Increased cyclooxygenase-2 (COX-2) expression is associated with chemotherapy resistance and outcome in ovarian cancer patients

G. Ferrandina1, L. Lauriola2, G. F. Zannoni2, A. Fagotti1, F. Fanfani1, F. Legge1, N. Maggiano2, M. Gessi2, S. Mancuso1, F. O. Ranelletti3 and G. Scambia1,+

Department of 1 Obstetrics and Gynecology, 2 Pathology and 3 Histology, Catholic University of the Sacred Heart, Rome, Italy

Received 9 November 2001; revised 8 February 2002; accepted 8 March 2002


    Abstract
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Background:

Cyclooxygenase-2 (COX-2) expression is associated with aggressive clinicopathological parameters and unfavourable prognosis in several human malignancies. The aim of this study was to investigate the expression of COX-2 and its association with clinicopathological parameters, response to treatment, and clinical outcome in ovarian cancer patients.

Patients and methods:

COX-2 expression was analysed by immunohistochemistry in 87 primary ovarian carcinomas from patients with measurable disease after primary laparotomy.

Results:

COX-2 immunoreaction was observed in 39 (44.8%) cases, and did not differ in distribution according to age, FIGO stage, debulking at time of surgery, presence of ascites, histotype or tumour grade. Both in patients cytoreduced at first surgery and in those undergoing only explorative laparotomy, the percentage of COX-2 positivity was significantly higher in non-responding than in patients responding to treatment (P = 0.043 and P = 0.0018, respectively). In multivariate analysis, only COX-2 positivity and older age retained an independent role in predicting a poor chance of response to treatment. There was no significant difference of clinical outcome according to COX-2 status in patients undergoing primary debulking while, in the subgroup of patients who underwent explorative laparotomy, COX-2-positive cases showed a shorter time to progression (P = 0.025) and overall survival (P = 0.025).

Conclusions:

The assessment of COX-2 status could provide additional information in order to identify ovarian cancer patients with a poor chance of response to chemotherapy and potentially candidates for more individualised treatments.

Key words: chemotherapy response, COX-2, ovarian cancer, prognosis


    Introduction
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Ovarian cancer is the most common gynaecological malignancy in western countries and represents the fifth leading cause of death due to cancer in women [1]. More than 70% of cases present with advanced stage of the disease at diagnosis and despite advances in cytoreductive surgery and chemotherapy regimens, chemotherapy failure and progression of disease occur [1]. Although several clinicopathological parameters have been reported to be of prognostic significance in ovarian cancer, including tumour FIGO (Federation Internationale de Gynecologie et d’Obstetrique) stage, volume of residual disease, presence of cytologically malignant ascites and grade of tumour differentiation [2], it is conceivable that the assessment of biochemical factors more strictly related to individual tumour cell biology and intrinsic aggressiveness could help in identifying high risk patients and facilitating management of this disease.

Recently, much attention has been focused on the involvement of cyclooxygenase (COX), the key enzyme in the conversion of arachidonic acid to prostaglandins, in critical steps of tumour onset and progression. Starting from epidemiological studies showing a 40–50% lower risk of colorectal cancer [3] and to a lesser extent of breast, prostate, and head and neck cancer [46] in people continuously taking non-steroidal anti-inflammatory drugs, which are well known COX inhibitors, much effort has been devoted to the understanding of the biological activities of COX.

Two COX isoforms have been characterised: COX-1, which is constitutively expressed in almost all tissues where it serves homeostatic functions, and COX-2, which is highly inducible by growth factors, prostaglandins and tumour promoters, and has been mainly associated with the inflammatory response [7]. More recently, several in vitro and pre-clinical studies showed that COX-2 overexpression is associated in colorectal cancer cells with bcl-2 overexpression, apoptosis inhibition, increased adhesion to extracellular matrix, and increased metastatic potential and neoangiogenesis [811]. Moreover, it has also been hypothesised that overexpression of COX-2 could impair host immune responses, as suggested by the ability of COX-2 inhibitors or COX-2 antisense constructs to revert tumour-induced immunosuppression [12, 13].

