Correlation between tumor angiogenesis and expression of thymidine phosphorylase, and patient outcome in uterine cervical carcinoma

Correlation Between Tumor Angiogenesis and Expression of Thymidine Phosphorylase, and Patient Outcome in Uterine Cervical Carcinoma MASATSUGU UEDA, MD, YOSHITO TERAI, MD, KOJI KUMAGAI, MD, KEN UEKI, MD, YOSHIAKI OKAMOTO, MD, AND MINORU UEKI, MD The relationship between tumor angiogenesis and the expression of thymidine phosphorylase (dThdPase) in uterine cervical carcinoma tissues, as well as patient outcome, were investigated. Primary tumor specimens surgically obtained from 54 patients (stages Ib to IIIb) receiving neither chemotherapy nor radiation therapy before surgery were examined. Intratumoral microvessel density (IMVD) and dThdPase expression were evaluated immunohistochemically using anti-CD34 and anti-dThdPase antibodies and were correlated with clinicopathologie parameters and prognosis. IMVD for the 54 tumors ranged from 24.8 to 118.6, with a median value of 57.7 (number/0.7386 nun2/field). For immtmoreactivity of dThdPase, 16 tumors were graded as - , 20 as 1+, and 18 as 2+. IMVD was significantly associated with the expression of dThdPase (P < .01). Both IMVD and dThdPase expression were well correlated with depth of myometrial invasion, endometrial invasion, and pelvic lymphnode

metastasis (P < .05). Overall survival rates for 18 patients with strong dThdPase-staining tumors, assessed as 2+, were lower than those for 36 with weak dThdPase staining tumors (P = .0108). However, there was no statistical correlation between IMVD and patient outcome. The expression of dThdPase plays a role in the promotion of angiogenesis and affects the patient's survival in uterine cervical carcinomas. HUM PATHOL30:1389-1394. Copyright © 1999 by W.B. Saunders Company Key words: angiogenesis, thymidine phosphorylase, prognosis, uterine cervical carcinoma. Abbreviations: VEGF, vascular endothelial growth factor; bFGF, basic fibroblast growth factor; TGF-cx, transforming growth factor-m; dThdPase, thymidine phosphorylase; PD-ECGF, platelet-derived endothelial cell growth factor; IMVD, intratumoral microvessel density; 5'-dFUrd, 5'-deoxy-5-fluorouridine; 5-FUra, 5-fluorouracil.

T u m o r angiogenesis is essential for t u m o r development. Prevascular tumors may remain d o r m a n t in situ for months to years, and the switching from a subgroup of prevascular t u m o r cells to an angiogenic phenotype enables rapid growth, progression, and metastasis. 1 T h e intratumoral microvessel c o u n t has been used to evaluate angiogenic activity in a variety of human solid rumors, including b r e a s t y prostate,4 gastric,5,6 colorectal,7 cervical, 8-13 endometrial, 14-16and ovarian 17-19carcinomas. Increased microvessel density appears to adversely affect prognosis, suggesting that angiogenic properties are correlated with t u m o r aggressiveness. 2,4-7,1°,12-15,17,1s Angiogenesis is induced by factors p r o d u c e d by the t u m o r or the nonmalignant cells that infiltrate the tumor. Various peptide growth factors such as vascular endothelial growth factor (VEGF), 2° basic fibroblast growth factor (bFGF),21 and transforming growth factor-e~ (TGF-oL)22 have been f o u n d to stimulate the proliferation and motility of endothelial cells, thus inducing new blood vessel formation. Although a number of angiogenic factors have been identified, 23 the angiogenic factors specific to t u m o r types remain unknown. A recent study has shown that thymidine phosphorylase (dThdPase) is almost identical to plateletderived endothelial cell growth factor (PD-ECGF). 24 dThdPase, an enzyme involved in pyrimidine nucleo-

