Schistosomiasis is a widespread parasitic infection caused by blood flukes of the genus Schistosoma and transmitted by specific freshwater snails. The infection is reported to have plagued humans since ancient times.[1] Some of the factors that influence the transmission of schistosomiasis in an endemic area include the presence of snail intermediate hosts of the parasites and human contact with the infected waters. Out of the 3 main human-infecting species of Schistosoma (S haematobium, S mansoni, and S japonicum), S haematobium is the predominant species in Africa – endemic in about 53 countries in Africa and the Middle East.[2,3] In Egypt, bladder cancer accounts for about 30% of all cancers, where it is the most common malignancy in men and the second most common malignancy in women after breast cancer,[4–6] and has been associated with many pathogenetic factors – most commonly bilharzial infestation, which is an endemic infection in the Nile River Valley.[6,7] In countries with a long history of schistosomiasis, research studies have identified the following risk factors for infection with S haematobium: male gender, an age < 20 years, living in smaller rural communities, exposure to canal waters, reagent strip detected hematuria and proteinuria, and a history of burning micturition.[1,8–10] Most investigators have accepted the association between schistosomiasis and bladder cancer since the work of Ferguson in 1911.[11] Evidence for this relationship derives from the geographic correlation between the conditions and the distinctive patterns of sex and age at diagnosis; the development of bladder cancer in a younger age group affects males more, is usually associated with schistosomal infection, and shows a high mortality rate and clinicopathologic features of schistosomal-associated bladder cancer (SABC).[4,12,13] The high frequency of squamous cell carcinoma (SqCC) is due to schistosomiasis-infested bladders that frequently show squamous metaplasia and dysplasia of the transitional epithelium.[12,13] A relative increase in the frequency of transitional cell carcinoma (TCC)-associated bladder cancer has been noted.[4]

Studies involving a range of human malignancies have shown significant overexpression of cyclooxygenase (COX)-2 in tumors compared with normal tissues, with an increase in the level of its downstream prostaglandin (PG) products.[14–16] COX catalyzes the conversion of arachidonic acid to PGs by 2 different COX isoforms, COX-1 and COX-2. COX-1 is constitutively expressed in most tissues and mediates the synthesis of PGs required for normal physiologic functions, whereas COX-2 is primarily responsible for PGs produced in inflammatory sites and undetectable in most tissues (normal tissues) under physiologic conditions, but it is induced by cytokines, growth factors, oncogenes, and tumor promoters.[17–20] The PGs produced by COX-2 promote tumor development by stimulating cell proliferation and angiogenesis and by suppressing programmed cell death and immune defense.[21]

Numerous reports from human epidemiologic studies, animal models, and in vitro cell culture experiments now suggest that nonsteroidal anti-inflammatory drugs and COX inhibitors have chemopreventive effects against colon cancer and reduce the risk for colorectal cancer.[16–19,22]

COX-2 has been found to be upregulated and overexpressed in tumors of the colon,[23,24] stomach, pancreas, and lung cancers as well as in bladder cancer, suggesting an important role for COX-2 in their tumorigenesis.[25–27] For the urinary bladder there have been several reports of COX-2 upregulation in nonschistosomal bladder cancer, and on the potential role of nonselective and selective COX-2 inhibitors in suppressing experimental tumorigenesis. This tumor-specific expression of COX-2 suggested to the authors the potential utility of this COX-2 promoter for the construction of a novel replication-selective adenovirus for the treatment of bladder cancer.[28–30]

In the present study, we investigated the predictive value of COX-2 in the development of schistosomal and nonschistosomal bladder lesions because of its potential therapeutic implications. We also investigated and compared the expression of COX-2 in patients with bladder cancer, chronic cystitis, and normal bladder tissue. The results were correlated to the classic prognostic factors, mainly tumor stage and grade, in a trial to determine the prognostic significance of COX-2 marker.

The study included 75 patients (64 men and 11 women) between ages 25 and 75 years. Patients were subjected to detailed history taking, full and complete clinical examination, urine analysis and routine laboratory investigations, urine cytology abdominal-pelvic ultrasonography, excretory urography, cystoscopic examination, and transurethral resection biopsies taken from the apparent lesions. Specimens were obtained from the urology department of Theodor Bilharz Research Institute (TBRI), Cairo, Egypt, and fixed in buffered formalin 10% and sent to the pathology department, TBRI. Serial sections were examined histopathologically and assessed for tumor grade according to the World Health Organization (WHO)[31] and pathologic tumor stage, in accordance with the Union Internationale Contre le Cancer (UICC).[32] Informed consent from all patients underwent cystoscopy, and biopsies from apparent growth and lesions were taken. The study protocol was approved by the Ethics Committee of TBRI according to the Institutional Committee for the Protection of Human Subjects and adopted by the 18th World Medical Assembly, Helsinki, Finland. Diagnosis of schistosomal infestation was based on detection of Schistosoma eggs in tissues and/or detection of circulating Schistosoma antibodies in sera of patients by enzyme-linked immunosorbent assay (ELISA).

