Constitutive B-Catenin Activation Induces Male-Specific

Published OnlineFirst August 8, 2013; DOI: 10.1158/0008-5472.CAN-12-4198

Cancer Molecular and Cellular Pathobiology Research

Constitutive b-Catenin Activation Induces Male-Speciﬁc Tumorigenesis in the Bladder Urothelium

Congxing Lin1, Yan Yin1, Kristina Stemler2, Peter Humphrey3, Adam S. Kibel4, Indira U. Mysorekar2,3, and Liang Ma1

Abstract The incidence for bladder urothelial carcinoma, a common malignancy of the urinary tract, is about three times higher in men than in women. Although this gender difference has been primarily attributed to differential exposures, it is likely that underlying biologic causes contribute to the gender inequality. In this study, we report a transgenic mouse bladder tumor model upon induction of constitutively activated b-catenin signaling in the adult urothelium. We showed that the histopathology of the tumors observed in our model closely resembled that of the human low-grade urothelial carcinoma. In addition, we provided þ evidence supporting the KRT5-positive;KRT7-negative (KRT5 ;KRT7 ) basal cells as the putative cells-of- origin for b-catenin–induced luminal tumor. Intriguingly, the tumorigenesis in this model showed a marked difference between opposite sexes; 40% of males developed macroscopically detectable luminal tumors in 12 weeks, whereas only 3% of females developed tumors. We investigated the mechanisms underlying this sexual dimorphism in pathogenesis and showed that nuclear translocation of the androgen receptor (AR) in the urothelial cells is a critical mechanism contributing to tumor development in male mice. Finally, we carried out global gene proﬁling experiments and deﬁned the molecular signature for the b-catenin–induced tumorigenesis in males. Altogether, we have established a model for investigating sexual dimorphism in urothelial carcinoma development, and implicated synergy between b-catenin signaling and androgen/AR signaling in carcinogenesis of the basal urothelial cells. Cancer Res; 73(19); 5914–25. 2013 AACR.

Introduction remained stable even when the smoking prevalence markedly Bladder cancer is the fourth most common cancer type increased in women (3). These observations suggested intrinsic among men, and ninth among women (1). Lifetime cost for mechanisms underpinning gender-based differential suscep- patients with bladder cancer is the highest among all cancer tibility to bladder cancer. types on a per-patient basis (2). Despite its high prevalence and The most common type of bladder cancer is urothelial burden, the etiology of bladder cancer is poorly understood. carcinoma arising from the bladder urothelium. Normally, One important feature of bladder cancer is that its incidence mature urothelium renews itself slowly but continuously by is three times higher in men than in women (1), despite the producing new cells in the basal layer, which differentiate into fact that the development and function of the bladder are intermediate cells and terminally differentiated apical umbrel- not thought to be linked to sex hormones. Historically, this la cells. However, under pathogenic conditions, the basal male predilection, similar to lung cancer, was explained by urothelial cells can rapidly proliferate to restore homeostasis higher rates of tobacco smoking in men. However, this male (4, 5). Therefore, the basal cell layer has been proposed to be the predilection existed in the nonsmoking population, and con- urothelial stem cell niche (4). Interestingly, the tumor-initiat- trary to lung cancers, the gender difference in bladder cancers ing cells isolated from urothelial carcinoma show basal cell characteristics (6, 7). However, the cells-of-origin for urothelial carcinoma remains to be determined. Abnormal activation of the canonical WNT/b-catenin path- Authors' Afﬁliations: 1Division of Dermatology, Department of Medicine, Departments of 2Obstetrics and Gynecology, and 3Pathology and Immu- way is implicated in many cancer types, especially those nology, Washington University in St. Louis, St. Louis, Missouri; and 4Divi- affecting epithelia of endodermal origin, such as prostate (8) sion of Urology, Department of Surgery, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts and colorectal cancers (9). Elevated WNT activity, possibly through promoter hypermethylation-mediated repression of Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/). WNT inhibitors, has been reported in a subset of bladder cancers (10, 11). Moreover, augmented expression of WNT Corresponding Author: Congxing Lin, Division of Dermatology, Depart- b ment of Medicine, Washington University School of Medicine, 660 South pathway genes (6) and nuclear localization of -catenin (12) Euclid Avenue, St. Louis, MO 63110. Phone: 314-454-8283; Fax: 314-454- have been observed in the bladder tumor-initiating cells. In 5626; E-mail: [email protected] mice, WNT/b-catenin signaling is required for basal urothelial doi: 10.1158/0008-5472.CAN-12-4198 proliferation during injury-induced regeneration (13), and 2013 American Association for Cancer Research. forced expression of a constitutively activated form of