tide metabolism, has been shown to have angiogenic activity.24,25 Haraguchi et aF 6 r e p o r t e d that dThdPase, through the degradation products of thymidine, stimulates the chemotaxis of endothelial cells and possibly other cells and thus indirectly stimulates angiogenesis. Several workers have f o u n d that the expression of dThdPase shows a significant correlation with t u m o r angiogenesis, a,5,16,19Very recently, we also reported that intraarterial infusion c h e m o t h e r a p y induced apoptotic cell death in locally advanced cervical carcinomas through the inhibition of t u m o r angiogenesis and dThdPase expression of tumor cells. 27 However, as far as we are aware, there has been no r e p o r t describing the possible association a m o n g t u m o r angiogenesis, dThdPase expression, and patient outcome in cervical carcinomas. In this study, we examined intratumoral microvessel density (IMVD) and the expression of dThdPase protein in cervical carcinoma tissues, and investigated the relationship between them with respect to clinicopathologic features and prognosis.

From the Department of Obstetrics and Gynecology, Osaka Medical College, Osaka, Japan. Accepted for publication July 29, 1999. Address correspondence and reprint requests to Masatsugu Ueda, MD, Department of Obstetrics and Gynecology, Osaka Medical College, 2-7, Daigakumachi, Takatsuki, Osaka 569-8686,Japan. Copyright © 1999 by W.B. Sannders Company 0046-8177/99/3011-0020510.00/0

MATERIALS AND METHODS

Patients and Tissue Samples A total of 54 primary cervical carcinomas, which had been resected in our department from August 1988 to January 1997, were used in this study. Patients had received neither chemotherapy nor radiation therapy before surgery. All patients underwent radical hysterectomy with bilateral adnexectomy and bilateral pelvic lymphadenectomy. Clinical stages and pathological diagnosis were decided according to the classification of the International Federation of Gynecology and Obstetrics. 2s Of these patients, 19 had stage Ib, 4 had stage 2a, 29 had stage 2b, 1 had stage 3a, and 1 had stage 3b

1389

HUMANPATHOLOGY Volume30, No. 11 (November 1999) disease. When positive surgical margins, parametrium involvement, or pelvic lymphnode metastasis were found in the histological sections after surgery, full pelvic and intracavitary radiation therapy was performed as a standard additional treatment. Data concerning patient outcome, including overall survival and development of metastasis, were available for all 54 patients. All were reviewed at 2- or 3-monthly intervals in the outpatient clinic. Those patients who failed to attend the clinic were traced by telephone. If the patient had died, the data and causes of death were recorded. Tumor specimens were fixed in 10% buffered formalin and embedded in paraffin wax. Histological features in resected tumors were assessed using standard hematoxylin and eosin-stained sections. Serial sections including the greatest diameter of the tumors from the operative specimens were used for the present study.

Determination of IMVD Immunohistochemical staining for CD34, to highlight endothelial cells] was performed using the avidin-biotinperoxidase complex method. Briefly, dewaxed and rehydrated tissue sections were incubated overnight at 4°C with mouse monoclonal anti-CD34 antibody (QB-END/10, Novocastra Laboratory, Newcastle, UK) at a 1:50 dilution and then washed with phosphate-buffered saline (PBS). Biotinylated horse anti-mouse immunoglobulin (DAKO, Kyoto,Japan) was then added to the sections for 30 minutes at room temperature. Peroxidase-conjugated avidin (DAKO) was applied after the sections were washed with PBS. Peroxidase activity was detected by exposure of the sections to the solution of 0.05% 3,3'-diaminobenzidine and 0.01% H202 in Tris-HC1 buffer at pH 7.6 for 3 to 6 minutes at room temperature. The sections were counterstained with hematoxylin. Normal mouse IgG was used as a substitute for the primary antibody for the negative controls. The sections showed a frequently heterogeneous staining pattern for anti-CD34 antibody. For the determination of IMVD, the 5 most vascular areas within a section were selected and counted under a light microscope with a 200-fold magnification (ie, 20× objective lens and 10× ocular lens; 0.7386 mm 2 per field) as described by Weidner et al. 2 The average numbers were recorded as IMVD for each case.

