Bikunin Loss in Urine as Useful Marker for Bladder Carcinoma

Bikunin Loss in Urine as Useful Marker for Bladder Carcinoma Ke-Hung Tsui,* Petrus Tang,* Chiao-Yun Lin, Phei-Lang Chang, Chih-Hao Chang and Benjamin Yat-Ming Yung† From the Chang Gung Bioinformatics Center, Department of Urology, Chang Gung Memorial Hospital (KHT, PLC, PT), and Departments of Pharmacology (CYL) and Basic Medical Sciences (CHC), College of Medicine, Chang Gung University, Taiwan, Republic of China, and Department of Health Technology and Informatics, Hong Kong Polytechnic University, Hong Kong, Special Administrative Region of the People’s Republic of China (YMC, BYMY)

Purpose: We searched for bladder tumor markers by analyzing urine samples from patients with bladder cancer and normal individuals. Materials and Methods: Proteins in urine samples of patients with cancer and normal subjects were systematically examined by 2-dimensional electrophoresis combined with matrix assisted laser desorption ionization time-of-flight mass spectrometry. Of the proteins bikunin expression was confirmed by Western blot analysis and further evaluated. To correlate urinary bikunin levels with clinical significance we examined urine samples from patients with bladder cancer and normal controls for bikunin expression in parallel with pro-urolinase-plasminogen activator, which was previously shown to be associated with advanced bladder carcinoma. Results: A significant relationship was established between the low level and absence of bikunin, and pro-urolinase-plasminogen activator in urine samples from patients with bladder tumors. Conclusions: Analysis of urinary proteomes may be a feasible, noninvasive and efficient strategy for searching for potential bladder tumor biomarkers. We identified bikunin loss in urine as a potential bladder carcinoma marker. Key Words: urinary bladder; carcinoma; proteomics; urine; tumor markers, biological T RANSITIONAL cell carcinoma of the bladder is the second most common malignancy of the genitourinary tract and the second most common cause of death from genitourinary tumors.1,2 Cytological analysis of voided urine is the most commonly used noninvasive method to detect transitional cell carcinoma but its usefulness is severely constrained by its low sensitivity.3,4 Several potential diagnostic markers for bladder cancer have been identified, including nuclear matrix protein 22, bladder tumor antigen, telomerase, survivin, bladder cancer antigen and ImmunoCyt®.5– 8 These new molecular markers seem to

have independent diagnostic significance for bladder cancer. Bladder cancer heterogeneity also implies that multiple rather than single biomarkers may be required for accurate diagnosis.5– 8 On the other hand, specific genetic alterations have been implicated in the molecular pathogenesis of transitional cell carcinoma with mutations reported in cell cycle regulatory genes, oncogenes and tumor suppressor genes.9 However, it has proven difficult to use these genetic alterations as diagnostic markers of bladder cancer because of their low sensitivity. All of these drawbacks in existing methods have

0022-5347/10/1831-0339/0 THE JOURNAL OF UROLOGY® Copyright © 2010 by AMERICAN UROLOGICAL ASSOCIATION

Vol. 183, 339-344, January 2010 Printed in U.S.A. DOI:10.1016/j.juro.2009.08.109

Abbreviations and Acronyms 2-D ⫽ 2-dimensional 2-DE ⫽ 2-D electrophoresis IPG ⫽ inositol phosphoglycan MALDI-TOF ⫽ matrix assisted laser desorption ionization timeof-flight PVDF ⫽ polyvinylidene difluoride RBC ⫽ red blood cell SDS ⫽ sodium dodecyl sulfate SDS-PAGE ⫽ SDS-polyacrylamide gel electrophoresis TBST ⫽ tris buffered salineTween u-PA ⫽ urolinase-plasminogen activator Submitted for publication April 8, 2009. Study received approval from the review board for the protection of human subjects. Supported by Chang Gung Memorial Hospital Research Funding Grants CMRPD150103 and CMRPG370121, National Science Council (Republic of China) Grant NSC96-2320-B-182-027-MY3 and HKPolyU Grant 155.B1.DD52. * Equal study contribution. † Correspondence: Department of Health Technology and Informatics, Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, Special Administrative Region of the People’s Republic of China (telephone: 852 3400 8683; FAX: 852 2362 4365; e-mail: [email protected]).

