Sonic Hedgehog (Shh) Signaling Promotes Tumorigenicity and Stemness via Activation of Epithelial-to-Mesenchymal Transition (EMT) in Bladder Cancer

Transcription

1 MOLECULAR CARCINOGENESIS Sonic Hedgehog (Shh) Signaling Promotes Tumorigenicity and Stemness via Activation of Epithelial-to-Mesenchymal Transition (EMT) in Bladder Cancer S.S. Islam, 1,2 R.B. Mokhtari, 1 A.S. Noman, 3 M. Uddin, 4 M.Z. Rahman, 5 M.A. Azadi, 6 A. Zlotta, 7 T. van der Kwast, 8 H. Yeger, 1 and W.A. Farhat 1,2 * 1 Developmental and Stem Cell Biology, The Hospital for Sick Children, Toronto, ON, Canada 2 Division of Urology, The Hospital for Sick Children, Toronto, ON, Canada 3 Department of Biochemistry and Molecular Biology, University of Chittagong, Chittagong, Bangladesh 4 Genetics and Genome Biology, The Hospital for Sick Children, Toronto, ON, Canada 5 Department of Pathology, Chittagong Medical College, Chittagong, Bangladesh 6 Faculty of Biological Sciences, University of Chittagong, Chittagong, Bangladesh 7 Department of Uro-Oncology, Mount Sinai Hospital, Toronto, ON, Canada 8 Department of Pathology, Laboratory Medicine Program, University Health Network, Toronto, ON, Canada Activation of the sonic hedgehog (Shh) signaling pathway controls tumorigenesis in a variety of cancers. Here, we show a role for Shh signaling in the promotion of epithelial-to-mesenchymal transition (EMT), tumorigenicity, and stemness in the bladder cancer. EMT induction was assessed by the decreased expression of E-cadherin and ZO-1 and increased expression of N-cadherin. The induced EMT was associated with increased cell motility, invasiveness, and clonogenicity. These progression relevant behaviors were attenuated by treatment with Hh inhibitors cyclopamine and GDC-0449, and after knockdown by Shh-siRNA, and led to reversal of the EMT phenotype. The results with HTB-9 were confirmed using a second bladder cancer cell line, BFTC905 (DM). In a xenograft mouse model TGF-b1 treated HTB-9 cells exhibited enhanced tumor growth. Although normal bladder epithelial cells could also undergo EMT and upregulate Shh with TGFb1 they did not exhibit tumorigenicity. The TGF-b1 treated HTB-9 xenografts showed strong evidence for a switch to a more stem cell like phenotype, with functional activation of CD133, Sox2, Nanog, and Oct4. The bladder cancer specific stem cell markers CK5 and CK14 were upregulated in the TGF-b1 treated xenograft tumor samples, while CD44 remained unchanged in both treated and untreated tumors. Immunohistochemical analysis of 22 primary human bladder tumors indicated that Shh expression was positively correlated with tumor grade and stage. Elevated expression of Ki-67, Shh, Gli2, and N-cadherin were observed in the high grade and stage human bladder tumor samples, and conversely, the downregulation of these genes were observed in the low grade and stage tumor samples. Collectively, this study indicates that TGF-b1-induced Shh may regulate EMT and tumorigenicity in bladder cancer. Our studies reveal that the TGF-b1 induction of EMT and Shh is cell type context dependent. Thus, targeting the Shh pathway could be clinically beneficial in the ability to reverse the EMT phenotype of tumor cells and potentially inhibit bladder cancer progression and metastasis Wiley Periodicals, Inc. Key words: EMT; Shh; Bladder cancer INTRODUCTION Bladder cancer is the fifth most common cancer in the developed countries and ninth worldwide with clear gender preference for males [1]. In 2012, an estimated new cases of bladder cancer were diagnosed, with bladder cancer related deaths in North America [2]. The incidence of bladder cancer is expected to increase in the future as a result of increasing global population, aging, and exposure to nicotine and occupational chemicals [3]. Greater than 90% of bladder cancers are transitional cell carcinomas (TCC) of urothelial origin. Approximately 70% of the cases are diagnosed as papillary, non-muscle invasive, and 30% as muscle invasive with a worse survival rate documented for the advanced stage and high-grade tumors [4,5]. Muscle invasive tumors penetrate deeply and 50% of patients will relapse with metastases to distant sites despite radical therapy [6]. Treatment failure in 95% of patients with advanced disease, and a poor 5 yr survival rate for metastatic bladder cancer necessitate a better understanding of bladder cancer progression. *Correspondence to: The Hospital for Sick Children, Division of Urology, Toronto, Ontario, M5G 1X8, Canada. Received 15 June 2014; Revised 6 January 2015; Accepted 14 January 2015 DOI /mc Published online in Wiley Online Library (wileyonlinelibrary.com) WILEY PERIODICALS, INC.

