taichung, Taiwan). Patients were identified froma single cohort registered in the Cancer Registry Database ofthe hospital. Access to the tissue samples was approved bythe Internal Review Board of the China Medical UniversityHospital (#DMR101-IRB2-276).

Cell culture

The human OVTC cell line (PA1), the human embryonic kidneycell line (HEK293T), and the human prostate adenocarcinomacell line (LnCap) were cultured in Dulbecco's modified Eaglemedium (DMEM) (GIBCO, CA, USA) with 10% fetal calf serum(FCS) (GIBCO, CA, USA) and 1% penicillin/streptomycin (GIBCO,CA, USA). PA1 and LnCap cells were provided courtesy ofDr. Min-Chie Hung (MD Anderson, TX, USA) and HEK293T cellswere obtained from Dr. Yuh-Pyng Shyr (Center of MolecularMedicine, China Medical University Hospital, Taichung,Taiwan). The cell lines were maintained at 37 °C in ahumidified atmosphere of 5% CO2.

Western blotting assay

Protein extraction and the immunoblot assay were performedas previously described (Ma et al., 2012b). Briefly, cells werewashed with 1×PBS and resolved in RIPA buffer (100 mM Tris,5 mM EDTA, 5% NP40; pH 8.0) with protease inhibitors (1 mMphenyl-methyl sulphonyl fluoride, 1 μg/ml aprotinin, 1 μg/mlleupeptin). Proteins were resolved by SDS–PAGE and thentransferred to PVDF membranes. Blocking of non-specificbinding was accomplished by adding 5% non-fat milk. Afterapplication of primary antibodies (AR, N-20 Santa Cruz, CA,USA; β-actin, Santa Cruz, CA, USA; p53 (FL-393), Santa Cruz,

26 W.-M. Chung et al.

CA, USA; p16 (F-12) Santa Cruz, CA, USA), secondary antibodies(1:3000, HRP-goat-anti-mouse and HRP-goat-anti-rabbit) wereapplied for 1 h at room temperature. Signals were enhancedusing an ECL chemiluminescence kit (Millipore, MA, USA) anddetected by ChemiDoc™ XRS+ (Bio-Rad, CA, USA).

Transfection and infection procedure

The lentiviral production and infection procedures were carriedout as reported previously, with minor modifications (Ma et al.,2012b; Lee et al., 2013). Briefly, cells were transfected withthe following lentivirus plasmids: psPAX2 packaging plasmid,pMD2G envelope plasmid (Addgen, MA, USA), pWPI-vector ctrl,and pWPI-hAR. Lentiviral plasmids carrying the GFP gene wereco-transfected with psPAX2 and pMD2G into HEK293T cells at aratio of 3:2:4 by lipofectamine 2000 (Invitrogen, CA, USA) asper the manufacturer's instructions. After 6 h, the medium wasreplaced with fresh DMEM/10% FBS medium and cells weremaintained at 37 °C in a humidified incubator in an atmosphereof 5% CO2 for 48 h. Medium containing virus was collected bycentrifugation and filtered through a 0.45 μm filter. Mediumcontaining 0.8 mg/ml polybrene (Sigma-Aldrich, MO, USA) wasthen added to culture dishes containing 106 PA1 cells. After16 h of infection, the medium containing virus was replacedwith fresh DMEM/10% FBSmedium and cells weremaintained at37 °C in a humidified incubator in an atmosphere of 5% CO2 for48 h. Infected cells were then collected and analyzed. Thegreen fluorescence protein (GFP)+ cells were measured byflowcytometry (BD, CA, USA, LSR II Flow Cytometry) todetermine infection efficiencies. GFP+ cells with infectionefficiencies greater than 85% were subjected to the followingexperiments.

Total RNA isolation and cDNA synthesis

RNA was extracted from PA1 cells as reported previously (Maet al., 2012b). Briefly, cells that had reached 80–90%confluence in 100-mm dishes were lysed with 1 ml Trizol(Invitrogen, CA, USA). Phenol/chloroform was then added andRNA-rich layers were separated by centrifugation. Soluble RNAwas precipitated with 2-propanol. RNA was then rinsed with75% ethanol and then dissolved in RNase-free water. Forfirst-strand cDNA synthesis, 5 μg of total RNA was used toperform reverse transcription PCR by the PrimeScript™ RTreagent kit (TAKARA Bio Inc., Kyoto, Japan). cDNA wassynthesized according to the manufacturer's instructions.

Quantitative real-time PCR analysis

A real-time detection system (Bio-Rad Laboratories, Inc.,CA, USA) and the KAPA™ SYBR FAST One-Step qRT-PCR Kit(KAPABIOSYSTEMS, MA, USA) were used according to themanufacturers' instructions. Relative gene expression wasdetermined by normalizing the expression level of the targetgene to the expression level of a housekeeping gene(β-actin). Threshold value (Ct) dynamics were used (2−ΔΔCt)for quantitation of gene expression. The qRT-PCR primersequences were as follows: CD133 forward 5′-TCT CTA TGTGGT ACA GCC G-3′, reverse 5′-TGA TCC GGG TTC TTA CCTG-3′; CD24 forward 5′-TTT ACA ACT GCC TCG ACA CAC ATAA-3′, reverse 5′-CCC ATG TAG TTT TCT AAA GAT GGA A-3′;

CD44 forward 5′-GAC CTC TGC AAG GCT TTC AA-3′, reverse5′-TCC GAT GCT CAG AGC TTT CTC-3′; CD117 forward 5′-CAAGGA AGG TTT CCG AAT GC-3′, reverse 5′- CCA GCA GGT CTTCAT GAT GT-3′; Nanog forward 5′-GGG CCT GAA GAA AACTAT CCA TCC-3′, reverse 5′-TGC TAT TCT TCG GCC AGT TGTTTT-3′; OCT-4 forward 5′-GGC CCG AAA GAG AAA GCG AACC-3′, reverse 5′-ACC CAG CAG CCT CAA AAT CCT CTC-3′; SCFforward 5′-ACT GAC TCT GGA ATC TTT CTC AGG-3′, reverse5′-GAT GTT TTG CCA AGT CAT TGT TGG-3′.

