Materials and methods
Cell lines and surgical specimens
Six transitional cell carcinoma cell lines (MGH-U1, MGH-U1R, MGH-U3, MGH-U4, TCC8704 and TSGH8301) were used. The cell lines were maintained in RPMI-1640 (Gibco) containing 10% heat-inactivated fetal bovine serum (FBS), 50 U/ml penicillin-G and 50 μg/ml streptomycin (Gibco), 2 mM L-glutamine and 1 mM sodium pyruvate (Gibco). The four human bladder cancer cell lines, MGH-U1, -U1R, -U3 and -U4, were generous gifts from Dr C.W. Lin (Massachusetts General Hospital, Boston, MA, USA). MGH-U1 and -U1R are sublines of T24, MGH-U3 was established from a grade 1, stage A bladder tumor, and MGH-U4 was established from a stage O bladder tumor with atypia. The remaining cell lines, TCC8704 and TSGH8301, were courtesy of Dr D.S. Yu (Department of Urology, Tri-Service General Hospital, Taipei, Taiwan, R.O.C.). Cells were incubated at 37°C in a CO2 incubator in humidified atmosphere containing 95% air and 5% CO2. Table I shows the grade and stage of tumors from which the cell lines were derived. Following Ethic approval and patients’ written consent, 30 primary bladder tumor specimens were resected at Chang Gung Memorial Hospital, Taiwan, and studied. Tumor staging was performed based on the 1997 tumor, nodes and metastasis (TNM) classification of bladder cancer, which was agreed upon by the American Joint Committee on Cancer (AJCC) and the Union Internationale Contra Cancer (UICC). Of the 30 resected specimens, 15 were superficial involvements (Ta-T1), 9 had muscle invasion (T2), 4 perivesical invasion (T3) and 2 distant metastases (M1). The histology was 5 grade 1, 14 grade 2 and 11 grade 3, respectively. Surgical specimens were immersed in plastic containers with optimum cutting temperature compound and stored in a −70°C refrigerator until use.
Chamber slide cultures and immunocytochemistry
Cells (5×104) were grown on each well of Lab-Tek® chamber slides (Nunc, Naperville, IL, USA) until confluence was achieved. Immunostaining was carried out at room temperature using the avidin-biotin-peroxidase complex (ABC) method (Vectastain ABC kit, Vector Laboratories, Burlingame, CA, USA), according to the manufacturer’s instructions. Briefly, 100 μl monoclonal antibodies at a concentration of 5 μg/ml, or at the dilution recommended by the manufacturer were added to each well. The slides were incubated for 1 h following 1 h of normal horse serum blocking. After washing twice with PBS solution, secondary antibodies (anti-mouse IgG) were added and incubated for 1 h. The slides were washed twice with PBS and then ABC reagent was added and incubated for another 40 min. After washing, the chromogen, 3-amino-9-ethylcarbazol (Aldrich Chemical Co., Inc., Milwaukee, WI, USA), containing 0.02% hydrogen peroxide, was used to detect the peroxidase conjugation. Gill’s hematoxylin solution (Fisher Scientific, Norcross, CA, USA) was used for counterstaining. Monoclonal antibodies (mAbs) used in this study were those against MMP-1 (41-1ES), MMP-2 (42-5D11; Calbiochem, San Diego, CA, USA), MMP-3 (55-2A4), MMP-9, TIMP-1 and TIMP-2 (Oncogene, Cambridge, MA, USA). A negative control (PBS instead of primary antibody, isotype-matched irrelevant mMAbs and normal mouse IgG) and positive controls (anti-HLA-A-B, C, W6/32) were included in each test.
Western blot analysis
Cell line extracts were obtained as previously described (2). Blots were incubated overnight with mouse-anti-MMP-2 (Lab Vision, USA) (1 μg/ml in 3% BSA) followed by incubation with goat anti-mouse IgG horseradish peroxidase, developed using the Santa Cruz enhanced chemiluminescence detection system (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and exposed by the Chemi-Smart 3000 (Vilber Lourmat) image system.
Gelatin zymography
Each cell line was grown in culture medium and protein extract was obtained at days 1, 3 and 5 for zymographic analysis. Gelatinolytic zymography was performed as described by Brown et al (3). Briefly, 30 μg of protein extract was mixed with non-reducing electrophoresis buffer on a 10% polyacrylamide gel containing gelatin (1 mg/ ml). After electrophoresis, the gels were incubated in a buffer containing Triton X-100 2.5%, 0.15 M NaCl, 10 mM CaCl2, 50 mM Tris-HCl buffer (pH 7.5) and 0.05% Coomassie brilliant blue and destained in 7% acetic acid and 5% methanol overnight with gentle rocking.
Immunohistochemical staining of bladder tumor tissues
Frozen tissue was cut (5 μm) and placed on gelatin-coated slides. The tissue was air-dried and fixed in chilled (4–5°C) acetone for 10 min. Immunostaining was carried out at room temperature using the ABC method, similar to that described above.
Statistical analysis
The significance of differences was calculated using Chi-square and Fisher’s exact probability tests. P<0.05 was considered to be statistically significant.