2 2 ISLAM ET AL. The epithelial-to-mesenchymal transition (EMT) phenotype is strongly correlated with drug resistance in many cancers including the muscle invasive and metastatic urothelial cancers [7]. EMT, a well-described normal developmental process, also contributes to cancer progression with alterations in cellular morphology, cellular architecture, adhesion, and migration capability. Transforming growth factorb1 (TGF-b1) is a major inducer of EMT during normal development also plays a similar role in many cancer cell types. These EMT phenotypic changes are considered an important mechanism in the early process of metastases and invasiveness of cancer cells [8,9]. The EMT in cancer induced by TGF-b1 supports development of a cancer stem cell like phenotype [10 12]. The EMT phenotype is associated with poor clinical outcome in patients with many different cancers including urothelial tumors [7]. However, the molecular mechanisms associated with TGF-b1 induced EMT in TCC remain largely unknown. Growing evidence implicates activation of the Hedgehog (Hh) signaling pathway in the tumorigenesis of many cancers including lung, ovarian, colon, medulloblastoma, and bladder cancer [13 15]. The role of the Hh pathway in tissue patterning, proliferation, differentiation, and EMT has been established [16]. Hedgehog signaling consists of three signaling proteins, Sonic Hedgehog (Shh), Indian Hedgehog (Ihh), and Desert Hedgehog (Dhh), which bind the 12-pass transmembrane patched1 (Patch1) receptor and repress the associated 7-pass transmembrane Smoothened (Smo). Upon activation, Smo translocates to primary cilia and initiates the activation of Gli transcription factors and ultimately upregulates hedgehog target genes [17,18]. In normal and malignant tissues, the TGF-b and Hh pathways have been shown to regulate key components of each other. TGF-b signaling activation is required for tumor progression in a mouse model of SMOmediated basal cell carcinoma (BCC) development [19]. Furthermore, Shh was shown to promote motility and invasiveness in human gastric cancer through TGF-b1-mediated activation of TbRI-Smad3 pathway [20]. TGF-b1 was also shown to induce Shh expression, which in turn activates Gli1 and Gli1 dependent EMT in non-small cell lung carcinoma [15]. Recent studies in breast, glioblastoma, and lung tumors also implicate involvement of Hh signaling in renewal and survival of cancer stem cells (CSCs) [21,22]. CSCs are considered more resistant to conventional therapies due to their self-renewal ability and intrinsic drug resistance [12]. Despite the current understanding of Shh signaling in cancer, how Shh signaling induced by TGF-b1 relates to acquisition of EMT and tumor aggressiveness in bladder cancer remains unknown. In this study we provide experimental evidence in vitro and in vivo for a critical role of Shh signaling in bladder cancer progression and acquisition of stemness. MATERIALS AND METHODS Cell Lines The human bladder transitional carcinoma HTB-9 cell line represents a grade II tumor and was a kind gift from Dr. Darius Bagli, The Hospital for Sick Children Toronto. Cells were maintained in RPMI-1640, supplemented with 10% FBS and 1% penicillin and streptomycin. The BFTC905 diffuse M, (designated DM here after) bladder cancer cell line, kindly provided by Dr. Jiri Hatina (Charles University in Pilsen, Czech Republic), was derived from a grade III TCC of urinary bladder and maintained in DMEM and 10% FBS. The UROtsa cell line representing immortalized normal human bladder epithelial was obtained from Dr. D. A. Sens, University of North Dakota. ND, USA, and maintained in low glucose DMEM containing 5% v/v FBS. Cells were treated with TGF-b1 (5 ng/ml) in the medium for 72 h. Cells were treated with Hh inhibitors cyclopamine (2 mm) and GDC-0449 (20 nm) for 48 h to block Hh signaling. Clinical Samples Paraffin-embedded tumor tissues were used from 22 patients after surgical excision for bladder cancer at the University of Chittagong and Chittagong Medical College Hospital between June 2012 and July This protocol was reviewed and permitted by Research Ethics Board of Medical College Hospital, Chittagong, Bangladesh. Reagents Reagents used were: recombinant human TGF-b1 protein from R & D Systems (Minneapolis, MN); cyclopamine and GDC-0449 from Selleck Chemicals (Houston, TX, USA) with a stock solution made up in dimethyl sulfoxide (DMSO). Antibodies used: Shh N- terminal antibody (sc-1194), rabbit anti-e-cadherin (sc-7870), goat anti-n-cadherin (sc-31031), goat anti- Gli2 (sc-20291), goat anti-glyceraldehyde 3-phosphate dehydrogenase [GAPDH] (sc-20357), all from Santa Cruz Biotechnology (CA, USA). Rabbit anti- ZO-1 ( ) was purchased from Invitrogen (CA, USA) and mouse anti-ki-67 (M7249) was from Dako, ON, Canada. Morphological Studies and Immunofluorescence Cells were first grown to 70% confluence in a 24- well plate in complete medium, washed twice, and serum starved overnight in serum free medium and then treated with TGF-b1 (5 ng/ml) for 72 h. Cells were then fixed in 4% paraformaldehyde for 15 min, washed with PBS, and observed under phase contrast microscopy. For immunofluorescence studies, cells were plated in a 48-well plate on sterile glass cover slips, treated with TGF-b1 for 72 h, fixed in methanol for 15 min at 208C and briefly washed with PBS. Cells were incubated with 4% Bovine Serum Albumin (BSA) for 30 min at room temperature, followed by three

3 SONIC HEDGEHOG (Shh) SIGNALING PROMOTES TUMORIGENICITY 3 times PBS wash 10 min each, and incubated with primary antibodies at 48C overnight. The primary antibodies used were: rabbit anti-e-cadherin PAb (1:150), rabbit anti-zo-1 PAb (1:500), rabbit anti-ncadherin PAb (1:150). Secondary antibodies used were Alexa Fluor 488 chicken anti-rabbit (H þ L) and Alexa Fluor chicken anti-goat (H þ L) (Invitrogen) diluted 1:1000 in 4% BSA with incubation for 60 min at room temperature in the dark. DAPI (Sigma) was used to counterstain the nucleus, applied for 2 min at room temperature, washed twice with PBS and mounted under cover slips with mounting medium (VectaShield) and observed under the fluorescence microscope. RNA Isolation, Purification, and Quantitative RT-PCR Analysis Total RNA from cell cultures and xenogaft tissues were isolated with DNeasy Mini Kit (Qiagen) and 1 mg total RNA reverse transcribed using superscript II Reverse Transcriptase (Invitrogen) according to manufacturer protocol. The primers sequences are: Shhsense: GTG GCC GAG AAG ACC CTA, antisense: AAG CGT TCA ACT TGT CCT TA, E-cadherin- sense: TCC ATT TCT TGG TCT ACG CC, antisense: CAC CTT CAG CCA ACC TGT TT, N-cadherin- sense: TGT TTG ACT ATG AAG GCA GTG G, antisense: TCA GTC ATC ACC TCC ACC AT, ZO-1- sense: AAG AGA AAG GTG AAA CAC TGC TGA, antisense: GGA AGA CAC TTG TTT TGC CAG GT, Gli2- sense: AGT TTG TTC TCG GGT GCT CTG, antisense: ACA TCT GTC ATC TGA AGC GGC, GAPDH- sense: GAA GGT GAA GGT CGG AGT CAA C, antisense: CAG AGT TAA AAG CAG CCC TGG T. