Weak expression of cyclooxygenase-2 is associated with poorer outcome in endemic nasopharyngeal carcinoma: analysis of data from randomized trial between radiation alone versus concurrent chemo-radiation (SQNP-01)
© Loong et al; licensee BioMed Central Ltd. 2009
Received: 10 April 2009
Accepted: 10 July 2009
Published: 10 July 2009
Over-expression of cyclooxygenase-2 (COX-2) enzyme has been reported in nasopharyngeal carcinoma (NPC). However, the prognostic significance of this has yet to be conclusively determined. Thus, from our randomized trial of radiation versus concurrent chemoradiation in endemic NPC, we analyzed a cohort of tumour samples collected from participants from one referral hospital.
58 out of 88 patients from this institution had samples available for analysis. COX-2 expression levels were stratified by immunohistochemistry, into negligible, weak, moderate and strong, and correlated with overall and disease specific survivals.
58% had negligible or weak COX-2 expression, while 14% and 28% had moderate and strong expression respectively. Weak COX-2 expression conferred a poorer median overall survival, 1.3 years for weak versus 6.3 years for negligible, 7.8 years, strong and not reached for moderate. There was a similar trend for disease specific survival.
Contrary to literature published on other malignancies, our findings seemed to indicate that over-expression of COX-2 confer a better prognosis in patients with endemic NPC. Larger studies are required to conclusively determine the significance of COX-2 expression in these patients.
Nasopharyngeal carcinoma (NPC) is the sixth most common male cancer in Singapore. The current standard of care for locally advanced NPC is concurrent chemo-radiation, which is associated with increased acute and long term morbidities [1, 2]. Increasing effort has been directed toward developing molecular targeted therapies for the treatment of NPC with increasing interest in cyclooxygenase-2 (COX-2) inhibitors.
COX-2 is a 68 kDA enzyme that catalyses the conversion of arachidonic acid to prostaglandins. Over-expression of COX-2 has been found in a variety of malignancies, both gastrointestinal (colon, oesophagus, stomach, pancreas) as well as outside the gastrointestinal tract (lung, breast, bladder and cervix), and shown to correlate with poorer outcomes [3–6].
We hereby describe a retrospective analysis of 58 samples from patients, diagnosed with endemic NPC, who had previously been randomized into a trial of radiotherapy (RT) alone versus concurrent chemo-radiation (CRT) . The aims of the study were to determine the expression level of COX-2 in our cohort of patients and to correlate this with known prognostic factors and overall and disease free survival. We thought the latter would be of particular interest given that studies pertaining to the prognostic significance of COX-2 expression in endemic NPC have so far delivered mixed results [8, 9].
Materials and methods
Between September 1997 to May 2003, 221 patients were accrued into a randomized phase III trial (SQNP01) comparing RT alone to CRT in patients with World Health Organization type II or III NPC . All patients had stage III or IVA/B NPC . Patients on the RT alone arm received standard-course RT to a dose of 70 Gy in 35 fractions using a modified Ho's technique. Patients on the CRT arm received 3 cycles of concurrent cisplatin on weeks 1, 4 and 7 of RT, followed by a further 3 cycles of adjuvant 5-fluorouracil and cisplatin.
Of the 221 patients, 88 were referred for treatment from a single institution following initial diagnosis of NPC. For logistic reasons, only patients from this hospital were included in this study. 58 out of these 88 patients had sufficient pre-treatment paraffin-embedded biopsy material available for analysis.
Institutional review board approval was obtained.
Archived paraffin blocks of tumor tissue biopsies were sectioned at 4 μm, dewaxed and rehydrated in a graded series of alcohol. This was followed by blockage of endogenous peroxidase in 3% hydrogen peroxide (H2O2) and 0.1% protease, digested for 2 minutes at room temperature. The sections were incubated with COX-2 mouse monoclonal antibody (Neomarkers RM9121-S, Clone SP21, Thermo Fisher Scientific, Cheshire, UK) diluted 1:500 overnight at room temperature. The slides were then washed in 3 changes of tris-buffered saline (pH 7.6) for 2 minutes each before incubation with Dako Envision+ System, Peroxidase (Dako, Glostrup, Denmark) for 30 minutes at room temperature. The peroxidase-catalyzed product was then visualized using Biogenex DAB Chromogen Kit (Biogenex, San Ramon, CA). The specimen was counterstained with Harris Haematoxylin, dehydrated, cleared and mounted in dibutyl-phthalate xylene (DPX) for analysis.
A semi-quantitive immunohistochemical (IHC) assay was used to determine the level of COX-2 expression. A single head and neck histopathologist was assigned to perform the scoring. She was blinded to all patient characteristics including the treatment received. The extent of COX-2 staining was scored from 0 to 3, and the intensity of staining scored from 1 to 4. The scores were then multiplied together and the final scores classified as follow: 0, negligible staining; 1–4, weak staining; 5–8, moderate staining; and 9–12, strong staining. For the purpose of statistical analysis, the cohort was grouped into tumors with negligible or weak staining (N = 34) versus tumors with moderate or strong staining (N = 24) as well as according to the 4 expression levels above.