www.jurology.com

339

340

BIKUNIN LOSS IN URINE AS USEFUL BLADDER CANCER MARKER

prompted the search for improved noninvasive markers for bladder cancer. Bikunin, a Kunitz-type protease inhibitor found in human amniotic fluid and urine, has anti-inflammatory and antimetastatic functions in animals and humans.10 The clinical efficacy of bikunin therapy has been investigated in patients with sepsis, pancreatitis, lung injury and advanced cancer. Several reports suggest that bikunin is useful and effective for controlling each of these diseases without any side effects. The broad spectrum of the biological functions of bikunin suggests the existence of multiple molecular targets that mediate diverse response to the compounds in cells. Current understanding of the bikunin mechanism of action comes mostly from studies of the suppression of several signaling cascades. Upon exposure to bikunin CD44 associates with an as yet unidentified receptor for bikunin and modulates the transcription of target genes through mitogen-activated protein kinase and phosphoinositide-3-kinase-dependent transcription. Thus, signaling activation suppression may account for the preventive action of bikunin against cancer.11,12 Bladder cancer is one of the most amenable carcinomas for tumor marker development because many tumor associated molecules are secreted in urine. To search for potential biomarkers of bladder carcinoma we previously systematically analyzed proteins secreted from bladder cancer cell lines. When evaluating the urinary levels of one of the secreted proteins, pro-u-PA, we implied that loss of pro-u-PA could be a potential marker of advanced bladder carcinoma.13 Establishing the profile of urinary proteins is an important step toward identifying molecular markers of carcinogenesis that would form the basis of a test for diagnosis and for monitoring prognosis. Proteins secreted by or shed from tumor cells are theoretically more likely than those residing in tumor cells to be detectable in body fluids such as urine. It is also possible that certain metabolites are secreted and, thus, detectable in urine in normal individuals while they are converted and/or absent in the urine of patients with cancer. To avoid misleading analysis of nonspecific variations among individual samples and have better, full use of urinary samples it becomes conceivable and important for us to evaluate proteomic changes in tumor sample compared with normal controls in groups. We analyzed urinary proteins in bladder cancer and normal control groups (5 individuals each) by 2-DE and MALDI-TOF mass spectrometry. Of identified proteins the level of bikunin was evaluated and compared by Western blotting in urine samples from more patients with cancer and normal controls. The association of urinary bikunin expression with

tumor was determined. This could possibly be useful as a bladder carcinoma marker.

MATERIALS AND METHODS Study Population and Urine Samples The study included 35 patients with bladder cancer treated with transurethral resection of bladder tumor between April 2004 and July 2005. Normal noncancer subjects, consisting of 15 healthy donors, were also included in this survey period. Studies were done with the approval of the institutional review board for the protection of human subjects. All tumors were reviewed by an experienced genitourinary pathologist to determine bladder transitional cell carcinoma. Pathological staging was done according to 2002 TNM staging14 and the WHO consensus classification was used to grade these tumors.15

Human Urine Sample Preparation Midstream void urine samples (50 ml) were collected from healthy donors and patients with bladder cancer, and 15 ml of each sample from a total of 5 patients with cancer or normal subjects were pooled together. There were 7 patients with cancer and 3 normal controls. Healthy donors and patients fasted at least 6 hours before samples were taken. To remove insoluble materials and RBC samples were immediately centrifuged at 3,000 rpm for 15 minutes at 4C. RBCs were noted in the pellets. Supernatants clear of RBCs were collected and centrifuged at 15,000 ⫻ gravity for 30 minutes at 4C. Final supernatants were loaded onto Centricon® 5 kDa membrane to concentrate proteins and remove small interference molecules. Briefly, Centricons were spun at 3,700 ⫻ gravity at 4C for 1 hour to reduce volume to 1 ml. Two volumes of cold 20% trichloroacetic acid were added to the concentrated urine samples. The mixture was stored in ice for 30 minutes and pellet was obtained by centrifugation at 13,000 ⫻ gravity for 20 minutes at 4C. The pellet was washed twice with cold acetone. Protein amounts in urine concentrates were measured using the bladder cancer antigen protein assay and frozen at ⫺80C.

