Between January 1996 and October 2012 radical cystectomies were performed on 324 consecutive patients with bladder cancer at the Department of Urology, Hirosaki University Graduate School of Medicine (Hirosaki, Japan). Histopathological grading was performed according to the World Health Organization system (6). Bladder tumor specimens were fixed with 10% buffered formalin for 12 h. The paraffin-embedded samples (3 μm) were subjected to hematoxylin and eosin staining. Written consent was obtained from all patients and the study was approved by the institutional ethics committees of Hirosaki University. Patients were divided into non-muscle invasive or muscle invasive groups according to the pathological state of invasion. The overall survival was estimated by the Kaplan-Meier method and compared between the two groups using the log-rank test.

Bladder tumors were surgically removed from certain genetically independent patients at the Department of Urology, Hirosaki University Graduate School of Medicine. Tumors were incubated with RPMI-1640 medium (Sigma-Aldrich, St. Louis, MO, USA) containing 5% fetal bovine serum (FBS; PAA Laboratories, Morningside, QLD, Australia) and 0.1% collagenase at 37°C for 16 h to prepare single-cell suspensions. The primary culture cells were subjected to various assays and immunostaining.

A human invasive and high-grade bladder cancer cell line, YTS-1, was provided by Dr H Kakizaki (Yamagata University, Yamagata, Japan). The RT4 cell line was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). YTS-1 and RT4 cells were maintained in RPMI-1640 medium supplemented with 10% FBS with 5% CO₂ at 37°C. Primary human urothelial cells (HUCs) were purchased from ScienCell Research Laboratories (Carlsbad, CA, USA) and maintained in urothelial cell media-basal (UCM-b; ScienCell Research Laboratories). All of the biochemical reagents were purchased from Sigma-Aldrich, unless otherwise stated. Anti-cortactin monoclonal antibody (clone EP1922Y) and anti-actin polyclonal antibody were purchased from Epitomics Inc. (Burlingame, CA, USA) and Sigma-Aldrich, respectively.

YTS-1 cells with reduced cortactin expression were generated by shRNA technology as described previously (2). An shRNA expression plasmid was constructed using pBAsi-hU6Neo DNA (Takara Bio Inc., Shiga, Japan). The shRNA sequence for cortactin was as follows: GATCCGCACGAGTCACAGAGAGATCTGTGAAGCCACAGATGGGATCTCTCTGTGACTCGTGCTTTTTTA (the siRNA sequence for cortactin is underlined). A human non-targeting siRNA sequence (Accell siRNA control; Thermo Scientific, Rockford, IL, USA) was used to prepare the control cells expressing non-targeting shRNA. The shRNA expression plasmids, knockdown and control constructs, together with pTK-HyB (Takara Bio Inc.) were introduced to the YTS-1 cells in a 10:1 molar ratio using Lipofectamine™ 2000 (Invitrogen Life Technologies, Carlsbad, CA, USA). Drug-resistant colonies were selected in the presence of 200 μg/ml hygromycin B (Sigma-Aldrich). Two knockdown clones (designated cortKD-1 and -2) were selected based on their cortactin expression levels. cortKD-1, -2 and a control clone (designated YTS control) were used for the assays described in the present study.

Total cancer cell lysates were prepared by solubilization in 50 mM Tris-HCl buffer, pH 7.5, containing 1% IgepalCA-630, 150 mM NaCl and proteinase inhibitors (10 μg/ml). The lysates were resolved by SDS-PAGE on an 8–16% gradient gel (Invitrogen Life Technologies), and transferred to a polyvinylidene fluoride membrane. Western blot analysis was performed using the specific primary antibodies (anti-cortactin antibody and anti-actin antibody) and a horseradish peroxidase-conjugated secondary antibody (GE Healthcare, Little Chalfont, UK). Signals were visualized using the ECL Plus detection system (GE Healthcare, Little Chalfont, UK).

Cells seeded on coverslips were fixed in 4% paraformaldehyde and permeabilized with phosphate-buffered saline (PBS) containing 0.1% saponin and 1% bovine serum albumin. Cells were stained with Alexa Fluor^® 568-labeled phalloidin (Invitrogen Life Technologies) together with the monoclonal antibody to cortactin, and Alexa Fluor^® 488-labeled secondary antibody (Invitrogen Life Technologies) was used for antibody staining. Cell staining was examined under an Olympus IX-71 fluorescence microscope (Olympus Inc., Tokyo, Japan) and LSM 710 laser scanning confocal microscope (Carl Zeiss, Oberkochen, Germany).