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an internal control for sample normalization. The PCR cycle conditions were as follows: 958C for 15 min, 948C for 60 s, 588C for 30 s, and 728C for 60 s, and total 40 thermal cycles. Each experiment was repeated in triplicate. Relative mrna levels were calculated based on Ct values. Western Blot Analysis Cells were washed twice with PBS and lysed at 48C in ice-cold RIPA buffer (Sigma, Canada, 50 mm Tris-HCL, 150 mm NaCl, 1% NP-40, 0.1 SDS, 0.5% sodium deoxycholate, 2 mm sodium fluoride, 2 mm Na 3 VO 4, 1 mm EDTA, 1 mm EGTA, and supplemented with protease inhibitor cocktail, Roche). Xenograft harvested tumor tissues were minced with a scalpel in ice-cold RIPA buffer. Lysates was incubated on ice for 30 min and then clarified at 14,500 RPM for 15 min. Protein concentration was measured with the BCA kit (Pierce) and 50 mg of protein/lane was separated in SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad). The membranes were blocked by blocking buffer containing 0.1% Tween 20 and 5% non-fat dry milk in PBS and incubated with primary antibodies. Primary antibodies and dilutions used were anti-e-cadherin (1:100), anti-n-cadherin (1:100), anti-shh (1:150), anti-gli2 (1:200), anti-cd-133 (1:300), anti-sox2 (1:400), anti-nanog (1:200), anti-oct4 (1:300), anticytokeratin5 [CK5] (sc-17090), anti-cytokeratin-14 [CK14] (sc-53253), anti-cd44 (ab-41478), and anti- GAPDH (1:200). After three times washing in 0.1% Tween 20, membranes were incubated with HRP conjugated secondary antibodies and immunodetection was performed using chemiluminiscence (ECL, Amersham Life Science) and the signals detected using X-ray film. Equal loading of protein was confirmed by reprobing membranes for GAPDH. Experiments were repeated three times and representative blots are presented. Wound Healing Assay Cells were seeded at a density of cells/well in 500 ml culture medium in triplicate in a 48-well plate and allowed to adhere overnight, treated with TGF-b1 (5 ng/ml), and incubated for 72 h. To inhibit Hh signaling, cells were treated with Hh inhibitors cyclopamine (2 mm) and GDC-0449 (20 nm) for 48 h or Shh knocked down with Shh specific sirna for 24 h. At 90% confluence cell monolayers were scratched across the monolayer in each well with a 200 ml pipette tip. Non-adherent cells were washed off with medium and fresh medium was added to the cells and incubated for up to 72 h. Phase contrast light microscopic images (4 magnification) were taken at time points 0, 48, and 72 h. The percent closure of the wound area by migratory cells was measured using the NIH- ImageJ program to quantify healing in the wound area. Migration and Matrigel Coated Invasion Assays Cells were first seeded in 10 cm tissue culture dishes and treated with or without TGF-b1 for 72 h in complete culture medium. Cells were treated with Hh inhibitors cyclopamine (2 mm) and GDC-0449 (20 nm) for 48 h and Shh knocked down with Shh specific sirna for 24 h. Cells were harvested and seeded into the upper chambers of uncoated and coated inserts at 50,000 cells/well. Uncoated inserts were used for migration assay and Matrigel TM coated chamber inserts for invasion assay (24-well plate, 8 mm pores, BD Bioscience) and cells were seeded in low serum (0.1% FBS) medium. Lower chamber contained regular complete medium with 10% FBS as chemoattractant. After 24 h of incubation, cells on the upper inserts were wiped off carefully using cotton swabs. Cells that invaded and migrated through the filters were stained with 1% crystal violet for 30 min at room temperature and washed with PBS. The number of invaded cells was determined by counting all cells attached to the bottom of the inserts under a phase contrast microscope at (4 magnification) and quantified using the NIH Image-J software. The experiments were done in triplicate.

4 4 ISLAM ET AL. Methylcellulose Clonogenic Assay The clonogenic assay was performed on HTB-9 and BFTC905 (DM) cells treated with TGF-b1 (5 ng/ml) for 72 h and after treatment with Hh inhibitors, cyclopamine (2 mm), and GDC-0449 (20 nm) for 48 h or knock down Shh with Shh specific sirna for 24 h. Then cells were trypsinized and re-suspended ( cells/ml) in a medium containing 40% methylcellulose (Stem Cell Technologies) supplemented with 49% RPMI-1640, 10% FBS and 1% antibiotics (100 IU/ ml penicillin and 100 mg/ml streptomycin). Suspended cells were plated in 35 m 2 tissue culture dishes (NUNC, Nalgen Nunc International, Cat # ) in triplicate and incubated in 5% CO 2 and 378 C. After 2 wk the number of colonies were counted (5 fields/ dish) by using a grading dish on a phase contrast microscope ( 10 magnification). Colony forming potential was determined as the number of colonies formed by number of cells seeded in five fields. The experiments were repeated three times and a representative experiment is presented. Shh sirna Design and Silencing of Gene Expression with Small Interfering RNA (sirna) Custom designed Shh specific sirna was purchased from Applied Biosystems, Ambion. The sequence for Shh sirna is: Sense- 5 0 CGACAUCAUAUUUAAGGAYtt3 0 and antisense- 5 0 AUCCUUAAAUAUGAUGUCGgg3 0. As a negative control non-targeting scrambled sirna was purchased from Applied Biosystems, Ambion. All experiments were performed in triplicate and a representative experiment is presented. The transfection mixture was prepared by mixing 10 ml of Lipofectamine-RNAiMAX and 190 ml Opti-MEM1 with 10 ml sirna and 190 ml Opti-MEM1 separately at room temperature. Combined diluted sirna duplex together with the Lipofectamine- RNAiMAX were incubated for 30 min at room temperature. Finally, the above transfection mixtures were added to 35 mm culture dishes. Transfection was performed for 24 h at 378C in a CO 2 incubator. Medium containing sirna was then replaced with fresh medium for a further 12 h and TGF-b1 (5 ng/ml) was added to the culture and incubated in 5% CO 2 for 48 h and then harvested for protein analysis. In Vivo Animal Study Four to six weeks old female NOD/SCID mice (Charles River, Quebec, Canada) were used in xenogarft studies. Briefly, cells were harvested and diluted in serum free medium, mixed with Matrigel and cells/mice were injected subcutaneously. When the tumors attained the diameter of 0.5 mm, the mice were randomized into two groups (five mice for each group). Tumor diameters were measured with slide calipers on a daily basis for 60 d. Tumor volume (cm 3 ) was calculated as V ¼ 0.5 D d 2. After euthanizing the mice on day 60 the tumors were resected, weighted, and fixed for further studies. Statistical Analysis All experiments were performed at least three times. The