Vol. 181, No. 4, Supplement, Monday, April 27, 2009308 THE JOURNAL OF UROLOGY®

expression level of claudin-7 in transitional cell carcinoma (TCC) of the urinary bladder, its relationship with conventional clinicopathological manifestations, and its effect on the progression of TCC.

METHODS: The study included 68 specimens of TCC of the bladder (35 with superficial, Ta-T1, and 33 with muscle invasive, T2-T4) and 26 of normal urothelium. Total RNA was extracted and then 1μg of it was reverse transcribed using a First-Standard cDNA. To examine the expression levels of the target gene (claudin-7) and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH), real-time quantitative RT-PCR was performed on the LightCycler® System using SYBR Green I dye. Immunohistochemical expression was examined using confocal laser scanning light-microscopy to validate the RT-PCR data. The difference between claudin-7 expression and several clinicopathological variables, such as stage and grade, were assessed using Student’s t-test with statistical significance indicated at P<0.05.

RESULTS: Claudin-7 was overexpressed in all normal urothelium samples and was significantly higher than that in TCC samples (P < 0.0001). Superficial (Ta-1) tumours had significantly higher claudin-7 expression than muscle-invasive (> or = pT2) tumours (P = 0.012). There was no significant difference between patients with G1-2 tumours and those with G3 tumours (P = 0.19). There was no significant difference between patients with recurrent superficial TCC and and those with no recurrence (P = 0.61). Immunohistochemistry showed claudin-7 overexpression in normal urothelium than TCC samples, and in superficial than invasive TCC.

CONCLUSIONS: Our data indicate that loss of the tight junction protein claudin-7 correlates with carcinogenesis and progression of TCC of the urinary bladder.

Source of Funding: None

863IMMUNOEXPRESSION OF ZEB1 AND SIP1 IN HUMAN BLADDER CANCER.

Raj P Pal, Andrew Ruddick, Hassan A.R Qazi, Nick J Mayer, Emre Sayan, Leicester, United Kingdom; Tamer Yagci, Ankara, Turkey; Marina Kriajevska, Eugene Tulchinsky, Thomas R.L Griffiths*, John K Mellon, Leicester, United Kingdom

INTRODUCTION AND OBJECTIVE: Epithelial mesenchymal transition (EMT) occurs in cancer cells and contributes towards a metastatic phenotype. Down regulation of E-cadherin is considered a hallmark of EMT. E-cadherin function can be inhibited by gene mutation, promoter hypermethylation or transcriptional repression. Currently described transcriptional repressors of E-cadherin include SIP1 and ZEB1. In bladder cancer, the role of these transcriptional repressors has not been fully evaluated. Our aim was to evaluate immunoexpression of E-cadherin, SIP1 and ZEB1 in a panel of human bladder cancer specimens.

METHODS: Paraffin sections from 134 patients with transitional cell carcinoma of the bladder were retrieved from our Pathology archive and re-graded (WHO 1973) and staged (UICC 2002). 77 were muscle-

invasive (12 grade 2, 65 grade 3) and 57 non-muscle invasive (41 pT1 and 16 pTa, all grade 3). Immunostaining was performed using an anti-E-cadherin mouse monoclonal antibody (Zymed, 4A2C7), a novel highly specific anti-SIP1 antibody and an anti-ZEB1 rabbit polyclonal antibody (SC Biotech, sc-25388).

RESULTS: Absent E-cadherin staining was apparent in 6/134 (4%), aberrant in 81/134 (60%) and positive in 47/134 (35%) specimens. Aberrant or absent expression of E-cadherin significantly correlated with higher tumour stage (>T2 v T1 and Ta; p=0.014; Chi-square test). 10/134 (7.5%) tumours displayed strong positive nuclear expression for ZEB1. 5/134 displayed diffuse ZEB1 positivity and 5/134 focal positivity. The extent of ZEB1 positivity inversely correlated with E-cadherin staining (p<0.0001, r=0.369, Spearman correlation). SIP1 staining, which was feasible in 128 samples, was predominantly nuclear but also cytoplasmic. 31/128 (24.2%) samples displayed strong or very strong SIP1 staining, which also inversely correlated with E-cadherin negativity (p=0.03, Fisher’s exact test). Increased intensity of SIP1 correlated with extent of ZEB1 staining (p=0.027, r=0.196).

CONCLUSIONS: This is the first study to demonstrate co-immunoexpression of SIP1, ZEB1 and E-cadherin. ZEB1 and SIP1 expression both correlate with down regulation of E-cadherin in human bladder cancer.

Source of Funding: British Urological Foundation

864THE FGFR3 MUTATION IS RELATED TO FAVORABLE PT1 BLADDER CANCER

Bas W Van Rhijn*, Bharati Bapat, Theo H Van der Kwast, Liyang Liu, Neil E Fleshner, Rati Vajpeyi, Toronto, ONCanada; Madelon N Van der Aa, Chris H Bangma, Ellen C Zwarthoff, Rotterdam, Netherlands; Michael A s Jewett, Alexandre R Zlotta, Toronto, ON, Canada

INTRODUCTION AND OBJECTIVE: Fibroblast growth factor receptor 3 (FGFR3) mutations have been recently reported at a very high frequency in pTa bladder cancer (BC) whereas these mutations are rare in high grade muscle-invasive BC. pT1-BC comprises a heterogeneous group of tumors for which different management options are advocated depending on the risk of progression to muscle-invasive disease. We have determined the frequency of FGFR3 mutations in a group of primary pT1-BC and correlated the FGFR3 mutation to various histo-pathological variables.

