1131 Expression of Thymidine Phosphorylase as an Indicator of Poor Prognosis for Patients with Transitional Cell Carcinoma of the Bladder Jun-ichirou Arima, M.D.1,2 Yoshiharu Imazono, M.D.1,2 Yuji Takebayashi, M.D.1,3 Kenryu Nishiyama, M.D.2 Tsutomu Shirahama, M.D.2 Suminori Akiba, M.D.4 Tatsuhiko Furukawa, M.D.1 Shin-ichi Akiyama, M.D.1 Yoshitada Ohi, M.D.2 1 Department of Cancer Chemotherapy, Institute for Cancer Research, Kagoshima University, Kagoshima, Japan. 2 Department of Urology, Kagoshima University, Kagoshima, Japan. 3 First Department of Surgery, Kagoshima University, Kagoshima, Japan. 4 Department of Public Health, Kagoshima University, Kagoshima, Japan. BACKGROUND. Thymidine phosphorylase (TP), which is identical to platelet-derived endothelial cell growth factor (PD-ECGF), stimulates chemotaxis of endothelial cells and is involved in the angiogenesis of human solid tumors. METHODS. The activity and expression of TP were examined in human transitional cell carcinomas (TCCs) of the bladder, and their association with clinicopathologic findings was determined. The activity of the enzyme in 37 TCCs and 12 adjacent nonneoplastic tissues was measured spectrophotometrically. The expression of TP was also examined by immunoblotting. Immunohistochemical analysis was performed on 108 TCCs. RESULTS. TP activity in the carcinomas was higher than that in adjacent normal tissues (P ⫽ 0.002). TP activity in Grade 3 tumors or those classified as pT2– 4 was higher than in Grade 1 and 2 tumors (P ⫽ 0.017) or those classified as pT1 (P ⫽ 0.007). The level of expression of TP detected by immunoblotting correlated well with TP activity. Immunohistochemical analyses showed that 62 of 108 cases (57.4%) were TP positive. There was a significant correlation between TP expression and histologic grade, infiltration pattern, local invasion, and lymph node metastasis. TP expression as a prognostic variable was studied using the Cox proportional hazards model. TP overexpression was an independent prognostic factor, as were lymph node metastasis and local invasion. CONCLUSIONS. These findings suggest that TP activity and its level of expression influence the progression of TCC and the prognoses of patients with this disease. Cancer 2000;88:1131– 8. © 2000 American Cancer Society. KEYWORDS: thymidine phosphorylase, transitional cell carcinoma, prognostic factor, angiogenesis. Supported in part by Grants-in-Aid from the Ministry of Education, Science and Culture, and the Ministry of Health and Welfare, Japan, and a Research Grant from the Princess Takamatsu Cancer Research Fund. The authors thank Ms. Y. Makiyama for technical assistance, and Drs. K. Shimoinaba (Shimoinaba Urological Clinic) and M. Ikoma (Ikoma Urological Clinic) for providing clinical samples. Address for reprints: Shin-ichi Akiyama, M.D., Department of Cancer Chemotherapy, Institute for Cancer Research, Kagoshima University, Sakuragaoka 8-35-1, Kagoshima 890-8520, Japan. Received April 26, 1999; revision received October 18, 1999; accepted November 22, 1999. © 2000 American Cancer Society ngiogenesis is a prerequisite for tumor growth,1 and angiogenic activity is correlated with a higher incidence of lymph node metastases and poor prognosis for patients with numerous different types of tumors, including bladder carcinoma.2,3 Thymidine phosphorylase (TP) is involved in pyrimidine nucleoside metabolism, catalyzing the reversible phosphorolysis of thymidine, deoxyuridine, and their analogues to their respective bases and 2-deoxyribose-1-phosphate.4 – 6 In mammals, the enzyme consists of 2 identical subunits, each with a molecular weight of about 55kD.7 We and others previously