data presented here were expressed as mean standard error of the mean (SEM), analyzed by student t test for difference between groups or variance analysis for multiple comparisons among groups. Development of xenograft tumors in NOD/SCID mice in the untreated and treated groups was analyzed by t test. P < 0.05 was considered as statistically significant. RESULTS TGF-b1 Promotes EMT, Migration, Invasion, and Clonogenicity in Bladder Cancer Cells In Vitro We have shown Smad2/3 dependent EMT in porcine bladder primary epithelial cells [23]. Here we sought to investigate whether bladder cancer TCC cells undergo and exhibit typical characteristics of EMT. In the absence of TGF-b1, most cells showed a more typical epithelial cobblestone like shape and tight cell- cell adhesion (Figure 1A). However, 72 h after TGF-b1 treatment, HTB-9 (designated as T-HTB-9 hereafter) cells exhibited EMT changes acquiring spindle type cell morphology and reduced cell- cell contacts (Figure 1A). To demonstrate the cellular localization of relevant EMT proteins, we immunolocalized E-cadherin, ZO-1, and N-cadherin by immunofluorescence, and representative images from multiple fields from the entire cell population are shown in Supplementary Figure 1. Figures show a significant reduction in immunostaining for epithelial markers E-cadherin and ZO-1 and increased expression of N-cadherin after TGF-b1 treatment consistent with EMT. In contrast, untreated HTB-9 cells showed strong expression of E-cadherin and ZO-1 and a much weaker expression of N-cadherin (Supplementary Figure 1). Figure 1B shows that E-cadherin and ZO-1 expression were downregulated by TGF-b1 treatment at the mrna level and increased expression was observed for N-cadherin. A similar quantitative change was determined at the protein level (Western blot). E-cadherin and ZO-1 expression were downregulated after TGF-b1 treatment while N-cadherin expression was markedly upregulated with a significant increase (55 60%) in expression after TGF-b1 treatment (Figure 1C). Taken together, the morphological changes and loss of E-cadherin and ZO-1 and gain of N-cadherin after TGF-b1 treatment is indicative of an EMT contributing to the migratory and invasive behavior of HTB-9 bladder cancer cells. Considering the effects TGF-b1 mediated EMT, we then investigated the migration, invasion, and clonogenic behavior of T-HTB-9 cells. Using the short term in vitro wound-healing assay we showed that T-HTB-9 cells migrated and covered approximately 95% of the scratch defect after 48 h (Figure 1D). By comparison, the untreated parental HTB-9 cells migrated with slower kinetics and only covered 45%

5 SONIC HEDGEHOG (Shh) SIGNALING PROMOTES TUMORIGENICITY 5 Figure 1. TGF-b1 promotes EMT, migration, invasion, and clonogenicity in bladder cancer cells: Cultured HTB-9 cells were treated with TGF-b1 (5 ng/ml) for 72 h and morphological, molecular, and functional parameters of EMT were assessed. (A) Phase contrast microscopic images (original magnification 10) of HTB-9 cells switch to a mesenchymal phenotype (referred as T-HTB-9 cells in later) exhibited by a shape change to elongated and non-polarized compare to epithelial like parental HTB-9 cells. (B) qrt-pcr results of parental HTB-9 and T-HTB-9 cells. TGF-b1 stimulation for 72 h showed lower expression of E-cadherin and ZO-1 and higher expression of N-cadherin at the mrna levels compared to parental HTB-9. Delta-delta Ct values were calculated, considering GAPDH as an internal control. (C) Western blot analysis confirmed the downregulation of E-cadherin, ZO-1, and upregulation of N-cadherin in T-HTB-9 cells compared to parental HTB-9 cells with quantitative densitometric analysis. (D) Wound healing assay with its quantitative analysis. T-HTB-9 cells showing increased motility compared to parental HTB-9 cells. T-HTB-9 cells migrated and covered approximately 90% of the scratch defects after 72 h (P < ). (E) Matrigel coated membrane migration assay. More T-HTB-9 cells migrated through the membrane compared to parental HTB-9 cells (P < ). (F) Increased invasion through Matrigel coated membranes (P < 0.005) and (G) Increased clonogenic capacity of T-HTB-9 cells compared to parental HTB-9 cells (P < 0.005) (original magnification 10). of the scratch defect (Figure 1D, P < ). We further determined the effects of TGF-b1 on chemotaxis in the T-HTB-9 cells using transwell migration assay with 10% serum as chemo-attractant. Significantly more T-HTB-9 cells migrated through the membrane compared to untreated parental HTB-9 cells (Figure 1E; P < , 184 þ / 15 cells/hpf vs. 83 þ / 9 cells/hpf). To determine T-HTB-9 cell invasiveness, we used the Matrigel invasion assay and quantified the numbers of invaded cells. T-HTB-9 cells showed an increased level of invasion compared to untreated parental HTB-9 cells (Figure 1F; (P < 0.005, T-HTB-9; 83 þ / 10 cells/hpf vs. HTB-9; 59 þ / 7 cells/hpf). We next investigated the clonogenic potential of T-HTB-9 cells. T-HTB-9 cells acquired a significantly higher degree of clonogenicity as well as producing larger colonies compared to untreated parental HTB-9 cells (Figure 1G; P < 0.005). These results suggest that TGF-b1 significantly increases motility, migration, invasiveness, and clonogenic capacity of human bladder HTB-9 cancer cells. T-HTB-9 Cells Show Upregulation Shh and Gli2 in mrna and Protein Level, and Hh Inhibitors Inhibits Migration, Invasion, and Clonogenicity We next examined whether HTB-9 cells upregulate Shh after TGF-b1 stimulation. qrt-pcr and Western blot results show that Shh was not detectable at both mrna and protein levels in the parental HTB-9 cultures. However, TGF-b1 stimulation induced Shh expression to a relatively high degree (Figure 2A, top left panel). To assess the extent of Hh pathway activation in the cultures, we analyzed for expression of the Hh transcription factor Gli2 by qrt-pcr and Western blots. Results showed that Gli2 expression was induced considerably relative to the untreated