5914 Cancer Res; 73(19) October 1, 2013

b-Catenin in Male-Speciﬁc Urothelial Tumorigenesis

b-catenin using UPII-Cre caused urothelial overgrowth, medium was replaced according to the treatment plan and although no tumor developed in the absence of additional MTT assay conducted on one plate every 24 hours as pre- mutations (14, 15). Together, these observations indicate a viously described. Each treatment was carried out in triplicates dynamic role for WNT/b-catenin signaling in regulating for at least three times, and representative results were homeostasis and carcinogenesis of the urothelium. presented. Androgen and androgen receptor (AR) signaling has also been linked to bladder cancers, as AR knockout mice Luciferase assay showed resistance to N-butyl-N-(4-hydroxybutyl)nitrosamine– Bladder cancer cells were seeded on white 96-well plates induced carcinogenesis (16). Cross-talk between WNT/ (10,000 cells/well) and let attach overnight. The next day, cells b-catenin and androgen–AR pathways has been reported were transfected with canonical WNT activity reporter LEF- both during development (17) and in carcinogenesis, partic- LUC and pRL-TK vector (Promega) using X-tremeGENE ularly in prostate cancer (8, 18, 19). Notably, a recent HP (Roche Applied Science) following the manufacturer's study showed a strong association between nuclear AR and instructions. Culture media containing transfection reagent b-catenin expression in human bladder cancers (20), and was replaced 24 hours later with conditioned media and the authors further revealed an interaction between these cells were incubated for additional 48 hours. Luciferase activity two proteins in bladder cancer cell lines. These findings was measured using the Dual-Glo luciferase assay system prompted us to hypothesize that b-catenin is involved in (Promega) and relative LEF-LUC activity was calculated as mediating sexual dimorphism in bladder tumorigenesis. the ratio of luminescence from LEF-LUC (firefly luciferase) to Herein, by using an inducible transgenic system to force pRL-TK (Renilla luciferase). persistent b-catenin activation in adult urothelium, we established a mouse bladder tumor model that exhibits strong Statistical analyses male predilection as a consequence of synergy between Student t tests were conducted to compare proliferative b-catenin and androgen/AR signaling. indices between experimental groups. Additional information is available in the Supplementary Materials and Methods Methods. Animal maintenance All animals were housed in the animal facility at Washing- Results ton University (St. Louis, MO) according to NIH (Bethesda, Forced expression of stabilized b-catenin in mature MD) and Animal Care and Use Committee guidelines. Doxy- urothelium induced luminal bladder tumors in male cycline food pellets 200 mg/kg were purchased from Bio- mice Serv. Castration was carried out following standard surgical Previousstudieshaveinvestigatedtheroleforsustained procedures. b-catenin activation in the urothelium by using a urothelial- specific UPII-Cre (14, 15). However, the urothelial expression Histologic analyses of this Cre is patchy and restricted to the upper layer of Bisected bladder were embedded in paraffin or optimum differentiated cells; the basal layer, which contains progeni- cutting temperature and cut into 5 mm sections. Immunohis- tor cells that have high proliferative reserve and therefore tochemical and immunofluorescence stainings were con- are more likely to be the cells-of-origin for tumors, is largely ducted according to standard protocols. excluded (Fig. 3C, b–d in ref. 21). Importantly, the basal layer is also the endogenous site of WNT activation during RNA isolation, microarray analysis, and real-time injury-induced regeneration. Therefore, we sought to use reverse transcriptase PCR an alternative genetic tool that can target basal urothelial The urothelium was physically separated from the bladder cells with high efficiency. We adopted an Msx2rtTA;tetO-Cre for RNA extraction using the RNeasy Kit (Qiagen). Illumina system, which confers Cre expression in Msx2-expressing mouse-6 chips were used for microarray analysis. cells upon doxycycline treatment (22). We first tested the mT/mG efficacy of the system using an R26 reporter, which Cell culture, treatments, and MTT assay expresses membrane-targeted enhanced GFP (EGFP) in Bladder cancer cell lines, UMUC3 and 5637, and WNT3a- the presence of Cre and red fluorescent tdTomato (mT) producing L-cells were obtained from American Type Culture protein in the absence of Cre (Fig. 1A; ref. 23). Trigenic mT/mG Collection. Cancer cells were maintained in phenol red–free Msx2rtTA;tetO-Cre;R26 mice were treated with doxycy- Dulbecco's Modified Eagle Medium (DMEM) supplemented cline at 4 weeks of age, and their bladders were collected with 10% charcoal-stripped FBS. A 10 mmol/L of testosterone 5 days later. Immunofluorescence analyses showed that all stock solution in ethanol was diluted 1,000-fold in culture cells above the basement membrane, marked by laminin medium to a final concentration of 10 nmol/L. WNT3a con- staining (Fig. 1B), including all cells positive for the basal-cell ditioned media was prepared according to the supplier's markerKRT5(Fig.1C),expressedEGFP.Conversely,all instructions. For MTT assays, bladder cancer cells were mesenchymal cells retained tdTomato (Fig. 1B and C). seeded on five 96-well plates (2,000 cells/well) in identical To test whether these EGFP-positive cells include uro- layouts and let attach overnight. The next day (time 0), normal thelial stem cells, we conducted lineage tracing in which

www.aacrjournals.org Cancer Res; 73(19) October 1, 2013 5915

Lin et al.

Figure 1. Forced b-catenin activation induces luminal tumors in mouse bladder urothelia. A, a schematic of Msx2rtTA;tetO-Cre; mT/mG R26 system. B–D, immunoﬂuorescence staining on reporter line using antibodies indicated showing ubiquitous EGFP expression throughout the urothelium 5 days (B and C) and 5 months (D) after doxycycline (Dox) treatment. E, macroscopic view of single tumor (left) and multifoci tumors (right) in GOF males 12 weeks after doxycycline treatment. F–H, histologic analyses of the tumor showing vascularization, polypoid structure (F and G), and E-cadherin expression (H) in tumors. I, tumor incidence in male and female GOF mutants. Scale bars in B–D represent 60 mm and in F–H represent 500 mm. H&E, hematoxylin and eosin.

the animals were treated with doxycycline for 5 days, and bladders were analyzed after 12 weeks. Thirty-five percent of bladders were collected 20 weeks later for expression anal- the GOF animals died during this period because they suffered ysis. Because the turnover timeformouseurotheliumis from a variety of conditions including alopecia and craniofacial around 40 weeks (24), the percentage of EGFP-positive malformations (Supplementary Fig. S1), reflecting the conse- cells would drop if stem cells were not labeled during the quences of abnormal b-catenin activation in the correspond- initial treatment. In contrast, we observed that all urothelial ing Msx2-expressing organs. Twenty-four percent of the cells remained EGFP-positive (Fig. 1D), indicating that remaining GOF mutants (29 of 122) had macroscopically urothelial stem cells were also targeted by this system. discernable tumors within the bladder lumen; 11 had a To assess the effects of persistent b-catenin activation, we single tumor and 18 had multifoci tumors (Fig. 1E). These Ex3 þ generated trigenic Msx2rtTA;tetO-Cre;b-Cat / animals luminal tumors closely resembled low-grade papillary uro- [hereafter referred to as gain-of-function (GOF) mutants]. The thelial carcinoma (Fig. 1F and G). They often exhibited pol- Ex3/þ b-Cat allele (25), which was also used in previous studies ypoid structure, and the urothelium lost its typical polarity to study b-catenin activation (14, 15), allows expression of and showed nuclear atypia, high mitogenic activity, and stabilized b-catenin upon Cre-mediated exon3 excision. These considerable vascular infiltration. The vast majority of tumor GOF mice, along with their littermate controls, were put on cells stained positive for E-cadherin, indicating their epithelial doxycycline-containing diet for at least 3 weeks to allow origin (Fig. 1H). However, no nuclear pleomorphism or complete Cre-mediated recombination, starting from postna- muscle invasion was observed. Notably, our results are in tal day 25 when the urothelium has fully matured, and their contrast with previous studies that used UPII-Cre to

5916 Cancer Res; 73(19) October 1, 2013 Cancer Research

b-Catenin in Male-Speciﬁc Urothelial Tumorigenesis

ex3 þ overexpress the same b-Cat allele but detected no tumors in The KRT5 ; KRT7 basal urothelial cells are putative the absence of additional mutations (14, 15). We suspect cells-of-origin for b-catenin–induced bladder tumors that this discrepancy reflects key difference in expression We next sought to identify the cells-of-origin for the pattern of the Cre-drivers. tumors.First,tobetterunderstand urothelial differentia- Intriguingly, we noticed a striking sex difference in the tion, immunofluorescence staining against three differen- rate of tumorigenesis, as we found that 45% (27 of 60) male tiation markers, keratin-5 (KRT5), keratin-7 (KRT7), and GOF mice developed tumors, compared with a mere 3% uroplakinIII (UPK3), was conducted to define different (2 of 62) in females (Fig. 1I). Histopathologic analysis also cell types. We observed four distinct cell populations: the þ showed that urothelia in GOF males were more severely KRT5-positive; KRT7-negative (KRT5 ;KRT7 ) basal cells þ þ affected than those in GOF females (Supplementary (Fig.2C,red),KRT5 ;KRT7 basal-intermediate cells (Fig. þ Fig. S2). Notably, this difference was not due to differential 2C,yellow),KRT5;KRT7 intermediate cells (Fig. 2C, þ expression of rtTA or b-catenin, as Western blot analyses green), and KRT5 ;UPK3 umbrella cells (Supplementary revealed equal b-catenin expression in GOF males and Fig. S4C, green). Similar results were also obtained from females (Supplementary Fig. S3). The fatality noted earlier staining using alternative basal marker CD44 and interme- was not due to sex-related causes, however, as females diate cell marker Loricrin (Supplementary Fig. S4D–S4I). and males died at similar rates (data not shown). This graded marker expression pattern, similar to that

