Clinicopathological and prognostic significance of thymidine phosphorylase and proliferating cell nuclear antigen in gastric carcinoma

Cancer Letters 166 (2001) 103±111

www.elsevier.com/locate/canlet

Clinicopathological and prognostic signi®cance of thymidine phosphorylase and proliferating cell nuclear antigen in gastric carcinoma Soichi Konno a, Yuji Takebayashi b,*, Motohiko Aiba c, Shin-ichi Akiyama d, Kenji Ogawa a b

a Department of Surgery, Tokyo Women's Medical University Daini Hospital, 2-1-10 Nishioku, Arakawaku, Tokyo 116, Japan Department of Pathology, Institute of Development, Aging & Cancer Tohoku University, 4-1 Seiryo-machi, Aobaku, Sendai 980-8575, Japan c Surgical Pathology, Tokyo Women's Medical University Daini Hospital, 2-1-10 Nishioku, Arakawaku, Tokyo 116, Japan d Department of Cancer Chemotherapy, Institute for Cancer Research, Faculty of Medicine, Kagoshima University, Sakuragaoka 8-35-1, Kagoshima 890, Japan

Received 12 December 2000; received in revised form 30 January 2001; accepted 30 January 2001

Abstract We have reported that thymidine phosphorylase (dThdPase) is identical to platelet derived-endothelial cell growth factor and that it has an angiogenic activity in vitro and in human carcinoma tissues as well as gastric carcinoma. Recently, we revealed that dThdPase may have an another function(s) besides angiogenesis in vitro and in human solid tumors. Using immunohistochemistry, we examined retrospectively whether the expression of dThdPase was correlated with tumor growth, comparing it with the proliferating cell nuclear antigen labeling index (PCNA LI) and examining their prognostic signi®cance in 116 patients with gastric carcinoma. A direct correlation of these two factors was observed (R ˆ 0:659, P , 0:001). A multivariate Cox regression analysis showed that both dThdPase positivity and PCNA LI were independent prognostic factors, as were depth of invasion and lymph node metastasis. Furthermore, the patients with dThdPase-positive/high PCNA LI tumors had the worst prognoses. The combination of dThdPase and PCNA expression is a better tool for predicting the prognosis of patients with gastric carcinoma than the expression of either of them alone. These results raise the possibility that dThdPase may have a function(s) involved in tumor growth besides angiogenesis. q 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Thymidine phosphorylase; Proliferating cell nuclear antigen; Gastric carcinoma; Prognostic factor

1. Introduction Thymidine phosphorylase (dThdPase; EC 2.4.2.4) catalyzes the reversible phosphorolysis of thymidine, deoxyuridine and their analogs to their respective bases and 2-deoxyribose-1-phosphates [1,2]. dThdPase also * Corresponding author. Tel.: 1 81-22-717-8511; fax: 181-22717-8512. E-mail address: [email protected] (Y. Takebayashi).

catalyzes the transfer of deoxyribose from one deoxynucleoside to another base to form a second deoxynucleoside [3,4]. In mammals, dThdPase consists of two identical subunits, each with a molecular weight of 55 kDa [5]. dThdPase is identical to platelet derivedendothelial cell growth factor (PD-ECGF) [6]. PDECGF stimulates chemotaxis and [ 3H] thymidine incorporation by endothelial cells in vitro and has angiogenic activity in vivo [7,8]. Recently, we have

0304-3835/01/$ - see front matter q 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0304-383 5(01)00432-3

