Ann Surg Oncol (2009) 16:352–360 DOI 10.1245/s10434-008-0242-2

ORIGINAL ARTICLE – THORACIC ONCOLOGY

Prognostic Role of PGE2 Receptor EP2 in Esophageal Squamous Cell Carcinoma Kuang-Tai Kuo, MD1,2, Hao-Wei Wang, MD3, Teh-Ying Chou, MD, PhD2,4, Wen-Hu Hsu, MD5, Han-Shui Hsu, MD, PhD5, Chi-Hung Lin, MD, PhD6, and Liang-Shun Wang, MD7,8 1 Division of Thoracic Surgery, Department of Surgery, Buddhist Tzu Chi General Hospital, Taipei Branch, Taipei, Taiwan; 2 Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan; 3Faculty of Medicine, School of Medicine, National Yang-Ming University, Taipei, Taiwan; 4Department of Pathology, Taipei Veterans General Hospital, Taipei, Taiwan; 5Division of Thoracic Surgery, Department of Surgery, Taipei Veterans General Hospital, Taipei, Taiwan; 6 Institute of Microbiology and Immunology, National Yang-Ming University, Taipei, Taiwan; 7Graduate Institute of Basic Medicine, Fu Jen Catholic University, Taipei, Taiwan; 8Department of Surgery, En Chu Kong Hospital, No.399, FuSing Rd., SanSia Township, Taipei 237, Taiwan

ABSTRACT Prostaglandin E2 (PGE2), a major cyclooxygenase-2 (COX-2) product, has been shown to affect numerous tumorigenic processes. PGE2 acts through Gprotein-coupled receptors designated as EPs. Recently it has been documented that PGE2 promotes colon cancer cell growth via EP2. However, the expression and the prognostic role of EP2 in esophageal squamous cell carcinoma (ESCC) remained unknown. From January 1995 to January 2001, tissue samples from 226 patients with ESCC who underwent esophagectomies at our institutions were collected and made into tissue core arrays for study. EP2 expression was examined by immunohistochemical staining and confirmed by Western blot. The clinicopathologic data were then analyzed. EP2 overexpression was observed in 43.4% (98/226) of ESCC. Overexpression of EP2 correlated positively with depth of tumor invasion (T status) (P = 0.016) and was associated with worse overall survival (P = 0.047). In patients without regional or distant lymph node metastasis (N0 or M0), EP2 overexpression was associated with worse overall survival (P = 0.033 and P = 0.003, respectively). Using Cox regression analysis, T status, N status, and M status were the independent factors of overall survival, but EP2 expression was not. However, when focusing on patients with T1-3N0M0 status, EP2 expression became an independent factor of overall survival (P = 0.048). Our results show that EP2 overexpression was associated with worse prognosis, and correlated positively Ó Society of Surgical Oncology 2008 First Received: 20 May 2008; Published Online: 3 December 2008 L.-S. Wang, MD e-mail: [email protected]

with T status in ESCC. Meanwhile, among those patients at earlier stages, EP2 overexpression significantly disclosed patients at high risks for poor prognosis. Esophageal cancer is one of the most aggressive human malignancies in the world, with overall 5-year survival rate below 20%.1 It is ranked the sixth most common cause of death among male cancer patients in Taiwan, responsible for approximately 1,180 deaths per year.2 Recently, it was reported that administration of preoperative chemoradiotherapy enabled more complete surgical resections and improved locoregional control in locally advanced esophageal cancers.3,4 However, most phase III trials could not demonstrate a survival benefit in those receiving preoperative chemoradiotherapy.5–7 Limited improvement in treatment outcomes by conventional therapies urges us to seek innovative strategies for treating esophageal cancers, especially those which are molecular targeted. Epidemiological studies have shown that chronic intake of nonsteroidal anti-inflammatory drugs (NSAIDs) reduced the incidence of many digestive cancers, and thereby many studies have been conducted to investigate the relationship between cyclooxygenase (COX), especially COX-2, and cancers.8 The reported overexpression rate of COX-2 in esophageal squamous cell carcinoma (ESCC) varied from 25% to 60%, but the prognostic significance of COX-2 overexpression in ESCC was debatable.9–12 Interestingly, the prognostic significance of COX-2 overexpression had been reported to be distinctly different in esophageal adenocarcinoma and ESCC by the same research group.12,13 COX-2 converts arachidonic acid to several prostaglandins including PGE2, and PGE2 has been implicated in

