Lung cancer has become the leading cause of cancer

Transcription

1 Prognostic Significance of Cyclooxygenase 2 Expression in 259 Cases of Non Small Cell Lung Cancer Alvaro C. Laga, MD; Dani S. Zander, MD; Philip T. Cagle, MD Context. Previous studies report that increased expression of cyclooxygenase 2 (COX-2) correlates with poor clinical outcome in several malignancies, including non small cell lung carcinoma (NSCLC). Cyclooxygenase 2 inhibitors have been reported to effectively inhibit carcinogenesis in colon cancer experimental models. Objective. We examined COX-2 expression in 259 cases of NSCLC to evaluate its prognostic significance. Design. Sections of NSCLC from patients with a median 5-year follow-up were immunostained with COX-2 monoclonal antibody (1:150) using the Dako mouse En- Vision system. Extent of COX-2 expression in neoplastic cells was recorded as follows: 0, 0% to 10% of cells positive; 1, 11% to 33% positive; 2, 34% to 66% positive; and 3, more than 66% positive. Intensity was scored as either increased ( ) or not increased ( ), compared to internal control smooth muscle and endothelial cells. Kaplan-Meier analysis was used to assess the relationship between survival and COX-2 expression, using the log-rank test for statistical significance. Results. No relationship was found between the extent and/or the intensity of COX-2 expression and patient survival when the entire cohort was considered. However, when separately analyzed according to disease stage and intensity of COX-2 expression, a significant relationship (P.03) between increased COX-2 expression and shortened patient survival was found only in patients with stage I and II NSCLC. Conclusions. To our knowledge, this is the largest series of NSCLCs in which COX-2 has been investigated as a prognostic marker. The findings in this large series support previous studies of smaller cohorts that reported that increased COX-2 expression predicts poor outcome in patients with early-stage NSCLC. (Arch Pathol Lab Med. 2005;129: ) Lung cancer has become the leading cause of cancer mortality in both men and women, and was responsible for an estimated deaths in 2003, accounting for approximately 28% of all cancer deaths. The 5-year survival rate for lung cancer is only.9%, compared to 62.3% for colon cancer, 86.6% for breast cancer, and 97.5% for prostate cancer. 1 The high lethality of lung cancer argues strongly for the investigation of new prognostic markers and therapeutic agents. Up to now, only stage and performance status have proved to be reliable variables in predicting survival prognosis in patients with non small cell lung carcinoma (NSCLC). 2 Cyclooxygenase 2 is a key inducible enzyme involved in the production of prostaglandins and eicosanoids, and has been associated with inflammation and carcinogenesis. 3 5 An increased expression of COX-2 has been reported in several human malignancies, including gastric, 6 colorectal, 7 esophageal, 8 breast, 9 and NSCLC, specifically pulmonary adenocarcinoma. 10 Furthermore, an association of Accepted for publication May 16, From the Department of Pathology, Baylor College of Medicine, Houston, Tex (Dr Laga); Department of Pathology and Laboratory Medicine, University of Texas Health Sciences Center, Houston Medical School, Houston (Dr Zander); and Department of Pathology, The Methodist Hospital, Houston, Tex (Dr Cagle). The authors have no relevant financial interest in the products or companies described in this article. Reprints: Philip T. Cagle, MD, Department of Pathology, The Methodist Hospital, 6565 Fannin St, Main Bldg, Room 227, Houston, TX ( pcagle@tmh.tmc.edu). COX-2 overexpression with poor prognosis in NSCLC has been proposed by various investigators, and selective and nonselective COX-2 inhibitors have been shown to inhibit primary tumor growth and the number and size of metastases in the lungs of experimental animal models, as well in various lung cancer cell lines in a dose-dependent manner. 18 However, a study by Marrogi et al 19 found no correlation between COX-2 expression and patient survival in 106 cases of NSCLC, and most recently, a study by Yamaguchi et al 20 evaluating the prognostic value of COX- 2 among other parameters found that any COX-2 expression decreased the hazard and increased the survival time in patients with resected lung adenocarcinoma. In addition, COX-2 overexpression has been found to protect, rather than sensitize, knockout mice to skin tumor development in one study. 21 The aim of this study was to further investigate the potential relationship between COX-2 expression and prognosis in a cohort of 265 patients with NSCLC. To our knowledge, this is the largest series of NSCLCs in which COX-2 has been investigated as a prognostic marker. MATERIALS AND METHODS Patients and Specimens Two hundred sixty-five cases of NSCLC with long-term followup (median of 5 years) were retrieved from the files of Baylor College of Medicine, Houston, Tex. The samples were obtained between 1975 and 1991 from attempted curative whole-tumor surgical excisions. All available tissue blocks for each patient were reviewed by a pulmonary pathologist (P.T.C.), and one best rep- Arch Pathol Lab Med Vol 129, September 2005 COX-2 Expression in Non Small Cell Lung Cancer Laga et al 1113