Cell growth analysis

The WST-1 assay (Roche, Basel, Switzerland) was used toassess cell growth. Briefly, 103 cells/100 μl/well were seededin 96-well plates with DMEM in 10% charcoal-dextran treatedFBS (CDFBS) and were incubated for designated time periods(1, 2, 4, 6, 8 days). Then, 10 μl of WST-1 solution was added toeach well and cells were allowed to incubate at 37 °C in anincubator for an hour. Cell viability was quantified bycolorimetric detection in an ELISA plate reader (BECKMAN, Il,USA, COULTER PARADIGM™ Detection Platform) at an absor-bance of 450 nm and 690 nm to generate an OD proportionalto the relative abundance of live cells in the given wells.

Colony formation assay and Standard cellnumber count

Two sets of 1.5 × 105 cells/dish were seeded in 6-cm plateswith DMEM in 10% CDFBS and incubated for 8 days. In one set ofcells, 1000 μl of 4% formaldehyde solution was added to fixedcells and the cells were allowed to incubate at roomtemperature for one hour. Crystal violet cell staining wasthen performed. After 1 h, crystal violate was washed from thecell culture dish and cell colonies were photographed. Theother set of cells were subjected for cell counting.

Luciferase assay

The assay was performed as previously described (Lyons etal., 2008). Briefly, pGL3-ARE (androgen responseelement-driven luciferase reporter plasmid) and pRL-TK(thymidine kinase promoter-driven renilla luciferase plas-mid) were transiently co-transfected into cells. After 6 h,medium was replaced with fresh medium and 10% CDFBS.Cells were then cultured for 48 h with or without DHT(10 nM). After 24 h, cells were washed with 1×PBS and thenincubated in the presence of 100 μl CCLR (cell culture lysisreagent) (Promega, WI, USA) at room temperature for30 min. Cell lysates were then placed in a microtube andcentrifuged at 14,000 rpm for 5 min. Supernatant (5 μl) wasthen mixed with 50 μl luciferase assay reagent. Luciferaseactivity was measured immediately using a luminescencemicroplate reader and presented as relative luminescenceunits.

Side population analysis

Side population analysis was preformed as described previ-ously by Goodell (Hirschmann-Jax et al., 2004). Briefly, 106

cells were detached and washed, and then incubated in DMEM

27Ligand-independent role AR in the development of OVTC

containing 2% FCS and 5 μg/ml Hoechst33342 dye (Sigma-Aldrich, MO, USA) for 90 min at 37 °C, either alone or in thepresence of 50 μM verapamil (Sigma-Aldrich, MO, USA). Aftercertain incubation time periods, cells were washed withice-cold 1× PBS and then resuspended with 1× PBS supple-mented with 2% FCS and 2 μg/ml propidium iodide (Sigma-Aldrich, MO, USA) at 4 °C for 5 min, to exclude dead cells. Thecells were analyzed using flowcytometry (BD, CA, USA, LSR IIFlowCytometry) with dualwavelength analysis (Hoechst-blue,424–444 nm; Hoechst-red, 675 nm) after excitation with350 nm UV light.

Verification and sorting of CD133+ cells

Cells were detached with 1% trypsin/EDTA and cell membranenon-specific binding to antibodies was blocked by 5% BSA for30 min at room temperature. Cells were then incubated withantibody CD133-APC (MACS, Miltenyi Biotech, Gladbach,Germany) for 30 min and then subjected to flowcytometry(FACS, BD, CA, USA, FACSAria) analysis. The CD133+/− cellswere collected by a cell sorter (BD, CA, USA, FACSAria).

Figure 1 AR promotes OVTC PA1 cell growth. A. Western blottinhigher in PA1 cells infected with AR cDNA (pWPI-hAR; AR) than in concontrol. The graph on the right-hand side is quantitation data of Aassays were performed at the indicated time points (1, 2, 4, 6, 8 daywave length at 450 nm deducted from that of 630 nm background repromotes cell growth. Cell number count was analyzed and photograAR expression was significantly lower in PA1 cells infected withLuciferase; siLuc). LnCap cells served as a positive control. The graexpression. E. AR siRNA suppressed cell growth. WST-1 assays wereY-axis indicates the absolute Abs. of siLuc and siAR infected cellsanalyzed and photographed on day 8. The results shown are from 3 inthan 0.05.

Sorting of CD133+ and CD133− populations

A CD133 MicroBead kit (Miltenyi Biotec, Gladbach, Germany)was used to sort PA1 CD133+ cells according to themanufacturer's instructions. Briefly, 108 cells were washed,resuspended in 300 μl of buffer (1X PBS/pH 7.2, 0.5% BSA/2 mM EDTA) with 100 μl FcR Blocking Reagent and 100 μl ofCD133 MicroBeads, mixed well, and then incubated for 30 minat 2–8 °C. The labeled cells were then washed, resuspendedin 500 μl buffer, and then applied to a MACS Column to collectunlabeled cells (CD133− cells). Finally, the column wasremoved from the magnet and placed in collection tubes tocollect magnetically-labeled cells (CD133+ cells).