104

S. Konno et al. / Cancer Letters 166 (2001) 103±111

demonstrated that the enzymatic activity of dThdPase is indispensable for its angiogenic activity [9]. Among the degradation products from the action of dThdPase on thymidine, 2-deoxy-D-ribose, a dephosphorylated product derived from 2-deoxy-D-ribose-1-phosphate, which had chemotactic activity in vitro and angiogenic activity in vivo [7]. These ®ndings suggested that the enzymatic products stimulate the chemotaxis of endothelial, and possibly, other cells, causing angiogenesis. The transfection of PD-ECGF/dThdPase into transformed ®broblasts in nude mice resulted in increased tumor vascularization [7]. Increased dThdPase expression correlates with tumor microvessel density in human breast cancer [10]. Areas of high blood velocity in ovarian tumors were associated with increased expression of PD-ECGF/dThdPase [11]. These observations suggested that dThdPase is really involved in angiogenesis in some human tumors. When compared with adjacent normal tissues, higher levels of dThdPase were observed in a variety of malignant tumors [10±19]. There are experimental evidences showing that tumor growth is dependent on angiogenesis. When tumors reach a size of a few millimeters, they are penetrated by new capillaries allowing for rapid tumor growth. Since these new vessels facilitate the entry of tumor cells into the vasculature, leading to metastasis, angiogenesis correlates with the probability of metastases [20,21]. Angiogenesis in human solid tumors is also a risk factor for metastasis and tumor recurrence [22±26]. The expression of proliferating cell nuclear antigen (PCNA) correlates with the late G1 and S phases of the cell cycle [27±29]. The prognostic signi®cance of PCNA expression as a marker of tumor growth has been investigated in a wide variety of human solid tumors, such as gastric carcinoma, breast cancer, colorectal carcinoma, lung carcinoma and renal cell carcinoma [30±33]. dThdPase seems to have another function besides angiogenesis in esophageal, colorectal and renal cell [15,17,19]. Recently, we demonstrated that dThdPase suppresses apoptosis induced by hypoxia in vitro [34]. PCNA plays an essential role in nucleic acid metabolism as a component of the replication and repair machinery [35]. The aim of this study was to investigate the property of dThdPase in tumor growth and the prognostic signi®cance of it using and comparing PCNA expression in gastric carcinoma.

2. Materials and methods 2.1. Patients and tumors The features of the 116 patients with gastric carcinoma investigated in this study are summarized in Table 1. All patients evaluated underwent surgical resection, between February 1986 through December 1992, in the Department of Surgery of Tokyo Women's University Daini Hospital. None had received prior chemotherapy or irradiation therapy. Patients received postoperative therapy; however, there was no difference in the outcome among the various treatment modalities. Of the 116 patients, 66 were males and 50 were females, and their average age was 59.7 years, ranging from 20 to 89 years. Histological typing and grading of tumors was performed according to the criteria established by the Japanese General Rules for Gastric Cancer Study [36]. Tumor specimens were collected after obtaining informed consent in accordance with institutional guidelines. The deepest invasive sites were selected for immunohistochemistry of dThdPase and PCNA. 2.2. Immunohistochemical detection and evaluation of dThdPase and PCNA Sections were deparaf®nized with xylene and dehydrated with 98% ethanol. Endogenous peroxidase was blocked by immersing the slides in 0.3% hydrogen peroxide in absolute methanol for 20 min at room temperature. After washing three times for 5 min each with PBS, the sections were blocked by soaking for 20 min at room temperature in PBS containing 1% bovine serum albumin. The blocked sections were incubated overnight at 48C with 0.5 mg/ml of a monoclonal antibody against dThdPase which was described previously [12,13,15,17±19]. After 12 h, the slides were incubated for 30 min at room temperature with biotinylated anti-mouse IgG diluted 100fold with PBS, washed three times for 15 min in PBS, and incubated for 30 min with avidin±biotin± peroxidase complex diluted with PBS [37]. After three washes for 15 min with PBS, the sections were incubated for 7 min with 0.5 mg/ml diaminobenzidine and 0.03% (v/v) H2O2 in PBS, and ®nally, counterstained with hematoxylin prior to mounting. Serial

S. Konno et al. / Cancer Letters 166 (2001) 103±111

105

Table 1 Relationship between dThdPase expression, PCNA LI and clinicopathological variables in gastric carcinomas a n

DThdPase positivity

PCNA Li

Variables

Negative (n ˆ 66)

Positive (n ˆ 50)

P value b,c

Low (n ˆ 75)

High (n ˆ 41)