All specimens were processed into paraffin blocks; 5-micrometer (mcm)-thick sections were cut on slides, which were treated with TESPA (3-aminopropyl-triethoxysilane, Sigma) for immunohistochemistry (IHC).

The expression of COX-2 was measured in 10 successive high-power fields (x400). COX-2 showed mostly cytoplasmic expression with condensation on the nuclear membrane. Two of the authors (O.H. and A.A.) analyzed the intensity, distribution, and pattern, and evaluated COX-2 immunoexpressions independently. If there was a discrepancy, a consensus was reached after further evaluation. The percentage of positively stained cells was determined semiquantitatively by assessing the whole tumor section, and each sample was assigned to one of the following categories according to the percentage of positive cells: 0 (0% to 4%), 1 (5% to 24%), 2 (25% to 49%), 3 (50% to 74%), or 4 (75% to 100%).[33] Specimens were regarded as COX-2-negative if less than 5% of the cells were stained and COX-2-positive if 5% or above of the cells were stained.

The median age (range) of the 75 patients (64 men and 11 women) was 50.24 (25-75) years. They were grouped accordingly.

• 7 cases with chronic nonschistosomal cystitis and
• 13 cases with chronic schistosomal cystitis, which include
  • 9 cases with premalignant lesions (3 cases with squamous metaplasia, 5 cases showing dysplastic changes, and 1 case with leukoplakia) and
  • 4 cases with simple schistosomal cystitis.

• SqCC: 16 cases; all cases were associated with schistosomiasis, and all were invasive (pT2 + PT4). Four tumors were grade 1 (well-differentiated), 6 grade 2 (moderately differentiated), and 6 grade 3 (poorly differentiated).
• TCC: 34 cases included 15 cases of schistosomal-associated TCC and 19 cases of non-schistosomal-associated TCC. For research purposes, pathologic staging of TCC was identified as 11 superficial tumors (pTa and pT1) and 23 invasive tumors (pT2, pT3, and pT4). Twelve tumors were grade I, 12 grade II, and 10 grade III.

In group I (control cases), there was no expression of COX-2 in the urothelial cells. However, 6 out of 20 cases (30%) in group II (chronic cystitis group) showed positive expression of COX-2 in the urothelial cells (5 of the 6 positive for COX-2 were chronic schistosomal cystitis). The urothelium in these positive cases had focal squamous metaplasia and/or showed areas of dysplasia. COX-2 reactivity was detected infrequently and at a weak intensity in the urothelium (Figures 1–3). In the dysplastic areas, staining was usually more intense in the basal layers of the epithelium. The COX-2 expression was significantly higher in schistosomal cystitis than in nonschistosomal cystitis (P < .05). The smooth muscles and endothelial cells within the sections with no tumor did not express COX-2 (Table 1).

Figure 1 Cyclooxygenase-2 expression and distribution of positive cases in different studied bladder lesions.

Figure 2 A control case, negative for cyclooxygenase (COX)-2 expression (immunostain for COX-2, diaminobenzidine, x40 micrometers [mcm]).

Figure 3 A case of chronic schistosomal cystitis, showing mild cytoplasmic expression of cyclooxygenase (COX)-2, involving the whole layers of the urothelium (immunostain for COX-2, diaminobenzidine, x40 micrometers [mcm]).

Table 1 COX-2 Expression and Distribution of Positive Cases in Different Studied Bladder Lesions

For group III, in all positive cases, COX-2 expression was more pronounced in the tumor cells than in the nonmalignant urothelium (P < .01). In tumor cells, the expression was mainly cytoplasmic and granular with occasional focal condensation around the nuclear envelope. Also, the expression of COX-2 positivity is markedly heterogeneous in most cases of TCC and SqCC, in which the percentage and intensity of positive staining markedly varied from one field of the tumor to another, whereas homogeneous staining was usually obtained in tumors of higher grade. COX-2 expression was significantly higher in TCC (22 of 34 positive cases for COX-2, 64.8%) than in SqCC (7 of 16 positive cases, 43.2%) (P < .01). Also, COX-2 expression was significantly higher in schistosomal TCC (13 of 15 positive cases, 85.2%) than in nonschistosomal TCC (9 of 19 positive cases, 47.3%) (P < .01). COX-2 expression was significantly higher in grade 3 bladder TCC than in grades 1 and 2 (P < .05 and P < .01, respectively) (Figures 4–7). Also, there was a significant difference in COX-2 expression level between the bladder TCCs at different clinical stages between the superficial and invasive tumors (P < .01) (Table 2). There was a significant and positive correlation between COX-2 expression and tumor grade (r_s = 0.38, P = .0052) and with tumor stage (pT) (r_s = 0.44, P = .0035).