Sphere formation assay

Cells were collected and washed to remove serum and thensuspended in serum-free DMEM supplemented with 20 ng/mlhuman recombinant epidermal growth factor (hrEGF), 10 ng/mlhuman recombinant basic fibroblast growth factor (hrbFGF),

g (left-hand side) showed that AR expression was significantlytrol cells (pWPI-vector ctrl; Vec). LnCap cells served as a positiveR protein expression. B. AR cDNA promotes cell growth. WST-1s). The Y-axis indicates the absolute light absorbance (Abs.; Abs.adings). Vec and AR infected cells were compared. C. AR cDNAphed on day 8. D. Western blotting (left-hand side) showed thatAR siRNA (pLKO.1-siAR; siAR) than in control cells (pLKO.1-siph on the right-hand side is the quantitation data of AR proteinperformed at the indicated time points (1, 2, 4, 6, 8 days). The. F. AR siRNA suppressed cell growth. Cell number count wasdependent experiments. *Indicates significance at a p-value less

Figure 2 Ligand-independent AR effects on OVTC PA1 cellgrowth. A. DHT effect on growth of PA1 cells. WST-1 assay wasperformed on DHT 10 nM treated PA1 cells at indicated time points(1, 2, 4, 6, 8 days). The Y-axis indicates the absorbance (Abs.).B. Endogenous AR transactivation activity was absent in PA1 cells.ARE-luciferase assaywas performed in DHT 10 nM treated PA1 cells.Thymidine kinase promoter-driven Renila construct (pRL-TK) wasco-transfected and measured as a transfection control. RelativeARE-luciferase activity (fold to EtOH) is presented as transactivationactivity of AR. AR transactivation function in LnCap cells served as apositive control. C. Exogenous AR transactivation was absent in PA1cells. AR cDNA (pWPI-hAR; AR) or vector control (pWPI-vector ctrl;Vec) was co-transfected with ARE-luciferase and pRL-TK in PA-1cells to observe ARE-luciferase activity. The transfection was alsoperformed in 293T cells as a positive control. The results shown arefrom 3 independent experiments. *Indicates a p-value less than0.05.

28 W.-M. Chung et al.

5 μg/ml insulin, and 0.4% BSA (Sigma, USA). The cells weresubsequently cultured in ultra low attachment 6-well plates(Corning Inc., Corning, NY, USA) at a density of less than 5000cells/well for 14 days. Spheres were observed under amicroscope and images were photographed under a phasecontrast fluorescence microscope (Nikon, ECLIPSE 80i).

Statistical analysis

Statistical analyses were performed using the Student'st-test. All experiments were repeated at least three times,and p-values less than 0.05 were considered to indicatestatistical significance.

Results

Ligand-independent AR actions promote OVTCcell growth

In order to test whether androgen/AR signaling plays a rolein the development of OVTC, we first used a lentiviral-baseddelivery system to engineer AR-knockdown PA1 cells and PA1cells that overexpress AR. Results of Western blot analysisshowed that AR protein levels were approximately 40-foldhigher in PA1 cells that were engineered to overexpress ARthan in vector control cells (Fig. 1A). In addition, we foundthat cells that overexpressed AR grew more rapidly thanvector control cells (Figs. 1B and C). In contrast, expressionof AR (Fig. 1D) and rate of growth (Figs. 1E and F) weresignificantly lower in AR-knockdown PA1 cells than in non-specific siRNA control cells.

We also tested the effects of androgen (5α-dihydrotes-tosterone; DHT 10 nM) on PA1 cell growth. As seen in Fig. 2A,DHT had little effect on PA1 cell growth. In order to betterunderstand ligand-independent AR functions, ARE (androgenreceptor response element) promoter assays were per-formed. As shown in Fig. 2B, DHT treatment did not induceluciferase activity in PA1 cells (Fig. 2B, 3rd vs. 4th lane) butresulted in significant activity in LnCaP cells (Fig. 2B, 1st vs.2nd lane). We further ectopically introduced wild type ARcDNA into PA1 cells to observe AR transactivation function(Fig. 2C). The result showed that luciferase activity was notinduced in pWPI-vector control cells or in pWPI-ARtransfected cells, although it resulted in marked luciferaseactivity in pWPI-AR transfected 293T cells.

The data in Figs. 1 and 2 demonstrate that AR promotesOVTC cell growth through a ligand-independent pathway.

AR promotes CSPCs characteristics

Studies have shown that CSPCs can trigger tumorigenesis andpromote cancer progression and that AR can promote thegrowth of CSPCs in various cancers. We, therefore, testedwhether AR regulates CSPCs in OVTC. We performed a cell-sorting assay to isolate PA1 cells that expressed theglycoprotein CD133, a well-established biomarker of CSPCsin ovarian cancer. Using a cell sorter and CD133 antibody, wewere able to distinguish between CD133− and CD133+ cellpopulations (Fig. 3A). The 5% extremes of the staining signalwithin the spectrum were defined as CD133+ and CD133−

cells. Then we confirmed CSPCs characteristics by examiningother CSPCs marker genes (Fig. 3B). We found that the levelsof expression of CD133, CD24, OCT-4, Nanog, CD117, and

Figure 3 AR over-expressed in PA1 CD133+ cells. A. Definition of PA1 CD133+/− cells using flowcytometry. The basalIgG-isotype-APC staining signal is presented on the left-hand side, and the CD133-APC staining signal is presented on the right-handside of the histogram. As indicated in P2 and P3, the 5% extremes of the staining signal within the spectrum were defined as CD133+and CD133− cells. Negatively stained cells (P2, left-hand side 5%) and positively stained cells (P3, right-hand side 5%) were subjectedto the cell sorter for analysis. B. CSPCs marker genes were more highly expressed in CD133+ cells than in CD133− cells. CD133, CD24,OCT-4, Nanog, CD117, and CD44 mRNA expressions were measured by quantitative real-time PCR, and the relative expressions werecalculated as 2−ΔΔCt. C. AR protein was more highly expressed in CD133+ cells than in CD133− cells. Western blotting assays (upperpanel) were performed to detect the expression of AR, and quantified data are shown in the lower panel. All data are from at leastthree independent experiments. The variations are due to standard deviation, and * indicates statistical significance at p b 0.05.