P value b,c

Age Gender Male Female Histopathological type Differentiated Undifferentiated Depth of invasion t1 t2 t3 Lymph node metastasis Negative Positive Lymphatic invasion Negative Positive Venous invasion Negative Positive Clinical stage I II III

58.2

61.2

NS NS

58.2

61.5

NS NS

a b c

72 44

40 (55.6) 26 (59.1)

32 (44.4) 18 (40.9)

47 69

26 (55.3) 40 (57.9)

21 (44.7) 29 (42.1)

30 44 42

24 (80.0) 25 (56.8) 17 (40.5)

6 (20.0) 19 (43.2) 25 (59.5)

54 62

38 (70.4) 28 (45.2)

16 (29.6) 34 (54.8)

39 77

32 (82.1) 34 (44.2)

7 (17.9) 43 (55.8)

78 38

50 (64.1) 16 (42.1)

28 (35.9) 22 (57.9)

42 28 46

31 (73.8) 18 (64.3) 17 (36.9)

11 (26.2) 10 (35.7) 29 (63.1)

NS P , 0.05

P , 0.01 P , 0.05 P , 0.05 P , 0.01

45 (62.5) 30 (68.2)

27 (37.5) 14 (31.8)

30 (63.8) 45 (65.2)

17 (36.2) 24 (34.8)

23 (76.7) 31 (70.5) 21 (50.0)

7 (23.3) 13 (29.5) 21 (50.0)

39 (72.2) 36 (58.1)

15 (27.8) 26 (41.9)

31 (79.5) 44 (57.1)

8 (20.5) 33 (42.9)

55 (70.5) 20 (52.6)

23 (29.5) 18 (47.4)

33 (78.6) 20 (71.4) 22 (47.8)

9 (21.4) 8 (28.6) 24 (52.2)

NS P , 0.05

NS P , 0.05 NS P , 0.02

Figures in parentheses represent percentage values. P values were conducted by the Chi-square test. NS, not signi®cant.

sections were also incubated with rabbit antiserum against human PCNA (PC-10, Nobocastra Lab., UK), diluted at 1:100 with PBS containing 5% goat serum. Antibody binding was detected by sequential incubation with biotinylated goat anti-rabbit serum. Other procedures were the same as those for dThdPase immunohistochemical staining. 2.3. Evaluation of the stained sections For microscopic analysis, we examined at least 200 carcinoma cells to determine whether the cells were positive for dThdPase at high power (£400) after screening for the areas with the highest intensity of staining at lower power (£100). When more than 5% of the carcinoma cells were stained, the specimen was de®ned as positive. For PCNA evaluation, we examined at least 200 nuclei to assess whether they were

positive for PCNA staining at high power (£400) after screening for areas with the highest intensity of staining at low power (£100). The proliferating cell nuclear antigen labeling index (PCNA LI) was de®ned as the percentage of cells examined which were positive. The high PCNA LI group was de®ned as that which displayed more than 50% stained cells and the low PCNA LI group as that which showed 50% or less stained cells. Immunohistochemical evaluation was performed by two investigators (S.K. and K.O.) independently without prior information of the clinicopathological features. 2.4. Statistical analysis Demographic and clinicopathological characteristics were compared between patients with dThdPase-positive and -negative tumors using the

106

S. Konno et al. / Cancer Letters 166 (2001) 103±111

Fig. 1. Immunostaining of gastric carcinoma was shown: (A), thymidine phosphorylase on a £200 ®eld; and (B), PCNA on a £200 ®eld.

S. Konno et al. / Cancer Letters 166 (2001) 103±111

107

type, lymph node metastasis or venous invasion. In contrast, PCNA LI and dThdPase positivity were correlated with depth of invasion, lymphatic invasion and clinical stage. Furthermore, a signi®cant correlation was also observed between dThdPase positivity and lymph node metastasis/venous invasion. 3.2. Coexpression of dThdPase and PCNA in gastric carcinoma

Fig. 2. The percentage of cells expressing PCNA as a function of the percentage of cells expressing dThdPase. Overlap accounts for less than 116 points appearing on the graph. The P value was calculated by the Spearman test.

