• non-muscle-invasive bladder cancer;
  • COX-2;
  • angiogenesis;
  • recurrence


  1. Top of page
  2. Abstract


To investigate the effect of cyclooxygenase-2 (COX-2) on microvessel density (MVD) and on the clinical prognosis in patients with non-muscle invasive urothelial carcinoma of the bladder, as COX-2 expression is significantly greater in epithelial tumours and there is increasing evidence that COX-2 might contribute to tumour neovascularization.


We assessed tumour samples from 110 patients undergoing transurethral resection for primary pTa/pT1 bladder carcinoma (pTa, 84; pT1, 26; grade 1, 22; grade 2, 81; grade 3, seven). Paraffin sections were assessed immunohistochemically using antibodies against COX-2, CD34 (endothelial cells) and CD105 (proliferating vessels). COX-2 expression was quantified by the number of stained cells (negative, +, ++) and the MVD calculated as vessels per field.


Of the 110 tumours, 45 (41%) had no immunostaining for COX-2, 40 had faint staining with at least isolated positive cells (+) and 25 stained ++. COX-2 positive tumours had significantly greater vascularization for proliferating vessels. In COX-2 negative tumours the MVD was 22.1, identified by CD34 immunostaining, and 3.4 for proliferating vessels (CD105), whereas COX-2 positive tumours had a MVD of 18.3 (CD34), and of 5.8, respectively (CD105). Complete follow-up data were available in 91 patients; after a mean follow-up of 25 months, 18 (20%) had tumour recurrences. There was no significant difference in the recurrence rates or disease-free survival between COX-2-positive (19%, 25.6 months) or -negative patients (21%, 25.2 months).


These results confirm the involvement of COX-2 in angiogenesis in bladder cancer, as COX-2 promoted blood vessel proliferation in the tumour zone, and indicate the usefulness of COX-2-inhibiting drugs in preventing and treating superficial bladder cancer.


  1. Top of page
  2. Abstract

Cyclooxygenases (COX) catalyse the conversion of arachidonic acid to prostaglandin peroxide, the rate-limiting step in the biosynthesis of prostaglandins and thromboxanes. Two isoforms of COX have been identified; COX-1 is constitutively expressed in normal tissues and involved in regulating constitutive cell processes. In contrast, COX-2 is frequently undetectable in normal tissues, but it can be induced by several stimuli, including cytokines and tumour promoters [1–3]. Overexpression of COX-2 is associated with inflammatory diseases such as rheumatoid arthritis [4], and with several kinds of tumours, including lung cancer, cancers of the gastrointestinal tract, breast cancer, bladder cancer and prostate cancer [5–13].

The complex mechanism of COX-2 in tumorigenesis is incompletely understood; the overexpression of COX-2 has been well documented in promoting tumour growth. Also, the inhibition of COX by NSAIDs can reduce the relative risk of death from colon cancer [14]. There is increasing evidence that COX-2 is a promoter for tumour angiogenesis. COX-2, similar to several cytokines and growth factors, has various effects on angiogenesis by regulating the urokinase type plasminogen activator [15]. Otherwise, it is well known that prostaglandin E2, the product that is catalysed by COX-2, is directly involved in the formation of new blood vessels [16,17].

The effect of COX-2 on the angiogenesis of bladder cancer and its prognostic value has been rarely studied. The aim of the present study was to investigate the correlation of COX-2 expression with microvessel density (MVD), assessed using antibodies against CD34 (endothelial cells) and CD105 (newly formed vessels). We also assessed the effect of COX-2 expression and MVD on the clinical follow-up of patients with pTa/pT1 bladder cancer.


  1. Top of page
  2. Abstract

The study included 110 patients undergoing transurethral resection for primary urothelial carcinoma of the bladder (median age 68 years, range 33–86; 25 women and 85 men). The pathological diagnosis was established using haematoxylin and eosin (H&E)-stained sections, and the tumours classified according to the Fifth edition of the UICC and the WHO [18]. The histopathology showed pTa tumours in 84 patients and pT1 in 26, with 22 tumours classified as grade 1, 81 as grade 2 and seven as grade 3. Adjuvant intravesical therapy with chemo- or immunotherapy was administered after surgery in 91 of the patients (83%).

The follow-up assessments were undertaken by office urologists according to the guidelines of the European Association of Urology, and included cystoscopy every 3 months. All tumour recurrences were histologically confirmed.