EP2 and Esophageal Squamous Cell Carcinoma

tumor growth through enhancement of cellular proliferation, promotion of angiogenesis, inhibition of apoptosis, and stimulation of tumor invasion.14–16 PGE2 elaborates its biological activities through some G-protein-coupled receptors called EPs, which consist of four isoforms: EP1, EP2, EP3, and EP4.17 The downstream signaling pathways of EPs are not the same: EP1 uses phospholipase C/inositol triphosphate as the second messenger signal, EP2 and EP4 act through increasing cyclic adenosine monophosphate (cAMP), whereas EP3 acts mainly through decreasing cAMP.18 Recently, the mechanism of PGE2-induced colon cancer progression has been elucidated, and EP2 was documented to play a key role in the signaling axis.19 In addition to this, there are several other studies showing that EP2 was associated with cell growth, angiogenesis and tumor development.20–23 Regarding esophageal cancers, reports are scanty but cell line studies have shown that activation of EP2 stimulated cell migration in esophageal adenocarcinoma and EP2 mediated the mitogenic effect of PGE2 in ESCC via activation of the extracellular signalregulated kinase (Erk)/activator protein-1 (AP-1) pathway.24,25 However, the clinical significance of EP2 overexpression in ESCC remains unclear. In the present study, we examined expressions of EP2 in 226 patients with resected ESCC by immunohistochemistry (IHC). The relationship between EP2 overexpression and clinicopathologic features were also analyzed.

MATERIALS AND METHODS Patients Approval for this retrospective study was obtained and the need for individual patient consent was waived by the Institutional Review Board. Tumor parts of surgical specimens from 226 patients with ESCC who underwent esophagectomy at our institutions from January 1995 to January 2001 were collected for study. Radical en bloc esophagectomy was carried out in every patient. None of these patients received neoadjuvant radiotherapy or chemotherapy, nor did they have distant organ metastasis in preoperative assessments. Stage of disease progression was classified according to the American Joint Committee on Cancer (AJCC) staging system. All stage IV patients in the present series were due to distant lymph node metastasis (M1a disease in U/3 and L/3 esophageal cancers; M1b disease in M/3 esophageal cancers). Postoperative adjuvant therapy was administered for patients at stage IIB or beyond, or when tumor recurrence was noticed. Irradiation dose was 60 Gy (10 Gy/5 fractions/week), and a combined chemotherapy regimen of choice consisted of cisplatin (20 mg/m2/day), 5-fluorouracil (600 mg/m2/day), and

353

leucovorin (120 mg/m2/day) administered by 24-h infusion for 4 days. After treatment, all patients were followed up with systemic examinations of biochemical tests, chest radiography, sonogram of abdomen and neck, and wholebody bone radioisotope scanning every 3–6 months. Computed tomography (CT) scan of chest and abdomen was also done if indicated. There were 17 patients lost to follow-up, and the follow-up rate was 92.5%. Tissue Core Array All of the pathological sections were reviewed by one pathologist (T.-Y.C.) and representative tumor parts on corresponding tissue blocks were marked. Cores of embedded tissue in demand were manually bored from the tissue blocks by a 10-gauge syringe needle. The cores were placed in a warm cast containing melted paraffin wax and were arranged into a 6 9 10 matrix with the core at the upper-left corner removed to serve as a marker for orientation. Each slide contained 59 cores and each patient had two to five cores of their tumor specimens, depending on the availability of tissue blocks. The positions of each core were recorded on reference sheets to facilitate data acquisition. Immunohistochemical Staining An immunoperoxidase procedure was used to detect the expression of EP2 in the pathological sections.26 Briefly, the tissue arrays were cut into 4-lm sections. The sections were deparaffinized and incubated with 3% hydrogen peroxide to inactivate endogenous peroxides. Specific polyclonal rabbit antibodies against EP2 (101750; Cayman Chemical Co., Ann Arbor, MI, USA) were used at a dilution of 1:50 and applied to tissue array sections. They were then washed and incubated with secondary antibody (LSAB II kit; DAKO, Carpinteria, CA, USA). Finally, the sections were counterstained with hematoxylin. Evaluation of EP2 Immunostaining All immunostained tissue array sections were evaluated in a coded manner without knowledge of the clinical and pathological background of the patients. Each stained core was evaluated according to the gross percentage of cells demonstrating a cytoplasmic immunoreactivity on the tumor part, and the assessment was accomplished independently by two examiners (K.-T.K. and H.-W.W.). Tumor specimens with high EP2 expression demonstrated on Western blot analysis were used as a positive control for the immunohistochemistry while an immunohistochemical staining without anti-EP2 antibodies applied was used as a negative control. Both positive and negative controls were

