Early Detection and Diagnosis
Quantitative analysis of survivin mRNA expression in urine and tumor tissue of bladder cancer patients and its potential relevance for disease detection and prognosis
Article first published online: 10 MAR 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 116, Issue 1, pages 100–104, 10 August 2005
How to Cite
Weikert, S., Christoph, F., Schrader, M., Krause, H., Miller, K. and Müller, M. (2005), Quantitative analysis of survivin mRNA expression in urine and tumor tissue of bladder cancer patients and its potential relevance for disease detection and prognosis. Int. J. Cancer, 116: 100–104. doi: 10.1002/ijc.21000
- Issue published online: 12 MAY 2005
- Article first published online: 10 MAR 2005
- Manuscript Accepted: 22 DEC 2004
- Manuscript Received: 5 OCT 2004
- Charité-Universitaetsmedizin Berlin, Campus Benjamin Franklin
- bladder cancer;
Suppression of apoptosis may favor the onset and progression of cancer. Survivin is an inhibitor of apoptosis that has been suggested as a novel diagnostic/prognostic marker of bladder cancer. In this study, survivin mRNA expression was measured by a sensitive real-time PCR assay in tumor tissue and urine from bladder cancer patients and assessed for its potential diagnostic and prognostic relevance. Specimens from 53 patients with bladder transitional cell carcinoma (TCC) were analyzed, the controls being normal urothelial tissues (n = 14) and urine from benign disease patients (n = 22) and healthy individuals (n = 14). Survivin transcripts were commonly detected in tumor tissues, but not in normal urothelium, and increasing mRNA levels correlate with progressing pathologic stage (p = 0.001) and grade categories (p < 0.004). Higher levels of expression were associated with a reduced time to recurrence in noninvasive TCCs (p = 0.027, log-rank test) and a trend toward shorter disease-free survival in muscle-invasive tumors (p = 0.067). Urinary survivin analysis detects TCC with higher sensitivity (68.6%) and equal specificity (100%) when compared with cytology (31.4% and 97.1%). Our results indicate that tissue levels of survivin mRNA predict disease-free survival in noninvasive TCC and may have a role in bladder cancer progression. When analyzed by RT-PCR in urine, survivin is a highly specific biomarker for TCC detection. © 2005 Wiley-Liss, Inc.
Deregulated expression of apoptosis inhibitors may favor the onset and progression of cancer by abnormally prolonging cell viability with accumulation of transforming mutations.1 Moreover, elevated expression of antiapoptotic molecules promotes aggressive tumor behavior and resistance to therapy.2 Survivin is an inhibitor of apoptosis (IAP) protein characterized by a highly conserved baculovirus IAP repeat (BIR) domain. In contrast to other members of the IAP protein family, which are widely expressed in normal tissues,3 survivin is selectively overexpressed in common human malignancies but barely detectable in normal differentiated adult tissues.4 In cancer cells, survivin is expressed in the G2/M phase of the cell cycle and counteracts apoptosis induction during mitosis by interfering with the function of caspases.5, 6 Its overexpression has been shown to be associated with an increased malignant potential and an unfavorable outcome in various malignancies. In bladder cancer, survivin expression has been demonstrated by immunohistochemistry and may correlate with accelerated recurrences.7, 8
The discovery of bladder cancer biomarkers may facilitate early and accurate detection as well as more effective monitoring of the disease. Furthermore, these markers may provide prognostic information and identify patients with an increased risk of disease recurrence and progression. Though numerous potential tumor markers have been identified for bladder TCC, their significance for the diagnosis and prognosis of the disease is still controversial.9, 10
Given the sharp differential expression of survivin in cancer vs. normal tissue, its detection appears to be a suitable tool for cancer diagnosis. Accordingly, testing for survivin in urine has shown promise in detecting bladder cancer.11, 12 In addition, survivin expression has been demonstrated by immunohistochemistry in bladder TCC and may correlate with accelerated recurrences.7, 8
Survivin gene expression has not yet been quantified in urine, and data on its tissue expression are still limited. The objectives of this study were to investigate quantitative survivin expression in tumor tissue and urine from TCC patients using reverse transcription-polymerase chain reaction (RT-PCR) assays and to assess its potential diagnostic and prognostic relevance.
Material and methods
Participants and sample collection
Over a period of 2 years (August 1999 to August 2001), voided urine samples and/or tumor tissue specimens were prospectively collected from a large series of consecutive patients with TCC of the bladder, the controls being patients with benign genitourinary disease and healthy individuals. Recruitment was primarily based on presenting symptoms and clinical evaluation and also depended on subsequent histopathologic verification in subjects with bladder cancer. Eighty-nine subjects were randomly selected for this study. They comprised 53 TCC patients, 22 patients with benign genitourinary diseases and 14 healthy individuals with a mean age of 68, 66 and 60 years, respectively.