2-DE Analysis Urinary protein samples (80 ␮g each) were separated by 13 cm Immobiline™ DryStrip 3-10 linear on the IPGphor™ IEF System in the first dimension. For soluble proteins from patients with early and late stage bladder carcinoma samples were separated by 13 cm Immobiline DryStrip 4-7 linear in the first dimension. After equilibration IPG gel strips were transferred onto 10% SDS-PAGE Hoefer SE600 vertical gels (Amersham Biosciences, Piscataway, New Jersey) for the second dimension. After fixing for 5 hours all gels were visualized with a silver staining technique. Protein spots were quantified using ImageMaster 2D Elite software (Amersham Biosciences). All experiments were repeated a minimum of 3 times.

Mass Spectrometric Analysis of Urinary Proteins Silver stained spots were excised and digested in gel with trypsin according to previously described procedures.16 Briefly, gels were destained by 1% potassium ferricyanide and 1.6% sodium thiosulfate (Sigma®), reduced with 25 mM NH4HCO3 containing 10 mM dithiothreitol (Bio-

BIKUNIN LOSS IN URINE AS USEFUL BLADDER CANCER MARKER

Table 1. Patient characteristics No. Pts Av age (range) Disease status: Localized Metastatic T stage: Ta T1 Pathological grade: High Low Tumor size (mm): 5 or Less Greater than 5 No. tumors: Single Multiple

68 (30–85) 35 0 15 20 25 10 8 27 3 32

synth®) at 60C for 30 minutes and alkylated with 55 mM iodoacetamide (Amersham Biosciences) at room temperature for 30 minutes. After reduction and alkylation proteins were digested with trypsin (Promega, Madison, Wisconsin) (20 ␮g/ml) at 37C overnight. After digestion tryptic peptides were acidified with 0.5% trifluoracetic acid and loaded onto an MTP AnchorChip™ 600/384 TF. MALDI-TOF mass spectrometry analysis was performed on an Ultraflex™ MALDI-TOF mass spectrometer. Monoisotopic peptide masses were assigned and used for database searches with the MASCOT search engine (Matrix Science, London, United Kingdom).

Western Blotting Cells were lysed in RIPA buffer composed of 1% Triton X-100, 1% SDS, 20 mM Na2HPO4, 100 mM NaCl and 0.2 mM phenylmethylsulfonyl fluoride. Lysates were boiled in

341

SDS sample buffer composed of 62.5 mM tris (pH 6.8), 5% ␤-mercaptoethanol (Merck, Whitehouse Station, New Jersey), 10% glycerol, 2% SDS and 0.001% bromophenol blue, and subsequently fractionated by 10% SDS-PAGE. Separated proteins in SDS-PAGE were electrotransferred to Hybond™ PVDF membrane. The PVDF membrane was soaked in blocking solution containing 5% (weight per volume) nonfat milk in TBST composed of 20 mM tris (pH 7.5), 0.5 M NaCl and 0.1% (volume per volume) Tween-20 for 1 hour at room temperature. To assess u-PA component levels the soaked PVDF membrane was incubated with polyclonal Ab against bikunin and pro-u-PA, diluted 1:1,000 in 5% (weight per volume) nonfat milk in TBST, for 4 hours at room temperature, washed with TBST 3 times for 15 minutes each and incubated in horseradishperoxidase conjugated goat anti-mouse IgG antibody, diluted 1:5,000 in TBST buffer, at room temperature for 1 hour. After washing immunobands were detected by enhanced chemiluminescence reaction (Amersham Pharmacia Biotech, Piscataway, New Jersey).

RESULTS To search for potential biomarkers of bladder carcinoma we systematically analyzed urinary proteins secreted from patients with bladder cancer and normal controls. Table 1 lists patient and tumor characteristics. After concentration proteins were resolved on 2-DE and silver stained (fig. 1). Spots were excised individually, digested in gel with trypsin and identified by MALDI-TOF mass spectrometry (fig. 1). A total of 20 urinary proteins were identified (table 2). While most of them were identified in the cancer and control groups, bikunin and lectin were highly expressed in cells of normal controls (figs. 1 and 2).

Figure 1. Silver stained 2-DE gels of 80 ␮g urine protein samples from patient with bladder cancer and normal controls were resolved by 2-DE, including first dimension with IPG (pH 3 to 10) and second dimension with 12% SDS-PAGE. Detected protein spots were marked, numbered and excised for further analysis.