6 6 ISLAM ET AL. Figure 2. T-HTB-9 cell show upregulation of Shh and Gli2 in mrna and protein level, and Hh inhibitors inhibits migration, invasion, and clonogenicity. (A) qrt-pcr and Western blot results showing the mrna and protein expression of Shh and Gli2 in T-HTB-9 cells compared to parental HTB-9 cells. (B) HTB-9 cells were cultured with and without TGF-b1 for 24 h. The cultured supernatants were collected and added to new HTB-9 cell cultures and incubated for 48 h. Cells were collected and lysed and analyzed the expression of Shh and Gli2 by Western blot. HTB-9 cells treated with conditioned medium show the higher expression of Shh and Gli2 compared to untreated cells. HTB-9 cells were treated with human recombinant Shh protein and assessed for expression of Shh and Gli2 and EMT phenotypes by Western blot. Human recombinant Shh protein upregulated Shh protein but failed to express EMT phenotypes as TGF-b1 treated HTB-9 cells upregulated both Shh protein and expressed EMT markers. (C) Wound healing, (D) Matrigel coated membrane invasion, and (E) Clonogenic characteristics with quantitative analysis. T-HTB-9 cells showing attenuation of migration, invasion, and clonogenic capacity. (F) Western blot analysis of the expression of epithelial marker E- cadherin and mesenchymal marker N-cadherin in T-HTB-9 cells treated with cyclopamine and GDC-0449 and showing the reversal of E- cadherin and downregulation of N-cadherin (P < 0.05). Patch1 and Smoothened was evaluated in the cyclopamine and TGF-b1 treated cells. No change in Patch1 expression was observed, but Smo expression was decreased in the cyclopamine treated cells. control (Figure 2A, top left panel). We then investigated whether EMT markers were originating from Shhþ and Gli2þ cells. We co-stained with Shh þ EMT markers and Gli2 þ EMT markers and found that, Shh and Gli2 were overwhelmingly expressed in the TGFb1 treated cells. Shh was localized in the cytoplasm, whereas, Gli2 was found to be localized in the nucleus in the same cell population. EMT markers (Ecadherin, ZO-1, and N-cadherin) co-localize with Shhþ and Gli2þ cells (Supplementary Figure 2), suggesting a strong role of Shh pathway components in the EMT mechanism. Since Shh is a secretable factor and is involved in an autocrine signaling in many cancers (14), we therefore determined whether upregulation of secretable Shh protein was present in the HTB-9 cell culture medium. First we cultured HTB-9 cells with and without TGF-b1. After 48 h of incubation the culture supernatants were collected and added to fresh HTB-9 cell cultures and incubated for 72 h. Western blot results showed a very low level (undetectable) of Shh protein expression in cells with unstimulated medium (Figure 2B; bottom left panel; not treated with supernatants). However, Shh protein expression was increased significantly in the cells treated with supernatants from TGF-b1 stimulated cells (Figure 2B, bottom left panel). Similarly constitutively expressed Gli2 was increased in the HTB-9 cells

7 SONIC HEDGEHOG (Shh) SIGNALING PROMOTES TUMORIGENICITY 7 treated with TGF-b1 compared to untreated HTB-9 cells (Figure 2B). These results suggest either active autocrine and/or paracrine Shh signaling in the T-HTB-9 cells. We next investigated whether soluble Shh protein alone could activate Shh signaling and EMT. We exposed HTB-9 cells to recombinant Shh-Nterminal (rshh-n) protein to the cultures with / þ TGF-b1. As shown in Figure 2B (bottom left panel), rshh-n protein treated cells show upregulation of Shh and Gli2 expression at the protein level but this did not reach the same level as in TGFb1 alone treated cells. The switching in the expression of EMT markers E-cadherin and N- cadherin was significant in the TGF-b1 only treated cells but not after treatment with rshh-n protein alone (Figure 2B, bottom left panel), suggesting a reduced EMT effect. Shh and Gli2 protein expressions were associated with decreased expression in E-cadherin and increased expression in N-cadherin (Figure 2B, bottom left panel). Together these data suggest that, transcriptional activation of Shh follows after the TGF-b1 treatment, which results in EMT phenotypic changes in T-HTB-9 cells. The data also suggest possible paracrine signaling. We next sought to verify if the Hh inhibitors could potentially inhibit the migration, invasion, and clonogenicity of T-HTB-9 cells. T-HTB-9 cells were treated with Hh pathway inhibitors cyclopamine and GDC-0449 and migration, invasion, and clonogenicity were assessed. Significant inhibition of wound closure was observed with both cyclopamine and GDC-0449 inhibitors (Figure 2C, top right panel). In contrast, 90 95% wound closure was seen after 48 h in T-HTB-9 cells (Figure 2C; P < 0.005). Blocking of Hh pathway by cyclopamine and GDC-0449 also significantly attenuated the invasive potential of T-HTB-9 cells (Figure 2D; middle right panel). We further found that both cyclopamine and GDC-0449 treatment caused a highly significant decrease in clonogenicity (Figure 2E; bottom right panel; cyclopamine- P < and GDC P < ) with a loss of the large colonies seen in controls. We then assessed the changes in EMT markers by Western blot and found that both cyclopamine and GDC treated T-HTB-9 cells revealed an upregulation of E-cadherin and a significant reduction in N- cadherin expression (Figure 2F, bottom right panel). We further determined the effects of Patched1 (Patch1) and Smoothened (Smo) in the T-HTB-9 cells. Treatment of T-HTB-9 cells with cyclopamine did not changed the Patch1 protein level, however, Smo protein level in the cyclopamine treated cells decreased moderately (Figure 2F). Thus blocking Shh activity with the inhibitors significantly inhibited migration, invasive ability, and clonogenic capacity on T-HTB-9 cells. The gain in E-cadherin and loss in N-cadherin by inhibitors suggests a critical threshold level required to support the alterations in tumor cell behavior. These results identify Shh signaling as a key regulator in the mechanism of TGF-b1-induced EMT. Upregulation of Shh Occurs Concomitant With TGF-b1 Induced EMT, and Migration, Invasion, and Clonogenicity are Attenuated by Shh-siRNA Having demonstrated that pharmacologically blocking Shh activation in HTB-9 cells countered TGF-b1 induction of EMT, we sought evidence at the transcriptional level. We first knocked-down Shh in HTB-9 cells with sirna targeting Shh (designated as Shh-siRNA) for 24 h before 24 h TGF-b1 treatment (total incubation period 48 h). Supplemental Figure 1 shows the transfection efficiencies in HTB-9 cells under phase contrast and fluorescence microscope. Results confirmed that HTB-9 cells transfected with Shh-siRNA maintained an epithelial morphology (Supplemental Figure 3). In contrast, scrambled sirna (sc-sirna) showed no change to the morphological changes acquired during the mesenchymal transition (Supplemental Figure 3). To obtain more uniform and enhanced affects of Shh-siRNA we re-transfected cells for a second round for 48 h before stimulating the cells with TGF-b1 for 48 h (total incubation period 96 h). Consistent with a single round transfection we observed that sc-sirna treatment revealed an enhancement of the EMT induction compared to cells transfected with Shh-siRNA, which was confirmed by morphological changes in HTB-9 cells (see Supplemental Figure 1). To identify the potential molecular mechanisms underlying the EMT process qrt-pcr analysis was performed. As shown in Figure 3A (top panel), the Shh expression was significantly decreased at the mrna level after TGF-b1 stimulation for 48 and 72 h in cells transfected with Shh-siRNA. In