mmunohist0chemical Staining for dThdPas e To evaluate the angiogenic activity of the tumor cells, we performed an immunohistochemical study for dThdPase, also known as PD-ECGF, which is an enzyme implicated in tumorassociated angiogenesis. ~4,25 A mouse monoclonal antidThdPase antibody was provided by the Nippon Roche Research Center (Kamakura, Japan). Tumor sections were heated twice in an oven at 70°C for 5 minutes and then incubated overnight at 4°C with anti-dThdPase antibody at a 1:100 dilution and washed with PBS. The following steps were the same as those used for the anti-CD34 protocol. The immunoreactivity of dThdPase is expressed as a percentage of dThdPase-positive cancer cells/total number of cancer cells, and assigned to 1 of 3 subgroups: - , <10%; 1+, 10% to 50%; 2+, >50%.

Statistical Analysis The correlation between IMVD and dThdPase immunoreactivity was examined by Welch's t-test. The clinical characteristics of the patients were compared with IMVD and the expression of dThdPase in the tumor cells, arm checked by the Mann-Whitney and chi-square tests. The relapse-free and survival curves were plotted according to the Kaplan-Meier

method, and their statistical differences were analyzed by the log-rank test. A level of P < .05 was accepted as statistically significant.

RESULTS Correlation Between Tumor Vascularity and dThdPase Expression T h e i m m u n o r e a c t i v i t y o f anti-CD34 a n t i b o d y was l o c a t e d only o n the cytoplasm o f endothelial cells a n d n o t o n t u m o r cells o r interstitial cells (Fig 1A). In contrast to the i m m u n o r e a c t i v i t y o f anti-CD34 antibody, that o f d T h d P a s e was o b s e r v e d only in the cytoplasm o f t u m o r cells (Figs 1B, 1C). T h e negative c o n t r o l slides f o r CD34 a n d d T h d P a s e exhibited n o specific staining. As shown in Figure 2, the IMVDs in the t u m o r s with dThdPase-positive staining f r o m 1 + to 2 + were significantly h i g h e r t h a n those in the t u m o r s negative for d T h d P a s e staining (P < .01).

Correlation Between Tumor Vascularity and Clinicopathologic Features IMVD for the 54 t u m o r s r a n g e d f r o m 24.8 to 118.6 with a m e d i a n value o f 57.7 for CD34 staining. A m e d i a n value o f 57.7 was taken as the c u t o f f p o i n t for discrimination o f the 54 patients into two subgroups: 30 patients with hypovascular t u m o r s a n d 24 with hypervascular tumors. A m o n g the clinicopathologic variables examined, age, clinical stage, histological grade, CPL classification, 29 ovarian metastasis, a n d Vaginal infiltration were n o t c o r r e l a t e d with t u m o r vascularity. However, there were statistical differences in histological type, d e p t h o f rnyometrial invasion, e n d o m e t r i a l invasion, a n d pelvic l y m p h n o d e metastasis b e t w e e n 2 subgroups, as shown in Table 1 (P < .05).

Correlation Between dThdPase Expression and Clinicopathologic Features T h e d i n i c o p a t h o l o g i c data f r o m 18 patients whose t u m o r s h a d 2 + dThdPase-positive staining a n d 36 whose t u m o r s h a d negative or 1 + dThdPase-positive staining, were c o m p a r e d . As shown in Table 2, the i m m u n o r e a c t i v i t y o f d T h d P a s e was well c o r r e l a t e d with d e p t h o f m y o m e t r i a l invasion, e n d o m e t r i a l invasion, pelvic l y m p h n o d e metastasis, a n d t u m o r vascularity ( P < .01).

