| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Imaging, Diagnosis, Prognosis |
Authors' Affiliations: 1 Department of Urology, Veterans Affairs Medical Center and University of California, San Francisco, San Francisco, California, 2 Department of Urology, Shimane University, Izumo, Japan, and 3 Department of Urology, Kagoshima University, Kagoshima, Japan
Requests for reprints: Rajvir Dahiya, Veteran Affairs Medical Center and University of California, San Francisco, 4150 Clement Street, San Francisco, CA 94121. Phone: 415-750-6964; Fax: 415-750-6639; E-mail: rdahiya{at}urol.ucsf.edu.
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Experimental Design: Samples (54 total) of bladder tumor and corresponding normal bladder mucosa were analyzed for the methylation and expression levels of six Wnt-antagonist genes (sFRP-1, sFRP-2, sFRP-4, and sFRP-5, Wif-1, and Dkk-3). To increase the sensitivity/specificity of bladder tumor detection, the methylation score (M score), a new method for multigene methylation analysis, was developed. The M score of each sample was calculated as the sum of the corresponding log hazard ratio coefficients derived from multivariate logistic regression analysis of the methylation status for each Wnt-antagonist gene. Receiver operator characteristic (ROC) curve analysis was used to determine the optimal sensitivity/specificity of the M score. Urine DNA from 24 matched patients with bladder tumor and 20 cancer-free volunteers was also used to investigate the methylation status of Wnt-antagonist genes.
Results: The methylation levels of Wnt-antagonists were significantly higher and mRNA levels were significantly lower in bladder tumor than in bladder mucosa. Each methylation level was inversely correlated with the corresponding mRNA level. In multivariate regression analysis, the methylation levels of sFRP-2 and Dkk-3 were significant independent predictors of bladder tumor (P < 0.05 and P < 0.01, respectively), whereas with sFRP-1, sFRP-5, and Wif-1 there was a trend towards significance as independent predictors. The M score of Wnt-antagonist genes was significantly higher in bladder tumor than in bladder mucosa (P < 0.05). Overall, the M score had a sensitivity of 77.2% and a specificity of 66.7% as a diagnostic biomarker (areas under the curve, 0.763). The M score could distinguish superficial from invasive bladder tumors with a sensitivity of 72.2% and a specificity of 61.1% as a staging biomarker (areas under the curve, 0.671). In patients with bladder tumor, 80.6% of the methylation-specific PCR results had identical methylation in samples of tumor- and urine-derived DNA. Most urine DNA in normal controls showed no aberrant methylation of the Wnt-antagonist genes.
Conclusions: Hypermethylation of Wnt-antagonist genes plays an important role in the pathogenesis of bladder tumor and can be detected using cellular DNA extracted from urine samples. This is the first report demonstrating that M score analysis of Wnt-antagonist genes could serve as an excellent epigenetic biomarker panel for bladder tumors.
80% of patients with primary bladder tumor have low-grade and superficial tumors, more than half of these patients relapse despite conservative transurethral resection with or without intravesical instillation of anticancer drugs. In addition, 30% of the patients may show tumor progression to a higher grade and/or stage of disease with fatal outcome (2, 3). Currently, cystoscopic examination is the most sensitive "gold" standard for bladder tumor detection. However, it is invasive, uncomfortable, and unpleasant to patients, with undesirable complications and significant cost. To make matters worse, in some cases of bladder tumor in situ, it may be difficult to distinguish bladder tumor from cystitis by cystoscopic examination. Conventional urine cytology has been the standard noninvasive method. However, urine cytology is of limited value because of operator dependency and low sensitivity (<50%; ref. 4). Therefore, noninvasive tests using urinary markers such as nuclear matrix protein 22 (NMP22) and bladder tumor antigen (BTA) have been developed and tested. These markers generally have a higher sensitivity but a much lower specificity than conventional cytology. In addition, inflammatory conditions may give false-positive results for BTAstat and NMP22 in >25% of cases (4). Therefore, the development of noninvasive and accurate diagnostic biomarkers of bladder tumor detection is imperative and crucial to improve bladder tumor prognosis.