The relationships between clinicopathological variables and dThdPase positivity or PCNA LI indicated the similar pro®les in depth of invasion, lymphatic invasion and clinical stage (Table 1). In order to analyze the relationship between dThdPase expression and PCNA LI more precisely, the percentage of cells expressing PCNA was plotted against those

Chi-square test or the Student's t-test. The Cox proportional hazards model was used in the multivariate survival analysis. Maximum likelihood parameter estimates and likelihood ratio statistics (LRS) in Cox proportional hazards models were obtained using a statistical package, EPICURE. We calculated Wald type con®dence intervals. Testing for statistical interaction was conducted by including a cross-product term of the two variables of interest in a model. All P values presented are two-sided.

3. Results 3.1. Expression of dThdPase and PCNA in gastric carcinoma The cytoplasm of many adenocarcinoma cells, as well as stromal cells that might be macrophages or ®broblasts, expressed dThdPase (Fig. 1A). In contrast, PCNA staining was con®ned to the nuclei in neoplastic cells. The mitotic features were highlighted by PCNA antibody (Fig. 1B). Expressions of both dThdPase and PCNA are associated with the extension of gastric carcinoma. Table 1 summarizes the correlation between clinicopathological variables and dThdPase/PCNA positivity in gastric carcinomas. No signi®cant correlation was found between dThdPase positivity and age, gender or histopathological type. We did not ®nd any signi®cant association between PCNA LI and age, gender, histopathological

Fig. 3. Survival curves by the Kaplan±Meier method: (A), between the patients with dThdPase-positive tumors and those with negative; and (B), between those with high and low PCNA tumors. P values were calculated by a generalized Wilcoxon test. All P values are two-sided.

108

S. Konno et al. / Cancer Letters 166 (2001) 103±111

Table 2 Results from multivariate Cox's proportional hazard model for survival times of 116 patients with gastric carcinomas a Variables (category) Depth of invasion t1 t2 t3 Lymph node metastasis Negative Positive dThdPase expression Negative Positive PCNA LI High ($50%) Low (,50%)

Hazard ratio

95% Con®dence interval

P value

1.696

1.010±2.847

0.0455

1.537

1.042±2.268

0.0303

3.056

1.506±6.205

0.0020

1.051

1.051±3.588

0.0341

a

The hazard ratio, 95% con®dence intervals (CIs), and two-sided P values were obtained from the Cox proportional hazards model. Covariables were: depth of invasion, lymph node metastasis, dThdPase expression, and PCNA LI.

Fig. 4. Survival curve by the Kaplan±Meier method among the patients with dThdPase-positive/high PCNA LI tumors, dThdPase-positive/low PCNA LI, dThdPase-negative/high PCNA LI and dThdPase-negative/low PCNA LI tumors. P values were calculated by a generalized Wilcoxon test. P values are two-sided.

S. Konno et al. / Cancer Letters 166 (2001) 103±111

expressing dThdPase. Fig. 2 showed the direct correlation of these two factors (R ˆ 0:659, P , 0:001). 3.3. Prognostic relevance and proportional hazards analysis Kaplan±Meier survival curves showed that the patients with dThdPase-positive gastric carcinomas or high PCNA LI tumors had a worse prognosis than those with negative dThdPase or low PCNA LI tumors (P , 0:001, P , 0:05), respectively (Fig. 3). Next, we wished to determine independent prognostic values on patient survival; a multivariate Cox regression model was constructed using dThdPase positivity, PCNA LI or established novel prognostic factors, such as depth of invasion and lymph node metastasis in clinic (Table 2). The analysis using this model demonstrated that both dThdPase positivity and PCNA LI were independent prognostic factors (hazard ratio, 3.056, 95% CI, 1.506±6.205, P ˆ 0:002; hazard ratio, 1.051, 95% CI, 1.051± 3.588, P ˆ 0:0341, respectively) as well as depth of invasion and lymph node metastasis. Furthermore, we investigated whether the combination of dThdPase positivity and PCNA LI could predict the prognosis of patients with gastric carcinomas or not, using Kaplan±Meier analysis after the categorization of patients such as dThdPase-positive/high PCNA LI, dThdPase-positive/low PCNA LI, dThdPase-negative/high PCNA LI and dThdPase-negative/low PCNA LI. Fig. 4 demonstrates that the patients with dThdPase-positive/high PCNA LI tumors showed the worst prognosis among these combinations. These data indicated that the combination of dThdPase and PCNA expression was a better tool for predicting the prognosis of patients with gastric carcinoma. 4. Discussion The present study revealed signi®cant correlations between dThdPase expression and several clinicopathological features of gastric carcinomas. The proportion of dThdPase-positive cells was higher in carcinomas with deep invasion and/or lymph node metastases (Table 1). Angiogenesis is needed for rapid tumor growth and the vascularized tumor extends vertically into the deep tissues beneath the basement membrane. The new proliferating capil-