Figure 4 A case of superficial papillary transitional cell carcinoma, grade 1, showing moderate cytoplasmic expression of cyclooxygenase (COX)-2 in about 50% of malignant urothelial cells (immunostain for COX-2, diaminobenzidine, x40 micrometers [mcm]).

Figure 5 A case of invasive transitional cell carcinoma, grade 2, showing marked cytoplasmic expression of cyclooxygenase (COX)-2 in the majority of malignant urothelial cells (immunostain for COX-2, diaminobenzidine, x40 micrometers [mcm]).

Figure 6 A case of schistosomal-associated moderately differentiated squamous cell carcinoma showing moderate cytoplasmic expression of cyclooxygenase (COX)-2 in about 70% of the malignant squamous cells (immunostain for COX-2, diaminobenzidine, x20 micrometers (more ...)

Figure 7 A case of well-differentiated squamous cell carcinoma showing focal cytoplasmic expression of cyclooxygenase (COX)-2 in the malignant squamous cells (immunostain for COX-2, diaminobenzidine, x20 micrometers [mcm]).

Table 2 COX-2 Expression and Distribution in Different Histopathologic Grades and Stages of TCC

Data on the role of COX enzymes in SABC are lacking. In our work, COX-2 was expressed in none of the 5 normal bladder specimens, in accordance with the results of Margulis and colleagues,[34] who showed no expression of COX-2 in all 9 control cases. We assessed the nonmalignant but chronically inflamed and metaplastic urothelium commonly associated with schistosomal infestation from patients with SABC. There was notable COX-2 positivity, and a significantly increased expression of COX-2 in schistosomal cystitis than in nonschistosomal cystitis, in contrast to previously published studies of nonmalignant nonschistosomal urothelium, in which completely negative staining has always been reported.[28,29] Upregulation of COX-2 expression may result from the severe and chronic inflammatory reaction commonly associated with chronic schistosomal infestation, and which has been described in active inflammation, eg, inflamed, nondysplastic colonic mucosa in both ulcerative colitis and Crohn's disease.[35]

In the present study, we found increased COX-2 positivity in group III in schistosomal-associated TCC and SqCC. The present results are consistent with those obtained by El-Sheikh and colleagues[36] in cases with SABC and those obtained by Kömhoff and coworkers,[26] Mohammed and coworkers,[27] Diamantopoulou and colleagues,[37] Wadhwa and coworkers,[38] Wild and coworkers,[39] and Gurocak and colleagues[40] in cases with non-SABC.

The overexpression of COX-2 in tumor cells can be attributed to transcriptional and posttranscriptional factors. The COX-2 gene promoter contains transcriptional regulatory elements linked to multiple signaling pathways downstream of growth factors and cytokines.[18] The proinflammatory cytokines interleukin-1, tumor necrosis factor (TNF)-alpha, and transforming growth factor (TGF)-beta, which are produced by activated macrophages at the sites of chronic inflammation, are inducers of COX-2 expression[41] and may contribute to mechanisms by which chronic inflammation initiates transformation of urothelial cells in SABC. However, the inflammatory reaction alone cannot account for upregulated COX-2 expression in SABC because the nonmalignant epithelium exposed to the same inflammatory stimuli had a significantly lower level of COX-2. Growth factors with tyrosine kinase second messengers, eg, epidermal growth factor (EGF) and TGF-alpha, are also known to be inducers of COX-2,[42] and both are involved in progression of bladder neoplasia, including SABC[43] and N-nitrosamine-induced bladder cancer.[44] In addition, activating H-ras mutations detected in cases of nonschistosomal and SABC[45] can upregulate the expression of COX-2 and increase the stability of COX-2 mRNA by 3.5-fold, providing an additional source of increased COX-2 activity by posttranscriptional regulation.[46]

N-Nitroso compounds, the chemical carcinogens present in higher levels in patients with urinary schistosomiasis,[47] are possible targets for this upgraded COX-2 activity. Moreover, increased COX-2 expression, partly brought about by the normal physiologic response to injury and inflammation, may accelerate genetic damage through the increased production of malondialdehyde, a mutagenic byproduct of COX-mediated PG synthesis and lipid peroxidation.[48]