contrast, Shh expression remained elevated in cells treated with sc-sirna. These results demonstrated that Shh-siRNA could efficiently downregulate the expression of Shh at the mrna level. To confirm whether EMT markers also changed due to silencing of Shh, we then investigated the changes in epithelial and mesenchymal markers after knockdown of Shh by Shh-siRNA. We observed a significant downregulation of E-cadherin and ZO-1 in cells treated with ShhsiRNA in contrast to the significant upregulation of E-cadherin and ZO-1 in the sc-sirna treated cells as confirmed by qrt-pcr (Figure 3B and C; top panel). On the other hand, knockdown of Shh by Shh-siRNA did not affect significantly N-cadherin mrna expression in the same time period after TGF-b1 treatment (Figure 3D; top panel). We next applied Shh-siRNA to assess the migration, invasion, and clonogenic potential. Results demonstrate that, corresponding to that seen with the Shh inhibitors cyclopamine and GDC-0449, a significant inhibition of wound healing was observed in ShhsiRNA treated cells after 48 h (Figure 3E; bottom left

8 8 ISLAM ET AL. Figure 3. Upregulation of sonic hedgehog (Shh) is concomitant with TGF-b1-induced EMT and migration, invasion, and clonogenicity is attenuated by Shh-siRNA: HTB-9 cells were transfected with Shh specific sirna (Shh-siRNA) and scrambled-sirna (sc-sirna) for 24 h and 48 h before addition of TGF-b1 (5 ng/ml) in the culture and cells were collected for experiments. (A) Downregulation of Shh by ShhsiRNA, (B and C) downregulation epithelial markers E-cadherin and ZO- 1 by sc-sirna and (D) Upregulation of mesenchymal marker N-cadherin showing the qrt-pcr based mrna expression after transfection of HTB-9 cells with compared to sc- sirna (P < 0.01). (E) T-HTB-9 cells were transfected with Shh-siRNA and cell motility was assessed by wound healing assay. The wound closure rate was monitored at 0 and 72 h. Shh-siRNA significantly inhibited the wound closure of T-HTB-9 cell compared to sc-sirna (P < 0.005). (F) Showing the Matrigel-coated membrane T-HTB-9 cells invasion with quantitative analysis. Significantly lower number of cells was invaded through Matrigel membrane (P < 0.005). (G) Clonogenic capacity with quantitative analysis. ShhsiRNA inhibited T-HTB-9 cells clonogenic capacity compared to scsirna (P < 0.005). panel; P < 0.006). In contrast, 95% wound closure was seen after 48 h in untreated T-HTB-9 cells. Similarly, significant inhibitions were observed in cell invasion and clonogenic capacity (Figures 3F and G; bottom right panel, invasion - P < and clonogenicity - P < 0.005). These results further support the importance of Shh in mediating the aggressive tumor cell phenotype via EMT. To verify the importance of Shh in bladder cancer, we studied another bladder TCC cell line BFTC905 (DM) and a normal transformed non-tumorigenic bladder epithelial cell line (UROtsa cells). First, the cells were treated with TGF-b1 and EMT potential was evaluated by immunofluorescence, qrt-pcr, and Western blot. Results demonstrated that, these cells underwent EMT as confirmed by downregulation of E-cadherin and upregulation of N-cadherin in mrna and protein levels (Figure 4A; top panel). Although these cells intrinsically expressed a low level of Shh, TGF-b1 further enhanced the expression of Shh shown by qrt-pcr and Western blot (Figure 4A; top panel). To verify the Shh effects on migration, invasion, and clonogenic potential, the cells were cultured with and without TGF-b1, and cells were treated with Shh inhibitors cyclopamine, GDC-0449, and Shh-siRNA for 24 h. We observed that the migration, invasion, and clonogenicity were decreased significantly compared to untreated cells (Figure 4A; top panel). These results lend further strong support for the role of Shh in the EMT. Supplemental Figure 4 shows this EMT switch by immunofluorescence staining for E-cadherin and N- cadherin after TGF-b1 treatment. In comparison, the normal transformed human bladder epithelial UR- Otsa cells also underwent EMT after TGF-b1 treatment as confirmed by the downregulation of E-cadherin and upregulation of N-cadherin (Figure 4B; bottom panel; also Supplemental Figure 4). These cells did

9 SONIC HEDGEHOG (Shh) SIGNALING PROMOTES TUMORIGENICITY 9 Figure 4. Autocrine Shh signaling reduce the migration, invasion, and clonogenicity in BFTC905 (DM) bladder cancer cell. Cultured BFTC905 (DM) cells were treated with TGF-b1 (5 ng/ml) for 72 h and morphological, molecular, and functional parameters of EMT were assessed by qpcr and Western blotting. (A; top panel) BFTC905 (DM) cells showing the downregulation of E-cadherin and upregulation of N- cadherin after TGF-b1 treatment and concomitantly upregulate Shh expression in the mrna and protein level. Treatment of these cells by cyclopamine and GDC-0449 and Shh-siRNA reduced the migration, invasion, and clonogenic capacity. (B; bottom panel) A non- malignant transformed bladder primary uroepthelial cells (UROtsa) showing the similar results found in Figure 4A. show expression of Shh without TGF-b1, however, the expression level was increased once the cells were treated with TGF-b1 suggesting a critical role for ligand dependent Shh activation (Figure 4B; bottom panel). However, in contrast to the tumor cell line, the UROtsa cells failed to form tumors in the NOD/SCID mice even when cells were treated with exogenous TGF-b1 (data not shown). Thus the Shh pathway supports the malignant potential of the tumor lines. T-HTB-9 Cells Show Increased Tumor Growth and Acquire Stemness Characteristics in NOD/SCID Xenografted Mice The clonogenic assay suggested a potent growth stimulating effect of TGF-b1 and requirement for Shh. We treated HTB-9 cells with and without TGF-b1 for 24 h in vitro and then washed, trypsinized and cells were injected subcutaneously into NOD/ SCID mice. The mice were monitored weekly and after 45 d the tumors were resected. Results demonstrate that TGF-b1 pre-treatment significantly accelerated tumor growth and size, when compared to the untreated group (Figure 5A). Tumor weights and tumor volumes were significantly higher in the TGFb1 treated group than the untreated counterpart with comparable kinetics suggesting an increase in the proliferative capacity (Figure 5A). EMT has been shown to be a critical step in the initiation and conversion of early stage tumors into invasive malignancies and is also associated with the stemness of cancer cells [24]. Having found that TGFb1 stimulated formation of large cell colonies in the clonogenic assay, we posited a possible stem cell like contribution. We therefore investigated whether the increased tumorigenic potential of TGF-b1 treated HTB-9 cells correlated with a possible increase in the stemness phenotype. We assessed for expression of putative stem cell markers, CD133, and embryonic stem cell transcription factors Sox2, Nanog, and Oct4