342

BIKUNIN LOSS IN URINE AS USEFUL BLADDER CANCER MARKER

Table 2. Urine proteins in bladder cancer and normal control groups identified by MALDI-TOF mass spectrometry Spot* (protein identified)

Mr

pl

Mascot Search Score/% Coverage†

Accession No.

1 (albumin, isoform CRA_t) 2 (similar to human albumin) 3 (lectin, mannose-binding 2) 4 (lectin, mannose-binding 2) 5 (complex-forming glycoprotein HC) 6 (Ig␬ light chain VLJ region) 7: Chain A, crystal structure of Fab fragment of human monoclonal IgM cold agglutinin IGKV1–5 PROTEIN IGKC protein Ig␬ light chain VLJ region 8: Ig␬ light chain VLJ region Chain A, crystal structure of Fab fragment of human monoclonal IgM cold agglutinin IGKC protein 9 (chain A, crystal structure of Fab fragment of human monoclonal IgM cold agglutinin) 10 (chain A, crystal structure of Fab fragment of human monoclonal IgM cold agglutinin) 11 (anti-tumor necrosis factor-␣ antibody light-chain Fab fragment) 12 (Ig␬ light chain variable region) 13 (bikunin) 14 (bikunin) 15 (Ig␬ chain C region [allotype Inv(1,2)])

60211 53416 26812 26812 20592 28907

6.66 5.84 5.12 5.12 5.84 6.73

288/77 267/80 108/53 108/53 110/65 83/27

gil1196260 gil763431 gil1196054 gil1196054 gil223373 gil2166945

23552 26503 25915 28548

5.75 5.75 5.75 5.75

79/36 77/36 76/36 74/36

gil1083579 gil2170666 gil4925811 gil2166931

28822 23552 25915 23552

6.73 5.75 6.15 5.75

83/28 76/28 72/36 77/41

gil2166946 gil1083579 gil4925811 gil1083579

23552

5.75

82/38

gil1083579

23787 21207 16164 16164 11161

6.19 7.60 4.76 4.76 5.61

71/37 78/22 69/25 69/25 74/29

gil1127530 gil4847543 gil585132 gil585132 gil106529

* Numbering of protein bands in figure 1. † Sequence coverage of matched peptides in the protein identified.

To verify the mass spectrometry assisted identification of bikunin and correlate urinary bikunin levels with clinical significance we examined urine samples from 7 patients with bladder cancer and 3 normal controls for bikunin expression (fig. 3). Since pro-u-PA loss in urine samples associated with tumor was previously shown to be associated with advanced bladder carcinoma,7 the level of pro-u-PA

Normal

Patients

Lectin

was also measured in parallel. Bikunin and prou-PA were abundant in controls but virtually not detected in the cancer groups. Our results indicate that there may be a significant relationship between the low or absent level of bikunin in the urine of patients with bladder carcinoma. To show the variation in 2-D gel patterns and obtain a global view of the protein expression profile in urine from patients with early and late stage bladder carcinoma urinary proteins were separated by 2-DE using a wide range IPG strip (pH 4 to 7). Bikunin was not expressed in early or late stage bladder carcinoma urine samples. N1 N2 P1 P2 P3 P4

N3 P5 P6 P7 bikunin

Bikunin

Figure 2. Cropped images show select proteins from 2-DE gel. Spots over-expressed in normal controls were identified as bikunin and lectin on MALDI-TOF-MS.

N1 N2 P1 P2 P3 P4

N3 P5 P6 P7 pro-uPA

Figure 3. Western blot analysis of bikunin and pro-u-PA in 80 ␮g urinary protein samples from patients 1 to 7 with cancer (P1 to P7) and 3 normal controls (N1 to N3). Samples were resolved on 12% SDS-PAGE, blotted onto PVDF membrane and probed with specific antibodies against bikunin and pro-u-PA.