TCC patients were included if tumor tissue specimens had been successfully collected by transurethral resection. All those with muscle-invasive tumors who had not received curative treatment, i.e., radical cystectomy, were excluded. Histopathologic classification of bladder cancer was based on the World Health Organization and 1997 TNM classification guidelines (International Union Against Cancer, 1997). TCCs were staged as pTa (n = 22), pT1 (n = 13), pT2 and pT3 (n = 18), and graded as G1 (n = 2), G2 (n = 20) and G3 (n = 31). Concomitant carcinoma in situ was diagnosed in 3 patients. Of the 35 patients with superficial TCC, 18 (8 with pTa tumors; 10 with pT1 tumors) received adjuvant intravesical instillation therapy with BCG or mitomycin C. All patients with muscle-invasive TCC (≥ pT2) underwent subsequent radical cystectomy at our institution, and 5 of them received adjuvant chemotherapy. Individuals with benign genitourinary disease, including benign prostatic hyperplasia (n = 6), urolithiasis (n = 6) and urinary tract infection (n = 10), were evaluated for a variety of genitourinary symptoms typically seen in urologic practice, including obstructive/irritative voiding symptoms and hematuria. Healthy individuals were recruited according to the self-reported current health status and had to be over 40 years of age. Control subjects were chosen to best match the age and sex distribution of the bladder cancer group. The study was approved by the ethics committee of the Free University of Berlin, and all patients gave their written informed consent.
All tissue specimens were snap-frozen immediately after collection and stored at −80°C. Samples of normal bladder mucosa were collected from 14 other patients during open surgery for benign prostatic hyperplasia. Exfoliated cells were obtained from 50 ml of spontaneously voided urine collected before cystoscopy or any other instrumentation of the urinary tract. Urine samples were processed immediately after collection. After separating a small aliquot for urinalysis, the urine specimen was centrifuged (1,000g). The cell pellet was then washed and divided into 2 parts: 1 was used for conventional cytology, while the other was snap-frozen and stored at −80°C until RNA extraction.
Total RNA was extracted from frozen sections of the biopsy material using RNAzolB (WAK Chemie Medical, Bad Homburg, Germany) according to the manufacturer's instructions. A representative section was HE-stained to confirm the preponderance of malignant cells. The concentration and quality of RNA (28S/18S ratio) were determined using the Agilent Bioanalyzer 2100 (Agilent Technologies, Waldbronn, Germany). Total RNA was isolated from exfoliated urinary cells using the high pure RNA isolation kit (Roche Diagnostics, Mannheim, Germany).
Real-time fluorescence RT-PCR was performed using the LightCycler instrument (Roche Molecular Systems, Alameda, CA) according to the manufacturer's instructions. The LightCycler allows online monitoring of rapid-cycle PCR using simultaneous fluorescence analysis as described in detail elsewhere.13 In a one-step RT-PCR reaction, 250 ng of total RNA were subjected to cDNA synthesis and subsequently amplified during 40 PCR cycles (0.5 sec at 95°C; 10 sec at 60°C; 10 sec at 72°C). The respective primers for survivin (Genbank accession number AF077350) were 5′-AAAGAGCCAAGAACAAAATTGC-3′ (sense) and 5′-GAGAGAGAAGCAGCCACTGTTAC-3′ (antisense), generating a 338 base pair product. Hybridization probes used are specific for an internal segment of the amplified survivin fragment (FL probe TGCTCTTGTTTTGTCTTGAAAGTGGC-FL and LC Red640-probe CCAGAGGTGCTTCTGCCTGTGC-PH). Only regularly spliced survivin and survivin-2B splice variants were detected using this PCR assay. In a separate one-step RT-PCR, mRNA encoding porphobilinogen deaminase (PBGD) was processed in the same manner for each sample (Fig. 1). The primers, probes and standard for PBGD analysis were supplied with the LightCycler h-PBGD Housekeeping Gene Set (Roche Diagnostics). Detection limits of the PCR assays were 103 copies for survivin and 102 for PBGD.