Tumor Vascularity, dThdPase Expression, and Patient Outcome O f the 54 patients e x a m i n e d in this study, 12 patients (22.2%) e x p e r i e n c e d recurrence, a n d 7 (13.0%) h a d died o f relapse o f the disease d u r i n g follow-up period. T h e r e was n o significant difference in relapsefree periods a n d survival rates b e t w e e n the hypovascular a n d hypervascular g r o u p s (Figs 3 a n d 4). T h e i m m u n o r e a c t i v i t y o f d T h d P a s e was n o t c o r r e l a t e d with relapse o f the disease (Fig 5); however, the overall survival rates for 18 patients whose t u m o r s h a d 2 + dThdPase-positive staining were significantly lower t h a n

1390

ANGIOGENESIS AND dThdPase IN CERVICAL CARCINOMA (Ueda et al)

FIGURE 1. Intratumoral microvessels and dThdPase expression in cervical carcinoma. Intratumoral microvessels are detected as consistent staining of endothelial ceils using antiCD34 antibody (A, original magnification x200). Immunoreactivity of dThdPase is identified in the cytoplasm of tumor cells, There is also a strong staining in squamous cell carcinoma (B, original magnification x200) and a d e n o c a r c i n o m a (C, original magnification x200).

those for:36 whose tumors had negative or 1 + dThdPasepositive staining (P = .0108) (Fig 6). DISCUSSION

P < .001

IMVD

I

I

P < .01

10o I

50

I

[

/

o -

n = 16

1+

n = 20

2+

n = 18

dThdPase Expression FIGURE 2. The relationship between IMVD and dThdPase expression. The IMVDs in the tumors with dThdPase-positive staining from 1+ to 2+ are significantly higher than those in the tumors negative for dThdPase staining (P < .01), 1391

Many clinical studies have shown that the intensity of angiogenesis, expressed as IMVD, predicts the probability Of metastasis, recurrence, and survival in a variety of tumors. 2,4-7'1°,~2-15'17a8 In gynecologic malignancies, the biology of t u m o r angiogenesis and its clinical significance have been reported in cervical, sq3,z7 endometrial, ~4q6 and ovarian 17-~9 carcinomas. Dinh et aP 2 showed that high IMVD might be used to predict recurrence in stage Ib squamous cell carcinoma of the cervix. B r e m e r et a113 indicated that IMVD was an i n d e p e n d e n t prognostic parameter for disease-free survival in patients with stages Ib and IIa cervical carcinomas. In early cervical lesions, IMVD was significantly increased in invasive carcinoma and in high-grade intraepithelial lesions compared with low-grade intraepithelial lesions and benign squamous epithelium, n Several investigators also f o u n d that t u m o r angiogenesis was used in the prognostic information for patients with endometria114a5 and ovarian ~7as carcinomas. In the current study, t u m o r vascularity in adenocarcinomas was significantly higher than that in squamous cell carcinomas, which might reflect the aggressive biological characteristic of cervical adenocarcinomas. Moreover, IMVD was well correlated with depth of myome-

HUMAN PATHOLOGY TABLE 1.

TABLE 2.

Correlation Between Tumor Vascularity and Clinicopathologic Features

Variable

Age

Volume 30, No. 11 (November 1999)

N50yr >50yr Clinical stage Stage 1 Stage 2 Stage 3 Histological type Squamous Adenosquamous and adeno Histological grade Well differentiated Moderately and poorly differentiated CPL classification* C type P type L type Depth of invasion+ 1-2 3-4 Endometrial invasion Negative Positive Pelvic lyrnphnode metastasis Negative Positive Ovarian metastasis Negative Positive Vaginal infiltration Negative Positive

Hypovascular Hypervascular (n = 30) (n = 24)

Correlation Between dThdPase Expression and Clinicopathologic Features

Variable

P

- / 1 + Expression 2+ Expression (n = 36) (n = 18)

P

Age 12 (40.0) 18 (60.0)

14 (58.3) 10 (41.7)

NS

10 (33.3) 19 (63.3) 1 (3.3)

9 (37.5) 14 (58.3) 1 (4.2)

NS

25 (83.3) 5 (16.7)

14 (58.3) 10 (41.7)

.0415

8 (26.7)

8 (33.3)

NS

22 (73.3)

16 (66.7)

17 (56.7) 2 (6.7) 11 (36.7)

6 (25.0) 1 (4.2) 17 (70.8)