29Ligand-independent role AR in the development of OVTC

CD44 were markedly higher in CD133+ cells than in CD133−cells. We then measured AR expression and found that ARexpression was significantly higher in CD133+ cells than inCD133− cells (Fig. 3C; 2nd vs. 3rd lane).

The data suggest that AR acts as a cancer progressionpromoter in CSPCs. In order to further investigate this finding,we performed a side population assay (Williams et al., 2010;Szotek et al., 2006) in PA1 cells that overexpressed AR cDNAand in AR-knockdown PA1 cells. Using verapamil, a calciumchannel blocker, to differentiate stem populations, we foundthat overexpression of AR cDNA increased the side populationfrom 2.3% to 7% (Fig. 4A) and that AR siRNA knockdowndecreased the side population from 2.3% to 1% (Fig. 4B).Moreover, we used flowcytometry to examine the effects ofAR on CD133 populations. As shown in Figs. 4C and D,overexpression of AR resulted in an increase in CD133+ cells(7.5% to 43%) and knockdown of AR by siRNA resulted in adecrease in CD133+ cells (7.7% to 1.5%). Furthermore, weused the sphere formation assay to confirm the role that ARplays in CSPCs characteristics. As shown in Fig. 4E, AR cDNAresulted in an increase in sphere number and size, while ARsiRNA had the opposite effect (Fig. 4F).

Since we observed that AR promotes CSPCs enrichment,we conducted further studies to better understand its rolein (progenitor) CSPCs and (non-progenitor) non-CSPCs. Usinga lentiviral-based gene delivery, we measured gene trans-duction efficiencies with N95% GFP+ cells using the pWPIGFP-carrying vector (Fig. 5A). Since the pLKO.1 vectordoes not carry the GFP gene, we treated PA1 cells withpuromycin (5 μM) for 2 weeks and then selectedpuromycin-resistant cells. Then we sorted out CD133+/−cells to perform the sphere formation assay. As shown inFig. 5B, CD133+ cells formed larger spheres than CD133−cells (Figs. 5B and C). Knockdown or overexpression of AR inCD133− cells showed little effect on sphere formation;however, knockdown of AR suppressed (Fig. 5B) andoverexpression of AR dramatically enhanced (Fig. 5C) sphereformation in CD133+ cells.

It has been documented that regulation of p53–p21 and Rb–p16 tumor suppression pathways are associated with cellularsenescence (Wu et al., 2011; Chen et al., 2011; Serrano et al.,1997) and cancer stemness (Armesilla-Diaz et al., 2009; Zhenget al., 2008; Akala et al., 2008; Strauss et al., 2012) bycontrolling cell cycle. Furthermore, our previous study

Figure 4 AR enriches CSPCs populations. A. AR cDNA increased side population (SP) of PA1 cells. The SP cells were defined byHoechst staining with or without verapamil treatment (circled area) in pWPI-vector control (Vec) and pWPI-hAR (AR) infected PA1cells (dot-plot graphs on the left). The percentage of SP cells is shown on the right-hand side. B. AR siRNA knockdown decreasedside-population (SP) of PA1 cells. The SP cells were defined by Hoechst staining with or without verapamil treatment (circled area) inpLKO.1-si Luciferase (siLuc) and pLKO.1-siAR (siAR) infected PA1 cells (dot-plot graphs on the left). The percentage of SP cells isshowed on the right-hand side C. AR enhanced CD133 expression and population. Flowcytometry was performed using CD133-APCconjugated antibody to measure Vec- and AR-infected PA1 CD133 expression (histogram on the left), and CD133 population(right-hand side of the bar graphs). D. AR knockdown suppressed CD133 expression and population. Flowcytometry was performed tomeasure siLuc- and siAR-infected PA1 CD133 expression (histogram on the left), and CD133 populations (right-hand side bar graphs).E. AR promotes CSPCs sphere formation. Sphere images of Vec- and AR-infected PA1 cells (day-1 and day-14) were taken by aphase-contrast microscope (100×) and are shown on the left-hand side. The quantitation of sphere numbers is shown on theright-hand side bar graph. F. AR knockdown suppressed CSPCs sphere formation. Sphere images of siLuc- and siAR-infected PA1 cells(day-1 and day-14) were taken by a phase-contrast microscope (100×) and are shown on the left-hand side. The quantitation of spherenumbers is shown on the right-hand side bar graph. All experiments were from at least 3 independent experiments where * indicatesp-values less than 0.05.

30 W.-M. Chung et al.

indicated that AR down-regulated p53 expression in hepato-cellular carcinoma (Ma et al., 2008). Therefore, we performedimmunoblot assays to see whether this was the case inOVTC CSPCs. As shown in Fig. 5D, we found that p53 and p16were downregulated by AR in CD133+ cells, which mightindicate that cancer stem cell stemness is associated withthe down regulation of p53 and p16 (4th vs. 3rd lane, and4th vs. 2nd lane). Meanwhile, this data was consisted withprevious studies that inactivation of both p53–p21 and p16–Rbsignaling pathway are involved in CSPCs progression (Chenet al., 2011).

Taken together, the results presented in Figs. 3 to 5demonstrate that AR promotes self-renewal of CSPCs inCD133+ PA1 cells.