To carry out a positive control and establish an external standard curve, all measurements included the determination of standards for both survivin and PBGD. The survivin cRNA standard was generated using a cloned survivin cDNA fragment whose sequence identity with Genbank entry AF077350 was confirmed on an ABI 377 Prism automatic sequencer. Survivin cRNA was synthesized during in vitro runoff transcription on the BamHI (Fermentas, Vilnius, Lithuania) linearized plasmid pGEM7Zf(+) using the RiboMax Large Scale RNA Production System T7 (Promega, Madison, WI) according to the manufacturer's instructions. cRNA was stabilized by adding bacteriophage MS2 RNA (Roche Diagnostics). The copy number of cRNA molecules was determined according to the LightCycler manual on relative quantification (Roche Diagnostics). Serial dilutions of the survivin RNA standard were used to plot the standard curve for accurate determination of the survivin copy number in analyzed samples.
The LightCycler Software (Roche Diagnostics) was used to analyze PCR kinetics and calculate quantitative data. In each sample, copy numbers of survivin mRNA were divided by those of PBGD mRNA to normalize survivin mRNA expression for sample-to-sample differences in RNA quantity, RNA quality and RT efficiency.14 Samples lacking housekeeping gene expression were excluded from analysis.
Voided urine cytology
Cytologic examination was performed using the Cytospin method (Shandon, ThermoElectron, Waltham, MA) and standard Pap staining. At least 2 slides were prepared from each patient's urine sample. All slides were examined by an experienced cytologist blinded to all patient characteristics.
To compare expression data of all histologic groups, the Kruskal-Wallis test for nonparametric analysis of variance (ANOVA) and the Jonckheere-Terpstra test for trend were performed using the SPSS software, version 10.5 (Chicago, IL). Significant differences between 2 groups were established by the Mann-Whitney U-test.
Urine samples were classified as either positive or negative for survivin. The chi-square test was used to relate these dichotomous variables to pathologic stage and grade categories. Sensitivity differences between urinary survivin and cytology were established by the McNemar test. All p-values are 2-sided.
Correlations were investigated between clinical follow-up parameters and both urinary and tissue levels of survivin expression using log-rank test and Kaplan-Meier estimates. For this purpose, patients were grouped by high vs. low survivin expression using cut points at the median expression level.
Tissue expression of survivin mRNA
Quantitative survivin mRNA expression was measured in specimens from bladder cancer patients (n = 53) with varying stage and grade categories and in nonneoplastic bladder mucosa samples (n = 14). RT-PCR analysis revealed sufficient PBGD expression in all samples investigated. Representative data obtained in cancer specimens are shown in Figure 1. All 53 tissue samples from bladder cancer patients were characterized by survivin mRNA expression. Table I shows the descriptive statistics for normalized survivin expression in tumor tissues. The Kruskal-Wallis ANOVA showed a significant correlation between normalized survivin values and histologic stage categories (p = 0.002). In addition, a significant trend toward higher normalized survivin values was observed with advancing pathologic stage (p < 0.001, test for trend). Significant differences were observed between pTa and pT1 tumors (p < 0.001, Mann-Whitney U-test) as well as between pTa and muscle-invasive tumors (p < 0.001). In view of the small number of grade 1 tumors (n = 2), they were pooled together with grade 2 tumors (n = 20). Normalized survivin expression was significantly higher in grade 3 (n = 31) than in grade 1/2 tumors (p = 0.004; Table I). In contrast, survivin expression was not detected in any of the normal bladder mucosa specimens.
|Histopathology||Survivin mRNA expression|
|Tumor tissue||Voided urine|
Survivin mRNA expression in exfoliated urinary cells
Voided urine samples were available from 73 of the 89 subjects investigated in this study. Sufficient PBGD expression was found in 68 of these samples, so that 5 subjects (3 healthy individuals and 2 bladder cancer patients) had to be excluded. The study design did not allow us to obtain a second urine sample from the same patient. Twenty-four of the 35 patients with histopathologically proven bladder cancer tested positive for urine survivin mRNA expression in the quantitative RT-PCR analysis. The RT-PCR assay used was at the detection limit for both survivin and PBGD in many samples. Table I shows the descriptive statistics for urinary survivin levels. Although the median normalized levels were approximately 10 times higher in exfoliated urinary cells than in tumor tissue, no significant correlation was observed between urine and tissue levels. In addition, urine survivin expression did not relate to pathologic stage or grade categories. Remarkably, none of the urine samples from 11 healthy individuals and 22 patients with benign genitourinary disease showed detectable survivin mRNA expression despite the presence of PBGD expression. These preliminary data disclosed a sensitivity of 68.6% (95% CI = 50.7–83.2%) and a specificity of 100% (90.3–100%) for noninvasive bladder cancer detection, while voided urine cytology had a sensitivity of 31.4% (16.8–49.3%) and a specificity of 97.1% (85.1–99.9%). Table II provides a comparison of urine survivin analysis with cytology.