NS

25 (83.3) 5 (16.7)

12 (50.0) 12 (50.0)

.0176

27 (90.0) 3 (10.0)

15 (62.5) 9 (37.5)

.0354

27 (90.0) 3 (10.0)

15 (62.5) 9 (37.5)

.0354

29 (96.7) 1 (3.3)

24 (100) 0 (0)

NS

20 (66.7) 10 (33.3)

16 (66.7) 8 (33.3)

NS

*The CPL classification proposed by Imai 2g is based on a variation in the pattern of growth at the advancing margin of the carcinoma. The C refers to cirrhotic sprouting (reactive sprouting), the P to progressive sprouting (sprouting of tumor cells without connective tissue reaction), and the L to lymphatic or blood vessel permeating forms. +Both the uterine cervix and parametrium are halved equally, and the deepest section of the primary tumor is classified from 1 to 4 of myometrial invasion.

trial invasion, endometrial invasion, and pelvic lymph n o d e metastasis. This is consistent with the findings of other studies using invasive cervical 13 and endometria115 carcinomas. However, there was no statistical correlation between IMVD and patient outcome. Not all investigators have f o u n d this association. An inverse correlation for IMVD for patients with cervical carcin o m a has b e e n reported, s,9 Kainz et al s showed that patients with tumors showing low IMVD had a significantly p o o r e r recurrence-free interval. Because all of these results have b e e n obtained from retrospective studies, a large prospective study will be n e e d e d to clarify whether quantitative pathology using microvessel counting is an i n d e p e n d e n t and important prognostic indicator of cervical carcinoma. Angiogenesis is a complex process that involves not only endothelial cell proliferation but also digestion of the extracellular matrix surrounding capillaries, endothelial cell migration, and differentiation into functioning capillaries. Abundant evidence supports the con-

_-<50yr >50 yr Clinical stage Stage 1 Stage 2 Stage 3 Histological type Squamous Adenosquamous and adeno Histological grade Well differentiated Moderately and poorly differentiated CPL classification* C type P type L type Depth of invasion+ 1-2 3-4 Endometrial invasion Negative Positive Pelvic lymph node metastasis Negative Positive Ovarian metastasis Negative Positive Vaginal infiltration Negative Positive IMVD <57.7 >57.7

17 (47.2) 19 (52.8)

9 (50.0) 9 (50.0)

NS

13 (36.1) 22 (61.1) 1 (2.8)

6 (33.3) 11 (61.1) 1 (2.8)

NS

27 (75.0) 9 (25.0)

12 (66.7) 6 (33.3)

NS

12 (33.3)

4 (22,2)

NS

24 (66.7)

14 (77.8)

19 (52.8) 1 (2.8) 16 (44.4)

4 (22.2) 2 (11.1) 12 (66.7)

NS

30 (83.3) 6 (16.7)

7 (38.9) l l (61.1)

.0009

32 (88.9) 4 (11.1)

10 (55.6) 8 (44.4)

.0055

32 (88.9) 4 (11.1)

10 (55.6) 8 (44.4)

.0055

35 (97.2) 1 (2.8)

18 (100) 0 (0)

NS

25 (69.4) 11 (30.6)

11 (61.1) 7 (38.9)

NS

27 (75.0) 9 (25.0)

3 (16.7) 15 (83.3)

.0001

*The CPL classification proposed by Imai 29 is based on a variation in the pattern of growth at the advancing margin of the carcinoma. The C refers to cirrhotic sprouting (reactive sprouting), the P to progressive sprouting (sprouting of tumor cells without connective tissue reaction), and the L to lymphatic or blood vessel permeating forms. +Both the uterine cervix and parametrium are halved equally, and the deepest section of the primary tumor is classified from 1 to 4 of myometrial invasion.