In the current work, we found that there was a statistically significant positive correlation between COX-2 expression and tumor grade (P = .01). COX-2 expression was significantly higher in grade 3 bladder TCC than in grades 1 and 2 (P < .05, P < .01). Correlation between COX-2 expression and progression of bladder TCC was observed; there was a significant difference in COX-2 expression level between the superficial TCC and invasive TCC (P < .01). These results are in accordance with El-Sheikh and colleagues[36] who confirmed COX-2 overexpression in SABC and a significant correlation between COX-2 and tumor grade. There was no correlation between the stage of SABC and COX-2 reactivity in their study.i

A previous study of 75 non-SABC TCCs by Kömhoff and coworkers[26] showed that COX-2 is highly expressed in malignant bladder tumors but not in benign bladder tissues; also, they detected no correlation between COX-2 expression and the individual clinical stages pTa-pT4, although such a correlation was present when stages pT1-pT4 were combined and compared with noninvasive TCC (pTa). Schmitz and colleagues[49] demonstrated in their study of cases with intrahepatic cholangiocarcinoma high levels of COX-2 expression associated with reduced apoptosis and increased proliferation of tumor cells, and concluded that COX-2 expression is an independent prognostic factor in resected intrahepatic cholangiocarcinoma, offering a potential adjuvant to therapeutic approach with COX-2 inhibitors.

There is much evidence that the COX-2 gene is involved in features of tumor aggressiveness, such as invasiveness and metastasis.[41] For example, COX-2 increases adhesion to the extracellular matrix and reduces the level of the cell-adhesion molecule, E-cadherin, in rat intestinal epithelial cells.[41] Human colon cancer cells transfected with a COX-2 expression vector have increased activity of metalloproteinase-2, which is necessary for the degradation of extracellular matrix, resulting in increased tumor cell migration.[50] However, there is no evidence that COX-2 is associated with invasiveness and metastasis in human tumors.

In the present study, COX-2 was expressed in muscle-invasive bladder tumors that are at a more advanced stage than superficial-type tumors; this is in accordance with the results of Wadhwa and coworkers,[38] Gurocak and colleagues,[40] and Margulis and colleagues.[34] Thus, it would be interesting to determine in additional studies whether COX-2 is a prognostic factor in bladder tumors.

The mechanism of elevated COX-2 expression in tumor cells may depend on the activation of oncogenes. Activation of the K-ras oncogene is associated with an elevated expression of COX-2,[32,33,41] and the K-ras oncogene is frequently activated in bladder tumors.[35] This particular mechanism may help explain the level of COX-2 expression found in bladder tumors. COX activates many carcinogens, one of which binds directly to hot spots for mutation in the p53 gene in lung[27] and bladder[26] cancer. Thus, COX may be involved in tumorigenesis by inactivating tumor suppressor genes, such as p53.

PGs have also been found to induce glutathione S-transferase-pi, an enzyme linked to cisplatin chemoresistance in bladder cancer.[51] In cases of SABC, the expression and activity of this enzyme were higher in tumor tissues than in uninvolved tissues.[52] The chemopreventive potential reported for nonsteroidal anti-inflammatory drugs, including aspirin, ketoprofen, piroxicam, and the recently developed specific COX-2 inhibitors, are thought to be mediated by the induction of apoptosis and are effective against the development of superficial rat urinary bladder carcinomas even after initiation with N-nitrosamines.[53]

In conclusion, TCC tissue in the bladder, especially the muscle-invasive type, frequently expresses COX-2. Thus, there may be a link between COX-2 expression and the development and progression of TCC in the bladder. Additional investigations are needed to determine whether COX-2 expression has any prognostic value and whether COX-2 inhibitors are useful for chemoprevention and cancer treatment of bladder tumors. Additional work is required to elucidate the possible downstream molecular targets of COX-2 products, and to assess the potential role of selective COX-2 inhibitors in preventing and treating SABC. Meanwhile, consideration of low-cost COX-2 inhibition in areas of endemic schistosomal infestation may be a simple and affordable intervention to limit the development and possibly the progression of SABC, diminishing such a devastating disease.

Olfat Ali Hammam, Theodor Bilharz Research Institute, Giza, Egypt Author's email: drolfathammam/at/hotmail.com.

Ahmed A. Aziz, Theodor Bilharz Research Institute, Giza, Egypt.

Mamdouh S. Roshdy, Theodor Bilharz Research Institute, Giza, Egypt.

Ahmed M. Abdel Hadi, Theodor Bilharz Research Institute, Giza, Egypt.