þ Figure 2. KRT5 ; KRT7 basal cells are the cells-of-origin for b-catenin–induced urothelial tumors. A–C, double immunofluorescence using antibodies indicated. Note the red KRT5þ; KRT7 basal cells, yellow KRT5þ; KRT7þ basal-intermediate cells, green KRT5; KRT7þ intermediate cells in C. D–D00, Histologic and double immunofluorescence analyses on adjacent sections of GOF urothelia 5 weeks after starting treatment showing dysplastic nodules (arrows in D–D00) with low-KRT5 expression (D0)andabsenceofKRT7(D00) expression. E and F, histologic and KRT5/7 double immunofluorescence analyses 8 (E–E00) and 12 (F–F00) weeks after treatment showing tumor progression along with expansion of the KRT5-low; KRT7 dysplastic cells. G–I, double immunofluorescence analyses using antibodies indicated. G, costaining of KRT7 and Ki67 revealing that the KRT7-negative nodular cells were highly proliferative. H and I, double immunofluorescence staining showing Ki67þ cells were localized close to basement membrane (H) and expressing KRT14 (I). J, a model for WNT-induced tumorigenesis of basal urothelial cells. Scale bars in A–C, 60 mm; in D–F, 250 mm; and in G–I, 125 mm. Dox, doxycycline.

www.aacrjournals.org Cancer Res; 73(19) October 1, 2013 5917

Lin et al.

observed in the prostate epithelium (26), reflects a conti- Aberrant AR expression in the b-catenin–GOF nuous differentiation process and a hierarchical relation- urothelium þ ship among urothelial cells, and suggests that the KRT5 ; Next, we sought to determine the mechanisms underlying KRT7 basal cells are likely early urothelial progenitor/ the male-predilection in this bladder tumor model. We stem cells. Next, we determined which cell population observed no difference in bladders between GOF males and is responsible for over proliferation driven by b-catenin females 3 weeks after doxycycline treatment (data not shown), mutation. Five days after doxycycline treatment, elevated but bladders of GOF males 5 weeks after treatment showed a b-catenin expression and nuclear localization of the protein higher proliferation index and more KRT7-negative cells than was observed throughout the GOF urothelium (Supplemen- those of GOF females at the same stage (Fig. 3A). This corre- tary Fig. S5B). The GOF urothelia were highly proliferative sponds to the time when circulating androgen level peaks in þ as 33% of urothelial cells were Ki67-positive (Ki67 ), com- male mice (29). Considering the potential interaction between pared with near-zero in controls (Supplementary Fig. S5C, b-catenin and androgen/AR signaling, we speculated that S5D, and S5G). In addition, we also detected considerable androgen signaling might function to enhance the male phe- mitotic activity in the GOF urothelium, evidenced by the notype. We first analyzed AR expression by immunohis- presence of phospho-histone H3-positive cells (Supplemen- tochemistry and found that in control males, AR was prefer- þ tary Fig. S5E and S5F). Notably, these Ki67 cells were entially expressed in the cytoplasm of the intermediate cells þ exclusively KRT5 basal cells, and this proliferation is (Fig. 3B). In contrast, prominent nuclear AR expression was þ accompanied by an expansion of the KRT5 ;KRT7 basal evident in the dysplastic/tumor cells of GOF males (Fig. 3C). cell layer (Supplementary Fig. S6). These results indicate Further analysis showed a mutually exclusive expression pat- that activation of b-catenin signaling promotes basal cell tern of KRT7 and nuclear AR (Fig. 3E). Indeed, nuclear AR proliferation. expression clearly defined the boundary between tumor/dys- To understand the dynamics of b-catenin–induced plastic epithelium and the neighboring epithelium (Supple- tumorigenesis, and define contribution from different mentary Fig. S8). In females, AR expression was much weaker urothelial cell types, we examined GOF mouse bladders in both controls and the GOF mutants as compared with males every 2 to 3 weeks over a period of 12 weeks. Five weeks (data not shown and Fig. 3D). Notably, this nuclear AR trans- after starting doxycycline, GOF urothelia exhibited re- location was not a result of elevated testosterone in the GOF markable histopathologic changes. They were thicker than males (Fig. 3F). controls (Fig. 2D and Supplementary Fig. S7B) and con- Next, we tested whether b-catenin overexpression is suf- tained histologically distinct cell nodules within the ficient to induce AR nuclear translocation in bladder cancer urothelium (arrows in Fig. 2D–D00 and Supplementary Fig. cells in a previously described AR-positive bladder cancer S7B). These nodular cells were small, and histologically cell line UMUC3 (16, 20). We grew these cells in charcoal- resembled the basal cells (Supplementary Fig. S7C). The stripped DMEM medium, and assayed for AR expression nodular cells displayed a unique signature of low KRT5 upon different treatment by immunofluorescence staining. (Fig. 2D00) and absence of KRT7 expression (Fig. 2D00). Without treatment, AR was mainly expressed in the cyto- Costaining of KRT7 and Ki67 revealed high-proliferative plasm. This cytoplasmic expression was not changed after 24 activities in these KRT5-low; KRT7–negative cells (Fig. 2G). hours of WNT3a treatment (Fig. 3G), which can markedly Moreover, these cells showed distinct signaling activity, increase b-catenin protein level (Fig. 4E). On the other hand, evidenced by p-S6 expression, an established target of treatment of 10 nmol/L testosterone induced nuclear AR the mTOR pathway (Supplementary Fig. S7D). These his- expression in around 15% of AR-positive cells. Intriguingly, tological/molecular signatures of the GOF urothelia largely cotreatment of 10 nmol/L testosterone and WNT3a marked- persisted throughout tumorigenesis (Fig. 2D–F, D0–F0,and ly increased the number of cells with nuclear AR expression D00–F00). Notably, the majority of the cells from fully devel- by more than 2-fold to above 40%. We further tested whether oped tumors are also KRT5-low;KRT7-negative, similar to this b-catenin–induced AR nuclear translocation rely on de the dysplastic nodules observed early on (Fig. 2F–F00). novo protein synthesis, by treating the cells with 25 mg/mL þ These results suggested that the KRT5 ;KRT7 basal cells cycloheximide along with testosterone and WNT3a. We were the cells-of-origin for b-catenin–induced luminal found that cycloheximide did not reduce the percentage of tumors (model in Fig. 2J). Notably, we found that although cells with nuclear AR expression (Fig. 3G and Supplementary most tumor cells retained their original keratin signature, Fig. S9). Altogether, these data indicated that b-catenin can only cells adjacent to the basement membrane were highly mediate AR nuclear translocation independent of its tran- proliferative (Fig. 2H), and most of these proliferating scriptional activity. Therefore, we speculate that the dys- cells were stained positive for KRT14 (Fig. 2I), an early plastic cells in the GOF urothelium with nuclear AR expres- progenitor cell marker in bladder cancers (27, 28). Our sion might express higher level of b-catenin protein com- finding that the basal cells are responsible for tumor pared with their neighboring cells. To test the hypothesis, we formation suggests that the failure of earlier studies to stained bladder sections from 2- and 4-month-old male detect tumors upon expression of activated b-catenin using b-catenin–GOF mice along with a wild-type (WT) control UPII-Cre may be explained by the differential oncogenic with an anti-b-catenin antibody. Intriguingly, the basal cells potential of progenitor and differentiated urothelial cell in the 2-month-old GOF mice clearly showed much higher types. expression (Fig. 3I, arrows) than their neighboring cells,