354

performed in each IHC experiment. The following scoring criteria had been applied previously: 0, no staining; 1?, weakly diffuse cytoplasmic staining (may contain stronger intensity in less than 10% of the cancer cells); 2?, heterogeneously granular cytoplasmic staining in 10–90% of the cancer cells; 3?, more than 90% of the cancer cells stained with strong intensity.10 For fitness of statistical analyses, 2? or 3? cytoplasmic immunoreactivity was defined as overexpression of EP2, namely EP2 high, whereas 0 or 1? cytoplasmic immunoreactivity was defined as EP2 low. The allocation of tumors to the EP2-low versus the EP2high category by the two investigators was similar (91% of the specimens were categorized identically). In cases of disagreement (n = 20), the slides were re-evaluated by a group of investigators (K-.T.K., H.-W.W. and L.-S.W.) using a multiheaded microscope. Western Blot Analysis IHC staining of EP2 was then confirmed by Western blot analysis in part of surgical specimens. Procedures for Western blot have been described previously.27 Briefly, paired frozen tissues from five patients with EP2 overexpression were firstly examined by microscope to ensure that at least 90% of each sample composed of tumor. They were thawed in ice-cold homogenization buffer, and the lysates were sonicated and centrifuged at 10,000 9 g for 10 min to sediment the particulate material. The protein content was measured and 100 lg of proteins were separated by sodium dodecyl sulfate (SDS) (10%) polyacrylamide gel electrophoresis (PAGE) and transferred to a nitrocellulose membrane. The membrane was then probed with rabbit polyclonal antibodies (101750; Cayman Chemical Co., Ann Arbor, MI, USA) specific to EP2 (1:2,000). The signal was amplified by biotin-labeled goat anti-rabbit IgG (1:3,000) and peroxidase-conjugated streptavidin (1:5,000). EP2 protein was visualized by exposing the membrane to an X-Omat film (Eastman Kodak, Rochester, NY, USA) with enhanced chemiluminescent regent (PIERCE, Rockford, IL, USA). Clinicopathologic Variables and Statistical Analysis Age, sex, history of smoking or habitual alcohol consumption, differentiation of tumor, lymphovascular invasion, tumor location, length of tumor, depth of tumor invasion (T status), status of lymph node metastasis (N status), status of distant metastasis (M status), and tumor stage in the database were reviewed. The relationships between clinicopathologic factors and EP2 expression status were analyzed by v2 test (or two-tailed Fisher’s exact test when the expected number in any cell was less than five cases). Survival curve was plotted, and median

K.-T. Kuo et al.

survival was estimated by the Kaplan–Meier method. Survival and the strength of associations between categories within a variable were compared with the log-rank test. Variables with a P value less than 0.1 were entered into the multivariate analysis by using a Cox proportional hazards stepwise model. Four variables were therefore analyzed: T status, N status, M status, and EP2 expression. Statistical significance was defined as a probability value less than 0.05. All statistical analyses were performed with SPSS software version 11.0 (SPSS, Inc., Chicago, IL, USA).