109

laries disrupt the basement membrane and are more easily penetrated by tumor cells than mature vessels. During the vascular phase, tumor cells may be shed into the circulation [21]. The clinical signi®cance of this has been documented in studies of breast [23], prostate [24], lung [25], gastric [26], and colorectal carcinomas [27], where microvessel density has been shown to be correlated with metastases and recurrence of the disease. With this as a model, the present study suggests that increased dThdPase level in gastric carcinomas may enhance the invasiveness and the ability of the tumors to metastasize by virtue of its angiogenic properties. We examined the expression of PCNA in gastric cancer in this study to elucidate the association between carcinoma cell proliferation and dThdPase expression, since PCNA is a marker of cell proliferation in vitro, and a prognostic marker in some human solid tumors as well as in gastric carcinoma [27±33]. Our results (Fig. 2) demonstrate a direct correlation between dThdPase and PCNA expression in gastric carcinoma. dThdPase also has an effect on tumor growth other than angiogenesis in head and neck squamous cell carcinoma [19]. Furthermore, dThdPase expression suppresses apoptosis induced by hypoxia in dThdPase-transfected KB 3-1 cells [34]. This phenomenon was also con®rmed in human colon and gastric carcinomas [38,39]. Recently, the suppression of cyclin D1 expression after the stable transfection of a cyclin D1 antisense construct in PANC-1 and COLO-357 human pancreatic cancer cells resulted in a decrease in thymidilate synthetase and an increase in dThdPase mRNA expression, as determined by reverse transcriptionpolymerase chain reaction (RT-PCR) [35]. However, the mechanism(s) is not still clear between apoptosis and dThdPase expression in carcinoma cells. Taken together, dThdPase may have an another role(s), such as cell cycle, DNA replication or repair, besides angiogenesis. Further study is needed to clarify these speculated effects of dThdPase besides angiogenesis on tumor growth in human solid tumors. The present study showed that dThdPase and PCNA expression in gastric carcinoma were independent prognostic factors in a multivariate Cox proportional hazards regression model. In esophageal [18], colorectal [15] and renal cell carcinoma [17], similar ®ndings were observed. We have shown that dThdPase has an effect on angiogenesis and prognosis

110

S. Konno et al. / Cancer Letters 166 (2001) 103±111

of patients with gastric carcinoma [13]. Maeda et al. also reported the relationship between dThdPase expression and hepatic metastasis and its prognostic signi®cance in gastric carcinoma [14]. In contrast, Tanigawa et al. reported that the expression of dThdPase was not associated with the overall survival of patients with gastric carcinoma [16]. This discrepancy may be due to interlaboratory differences in tissue ®xation processing, and staining techniques, experience of the observer in selecting hot spot, the technique of counting the dThdPase-positive cells and the antibodies used in each study. Recently, dThdPase activity and expression have been reported to be involved in the invasiveness of gastric carcinoma from other laboratories [40,41]. With these ®ndings together, dThdPase expression may be involved in the prognosis of patients with gastric carcinoma. Interestingly, the combination of dThdPase and PCNA expression can be a better tool for predicting the prognosis of patients with gastric carcinoma rather than either of them alone. This suggests that the evaluation of both molecules could better predict the prognosis of patients with gastric carcinomas. A dThdPase inhibitor, 6 amino-5-chrolouracil, inhibited angiogenesis induced by dThdPase in a CAM assay and a mouse gelatin sponge implant assay [7,9]. Recently, Takao et al. demonstrated that inhibition of dThdPase suppressed metastasis of dThdPase-transfected KB-3-1 cells in vivo [42]. Taken together with this evidence and our results, the growth of tumor expressing dThdPase may be inhibited by suppressing dThdPase expression. Acknowledgements The authors would like to thank Dr Tatsuhiko Fukuiwa (Kagoshima University, Japan) for his critical reading of our manuscript.