10 10 ISLAM ET AL. Figure 5. T-HTB-9 cells show increased tumor growth and stemness characteristics in NOD/SCID xenograft mice: Cultured HTB-9 cells were treated with TGF-b1 (5 ng/ml) for 24 h and trypsinized. Two groups of mice (five mice in each group) were implanted tumors separately. Each NOD/SCID mouse was injected cell suspension into abdomen of each mouse mixed with serum free media and high concentration growth factor Matrigel matrix. The mice were monitored weekly for the growth of tumors and measured the tumor by Slide Calipers for 45 d. (A) Morphology of tumor growth, tumor volume. qpcr analysis of (B) CD133, (C) Sox2, (D) Nanog, (E) Oct4, (F) CK5, (G) CK14, and (H) CD44 5D. (I) Western blot analysis showing the expression of stem cell markers (as analyzed in Figure 5B H by qpcr) in xenograft mouse tumor tissues with the quantification of expression changes by densitometric analysis normalized by GAPDH. and bladder cancer specific stem cell markers, cytokeratin5 (CK5), cytokeratin14 (CK14), and CD44 by qrt- PCR and Western blot in xenograft tumor tissues. Twofold expression of CD133 was detected at the mrna level in the TGF-b1 treated tumor group compared to untreated tumor (Figure 5B). We further observed that TGF-b1 treated tumors showed strong staining of Sox2 and Nanog and a 2-fold higher expression of Sox2 (Figure 5C) and Nanog (Figure 5D) at the mrna. Oct4 was not detected in the untreated tumors, however, its expression was found to be elevated in the treated tumor group (Figure 5E). The expression levels of CK5, CK14, and CD44 at the mrna levels show very similar expression levels with a 2-fold change in the mrna level in the TGF-b1 untreated and treated xenograft tumors (Figure 5F H). These relative increases in vivo were further confirmed by Western blot (Figure 5I) with quantification by densitometry analysis. These observations support the idea of a transition to a more stem cell like and mesenchymal like, more proliferative, and invasive phenotype under the influence of TGF-b1. Elevated Shh and Gli2 Expression is Correlated With Advanced Clinical Stage and Poor Prognosis in Patients With Bladder Cancer We next asked if patient tumors would exhibit expressions of Ki-67, Shh, Gli2, and EMT markers as evidenced in the bladder cancer cell lines. In addition, to further correlate we assessed the expression levels of Shh and proliferation and EMT markers in the xenogaft tumors by qrt-pcr and Western blot. We evaluated 22 bladder tumor patient samples and stained them for the expression of Ki-67, Shh, Gli2, E-cadherin, and N-cadherin. The 22 cases represented low to high malignant stages of bladder cancer. Five representative cases are presented here. Table 1 lists the tumor grade and stage, positive staining and

11 Table 1. Shh, Gli2, Ki-67, E-cadherin, and N-cadherin are Enhanced in Primary Human Bladder Tumor Specimens Case no. Pathology Shh Gli2 Ki-67 E-cad N-cad 1 High grade/high stage (G4/T4) þþþ þþþ þþþ þþþ 2 High grade/high stage (G4/T4) þþþ þþþ þþþ þþþ 3 High grade/high stage (G3/T3) þþþ þþþ þþþ þ þþþ 4 High grade/high stage (G3/T3) þþþ þþþ þþþ þþþ 5 High grade/low stage (G3/T2) þþ þþþ þþþ þþþ 6 High grade/low stage (G3/T2) þþ þþ þþ þ þþ 7 High grade/low stage (G3/T2) þþ þþ þþ þ þþ 8 Low grade/low stage (G2/T2) þþþ þþþ þþ þþþ 9 Low grade/low stage (G2/T2) þþ þþ þþþ 10 Low grade/low stage (G2/T2) þþ þþ þþþ þ 11 Low grade/low stage (G2/T2) þ þþ þþ þþþ þ 12 Low grade/low stage (G2/T2) þ þþ þþ þþþ þ 13 Low grade/low stage (G2/T2) þ þþ þþ þþ 14 Low grade/low stage (G2/T2) þþ þþ þþþ 15 Low grade/low stage (G2/T2) þ þþþ þ 16 Low grade/low stage (G2/T2) þþ þþ þþ 17 Low grade/low stage (G2/T2) þþ þ þþþ þ 18 Low grade/low stage (G2/T2) þ þþ þþþ þ 19 Low grade/low stage (G2/T2) þþ þþ þþþ þ 20 Low grade/low stage (G2/T2) þþ þþ þþþ 21 Low grade/low stage (G1/T2) þþ þþ þþþ 22 Low grade/low stage (G1/T1) þþ þ þþþ Summary of the expression SONIC HEDGEHOG (Shh) SIGNALING PROMOTES TUMORIGENICITY 11 UC tumor Shh-positive; Gli2-positive 11 Shh-positive; Gli2-negative 1 Shh-negative; Gli2-negative 8 Shh-negative; Gli2-negative 2 Total 22 expression intensity. The proliferation marker Ki-67 was expressed in all grades and stages of tumors, however the G1/T2 tumor showed a comparatively reduced expression (Figure 6A; left panel), suggesting that Ki-67 expression was significantly positively correlated with advanced tumor grade. Similarly, increased levels of Shh and Gli2 expression were associated with high grade (G4, G3, and G2) and high stages (T4, T3, and T2) tumors, whereas Shh expression intensity was weaker in the low grade (G1) and stage tumors (T1). Shh expression was found to be higher in cells at the tumor stroma border and particularly in those cells invading the surrounding stroma. However, nuclear localized Shh positive cells in the center of the tumor were found to be comparatively lower in G1/T2 versus the higher stages (Figure 6A; left panel). Similarly, nuclear localized Gli2 expression was found to be relatively higher in all tumor samples examined in the study (Figure 6A; left panel). To determine the EMT phenotype in the human bladder tumor samples, we examined the E-cadherin and N-cadherin expression levels. In general, E-cadherin expression was weakly positive in the G2/T3, G3/T2, and G4/T4 tumors specifically at the invasive front E-cadherin expression was faint or even absent. In contrast, in the G1/T2 tumor was strongly positive for E-cadherin at the cell-cell contacts (Figure 6A; left panel). Whereas, N-cadherin expression was significantly upregulated in all high grade and stage tumors, the G1/T2 tumor showed weak or even no expression of N-cadherin. N-cadherinpositive tumor cells were concentrated primarily in areas of tumors-stroma interaction and were scattered throughout the stroma (Figure 6A; left panel), suggesting a strong correlation between tumor grade, EMT, and metastatic potential. To validate the expression patterns found by immunohistochemistry in the patient tumors we analyzed the expression of these genes by qpcr and Western blot in our xenograft tumor sample. We found that Ki-67, Shh, and N-cadherin expression were upregulated in the TGF-b1 treated tumors, however, E-cadherin and ZO-1 expression were downregulated at the mrna and protein levels (Figure 6B; right panel). Again no obvious changes were observed in Gli2 expression where Gli2 expression was equivalent in TGF-b1 treated and untreated tumors (Figure 6B; right panel). These findings suggest that Shh positive tumors cells associate closely with stromal cells present at the borders within the tumor where Gli2 expression is induced, and is likely that ligand-dependent Shh activation occurs resulting in increased invasion and subsequent metastasis.