RESULTS There were 211 males and 15 females, with mean age of 64 years (range 36–86 years). One hundred and eighty-four (81.4%) patients were smokers and 163 patients (72.1%) consumed alcohol habitually. The surgical mortality rate was 5.3% (12 patients), and the causes of surgical mortality in 12 patients included pulmonary complications (11 patients) and myocardial infarction (1 patient). The median follow-up period for the remaining 214 patients was 23.1 months (range 3.5–135.5 months). The overall median survival time was 21.6 months, and the cumulative 5year survival rate was 29.4%. The cumulative 5-year survival rate for N0 patients (n = 96) was 44.9%, and for N1 patients (n = 130) was 15.9%. Immunohistochemically, EP2 staining was observed in both membrane and cytoplasm of tumor cells. In the current series, EP2 immunoreactivity (1? to 3?) was detected in 120 of 226 ESCC specimens (53%). EP2 overexpression (3?, Fig. 1a and 2?, Fig. 1b) was observed in 98 patients (43%), whereas only weak (1?, Fig. 1c) or negative (0, Fig. 1d) EP2 expression was observed in 128 patients (57%). The overexpression of EP2 was further confirmed by Western blot analysis. As shown in Fig. 2, EP2 was indicated as a 53-kDa band. The band was not exclusively present in tumorous tissues but also could be detected in nontumorous counterparts. As shown in Table 1, EP2 overexpression had a significantly positive correlation with depth of invasion (T factor) (P = 0.016). Otherwise, no significant difference in other clinicopathologic variables was observed between tumors with and without EP2 overexpression. Overall survival was significantly different between patients with and without EP2 overexpression (P = 0.047) (Fig. 3). Meanwhile, among the patients with N0 status (n = 96), EP2 overexpression was a significant indicator of worse prognosis (P = 0.033), whereas among the patients with N1 status (n = 130) this significance did not exist (P = 0.615) (Fig. 4). Similarly, among the patients with M0 status (n = 186), EP2 overexpression also indicated a

EP2 and Esophageal Squamous Cell Carcinoma

355

FIG. 1 Representative examples of EP2 immunohistochemical stains. a Diffusely strong (3?) immunoreactivity in the tumor cells. b Heterogeneously strong (2?) immunoreactivity in the tumor cells; tumors with 3? and 2? were categorized as EP2 high. c Diffusely weak (1?) immunoreactivity in the tumor cells. d Negative (0) immunoreactivity in the tumor cells. Tumors with 1? and 0 were categorized as EP2 low (original magnification 409)

(P = 0.018) and EP2 expression (P = 0.048) were independent factors. DISCUSSION

FIG. 2 Western blot analysis of EP2 expression in tumorous (T) and nontumorous (N) tissue pairs from five representative patients with ESCC

significantly worse prognosis (P = 0.003), whereas such an association disappeared (P = 0.507) among the patients with M1 status (n = 40) (Fig. 5). Furthermore, among the patients who did not receive postoperative adjuvant therapy, namely those with T1-3N0M0 status (n = 84), EP2 overexpression was also associated with a significantly worse prognosis (P = 0.027, Fig. 6) In the univariate analysis (Table 2), T status (P \ 0.001), N status (P \ 0.001), M status (P \ 0.001), and EP2 expression (P = 0.047) were significantly prognostic factors for survival. When we used the Wald backward stepwise method for multivariate analysis, T status (P = 0.004), N status (P = 0.003), and M status (P = 0.002) were independent factors but EP2 expression was not (P = 0.194). However, when we narrowed the range to patients with T1-3N0M0 status (n = 84), T status

The crucial role of PGE2 in cancer development and progression has gradually been recognized, but most of the involved molecular mechanisms are not fully clarified. Since PGE2 elaborates its biological activities through activation of EPs, investigation of EPs in tumors may lead to understanding of PGE2 downstream pathways and result in therapeutic strategies. From previous reports, we know that only selective EPs are dominant in specific tissues.28–30 Therefore, we speculate that most tumors will also have specifically dominant EP(s) and such EP(s) may have a determinant role in PGE2-related cancer progression. In this study, a significantly positive correlation between EP2 overexpression and depth of tumor invasion (T status) in ESCC was found, namely tumors with EP2 overexpression usually had deeper invasions. Similar finding had been reported by Miyata et al., who showed that the frequency of positivity for EP4R expression increased in parallel with depth of invasion in transitional cell carcinoma of the upper urinary tract.31 They also found a significant association between EP4R positivity and histological grade, but this relationship was not shown in our data.

356

K.-T. Kuo et al.