demonstrated that TP is identical to plateletderived endothelial cell growth factor (PD-ECGF).8 –11 PD-ECGF/TP has angiogenic activity for which the enzymatic activity of TP is needed. Among the products of degradation of thymidine by TP, 2-deoxy-D-ribose, a dephosphorylated product derived from 2-deoxy-D-ribose-1-phosphate, has chemotactic activity in vitro and an- A 1132 CANCER March 1, 2000 / Volume 88 / Number 5 giogenic activity in vivo.12 This suggests that TP phosphorolysis product(s) may stimulate chemotaxis of endothelial cells and possibly other cells, causing angiogenesis. Compared with adjacent normal tissues, TP activity has been reported to be increased in a variety of malignant tumors.13–16 Increased TP expression correlates with microvessel density in several solid turmors, such as breast carcinomas,17 ovarian carcinomas,15 renal cell carcinomas (RCCs),18 and colorectal carcinomas.19 In RCCs and colorectal carcinomas, TP positivity, but not microvessel density, is an independent prognostic factor.18,19 These findings suggest that TP is involved in angiogenesis and the progression of some solid tumors, and that TP has effects on processes other than just angiogenesis. O’Brien et al. reported that tumor cell TP expression in bladder carcinomas correlated with tumor grade but not with tumor vascularity, relapse free survival, or overall survival.20 On the other hand, Mizutani et al. reported that patients with Ta bladder carcinoma with a low level of PD-ECGF/TP expression had a longer postoperative tumor free period than those with a high level of expression during a 3-year follow-up study.21 Thus, we examined retrospectively the expression of TP in transitional cell carcinoma (TCC) and its association with clinicopathologic findings and prognosis. MATERIALS AND METHODS Patients We examined 108 patients with TCC of the bladder (84 males and 24 females) treated by radical cystectomy from May 1984 to June 1996 at the Department of Urology, Kagoshima University Hospital, Kagoshima, Japan. All experiments were performed after obtaining informed consent from patients according to institutional rules. The average age of patients at the time of surgery was 64.5 years (range, 35–79). Follow-up of the patients, including survival analysis, was updated in March 1997 (the median follow-up was 51 months [range, 2–151]). The immunohistochemical evaluations were completed in March 1997. At that time, 27 patients had died of TCC, 54 patients were alive, and 27 patients had died of other diseases. None of the patients had received prior chemotherapy or irradiation. Twelve patients were treated by transurethral resection before radical cystectomy. Tumors were graded according to the criteria of the Japanese Urological Association22 and the TNM system.23 Histologic grades were classified into two groups, Grades 1 ⫹ 2 and Grade 3, and a tumor was given the predominant grade if it had mixed grades. Infiltration patterns were classified into two groups, INF ␣ for expanded growth and INF ␤⫹␥ for infiltration with cell clusters and single cells or microclusters. Venous involvement was classified as pV0 for no venous invasion and pV1 for the presence of venous invasion. Lymph duct involvement was classified into two groups, pL0 for no lymph duct invasion and pL1⫹2 for the presence of lymph duct invasion up to the superficial muscle layer and beyond it. Local invasion was also classified into two groups, pT1 and pT2– 4. Lymph node metastases were classified into two groups, pN0 and pN1–3. The pN category was determined by intraoperative and preoperative findings from computed tomography scans, ultrasonography, and other radiographic studies. The