COX-2 has been found to be overexpressed in the vast majority of colorectal cancers, as well as in other solid tumours, and has been associated with clinicopathological parameters of aggressiveness and unfavourable prognosis [1420]. In particular, we found that COX-2 overexpression is a strong predictor of chemotherapy response and poor outcome in locally advanced cervical cancer patients [21]. As far as ovarian cancer is concerned, only preliminary data have reported that COX-2 is constitutionally expressed in ovarian cancer cell lines and that it is inducible by prostaglandins, resulting in protection from apoptosis [22].

To our knowledge, no data have been reported until now about the expression of COX-2 and its possible clinical significance in ovarian cancer. The aim of the study therefore was to investigate by immunohistochemistry the expression of COX-2 and its association with clinicopathological parameters, response to treatment and clinical outcome in a single institutional series of primary untreated ovarian cancer patients.


    Patients and methods
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
Patients
The study included 87 ovarian cancer patients admitted to the Department of Obstetrics and Gynecology (Division of Gynecologic Oncology, Catholic University of Rome) between February 1995 and April 2000. Staging was performed according to the FIGO classification. In order to make the analysis of the association with response to chemotherapy and outcome as reliable as possible, a highly homogenous series of stage IIIC–IV ovarian cancer patients with measurable disease at time of first surgery were included in the study. Median age was 57 years (range 27–81 years). Sixty-eight (78.2%) patients had stage IIIC, and 19 (21.8%) patients had stage IV disease. Most of the tumours (74.7%) were serous adenocarcinomas and showed a poor grade of differentiation (81.5%). Other clinicopathological characteristics are listed in Table 1.


View this table:
[in this window]
[in a new window]
 
Table 1.  COX-2 expression according to the clinicopathological characteristics of the ovarian cancer population
 
As summarised in Figure 1, showing the flow chart of the overall population, patients were divided into two groups according to the possibility of performing primary debulking: 52 (59.8 %) cases underwent primary debulking with <2 cm or >=2 cm residual disease (n = 40 and n = 12, respectively), while 35 (40.2%) cases were considered unresectable and submitted only to multiple biopsies. Most of these patients had upper abdominal or peritoneal tumours with infiltration of the upper gastrointestinal tract and/or major vessels. All patients, independent of age, underwent four to six cycles of cisplatin-based chemotherapy (total dose at least 400 mg/m2) 2–3 weeks after primary surgery. In particular, patients cytoreduced at first surgery received six cycles of chemotherapy unless they showed clinical progression during treatment. As far as patients undergoing explorative laparotomy are concerned, they received three or four cycles of chemotherapy before attempting a second cytoreductive surgery, unless they showed clinical progression during treatment.



View larger version (17K):
[in this window]
[in a new window]
 
Figure 1. Flow chart of the patient population.

 
In patients cytoreduced at first surgery, response to chemotherapy was assessed by clinical (gynaecological) and ultrasound examination and analysis of CA125 levels, and was recorded according to World Health Organisation (WHO) criteria [23].

In the subgroup of patients who had tumour secondary debulking, response to chemotherapy was assessed by ultrasound examination, computed tomography (CT) scan and CA125 analysis performed before and after chemotherapy. Moreover, a direct assessment of the extent of response to chemotherapy was carried out at the time of the second laparotomy, taking into account the findings of the primary explorative laparotomy.

Immunohistochemistry
Tumour tissues biopsies from primary tumours were obtained at first surgery in all cases. Tissue specimens were fixed in formalin and paraffin-embedded according to standard procedures. Four micrometres of representative blocks from each case were deparaffinised in xylene, rehydrated, treated with 0.3% H2O2 in methanol for 10 min to block endogenous peroxidase activity, and subjected to heat-induced epitope retrieval in a microwave oven using the Dako ChemMate detection kit (Dako, Glostrup, Denmark) according to the manufacturer’s instructions. Slides from all cases studied were then processed simultaneously for immunohistochemistry on the TechMate Horizon automated staining system (Dako) using the Vectastain ABC peroxidase kit (Vector Laboratories, Burlingame, CA, USA). Endogenous biotin was saturated by a biotin blocking kit (Vector Laboratories). Sections were incubated with normal rabbit serum for 15 min, then with rabbit polyclonal antiserum against COX-2 (Cayman, Ann Arbor, MI, USA) diluted 1:300 for 1 h. Negative controls were performed using non-immunised rabbit serum or by omitting the primary antiserum.