TABLE 1 Relationship between EP2 expression and clinicopathologic parameters Parameters

EP2 high (n = 98)

EP2 low (n = 128)

Age (years)

63.94 ± 10.16 64.29 ± 10.63 0.803

Sex Male Female

117 11

Yes

78

106

No

20

22

Yes

73

90

No

25

38

Smoking

0.8

log-rank p = 0.047

0.6

0.539

Alcohol consumption

0.490

Tumor differentiation 15

11

Moderate

68

101

Poor

15

16

Yes

26

29

No

72

99

18

29

Lymphovascular invasion

58

63

Lower thirdb

22

36

\ 5 cm

71

82

C5 cm

27

46

Length of tumor

0.880

0.183

Depth of invasion (T)

0.016*

T1

8

16

T2

14

38

T3

61

58

T4

15

16

Nodal status (N)

0.128

N0

36

60

N1

62

68

81 17

105 23

I

7

13

II

27

57

III

48

36

IV

16

22

Metastasis (M)

0.904

Stage

0.057

a

Including one tumor over cervical esophagus

b

Including three tumors over esophagogastric junction

0.2

20

40

60

80

100

120

140

Time (Months) FIG. 3 Survival curves of patients with or without EP2 overexpression (n = 226; P = 0.047)

0.604

Middle third

0.4

0

0.564

Well

M0 M1

EP2 low EP2 high

0.178 94 4

Tumor location Upper thirda

P value

Cumulative Overall Survival 1.0

* P \ 0.05

In fact, the link between PGE2 and tumor invasion is now becoming more elucidated. PGE2 may enhance invasiveness either by transactivating epidermal growth

factor receptor (EGFR) and its downstream process or by regulating vascular endothelial growth factor (VEGF) production.32–35 In the aforementioned study using esophageal adenocarcinoma cells (cell line OE33), treatment with either the selective EP2 agonist Butaprost or 16,16dimethyl PGE2 significantly inhibited butyrate-induced apoptosis and stimulated OE33 cell migration.24 Those authors documented that the importance of EP receptors in esophageal cancer cell aggressiveness was no less than that of PGE2. Corresponding to their findings, our current result implied that EP2 might play an important role in tumor invasiveness in ESCC because overexpression of EP2 was positively correlated with the depth of tumor invasion (T status). We also found that EP2 overexpression was associated with a worse overall survival. This could be partly explained by the abovementioned finding that EP2 overexpression correlated positively with the T status but there could be other factors not defined yet. Nevertheless, in the multivariate analysis, EP2 overexpression was not an independent factor for overall survival. This was also consistent with the result reported by Miyata et al. showing that EP4R expression was not an independent factor for cause-specific survival in a multivariate model.31 We considered that the influence of EP2 overexpression was somehow weakened by the T status in multivariate analysis and hence lost its significance. It was a novel finding that, among the patients without evident lymph node metastasis (N0 and M0), EP2 overexpression was an indicator of poor prognosis. Furthermore, EP2 expression became an independent factor of overall survival in the patients with T1-3N0M0 status. We considered that EP2 overexpression might be associated with a potential for ongoing disease progression

EP2 and Esophageal Squamous Cell Carcinoma FIG. 4 Survival curves of patients at a N0 status (n = 96; P = 0.033) and b N1 status (n = 130; P = 0.615), with or without EP2 overexpression

357

a

b

Cumulative Overall Survival

Cumulative Overall Survival EP2 low

1.0

1.0

EP2 low

EP2 high

0.8

log-rank p = 0.033

EP2 high

0.8

0.6

0.6

0.4

0.4

0.2

0.2 0

20

40

60

80

100

120

140

log-rank p = 0.615

0

20

40

Time (Months) FIG. 5 Survival curves of patients at a M0 status (n = 186; P = 0.003) and b M1 status (n = 40; P = 0.507), with or without EP2 overexpression

a

b

Cumulative Overall Survival

Cumulative Overall Survival EP2 low EP2 high

1.0 0.8

log-rank p = 0.003

0.4

0.4

0.2

0.2 40

60

80

100

Time (Months)