For immunohistochemical investigations, formalin-fixed, paraffin-embedded, 4 µm sections were used; after dewaxing and rehydrating, endogenous peroxidase was blocked with 3% H2O2. For COX-2 immunohistochemistry, the COX-2 antibody was used (Oxford Biochemicals, Oxford, MI) with the TSA detection system (Perker-Elmer, Foster City, CA). Sections from patients with a chronic inflammatory disease (ulcerative colitis) were used as positive controls, with negative controls comprising benign urothelium and omitting the first antibody.

After blocking the endogenous peroxidase, the cells were permeabilized with trinitrobenzenesulphonic acid, 2% saponin and 0.3% Triton-X for 30 min at room temperature. After washing with Tris-buffered saline + 0.5% Tween 20 (TNT), the primary antibody was applied (diluted 1 : 200) and incubated overnight at 4 °C. Horseradish peroxidase-labelled secondary antibody (Dianova, Hamburg, Germany) was then applied and the slides incubated for 30 min at room temperature. After washing with TNT, the slides were incubated with tyramine-streptavidine for 10 min at room temperature. Two un-amplified controls were used for every sample. The slides were then incubated with avidin-biotin, the reaction detected with diaminobenzidine, counterstained with H&E, and the slides dehydrated, clarified in xylene and mounted. The slides were examined at × 200 and × 400; samples with 1–10% stained cells were classified as faint (+) and those with ≥ 10% stained cells as positive (++).

For CD34 immunohistochemistry, the Class II Clone QBEnd 10 mouse antihuman monoclonal antibody was used (Dako, Carpenteria, CA). Antigen was unmasked by microwave heating of the slides three times for 5 min in citrate buffer (pH 6.0) at 700 W. After washing in PBS, the primary antibody was applied for 1 h at room temperature. For the detection system, LSAB2 (DAKO, K0675) was used with secondary antibody as the biotinylated link (30 min) followed by streptavidin-peroxidase (30 min). After every step, the slides were washed in PBS (pH 7.4). The reaction was detected by incubating with diaminobenzidine for 5 min. After counterstaining with H&E, the slides were dehydrated, clarified and mounted, and examined at × 400. The stained vessels were counted in five consecutive fields from the representative tumour zone and the mean value considered as the MVD.

For CD105, antigen was unmasked with trypsin 2% for 30 min at 37 °C. After washing with PBS (pH 7.4), the anti-CD105 (GP160, Dako M3537; diluted 1 : 15) primary antibody was applied for 1 h at room temperature. For the detection system, LSAB+ (Dako, K0690) as used with the secondary antibody as the biotinylated link (30 min), followed by streptavidine-peroxidase (30 min). Positivity was detected using diaminobenzidine for 5 min and the slides counterstained with H&E. After dehydrating and clarifying, the slides were mounted and examined at ×400. Positive vessels were counted in five consecutive microscopic fields, considering the mean value as the MVD.

The results were assessed statistically with commercial software, using Fisher's exact test to assess the correlation between immunohistochemical findings with conventional clinical features, e.g. pathological stage, histological stage and tumour recurrence rate. The Kaplan-Meier method was used to derive the recurrence-free survival and the log-rank test to compare curves for two or more groups; P < 0.05 was considered to indicate significant differences between groups [10].


  1. Top of page
  2. Abstract

Follow-up data for 91 of the 110 (83%) patients were available for further analysis; the median (range) follow-up was 26 (4–52) months. Of the 91 patients, 18 (20%) had tumour recurrences; one recurrent carcinoma progressed to muscle-invasive disease, and this patient underwent radical cystectomy and reconstruction of an ileal neobladder.


All negative control samples stained negatively for COX-2, as did random bladder samples from benign bladder mucosa. The positive controls (samples from ulcerative colitis) showed homogenous and strongly positive COX-2 expression. In contrast to the positive controls, the samples of urothelial carcinoma showed a weaker and more heterogeneous staining (Fig. 1) . The details of staining with stage and grade are shown in Table 1.


Figure 1. pTa grade II urothelial carcinoma showing positive COX-2 staining.