Tumour sections showing immunoreaction were scored as positive. The analysis of all tissue sections was performed without any prior knowledge of the clinical parameters by two different pathologists (L.L. and F.O.R.) by means of light microscopy. The proportion of immunostained cells was scored at low magnification (5x objective lens) by evaluating the entire tumour area. When tumour area with positive immunostaining was >10% of the total tumour area, the case was scored as positive. Intensity of staining was also evaluated subjectively using a range from 0 (none) to 1 (feint) to 2 (strong). Cases in which the intensity of staining was scored <2 were considered negative. In case of disagreement (n = 7, 8%), the sections were subjected to a joint re-evaluation.

Statistical analysis
Fisher’s exact test or {chi}2 test were used to analyse the distribution of COX-2-positive cases according to several clinicopathological features.

Overall survival (OS) and time to progression (TTP) were calculated from the date of diagnosis to the date of death/progression or date last seen. Medians and life tables were computed using the product-limit estimate by the Kaplan–Meier method [24], and the log-rank test was employed to assess statistical significance [25]. Multiple logistic analysis [26] was used to analyse the role of clinicopathological parameters and COX-2 staining as a predictor of response to treatment. Statistical analysis was carried out using SOLO (BMDP Statistical Software, Los Angeles, CA, USA).


    Results
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
COX-2 immunostaining
COX-2 immunostaining was observed mainly in the cytoplasm of tumour cells. Figure 2 shows a representative example of an ovarian tumour with intense COX-2 immunostaining (A), compared with a tumour with a very low percentage of COX-2-stained tumour cells (B). Stromal lymphoid cells sometimes showed different intensities of COX-2 immunoreaction. Thirty-nine cases (44.8%) were scored as COX-2 positive.



View larger version (84K):
[in this window]
[in a new window]
 
Figure 2. COX-2 immunostaining in primary ovarian carcinoma. (A) COX-2-positive tumour showing intense cytoplasmatic immunoreaction in tumour cells. (B) COX-2-negative tumour showing only scattered COX-2-positive cells. Bar = 35 µm.

 
Correlation with clinicopathological parameters and response to chemotherapy
No significant association between the extent of residual tumour at surgery and stage of disease was observed, while patients with greater tumour volume showed ascites more frequently than patients with lesser residual disease (P = 0.04). The extent of residual disease was also shown to be associated with poor grade of differentiation (P = 0.04) and older age (P = 0.013). A direct association between older age and poor grade of differentiation was found (P = 0.016). Finally, the presence of ascites was found more frequently in the case of serous adenocarcinomas than in other histotypes (P = 0.013).

Table 1 shows the distribution of COX-2 positivity according to clinicopathological characteristics. COX-2 positivity was not distributed differently according to age, FIGO stage, debulking at time of surgery, presence or absence of ascites, or tumour grade. COX-2 positivity did not show any significant variation according to different histoypes, although it is noteworthy that the only two cases with a mucinous histotype were both COX-2 positive.

With respect to response to chemotherapy, in the group of 52 patients in which cytoreduction was performed there were 41 (78.8%) responses (29 complete and 12 partial responses), while 11 (21.1%) patients were classified as non-responders. In patients who underwent only explorative laparotomy, clinical response was documented in 23 (65.7%) cases, while 12 (34.3%) patients experienced disease progression.

In the group of 52 patients who were cytoreduced at first surgery, COX-2 positivity was found in a statistically significantly higher percentage of non-responding cases (n = 8/11; 72.7%) than in patients responding to chemotherapy (n = 15/41; 36.6%) (P = 0.043). Similar results were obtained in the subgroup of patients undergoing only exploratory laparotomy since the percentage of COX-2 positivity was higher in non-responders (n = 10/12; 83.3%) than in patients responding to treatment (n = 6/23; 26.1%) (P = 0.0018).