Cumulative Overall Survival 1.0

EP2 low EP2 high

0.8

log-rank p = 0.027

0.6

0.4

0.2

0

20

40

60 80 Time (Months)

100

120

140

FIG. 6 Survival curves of patients who did not receive postoperative adjuvant therapy (T1-3N0M0 status, stage I and IIA), with or without EP2 overexpression (n = 84; P = 0.027)

and thereby worsened survival. Although the underlying causes were not defined from our current data, the clinical importance of this finding should be highlighted. In

100

120

140

120

140

EP2 low EP2 high

0.8 0.6

20

80

1.0

0.6

0

60

Time (Months)

log-rank p = 0.507

0

20

40

60

80

100

120

140

Time (Months)

general, these T1-3N0M0 patients are considered at earlier stages (stage I and IIA) and usually receive follow-up only until the discovery of tumor recurrence. From our current data, we showed that EP2 overexpression could signal high-risk patients for worse prognosis in such a patient group and may enable them to undergo postoperative adjuvant therapy earlier in order to achieve better outcomes. Compared with our previous study showing that COX-2 overexpression was not associated with a prognostic significance, the current study revealed that EP2 overexpression has a significantly negative influence on overall survival.10 Although the patient population and sample size between these two studies were different and could have some impacts on the results, we still considered that there was a logical explanation from biological viewpoints. COX-2 is an upstream enzyme, and overexpression of COX-2 implies increased production of PGE2 at most. However, the expression status of EP receptors on tumors is variable and increased PGE2 will not lead to expected overreaction once the EP receptors are underexpressed or even not expressed. On the contrary, the EP receptors are the real executors of PGE2 stimulation and overexpression of EP

358 TABLE 2 Relationship between clinicopathologic characteristics and survival

K.-T. Kuo et al.

Factors

No.

MST (months)

211

21.6

15

27.5

Yes

184

19.3

No

42

12.3

Yes

163

19.0

No

63

23.7

26 169

67.6 19.3

31

7.9

Yes

55

19.3

No

171

22.0

Upper thirda

54

8.8

Middle third

121

22.0

Lower thirdb

58

17.3

\5 cm

153

19.3

C5 cm

73

17.3

Sex Male Female Smoking

Alcohol consumption

Tumor differentiation Well Moderate Poor Lymphovascular invasion

Univariate analysis, P value

Multivariate analysis, P value

0.918

–

0.357

–

0.980

–

0.626

–

0.395

–

0.135

–

0.901

–

Tumor location

Length of tumor

Depth of invasion (T) T1

24

67.6

T2 T3

52 119

26.9 17.3

T4

31

8.8

Nodal status (N) N0

96

36.0

N1

130

14.3

M0

186

24.5

M1

40

11.6

High

98

15.2

Low

128

28.9

High

30

30.0

Low

54

68.4

Metastasis (M)

EP2 expression MST median survival time a

Including one tumor over cervical esophagus

b

Including three tumors over esophagogastric junction * P \ 0.05

\0.001*

0.004*

\0.001*

0.003*

\0.001*

0.002*

0.047*

0.194

0.027*

0.048*

T1–3N0M0 patients (n = 84) EP2 expression

receptors guarantees the downstream reaction of PGE2. Therefore, it is not surprising that EP2 overexpression was associated with worse survival but COX-2 overexpression was not. From these results, we also assume that combination of EP2 expression status and COX-2 expression status may offer more information on predicting patient outcomes in

ESCC. However, further studies are needed to confirm our speculation. In conclusion, our results showed that EP2 was overexpressed in 43.4% of ESCC specimens. EP2 overexpression was significantly associated with a worse overall survival and correlated positively with the T status

EP2 and Esophageal Squamous Cell Carcinoma

in ESCC. In patients without regional or distant lymph node metastasis, EP2 overexpression indicated a worse prognosis. Furthermore, in the patients with T1-3N0M0 status, EP2 expression was an independent factor of overall survival. These findings suggested that EP2 could serve as a useful molecular marker for ESCC and may be applied as a reference for adjuvant treatment in selective ESCC patients. ACKNOWLEDGEMENTS This study was supported by grants from National Science Council (NSC94-2314-B-010-071) and Tzu Chi Foundation Medicine Mission (TCRD-I9603-03). The authors are indebted to Ms. Li-Ling Yang and Ms. Chang-Yi Lee for their excellent technical assistance.