cept that tumors can induce angiogenesis through a variety of angiogenic molecules, such as VEGF, bFGF, TGF-et, and dThdPase/PD-ECGF, released from the tumor cells or rumor-associated inflammatory cells. 2°-22,24-z6 Among several potential angiogenic factors, expression of dThdPase has previously been demonstrated in breast, ~,3° bladder, ~1 gastric, 5,32,a~ colorectal, a4 endometrial, 16 and ovarian 19 carcinomas. O u r study using invasive cervical carcinomas showed that dThdPase expression was closely associated with IMVD, tumor invasion, pelvic l y m p h n o d e metastasis, and overall survival. Recent studies have also f o u n d that immunohistochemical expression of dThdPase is correlated with t u m o r vascularity, 3,5,16,19 the depth of t u m o r invasion, 34 lymphnode metastasis, 3a,34 and hematogenous metastasis. 5,32 Conversely, the correlation between dThdPase expression and prognosis remains controversial. Fox et al so re-

1392

ANGIOGENESIS AND dThdPase IN CERVICAL CARCINOMA (Ueda et al) Relapse-free

Relapse-free Rate

Rate 1

I

hypovascular (n = 30) .8

-/1 + dThdPase expression (n = 36)

.8

hypervascular (n = 24)

.6

.6

.4

.4

.2

.2

[ 2+ dThdPase expression (n = 18)

0

0 I

0

I

20

I

I

40

60

I

80

I

1 O0

I

[

I

!

I

~

p

0

20

40

60

80

1 O0

120

I

120

Months

Months FIGURE 3. Kaplan-Meier analysis of relapse-free rate of 54 patients according to tumor vascularity. There is no statistical difference between the 2 groups.

ported that dThdPase expression was associated with a favorable prognosis in patients with node-positive breast carcinoma. O'Brien et a131 failed to find a correlation between dThdPase expression and prognosis in patients with bladder carcinoma. Fujiwaki et a116 and Nakanishi et a119 also showed that dThdPase expression was not correlated with prognosis in endometrial and ovarian carcinoma, respectively. However, immunoreactivity of dThdPase was inversely correlated with survival, as d e t e r m i n e d by univariate analysis in patients with gastric ~2,33 and colorectaP 4 carcinoma. The reason for this discrepancy is not clear, but the association between dThdPase expression and patient outcome may d e p e n d on the histological type and the evaluation of staining results. In recent years, dThdPase has received considerable attention as a m o d u l a t o r o f chemotherapeutic Survival Rate 1

FIGURE 5. Kaplan-Meier analysis of relapse-free rate of 54 patients according to dThdPase expression. There is no statistical difference between the 2 groups.

drugs. 5'-Deoxy-5-fluorouridine (5'-dFUrd) is converted to 5-fluorouracil (5-FUra) and 5-FUra is converted to fluoro-deoxyuridine m o n o p h o s p h a t e by dThdPase in the t u m o r tissue. Haraguchi et aP 5 showed that dThdPase-transfected KB cells were more sensitive to 5'-dFUrd than untransfected KB cells. We previously r e p o r t e d that epidermal growth factor and TGF-~ upregulated the production of dThdPase in h u m a n cervical carcinoma cells and consequently e n h a n c e d the anti-proliferative action of 5'-dFUrd. s6 Thus, with respect to t u m o r angiogenesis, the antitumor activities of 5'-dFUrd and 5-FUra, being widely used in the treatment of various kinds of tumors, may be important in the selective inhibition of those tumor cells expressing dThdPase. Moreover, we also demonstrated that antitumor effects of preoperative intraarterial chemotherapy for patients with cervical carcinoma were closely associated with the inhibition of tumor angiogenesis Survival Rate

"•

hypovascular (n = 30) I

I,

.8

]

-/1 + dThdPase expression (n = 36)

1

=

hypervascular (n = 24)

.8

[ 2+ dThdPase expression (n = 18) .6

.6

.4

.4

.2

.2

0 0 I

20

40

i

I

I

c

60

80

1 O0

1 20

I

~

i

0

20

40

"'

I''

I

I

i

60

80

I00

120 Months

Months FIGURE 4. Kaplan-Meier analysis of survival rate of 54 patients according to tumor vascularity. There is no statistical difference between the 2 groups. 1393

FIGURE 6. Kaplan-Meier analysis of survival rate of 54 patients according to dThdPase expression. The overall survival rates for 18 patients whose tumors have 2+ dThdPase-positive staining are significantly lower than those for 36 whose tumors have negative or 1+ dThdPase-positive staining (P = .0108).