Co-expression of CD133 and AR in OVTC tissue

The in vitro data show that AR promotes OVTC cell growth bypromoting self-renewal of CSPCs. We further examined therole AR has on CD133 expression in human OVTC tissuesamples using an immunohistochemical approach. Expres-sion levels of AR and CD133 in cancer tissue were comparedwith those in normal ovarian tissue taken from the samepatient. As seen in Fig. 6, it is markedly lower expression ofAR and fewer co-expression of CD133 in normal ovarianepithelial cells (Figs. 6A and C) compared to that in ovariantumor epithelial cells. Furthermore, the co-expression of ARand CD133 were more obviously in cytoplasm than nucleus(Figs. 6B and D). Those observations strengthen our

Figure 5 AR promotes CD133 CSPCs self-renew. A. AR cDNA infection efficiency in CD133− and CD133+ PA1 cells. The greenfluorescence protein (GFP) histograms (left-hand side) of infected or non-infected cells were measured by flowcytometry. Thequantitation of GFP populations is presented on the right-hand side. B. AR knockdown sphere images of the CD133+ and CD133− PA1cell colonies. The sphere images at day 14 were taken by a phase contrast microscope (40×). The images of CD133− and CD133+ cellsinfected with siLuc (a, c) or siAR (b, d) were presented. C. AR over expression sphere images of the CD133+ and CD133− PA1 cellcolonies. The sphere images were taken on day 14 by a phase contrast microscope (40×). The images of CD133− and CD133+ infectedwith Vec (a, c) and AR cDNA (b, d) were showed. D. Protein expression of p53, and p16 in CD133+ and CD133− cells by Westernblotting. The expression of actin served as a loading control. The relative band density (fold) quantitation of Western blots images areshown below. All data were from at least three independent experiments. The variations were from the calculation of standarddeviation and the scale bar inlet indicated 50 μm. *Indicates significance at a p-value less than 0.05.

31Ligand-independent role AR in the development of OVTC

hypothesis that ligand-independent AR function involved inthe disease progression.

Discussion

Ligand-independent AR function in cancer growthand CSPCs self-renewal

In this study, we have demonstrated the first in vitroevidence that AR promotes OVTC development through aligand-independent pathway.

The functions of AR in cancer development have mostlybeen studied in prostate cancer (Heinlein and Chang, 2004).The current therapeutic paradigm for prostate cancer isandrogen ablation therapy (Pienta and Bradley, 2006).However, some studies have shown that ligand-independentactivation of AR is responsible for relapsed prostate cancerafter anti-androgen therapy (Culig, 2004; Taplin and Balk,2004; Jenster, 1999, 2000). Studies in prostate cancer cellsreported that AR transactivation was mediated by growthfactors, such as IGF-I, KGF, EGF (Culig et al., 1994), and theCXCL12/CXCR14 axis (Kasina and Macoska, 2012) in theabsence of androgen. Kasina and Mascoska's (2012)studyillustrated that intracellular signals of those growth factors

Figure 6 AR and CD133 expression in human ovarian teratocarcinoma tissues. A, B. AR expression in normal (A) and OVTC (B) tissuesof same patient (×40). C, D. Serial sections of CD133 expressions in normal (C) and OVTC (D) tissues of same patient (×40). Blue colorrepresents the nucleus and brown color represents AR antibody staining or CD133 antibody staining.

32 W.-M. Chung et al.

i.e. Rho GTPase (Lyons et al., 2008), MAPK, or AKT pathway(Feldman and Feldman, 2001) play a role in the ligand-independent function of AR in prostate cancer cells. There-fore, ligand-independent functions of AR appear to betriggered by intracellular signaling molecules that translocateAR into the nucleus (Culig et al., 1994; Lyons et al., 2008;Kasina and Macoska, 2012; Feldman and Feldman, 2001). Theresults of those studies imply that targeting of those signals aswell as AR would improve the effectiveness of therapy foradvanced prostate cancer. In addition, our previous workdemonstrated that AR enhanced CSPCs in specimens ofendometrial cancer associated with Cisplatin resistance(Chen et al., 2014).

In this study, we found two interesting, yet previouslyunreported phenomena: AR can exert biological functions inOVTC cells in a ligand-independent manner and AR functioncan promote cancer cell growth through a pathway that doesnot involve transactivation. One possibility of this ligand-independent AR function might be that AR acts as transcrip-tional factor co-factor to regulate gene expressions.

Pathophysiological roles of AR in OVTC

Few studies have provided evidence that androgen/AR play arole in OVTC. Previous studies (Chadha et al., 1993; Cardillo et

al., 1998; Shintaku et al., 2011) showed that 84% of benign andmalignant ovarian tumors expressed AR; yet, there is no cleardistinction between AR expression and ovarian cancers ofdifferent cellular origins. Because of the rarity of OVTC, fewstudies have been conducted on the pathogenesis of thedisease. Herein, we have demonstrated that AR promotes thedevelopment and progression of OVTC by influencing tumor-igenic CD133+ CSPCs populations. Furthermore, this is thefirst study showing that a male hormone receptor promotesovarian cancer in a ligand-independent manner.

p53 and p16 are involved inligand-independent-AR-regulatedOVTC CSPC progression

The tumor suppressor protein p53 plays a pivotal role in cellcycle arrest, senescence, DNA repair and programmed celldeath. An abundance of evidence shows that p53 plays acritical role in self-renewal and differentiation of variouscancer stem cells by controlling the cell cycle (Armesilla-Diazet al., 2009; Zheng et al., 2008; Akala et al., 2008).Furthermore, our previous study demonstrated that ARsuppressed p53 expression in murine hepatic cells and thatloss of hepatic AR promoted p53-mediated DNA damagesensing and repair systems as well as apoptosis (Ma et al.,

Figure 7 Illustration showing how AR promotes OVTC cell growth through CSPCs self-renewal. Ligand-independent AR function inCSPCs cells (e.g. CD133+ cells) facilitates OVTC cell growth. The expression of AR in CSPCs cells is accompanied by down-regulation ofp53 and p16.

33Ligand-independent role AR in the development of OVTC

2008). Our findings demonstrated that AR expression promotesself-renewal of CSPCs. Studies have shown that expression ofthe tumor suppressor gene p16 results in cell cycle arrest byinhibiting CDK4 and CDK6 (Akala et al., 2008; Strauss et al.,2012). Its downregulation was also reported to be essential forcellular quiescence in cancer stem cells (Wu et al., 2011; Chenet al., 2011; Armesilla-Diaz et al., 2009; Strauss et al., 2012).Notably, the p53–p21 and p16–Rb pathways are two importantarms converging cellular senescence that attenuates tumori-genicity while they were inactivated (Wu et al., 2011; Chenet al., 2011; Serrano et al., 1997; Armesilla-Diaz et al., 2009;Zheng et al., 2008; Akala et al., 2008).