Aberrant promoter hypermethylation of known or putative tumor suppressor genes occurs frequently during the pathogenesis of human cancers and has been found to be one of the primary mechanisms in the down-regulation of these genes (5, 6). The recent development of methylation-specific PCR (MSP) can detect these epigenetic changes and could be used for cancer detection (7). Considering that detection and quantification of promoter CpG methylation in body fluid DNA is feasible and noninvasive, combined MSP analyses of multiple genes in voided urine can provide a reliable way to improve cancer diagnosis (8).
In this regard, several antagonists of Wnt signaling have been identified (9) and can be divided into two functional classes, the secreted frizzled-related protein (sFRP) class and the Dickkopf (Dkk) class. The former class, which includes the sFRP gene family (sFRP-1 to sFRP-5), Wnt inhibitory factor-1 (Wif-1), and Cerberus, inhibits Wnt signaling by directly binding to Wnt molecules instead of Fz (9). The latter class, which comprises certain Dickkopf family proteins (Dkk-1 to Dkk-4), inhibits Wnt signaling by binding to the LRP5/LRP6 component of the Wnt receptor complex. Thus, the functional loss of Wnt antagonists can contribute to activation of the Wnt pathway and result in carcinogenesis through dysregulation of cell proliferation and differentiation. Recent publications from our laboratory and others have shown that impaired regulation by promoter hypermethylation of Wnt-antagonists such as sFRP, Wif-1, and Dkk-3 was found in bladder tumors (10–13). However, no comprehensive and combined analysis of Wnt-antagonist gene methylation has been reported for any types of cancer.
Therefore, we assessed (a) how the methylation status of all Wnt-antagonist genes changes during bladder carcinogenesis and (b) whether multigene methylation analysis of Wnt-antagonists could serve as a potential epigenetic biomarker panel for the diagnosis or staging of bladder tumor. In addition, we also investigated whether promoter hypermethylation of these genes could be detected in urine sediments from patients with bladder tumor.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Urine samples. Paired voided urine samples were collected before surgery from 24 patients with bladder tumor. The samples (50 mL fresh urine) were spun down by centrifugation at 1,500 rpm for 15 minutes, the supernatant was decanted, and genomic DNA from urine sediments was purified using QIAamp DNA Mini Kit (Qiagen, Valencia, CA). In addition, 20 normal voided urine sediments from age- and sex-matched healthy volunteers with no smoking history were included.
Nucleic acid extraction. Genomic DNA and total RNA was extracted from bladder tumor and matched bladder mucosa tissue samples using a Qiagen kit (Qiagen) after microdissection (15). The concentrations of DNA and RNA were determined with a spectrophotometer and their integrity was assessed by gel electrophoresis.
cDNA preparation and RT-PCR analysis. cDNA was prepared using total RNA (1 µg) and stored at –20°C until used. Primer pairs were designed to detect sFRP-1, sFRP-2, sFRP-4, and sFRP-5, Wif-1, and Dkk-3 mRNA expression. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal control. The primer sequences, annealing temperatures, and PCR cycles are shown in Table 1 . The PCR products were electrophoresed in 1.5% agarose gels. The expression levels of the genes were evaluated by ImageJ software, and the areas under the curve (AUC) were calculated and analyzed to determine the relative levels of target gene expression to GAPDH levels (arbitrary units). As reported previously (16, 17), for semiquantitative analysis of PCR products, a suitable number of PCR cycles for target genes and GAPDH were determined so that it was within the exponential phase. This method was used to determine the suitable number of cycles for all RT-PCRs in this study (Table 1).
|
Bisulfite DNA sequencing. Bisulfite-modified DNA was amplified using a pair of universal primers. Direct bisulfite DNA sequencing of the PCR products using either forward universal primer or reverse primer was done according to the manufacturer's instructions (Applied Biosystems, Foster City, CA).