2 resentative block containing viable tumor was selected for each case. Exclusion of 1 case because of diagnosis of mixed small cell and non small cell carcinoma, 3 cases with a diagnosis of adenosquamous carcinoma, and 2 other cases because of unverifiable follow-up yielded a cohort of 259 cases with clinical followup that were assessable for COX-2 expression. Of these, 150 were men and 109 were women, with a mean age of 64 years at the time of diagnosis. Distribution according to stage and cell type was as follows: stage I, n 188; stage II, n 35; stage III, n 22; stage IV, n ; adenocarcinoma, n 172; squamous cell carcinoma, n 52; and large cell carcinoma, n 35. Nineteen cases of previously diagnosed bronchioloalveolar carcinoma were reclassified as adenocarcinoma of mixed type under the new World Health Organization definition, leaving no true bronchioloalveolar carcinomas in the study. The median follow-up was 60 months (range, months). Histologic classification was performed according to World Health Organization criteria, and pathologic staging was performed according to the international staging system for lung cancer. 22,23 The Institutional Review Board at Baylor College of Medicine approved this project. Immunohistochemistry Four-to5- m-thick, formalin-fixed, paraffin-embedded tissue sections were deparaffinized in xylene, treated with 0.3% hydrogen peroxide to block endogenous peroxidase activity, and steamed in phosphate-buffered saline (ph 7.2) for antigen retrieval. Mouse monoclonal anti COX-2 antibody (Cayman Chemical, Ann Arbor, Mich) was used as the primary antibody at a dilution of 1:150 using the Dako mouse EnVision system (DakoCytomation, Carpinteria, Calif). Purified normal mouse serum was used for the negative control. Sections of colon adenocarcinoma were used as positive controls. Two observers blinded to all clinical and histopathologic data evaluated the results independently. In cases of discrepancy, consensus was achieved after discussion of the findings with the aid of a multiheaded microscope. The proportion of COX-2 immunoreactive neoplastic cells was recorded as 0, 1, 2, and 3, with 0 indicating 0% to 10% of neoplastic cells immunopositive; 1, 11% to 33% of tumor cells positive; 2, 34% to 66% of tumor cells positive; and 3, more than 66% of tumor cells positive. Intensity was scored either as increased ( ) or not increased ( ) compared to internal control smooth muscle and endothelial cells, as described in a previous study 11 (Figure 1). In cases in which 2 or more patterns of immunohistochemical intensity were observed, the predominant pattern was registered. The degree of tumor staining for COX-2 was also graded according to the sum of the proportion (0 3) and intensity (1 for no increased expression, 2 for increased expression) of cell staining, yielding a potential score of 1 to 5. Statistics Kaplan-Meier analysis was used to assess the relationship between survival and COX-2 expression, with the log-rank test (Cox-Mantel) for statistical significance using WinStat for Excel. The 2 test was used to investigate an association between elevated COX-2 expression and clinicopathologic features. Multivariate analysis of survival time was performed using the Cox proportional hazards modeling of potential factors related to patient survival. RESULTS Of the 259 patients, 44 (17%) showed less than 10% COX-2 immunoreactive neoplastic cells, 30 (11.5%) showed COX-2 immunopositivity in 11 to 33% of tumor cells, 48 (18.5%) showed COX-2 immunopositivity in 34% to 66% of tumor cells, and 137 (53%) showed COX-2 immunopositivity in more than 66% of the tumor cells. There were 136 tumors with markedly increased COX-2 immunoreactivity (Figure 1, A), whereas the remaining 123 cases did not show an increase in COX-2 intensity (Figure 1, B) compared to internal control cells (Figure 1, C). The median staining score with COX-2 for tumors was 4 (range, 1 5). The results showed 44 cases (17%) with a score of 1, 13 cases (5%) with a score of 2, 41 cases (16%) with