In this study, we found that overexpression of AR inCD133+ cells was associated with down-regulation of p53and p16 (Fig. 5D). This result not only supports our previousfinding that AR plays an important role in the progression ofhepatoma (Ma et al., 2008), but also provides evidence thatCSPCs progression is mediated by ligand-independentAR-p53/p16 signaling. Taken together, our data indicatethat AR controls OVTC CSPC progression by regulating, in aligand-independent manner, p53- and p16-related events,such as cell cycle regulation.

Conclusion

The present study provides evidence that expression of AR inthe absence of increased AR-mediated transcription exertsbiological functions which indicates that ligand-independentAR function in CSPCs (e.g. CD133+ cells) facilitates OVTCcell growth, and that the expression of AR in CSPCs isaccompanied by down-regulation of p53 and p16 (Fig. 7).Our data not only demonstrate a novel mechanism by whichAR regulates cancer cell behavior, but also providespathological insights into germ cell tumors.

Contribution of authors

W.M. Chung and W.C. Chang executed the experiments andcomposed the draft of the article. L. Chen participated inthe experimental design and provided critical scientificopinions of this work. T.Y. Lin and L.C. Chen participated in

immunohistochemistry studies and provided histology andpathology suggestions. W.L. Ma and Y.C. Hung designed,supported, and edited the article.

Acknowledgments

We are grateful for Dr. Min-Chie Hung and Yuh-Pyng Sher'sgenerosity for providing cell lines, and also thanks to JeffreyConrad for language enhancement. This work was supportedby the National Science Council grants: NSC101-2314-B-039-027-MY3 and NSC101-2314-B-039-014; and Taiwan Ministryof Health and Welfare Clinical Trial and Research Center ofExcellence (DOH102-TD-B-111-004).

References

Ahmad, N., Kumar, R., 2011. Steroid hormone receptors in cancerdevelopment: a target for cancer therapeutics. Cancer Lett.300, 1–9.

Akala, O.O., Park, I.K., Qian, D., Pihalja, M., Becker, M.W., Clarke,M.F., 2008. Long-term haematopoietic reconstitution byTrp53−/−p16Ink4a−/−p19Arf−/− multipotent progenitors. Na-ture 453, 228–232.

Armesilla-Diaz, A., Bragado, P., Del Valle, I., Cuevas, E., Lazaro, I.,Martin, C., Cigudosa, J.C., Silva, A., 2009. p53 regulates the self-renewal and differentiation of neural precursors. Neuroscience158, 1378–1389.

Baba, T., Convery, P.A., Matsumura, N., Whitaker, R.S., Kondoh, E.,Perry, T., Huang, Z., Bentley, R.C., Mori, S., Fujii, S., et al.,2009. Epigenetic regulation of CD133 and tumorigenicity ofCD133+ ovarian cancer cells. Oncogene 28, 209–218.

Baron, S., Manin, M., Beaudoin, C., Leotoing, L., Communal, Y.,Veyssiere, G., Morel, L., 2004. Androgen receptor mediatesnon-genomic activation of phosphatidylinositol 3-OH kinasein androgen-sensitive epithelial cells. J. Biol. Chem. 279,14579–14586.

Bonaccorsi, L., Marchiani, S., Ferruzzi, P., Muratori, M., Crescioli,C., Forti, G., Maggi, M., Baldi, E., 2006. Non-genomic effects ofthe androgen receptor and vitamin D agonist are involved insuppressing invasive phenotype of prostate cancer cells. Steroids71, 304–309.

Cardillo, M.R., Petrangeli, E., Aliotta, N., Salvatori, L., Ravenna, L.,Chang, C., Castagna, G., 1998. Androgen receptors in ovariantumors: correlation with oestrogen and progesterone receptors

34 W.-M. Chung et al.

in an immunohistochemical and semiquantitative image analysisstudy. J. Exp. Clin. Cancer Res. 17, 231–237.

Chadha, S., Rao, B.R., Slotman, B.J., van Vroonhoven, C.C., van derKwast, T.H., 1993. An immunohistochemical evaluation ofandrogen and progesterone receptors in ovarian tumors. Hum.Pathol. 24, 90–95.

Chang, C., Lee, S.O., Wang, R.S., Yeh, S., Chang, T.M., 2014.Androgen receptor (AR) physiological roles in male and femalereproductive systems: lessons learned from ARKO mice lackingAR in selective cells. Biol. Reprod. 21 (3), 386–394.

Chen, H., Zhou, L., Dou, T., Wan, G., Tang, H., Tian, J., 2011.BMI1'S maintenance of the proliferative capacity of laryngealcancer stem cells. Head Neck 33, 1115–1125.

Chen, L., Chang, W.C., Hung, Y.C., Chang, Y.Y., Bao, B.Y., Huang,H.C., Chung, W.M., Shyr, C.R., Ma, W.L., 2014. Androgenreceptor increases CD133 expression and progenitor-like popu-lation that associate with cisplatin resistance in endometrialcancer cell line. Reprod. Sci. 21 (3), 386–394.

Chung, W.M., Chang, W.C., Chen, L., Chang, Y.Y., Shyr, C.R., Hung,Y.C., Ma, W.L., 2013. MicroRNA-21 promotes the ovarianteratocarcinoma PA1 cell line by sustaining cancer stem/progenitor populations in vitro. Stem Cell Res. Ther. 4, 88.

Culig, Z., 2004. Androgen receptor cross-talk with cell signallingpathways. Growth Factors 22, 179–184.