M category was based on the results of chest X-rays, skeletal surveys, bone scans, computed tomography, and other radiographic studies. Patient and tumor profiles are summarized in Table 1. Preparation of Monoclonal Antibody against TP Monoclonal antibodies were raised against a glutathione S-transferase (GST) TP fusion product containing 244 amino acids of TP (residues 7–250), as previously described.16 Tissue Specimens and Preparation of Homogenates TCC samples (10 renal pelvic and/or ureter tumors consisting of 2 pT1G2, 2 pT2G3, 5 pT3G2, and 2 pT3G3, and 27 bladder tumors) were obtained at surgery from 37 patients. Twelve noncancerous urothelia (8 renal pelvic urothelia from the patients with RCC and 4 bladder urothelia from the patients with benign prostatic hypertrophy) were obtained at surgery. Specimens were frozen at ⫺80 °C within 10 minutes and then homogenized in lysis buffer (50mM Tris-HCl, pH 6.8, containing 1% Triton X-100, 2 mM PMSF, and 0.02% 2-mercaptoethanol). The lysates were centrifuged at 15,000 x g for 30 minutes at 4 °C. The activities of TP in the supernatants were assayed, and the supernatants were also used for immunoblot analyses. Protein content was determined by the method of Bradford.24 Assay for TP Activity TP activity was assayed by the spectrophotometric procedure described by Friedkin and Roberts with some modifications.4 Supernatants from homogenates, containing 100 g protein, were incubated with 0.1 M Tris-arsenate buffer (pH 6.5) containing 10 mM thymidine in a total volume of 0.1 mL. After incubation for 1 hour at 37 °C, the reaction was stopped by adding 1 mL of 0.2 N NaOH, and the amount of thymine formed was determined by measuring the absorbance at 300 nm. TP Thymidine Phosphorylase and TCC/Arima et al. TABLE 1 Patient and Tumor Profiles 1133 Immunohistochemistry Characteristics No. of patients (%) Total no. Gender Male Female Age (yrs) Median Range Grade G1 G2 G3 Infiltration pattern INF␣ INF␤ INF␥ Venous involvement pV0 pV1 Lymph duct involvement pL0 pL1⫹2 pT category pT1 pT2 pT3 pT4 pN category pN0 ⱖ pN1 108 (100) 84 (77.8) 24 (22.2) 61 35–79 5 (4.6) 46 (42.6) 57 (52.8) 20 (18.5) 67 (62.0) 21 (19.5) 76 (70.4) 32 (29.6) 49 (45.4) 59 (54.6) 49 (45.4) 15 (13.9) 23 (21.3) 21 (19.4) 91 (84.3) 17 (15.7) INF: infiltration. activity is expressed as the amount of thymine formed per g protein per hour. Immunoblotting Samples containing 50 g protein were resolved by 11% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and the proteins in the gel were electrophoretically transferred to a sheet of PVDF as previously described.25 After transfer, the membrane was placed in TTBS (400 mM NaCl, 20 mM Tris-HCl, pH 8.0, 0.05% Tween-20) containing 5.0% (weight/vol) skimmed milk for 1 hour. The blots were incubated overnight with TTBS containing 5% skimmed milk and a monoclonal antibody against TP (1:1000 dilution). After washing 4 times in TTBS for 10 minutes each, the sheet was incubated for 1 hour with horseradish-peroxidase (HRP)– conjugated horse antimouse immunoglobulin (Ig)G (Amersham, Buckinghamshire, UK) diluted 1:1000 in TTBS. After 3 washes in TTBS for 10 minutes each, the HRP activities were detected with the ECL Western blot detection system (Amersham). TP was investigated immunohistochemically using a monoclonal antibody against TP and an avidin-biotinperoxidase complex (ABC-PO Vectastain kit, Vector Laboratories, Inc., Burlingame, CA). Samples were fixed with 10% formaldehyde in phosphate-buffered saline (PBS), embedded in paraffin, and cut into 3-m-thick sections. The sections were deparaffinized with xylene and treated with 98% ethanol. Endogenous peroxidase was blocked by immersing the