|TCC||35||24 (68.6%)4||11 (31.4%)4||0.001|
|Ta||16||9 (56.2%)||4 (25.0%)||0.125|
|T1||9||6 (66.7%)||3 (33.3%)||0.250|
|≥ T2||10||9 (90.0%)||4 (40.0%)||0.063|
|G 1/2||17||9 (52.9%)||3 (17.6%)||0.031|
|G3||18||15 (83.3%)||8 (44.4%)||0.039|
|Controls||33||0 (100%)5||1 (97.1%)5||1.0|
Correlation between survivin expression levels and patient outcome
The effect of tissue survivin levels on time to first recurrence was analyzed in patients with superficial bladder cancers (Ta and T1; n = 32). The median expression level in these patients was 41.1, and groups of high vs. low survivin expression were formed using a cut point of 40. A trend toward reduced recurrence-free survival was observed in patients with high tissue survivin levels (27.6 vs. 48.1 months), though this was not statistically significant (p = 0.066, log-rank test). This trend did not appear to correlate with intravesical instillation treatment, since 6 of 16 patients with low and 12 of 16 with high expression levels had received BCG or mitomycin C therapy. When patients with pTa tumors were analyzed as a separate group (n = 21), the mean time to first recurrence was 39.3 months (95% CI = 34.8–43.9 months) in those with a survivin level below the selected cutoff of 40 but only 29.0 months (13.4–34.6) in those with higher expression levels (p = 0.027). The Kaplan-Meier estimates of recurrence-free survival for this group of patients are given in Figure 2. The 3-year recurrence-free survival rates for these patients were 67% and 37%, respectively. Intravesical instillation treatment had been completed in 3 of 11 patients with high and 5 of 10 with low expression levels. Tumor grade had no apparent effect on recurrence-free survival (p = 0.220), probably due to the limited number of high-grade pTa tumors.
The mean time to metastasis or relapse was 24.6 months in patients with muscle-invasive bladder cancers who underwent radical cystectomy at first diagnosis. When these patients were grouped using a cut point near the median survivin expression, disease-free survival appeared to be shorter in patients with higher (> 70) than in those with lower tissue levels of survivin expression (15.6 vs. 30.6 months). However, this was not statistically significant (p = 0.21), and differences in disease-free survival tended to be more pronounced (13.6 vs. 32.3 months; p = 0.067) using a cut point of 90. Patients grouped by a cut point of 90 had a nearly equal distribution of adjuvant chemotherapy (2/7 vs. 3/11). When all bladder cancer patients were analyzed, tissue survivin levels appeared to correlate with overall survival. However, the tissue survivin levels were not predictive of overall survival when adjusting for the prognostic effect of tumor stage. Urinary survivin expression did not correlate with patient outcome, whether analyzed as a dichotomous variable (positive vs. negative) or as a quantitative variable.
Elevated expression of antiapoptotic molecules is a mechanism by which tumor cells escape apoptosis. IAP family proteins are important apoptosis regulators, and some have been implicated in the initiation and progression of cancer.3 Survivin is unique among IAP proteins for its sharp differential expression in cancer but not in normal tissues. In this study, we quantified survivin mRNA expression in tumor tissue of bladder cancer patients and in normal reference tissue. Our data indicate strong survivin mRNA expression in bladder cancer tissue but not in normal bladder urothelium. A high incidence of survivin expression in bladder cancer was previously determined by immunohistochemistry7, 8 and RT-PCR assays.15, 16 However, some reports found the expression to be dependent on the tumor grade with a low percentage of grade 1 tumors expressing survivin.8, 15 On the other hand, studies using quantitative or semiquantitative approaches to detect survivin demonstrated its expression even in low-grade tumors, albeit at significantly decreased levels.7, 16 Since the present study had a very limited number of grade 1 tumors, no conclusions can be drawn on the extent of survivin expression in this tumor category. Our quantitative data show that survivin expression increases with the stage and grade of bladder TCC. We are first to demonstrate a higher expression level in invasive superficial TCC (pT1) than in the noninvasive type (pTa). These findings suggest that survivin may influence the malignant potential of transformed transitional cells and may thus play a role in the progression of bladder cancer. This is in accordance with the hypothesis that survivin participates in tumor progression rather than oncogenic transformation.17 An elevated survivin expression has been reported in invasive and poorly differentiated bladder TCC.16 Whether the survivin expression relates to the grade and stage of superficial bladder cancers is still debatable.7, 8 However, the quantitative RT-PCR approach used in our study may be sensitive enough to detect differences in expression levels among grade and stage categories. These differences are likely to translate to the protein level, since a close correlation between survivin mRNA and protein has been demonstrated in other tumors,18, 19 and survivin protein expression has been shown to be transcriptionally regulated, at least in part.20, 21