In Table 2 the univariate and multivariate analysis of clinicopathological parameters and COX-2 status as predictors of response to chemotherapy in the whole population are summarised.


View this table:
[in this window]
[in a new window]
 
Table 2.  Univariate and multivariate analysis of clinicopathological parameters and COX-2 status as predictors of chemotherapy response in ovarian cancer patients
 
When logistic regression was applied, only COX-2 positivity and older age retained an independent role in predicting a poor chance of response to treatment.

Similar results were obtained after subgrouping in patients undergoing cytoreduction or exploratory laparotomy (data not shown).

Survival analysis
Follow-up data were available for 87 patients. As of May 2001, the median follow-up period was 25 months (range 4–147 months). During the follow-up period, progression and death of disease were observed in 64 (73.6%) and 42 (48.3%) cases, respectively. Figure 3 shows the TTP and OS curves in our population according to COX-2 status.



View larger version (24K):
[in this window]
[in a new window]
 
Figure 3. TTP and OS curves according to COX-2 status in ovarian cancer in cytoreduced cases (A and B) and in cases undergoing explorative laparotomy (C and D).

 
There was no significant difference of clinical outcome in terms of both TTP and OS according to COX-2 status in patients undergoing primary debulking (Figure 3A and B), while in the subgroup of patients who underwent explorative laparotomy (Figure 3C and D), COX-2-positive cases demonstrated a shorter TTP (median 11 months compared with 17 months in COX-2-negative cases; P = 0.025) and OS (median 22 months compared with >50% survival in COX-2-negative cases at last follow-up; P = 0.025).


    Discussion
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
This is the first study demonstrating the association between COX-2 and reduced susceptibility to chemotherapy and poor outcome in a large series of primary advanced ovarian cancer with measurable disease at first surgery.

Overexpression of COX-2 has been reported in tumour cells compared with normal cells, which showed low or no COX-2 expression in several tissue types [16, 17, 19]. Although the mechanism of COX-2 up-regulation is unknown, recent evidence suggests that it could result from the dysregulation of key steps in the epidermal growth factor receptor signalling pathway, namely of ras and mitogen-activated protein kinases [18, 27].

Although we failed to demonstrate an association between COX-2 status and any of the clinicopathological characteristics, a very strong association between COX-2 positivity and a poor chance of response to treatment was observed. More importantly, the ability of COX-2 expression to predict ovarian tumour susceptibility to chemotherapy was retained in multivariate analysis. Furthermore, older age retained an independent role in predicting a poor chance of response to treatment. This could be explained by the correlation between older age and both poor tumour grade of differentiation and the reduced feasibility of cytoreduction.

We reported recently that high COX-2 expression is associated with a poor chance of response to neoadjuvant cisplatin-based chemotherapy in locally advanced cervical carcinoma [21]. Moreover, it is noteworthy that COX inhibitors are able to enhance the cytotoxicity of several chemotherapeutic agents in vitro [28] and to potentiate tumour cell radiosensitivity in vivo [29].

In this context, our current findings support the hypothesis that the association between COX-2 expression and chemoresistance could be a generalised phenomenon that is yet to be clarified on a biochemical level. It is conceivable that the involvement of COX-2 in the biochemical pathways influencing tumour cell susceptibility to cytotoxic agents could play a major role: in particular, COX-2 overexpression has been associated with the function of Her2/neu, which is renowned for inducing resistance to several cytotoxic agents. Moreover, COX-2 has been reported to induce the antiapoptotic bcl-2 protein [30] and to be associated with neoangiogenesis in tumour-bearing mice [31]. Since both inhibition of apoptosis and promotion of neoangiogenesis are strictly related to chemotherapy resistance [32, 33], it is conceivable that COX-2 expression could play a role as an indicator of chemoresistance in ovarian cancer.

The correlation between COX-2 overexpression and unfavourable clinical outcome has been demonstrated in several human malignancies and has been associated with clinicopathological characteristics of tumour aggressiveness such as advanced stage, node involvement or invasive properties [12, 1416, 1820].

It remains to be verified why, although COX-2 positivity seems to identify chemoresistant tumours in patients undergoing both cytoreduction and explorative laparotomy, COX-2 positivity behaves as a marker of poor TTP and OS only in the latter patient group.