Download figure to PowerPoint

Table 1.  COX-2 expression with stage and grade in 110 samples
GroupNStain, n (%)
pTa  8439 (46)28 (33)17 (20)
pT1  26  6 (23)12 (46)  8 (31)
G1  2212 (55)  8 (36)  2 (9)
G2  8132 (40)28 (35)21 (26)
G3  7  1  4  2
Total11045 (41)40 (36)25 (23)

For vascular markers, CD34 staining gave a mean (sd) MVD for all investigated tumours of 19.49 (8.78) vessels/field, but the MVD was not significantly different between pTa, at 19.97 (8.49), and pT1 tumours, at 18.82 (10.05) vessels/field. For CD105 staining, the mean MVD of all tumours was 4.6 (4.6) vessels/field; there was no difference in MVD between pTa, at 4.5 (4.6) and pT1 tumours, at 5.0 (6.2) vessels/field.


Tumours with negative COX-2 staining had a MVD (CD34) of 22.5 (13.7) and those with faint or positive COX-2 expression 24.1 (13.7) vessels/field (not significant, P = 0.98). For proliferating vessels (CD105) the respective values were 5.5 (4.3) and 6.7 (5.8) (not significant); only tumours with COX-2 expression of > 10% had a significantly higher MVD than the COX-2-negative tumours, at 7.6 (P = 0.04).

In patients with tumours completely negative or faint/positive staining for COX-2 the recurrence rates were 21% and 19%. with recurrence-free survivals of 25.2 and 25.6 months, respectively. Kaplan-Meier analysis showed no significant difference (P > 0.05).


  1. Top of page
  2. Abstract

There is increasing evidence for the important role of COX-2 in several types of tumours; experimental and epidemiological studies show that tumour development can be reduced by inhibitors of COX-2. Although COX-2 catalyses the initial step in the formation of prostaglandins, the complex mechanism of COX-2 action remains unclear. Many reports suggest that one of the major issues in the role of COX-2 is in promoting angiogenesis. In a corneal model, Masferrer et al.[19] shows that neovascularization could be blocked by celecoxib, but not by a COX-1 inhibitor. In a prostate cancer mouse model, Liu et al.[13] described an inhibition of tumour growth and decreasing MVD after treatment with a selective COX-2 inhibitor. In a clinical study, Marrogi et al.[20] showed a correlation between COX-2 expression and MVD detected with the CD31 antibody in patients with non-small cell lung carcinoma. Other authors describe an association of COX-2 expression with increased angiogenesis in breast cancer and sporadic colorectal adenomas [12,21].

Few results are available for the role of COX-2 in bladder cancer; Kömhoff et al.[22] investigated COX-2 expression in specimens of bladder cancer with immunohistochemistry, e.g. by in situ hybridization. They found that high-grade tumours had significantly higher COX-2 expression than low- and intermediate-grade cancers. Mohammed et al.[7] investigated the expression of COX-2 in invasive bladder carcinoma, e.g. in samples of carcinoma in situ, finding that most overexpressed COX-2; 86% of invasive carcinomas and 75% of carcinoma in situ overexpressed COX-2. Yoshimura et al.[23] investigated the expression of COX-2 in bladder carcinoma and found that COX-2 expression correlated well with increasing tumour stage and grade. In the present study, MVD was assessed using anti-CD34 and anti-CD105 antibodies, the former detecting endothelial cells of neo-angiogenetic vessels and pre-existing vessels, and the latter reportedly preferentially binding to endothelial cells of newly formed vessels [24,25]. We confirmed that COX-2 expression is associated with invasion of the lamina propria. In contrast to the data of Kömhoff et al.[22], the present results showed no association with tumour grade. When correlating COX-2 expression with MVD, the results differed for the different antibodies; for CD34 there was no difference among tumours with negative, faint or positive COX-2 staining, but for proliferating vessels, positive tumours had a significantly higher MVD than those with negative or only faint expression, indicating that COX-2 is more effective in early vascular formation than in vascular stabilization.

The finding that inhibition of COX-2 may reduce the risk of cancer development indicates that the expression of COX-2 in tumour samples might be used as a prognostic factor. Several authors have investigated the prognostic role of COX-2 in different types of tumours, reporting prognostic value for gynaecological cancers, including squamous cell carcinoma of the cervix [26,27], breast cancer [28] and ovarian cancer [29]. COX-2 was also reported as a prognostic marker for tumours of the respiratory tract, including non-small cell carcinoma [30] or adenocarcinoma of the lung [31]. In addition, for patients with colorectal carcinoma, Konno et al.[32] reported that the COX-2 level was significantly associated with lymphatic node invasion, and was a prognostic factor in a univariate, but not a multivariate, analysis.