[4] [5] [6] [7] [8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

References [16] [1] M.H. Iltzssch, M.H. Kouni, S. Cha, Kinetic studies of thymidine phosphorylase from mouse liver, Biochemistry 24 (1985) 6799±6807. [2] D. Friedkin, D. Roberts, The enzymatic synthesis of nucleosides, J. Biol. Chem. 207 (1954) 245±256. [3] T.A. Krenistsky, G.W. Koszalka, J.V. Tuttle, Purine nucleo-

[17]

side synthesis, an ef®cient method employing nucleoside phosphorylase, Biochemistry 20 (1981) 3615±3621. M. Zimmerman, J. Seidenberg, Deoxyribosyl transfer, J. Biol. Chem. 230 (1964) 2618±2621. R.C. Gallo, S. Perry, T.R. Breitman, The enzymatic mechanisms for deoxythymidine synthesis in human leukocytes, J. Biol. Chem. 242 (1967) 5059±5068. T. Furukawa, A. Yosimura, T. Sumizawa, M. Haraguchi, S. Akiyama, K. Fukui, M. Ishizawa, Y. Yamada, Angiogenic factor, Nature 356 (1992) 668. M. Haraguchi, K. Miyadera, K. Uemura, T. Sumizawa, T. Furukawa, K. Yamada, S. Akiyama, Y. Yamada, Angiogenic activity of enzymes, Nature 368 (1994) 198. F. Ishikawa, K. Miyazono, U. Hellman, M. Drexler, C. Wernstedt, K. Hagiwara, K. Usuki, F. Takaku, W. Risau, C.H. Heldin, Identi®cation of angiogenic activity and the cloning and expression of platelet-derived endothelial cell growth factor, Nature 338 (1989) 556±562. K. Miyadera, T. Sumizawa, M. Haraguchi, H. Yoshida, W. Konstanty, Y. Yamada, S. Akiyama, Role of thymidine phosphorylase activity in the angiogenic effect of platelet-derived endothelial cell growth factor/thymidine phosphorylase, Cancer Res. 55 (1995) 1687±1690. M. Toi, S. Hoshina, T. Taniguchi, Y. Yamamoto, H. Ishituka, T. Tominaga, Expression of platelet-derived endothelial cell growth factor/thymidine phosphorylase in human breast cancer, Int. J. Cancer 14 (1995) 79±82. K. Reynold, F. Farzaneh, W.P. Collins, A.L. Harris, Association of ovarian malignancy with expression of platelet-derived endothelial cell growth factor, J. Natl. Cancer Inst. 86 (1995) 1234±1238. Y. Takebayashi, K. Yamada, K. Miyadera, T. Sumizawa, T. Furukawa, F. Kinoshita, D. Aoki, H. Okumura, Y. Yamada, S. Akiyama, T. Aikou, The activity and expression of thymidine phosphorylase in human solid tumors, Eur. J. Cancer 32 (1996) 1227±1232. Y. Takebayashi, K. Miyadera, S. Akiyama, S. Hokita, Y. Yamada, S. Akiba, Y. Yamada, T. Sumizawa, T. Aikou, Expression of thymidine phosphorylase in human gastric carcinoma, Jpn. J. Cancer Res. 87 (1995) 288±295. K. Maeda, Y.S. Chung, Y. Ogawa, S. Takatsuka, S.M. Kang, M. Ogawa, T. Sawada, N. Onda, Y. Kato, M. Sowa, Thymidine phosphorylase/platelet-derived endothelial cell growth factor expression associated with hepatic metastasis in gastric carcinoma, Br. J. Cancer. 73 (1996) 884±888. Y. Takebayashi, S. Akiyama, S. Akiba, K. Yamada, K. Miyadera, T. Sumizawa, Y. Yamada, F. Murata, T. Aikou, Clinicopathologic and prognostic signi®cance of an angiogenic factor, thymidine phosphorylase, in human colorectal carcinoma, J. Natl. Cancer Inst. 88 (1996) 1110±1117. N. Tanigawa, H. Amaya, M. Matsmura, Y. Katoh, A. Kitaoka, T. Aotake, T. Shimomatsya, O.A. Rosenwasser, M. Iki, Tumor angiogenesis and expression of thymidine phosphorylase/platelet derived endothelial cell growth factor in human gastric carcinoma, Cancer Lett. 108 (1996) 281±290. Y. Imazono, Y. Takebayashi, K. Nishiyama, S. Akiba, K. Miyadera, Y. Yamada, S. Akiyama, Y. Ohi, Correlation