REFERENCES 1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Smigal C, et al. Cancer statistics, 2006. CA Cancer J Clin. 2006;56:106–30. 2. Department of Health, the Executive Yuan, Republic of China. Statistics of Causes of Death and Cancer Registry Annual Report in Taiwan area, 2005. 3. Law S, Kwong DL, Kwok KF, Wong KH, Chu KM, Sham JS, et al. Improvement in treatment results and long-term survival of patients with esophageal cancer: impact of chemoradiation and change in treatment strategy. Ann Surg. 2003;238:339–47. 4. Berger AC, Farma J, Scott WJ, et al. Complete response to neoadjuvant chemoradiotherapy in esophageal carcinoma is associated with significantly improved survival. J Clin Oncol. 2005;23:4330–7. 5. Le Prise E, Etienne PL, Meunier B, et al. A randomized study of chemotherapy, radiation therapy, and surgery versus surgery for localized squamous cell carcinoma of the esophagus. Cancer. 1994;73:1779–84. 6. Bosset JF, Gignoux M, Triboulet JP, et al. Chemoradiotherapy followed by surgery compared with surgery alone in squamouscell cancer of the esophagus. N Engl J Med. 1997;337:161–7. 7. Urba SG, Orringer MB, Turrisi A, Iannettoni M, Forastiere A, Strawderman M. Randomized trial of preoperative chemoradiation versus surgery alone in patients with locoregional esophageal carcinoma. J Clin Oncol. 2001;19:305–13. 8. Thun MJ, Namboodiri MM, Calle EE, Flanders WD, Heath CW Jr. Aspirin use and risk of fatal cancer. Cancer Res. 1993;53:1322–7. 9. Nozoe T, Ezaki T, Kabashima A, Baba H, Maehara Y. Significance of immunohistochemical expression of cyclooxygenase-2 in squamous cell carcinoma of the esophagus. Am J Surg. 2005;189:110–5. 10. Kuo KT, Chow KC, Wu YC, Lin CS, Wang HW, Li WY, et al. Clinicopathologic significance of cyclooxygenase-2 overexpression in esophageal squamous cell carcinoma. Ann Thorac Surg. 2003;76:909–14. 11. Heeren P, Plukker J, van Dullemen H, Nap R, Hollema H. Prognostic role of cyclooxygenase-2 expression in esophageal carcinoma. Cancer Lett. 2005;225:283–9. 12. Sivula A, Buskens CJ, van Rees BP, Haglund C, Offerhaus GJ, van Lanschot JJ, et al. Prognostic role of cyclooxygenase-2 in neoadjuvant-treated patients with squamous carcinoma of the esophagus. Int J Cancer. 2005;116:903–8. 13. Buskens CJ, van Rees BP, Sivula A, et al. Prognostic significance of elevated cyclooxygenase 2 expression in patients with adenocarcinoma of the esophagus. Gastroenterology. 2002;122: 1800–7.