Culig, Z., Hobisch, A., Cronauer, M.V., Radmayr, C., Trapman, J.,Hittmair, A., Bartsch, G., Klocker, H., 1994. Androgen receptoractivation in prostatic tumor cell lines by insulin-like growthfactor-I, keratinocyte growth factor, and epidermal growthfactor. Cancer Res. 54, 5474–5478.

Fauser, B.C., Laven, J.S., Tarlatzis, B.C., Moley, K.H., Critchley, H.O.,Taylor, R.N., Berga, S.L., Mermelstein, P.G., Devroey, P.,Gianaroli, L., et al., 2011. Sex steroid hormones and reproductivedisorders: impact on women's health. Reprod. Sci. 18, 702–712.

Feldman, B.J., Feldman, D., 2001. The development of androgen-independent prostate cancer. Nat. Rev. Cancer 1, 34–45.

Fong, M.Y., Kakar, S.S., 2010. The role of cancer stem cells and theside population in epithelial ovarian cancer. Histol. Histopathol.25, 113–120.

Fujimoto, N., Mizokami, A., Harada, S., Matsumoto, T., 2001.Different expression of androgen receptor coactivators in humanprostate. Urology 58, 289–294.

Gao, M.Q., Choi, Y.P., Kang, S., Youn, J.H., Cho, N.H., 2010. CD24+cells from hierarchically organized ovarian cancer are enrichedin cancer stem cells. Oncogene 29, 2672–2680.

Guo, R., Wu, Q., Liu, F., Wang, Y., 2011. Description of the CD133+subpopulation of the human ovarian cancer cell line OVCAR3.Oncol. Rep. 25, 141–146.

Heinlein, C.A., Chang, C., 2002. The roles of androgen receptors andandrogen-binding proteins in nongenomic androgen actions. Mol.Endocrinol. 16, 2181–2187.

Heinlein, C.A., Chang, C., 2004. Androgen receptor in prostatecancer. Endocr. Rev. 25, 276–308.

Hirschmann-Jax, C., Foster, A.E., Wulf, G.G., Nuchtern, J.G., Jax,T.W., Gobel, U., Goodell, M.A., Brenner, M.K., 2004. A distinct“side population” of cells with high drug efflux capacity in humantumor cells. Proc. Natl. Acad. Sci. U. S. A. 101, 14228–14233.

Jenster, G., 1999. The role of the androgen receptor in the developmentand progression of prostate cancer. Semin. Oncol. 26, 407–421.

Jenster, G., 2000. Ligand-independent activation of the androgenreceptor in prostate cancer by growth factors and cytokines. J.Pathol. 191, 227–228.

Jiao, J., Huang, L., Ye, F., Shi, M., Cheng, X., Wang, X., Hu, D., Xie,X., Lu, W., 2013. Cyclin D1 affects epithelial–mesenchymaltransition in epithelial ovarian cancer stem cell-like cells. Onco.Targets Ther. 6, 667–677.

Kasina, S., Macoska, J.A., 2012. The CXCL12/CXCR4 axis promotesligand-independent activation of the androgen receptor. Mol.Cell. Endocrinol. 351, 249–263.

Koshy, M., Vijayananthan, A., Vadiveloo, V., 2005. Malignantovarian mixed germ cell tumour: a rare combination. Biomed.Imaging Intervention J. 1, e10.

Kousteni, S., Bellido, T., Plotkin, L.I., O'Brien, C.A., Bodenner, D.L.,Han, L., Han, K., DiGregorio, G.B., Katzenellenbogen, J.A.,Katzenellenbogen, B.S., et al., 2001. Nongenotropic, sex-nonspecific signaling through the estrogen or androgen receptors:dissociation from transcriptional activity. Cell 104, 719–730.

Lee, M., Nam, E.J., Kim, S.W., Kim, S., Kim, J.H., Kim, Y.T., 2012.Prognostic impact of the cancer stem cell-related marker NANOG inovarian serous carcinoma. Int. J. Gynecol. Cancer 22, 1489–1496.

Lee, S.O., Ma, Z., Yeh, C.R., Luo, J., Lin, T.H., Lai, K.P., Yamashita,S., Liang, L., Tian, J., Li, L., et al., 2013. New therapy targetingdifferential androgen receptor signaling in prostate cancerstem/progenitor vs. non-stem/progenitor cells. J. Mol. CellBiol. 5, 14–26.

Losel, R.M., Falkenstein, E., Feuring, M., Schultz, A., Tillmann, H.C.,Rossol-Haseroth, K., Wehling, M., 2003. Nongenomic steroid action:controversies, questions, and answers. Physiol. Rev. 83, 965–1016.

Luo, L., Zeng, J., Liang, B., Zhao, Z., Sun, L., Cao, D., Yang, J.,Shen, K., 2011. Ovarian cancer cells with the CD117 phenotypeare highly tumorigenic and are related to chemotherapyoutcome. Exp. Mol. Pathol. 91, 596–602.

Lyons, L.S., Rao, S., Balkan, W., Faysal, J., Maiorino, C.A.,Burnstein, K.L., 2008. Ligand-independent activation of andro-gen receptors by Rho GTPase signaling in prostate cancer. Mol.Endocrinol. 22, 597–608.

Ma, W.L., Hsu, C.L., Wu, M.H., Wu, C.T., Wu, C.C., Lai, J.J., Jou, Y.S.,Chen, C.W., Yeh, S., Chang, C., 2008. Androgen receptor is a newpotential therapeutic target for the treatment of hepatocellularcarcinoma. Gastroenterology 135, 947–955 (955 e941-945).

Ma, W.-L., Yeh, C.-C., Jeng, L.-B., Chang, C., 2012a. Androgen andandrogen receptor signals jamming monocyte/macrophage func-tions in premalignant phase of livers. Biomedicine 2, 155–159.