It is conceivable that cytoreduction can modify the tumour–host interactions and then tumour biology, hindering the prognostic impact of COX-2 positivity.

An alternative hypothesis is that the selection of chemosensitive cases based on the feasibility of interval debulking surgery (IDS) could be more reliable than in patients optimally debulked at first surgery.

In conclusion, these findings need to be confirmed in a larger series, mainly because the assessment of response rate is a difficult end point in this tumour. However, our study showed that the assessment of COX-2 status could provide additional information in order to identify ovarian cancer patients with a very poor chance of response to chemotherapy, and therefore who are potentially candidates for more individualised treatments.


    Acknowledgements
 
This work was supported financially by grants from the Associazione Italiana per la Ricerca sul Cancro (AIRC).


    Footnotes
 
+ Correspondence to: Dr G. Scambia, Department of Obstetrics and Gynecology, Catholic University of the Sacred Heart, Largo A. Gemelli 8, 00168 Rome, Italy. Tel/Fax: +39-06-35508736; E-mail: giovanni.scambia@libero.it Back


    References
 Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 References
 
1. Lundis SH, Murray T, Bolden S, Wingo PA. Cancer Statistics. Cancer J Clin 1998; 48: 6–13.[Abstract/Free Full Text]

2. Hoskins WJ, Perez CA, Young RC. Principles and Practice of Gynecologic Oncology. Philadelphia, PA: Lippincott Williams and Wilkins 2000; 1005–1007.

3. DuBois RN, Giardiello FM, Smalley WE. Non steroidal anti-inflammatory drugs, eicosanoids and colorectal cancer. Gastroenterol Clin North Am 1996; 25: 773–791.[ISI][Medline]

4. Harris RE, Namboodiri KK, Farrar WB. Non steroidal anti-inflammatory drugs and breast cancer. Epidemiology 1996; 7: 203–205.[ISI][Medline]

5. Norrish AE, Jackson RT, McRae CU. Non steroidal anti-inflammatory drugs and prostate cancer progression. Int J Cancer 1998; 77: 511–515.[ISI][Medline]

6. Funkhouser EM, Sharp GB. Aspirin and reduced risk of esophageal carcinoma. Cancer 1995; 76: 1116–1119.[ISI][Medline]

7. Williams CS, DuBois RN. Prostaglandins endoperoxide synthase: why two isoforms? Am J Physiol 1996; 270: G393–G400.[Abstract/Free Full Text]

8. Tsujii M, DuBois RN. Alterations in cellular adhesion and apoptosis in epithelial cells overexpressing prostaglandin endoperoxide synthase 2. Cell 1995; 83: 493–501.[ISI][Medline]

9. Tsujii M, Kawano S, DuBois RN. Cyclooxygenase-2 expression in human colon cancer cells increases metastatic potential. Proc Natl Acad Sci USA 1997; 94: 3336–3340.[Abstract/Free Full Text]

10. Tsujii M, Kawano S, Tsuji S et al. Cyclooxygenase-2 regulates angiogenesis induced by colon cancer cells [published erratum appears in Cell 1998; 94: 271]. Cell 1998; 93: 705–716.[ISI][Medline]

11. Liu XH, Kirschenbaum A, Yao S et al. Inhibition of cyclooxygenase-2 suppresses angiogenesis and the growth of prostate cancer in vivo. J Urol 2000; 164: 820–825.[ISI][Medline]

12. Huang M, Stolina M, Sharma S et al. Non small cell lung cancer cyclooxygenase-2-dependent regulation of cytokine balance in lymphocytes and macrophages: up-regulation of interleukin 10 and down-regulation of interleukin 12 production. Cancer Res 1998; 58: 1208–1216.[Abstract]

13. Duff M, Mestre JR, Smyth G. COX-2 inhibitor and interferon act synergistically to reduce melanoma growth in a murine model. Proc Am Assoc Cancer Res, 2001; 42: 3670 (Abstr).