a score of 3, 66 cases (25%) with score of 4, and 95 cases (37%) with a score of 5. A diffuse cytoplasmic immunostaining pattern was observed in all cases. A significant relationship was found only between elevated COX-2 expression and patient sex, with women demonstrating increased expression more commonly than men (P.01). A trend toward an association between elevated COX-2 expression and tumor cell type was observed, with adenocarcinomas more often demonstrating an increase in COX-2 expression (P.16). There was no significant association between elevated COX-2 expression and age, disease stage, or pack-years of smoking (Table 1). Survival analysis demonstrated no association between COX-2 expression and poor outcome when the proportion of positive neoplastic cells was considered, either in the entire cohort or after controlling for disease stage and/or tumor cell type. We also evaluated survival according to a score combining the proportion of COX-2 positive cells and the intensity of COX-2 expression, and no significant association between high score and shortened survival was shown in the entire cohort (P.43) or after controlling for stage (P.27) of disease and tumor cell type (data not shown). After analyzing the entire cohort according to COX-2 expression intensity, with increased COX-2 expression intensity designated with a plus sign and the lack of increase in COX-2 expression designated with a minus sign (as described previously in a study by Achiwa et al 11 ), no significant relationship between increased COX-2 expression and decreased survival was found (P.28; Figure 2, A). However, after controlling COX-2 expression intensity by disease stage, we observed a significant association between increased ( ) COX-2 immunohistochemical expression and poor survival for patients with NSCLC stage I and II disease. Patients with stage I and II NSCLC and increased COX-2 expression had a 64% survival at 5 years, compared to a 72% 5-year survival rate for patients without increased COX-2 expression (P.03; Figure 2, B). We also performed multivariate analysis to identify which factor(s) would be significantly related to survival time in patients with stage I and stage II NSCLC, using age, sex, cell type, stage, and COX-2 expression as covariates. This analysis showed a significant effect for stage (coefficient [stage II/stage I] ; 95% CI, ; P.001) and increased COX-2 expression (coefficient [COX-2 /COX-2 ] ; 95% CI, ; P.03; Table 2). COMMENT Currently, clinicopathologic staging is the standard system providing the most important prognostic information for patients diagnosed with NSCLC; therefore, it is the basis for therapy in patients with NSCLC. However, this system has proved to be far from ideal. Approximately 40% of patients with NSCLC T1 T2 without nodal involvement and/or metastasis will still experience recurrent disease after apparently complete resection of their primary tumor. 2 In this context, various investigators have evaluated different molecular markers, including COX-2, that are thought to play an important role in the devel- 11 Arch Pathol Lab Med Vol 129, September 2005 COX-2 Expression in Non Small Cell Lung Cancer Laga et al

3 Figure 1. Representative sections of lung adenocarcinoma both with (A) and without (B) increased cyclooxygenase 2 immunoreactivity, as compared to endothelial cells used as internal control (C) (diffuse cytoplasmic staining, original magnification 100). Figure 2. Survival curve for all 259 patients according to the presence or absence of increased cyclooxygenase 2 (COX-2) expression. A, Overall survival was better for patients without increased COX-2 expression, but the difference was not significant (P.28). B, Survival curve for 223 patients with stage I and stage II disease by COX-2 status. We found a statistically significant shortened survival time for patients with increased COX-2 expression as compared to patients without such an increase (P.03). Arch Pathol Lab Med Vol 129, September 2005 COX-2 Expression in Non Small Cell Lung Cancer Laga et al 1115