There are few studies of the prognostic effect of COX-2 expression in bladder carcinoma. Kim et al.[33] investigated the prognostic value of COX-2 in selected patients with pT1G3 bladder carcinoma, who were treated with transurethral resection and intravesical immunotherapy. In the latter group COX-2 was significantly associated with tumour recurrence and progression. Shirahama et al.[34] investigated the prognostic value of COX-2 in patients undergoing cystectomy for advanced bladder carcinoma. In contrast to Kim et al., they could not confirm the prognostic value in bladder cancer for these patients. In 108 patients undergoing cystectomy COX-2 was expressed in 31% of the tumours. In a univariate analysis, COX-2 expression was significantly associated with local, lymphatic and venous invasion, but it was not associated with the patients’ prognosis. The authors concluded that COX-2 was important in bladder cancer development but not in the prognosis. In the present homogenous group of 110 patients with primary non-muscle invasive bladder cancer, who were treated with transurethral resection and intravesical chemo- or immunotherapy, there was no correlation between COX-2 expression and either the recurrence rate or the recurrence-free interval. COX-2 expression was associated with tumour invasion but not with recurrence or progression. That COX-2 expression is associated with invasion of the lamina propria underlines the effect of COX-2 on the development of early bladder carcinoma; COX-2 also seems to be involved in the development of tumour vascularization.

In conclusion, in contrast to other cancers, COX-2 does not seem to be a useful prognostic marker for tumour recurrence or progression in pTa/pT1 bladder cancer. The present study indicates the involvement of COX-2 in the development and angiogenesis of non-muscle invasive bladder cancer. These data support the possible utility of COX-2 inhibitors in the prevention and prophylaxis of recurrence in superficial bladder cancer.