HUMAN PATHOLOGY

Volume 30, No. 11 (November 1999)

and dThdPase expression of t u m o r cells. 27 Ascertainm e n t of factors that are highly relevant to t u m o r angiogenesis, based on the data of t u m o r vascularity as well as clinical outcomes, may have an influence on future development of new treatment modalities that may inhibit t u m o r growth and metastatic formation by inhibiting angiogenesis.

Acknowledgment. T h e authors thank N i p p o n R o c h e Co. for the gift of anti-dThdPase antibody. We also thank several colleagues fo r c o l l e c t i n g clinical materials, and E. Shintani and K. Sato for their technical assistance.

REFERENCES 1. FolkmanJ: Clinical appfications of research on angiogenesis: N EnglJ Med 333:1757-1763 1995 2. Weidner N, Semple JP, Welch WR, et al: Tumor angiogenesis and metastasis: Correlation in invasive breast carcinoma. N Engl J Med 324:1-8, 1991 3. Toi M, Hoshina S, Taniguchi T, et al: Expression of plateletderived endothelial cell growth factor/thymidine phosphorylase in human breast cancer. IntJ Cancer 64:79-82, 1995 4. Weidner N , Carrol PR, Flax J, et al: Tumor angiogenesis correlated with metastasis in invasive prostate carcinoma. AmJ Patho! 143:401-409, 1993 ~ 5. TanigawaN, Amaya H, Matsiamura M, et al: TumOr angiogenesis and expression of thymidine phosphorylase/platele~ derived endothelial cell growth factor in human gastric carcinoma. Cancer Lett 108:281-290~ 1996 6. Tamgawa N, Amaya H, Matsumura M,'. et al: Correlation between expression of vascular endothelial growth factor and tumor vascularity, and patient outcome in human gastric carcinoma. J Clin Onco115:826-832, 1997 7. Tanigawa N, Amaya H, Matsumura M, et al: Tumor angiogenesis and mode of metastasis in patients with colorectal cancer. Cancer Res 57:1'043-1046, 1997 8. Kainz C, Speiser P, Wanner C, et al: Prognostic value of tumor microvessel density in cancer of the uterine cervix stage IB to IIB. Anticancer Res 15:1549-1551, 1995 9. Rutgers JL, Mattox TF, Vargas P: Angiogenesi's in uterine cervical squamous cell carcinoma. I n t J Gynecol Pathol 14:114-118, 1995 10. Wiggins DL, Granai CO, Steinhoff MM, et al: Tumor anglogenesis as a prognostic factor in cervical carcinoma. Gynecol Oncol 56:353-356, 1995 11. Guidi P~J,Abu-Jawdhh G, Berse B, et al: Vascular permeability factor (vascular endothelial growth factor) expression and angiogenesis in cervical neoplasia.J Natl Cancer Inst 87:1237-1245, 1995 12. Dinh TV, Hannigan EV, Smith ER, et al: Tumor angiogenesis as a predictor of recurrence in stage Ib squamous cell carcinoma of the cervix. Obste~ Gynecol 87:751-754, 1996 13. Bremer GL, Tiebosch A, Putten H, et al: Tumor angiogenesis: An independent prognostic parameter in cervical cancer. Am J Obstet Gynecol 174:126-131, 1996 14. Kirschner CV, Alanis-AmezuaJM, Martin VG, et al: Angiogenesis factor in endometrial carcinoma: A new prognostic indicator? Am J Obstet Gyneco1174:1879-1884, 1996 15. Kaku T, Kamura T, Kinukawa N, et al: Angiogenesis in endometrial carcinoma. Cancer 80:741-747, 1997