S. Konno et al. / Cancer Letters 166 (2001) 103±111

[18]

[19]

[20] [21]

[22] [23] [24] [25]

[26] [27]

[28] [29] [30] [31]

between thymidine phosphorylase expression and prognosis in human renal cell carcinoma, J. Clin. Oncol. 15 (1997) 2570± 2578. Y. Takebayashi, S. Natsugoe, M. Baba, S. Akiba, T. Fukumoto, K. Miyadera, Y. Yamada, S. Takao, S. Akiyama, T. Aikou, Thymidine phosphorylase in esophageal squamous carcinoma, Cancer 85 (1999) 282±289. T. Fukuiwa, Y. Takebayashi, S. Akiba, T. Matsuzaki, Y. Hanamure, K. Miyadera, Y. Yamada, S. Akiyama, Expression of thymidine phosphorylase and vascular epithelial cell growth factor in human head and neck squamous cell carcinoma and their different characteristics, Cancer 85 (1999) 960±969. J. Folkman, What is the evidence that tumors are angiogenesis dependent? J. Natl. Cancer Inst. 82 (1990) 4±6. L. Liotta, J. Kleinerman, G. Saidel, Quantitative relationships of intravascular tumor cells, tumor vessels, and pulmonary metastases following tumor implantation, Cancer Res. 34 (1974) 997±1004. N. Weidner, J.P. Semple, W.R. Welch, J. Folkman, Tumor angiogenesis and metastasis ± correlation in invasive breast carcinoma, N. Engl. J. Med. 324 (1991) 1±8. N. Weidner, P.R. Carroll, J. Flax, W. Blumenfeld, J. Folkman, Tumor angiogenesis correlates with metastasis in invasive prostate carcinoma, Am. J. Pathol. 143 (1993) 401±409. P. Macchiarini, G. Fontanini, M.J. Hardin, F. Squartini, C.A. Angeletti, Relation of neovascularization to metastasis of nonsmall-cell lung cancer, Lancet 340 (1992) 145±146. K. Maeda, Y.S. Chung, S. Takatsuka, Y. Ogawa, T. Sawada, Y. Kondo, M. Sowa, Tumor angiogenesis as a predictor of recurrence in gastric carcinoma, J. Clin. Oncol. 13 (1995) 477±481. Y. Takebayashi, S. Akiyama, K. Yamada, S. Akiba, T. Aikou, Angiogenesis as an unfavorable prognostic factor in human colorectal carcinoma, Cancer 78 (1996) 226±231. J.E. Celis, A. Celis, Cell cycle-dependent variations in the distribution of the nuclear protein cyclin proliferating cell nuclear antigen in cultured cell: subdivision of S phase, Proc. Natl. Acad. Sci. USA 82 (1985) 3262±3266. G.F. Morris, M.B. Mathews, Regulation of proliferating cell nuclear antigen during the cell cycle, J. Biol. Chem. 264 (1985) 13856±13864. K.S. Moore, K. Sullivian, E.M. Tan, M.B. Prystowaky, Proliferating cell nuclear antigen/cyclin is an interleukin 2-responsive gene, J. Biol. Chem. 264 (1987) 8477±8480. K. Maehara, Y. Emi, H. Baba, K. Adachi, Y. Ichiyoshi, K. Akazawa, K. Sugmicachi, Recurrences and related characteristics of gastric cancer, Br. J. Cancer 74 (1996) 975±979. M. Mori, Y. Kakeji, S. Moriguchi, Y. Mahara, K. Sugimachi,