Student's t-test was used to compare the age between patients with COX-2 IHC and those without COX-2 IHC. Similarly, Fisher's exact test was performed to compare the sex, T status, N status, TNM stage and treatment received between these two groups of patients. Among patients with COX-2 IHC, Fisher's exact test was used to investigate the distribution of IHC scores among those with different N stage, T stage, TNM stage and treatment received. Overall survival and disease specific survival (DSS) (defined as the period from the date of randomization to the date of death due to the disease or the date of the last follow up, whichever is earlier) was analyzed using Kaplan-Meier method and compared using log-rank test. Hazards ratio (HR), together with 95% confidence interval (CI), was reported by means of Cox regression.
Patients with COX-2 IHC
No. (%) (N = 58)
Patients without COX-2 IHC
No. (%) (N = 163)
COX-2 expression, its correlation with known prognostic factors and its impact on overall and disease specific survivals
Distribution of patients with different IHC score by treatment
Distribution of T and N status by IHC scores
To the best of our knowledge, this is only the second study examining COX-2 expression in NPC that involved a cohort of patients treated uniformly as part of a clinical trial. Although the 58 patients were only a subgroup of the entire randomized cohort, they were nonetheless representative as there was no difference in patient, tumor or treatment characteristics between them and the remainder of patients where no COX-2 analysis was performed.
Based on our findings, 71% of NPC expressed COX-2, in keeping with other series which reported similar proportions in the range of 62% to 83% [8, 9, 11, 12]. In the only other report based on a cohort of patients receiving treatment as part of a randomized trial, Chan et al.  reported that the proportion of biopsies showing negligible, weak, moderate or strong intensity of COX-2 staining was 17%, 24%, 32% and 27% respectively; whereas in our study the corresponding figures were 29%, 29%, 14% and 28%. In that same study, univariate analysis showed that patients whose tumors co-expressed COX-2 and hypoxia-inducible factor-1alpha (HIF-1alpha) experienced inferior progression free survival, though this finding was not significant on multivariate analysis. This is in contrast to our findings where patients with weak COX-2 expression were associated with inferior overall and disease specific survival. Unfortunately, our small sample size precluded multivariate analysis, but test for correlation found a statistically significant imbalance in the T status, with T1 tumors having only negligible or weak COX-2 expression. This finding could have suggested that perhaps, insignificant COX-2 expression may be associated with a smaller primary tumour, hence theoretically resulting in better outcomes. However, our results indicated otherwise.
There had been numerous studies analyzing COX-2 expression and their prognostic significance across a multitude of malignancies. They included studies with large sample sizes treated with uniform protocols in a randomized clinical trial setting [13, 14]. In those studies, overexpression of COX-2 was reported to confer a worse prognosis, specifically in those with certain tumour characteristics and having underwent specific treatment modality. These were chiefly prostate tumors treated by radiotherapy and short-term hormonal treatment, breast cancers which were estrogen-receptor positive and treated by breast conserving surgery and radiotherapy, and rectal cancers which received preoperative radiotherapy. To allow us to more appropriately apply COX-2 as a prognostic marker, a better understanding of the mechanisms of interaction between COX-2 and the individual treatment modalities is much needed.
The possible mechanisms that underlie the association of weak, rather than COX-2 overexpression with worse overall survival in endemic NPC is outside the scope of our present study. A plausible explanation could be suggested by a separate finding described in hepatocellular carcinoma, also a viral-associated cancer endemic in our population, where COX-2 was found to be overexpressed in the well-differentiated sub-types and was correlated with the presence of pro-inflammatory cells, macrophages and mast cells [15, 16]. Alternatively, COX-2 overexpression had been shown to confer a growth disadvantage by inducing cell cycle arrest via a prostaglandin-independent mechanism .
Given the discrepancy between the 2 studies in endemic NPC for which analysis was performed on defined patient cohorts treated uniformly in a clinical trial (albeit the study by Chan et al.  included stage II patients while our study only involved stage III and IV patients), future studies with a larger sample size (assuming the same proportion of COX-2 expression in NPC as reported in our series, approximately 160 patients will be required for multivariate analysis) should be performed to show conclusively if weak COX-2 expression independently confers a worse prognosis in these patients and if so, elucidate the mechanisms involved.
This study was funded by the National Medical Research Council Singapore.