Statistical analysis. Initially, the optimal cutoff values for the relative methylation level necessary to distinguish bladder tumor from normal bladder mucosa were determined for all investigated Wnt-antagonist genes using receiver operator characteristic (ROC) curve analysis. Next, using a previously reported analytic technique, we calculated the M score for each sample, defined as the sum of the corresponding log hazard ratio (HR) coefficients for each gene, which were derived from multivariate logistic regression analyses in the bladder tumor and corresponding bladder mucosa samples. The optimal sensitivity and specificity of the M score for diagnosis of bladder tumor and for staging was determined by ROC curve analysis using MedCalc software (MedCalc Software, Mariakerke, Belgium). A pairwise comparison was employed to test for significance using the AUC analysis. All data, except for ROC curve analysis, were analyzed by the StatView V statistical package (SAS Institute, Inc., Cary, NC). Statistical analysis was done using the Mann-Whitney test, Kruskal-Wallis test, 
2 test, Fisher's exact test, and multivariate logistic regression analysis. P < 0.05 was regarded as statistically significant.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
Expression of Wnt-antagonist gene transcripts in clinical tissue samples. Representative RT-PCR results of Wnt-antagonist gene expression are shown in Fig. 1B. As shown in Table 2A, the expression level of Wnt-antagonist mRNA transcripts was lower in bladder tumor than in bladder mucosa. The expression level of mRNA transcripts in all investigated Wnt-antagonist genes was inversely correlated with the relative methylation levels in these promoters (P < 0.05).
Evaluation of M score; multigene methylation analysis with Wnt-antagonists for distinguishing bladder tumor from normal bladder mucosa. The optimal cutoff point (MSP-negative versus MSP-positive) of each relative methylation level was established using ROC curve analysis in order to distinguish bladder tumor from normal bladder mucosa. However, using these optimal cutoff values, none of the Wnt-antagonist genes could be used as a single reliable biomarker for bladder tumor because of their limited sensitivity/specificity and AUC (Table 2B). On the other hand, mRNA expression levels of all Wnt-antagonist genes were significantly higher in the MSP-negative groups than in the corresponding MSP-positive groups in bladder tumor (Table 2C).
First, we confirmed that the relative methylation levels of Wnt-antagonist genes were not related to age and sex. Then we did a multivariate logistic analysis to determine which genes are independent predictors of pathology, comparing bladder tumor and normal bladder mucosa using multigene methylation analysis. As shown in Table 2D, multivariate logistic regression analysis revealed that the relative methylation levels of sFRP-2 (P < 0.05) and Dkk-3 (P < 0.01) were significant independent predictors of bladder tumors. Although, sFRP-1, sFRP-5, and Wif-1 were not significant, they showed trends as independent predictors of bladder tumor (P < 0.1). The individual gene HR for pathology (bladder tumor versus bladder mucosa) were different from one another. However, sFRP-4 was clearly not an independent predictor of bladder tumor in multivariate logistic analysis and was eliminated from the M score. For all patients, the M score, determined by the sum of the corresponding log HRs, was significantly higher in bladder tumors than in normal bladder mucosa (P < 0.05; Fig. 2A ). The M score had a sensitivity of 77.2% and a specificity of 66.7% (AUC, 0.763) using the optimal cutoff point (7.32) by ROC curve analysis (Fig. 2B). When used as a single variable, the methylation levels of Wnt-antagonist genes were not significantly different from one another using a pairwise comparison test for AUC, however, the M score was a better biomarker even compared with independent predictors of bladder tumor such as sFRP-2 (P = 0.001) or sFRP-5 (P < 0.05). As shown in Fig. 2B, M score was the best variable compared with other single variables.
|
Detection of methylation of Wnt-antagonist genes in urine sediments. Typical MSP results for Wnt-antagonist genes in matched bladder tumor and bladder mucosa samples and urine from patients with bladder tumor (patients 1 and 2) are shown in Fig. 1A. Representative MSP results of urine DNA from normal controls are also shown in Fig. 1A. Identical methylation of all Wnt-antagonist genes in bladder tumor and urine samples was found for 12 of 24 cases (50.0%). However, identical methylation status were found in bladder tumor and urine samples in 116 (80.6%) of a total 144 MSP results. These MSP results are summarized in Table 3A . The prevalence of methylation in urine samples from patients with bladder tumor (61.1% in total) and from normal controls (6.7% in total) is shown in Table 3B. All 24 bladder tumor cases (100%) had at least one CpG methylated promoter out of the six Wnt-antagonist genes. Bladder tumor samples that were methylation-negative were also negative in the matched urine DNA except for only one (sFRP-5 in patient 32; 1 of 30, 3.3%). Overall, the frequency of methylation in urine DNA from patients with bladder tumor was lower in superficial or low-grade bladder tumors when compared with invasive or high-grade bladder tumors (P < 0.0001 and P = 0.1, respectively; Table 3B).