[32]

[33]

[34]

[35] [36] [37]

[38]

[39]

[40]

[41]

[42]

111

J.M. Jessup, L.B. Chen, G.D. Steel Jr, The prognostic signi®cance of proliferating cell nuclear antigen in clinical gastric cancer, Surgery 113 (1993) 683±690. M.E. Harper, J.B. Glynne, L. Goddard, D.W. Wilson, S.S. Matenhelia, I.G. Conn, W.B. Peeling, K. Grif®ths, Relationship of proliferating cell nuclear antigen (PCNA) in prostatic carcinomas to various clinical parameters, Prostate 20 (1992) 243±247. B.A. Robbins, D. del la Vega, K. Ogata, E.M. Tan, R.M. Nakamura, Immunohistochemical detection of proliferating cell nuclear antigen in solid human malignancies, Arch. Pathol. Lab. Med. 111 (1987) 841±845. M. Kitazono, Y. Takebayashi, K. Ishitsuka, S. Takao, A. Tani, T. Furukawa, K. Miyadera, Y. Yamada, T. Aikou, S. Akiyama, Prevention of hypoxia-induced apoptosis by the angiogenic factor thymidine phosphorylase, Biochem. Biophys. Res. Commun. 253 (1998) 797±803. Z. Kelman, PCNA: structure, functions and interactions, Oncogene 14 (1997) 629±640. Japanese Research Society for Gastric Cancer, The General Rules for Gastric Cancer Study, 12th Edition, Kanehara Public Co, Tokyo, 1993. S.M. Hsu, L. Raine, A.H. Fanger, A comparative study of the peroxidase method and an avidin±biotin complex method for studying polypeptide hormones with radioimmunoassay antibody, Am. J. Clin. Pathol. 75 (1981) 734±738. T. Matsuura, I. Kuratate, K. Teramachi, M. Osaki, Y. Fukuda, H. Ito, Thymidine phosphorylase expression is associated with both increase of intratumoral microvessels and decrease of apoptosis in human colorectal carcinomas, Cancer Res. 59 (1999) 3505±3511. M. Osaki, T. Sakatani, E. Okamoto, E. Goto, H. Adachi, H. Ito, Thymidine phosphorylase expression results in a decrease in apoptosis and increase in intratumoral microvessel density in human gastric carcinomas, Virchows Arch. 437 (2000) 31± 60. S. Shimaoka, S. Matsshita, T. Nitanda, A. Matsuda, T. Niho, T. Suenaga, Y. Nishimata, S. Akiba, S. Akiyama, H. Nishimata, The role of thymidine phosphorylase expression in the invasiveness of gastric carcinoma, Cancer 88 (2000) 2220± 2227. B. Nakata, K. Hirakawa-Ys Chung, Y. Kato, N. Onoda, K. Maeda, M. Sowa, Thymidine phosphorylase activity in tumor correlates with venous invasion, Int. J. Mol. Med. 2 (1998) 461±464. S. Takao, S. Akiyama, A. Nakajo, H. Yoh, M. Kitazono, S. Natsugoe, K. Miyadera, M. Fukushima, Y. Yamada, T. Aikou, Suppression of metastasis by thymidine phosphorylase inhibitor, Cancer Res. 60 (2000) 5345±5348.