slides in 0.02% H2O2 in absolute methanol for 20 minutes at room temperature. The sections were washed 3 times with PBS for 5 minutes each, incubated with 3% skimmed milk in PBS for 30 minutes at room temperature, and then incubated at 4 °C overnight with a monoclonal antibody against TP diluted 400-fold with 3% skimmed milk in PBS. After washing 3 times with PBS for 5 minutes each, they were then incubated for 30 minutes with biotinylated antimouse IgG at room temperature. After washing 3 or more times with PBS for 5 minutes each, the sections were incubated for 30 minutes with avidin-biotinylated horseradish peroxidase complex and then washed 3 times with PBS for 15 minutes each. The immune complexes were visualized by incubating the sections with 0.5 mg/mL diaminobenzidine and 0.03% (vol/vol) H2O2 in PBS for 7 minutes. The sections were counterstained with hematoxylin and mounted. We examined at least 200 carcinoma cells by light microscopy to determine whether or not the cells were positive for TP. The evaluation of TP expression was done by two investigators (J.A. and Y.I.) who had no knowledge of the clinicopathologic characteristics, using a “two-headed” light microscope. We separated the patients into two groups, each with about the same number of cases. As a result, a cutoff point of 5% was selected; samples were considered to be TP positive when more than 5% of the parenchyma cells were stained and TP negative when less than 5% of the cells were stained. Statistical Methods Demographic and clinicopathologic characteristics were compared using the chi-square test or the Student t test.26 We used the Kaplan–Meier method to estimate survival rates. The Cox proportional hazards models were used in survival analysis. All P values presented are two-sided. RESULTS TP Activity in Tumors and Normal Urothelium We measured the TP activity in 37 TCCs and 12 adjacent nonneoplastic tissues, including those from 4 1134 CANCER March 1, 2000 / Volume 88 / Number 5 FIGURE 1. The correlations between thymidine phosphorylase (TP) activity and (A) nonneoplastic or cancer tissues, (B) tumor grade, and (C) tumor stage, analyzed by the unpaired Student t test, are shown. Boxes correspond to interquartile ranges. Lines in boxes represent mean values. The open circles represent the outliers. benign urologic diseases. The average TP activity in TCC (8.54 nmol/100 g protein/hour) was 10.2-fold higher than in normal urothelia (0.84 nmol/100 g protein/hour) (P ⫽ 0.002). The average TP activity in G3 tumors (15.0 nmol/100 g protein/hour) was 3.3fold higher than in G1–2 tumors (4.6 nmol/100 g protein/hour) (P ⫽ 0.017), and the average in pT2– 4 tumors (11.7 nmol/100 g protein/hour) was 6.0-fold higher than in T1 (2.0 nmol/100 g protein/hour) (P ⫽ 0.007) (Fig. 1). Immunoblotting of TP from Tumors and Normal Urothelium Homogenates from 7 bladder TCCs and one noncancerous urothelium were subjected to immunoblot analysis. A monoclonal antibody against TP was used to examine the levels of TP expression in tissues. We detected TP with a molecular weight of 55,000 in carcinomas and nonneoplastic tissues. The level of TP expression in cancer tissues was higher than in nonneoplastic tissues. TP expression as well as TP activity seemed to be correlated with tumor grade and stage. In high grade tumors (G2, pT3 or G3, pT3), it tended to be higher than in low grade tumors (G1, pT1 or G2, pT1) (Fig. 2). More experiments and statistical analyses are needed to confirm these differences. Correlations among Pathologic Findings, Clinical Outcome, and TP Expression The expression of TP in carcinomas was examined by immunohistochemical staining with a monoclonal antibody