  1. Top of page
  2. Abstract
  • 1
    Needleman P, Turk J, Jakschik BA, Morrison AR, Lefkowith JB. Arachidonic acid metabolism. Ann Rev Biochem 1986; 55: 69102
  • 2
    Herschman HR. Prostaglandin synthase 2. Biochim Biophys Acta 1996; 1299: 12540
  • 3
    Masferrer JL, Isakson PC, Seibert K. Cyclooxygenase-2 inhibitors. a new class of anti-inflammatory agents that spare the gastrointestinal tract. Gastroenterol Clin North Am 1996; 25: 36372
  • 4
    Simon LS. Role and regulation of cyclooxygenase-2 during inflammation. Am J Med 1999; 106: 37S42S
  • 5
    Lim HY, Joo HJ, Choi JH et al. Increased expression of cyclooxygenase-2 protein in human gastric carcinoma. Clin Cancer Res 2000; 6: 51925
  • 6
    Yip-Schneider MT, Barnard DS, Billings SD et al. Cyclooxygenase-2 expression in human pancreatic adenocarcinomas. Carcinogenesis 2000; 21: 13946
  • 7
    Mohammed SI, Knapp DW, Bostwick DG et al. Expression of cyclooxygenase-2 (COX-2) in human invasive transitional cell carcinoma (TCC) of the urinary bladder. Cancer Res 1999; 59: 564750
  • 8
    Hixson L, Alberts D, Krutzsch M et al. Antiproliferative effect of nonsteroidal anti-inflammatory drugs (NSAIDs) against human colon cancer cells. Cancer Epidemiol Biomarkers Prev 1994; 3: 4338
  • 9
    Hida T, Yatabe Y, Achiwa H et al. Increased expression of cyclooxygenase 2 occurs frequently in human lung cancers, specifically in adenocarcinomas. Cancer Res 1998; 58: 37614
  • 10
    Wolff H, Saukkonen K, Anttila S et al. Expression of cyclooxygenase-2 in human lung carcinoma. Cancer Res 1998; 58: 49975001
  • 11
    Tucker ON, Dannenberg AJ, Yang EK et al. Cyclooxygenase-2 expression is up-regulated in human pancreatic cancer. Cancer Res 1999; 59: 98790
  • 12
    Costa C, Soares R, Reis-Filho JS, Leitao D, Amendoeira I, Schmitt FC. Cyclo-oxygenase 2 expression is associated with angiogenesis and lymph node metastasis in human breast cancer. J Clin Pathol 2002; 55: 42934
  • 13
    Liu XH, Kirschenbaum A, Yao S, Lee R, Holland JF, Levine AC. Inhibition of cyclooxygenase-2 suppresses angiogenesis and the growth of prostate cancer in vivo. J Urol 2000; 164: 820 5
  • 14
    Dubois RN, Abramson SB, Crofford L et al. Cyclooxygenase in biology and disease. FASEB J 1998; 12: 106373
  • 15
    Fox SB, Taylor M, Grondahl-Hansen J, Kakolyris S, Gatter KC, Harris AL. Plasminogen activator inhibitor-1 as a measure of vascular remodelling in breast cancer. J Pathol 2001; 195: 23643
  • 16
    Form DM, Auerbach R. PGE2 and angiogenesis. Proc Soc Exp Biol Medical 1983; 172: 2148.
  • 17
    Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during tumourigenesis. Cell 1996; 86: 35364
  • 18
    Sobin LH, Wittekind C. TNM Classification of Malignant Tumours, 5th edn. New York: John Wiley & Sons, 1997
  • 19
    Masferrer JL, Leahy KM, Koki AT et al. Antiangiogenic and antitumour activities of cyclooxygenase-2 inhibitors. Cancer Res 2000; 60: 130611
  • 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: 473944
  • 21
    Chapple KS, Scott N, Guillou PJ, Coletta PL, Hull MA. Interstitial cell cyclooxygenase-2 expression is associated with increased angiogenesis in human sporadic colorectal adenomas. J Pathol 2002; 198: 43541
  • 22
    Komhoff M, Guan Y, Shappell HW et al. Enhanced expression of cyclooxygenase-2 in high grade human transitional cell bladder carcinomas. Am J Pathol 2000; 157: 2935
  • 23
    Yoshimura R, Sano H, Mitsuhashi M, Kohno M, Chargui J, Wada S. Expression of cyclooxygenase-2 in patients with bladder cancer. J Urol 2001; 165: 146872
  • 24
    Tanaka F, Otake Y, Yanagihara K et al. Evaluation of angiogenesis in non-small cell lung cancer. Cancer Clin Res 2001; 7: 34105
  • 25
    Kumar S, Ghellal A, Li C et al. Breast carcinoma: vascular density determined using CD105 antibody correlates with tumour prognosis. Cancer Res 1999; 59: 85661
  • 26
    Kim YB, Kim GE, Cho NH et al. Overexpression of cyclooxygenase-2 is associated with a poor prognosis in patients with squamous cell carcinoma of the uterine cervix treated with radiation and concurrent chemotherapy. Cancer 2002; 95: 5319
  • 27
    Ferrandina G, Lauriola L, Zannoni GF et al. Expression of cyclooxygenase-2 (COX-2) in tumour and stroma compartments in cervical cancer: clinical implications. Br J Cancer 2002; 87: 114552
  • 28
    Ristimaki A, Sivula A, Lundin J et al. Prognostic significance of elevated cyclooxygenase-2 expression in breast cancer. Cancer Res 2002; 62: 6325
  • 29
    Ferrandina G, Lauriola L, Zannoni GF et al. Increased cyclooxygenase-2 (COX-2) expression is associated with chemotherapy resistance and outcome in ovarian cancer patients. Ann Oncol 2002; 13: 120511
  • 30
    Khuri FR, 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: 8617
  • 31
    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: 10015
  • 32
    Konno H, Baba M, Shoji T, Ohta M, Suzuki S, Nakamura S. Cyclooxygenase-2 expression correlates with uPAR levels and is responsible for poor prognosis of colorectal cancer. Clin Exp Metastasis 2002; 19: 52734
  • 33
    Kim SI, Kwon SM, Kim YS, Hong SJ. Association of cyclooxygenase-2 expression with prognosis of stage T1 grade 3 bladder cancer. Urology 2002; 60: 81621
  • 34
    Shirahama T, Arima J, Akiba S, Sakakura C. Relation between cyclooxygenase-2 expression and tumour invasiveness and patient survival in transitional cell carcinoma of the urinary bladder. Cancer 2001; 92: 188 93



microvessel density


haematoxylin and eosin


Tris-buffered saline + 0.5% Tween 20.