16. Fujiwaki R, Ham K, Iida K, et al: Immunohistochemical expression of thymidine phosphorylase in human endometrial cancer. Gynecol Oncol 68:247-252, 1998 17. Hollingsworth HC, Kohn EC, Steinberg SM, et al: Tumor angiogenesis in advanced stage ovarian carcinoma. Am J Pathol 147:33-41, 1995 18. Schoell WMJ, Pieber D, Reich O, et al: Tumor angiogenesis as a prognostic factor in ovarian carcinoma. Cancer 80:2257-2262, 1997 19. Nakanishi Y, KodamaJ, Tokumo K, et al: The expression of platelet-derived endothelial cell growth factor/thymidine phosphorylase associates with angiogenesis in epithelial ovarian cancer. IntJ Clin Oncol 2:219-223, 1997 20. Connolly DT: Vascular permeability factor: A unique regulator of blood vessel function.J Cell Biochem 47:219-223, 1991 21. Ohtani H, Nakamura S, Watanabe Y, et al: Immunocytochemical localization of basic fibroblast growth factor in carcinomas and inflammatory lesions of the human digestive tract. Lab Invest 68:520527, 1993 22. Malden LT, Novak U, Burgess AW: Expression of transforming growth factor alpha messenger RNA in the normal and neoplastic gastro-intestinal tract. IntJ Cancer 43:380-384, 1989 23. FolkmanJ, Klagsburn M: Angiogenic factors. Science 235:442447, 1987 24. Furukawa T, Yoshimura A, Sumizawa T, et al: Angiogenic factor. Nature (Lond) 356:668, 1992 25. Moghaddam A, Zhang HT, Fan TPD, et al: Thymidine phosphorylase is angiogenic and promotes tumor growth. Proc Natl Acad Sci U S A 92:998-1002, 1995 26. Haraguchi M, Miyadera K, Uemura K, et al: Angiogenic activity of enzymes. Nature (Lond), 368:198, 1994 27. Ueda M, Ueki K, Kumagai K, et al: Apoptosis and tumor angiogenesis in cervical cancer after preoperative chemotherapy. Cancer Res 58:2343-2346, 1998 • 28. International Federation of Gynecology and Obstetrics Cancer Committee: Staging ann'0uncement. Gynec01 Oncol 25:383-385, 1986 29. Imai T: Growth patterns in human carcinoma. Obstet Gynecol 16:296-308, 1960 30. Fox SB, Westwood M, Moghaddaln A, et al: The antigenic factor platelet-derived endothelial cell growth factor/thymidine phosphorylase is up-regulated in breast cancer epithelium and endothelium. BrJ Cancer 73:275-280, 1996 31. O'Brien TS, Fox SB, Dickison AJ, et al: Expression of the angiogenic factor thymidine phosphorylase/platelet-derived endothelial cell growth factor in primary bladder cancers. Cancer Res 56:47994804, 1994 32. Maeda K, Chung YS, Ogawa S, et al: Thymidine phosphorylase/platelet-derived endothelial cell growth factor expression associated with hepatic metastasis in gastric carcinoma. Br J Cancer 73:884-888, 1996 33. Takebayashi Y, Miyadera K, Akiyama S, et al: Expression of thymidine phosphorylase in human gastric carcinoma. J p n J Cancer Res 87:288-295, 1996 34. Takebayashi Y, Akiyama S, Akiba S, et al: Clinicopathologic and prognostic significance of an angiogenic factor, thymidine phosphorylase, in human colorectal carcinoma.J Natl Cancer Inst 88!11101117, 1996 35. Haraguchi M, Fm:ukawa T, Sumizawa T, et al: Sensitivity of human KB cells expressing! platelet-derived endothelial cell growth factor to pyrimidine antimetabolites. Cancer Res 53:5680-5683, 1993 36. Ueda M, Ueki M, Terai Y, et al: Stimulatory effects of EGF and TGF-~ on invasive activity and 5'2deoxy-5-fluorouridine sensitivity in uterine cervical carcinoma SKG-IIIb cells. Int J Cancer 72:10271033, 1997

1394