- Hughes PJ, Scott PM, Kew J, Cheung DM, Leung SF, Ahuja AT, van Hasselt CA: Dysphagia in treated nasopharyngeal cancer. Head Neck 2000, 22: 393-7. 10.1002/1097-0347(200007)22:4<393::AID-HED13>3.0.CO;2-2View ArticlePubMedGoogle Scholar
- Chambers MS, Rosenthal DI, Weber RS: Radiation-induced xerostomia. Head Neck 2007, 29: 58-63. 10.1002/hed.20456View ArticlePubMedGoogle Scholar
- Matsubayashi H, Infante JR, Winter J, Klein AP, Schulick R, Hruban R, Visavanathan K, Goggins M: Tumour COX-2 expression and prognosis of patients with resectable pancreatic cancer. Cancer Biol Ther 2007, 6: 1569-75.View ArticlePubMedGoogle Scholar
- 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. 10.1016/j.amjsurg.2004.03.019View ArticlePubMedGoogle Scholar
- Sackett MK, Bairati I, Meyer F, Jobin E, Lussier S, Fortin A, Gélinas M, Nabid A, Brochet F, Têtu B: Prognostic significance of cyclooxygenase-2 overexpression in glottic cancer. Clin Cancer Res 2008, 14: 67-73. 10.1158/1078-0432.CCR-07-2028View ArticlePubMedGoogle Scholar
- Haffty BG, Yang Q, Moran MS, Tan AR, Reiss M: Estrogen-dependent prognostic significance of cyclooxygenase-2 expression in early-stage invasive breast cancers treated with breast-conserving surgery and radiation. Int J Radiat Oncol Biol Phys 2008, 71: 1006-13.View ArticlePubMedGoogle Scholar
- Wee J, Tan EH, Tai BC, Wong HB, Leong SS, Tan T, Chua ET, Yang E, Lee KM, Fong KW, Tan HS, Lee KS, Loong S, Sethi V, Chua EJ, Machin D: Randomized trial of radiotherapy versus concurrent chemoradiotherapy followed by adjuvant chemotherapy in patients with American Joint Committee on Cancer/International Union against cancer stage III and IV nasopharyngeal cancer of the endemic variety. J Clin Oncol 2005, 23: 6730-8. 10.1200/JCO.2005.16.790View ArticlePubMedGoogle Scholar
- Chan CM, Ma BB, Hui EP, Wong SC, Mo FK, Leung SF, Kam MK, Chan AT: Cyclooxygenase-2 expression in advanced nasopharyngeal carcinoma-a prognostic evaluation and correlation with hypoxia inducible factor-1alpha and vascular endothelial growth factor. Oral Oncol 2007, 43: 373-8. 10.1016/j.oraloncology.2006.04.004View ArticlePubMedGoogle Scholar
- Chen WC, McBride WH, Chen SM, Lee KF, Hwang TZ, Jung SM, Shau H, Liao SK, Hong JH, Chen MF: Prediction of poor survival by cyclooxygenase-2 in patients with T4 nasopharyngeal cancer treated by radiation therapy: clinical and in vitro studies. Head Neck 2005, 27: 503-12. 10.1002/hed.20178View ArticlePubMedGoogle Scholar
- Greene FL, Page DL, Fleming ID, Fritz AG, Balch CM, Haller DG, Morrow M, editors: American Joint Committee on Cancer AJCC cancer staging manual. 6th edition. New York: Springer; 2002.
- Soo R, Putti TC, Tao Q, Goh BC, Lee KH, Kwok-Seng L, Tan L, Hsieh WS: Overexpression of cyclooxygenase-2 in nasopharyngeal carcinoma and association with epidermal growth factor receptor expression. Arch Otolaryngol Head Neck Surg 2005, 131: 147-52. 10.1001/archotol.131.2.147View ArticlePubMedGoogle Scholar
- Tan KB, Putti TC: Cyclooxygenase-2 expression in nasopharyngeal carcinoma: immunohistochemical findings and potential implications. J Clin Pathol 2005, 58: 535-8. 10.1136/jcp.2004.021923PubMed CentralView ArticlePubMedGoogle Scholar
- de Heer P, Gosens MJ, de Bruin EC, Dekker-Ensink NG, Putter H, Marijnen CA, Brule AJ, van Krieken JH, Rutten HJ, Kuppen PJ, Velde CJ, Dutch Colorectal Cancer Group: Cyclooxygenase-2 expression in rectal cancer is of prognostic significance in patients receiving preoperative radiotherapy. Clin Cancer Res 2007, 13: 2955-60. 10.1158/1078-0432.CCR-06-2042View ArticlePubMedGoogle Scholar
- Khor LY, Bae K, Pollack A, Hammond ME, Grignon DJ, Venkatesan VM, Rosenthal SA, Ritter MA, Sandler HM, Hanks GE, Shipley WU, Dicker AP: COX-2 expression predicts prostate-cancer outcome: analysis of data from RTOG 92-02 trial. Lancet Oncol 2007, 8: 912-20. 10.1016/S1470-2045(07)70280-2PubMed CentralView ArticlePubMedGoogle Scholar
- Cervello M, Montalto G: Cyclooxygenases in hepatocellular carcinoma. World J Gastroenterol 2006, 12: 5113-21.PubMed CentralPubMedGoogle Scholar
- Bae SH, Jung ES, Park YM, Kim BS, Kim BK, Kim DG, Ryu WS: Expression of cyclooxygenase-2 (COX-2) in hepatocellular carcinoma and growth inhibition of hepatoma cell lines by a COX-2 inhibitor, NS-398. Clin Cancer Res 2001, 7: 1410-8.PubMedGoogle Scholar
- Trifan OC, Smith RM, Thompson BD, Hla T: Overexpression of cyclooxygenase-2 induces cell cycle arrest: Evidence for a prostaglandin-independent mechanism. J Biol Chem 1999, 274: 34141-7. 10.1074/jbc.274.48.34141View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.