Discussion
Bladder cancer is the fourth most common malignant neoplasm in men and the eighth most common in women among Americans. Bladder cancer can be classified as superficial and invasive. Additionally, the majority of the bladder tumors are primarily superficial, but 70% of them will recur and 20–30% result in progression and metastasis (4). The aims of management of bladder cancer are two-fold: i) to detect relapse of the disease prior to the development of overt symptoms, such as gross hematuria or pain, and ii) to identify tumors that potentially indicate early recurrence, invasion and dissemination. Several studies have attempted to define the most predictable markers for recurrence and metastasis (5). The most common prognostic factors are staging and grading according to pathological characteristics. However, 36% of patients with urothelial cancer lack these characteristics. Previously, cytological analysis was used for transitional cell carcinomas. However, new tumor markers and molecules are currently under investigation (5). Notably, when considering treatment modalities for patients with transitional cell carcinoma of the bladder it is crucial to identify tumors that are likely to progress to invasive disease. Metastasis begins with the growth of tumor cells and invasion into the stroma surrounding the primary neoplasm. Degradation of the extracellular matrix and basement membrane are believed to be associated with tumor invasion and metastasis (1,6,7). The principal intrinsic components of the basement membrane are laminine and type IV collagen. Type IV collagen differs from the interstitial collagens (type I, II and III) in that it is localized exclusively in the basement membrane. Various types of matrix degradative enzymes or proteases are expressed and/or secreted by tumor cells. These include the metalloproteinases (8), serine proteinases (9) and lysosomal proteases (10,11). MMPs are a family of peptidae enzymes involved in remodeling extracellular components (12), including collagen, gelatin, fibronectin, laminin and proteoglycan. MMPs are complex regulators of multiple cell functions. Different cell types express various MMPs in cancer progression and metastasis (13). Kitagawa et al (14) noted that MMP-2 or gelatinase A is able to degrade type IV collagen and plays a role in tumor angiogenesis and metastasis. Numerous studies focused on the roles of MMPs in tumor invasion and metastasis (15,16). The roles of MMPs in angiogenesis are dual and complex and the increased expression of MMPs has been reported in various human malignant tumors (17–20). MMP-2 and -9 are critical for the angiogenic switch when the tumors become vascularized (21). MMPs were secreted in urine or serum by the tumors, and a zymographic analysis revealed that the urinary levels of MMP-9 and activated MMP-2 were higher in invasive bladder cancers than in superficial ones (22,23). However, in this study, the zymographic analysis of the cell lines did not yield similar results.
TIMP was originally identified as a mammalian collagenase inhibitor and is a glycoprotein with a molecular mass of approximately 28 kDa (24,25). TIMP-2 is a non-glycosylated protein with a molecular mass of approximately 21 kDa and a structure homologous to TIMP. TIMPs are secreted by many types of cells, including tumor cells, and inhibit the activities of various MMPs. In addition, these inhibitors suppress tumor cell invasion in vitro (26,27) and experimental metastasis in vivo (3,27,28). The hypothesis that TIMPs act as tumor suppressor genes due to their anti-metalloproteinase activity and their protective role on the extracellular matrix has been noted (29,30). On the other hand, it was proposed that the simultaneous cellular expression of MMPs and TIMPs in patients with breast cancer be determined as a predictor of clinical outcome (31). No significance of the MMP/TIMP ratio in predicting invasiveness was found in this study.
Our immunohistochemical results showed that MMP-9 was associated with the invasiveness of TCC and MMP-2 was associated with the grade of TCC. A discrepancy was found between the results of frozen sections and chamber slide cell stain, in which tumors of higher grade were stained negatively with MMP-2. Possible causes for the discrepancy in the results may be that the number of cell lines used were insufficent to reflect patient tumors. Additionally, cell lines able to grow in vitro may be highly selective subpopulations of the original tumors. The three-dimensional structure noted in in vivo tumors, as well as the interaction between tumors and stromal cells may be required for MMP-2 expression. Another explanation is that these cell lines were maintained in vitro for more than 2 years and loss of MMP-2 expression occurred following such a long-term in vitro culture. Early stage immunohistochemistry is therefore required to verify this assumption. MMP-9 has been reported to be overexpressed in invasive bladder cancers (23). However, MGH-U1 and -U1R cells were stained negatively with MMP-9. The explanation in MMP-2 may be applied to that of MMP-9. In a given tumor microenvironment, the interaction among tumor cells in situ and tumor-associated cells, such as neutrophils, macrophages, lymphocytes and endothelial cells, as well as environmental factors (hypoxia and pH), cytokines and growth factors released by these cells may be required for TCC expression of selective MMPs and TIMPs. The selective expression of these molecules then regulates tumor progression and angiogenesis. Therefore, the immunohistochemical results of the expression of MMPs and TIMPs in TCC tumor specimens may have greater clinical relevance than those obtained with the limited number of TCC cell lines in this study.
CD44 is a cell surface transmembrane glycoprotein that participates in cell motility and adherence of cells to extracellular matrix (ECM). CD44 also modulates the secretion and activation of MMP-2 (32). The collaboration between MMPs and CD44 at the cell surface may be essential in mediating degradation of the ECM to facilitate cell migration (33). In our previous study, a clear correlation of weak or negative staining of CD44v5 and surgical specimen tumor grades and stages was observed (2). The combined weight of the previous (2) and present results indicate that loss of CD44v5 expression may be induced by MMP-2 expressed by high-grade urothelial carcinomas, exhibiting a more invasive phenotype.
In conclusion, immunohistochemical testing of MMP and TIMP expression on 30 TCC surgical specimens revealed that the overexpression of MMP-2 was correlated with tumor grade, while that of MMP-9 was correlated with tumor stage. However, the expression of TIMP-1 or TIMP-2 did not significantly correlate with any of the disease states analyzed, although there was a tendency for TIMP-2 to correlate with tumor grade.