5918 Cancer Res; 73(19) October 1, 2013 Cancer Research

b-Catenin in Male-Speciﬁc Urothelial Tumorigenesis

Figure 3. Misregulation of AR in the urothelial cells upon elevated b-catenin expression. A, Ki67- index and KRT7-index showing that GOF male urothelia are more severely affected; , P < 0.025. B–D, IHC against AR showing cytoplasmic expression in the intermediate cells of control males (B), nuclear expression in the dysplastic urothelial cells of GOF males (C), and weak cytoplasmic expression in GOF females (D). E, double immunofluorescence showing complementary expression of AR and KRT7. F, serum testosterone level of control and GOF-mutant animals 8 weeks after doxycycline treatment measured by ELISA assay. Note that there is no significant difference between control males and GOF males. N.S., not statistically significant. G, immunofluorescence AR staining in UMUC3 cells upon different treatment as indicated. All cells were collected 24 hours after treatment. H–J, b-catenin IHC staining on urothelia of control (H), 2-month-old GOF (I), and 4-month- old GOF (J) male animals. Note the elevated b-catenin expression in the dysplastic basal cells (arrows in I) and tumor cells (arrows in J) in the GOF animals. All scale bars represent 60 mm.

which also showed stronger staining than the control (Fig. gen–AR signaling activity. Next, we examined the bladders 3H). The uneven b-catenin protein expression was also from these castrated GOF animals after 3 months of doxycy- observed in the 4-month-old GOF male mouse with multifoci cline treatment, and found that only 12.5% (2 of 16) of castrated bladder tumors. The tumor cells and some dysplastic basal males developed luminal tumors (Fig. 4C). In addition, cell cells (Fig. 3J), the same cell population with nuclear AR proliferation of castrated GOF males was 20% less than non- expression (Supplementary Fig. S8), clearly showed a stron- castrated counterparts 3 weeks after castration (data not ger b-catenin staining than the rest of the urothelium. These shown). These results indicated that misregulation of andro- data suggest that the nuclear AR translocation may be gen–AR signaling contributed to b-catenin–induced male- mediated by high level of b-catenin protein expression. specific bladder tumorigenesis. To test whether this finding is relevant in human bladder Androgen–AR signaling enhances b-catenin–induced carcinogenesis, we analyzed the roles for AR and b-catenin bladder tumorigenesis signaling in two human bladder cancer cell lines, the AR- To determine whether abnormally expressed AR plays a positive UMUC3 and the AR-negative 5637 cells. Both cell role in promoting b-catenin–induced tumorigenesis in male lines were treated with either WNT3a conditioned medium, mice, we castrated GOF male mice 15 days after starting 10 nmol/L testosterone, or the combination of both for doxycycline treatment and determined its effect on tumor 96 hours, and cell growth and viability were assessed by development (Fig. 4A). We first analyzed AR expression by MTT assay in 24-hour intervals. In neither cell line did 10 real-time reverse transcriptase PCR (RT-PCR) analysis, and nmol/L testosterone treatment show any overt effect on found that its urothelial expression in castrated male GOF cell growth. On the other hand, WNT3a conditioned medi- mice was more than 3-fold lower than noncastrated counter- um, which led to increased b-catenin protein expression parts (Fig. 4B), indicating a markedly reduced level of andro- (Fig.4DandE),significantly enhanced growth of both cell

www.aacrjournals.org Cancer Res; 73(19) October 1, 2013 5919

Lin et al.

Figure 4. AR enhances b-catenin– induced bladder tumorigenesis. A, a schematic diagram of the time line of doxycycline (Dox) regimen and castration schedule. B, quantitative real-time RT-PCR showing reduced AR expression in the castrated GOF males compared with noncastrated GOF males. C, tumor incidence in castrated and noncastrated GOF males. D and E, b-catenin immunoﬂuorescence showing markedly increased protein level in UMUC3 cells after incubation in WNT3a conditioned media. F, representative MTT assays showed increased cell growth in both cell lines by WNT3a treatment at 96 hours (1,087.7% 27.7% WNT3a vs. 943.1 35.1 control, P < 0.01 in UMUC3; 1,182.1% 19.7% WNT3a vs. 1,026.2% 72.4%, P < 0.03 in 5637 cells). Testosterone alone did not affect cell growth in either cell line (P ¼ 0.71 and 0.77, respectively). Combined treatment of WNT3a and testosterone only augmented the WNT3a effect in UMUC3 cells (1,215.6% 20.0% vs. 1,087.7% 27.7%, P < 0.02 in UMUC3; 1,182.1% 19.7% vs. 1,174.4% 29.8%, P ¼ 0.86 in 5637 cells). TST, testosterone.

lines at 96 hours (Fig. 4F, green triangles). Combined Androgen–AR signaling enhances the transcriptional treatment of testosterone and WNT3a, however, augmented activity of b-catenin–mediated canonical WNT signaling the growth-promoting effect of WNT3a conditioned media pathway only in AR-positive UMUC3 cells at 96 hours (1,215.6% To gain insights into the molecular mechanisms through 20.0% vs. 1,087.7% 27.7%, P < 0.02 in UMUC3; 1,182.1% which AR synergizes with b-catenin in promoting tumori- 19.7% vs. 1,174.4% 29.8%, P ¼ 0.86 in 5637 cells). Collec- genesis, we conducted expression proﬁling of urothelial cells tively, these data showed that androgen–AR signaling from control and GOF animals. Urothelial samples were synergize with b-catenin to promote growth of the bladder collected from animals 3 weeks after starting doxycycline cancer cells. treatment, a stage when androgen starts to elevate but