Gelatin zymography was performed in 10% Novex^® Zymogram precast SDS-PAGE gel (Invitrogen Life Technologies) in the presence of 0.1% gelatin under non-reduced conditions. Cells (1×10⁶) in 2 ml RPMI-1640 containing 10% FBS were placed in a single well (6-well dish) and grown to 80% confluence. Cells were washed with PBS and subsequently incubated with 2 ml Opti-MEM^® (Invitrogen Life Technologies) for 24 h. Conditioned media were collected and subjected to SDS-PAGE. Gels were washed in 2.5% Triton X-100 for 30 min at room temperature in order to remove SDS, incubated at 37°C overnight in a substrate buffer containing 50 mM Tris-HCl, pH 8.0, 5 mM CaCl₂ and subsequently stained with 0.5% Coomassie Brilliant Blue R-250 in 50% methanol and 10% acetic acid for 1 h.

The Matrigel matrix gel invasion assay was performed using a Transwell system (BD Biosciences, San Jose, CA, USA). The bottom surface of the filter (8-μm pore size) of the upper chamber was coated with 100 μg/ml fibronectin and the top surface was covered with 1 mg/ml Matrigel matrix. The lower chamber was filled with serum-free RPMI-1640 medium. Bladder cancer cells (5×10⁴) were labeled using a Vybrant^® carboxyfluorescein diacetate succinimidyl ester (CFDA SE) Cell Tracer kit (Invitrogen Life Technologies) and subsequently placed into the upper chamber. Following incubation at 37°C for 24 h, non-migrated cells that remained on the top surface of the filter were carefully removed with cotton swabs. Migrated cells on the bottom surface were fixed with 4% paraformaldehyde and counted under a Olympus IX-71 fluorescence microscope.

The transurothelial invasion assay was performed using the Transwell system. The upper surface of the filter (8-μm pore size) of the upper chamber was coated with 100 μg/ml collagen I. HUCs (1×10⁵) were placed onto the upper chamber and cultured for 2 days in order to form a monolayer. CFDA SE-labeled bladder cancer cells (5×10⁴) were re-suspended in UCM-b medium and placed onto the HUC monolayer. Following incubation at 37°C for 24 h, non-migrated cells were removed. Transinvaded cells were fixed and counted.

The statistical program SPSS 12.0 (SPSS, Chicago, IL, USA) was used to perform the statistical analyses. Statistically significant differences were determined using Student’s t-test and P<0.05 was considered to indicate a statistically significant difference.

Patients receiving radical cystectomies (n=324) were divided into two groups (non-muscle invasion, n=104 and muscle invasion, n=204) according to the pathological state of invasion. Fig. 1A and B show the representative images of non-muscle invasive bladder cancer and muscle invasive bladder cancer, respectively. Non-muscle invasive bladder cancer exhibited papillary growth (Fig. 1A), whereas numerous cancer cells were observed in the muscle layer in muscle invasive bladder cancer as indicated by the black arrows in Fig. 1B. Of the 324 patients, cancer recurrence was exhibited in 84 patients. The recurrence rate was higher in the muscle invasion group (87%) compared with the non-muscle invasion group (13%; Fig. 1C). Furthermore, the Kaplan-Meier curve demonstrated that overall survival was significantly shorter in the muscle invasion group compared with the non-muscle invasion group (P<0.0001; Fig. 1D). These results indicate that the muscle invasiveness of bladder cancer correlates with a higher recurrence rate and a worse prognosis.

To further characterize muscle invasive and non-muscle invasive bladder cancers, primary cultures of bladder cancers were established for various assays. Primary culture cells were prepared from bladder tumor specimens of four patients. Patients 1 and 3 (P1 and P3) had muscle invasive cancer and P2 and P4 had non-muscle invasive cancer (Fig. 2A). The cells were initially tested for their Matrigel invasion capacity. For comparison, two established bladder cancer cell lines, RT4 and YTS-1, were adopted. RT4 was derived from a well-differentiated papillary tumor and YTS-1 was established from an undifferentiated high-grade urothelial carcinoma, which is muscle invasive. YTS-1 exhibited a high Matrigel invasion capacity, however, the invasion capacity of RT4 was particularly low (Fig. 2B). The primary culture cells from P1 and P3 with muscle invasive cancer exhibited high Matrigel invasion capacities, which are equivalent to that of YTS-1 (Fig. 2B). By contrast, the invasion capacities of the primary culture cells from P2 and P4 with non-muscle invasive cancer were significantly lower than those of P1 and P3 (Fig. 2B). These results indicate that muscle invasive bladder cancer cells exhibited a higher capacity for Matrigel invasion compared with non-invasive bladder cancer cells.

For bladder cancer cells to invade the muscle layer, they must invade through a urothelial cell monolayer. In order to evaluate the capacity of invasion through the urothelial cells, we established a transurothelial invasion system using HUCs (Fig. 3A). HUCs, the cells lining the surface of the bladder, are the first line of defense for the bladder against infection by various pathogens and cancer cell invasion. Bladder cancer cells were placed onto a monolayer of HUCs (Fig. 3B) and 24 h later the transurothelial invasion capacity of the bladder cancer cells was measured by counting the number of cells that had transinvaded through the HUC monolayer. YTS-1 exhibited a higher capacity for transurothelial invasion capacity and the invasion capacity of RT4 was particularly low (Fig. 3C). Primary culture cells from muscle invasive bladder cancer (P1 and P3) also exhibited a significantly higher capacity for transurothelial invasion compared with the primary culture cells from non-muscle invasive bladder cancer (P2 and P4) (Fig. 3C). These results indicate that muscle invasive bladder cancer cells possess a significantly higher capacity for transurothelial invasion compared with non-muscle invasive cells.