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
In multivariate logistic regression analysis, the methylation status of most Wnt-antagonist genes can be independent predictors of bladder tumor. However, none of them were reliable biomarkers for bladder tumors as a single variable because of their limited sensitivity/specificity and AUC. The M score is a new tool, which we initially developed for prostate cancer samples to integrate the methylation status of multiple genes (22). In fact, Marsit et al. have recently reported a significant relationship between the odds risk of invasive bladder tumor as compared with superficial bladder tumor and increasing numbers of methylated genes (23). Similarly, Ray et al. employed multivariate Cox proportional hazards models for their multigene methylation analysis in the medulloblastoma, and used the sum of the log HR as a risk score for each patient (24). When an M score derived from Wnt-antagonist genes was applied to our bladder tumor samples, the sensitivity and specificity of bladder tumor detection were significantly improved. In a review of the literature, the sensitivity of urine cytology to detect bladder tumor seems to be <50% (4), whereas urinary bladder tumor markers approved by the Food and Drug Administration [BTAstat, BTAtrak, NMP22, FDP, ImmunoCyt, and fluorescence in situ hybridization (UroVysion, Abbott Laboratories Des Plaines, IL)], have a sensitivity of 50% to 70% with 50% to 70% specificity (4). In addition, bacterial cystitis or hematuria from urolithiasis or inflammation may give false-positive results for BTAstat and NMP22 in >25% of cases (4). However, Dulaimi et al. have reported that they observed no hypermethylation in tumor suppressor genes in urine DNA from patients with cystitis (7). Other candidate epigenetic biomarkers for detecting bladder tumor such as methylation of DAPK, RARß, and RASSF1A, etc., also showed lower sensitivity (<70%; refs. 7, 8) compared with the M score used in this study. Therefore, the M score derived from Wnt-antagonist genes has a higher sensitivity and specificity compared with other bladder tumor markers.
Conventional MSP analysis is sensitive in that it can detect 0.1% of methylated cancer cell DNA in a heterogeneous cell population (25). In addition, it has been reported in other studies that bladder tumor cells and renal and prostate cancer cells can be detected by methylation analysis of voided urine (26, 27). Therefore, we investigated whether cancer cells could be detected using conventional MSP analysis of Wnt-antagonist genes in the voided urine of patients with bladder tumor. Based on the high sensitivity, specificity, and utility of the Wnt-antagonist gene M score to detect bladder tumor, we hypothesized that MSP analysis of voided urine from patients with bladder tumor could provide highly sensitive and specific epigenetic information for bladder tumor detection. Our results from urine samples showed a high percentage of identical methylation with tumor-tissue DNA. Conversely, >90% of the urine DNA from normal controls had no aberrant methylation. In all patients with bladder tumor, at least one of the six Wnt-antagonist genes was hypermethylated. These findings clearly suggested that (a) methylation detection of the Wnt-antagonist genes was feasible and reliable and (b) the urine M score of the Wnt-antagonist genes could be used as an excellent noninvasive diagnostic biomarker for bladder tumor. In fact, methylation of the sFRP1, sFRP2, sFRP4, and sFRP5 genes in fecal DNA isolated from stool samples has been used to screen for colorectal cancer (28).
In the present study, the M score of Wnt-antagonist genes in bladder tissues was significantly higher in invasive bladder tumors than in superficial bladder tumors. Likewise, methylation of urine DNA from patients with bladder tumors was lower in superficial or low-grade bladder tumors when compared with invasive or high-grade bladder tumors. These findings indicate a positive association between methylation of the Wnt-antagonist genes and invasive disease. In fact, methylation of sFRP genes in bladder tumor seemed to be related to poor prognosis and higher tumor stage and grade (10, 22). Multiple Wnt-antagonist genes may gradually become methylated during the process of bladder tumor development and progression. In turn, the M score of the Wnt-antagonist genes might reflect the presence of bladder tumor that progresses to invasive disease requiring future aggressive treatment. Although these results could be extended to larger clinical studies, the detection of methylated Wnt-antagonist genes in urine provides a new way to detect bladder tumor, especially in high-risk individuals. We believe that an optimal hypermethylation panel of Wnt-antagonist genes can contribute significantly to early detection of bladder tumor and also predict bladder tumor aggressiveness.