5920 Cancer Res; 73(19) October 1, 2013 Cancer Research

b-Catenin in Male-Speciﬁc Urothelial Tumorigenesis

Figure 5. Sexual dimorphic gene regulation by hyperactive WNT/ b-catenin signaling in the bladder urothelium. A, a schematic for microarray analyses. B, hierarchical clustering heat map of the 130 male tumor genes. Note that most genes are not distinguishable between WT males and females, but are changed prominently in GOF males. C, composition of male tumor genes. D, star glyph distribution of the fold-change of the 85 male tumor WNT targets revealed sexual dimorphic response to WNT activation. Log2 values of fold change for each gene were shown. Note that the peaks were eminent in male for all genes except for two (asterisks). This change reﬂects a higher WNT- responsiveness in males. E, classiﬁcation of the 85 genes based on Gene Ontology biologic processes. Note that some genes have multiple functions and may be represented in more than one category.

sexual dimorphism in bladder phenotype has yet to develop. group, among which were 68 b-catenin targets, an additional We analyzed four groups of samples representing different 17 possible (changed in the GOF males relative to the control sex and genotypes, each consisting of three pooled samples males by 1.5- to 2-fold) b-catenin targets, and 11 genes (Fig. 5A and B), and identified three clusters of differentially differentially expressed in opposite sexes in controls (Fig. regulatedgenes(DRG;Fig.5A):(i)genesdifferentiallyreg- 5B and C and Supplementary Table S2). Intriguingly, 83 out ulated between genotypes, that is, b-catenin targets; (ii) of 85 b-catenin targets showed the same trend of, but greater genes differentially regulated between opposite sexes, that changes in the GOF males (Fig. 5D), suggesting that b-cate- is, sex-specific genes; and (iii) genes differentially expressed nin responsiveness is higher in males than in females. These between GOF males and GOF females, that is, genes causing genes were further categorized by their biologic functions male-predilection in b-catenin–induced bladder tumorigen- (Fig. 5E). Notably, among these genes there were several esis.Wesoughttodeterminetherelationshipamongthese well-characterized direct downstream targets of canonical three groups of DRGs, to understand the potential mechan- TCF/LEF–dependent WNT signaling pathway including isms through which AR impacts the downstream signaling of Axin2, Stra6, Sp5,andGpr49, suggesting that AR might b-catenin. Comparing the data acquired from GOF to con- synergize with b-catenin by enhancing b-catenin–mediated trol, we uncovered 576 genes (b-catenin targets; Supplemen- transcriptional activity. To test this hypothesis, we treated tary Fig. S1) that differed in expression level by more than 2- the bladder cancer cells with 10 nmol/L of testosterone to fold, including 24 previously identified direct canonical T- induce nuclear AR activity, and then determined its impact cell factor/lymphoid enhancer factor (TCF/LEF)–dependent on the transcriptional activity of TCF/LEF–dependent WNT signaling targets (highlighted in Supplementary Table canonical WNT/b-catenin signaling by assaying for lucifer- S1). This result validated our model and indicated that ase activity of the well-characterized LEF-LUC reporter abnormal activation of canonical WNT signaling is one of (TOPFLASH; ref. 30). LEF-LUC contains a luciferase gene the events caused by b-catenin–GOF mutation. The rest driven by a minimal promoter and enhancer with seven majority of the genes were likely either targets of non- TCF/LEF–binding sites, and therefore is used to analyze WNT–related b-catenin signaling, or genes further down- activity of the TCF/b-catenin–mediated transcription. As a stream of the b-catenin pathway. Next, we compared data- positive readout, WNT3a treatment enhanced luciferase sets from GOF males to GOF females to identify genes activity in both cancer cell lines (fold-change is 4.25 underpinning b-catenin–induced sexual dimorphic tumori- 1.81 and 3.36 0.72 in UMUC3 and 5637 cells, genesis. We found 130 DRGs (male tumor genes) in this respectively;Fig.6A).Ontheotherhand,10nmol/L

www.aacrjournals.org Cancer Res; 73(19) October 1, 2013 5921

Lin et al.

Figure 6. AR enhances b-catenin– mediated transcriptional activity. A, LEF-LUC luciferase activity was induced 4.25-fold by WNT3a conditioned medium in UMUC3 cells and was increased 7.40-fold by cotreatment with TST (WNT3a 4.25 1.81, P < 0.03 compared with control; WNT3aþTST 7.40 0.34, P < 0.05 compared with WNT3a alone). This synergistic effect was not observed in AR- negative 5637 cells (WNT3a 3.36 0.72; WNT3aþTST 2.92 0.58, P ¼ 0.46). B, expression of canonical WNT downstream target, Axin2, upon same treatments measured by real-time RT-PCR. C, Western blot analysis against b-catenin revealed no increase in protein level in nuclear extracts of UMUC3 cells treated with control or TST media. D and E, immunoﬂuorescence showing similar levels of b-catenin in UMUC3 cells with or without TST treatment. F and G, immunoﬂuorescence showing increased LEF1 protein in the TST- treated UMUC3 cell nuclei (G compared with F). H and I, IHC showing nuclear LEF1 expression in the b-catenin GOF urothelial cells (I), but not control urothelial cells (H). J, real-time RT-PCR analyses on putative AR downstream targets using urothelial RNA isolated from animals with sex and genotype indicated.

of testosterone alone did not have any impact on either cell 7.40 0.34; P < 0.05), but had no additional effect on line (Fig. 6A). Intriguingly, cotreatment of WNT3A and WNT3A-treated AR-negative 5637 cells. Similar results were testosterone augmented growth of AR-positive UMUC3 also obtained when we analyzed the expression of direct cells for an additional 74.1% than WNT3A alone (fold change TCF/b-catenin downstream target Axin2 in these cells