against TP. The cytoplasms of TP positive carcinoma cells were as intensely stained as those of FIGURE 2. Expression of thymidine phosphorylase (TP) in transitional cell carcinomas (TCCs), detected by immunoblotting with a monoclonal antibody against TP, is shown. Lane 1, nonneoplastic tissue; Lanes 2– 8, TCCs ( 2, G1pT1; 3, G2pT1; 4, G3pT1; 5, G2pT2; 6, G3pT2; 7, G2pT3; 8, G3pT3). stromal cells (Fig. 3). The correlations between clinical or pathologic features and TP expression in human TCC are summarized in Table 2. Sixty-two (57.4%) of 108 patients were TP positive (Fig. 4). There was a significant correlation between TP expression and histologic grade, infiltration pattern, local invasion, and lymph node metastasis or extent of disease. No significant association was found between TP expression and gender, venous involvement, or lymph duct invasion. Prognostic Relevance of TP Expression Kaplan–Meier product limit estimates of overall survival are plotted in Figure 5. Patients with TP positive carcinomas had significantly (P ⬍ 0.001) poorer survival than those with negative tumors. Thus, we investigated whether TP positivity is an independent prognostic variable. A Cox proportional hazards model was constructed using established prognostic factors and TP expression. Univariate analysis showed that histologic grade (P ⬍ 0.001), infiltration pattern (P ⫽ 0.001), local invasion (P ⬍ 0.001), venous involvement (P ⫽ Thymidine Phosphorylase and TCC/Arima et al. 1135 FIGURE 3. Immunoreactivity to a monoclonal antibody against thymidine phosphorylase (TP) of transitional cell carcinoma (TCC) tissue is shown. (A) TP negative TCC, (B) TP positive TCC. 0.001), lymph duct invasion (P ⬍ 0.001), N category (P ⬍ 0.001), and TP expression (P ⫽ 0.001) were significant prognostic factors (Table 3). Multivariate survival analysis showed that lymph node metastasis (P ⫽ 0.019), local invasion (P ⫽ 0.046), and TP expression (P ⫽ 0.047) were significant prognostic factors (Table 4). DISCUSSION Angiogenic activity appears to be an important determinant of the behavior of bladder carcinomas because increased angiogenesis, as assessed by vessel counts using immunohistochemistry, is associated with shorter overall survival for patients with invasive TCC.2 The expression of several angiogenic factors has been examined in bladder carcinoma. Basic fibroblast growth factor (bFGF) was found in the urine of patients with bladder carcinoma, particularly in those with metastatic disease.27 Immunohistochemical studies demonstrated increased expression of acidic FGF in bladder carcinomas compared with normal bladders, and the strongest expression of acidic FGF was found in high grade tumors.28 Expression of vascular endothelial growth factor (VEGF) in superficial bladder tumors was significantly higher than in invasive tumors and in normal bladder.29 TP expression in invasive bladder tumors was reported to be higher than in superficial tumors and normal bladder.29 Our study also shows that TP expression is higher in human transitional cell carcinoma than in normal urothelium, and that TP activity TABLE 2 Relations between TP Expression in TCCs and Clinicopathologic Characteristics No. of patients Gender Male Female Histologic grade G1⫹2 G3 Infiltration pattern INF␣ INF␤ ⫹ ␥ Local invasion pT1 pT2–4 Venous involvement pV0 pV1 Lymph duct involvement pL0 pL1⫹2 Lymph node metastasis pN0 pN1–3 Disease extention pT1–3 and pN0 pT4 or pN1–3 No. of patients Positive Negative P valuea 84 24 51 (60.7%) 11 (45.8%) 33 (39.3%) 13 (54.2%) NS 51 57 23 (45.1%) 39 (68.4%) 28 (54.9%) 18 (31.6%) P ⫽ 0.014 20 88 7 (35.0%) 55 (62.5%) 13 (65.0%) 33 (37.5%) P ⫽ 0.025 49 59 20 (40.8%) 42 (71.2%) 29 (59.2%) 17 (28.8%) P ⫽ 0.002 76 32 40 (52.6%) 22 (68.8%) 36 (47.4%) 10 (31.2%) NS 49 59 25 (51.0%) 37 (62.7%) 24 (49.0%) 22 (37.3%) NS 91 17 47 (51.6%) 15 (88.2%) 44 (48.4%) 2 (11.8%) P ⫽ 0.003 87 21 44 (50.6%) 18 (85.7%) 43 (49.4%) 3 (14.3%) P ⫽ 0.002 INF: infiltration; NS: not significant. a P values were obtained by the chi-square test. 1136 CANCER March 1, 2000 / Volume 88 / Number 5 TABLE 3 Results of Univariate Survival Analysis FIGURE 4. Percentages of carcinoma cells obtained from 108 patients with transitional cell carcinoma (TCC) expressing thymidine phosphorylase (TP) are shown. Gender Male Female Histologic grade G1⫹2 G3 Infiltration pattern INF␣ INF␤ ⫹ ␥ Local invasion pT1 pT2–4 Venous involvement pV0 pV1 Lymph duct invasion pL0 pL1–2 Lymph node metastasis pN0 pN1–3 TP expression Negative Positive No. of patients HRa 95% CIa P valuea 84 24 0.91 Reference 0.39–2.15 0.845 51 57 Reference 7.46 2.25–24.72 20 88 Reference 9.76 1.32–71.77 49 59 Reference 7.54 2.63–21.68 76 32 Reference 3.46 1.68–7.14 49 59 Reference 6.65 2.53–17.44 91 17 Reference 7.82 3.59–17.02 46 62 Reference 4.95 1.89–12.98 ⬍ 0.001 0.001 ⬍ 0.001 0.001 ⬍ 0.001 ⬍ 0.001 0.001 INF: infiltration; HR: hazard ratio; CI: confidence interval. a Hazard ratios and 95% confidence intervals were obtained from the Cox proportional hazards model. b P values were obtained by the chi-square test. TABLE 4 Results of Multivariate Survival Analysis FIGURE 5. Kaplan–Meier survival curves of patients with transitional cell carcinoma are shown. TP: thymidine phosphorylase. in invasive bladder carcinoma classified as pT2– 4 is significantly higher than in superficial bladder carcinoma classified as pT1. TP expression has also been associated with the extent of invasion in colorectal carcinoma,19 squamous cell carcinoma of the esophagus,30,31 and extrapancreatic neural plexus invasion in ductal adenocarcinoma of the pancreas.32 These findings suggest that TP plays an important role in invasion of various kinds of tumor cells as well as TCC cells. TP expression was also correlated with histologic grade, infiltration pattern, local invasion, and lymph node metastasis in this study, suggesting that TP expression has an important role in tumor growth and metastasis in bladder carcinoma. Tumor vascularity is an important indicator of a poor prognosis in patients with invasive TCC,2 and TP positivity was a predictor of poor prognosis in patients with TCC in our study. Mizutani et al. reported that Histologic grade Local invasion Lymph node metastasis TP expression HRa 95% CIa P valueb 2.33 3.28 2.76 2.68 0.54–9.86 0.92–11.67 1.20–6.33 0.94–7.59 0.230 0.046 0.019 0.047 HR: hazard ratio; CI: confidence interval; TP: thymidine phosphorylase. a Hazard ratios and 95% confidence intervals were obtained from the Cox proportional hazards model, adjusted for variables shown. b P values were obtained by the chi-square test. elevated PD-ECGF/TP expression predicted early recurrence of Ta bladder carcinoma.21 On the other hand, TP expression in bladder carcinomas treated by transurethral resection with or without radiotherapy was reported to have no prognostic value.20 The difference between these findings may be attributed to the different antibodies used, differences in the tumors and the treatments (i.e., radical cystectomy vs. transurethral resection with or without radiotherapy), and/or differences in the cutoff values used to deter- Thymidine Phosphorylase and TCC/Arima et al. mine TP positivity. PD-ECGF/TP expression in bladder carcinomas was previously reported not to be correlated with tumor vascularity.20 However, this does not necessarily mean that PD-ECGF/TP is not correlated