12 12 ISLAM ET AL. Figure 6. Expression pattern of Shh, Gli2, Ki-67, and EMT markers in human bladder tumor samples. Low and high stage tumors were isolated form bladder cancer patients and stained for markers expression. (A) Immunocytochemistry staining showing the expression of Ki-67, Shh, Gli2 in the high grade, and high stage patient tumors. The EMT marker E-cadherin shows high expression in low stage and low-grade tumors, however, expression decreased in high stage and high-grade tumor. Conversely N-cadhein expression was low in lower stage/grade tumors and its expressions increased in high stage/grade tumors (insert-higher magnification; original magnification 40). (B) qpcr (top) and Western blots (middle) (with densitometric quantification) analysis of the expression of Ki-67, Shh, Gli2, E- cadherin, ZO-1, N- cadherin, from xenograft tumors. (C and D) RNA-Sequence expression profiling of 67 (sixty seven) human muscle invasive bladder cancer tumor samples: A gene signature consisting of 20,532 genes expressed in tumor samples correlated with genes involved in UC progression. A meta-analysis of 67 (sixty seven) UC samples were performed from publicly available bladder cancer studies curated in TCGA. (C) Boxplots showing the RPKM values from RNA-seq samples of all analyzed Shh, Gli2, E-cadherin, N-cadherin, fibronectin, and vimentin from TCGA curated genes. The 50th and 75th percentile values computed as 2.68 and 11.83, respectively. Any values falls below the 2.68 and RPKM values considered as the downregulation of gene expression. (D) Scatter plot shows the correlation between Shh and Gli2 expression with the EMT markers.

13 SONIC HEDGEHOG (Shh) SIGNALING PROMOTES TUMORIGENICITY 13 In order to validate whether Shh, Gli2, and EMT markers are indeed upregulated in the bladder tumor samples based on our cohort, we took advantage of publicly available RNA-seq data (67 samples) from The Cancer Genome Atlas (TCGA) and compared the average expression of Shh, Gli2, E-cadherin, N- cadherin, fibronectin, and vimentin (RPKM values). The pre-computed Reads Per Kilobase of Transcripts per Million (RPKM) from public RNA-Seq samples were obtained for 20,532 genes in the genome. The data were normalized using quantile normalization. To compare the RPKM values of our genes of interest, we computed the 50th and 75th percentile RPKM values from the entire dataset (all genes from all samples). Any RPKM value for a gene greater than 50th percentile was considered a highly expressed gene due to upregulation (Figure 6C). The values for 50th and 75th percentile threshold levels computed from the whole genome genes (20,532 genes) expression are 2.68 and 11.83, respectively. The expression values for Shh, Gli2, N-cadherin, fibronectin, and vimentin were significantly higher demonstrating that these genes are mostly upregulated in UC tumors (Figure 6C and D). These results suggest that the Shh dependent cells are represented to a greater extent in the more tumorigenic cell population found in high grade and stage muscle invasive bladder cancers. DISCUSSION We have provided strong evidence for a role of Shh signaling in mediating the tumorigenic potential and malignant phenotype in bladder cancer cells both in vitro and in vivo. Essentially, bladder cancer cells undergo TGF-b1-induced EMT, exogenous TGF-b1 treatment activates the Shh signaling pathway, and results obtained from in vitro and in vivo studies demonstrate that Shh may function as an inducer of bladder cancer progression. Shh expression correlates with advanced clinical stages and poor prognostic bladder cancer. We show that human bladder TCC cells (in two bladder cancer cell line models) undergo TGF-b1 mediated EMT with decreased expression of E-cadherin and ZO-1 and increased expression of N-cadherin. A key finding here is that prolonged exposure to TGF-b1 increased ability of bladder cancer cells to migrate, invade, and form colonies. This is consistent with recent data on a lung cancer cell line A549, where TGF-b1 treatment also induced EMT and showed increased ability for migration, invasion, and tumorigenicity [15]. Increase in cloncogenicity was also reflected in a markedly increased ability for tumorigenicity suggesting that exposure to TGF-b1 in the tumor microenvironment could enhance bladder tumor growth. Our data convincingly implicate upregulation of Shh pathway signaling in this process. Gli2 expression also increased above endogenous levels in vitro. This suggests that an increased level of Gli2 may be associated with TGF-b1-induced HTB-9 cell migration, invasion, and clonogenicity. The role of Gli2 transcription factor in growth and tumorigenicity in prostate cells was recently described [25]. Ligand-dependent activation of Hh pathway has been shown in small cell lung carcinoma, esophageal carcinoma, and gastric, and pancreatic carcinoma [13,15]. Our results, together with other previous studies, suggest that urothelial bladder cancer cells appear to mediate intrinsic Hh signaling through the expression of Shh ligand in an autocrine fashion. A recent study [26] revealed that bladder tumors showed weak to strong expression of Shh protein. Our survey of patient bladder cancers further supports a role of Shh in tumor progression and EMT. We demonstrated that the activation of TGF-b1-induced Shh signaling leads to increase tumor cell motility, invasion, and tumorigenic capability of bladder cancer cells. Inhibition of the Shh signaling pathway is considered to have therapeutic value in the treatment of cancers. Currently several Hh inhibitors are under clinical and pre-clinical development and are being tested for efficacy [27]. TGF-b1 driven EMT in non-small cell lung carcinoma cells was shown to inhibited by pharmacologic inhibitors of the Hh pathway, and in turn hindered Gli1 dependent EMT [15]. Our results found that blocking TGF-b1 induced Shh signaling by both Shh-siRNA and chemical inhibitors cyclopamine and GDC-0449 reduced in vitro cell motility, invasiveness, and clonogenic capacity. Taken as a whole, our findings support the idea that Shh inhibitors could be therapeutically promising for regulating EMT and metastatic bladder cancer treatment. Shh signaling appears to serve an important role in normal bladder urothelial stem cell self-renewal together with Wnt signaling [28]. Furthermore, data from many human tumors including, glioblastoma, breast cancer, pancreatic cancer, and bladder cancer suggest that Hh signaling modulates cancer stem cells (CSCs). Active Hh signaling has also been identified in glioblastoma CSCs and pathway inhibition with cyclopamine or sirna directed against pathway components results in loss of tumorigenic potential [29]. The thinking is that CSCs are responsible for tumor initiation and progression [30] and share many characteristics with normal stem cells, including selfrenewal and differentiation. Several CSC markers have been identified including CD133. CD133 þ positive cells show differentiation into cells that phenotypically resembled the original tumor in culture, while CD133 negative cells failed to do so [31]. Furthermore, CD133 þ positive cells are able to form tumors in immunocompromised mice. Recent studies demonstrated that various solid tumors, such as colon, brain, and prostate are derived from the CD133 þ CSCs [32,33]. In our studies we found increased levels of CD133 þ cell expression at mrna and protein level in TGF-b1 treated xenograft