Ma, W.L., Hsu, C.L., Yeh, C.C., Wu, M.H., Huang, C.K., Jeng, L.B.,Hung, Y.C., Lin, T.Y., Yeh, S., Chang, C., 2012b. Hepaticandrogen receptor suppresses hepatocellular carcinoma metas-tasis through modulation of cell migration and anoikis.Hepatology 56, 176–185.

Ma, W.L., Lai, H.C., Yeh, S., Cai, X., Chang, C., 2014. Androgenreceptor roles in hepatocellular carcinoma, fatty liver, cirrhosisand hepatitis. Endocr. Relat. Cancer. http://dx.doi.org/10.1530/ERC-13-0283.

Matsumoto, T., Shiina, H., Kawano, H., Sato, T., Kato, S., 2008.Androgen receptor functions in male and female physiology. J.Steroid Biochem. Mol. Biol. 109, 236–241.

Meng, E., Long, B., Sullivan, P., McClellan, S., Finan, M.A., Reed, E.,Shevde, L., Rocconi, R.P., 2012. CD44+/CD24− ovarian cancercells demonstrate cancer stem cell properties and correlate tosurvival. Clin. Exp. Metastasis 29, 939–948.

Moniaga, N.C., Randall, L.M., 2011. Malignant mixed ovarian germcell tumor with embryonal component. J. Pediatr. Adolesc.Gynecol. 24, e1–e3.

Oliveira, F.G., Dozortsev, D., Diamond, M.P., Fracasso, A.,Abdelmassih, S., Abdelmassih, V., Goncalves, S.P., Abdelmassih,R., Nagy, Z.P., 2004. Evidence of parthenogenetic origin ofovarian teratoma: case report. Hum. Reprod. 19, 1867–1870.

Pienta, K.J., Bradley, D., 2006. Mechanisms underlying thedevelopment of androgen-independent prostate cancer. Clin.Cancer Res. 12, 1665–1671.

Serrano, M., Lin, A.W., McCurrach, M.E., Beach, D., Lowe, S.W.,1997. Oncogenic ras provokes premature cell senescenceassociated with accumulation of p53 and p16INK4a. Cell 88,593–602.

Shintaku, M., Nakagawa, E., Tsubura, A., 2011. ExtramammaryPaget's disease arising in a mature cystic teratoma of the ovary:immunohistochemical expression of androgen receptor. Pathol.Int. 61, 498–500.

35Ligand-independent role AR in the development of OVTC

Simoncini, T., Genazzani, A.R., 2003. Non-genomic actions of sexsteroid hormones. Eur. J. Endocrinol. 148, 281–292.

Singh, P., Uzgare, A., Litvinov, I., Denmeade, S.R., Isaacs, J.T.,2006. Combinatorial androgen receptor targeted therapy forprostate cancer. Endocr. Relat. Cancer 13, 653–666.

Skubitz, A.P., Taras, E.P., Boylan, K.L., Waldron, N.N., Oh, S.,Panoskaltsis-Mortari, A., Vallera, D.A., 2013. Targeting CD133 inan in vivo ovarian cancer model reduces ovarian cancer progres-sion. Gynecol. Oncol. 130 (3), 579–587.

Strauss, R., Hamerlik, P., Lieber, A., Bartek, J., 2012. Regulation ofstem cell plasticity: mechanisms and relevance to tissue biologyand cancer. Mol. Ther. 20, 887–897.

Sviracevic, B., Sedlar, S., Malobabic, D., Cuk, D., 2011. Mixedmalignant germ cell tumor of ovary. Med. Pregl. 64, 93–95.

Szotek, P.P., Pieretti-Vanmarcke, R., Masiakos, P.T., Dinulescu, D.M.,Connolly, D., Foster, R., Dombkowski, D., Preffer, F., Maclaughlin,D.T., Donahoe, P.K., 2006. Ovarian cancer side population definescells with stem cell-like characteristics and Mullerian InhibitingSubstance responsiveness. Proc. Natl. Acad. Sci. U. S. A. 103,11154–11159.

Taplin, M.E., Balk, S.P., 2004. Androgen receptor: a key molecule inthe progression of prostate cancer to hormone independence. J.Cell. Biochem. 91, 483–490.

Walters, K.A., Allan, C.M., Handelsman, D.J., 2008. Androgenactions and the ovary. Biol. Reprod. 78, 380–389.

Wang, L., Hsu, C.L., Chang, C., 2005. Androgen receptor corepressors:an overview. Prostate 63, 117–130.

Wang, R.S., Yeh, S., Tzeng, C.R., Chang, C., 2009. Androgen receptorroles in spermatogenesis and fertility: lessons from testicular cell-specific androgen receptor knockout mice. Endocr. Rev. 30,119–132.

Wasim, T., Majrroh, A., Siddiq, S., 2009. Comparison of clinicalpresentation of benign and malignant ovarian tumours. JPMA J.Pak. Med. Assoc. 59, 18–21.

Williams, A., Datar, R., Cote, R., 2010. Technologies and methodsused for the detection, enrichment and characterization ofcancer stem cells. Natl. Med. J. India 23, 346–350.

Wu, X., Gao, H., Ke, W., Hager, M., Xiao, S., Freeman, M.R., Zhu, Z.,2011. VentX trans-activates p53 and p16ink4a to regulate cellularsenescence. J. Biol. Chem. 286, 12693–12701.

Zhang, S., Balch, C., Chan, M.W., Lai, H.C., Matei, D., Schilder, J.M.,Yan, P.S., Huang, T.H., Nephew, K.P., 2008. Identification andcharacterization of ovarian cancer-initiating cells from primaryhuman tumors. Cancer Res. 68, 4311–4320.

Zheng, H., Ying, H., Yan, H., Kimmelman, A.C., Hiller, D.J., Chen,A.J., Perry, S.R., Tonon, G., Chu, G.C., Ding, Z., et al., 2008.p53 and Pten control neural and glioma stem/progenitor cellrenewal and differentiation. Nature 455, 1129–1133.