with prognosis. In RCC and colorectal carcinoma, TP positivity, but not microvessel density, is an independent prognostic factor.18,19 Recently, we proposed that TP has some roles, other than in angiogenesis, in the progression of solid tumors.19 In an experimental in vitro system, TP conferred resistance to apoptosis induced by hypoxia, and the degradation products of thymidine were involved in the resistance.33 This effect of TP may confer invasive and metastatic properties to tumor cells. Expression of PD-ECGF/TP may play an important role in the progression of invasive TCC; and inhibitors of PD-ECGF/ TP, such as 5-chloro-6-[1-C2-iminopyrrolidinyl) methyl] uracil hydrochloride (TPI),34,35 may suppress the growth and metastasis of the tumors. 13. 14. 15. 16. 17. 18. REFERENCES 1. Folkman J, Shing Y. Angiogenesis. J Biol Chem 1992;267: 10931– 4. 2. Dickinson A, Fox S, Persad R, Hollyer J, Sibley G, Harris A. Quantitation of angiogenesis is an independent predictor of prognosis in invasive bladder cancer. Br J Urol 1994;74: 762– 6. 3. Bochner BH, Cote RJ, Weidner N, Groshen S, Chen SC, Skinner DG, et al. Angiogenesis in bladder cancer: relationship between microvessel density and tumor prognosis. J Natl Cancer Inst 1995;87:1603–12. 4. Friedkin M, Roberts D. The enzymatic synthesis of nucleosides, an efficient method employing nucleoside phosphorylase. J Biol Chem 1954;207:245–56. 5. Krenitsky TA, Koszalka GW, Tuttle JV. Purine nucleoside synthesis, an efficient method employing nucleoside phosphorylases. Biochemistry 1981;20:3615–21. 6. Iltzsch MH, Kouni MH, Cha S. Kinetic studies of thymidine phosphorylase from mouse liver. Biochemistry 1985;24: 6799 – 807. 7. Desgranges C, Razaka G, Rabaud H, Bricaud H. Catabolism of thymidine in human blood platelets: purification and properties of thymidine phosphorylase. Biochim Biophys Acta 1981;654:211– 8. 8. Furukawa T, Yoshimura A, Sumizawa T, Haraguchi M, Akiyama S. Angiogenic factor. Nature 1992;356:668. 9. Usuki K, Saras J, Waltenberger J, Miyazono K, Pierce G, Thomason A, et al. Platelet-derived endothelial cell growth factor has thymidine phosphorylase activity. Biochem Biophys Res Commun 1992;184:1311– 6. 10. Moghaddam A, Bicknell R. Expression of platelet-derived endothelial cell growth factor in Escherichia coli and confirmation of its thymidine phosphorylase activity. Biochemistry 1992;31:12141– 6. 11. Sumizawa T, Furukawa T, Haraguchi M, Yoshimura A, Ishizawa M, Yamada Y, et al. Thymidine phosphorylase activity associated with platelet-derived endothelial cell growth factor. J Biochem 1993;114:9 –14. 12. Haraguchi M, Miyadera K, Uemura K, Sumizawa T, Fu- 19. 20. 21. 22. 23. 24. 25. 26. 27. 1137 rukawa T, Yamada K, et al. Angiogenic activity of enzymes [letter]. Nature 1994;368:198. Pauly JL, Schuller MG, Zelcer AA, Kris TA, Gore SS. Identification and comparative analysis of thymidine phosphorylase in the plasma of healthy subjects and cancer patients: brief communication. J Natl Cancer Inst 1997;58:1587–90. Yoshimura A, Kuwazuru Y, Furukawa T, Yoshida H, Yamada K, Akiyama S. Purification and tissue distribution of human thymidine phosphorylase: high expression in lymophocytes, reticulocytes, and tumors. Biochim Biophys Acta 1990;1034: 107–13. Reynolds K, Farzaneh F, Collins WP, Campbell S, Bourne TH, Lawton F, et al. Association of ovarian malignancy with expression of platelet-derived endothelial cell growth factor. J Natl Cancer Inst 1994;86:1234 – 8. Takebayashi Y, Yamada K, Miyadera K, Sumizawa T, Furukawa T, Kinoshita F, et al. The activity and expression of thymidine phosphorylase in human solid tumours. Eur J Cancer 1996;32:1227–32. Toi M, Hoshina D, Taniguchi