5922 Cancer Res; 73(19) October 1, 2013 Cancer Research

b-Catenin in Male-Speciﬁc Urothelial Tumorigenesis

(Fig. 6B). These data are largely consistent with a recent Notably, the bladder phenotype of this model is clearly independent publication by Li and colleagues (20), in which different from previous report showing that forced expression ex3 the authors also indicate this synergy is at least in part of b-Cat allele alone using UPII-Cre does not cause malig- mediated by elevated nuclear b-catenin expression upon AR nancy (14). This phenotypic variation is likely caused by key activation. We next tested whether this is also the case in differences in spatiotemporal expression pattern of the Cre our study by assaying nuclear expression of b-catenin by drivers. The endogenous UPKII is preferentially expressed in both Western blot analysis (Fig. 6C) and immunofluores- the upper urothelial layers (21). Therefore, we suspect that cence staining (Fig. 6E). However, we found no evidence of the UPII-Cre, although driven by a partially inverted promoter enhanced b-catenin protein expression in either assay (Fig. (31), targets a more differentiated cell population with 6C–E).Therefore,wespeculatethattheactivationoftran- limited proliferative reserve, which is more resilient to onco- scription by AR may be mediated through regulation of genic mutation than the stem/progenitor cells targeted by b-catenin cofactor TCF/LEF family proteins. We examined Msx2-rtTA. Indeed, from the low-magnification images pre- the expression of LEF1 in testosterone-treated UMUC3 cells sented in the original article, it seemed that UPII-Cre expres- by immunofluorescence (Fig. 6F and G). Indeed, the nuclear sion is patchy and the basal cells were largely spared (Fig. 3C, LEF1 expression was highly elevated after testosterone b–d in ref. 21). On the other hand, our findings have strongly treatment (Fig. 6G). We next analyzed LEF1 expression suggested that the cells-of-origin for the urothelial tumors þ in control and b-catenin–GOF male bladders, and found in the Msx2-rtTA model are indeed the KRT5 ; KRT7 basal that the nuclear LEF1 expression can also be observed in the cells. Moreover, the UPII-Cre is expressed during embryonic/ dysplastic cells in the GOF urothelium (Fig. 6I) but not the neonatal development when urothelium acquires differentia- control urothelium (Fig. 6H). These data suggested that AR tion. In contrast, Cre activity is not induced in the Msx2-rtTA may enhance b-catenin–mediated transcription by activat- model until doxycycline is given after urothelia have matured. ing its cofactor LEF1. Whether b-catenin responsiveness of the developing urothe- In addition to facilitating transcription mediated by lium channels that of the adult urothelium remains un- b-catenin, AR might also contribute to tumor growth by known. Notably, an AhCreERT (14) was also used to conditional regulating its own downstream target genes. Therefore, we activate b-catenin in the urothelium. However, the intestinal compared the 130 genes differentially regulated between hyperplasia caused by intestinal Cre expression in this model GOF males and females, with existing microarray databases precluded the use of AhCreERT for long-term tumor studies and found eight known downstream targets of AR in the (14). Collectively, the phenotypic differences among these male tumor genes (Supplementary Table S3). Some of these models likely reflect the dynamics of the urothelium in genes play important roles in mediating normal biologic response to oncogenic b-catenin mutation. Unfortunately, processes and carcinogenesis (Supplementary Table S3), the detailed stage and cell type–specific characterization of and their expression was further verified by real-time PCR the UPII-Cre expression has not been published. Further anal- (Fig. 6J). It is possible that aberrant expression of these yses determining spatiotemporal expression pattern of the genes might have contributed to enhanced bladder tumor- UPII-Cre within the urothelium will provide valuable insights igenesis in males. Together, these data suggested two into cellular mechanisms underlying the differential tumori- potential mechanisms mediated by AR and contribute to genesis in these models. b-catenin–induced male-specific bladder tumorigenesis: (i) AR augments b-catenin responsiveness in males, and (ii) AR Basal cells as cells-of-origin for urothelial carcinoma activates a distinct set of downstream target genes in the Bladder urothelium contains heterogeneous cell popula- presence of ectopic b-catenin. Notably, we did not observe tions with distinct functions. Previous studies have shown coactivation of sonic hedgehog (SHH) and fibroblast growth that the urothelial stem cells or progenitor cells reside in the þ þ factor (FGF) signaling, indicating that b-catenin–induced KRT5 ; SHH basal layers, and these cells are capable of bladder tumor development is independent of these onco- differentiating into umbrella cells and reconstitute urothelium genic pathways (Supplementary Fig. S10). under both homeostatic and wound conditions (4, 13). In support of this notion, our marker analyses have revealed a graded shifting of cell surface markers from basal cells to the Discussion fully differentiated luminal umbrella cells (Fig. 2A–C and A novel mouse model for bladder cancer research Supplementary Fig. S4), which is consistent with a gradual þ In this study, we established a bladder tumor model in basal to apical differentiation process. However, the KRT5 which a constitutively activated form of b-catenin is induced cells constitute more than half of all urothelial cells and likely to express in all urothelial cells in adult mice. This mouse contain not only progenitor cells including urothelial stem model has the following distinctive features: (i) the genetics is cells/transit-amplifying cells, but also intermediate cell types. þ defined, and the histopathology of the luminal tumors faith- Our analysis revealed a KRT5 ; KRT7 subpopulation that fully recapitulates that of the low-grade papillary urothelial represents a small fraction of KRT5 cells, which localizes right carcinoma, (ii) the tumor development is relatively fast, and adjacent to the basement membrane. Whether they are the (iii) the incidence is much higher in males, which makes it an true urothelial stem cells and/or early transit-amplifying excellent tool to investigate sexual dimorphism in bladder cells remains to be tested by lineage tracing experiments. pathogenesis. Nonetheless, this cell population shows the most primitive

www.aacrjournals.org Cancer Res; 73(19) October 1, 2013 5923

Lin et al.