A correlation of survivin expression with aggressive tumor behavior and poor outcome has been demonstrated for a number of malignancies, including colorectal carcinoma,22 neuroblastoma19 and breast cancer.23 According to our data, survivin expression in bladder cancer is correlated with clinicopathologic features of a poor prognosis. Our findings indicate that tissue survivin expression predicts time to recurrence in patients with noninvasive bladder cancer. As a quantitative marker, survivin may thus provide prognostic information, which may help to simplify the current intensive follow-up protocols in these patients. Moreover, the finding of high survivin expression in noninvasive bladder tumors may, if validated in larger studies, provide a rationale for a more aggressive approach regarding intravesical instillation treatment. Our data are in accordance with an earlier study showing the prognostic relevance of survivin expression in low-grade tumors,8 which are generally noninvasive. However, reports on the prognostic significance of survivin in bladder cancer have been controversial.7, 15, 16 Our preliminary data indicate that survivin analysis in tumor tissue may even identify patients at an increased risk for tumor progression and metastatic disease following radical cystectomy. This subset of patients may benefit most from adjuvant treatment regimens. Interestingly, survivin is one of the few genes transcriptionally repressed by p53,24, 25 and it is also positioned in a p53-dependent pathway that controls mitotic progression.26 Alterations in the p53 pathway that were shown to correlate with bladder cancer progression and clinical outcome27, 28 may thus influence expression of the survivin gene.
Though urinary survivin levels did not correlate with clinicopathologic characteristics or follow-up parameters, survivin mRNA detection by RT-PCR was a highly specific marker for bladder cancer. This is in accordance with previously published reports on urinary survivin analysis using other detection systems.11, 12 Consistent with our results, survivin mRNA analysis in bladder washings proved suitable for detecting tumor recurrence in patients with superficial bladder cancer.29 Moreover, promising results have been obtained when using a nested PCR assay for survivin mRNA detection in urine specimens of bladder carcinoma patients.30 Our report is the first to quantify survivin mRNA expression in both exfoliated urinary cells and corresponding tumor tissues. In our study, no correlation could be demonstrated between tissue and urinary survivin levels. Surprisingly, however, exfoliated urinary cells had an approximately 10-fold higher relative survivin expression than tumor tissues. The reasons for these differences are not known and may even be due to artifacts, such as a preferential decay of PBGD mRNA expression in urine. Alternatively, exfoliated urinary cells may be enriched for very aggressive tumor cells with tremendous survivin overexpression, whereas tumor tissue is more heterogeneous with relatively high proportions of resting or even normal cells. The sensitivity of urinary survivin for bladder cancer detection was lower than initially reported8 but in good accordance with a recent study using a semiquantitative immunoassay.11 Regardless of whether survivin detection strategies in urine samples are based on protein or mRNA analysis, they should yield comparable results, since survivin is a short-lived nonsecreted protein,17, 31 and detection is dependent on its abundance in exfoliated malignant cells. Our RT-PCR-based detection method was markedly superior to voided urine cytology in its diagnostic accuracy. However, larger studies are needed to validate the diagnostic value of urinary survivin. Moreover, further investigations should focus on optimizing strategies for detecting survivin in urine. This appears especially promising in view of the excellent specificity repeatedly reported for urinary survivin.
Taken together, our findings demonstrate that expression of the IAP survivin commonly occurs in bladder cancer tissue. Moreover, the expression levels relate to main clinicopathologic variables of bladder cancer and appear to predict time to recurrence in noninvasive tumors. RT-PCR-based analysis of survivin mRNA is an objective tool for determining the expression level of this IAP protein in individual tumors, which may provide further prognostic information. When applied in exfoliated urinary cells, this approach has excellent specificity in detecting bladder cancer. Given the emerging data on the important role of survivin in promoting tumor cell survival and proliferation, our results support the search for survivin targeting strategies as novel approaches in the management of bladder cancer. Promising results have already been reported for survivin targeting in bladder cancer cell lines.32, 33
The authors thank Ms. Petra von Kwiatkowski and Ms. Birgit Schak from the Department of Urology for their excellent technical support in collecting and assessing the samples. They are also grateful to Dr. Werner Hopfenmüller, Department of Medical Statistics, for help in performing the statistical analysis and to Dr. Joanne Weirowski for her linguistic advice.