In conclusion, hypermethylation of the Wnt-antagonist genes plays an important role in the pathogenesis of bladder tumor and can be readily detected in the voided urine of patients with bladder tumor. We have shown that a novel M score analysis using methylated Wnt-antagonist genes as molecular markers can accurately detect bladder tumors. To the best of our knowledge, this is the first report demonstrating M score analysis of Wnt-antagonist genes as a valuable new detection tool that can serve as an excellent biomarker for bladder tumor, including invasive disease.
|
Footnotes |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 11/11/05; revised 12/17/05; accepted 1/19/06.
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionReferences |
|---|
-catenin gene through epigenetic and genetic pathways in human prostate cancer. Cancer Res 2005;65:2130–8.This article has been cited by other articles:
|
|
 |
|
  J. Veeck, C. Geisler, E. Noetzel, S. Alkaya, A. Hartmann, R. Knuchel, and E. Dahl Epigenetic inactivation of the secreted frizzled-related protein-5 (SFRP5) gene in human breast cancer is associated with unfavorable prognosis Carcinogenesis, May 1, 2008; 29(5): 991 - 998. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Yue, Q. Sun, S. Dacic, R. J. Landreneau, J. M. Siegfried, J. Yu, and L. Zhang Downregulation of Dkk3 activates {beta}-catenin/TCF-4 signaling in lung cancer Carcinogenesis, January 1, 2008; 29(1): 84 - 92. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Yu, T. Zhu, Z. Wang, H. Zhang, Z. Qian, H. Xu, B. Gao, W. Wang, L. Gu, J. Meng, et al. A Novel Set of DNA Methylation Markers in Urine Sediments for Sensitive/Specific Detection of Bladder Cancer Clin. Cancer Res., December 15, 2007; 13(24): 7296 - 7304. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Sato, H. Suzuki, M. Toyota, M. Nojima, R. Maruyama, S. Sasaki, H. Takagi, Y. Sogabe, Y. Sasaki, M. Idogawa, et al. Frequent epigenetic inactivation of DICKKOPF family genes in human gastrointestinal tumors Carcinogenesis, December 1, 2007; 28(12): 2459 - 2466. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Suzuki, H. Shigematsu, T. Nakajima, R. Kubo, S. Motohashi, Y. Sekine, K. Shibuya, T. Iizasa, K. Hiroshima, Y. Nakatani, et al. Synchronous Alterations of Wnt and Epidermal Growth Factor Receptor Signaling Pathways through Aberrant Methylation and Mutation in Non Small Cell Lung Cancer Clin. Cancer Res., October 15, 2007; 13(20): 6087 - 6092. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Bruins, M. J. Jonker, O. Bruning, J. L.A. Pennings, M. M. Schaap, E. M. Hoogervorst, H. van Steeg, T. M. Breit, and A. de Vries Delayed expression of apoptotic and cell-cycle control genes in carcinogen-exposed bladders of mice lacking p53.S389 phosphorylation Carcinogenesis, August 1, 2007; 28(8): 1814 - 1823. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Q. Feng, S. E. Hawes, J. E. Stern, A. Dem, P. S. Sow, B. Dembele, P. Toure, P. Sova, P. W. Laird, and N. B. Kiviat Promoter Hypermethylation of Tumor Suppressor Genes in Urine from Patients with Cervical Neoplasia Cancer Epidemiol. Biomarkers Prev., June 1, 2007; 16(6): 1178 - 1184. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Urakami, H. Shiina, H. Enokida, H. Hirata, K. Kawamoto, T. Kawakami, N. Kikuno, Y. Tanaka, S. Majid, M. Nakagawa, et al. Wnt Antagonist Family Genes as Biomarkers for Diagnosis, Staging, and Prognosis of Renal Cell Carcinoma Using Tumor and Serum DNA Clin. Cancer Res., December 1, 2006; 12(23): 6989 - 6997. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