4 Table 1. Correlation Between Elevated Cyclooxygenase 2 (COX-2) Expression and Clinicopathologic Features Clinical Feature Total No. of Cases COX-2* COX-2* P Value All cases Sex Male Female Age, y Stage I II III IV Pack-years Tumor cell type AC SqCC LCC * Increased ( ) and normal ( ) COX-2 expression compared to internal control cells. Smoking history was available for 168 patients; all were smokers. AC indicates adenocarcinoma; SqCC, squamous cell carcinoma; LCC, large cell carcinoma. opment of lung cancer, in an attempt to identify 1 or more markers with prognostic and therapeutic potential. At least 3 lines of evidence strongly support the proposal that COX-2 plays a major role in lung cancer development. First, COX-2 has been shown to be overexpressed in lung cancer ; second, selective and nonselective nonsteroidal anti-inflammatory drugs (NSAIDs) have been reported to prevent lung cancer in a dose-dependent manner in vitro and in experimental animal models 18 ;and third, some epidemiologic studies suggest that regular use of NSAIDs may reduce the incidence of lung cancer. 24 Several mechanisms by which COX-2 contributes to tumorigenesis have been identified. Among these, inhibition of apoptosis, increased angiogenesis, and increased invasive and metastatic potential are thought to be the most important Our results in this large cohort of cases continue to support the concept that COX-2 overexpression predicts shortened survival in patients with early-stage NSCLC. Hida et al 10 reported that COX-2 overexpression occurs frequently in lung adenocarcinomas. Interestingly, we found a trend (P.16) for an association between expression of COX-2 and histologic tumor type. We did not find a significant association between increased COX-2 expression and age, disease stage, and pack-years of smoking. Previous studies by Achiwa et al 11 and Khuri et al 13 found that COX-2 overexpression in stage I adenocarcinomas and NSCLC, respectively, predicted poor patient survival. In addition, a study by Barbender et al 12 investigated COX- 2 messenger RNA expression with real-time polymerase chain reaction in 89 cases of curatively resected NSCLC and concluded that high COX-2 expression was associated with inferior survival. Most recently, a study by Yamaguchi et al 20 evaluated the prognostic significance of COX- 2 among other markers in a cohort of 117 patients with resected adenocarcinoma of the lung. These authors found that any COX-2 staining decreased the hazard and increased the survival time. An explanation for this discrepancy could be simply the method by which they registered and analyzed COX-2 expression. It is well recognized that immunohistochemistry may produce different results depending on the fixative used to preserve tissue, the method of antigen retrieval, and the use of automated versus manual performance, among other factors. In contrast to the studies cited, we used the Dako Envision Plus detection system, which has been reported to have higher sensitivity, allowing for higher antibody dilutions for comparable staining to other techniques, such as the avidinbiotin complex. 30 When we analyzed this large cohort of cases by quantifying the proportion of neoplastic cells staining with COX-2, alone or in combination with COX- 2 intensity and after controlling for stage, we did not find significant results. When we considered the intensity of COX-2 expression alone as compared to internal control cells based on a previous study, 11 we observed that patients with stage I and stage II NSCLC and increased COX-2 expression had a poorer survival than their counterparts without increased COX-2 expression. In addition, our study included a well-defined patient population with not only adenocarcinomas, but also squamous cell carcinoma and large cell carcinoma. In our study, increased immunohistochemical expression of COX-2 proved to predict shortened patient survival for patients with stage I and stage II NSCLC. As is depicted in Table 2, COX-2 expression is an independent prognostic factor for stage I and stage II NSCLC in our study, which also supports the same finding by others The results of this study support the concept that immunohistochemical analysis of molecular markers, namely COX-2, can add prognostic information to that provided by standard staging. This conclusion may serve as a foundation for new therapeutic modalities. We found that increased immunohistochemical expression of COX-2 correlates with shortened survival and Table 2. Cox Proportional Hazards Model for Potential Prognostic Factors in Patients With Stage I and II Non Small Cell Lung Carcinoma Variable Adverse/Favorable Coefficient (95% CI) SE* P Value Age, y Sex Cell type Stage COX-2 64/ 64 Male/female AC/all others Stage II/stage I Increased/not increased * SE indicates standard error of the coefficient ( ) ( ) ( ) ( ) ( ) Arch Pathol Lab Med Vol 129, September 2005 COX-2 Expression in Non Small Cell Lung Cancer Laga et al