T, Yamamoto Y, Ishituka H, Tominaga T. Expression of platelet-derived endothelial cell growth factor/thyimidine phosphorylase in human breast cancer. Int J Cancer 1995;64:79 – 82. Imazono Y, Takebayashi Y, Nishiyama K, Akiba S, Miyadera K, Yamada Y, et al. Correlation between thymidine phosphorylase expression and prognosis in human renal cell carcinoma. J Clin Oncol 1997;15:2570 – 8. Takebayashi Y, Akiyama S, Akiba S, Yamada K, Miyadera K, Sumizawa T, et al. Clinicopathologic and prognostic significance of an angiogenic factor, thymidine phosphorylase, in human colorectal carcinoma. J Natl Cancer Inst 1996;88: 1110 –7. O’Brien TS, Fox SB, Dickinson AJ, Turley H, Westwood M, Moghaddam A, et al. Expression of the angiogenic factor thymidine phosphorylase/platelet-derived endothelial cell growth factor in primary bladder cancers. Cancer Res 1996; 56:4799 – 804. Mizutani Y, Okada Y, Yoshida O. Expression of plateletderived endothelial cell growth factor in bladder carcinoma. Cancer 1997;79:1190 – 4. The Japanese Urological Association, The Japanese Pathological Society and Japanese Radiological Society. General rule for clinical and pathological studies on renal cell carcinoma [in Japanese]. 2nd edition. Tokyo: Kanehara Press, 1992. Hermanek P, Sobin LH, editors. UICC TNM classification of malignant tumors. 4th edition, 2nd revision. Berlin: SpringVerlag, 1992. Bradford A. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72:248 –54. Miyadera K, Sumizawa T, Haraguchi M, Yoshida H, Konstanky W, Yamada Y, et al. Role of thymidine phosphorylase activity in angiogenic effect of platelet-derived endothelial cell growth factor/thymidine phosphorylase. Cancer Res 1995;55:1687–90. Zar JH. Biostatistical analysis. 3rd edition. New Jersey: Prentice Hall, 1996. Nguyen M, Watanabe H, Budson AE, Richie JP, Folkman J. Elevated levels of the angiogenic peptide basic fibroblast growth factor in urine of bladder cancer. J Natl Cancer Inst 1993;85:241–2. 1138 CANCER March 1, 2000 / Volume 88 / Number 5 28. Chopin DK, Caruelle JP, Colombel M, Palcy S, Ravery V, Caruelle D, et al. Increased immunodetection of acidic fibroblast growth factor in bladder cancer, detectable in urine. J Urol 1993;150:1126 –30. 29. O’Brien T, Cranston D, Fuggle S, Bicknell R, Harris AL. Different angiogenic pathways characterize superficial and invasive bladder cancer. Cancer Res 1995;55:510 –3. 30. Igarashi M, Dhar DK, Kubota H, Yamamoto A, El AO, Nagasue N. The prognostic significance of microvessel denstiy and thymidine phosphorylase expression in squamous cell carcinoma of the esophagus. Cancer 1998;82:1225–32. 31. Takebayashi Y, Natsugoe S, Baba M, Akiba S, Furukawa T, Miyadera K, et al. Thymidine phosphorylase in human esophageal squamous cell carcinoma. Cancer 1999;85:282–9. 32. Takao S, Takebayashi Y, Xiangming C, Shinchi H, Natsugoe S, Miyadera K, et al. Expression of thymidine phosphorylase is associated with a poor prognosis in patients with ductal adenocarcinoma of the pancreas. Clin Cancer Res 1998;4: 1619 –24. 33. Kitazono M, Takebayashi Y, Ishitsuka K, Takao S, Tani A, Furukawa T, et al. Prevention of hypoxia-induced apoptosis by the angiogenic factor thymidine phosphorylase. Biochem Biophys Res Commun 1998;253:797– 803. 34. Miyadera K, Suzuki N, Akiyama S, Fukushima M, Yamada Y. Novel functional nucleoside TAS-102, combined from of F3dThd and its modulator (2): inhibitory effect of TPI on tumor-derived angiogenesis and metastasis. Proc Am Assoc Cancer Res 1998;39:609. 35. Matsushita S, Nitanda T, Furukawa T, Sumizawa T, Tani A, Nishimoto K, et al. The effect of thymidine phosphorylase inhibitor on angiogenesis and apoptosis in tumors. Cancer Res 1999;59:1911– 6.