14. Tomozawa S, Tsuno NH, Sunamio E et al. Cyclooxygenase-2 overexpression correlates with tumor recurrence, especially haematogenous metastasis of colorectal cancer. Br J Cancer 2000; 83: 324–328.[ISI][Medline]

15. Madaan S, Abel PD, Chaudhari KS et al. Cytoplasmic induction and overexpression of cyclooxygenase-2 in human prostate cancer: implications for prevention and treatment. Br J Urol 2000; 86: 736–741.

16. Tucker ON, Dannemberg AJ, Yang EK et al. Cyclooxygenase-2 expression is upregulated in human pancreatic cancer. Cancer Res 1999; 59: 987–990.[Abstract/Free Full Text]

17. Hwang D, Skollard D, Byerne J, Levine E. Expression of cyclooxygenase 1 and cylooxygenase-2 in human breast cancer. J Natl Cancer Inst 1998; 90: 455–460.[Abstract/Free Full Text]

18. Ohno R, Yoshinaga K, Fujita T et al. Depth of invasion parallels increased cyclooxygenase 2 levels in patients with gastric carcinoma. Cancer 2001; 91: 1876–1881.[ISI][Medline]

19. Sheehan KM, Sheahan K, O’Donoghue DP et al. The relationship between cyclooxygenase-2 expression and colorectal cancer. J Am Med Assoc 1999; 282: 1254–1257.[Abstract/Free Full Text]

20. Kulkarni S, Rader JS, Zhang F et al. Cyclooxygenase-2 is overexpressed in human cervical cancer. Clin Cancer Res 2001; 7: 429–434.[Abstract/Free Full Text]

21. Ferrandina G, Lauriola L, Distefano MG et al. Increased cylooxygenase-2 (COX-2) expression is associated with chemotherapy resistance and poor survival in cervical cancer patients. J Clin Oncol 2002; 15: 973–981.

22. Munkarah A, Morris RT, Baumann P et al. Prostaglandins induced cyclooxygenase-2 expression and reduced apoptosis in epithelial ovarian cancer cells. Proc Soc Gynecol Oncol 2001; 176 (Abstr).

23. World Health Organization. WHO Handbook for Reporting Results of Cancer Treatment. Offset publication no. 48. Geneva: WHO 1979; 16–21.

24. Kaplan E, Meyer P. Non parametric estimation from incomplete observations. J Am Stat Assoc 1958; 53: 457–481.[ISI]

25. Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966; 50: 163–170.[Medline]

26. Cox DR. Analysis of binary data. London: Methven 1970.

27. Sheng H, Shao J, DuBois RN. K-Ras-mediated increase in cyclooxygenase-2 mRNA stability involves activation of the protein kinase B. Cancer Res 2001; 61: 2670–2675.[Abstract/Free Full Text]

28. Hida T, Kozaki K, Muramatsu H et al. Cyclooxygenase-2 inhibitor induces apoptosis and enhances cytotoxicity of various anticancer agents in non small cell lung cancer cell lines. Clin Cancer Res 2000; 6: 2006–2011.[Abstract/Free Full Text]

29. Kishi K, Petersen S, Petersen C et al. Preferential enhancement of tumor radioresponse by cyclooxygenase-2 inhibitors. Cancer Res 2000; 60: 1326–1331.[Abstract/Free Full Text]

30. Liu XH, Yao S, Kirschenbaum A, Levine AC. NS398, a selective cyclooxygenase inhibitor, induces apoptosis and down regulates bcl-2 expression in LNCaP cells. Cancer Res 1998; 58: 4245–4249.[Abstract]

31. Masferrer JL, Leahy KM, Koki AT et al. Antiangiogenic and antitumor activities of cyclooxygenase-2 inhibitors. Cancer Res 2000; 60: 1306–1311.[Abstract/Free Full Text]

32. Chiung EJ, Seong J, Jang WI et al. Spontaneous apoptosis as predictor of radiotherapy in patients with stage IIB squamous cell carcinoma of the uterine cervix. Acta Oncol 1999; 38: 449–454.[ISI][Medline]

33. Cooper RA, Wilks DP, Logue JP et al. High tumor angiogenesis is associated with poorer survival in carcinoma of the cervix treated with radiotherapy. Clin Cancer Res 1998; 4: 2795–2800.[Abstract]