5 therefore may help further characterize a subset of patients with stage I and stage II NSCLC who may benefit from therapy with NSAIDs and/or selective COX-2 inhibitors. Further prospective studies are warranted to gain more insight into the role of COX-2 inhibitors as adjuvants in the treatment of NSCLC. References 1. American Lung Association. Trends in lung cancer morbidity and mortality. Available at: Accessed February 10, Mountain CF. Staging classification of lung cancer: a critical evaluation. Clin Chest Med. 2002;23: Eberhart CE, Coffey RJ, Radhika A, Giardiello FM, Ferrenbach S, DuBois RN. Up-regulation of cyclooxygenase 2 expression in human colorectal adenomas and adenocarcinomas. Gastroenterology. 1994;107: Vane JR, Bakhle YS, Botting RM. Cyclooxygenases 1 and 2. Annu Rev Pharmacol Toxicol. 1998;38: Oshima M, Dinchuk JE, Kargman SL, et al. Suppression of intestinal polyposis in Apc knockout mice by inhibition of cyclooxygenase 2 (COX- 2). Cell. 1996;87: Ristimaki A, Honkanen N, Jankala H, Sipponen P, Harkonen M. Expression of cyclooxygenase-2 in human gastric carcinoma. Cancer Res. 1997;57: Sano H, Kawahito Y, Wielder RL, et al. Expression of cyclooxygenase-1 and -2 in human colorectal cancer. Cancer Res. 1997;55: Zimmermann KC, Sarbia M, Weber AA, Borchard F, Gabbert HE, Shror K. Cyclooxygenase expression in human esophageal carcinoma. Cancer Res. 1999; 59: Hwang D, Scollard D, Byrne J, Levine E. Expression of cyclooxygenase-1 and cyclooxygenase-2 in human breast cancer. J Natl Cancer Inst. 1998;90: Hida T, Yatabe Y, Achiwa H, et al. Increased expression of cyclooxygenase- 2 occurs frequently in human lung cancer, specifically in adenocarcinomas. Cancer Res. 1998;58: Achiwa H, Yatabe Y, Hida T, et al. Prognostic significance of elevated cyclooxygenase 2 expression in primary, resected lung adenocarcinomas. Clin Cancer Res. 1999;5: Barbender J, Park JM, Metzger R, et al. Prognostic significance of cyclooxygenase 2 mrna expression in non-small cell lung cancer. Ann Surg. 2002; 235: Khuri F, Wu H, Lee JJ, et al. Cyclooxygenase-2 overexpression is a marker of poor prognosis in stage I non-small cell lung cancer. Clin Cancer Res. 2001; 7: Duperron C, Castonguay A. Chemopreventive efficacy of aspirin and sulindac against lung tumorigenesis in A/J mice. Carcinogenesis. 1997;18: Rioux N, Castonguay A. Prevention of NKK-induced lung tumorigenesis in A/J mice by acetylsalicylic acid and NS-398. Cancer Res. 1998;58: Yao R, Rioux N, Castonguay A, et al. Inhibition of COX-2 and induction of apoptosis: two determinants of nonsteroidal anti-inflammatory drugs chemopreventive efficacies in mouse lung tumorigenesis. Exp Lung Res. 2000;26: Masferrer JL, Leahy KM, Koki AT, et al. Antiangiogenic and antitumor activities of cyclooxygenase-2 inhibitors. Cancer Res. 2000;60: Tsubouchi Y, Mukai S, Kawahito Y, et al. Meloxicam inhibits the growth of non-small cell lung cancer. Anticancer Res. 2000;20: Marrogi AJ, Travis WD, Welsh JA, et al. Nitric oxide synthase, cyclooxygenase 2, and vascular endothelial growth factor in the angiogenesis of non-small cell lung carcinoma. Clin Cancer Res. 2000;6: Yamaguchi NH, Lichtenfels AJ, Demarchi LM, et al. COX-2, MMP-9, and Noguchi classification provide additional prognostic information about adenocarcinoma of the lung. Am J Clin Pathol. 2004;121: Bol DK, Rowley RB, Ho CP, et al. Cyclooxygenase-2 overexpression in the skin of transgenic mice results in suppression of tumor development. Cancer Res. 2002;62: Travis WD, Brambilla E, Müller-Hermelink HK, Harris CC, eds. Pathology and Genetics of Tumours of the Lung, Pleura, Thymus and Heart. Lyon, France: IARC Press; World Health Organization Classification of Tumours; vol American Joint Committee on Cancer. Lung. In: Greene FL, ed. Cancer Staging Manual. 6th ed. New York, NY: Springer Verlag; 2002: Harris RE, Beebe-Donk J, Schuller HM. Chemoprevention of lung cancer by non-steroidal anti-inflammatory drugs among cigarette smokers. Oncol Rep. 2002;9: Strauss GM, Kiatkowski DJ, Harpole DH, Lynch TJ, Skarin AT, Sugarbaker DJ. Molecular and pathologic markers in stage I non-small cell carcinoma of the lung. J Clin Oncol. 1995;13: 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: Castelao JE, Bart RD III, DiPerna CA, Sievers EM, Bremmer RM. Lung cancer and cyclooxygenase-2: review article. Ann Thorac Surg. 2003;76: Nie D, Honn KV. Cyclooxygenase, lipoxygenase and tumor angiogenesis. Cell Mol Life Sci. 2002;59: Nie D, Lamberti M, Zacharek A, et al. Thromboxane A(2) regulation of endothelial cell migration, angiogenesis, and tumor metastasis. Biochem Biophys Res Commun. 2000;267: Sabattini E, Bisgaard K, Ascani S, et al. The EnVision system: a new immunohistochemical method for diagnostics and research: critical comparison with the APAAP, ChemMate, CSA, LABC, and SABC techniques. J Clin Pathol. 1998;51: Arch Pathol Lab Med Vol 129, September 2005 COX-2 Expression in Non Small Cell Lung Cancer Laga et al 1117