BIKUNIN LOSS IN URINE AS USEFUL BLADDER CANCER MARKER

DISCUSSION The proteome is much more complex and dynamic than the genome. Thus, it could potentially overcome some limitations of other approaches to identifying new marker molecules. Proteomic approaches use MALDI-TOF mass spectrometry, which screens for serum protein patterns with high sensitivity and specificity for identifying patients with bladder cancer regardless of tumor stage.17 Proteins secreted from tumor cells are potential biomarkers for disease diagnosis and/or prognosis. In this study a set of about 20 urinary proteins from patients with bladder cancer and normal controls was systematically identified by a straightforward proteomic approach, consisting of 2-D SDSPAGE and MALDI-TOF mass spectrometry. Expression profiles of urinary proteins in the cancer and normal control groups were similar but not identical. Bikunin was unique to the control group. Cystoscopy with cytology is the mainstay for diagnosing bladder cancer. Cytology is specific for diagnosing bladder carcinoma but less sensitive, particularly for detecting low grade disease. On the other hand, cystoscopy is an invasive, relatively costly technique that may also be inconclusive at times, particularly in cystitis cases. Thus, a simple, noninvasive marker to detect bladder cancer would be beneficial. Bikunin identified in this study was differentially expressed in the urine of patients with bladder cancer compared to that in normal controls. Bikunin was detected in control urine samples. A significant relationship was found between the low level and absent bikunin in the urine of patients with bladder carcinoma. Results suggest that a low level of urine bikunin may be associated with bladder carcinoma. Bikunin is a plasma proteinase inhibitor whose physiological function is only now beginning to be revealed.18 –19 The broad spectrum of the biological function of bikunin suggests the existence of multiple molecular targets that mediate diverse responses to compounds in cells. Our current understanding of the bikunin mechanism of action comes mostly from studies of the suppression of several signaling cascades.11,20 Bikunin may block tumor cell invasion by inhibiting tumor cell associated plasmin activity and u-PA expression at the gene and protein levels, possibly by expression of the mitogen-activated protein kinase signaling cascade. Suppressing signaling activation may explain cancer prevention by bikunin. Treatment of patients with cancer with bikunin may be beneficial in the adjuvant setting to delay the onset of metastasis and/or in combination with cytotoxic agents to improve treatment efficiency in those with advanced cancer.10 Extracellular proteolytic enzymes have been implicated in cancer metastasis. The release of proteolytic enzymes in tumors facilitates cancer

343

cell invasion into surrounding normal tissue by the breakdown of basement membranes and the extracellular matrix.21 Plasminogen activation may have an important role in this process. The plasminogen activator u-PA is capable of catalyzing conversion of the inactive zymogen plasminogen to the active proteinase plasmin, which can then degrade most extracellular proteins.21 Studies have established that the level of u-PA in malignant tumors is significantly higher than in corresponding normal tissue or in benign tumors of the same tissue.22 For human malignancy u-PA was the first proteinase shown to be a prognostic marker. Duffy et al reported that patients with breast tumors with high u-PA enzyme activity had a significantly shorter disease-free interval than patients with tumors with low levels of activity.23 Higher u-PA antigen levels were subsequently found to correlate with shortened overall survival in patients with this disease.24 For breast cancer u-PA antigen levels appear to be among the most relevant, direct and strongest prognostic factors. Apart from breast cancer u-PA was also shown to be a prognostic marker of other malignancies, including cancer of the lung,25 bladder,26 stomach, ovary and brain. In our recent study u-PA was virtually not detected in normal urine or was minimally present in the few cancer urine samples but loss of pro-u-PA in urine was associated with advanced bladder carcinoma.13 This low expression likely suggests that pro-u-PA may be more readily converted to u-PA in cancer cells, resulting in the loss of pro-u-PA in the urine of patients with cancer. Similar to pro-u-PA, bikunin was highly expressed in the urine of normal controls. Since bikunin has a critical role in inhibiting uPA expression, loss of bikunin results in high uPA expression and loss of pro-u-PA in cancer cases. Thus, our results indicate that there may be an important linkage among bikunin, pro-u-PA and u-PA in bladder carcinoma. Based on our observation the loss of bikunin and pro-u-PA in urine may serve as a highly reliable cancer biomarker.

CONCLUSIONS Identifying the urinary proteomes developed in this study may serve as an ideal, efficient method to establish a panel of potential biomarkers. Noninvasive detection of biomarkers in urine may be useful for clinical application in bladder cancer diagnosis and prognosis. Proteins secreted from bladder tumor or normal cells and present in urine have potential diagnostic or prognostic value. The biomarkers identified in this study may be therapeutic targets for the future development of novel antitumor agents.

344

BIKUNIN LOSS IN URINE AS USEFUL BLADDER CANCER MARKER

REFERENCES 1. Raghaven D, Shipley WU, Garnick MB et al: Biology and management of bladder cancer. New Engl J Med 1990; 322: 1129.