METHODS: We included 132 patients from two academic centers (N=60 in Rotterdam / 72 in Toronto) with primary (first diagnosis) pT1-BC. Mean age was 68.7 years (SD: 9.9 y). An experienced uro-pathologist reviewed the slides for grade (1973 and 2004 classification systems) and determined if the tumor was micro-invasive (<0.5 mm; pT1m) or extensive invasive (multiple spots with invasion and/or >0.5 mm; pT1e). The FGFR3 mutation status was examined by SNaPshot analysis and correlated to pathological parameters.

RESULTS: FGFR3 mutations were detected in 37/132 (28%) pT1-BC. The most frequent mutation in the FGFR3 gene (S249C) was observed 21 times whereas other FGFR3 mutations, R248C, G372C, Y375C, G382R were observed 5, 2, 5 and 1 time(s), respectively. The table shows that presence of a FGFR3 mutation was highly correlated with favorable disease characteristics in pT1-BC.

FGFR3 mutations in pT1 Bladder CancerFGFR3 mutant FGFR3wild type P-value(chi-square)

Grade 1973 G2 25 31 <0.0001G3 12 64

Grade 2004 Low-grade 15 11 <0.0001High-grade 22 84

Sub-stage pT1m 18 22 =0.004pT1e 19 73

Total: 37 95

CONCLUSIONS: The FGFR3 mutation selectively identifies pT1-BC patients with favorable disease characteristics. Further study may confirm that this molecular marker is able to select patients who will

Vol. 181, No. 4, Supplement, Monday, April 27, 2009 THE JOURNAL OF UROLOGY® 309

benefit from a conservative approach to their disease.

Source of Funding: Hospital / Dutch Cancer Society

865CPG HYPERMETHYLATION OF COL1A2 CONTRIBUTES TO PROLIFERATION AND MIGRATION ACTIVITY OF BLADDER CANCER

Katsuhisa Mori*, Hideki Enokida, Ichiro Kagara, Kazumori Kawakami, Kenryu Nishiyama, Kagoshima, Japan; Kazuya Kawahara, Aira-cho, Japan; Naohiko Seki, Chiba, Japan; Masayuki Nakagawa, Kagoshima, Japan

INTRODUCTION AND OBJECTIVE: Gene silencing due to aberrant DNA methylation plays an important role in carcinogenesis. For genome-wide screening for methylated genes in bladder cancer (BC), we performed cDNA microarray analysis of the BC cell line (BOY) treated with demethylating agent (5-Aza-dC). Collagen type 1 alpha 2 (COL1A2) gene was identified as the most up-regulated one of the 30,144 genes screened. We hypothesize that inactivation of the COL1A2 gene through CpG methylation contributes to proliferation and migration activity of human BC by changing extra-cellular matrix (ECM).

METHODS: To test this hypothesis, we extracted DNA and RNA from 67 BC samples and 10 normal bladder epitheliums (NBEs) from cystectomy and transurethral resection. The CpG hypermethylation of COL1A2 promoter was analyzed by quantitative methylation-specific PCR (QMSP) and confirmed by bisulfite-modified DNA sequencing. Real-time RT-PCR was carried out to evaluate COL1A2 mRNA expression. We established a stable COL1A2 transfectant for gain of function studies, for which cellular proliferation and migration were examined by MTT assay and wound healing assay, respectively. The protein secretion of type 1 collagen in culture medium was detected by immunoblot.

RESULTS: Bisulfite DNA sequencing demonstrated that the promoter hypermethylation of COL1A2 was a frequent event in clinical BCs. The methylation index was significantly higher in the 67 BCs than in the10 NBEs (42.3 ± 28.9 vs. 0.07 ± 0.01, p = 0.0011). Conversely, COL1A2 mRNA expression was significantly lower in the BCs than in the NBEs (15.8 ± 3.8 vs. 32.0 ± 22.0, p = 0.0052). In vitro, after 5-aza-dC treatment, the COL1A2 mRNA expression was markedly increased in BOY. Our cell proliferation assays consistently demonstrated a significant growth inhibition in the COL1A2 transfectant compared with control and wild-type BOY cells (on day 5; 0.791 ± 0.020, 1.056 ± 0.034, and 1.107 ± 0.027, respectively, p < 0.0001). Wound healing assays also showed a significant wound healing inhibition in the COL1A2 transfectant compared to the counterparts (% of wound closure; 75.5 ± 0.2, 94.3 ± 2.9, and 96.9 ± 3.1, respectively, p = 0.0016). Immunoblot demonstrated that type 1 collagen protein markedly increased in culture medium of the COL1A2 transfectant.