differentiation state evidenced by the lack of KRT7, Loricrin, b-catenin cofactor LEF1 protein in the nuclei (Fig. 6F–I). and UPK3 expression (Fig. 2A–C and Supplementary Fig. S4). Whether this LEF1 activation plays a role in the augmented In addition, through histologic analyses at sequential time canonical WNT signaling activity, and how AR regulates LEF1 points, we found that the exophytic bladder tumors developed and/or other TCF factors, are currently being investigated. It is in the GOF mutant were indeed evolved from small but also noteworthy that b-catenin can activate not only TCF- þ highly proliferative clusters of dysplastic KRT5 ; KRT7 cells dependent WNT signaling, but also other downstream signal- (Fig. 2). Despite all these studies, the definitive evidence of ing cascades. In our microarray analyses, we also found many these cells as cells-of-origins for bladder tumors still has to non-WNT–related genes that showed greater expression come from lineage tracing experiments, which require genetic changes in GOF males than females (Supplementary Table tools (Cre lines) that can specifically target these cells. Estab- S1). It is plausible that the binding of AR to b-catenin can lishing these tools should be an important aim of future recruit other cofactors, such as histone methyltransferases or studies. histone acetyltransferases, to modulate downstream gene expression. Molecular bases for male-predilection in bladder Another interesting finding is that b-catenin accumula- tumorigenesis tion can mediate AR nuclear translocation, but only in basal Male-predilectionisanimportantissueinbladdercancer cells–derived dysplastic/tumor urothelial cells (Fig. 3D–E). research because more than three quarters of patients with This differential AR expression could be a result of higher bladder cancer are men. Interestingly, this gender difference amount of b-catenin observed in these dysplastic cells in disease susceptibility has also been observed in BBN- (Fig.3IandJ).Importantly,theb-cateninex3 allele is tran- induced mouse model (16) and our model, suggesting it scribed not only in basal cells and their derivatives but in cannot be simply explained byhigherratesoftobacco all urothelial cells. Therefore, the elevated protein expres- smoking in men. The involvement of androgen–AR signaling sion may reflect higher transcription from the b-catenin in bladder carcinogenesis has started to be revealed by locus and/or increased protein stability, in these parti- several recent studies (16, 20, 32). However, the exact mech- cular cells. Alternatively, the nuclear translocation of AR anism(s) through which AR promotes bladder tumor devel- may require not only b-catenin, but also other cofactor(s) opment remain to be elucidated. Through both in vivo and in that only present in this particular cell type, which are vitro analyses, our study shows a profound synergy between yettobeidentified. These speculations require further AR and b-catenin in bladder tumorigenesis, which is sup- investigations. ported by the following three important discoveries: (i) Altogether, our study indicates that the synergistic acti- overexpression of b-catenin protein in the urothelial cells vation of b-catenin and AR may be the underlying mechan- induces nuclear translocation of the AR, (ii) coordinated AR isms for male-predilection in bladder carcinogenesis. The and b-catenin activation enhances bladder tumor growth, clinical relevance of this findingissupportedbyarecent and (iii) AR facilitates b-catenin–induced bladder tumori- article showing a strong association between nuclear genesis by enhancing its transcriptional activity. These expression of these two proteins in human bladder can- findings clearly indicate a multifaceted interaction between cers (20). However, abnormal b-catenin activation is only AR and b-catenin during bladder carcinogenesis. observed in a small fraction of the urothelial cancers The function of AR has been linked to b-catenin by numer- (11, 12), which is obviously not sufficient to explain the ous studies on prostate cancer cells (18, 33–39). b-Catenin can 3:1 male predilection. Whether and how androgen–AR directly bind to liganded AR (18, 33) and enhances AR-medi- signaling also contributes to the rest of the male bladder ated transcriptional activity (18, 33) in addition to regulate cancers without b-catenin activation remains to be deter- AR expression (39). On the other hand, liganded AR can mined by future studies. facilitate nuclear translocation of b-catenin (34) and its binding to AR-responsive promoter elements (34). It has been Disclosure of Potential Conflicts of Interest shown that AR can compete with TCF factors (35, 36), in No potential conflicts of interest were disclosed. particular TCF4 for b-catenin. Consequently, activation of AR can suppress the transcriptional activity of the canonical Authors' Contributions – b Conception and design: C. Lin, A.S. Kibel, L. Ma TCF/LEF dependent WNT/ -catenin signaling. Our study, Development of methodology: C. Lin, Y. Yin in contrast, has shown that nuclear AR can enhance the Acquisition of data (provided animals, acquired and managed patients, activity of b-catenin–mediated canonical WNT signaling in provided facilities, etc.): C. Lin, Y. Yin in vivo Analysis and interpretation of data (e.g., statistical analysis, biostatistics, urothelial cells, both (Fig. 5 and Supplementary computational analysis): C. Lin, Y. Yin, P. Humphrey, I.U. Mysorekar, L. Ma Table S2) and in vitro (Fig. 6A and B). These data indicate that Writing, review, and/or revision of the manuscript: C. Lin, Y. Yin, K. Stemler, the signaling output of this reciprocal interaction between P. Humphrey, A.S. Kibel, I.U. Mysorekar, L. Ma Administrative, technical, or material support (i.e., reporting or orga- AR and b-catenin is highly cell context–dependent. Notably, nizing data, constructing databases): Y. Yin, K. Stemler same result has been obtained by a recent independent Study supervision: C. Lin, A.S. Kibel, I.U. Mysorekar, L. Ma report by Li and colleagues (20). However, unlike what was fi Acknowledgments described in that study, we did not nd an associated ex3 b The authors thank Dr. Fanxin Long for providing b-Cat animal, TOPFLASH increase in nuclear -catenin expression upon AR activation plasmid, and L-Wnt3a cells. The authors also thank Dr. Douglas Covey for help on (Fig. 6C–E). Instead, we observed an increased expression of experiments involving testosterone.

5924 Cancer Res; 73(19) October 1, 2013 Cancer Research

b-Catenin in Male-Speciﬁc Urothelial Tumorigenesis

Grant Support advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate This work is supported by American Urological Association Research Scholar this fact. Program (C. Lin) and NIH R01 ES01659701 (L. Ma). The costs of publication of this article were defrayed in part by the Received November 12, 2012; revised June 13, 2013; accepted July 2, 2013; payment of page charges. This article must therefore be hereby marked published OnlineFirst August 8, 2013.