10. Kobayashi H, Suzuki M, Hirashima Y et al: The protease inhibitor birkunin, a novel anti-metastatic agent. Biol Chem 2003; 384: 749.

19. Yagyu T, Kobayashi H, Matsuzaki H et al: Enhanced spontaneous metastasis in bikunin-deficient mice. Int J Cancer 2006; 118: 2322.

2. Newling DW: Preventing recurrence and progression in superficial bladder cancer. Curr Opin Urol 1996; 6: 272.

11. Matsuzaki H, Kobayashi H, Yagyu T et al: Birkunin inhibits lipopolysaccharide-induced tumor necrosis factor alpha induction in macrophages. Clin Diagn Lab Immunol 2004; 11: 1140.

3. Konety BR and Getzenberg RH: Urine based markers of urological malignancy. J Urol 2001; 165: 600.

12. Matsuzaki H, Kobayashi H, Tanakan Y et al: Bikunin target genes in ovarian cancer cells identified by microarray analysis. J Biol Chem 2003; 278: 14640.

20. Kobayashi H, Suzuki M, Tanaka Y et al: Suppression of urokinase expression and invasiveness by urinary trypsin inhibitor is mediated through inhibition of protein kinase C- and MEK/ERK/c-Jundependent signaling pathways. J Biol Chem 2001; 276: 2015.

4. Konety B and Lotan Y: Urothelial bladder cancer: biomarkers for detection and screening. BJU Int 2008; 102: 1234. 5. van Tilborg AA, Bangma CH and Zwarthoff EC: Bladder cancer biomarkers and their role in surveillance and screening. Int J Urol 2009; 16: 23. 6. Ramakumar S, Bhuiyan J, Besse JA et al: Comparison of screening methods in the detection of bladder cancer. J Urol 1999; 161: 388. 7. Konety BR: Molecular markers in bladder cancer: a critical appraisal. Urol Oncol Semin Orig Invest 2006; 24: 326. 8. Quek ML, Sanderson K, Daneshmand S et al: New molecular markers for bladder cancer detection. Curr Opin Urol 2004; 14: 259. 9. Sidransky D, Von Eschenbach A, Tsai YC et al: Identification of p53 gene mutations in bladder cancers and urine samples. Science 1991; 252: 706.

13. Lin CY, Tsui KH, Yu CC et al: Searching cell secreted proteomes for potential urinary bladder tumor markers. Proteomics 2006; 6: 4381. 14. Greene FL, Page DL, Flemming ID et al: AJCC Cancer Staging Manual, 6th ed. New York: Springer-Verlag 2002. 15. Mostofi FK, Davis CJ and Sesterhenn IA: Histological typing of urinary bladder tumors. In: International Histological Classification of Tumors. Geneva: World Health Organization 1998. 16. Shevchenko A, Wilm M, Vorm O et al: Mass spectrometric sequencing of proteins silverstained polyacrylamide gels. Anal Chem 1996; 68: 850. 17. Schwamborn K, Krieg RC, Grosse J et al: Serum proteomic profiling in patients with bladder cancer. Eur Urol 2009; Epub ahead of print. 18. Fries E and Blom AM: Bikunin—not just a plasma proteinase inhibitor. Int J Biochem Cell Biol 2000; 32: 125.

21. Andreasen PA, Kjoller L, Christensen L et al: The urokinase-type plasminogen activator system in cancer metastasis: a review. Int J Cancer 1997; 72: 1. 22. Duffy ML: Proteases as prognostic markers in cancer. Clin Cancer Res 1996; 2: 613. 23. Duffy MJ, O’Grady P, Devaney D et al: Urokinaseplasminogen activator, a marker for aggressive breast cancers. Cancer 1988; 62: 531. 24. Duffy MJ, Reilly D, O’Sullivan C et al: Urokinaseplasminogen activator, a new and independent prognostic marker in breast cancer. Cancer Res 1990; 50: 6827. 25. Oka T, Ishida T, Nishino T et al: Immunohistochemical evidence for Urokinase plasminogen activator in primary and metastatic tumors for pulmonary carcinoma. Cancer Res 1991; 51: 3522. 26. Hasui Y, Marutsuka K, Suzumiya J et al: The content of urokinase-type plasminogen activator as a prognostic factor in bladder cancer. Int J Cancer 1992; 50: 871.