CONCLUSIONS: Our data suggest that inactivation of the COL1A2 gene through CpG methylation may change an ECM binding on epithelial collagen and contributes to proliferation and migration activity of human bladder cancer.

Source of Funding: None

866A GAIN OF 5P15.33 BY ARRAY CGH AND FISH MAY BECOME A NOVEL MARKER FOR PREDICTING DISEASE PROGRESSION IN BLADDER CANCER

Yoshiaki Yamamoto*, Satoshi Eguchi, Yasuyo Chochi, Kohsuke Sasaki, Hideyasu Matsuyama, Ube, Japan

INTRODUCTION AND OBJECTIVE: We identified a gain of 5p15.33, including TPPP (tubulin polymerization promoting protein) and ZDHHC11 (Zinc finger DHHC domain-containing protein 11) genes, that distinguished low-grade from high-grade bladder cancer in array comparative genomic hybridization (CGH) analysis. To establish the clinical usefulness of 5p15.33 status, we examined the relationship of 5p15.33 gain to biological characteristics in 100 urothelial carcinomas

of the urinary bladder. METHODS: We screened 18 bladder cancers for genome-wide

copy number changes using array CGH onto an array consisting of 4030 bacterial artificial chromosome clones. The number of 5p15.33 spots was measured by fluorescence in situ hybridization (FISH) in 100 urothelial carcinomas, and compared with clinicopathological data.

RESULTS: Gain of 5p15.33 was detected by array CGH analysis in 5/9 (55.6%) high-grade bladder cancers and 0/9 (0%) low-grade bladder cancers. In FISH analysis, gain of 5p15.33 was significantly correlated with higher histological grade (P<0.0001), advanced pathological stage (P=0.0284), and multiple tumors (P=0.0019). Tumors with a gain of 5p15.33 had a significantly higher progression-free survival rate than those without (P=0.0006, log-rank test). Multivariate analysis revealed that gain of 5p15.33 was a predictor for disease progression in bladder cancer (hazard ratio: 1.887, 95% CI: 1.215-2.968, P=0.0050).

CONCLUSIONS: These data suggest that gain of 5p15.33 (TPPP and ZDHHC11) is associated with progression in bladder cancer, and is a potential biomarker for estimating biological characteristics.

Source of Funding: None

867ABSENCE OF CD44 EXPRESSION IS AN INDEPENDENT PREDICTOR OF POOR OUTCOME FOR PATIENTS WITH UROTHELIAL BLADDER CANCER

Tobias Klatte*, David B Seligson, Jian Yu Rao, Hong Yu, Michela de Martino, Kelly Kawaoka, Isla P Garraway, Steven G Wong, Arie S Belldegrun, Allan J Pantuck, Los Angeles, CA

INTRODUCTION AND OBJECTIVE: CD44 is a cell surface protein involved in cell adhesion, migration and lymphocyte activation by interacting with hyaluronic acid, collagens, matrix metalloproteinases and other proteins that are components of the extracellular matrix. Previous studies suggest that CD44 is a putative cancer stem cell marker with functional importance for tumorigenesis, and absence of its expression has been identified as a poor prognostic feature in several cancers. Here we report on CD44 expression in bladder cancer and its association with survival.

METHODS: A tissue microarray was constructed containing 411 primary urothelial bladder cancers, each in triplicate. Immunohistochemical staining was performed with a commercially available antibody. The percentage of tumor cells staining positive for CD44 was evaluated by two independent pathologists in a blinded fashion. Associations of staining with tumor stage, grade, and survival were assessed with t-tests, Kaplan-Meier survival estimates and Cox proportional hazards models.

RESULTS: CD44 was expressed in 81% of the tumors. CD44 expression was higher in non-invasive (Ta) vs. invasive (T1-T4) tumors (p<0.001) and decreased with increasing grade (p<0.001). In patients undergoing TUR-B, absence of CD44 expression was associated with a 2.3 fold increased risk of recurrence (p=0.005). Median time to recurrence for tumors with vs. without CD44 expression was 23 vs. 9 months, respectively (p=0.0032). In a multivariate Cox model, absence of CD44 expression (HR 2.33, p=0.006) and increasing T stage (HR 1.32, p=0.047) were the only significant, adverse prognostic factors for tumor recurrence. Likewise, in patients undergoing cystectomy, median overall survival for CD44 non-expressors vs. expressors was 30 months vs. 75 months, respectively (p=0.0027), corresponding to a 1.77 fold increased risk of death if CD44 expression was absent. Again, absent CD44 expression was retained as an independent prognostic factor of overall survival (HR 1.55, p=0.042).

CONCLUSIONS: CD44 is differentially expressed in non-invasive vs. invasive and low grade vs. high grade urothelial bladder cancer. Absence of CD44 expression is an independent adverse predictor for recurrence and overall survival. Routine evaluation of CD44 expression may allow the identification of high-risk patients who require more intensive surveillance or aggressive therapy. Targeting of CD44 with monoclonal antibodies may provide new avenues for bladder cancer imaging and treatment.

Source of Funding: None