Invasive cancer cells form invadopodia to invade the surrounding tissues; these filamentous actin (F-actin)-based membrane protrusions are required for the degradation of the extracellular matrix (ECM) and migration through the tissues (7–9). The primary culture cells were analyzed for the formation of invadopodia. Cells were double-stained with phalloidin and a cortactin monoclonal antibody (clone EP1922Y), which acts as an invadopodium marker (10–12). Invadopodia are visualized as yellow puncta, since the cores of invadopodia are composed of F-actin and cortactin co-localizes with F-actin. The typical invadopodia are denoted by white arrows in Fig. 4. Few invadopodia were observed in RT4 cells (Fig. 4A–C), however, numerous invadopodia were formed in the YTS-1 cells (Fig. 4D–F). Furthermore, primary culture cells from muscle invasive bladder cancer (P1 and P3) formed numerous invadopodia (Fig. 4G–I and M–O). By contrast, primary culture cells from non-muscle invasive bladder cancer (P2 and P4) formed very few invadopodia (Fig. 4J–L and P–R). These results indicate that the muscle invasiveness of bladder cancer cells correlates with the ability to form invadopodia.

In order to degrade surrounding tissues and the ECM for invasion, invadopodia secrete MMPs, which is a primary function of invadopodia. It has previously been demonstrated that the ability of bladder cancer cells to form invadopodia correlated with the quantity of MMP-2 secretion (13). The secretion of MMP-2 by bladder cancer cells was subsequently examined by gelatin zymography. A large quantity of MMP-2 was secreted by YTS-1 and the muscle invasive bladder cancer cells from P1 and P3 (Fig. 5, lanes 3, 4 and 6). By contrast, the non-muscle invasive cancer cells, such as the RT4 cells and those from P2 and P4 only secrete a marginal quantity of MMP-2, if any (Fig. 5, lanes 2, 5 and 7).

It was demonstrated that YTS-1 and muscle invasive bladder cancer cells from P1 and P3 exhibited a greater ability to invade through the urothelial cell monolayer (Fig. 3) and that these cells formed functional invadopodia (Figs. 4 and 5), indicating that invadopodia are important in bladder cancer cell invasion through the urothelial cell monolayer. In order to determine the importance of invadopodia in transurothelial invasion, the formation of invadopodia in YTS-1 cells was reduced and the effect of this reduction on transurothelial invasion was examined. To lower the formation of invadopodia, the expression of cortactin in YTS-1 cells was silenced, as cortactin is a key regulator for invadopodia formation (10–12).

Several stable cortactin knockdown cell lines were established using an invasive bladder cancer cell line, YTS-1. A YTS-1 cell line expressing non-targeting siRNA served as a control (designated YTS control). Western blotting revealed that cortactin expression was lower in a knockdown cell line (designated cortKD) compared with the YTS control (Fig. 6A). The results from the assays using cortKD are shown in the present study, since the results yielded from other cortactin knockdown cell lines were almost identical to those from cortKD in all of the assays. Invadopodia in the YTS control cells were visualized using Alexa Fluor^® 568-labeled phalloidin staining as F-actin-rich puncta (Fig. 6B, left panel). A section of cortactin staining also exhibited a punctate pattern (Fig. 6B, central panel) and co-localization of cortactin with F-actin puncta, which indicates that these puncta are invadopodia (Fig. 6B, right panel). Certain typical invadopodia are indicated by the white arrows (Fig. 6B, right panel). In cortKD cells, no clear invadopodia were observed (Fig. 6C). These results indicate that the formation of invadopodia was impaired in cortKD cells. The cells were subsequently assayed for the secretion of MMPs. According to gelatin zymography, a large quantity of MMP was secreted by the YTS control cells (Fig. 6D, lane 1), however, the secretion of MMP-2 by cortKD cells was significantly lower than that secreted by the YTS control cells (Fig. 6D, lane 2).

cortKD cells impaired the formation of invadopodia and the secretion of MMPs (Fig. 6A–D). To examine whether impairing invadopodia formation affects transurothelial invasion, the transurothelial invasion capacity of cortKD and YTS control cells was measured. The transurothelial invasion capacity of cortKD cells was markedly reduced compared with that of the YTS control cells (Fig. 6E). This indicates that impaired invadopodia formation and reduced MMP-2 secretion result in reduced transurothelial invasion, which indicates that invadopodia are essential in the invasion of bladder cancer cells through the HUC monolayer.