359 14. Pai R, Szabo IL, Soreghan BA, Atay S, Kawanaka H, Tarnawski AS. PGE(2) stimulates VEGF expression inendothelial cells via ERK2/JNK1 signaling pathways. Biochem Biophys Res Commun. 2001;286:923–8. 15. Sheng H, Shao J, Washington MK, DuBois RN. Prostaglandin E2 increases growth and motility of colorectal carcinoma cells. J Biol Chem. 2001;276:18075–81. 16. Buchanan FG, Wang D, Bargiacchi F, DuBois RN. Prostaglandin E2 regulates cell migration via the intracellular activation of the epidermal growth factor receptor. J Biol Chem. 2003;278: 35451–7. 17. Coleman RA, Smith WL, Narumiya S. International Union of Pharmacology classification of prostanoid receptors: properties, distribution, and structure of the receptors and their subtypes. Pharmacol Rev. 1994;46:205–29. 18. Dey I, Lejeune M, Chadee K. Prostaglandin E2 receptor distribution and unction in the gastrointestinal tract. Br J Pharmacol. 2006;149:611–23. 19. Castellone MD, Teramoto H, Williams BO, Druey KM, Gutkind JS. Prostaglandin E2 promotes colon cancer cell growth through a Gs-axin-beta-catenin signaling axis. Science. 2005;310:1504–10. 20. Sonoshita M, Takaku K, Sasaki N, et al. Acceleration of intestinal polyposis through prostaglandin receptor EP2 in Apc (Delta 716) knockout mice. Nat Med. 2001;7:1048–51. 21. Han SW, Roman J. Suppression of prostaglandin E2 receptor subtype EP2 by PPAR gamma ligands inhibits human lung carcinoma cell growth. Biochem Biophys Res Commun. 2004;314:1093–9. 22. Kamiyama M, Pozzi A, Yang L, DeBusk LM, Breyer RM, Lin PC. EP2, a receptor for PGE2, regulates tumor angiogenesis through direct effects on endothelial cell motility and survival. Oncogene. 2006;25:7019–28. 23. Sung YM, He G, Hwang DH, Fischer SM. Overexpression of the prostaglandin E2 receptor EP2 results in enhanced skin tumor development. Oncogene. 2006;25:5507–16. 24. Piazuelo E, Jime´nez P, Strunk M, Santander S, Garcı´a A, Esteva F, et al. Effects of selective PGE2 receptor antagonists in esophageal adenocarcinoma cells derived from Barrett’s esophagus. Prostaglandins Other Lipid Mediat. 2006;81:150– 61. 25. Yu L, Wu WK, Li ZJ, et al. E series of prostaglandin receptor 2mediated activation of extracellular signal-regulated kinase/activator protein-1 signaling is required for the mitogenic action of prostaglandin E2 in esophageal squamous-cell carcinoma. J Pharmacol Exp Ther. 2008;327:258–67. 26. Wang LS, Chow KC, Wu CW. Expression and up-regulation of interleukin-6 in oesophageal carcinoma cells by n-sodium butyrate. Br J Cancer. 1999;80:1617–22. 27. Wang YF, Chow KC, Chang SY, Chiu JH, Tai SK, Li WY, et al. Prognostic significance of nm23-H1 expression in oral squamous cell carcinoma. Br J Cancer. 2004;90:2186–93. 28. Watabe A, Sugimoto Y, Honda A, et al. Cloning and expression of cDNA for a mouse EP1 subtype of prostaglandin E receptor. J Biol Chem. 1993;268:20175–8. 29. Honda A, Sugimoto Y, Namba T, et al. Cloning and expression of a cDNA for mouse prostaglandin E receptor EP2 subtype. J Biol Chem. 1993;268:7759–62. 30. Sugimoto Y, Namba T, Honda A, Hayashi Y, Negishi M, Ichikawa A, et al. Cloning and expression of a cDNA for mouse prostaglandin E receptor EP3 subtype. J Biol Chem. 1992;267:6463–6. 31. Miyata Y, Kanda S, Nomata K, Eguchi J, Kanetake H. Expression of cyclooxygenase-2 and EP4 receptor in transitional cell carcinoma of the upper urinary tract. J Urol. 2005;173:56–60. 32. Pai R, Nakamura T, Moon WS, Tarnawski AS. Prostaglandins promote colon cancer cell invasion; signaling by cross-talk

360 between two distinct growth factor receptors. FASEB J. 2003;17:1640–7. 33. Spinella F, Rosano L, DiCastro V, Natali PG, Bagnato A. Endothelin-1-induced prostaglandin E2-EP2, EP4 signaling regulates vascular endothelial growth factor production and ovarian carcinoma cell invasion. J Biol Chem. 2004;279:46700–5. 34. Han C, Wu T. Cyclooxygenase-2-derived prostaglandin E2 promotes human cholangiocarcinoma cell growth and invasion

K.-T. Kuo et al. through EP1 receptor-mediated activation of the epidermal growth factor receptor and Akt. J Biol Chem. 2005;280:24053– 63. 35. Han C, Michalopoulos GK, Wu T. Prostaglandin E2 receptor EP1 transactivates EGFR/MET receptor tyrosine kinases and enhances invasiveness in human hepatocellular carcinoma cells. J Cell Physiol. 2006;207:261–70.