References 1. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2012. CA Cancer 21. Mo L, Cheng J, Lee EY, Sun TT, Wu XR. Gene deletion in urothelium by J Clin 2012;62:10–29. specific expression of Cre recombinase. Am J Physiol Renal Physiol 2. Botteman MF, Pashos CL, Redaelli A, Laskin B, Hauser R. The health 2005;289:F562–8. economics of bladder cancer: a comprehensive review of the pub- 22. Lin C, Yin Y, Chen H, Fisher AV, Chen F, Rauchman M, et al. Con- lished literature. PharmacoEconomics 2003;21:1315–30. struction and characterization of a doxycycline-inducible transgenic 3. Freedman ND, Silverman DT, Hollenbeck AR, Schatzkin A, Abnet CC. system in Msx2 expressing cells. Genesis 2009;47:352–9. Association between smoking and risk of bladder cancer among men 23. Muzumdar MD, Tasic B, Miyamichi K, Li L, Luo L. A global double- and women. JAMA 2011;306:737–45. fluorescent Cre reporter mouse. Genesis 2007;45:593–605. 4. Mysorekar IU, Isaacson-Schmid M, Walker JN, Mills JC, Hultgren SJ. 24. Jost SP. Cell cycle of normal bladder urothelium in developing and Bone morphogenetic protein 4 signaling regulates epithelial renewal in adult mice. Virchows Arch B Cell Pathol Incl Mol Pathol 1989;57:27–36. the urinary tract in response to uropathogenic infection. Cell Host 25. Harada N, Tamai Y, Ishikawa T, Sauer B, Takaku K, Oshima M, et al. Microbe 2009;5:463–75. Intestinal polyposis in mice with a dominant stable mutation of the 5. Mulvey MA, Lopez-Boado YS, Wilson CL, Roth R, Parks WC, Heuser J, beta-catenin gene. EMBO J 1999;18:5931–42. et al. Induction and evasion of host defenses by type 1-piliated 26. van Leenders G, Dijkman H, Hulsbergen-van de Kaa C, Ruiter D, uropathogenic Escherichia coli. Science 1998;282:1494–7. Schalken J. Demonstration of intermediate cells during human pros- 6. He X, Marchionni L, Hansel DE, Yu W, Sood A, Yang J, et al. Differ- tate epithelial differentiation in situ and in vitro using triple-staining entiation of a highly tumorigenic basal cell compartment in urothelial confocal scanning microscopy. Lab Invest 2000;80:1251–8. carcinoma. Stem Cells 2009;27:1487–95. 27. Volkmer JP, Sahoo D, Chin RK, Ho PL, Tang C, Kurtova AV, et al. Three 7. Chan KS, Volkmer JP, Weissman I. Cancer stem cells in bladder can- differentiation states risk-stratify bladder cancer into distinct subtypes. cer: a revisited and evolving concept. Curr Opin Urol 2010;20:393–7. Proc Natl Acad Sci U S A 2012;109:2078–83. 8. Kypta RM, Waxman J. Wnt/beta-catenin signalling in prostate cancer. 28. Ho PL, Lay EJ, Jian W, Parra D, Chan KS. Stat3 activation in urothelial Nat Rev Urol. 2012 Jun 19. [Epub ahead of print]. stem cells leads to direct progression to invasive bladder cancer. 9. Bienz M, Clevers H. Linking colorectal cancer to Wnt signaling. Cell Cancer Res 2012;72:3135–42. 2000;103:311–20. 29. Selmanoff MK, Goldman BD, Ginsburg BE. Developmental changes in 10. Urakami S, Shiina H, Enokida H, Kawakami T, Kawamoto K, Hirata H, serum luteinizing hormone, follicle stimulating hormone and androgen et al. Combination analysis of hypermethylated Wnt-antagonist family levels in males of two inbred mouse strains. Endocrinology 1977;100: genes as a novel epigenetic biomarker panel for bladder cancer 122–7. detection. Clin Cancer Res 2006;12:2109–16. 30. Korinek V, Barker N, Morin PJ, van Wichen D, de Weger R, Kinzler KW, 11. Urakami S, Shiina H, Enokida H, Kawakami T, Tokizane T, Ogishima T, et al. Constitutive transcriptional activation by a beta-catenin-Tcf et al. Epigenetic inactivation of Wnt inhibitory factor-1 plays an impor- complex in APC/ colon carcinoma. Science 1997;275:1784–7. tant role in bladder cancer through aberrant canonical Wnt/beta- 31. Ayala de la Pena F, Kanasaki K, Kanasaki M, Tangirala N, Maeda G, catenin signaling pathway. Clin Cancer Res 2006;12:383–91. Kalluri R. Loss of p53 and acquisition of angiogenic microRNA profile 12. Chan KS, Espinosa I, Chao M, Wong D, Ailles L, Diehn M, et al. are insufficient to facilitate progression of bladder urothelial carcinoma Identification, molecular characterization, clinical prognosis, and ther- in situ to invasive carcinoma. J Biol Chem 2011;286:20778–87. apeutic targeting of human bladder tumor-initiating cells. Proc Natl 32. Johnson AM, O'Connell MJ, Miyamoto H, Huang J, Yao JL, Messing Acad Sci U S A 2009;106:14016–21. EM, et al. Androgenic dependence of exophytic tumor growth in a 13. Shin K, Lee J, Guo N, Kim J, Lim A, Qu L, et al. Hedgehog/Wnt feedback transgenic mouse model of bladder cancer: a role for thrombospondin- supports regenerative proliferation of epithelial stem cells in bladder. 1. BMC Urol 2008;8:7. Nature 2011;472:110–4. 33. Truica CI, Byers S, Gelmann EP. Beta-catenin affects androgen recep- 14. Ahmad I, Morton JP, Singh LB, Radulescu SM, Ridgway RA, Patel S, tor transcriptional activity and ligand specificity. Cancer Res 2000; et al. b-Catenin activation synergizes with PTEN loss to cause bladder 60:4709–13. cancer formation. Oncogene 2011;30:178–89. 34. Mulholland DJ, Cheng H, Reid K, Rennie PS, Nelson CC. The androgen 15. Ahmad I, Patel R, Liu Y, Singh LB, Taketo MM, Wu XR, et al. Ras receptor can promote beta-catenin nuclear translocation independently mutation cooperates with beta-catenin activation to drive bladder of adenomatous polyposis coli. J Biol Chem 2002;277:17933–43. tumourigenesis. Cell Death Dis 2011;2:e124. 35. Mulholland DJ, Read JT, Rennie PS, Cox ME, Nelson CC. Functional 16. Miyamoto H, Yang Z, Chen YT, Ishiguro H, Uemura H, Kubota Y, et al. localization and competition between the androgen receptor and T-cell Promotion of bladder cancer development and progression by andro- factor for nuclear beta-catenin: a means for inhibition of the Tcf gen receptor signals. J Natl Cancer Inst 2007;99:558–68. signaling axis. Oncogene 2003;22:5602–13. 17. Miyagawa S, Satoh Y, Haraguchi R, Suzuki K, Iguchi T, Taketo MM, 36. Chesire DR, Isaacs WB. Ligand-dependent inhibition of beta-catenin/ et al. Genetic interactions of the androgen and Wnt/beta-catenin TCF signaling by androgen receptor. Oncogene 2002;21:8453–69. pathways for the masculinization of external genitalia. Mol Endocrinol 37. Amir AL, Barua M, McKnight NC, Cheng S, Yuan X, Balk SP. A direct 2009;23:871–80. beta-catenin-independent interaction between androgen receptor and 18. Yang F, Li X, Sharma M, Sasaki CY, Longo DL, Lim B, et al. Linking beta- T cell factor 4. J Biol Chem 2003;278:30828–34. catenin to androgen-signaling pathway. J Biol Chem 2002;277:11336–44. 38. Schweizer L, Rizzo CA, Spires TE, Platero JS, Wu Q, Lin TA, et al. The 19. Verras M, Brown J, Li X, Nusse R, Sun Z. Wnt3a growth factor induces androgen receptor can signal through Wnt/beta-Catenin in prostate androgen receptor-mediated transcription and enhances cell growth cancer cells as an adaptation mechanism to castration levels of in human prostate cancer cells. Cancer Res 2004;64:8860–6. androgens. BMC Cell Biol 2008;9:4. 20. Li Y, Zheng Y, Izumi K, Ishiguro H, Ye B, Li F, et al. Androgen activates 39. Yang X, Chen MW, Terry S, Vacherot F, Bemis DL, Capodice J, et al. beta-catenin signaling in bladder cancer cells. Endocr Relat Cancer Complex regulation of human androgen receptor expression by Wnt 2013;20:293–304. signaling in prostate cancer cells. Oncogene 2006;25:3436–44.

www.aacrjournals.org Cancer Res; 73(19) October 1, 2013 5925

Constitutive β-Catenin Activation Induces Male-Specific Tumorigenesis in the Bladder Urothelium

Congxing Lin, Yan Yin, Kristina Stemler, et al.

Cancer Res 2013;73:5914-5925. Published OnlineFirst August 8, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-12-4198

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2013/08/09/0008-5472.CAN-12-4198.DC1

Cited articles This article cites 38 articles, 14 of which you can access for free at: http://cancerres.aacrjournals.org/content/73/19/5914.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/73/19/5914.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/73/19/5914. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.