We extracted tissue samples from 155 NMIBC and 6 MIBC patients who had undergone TURBT between April 2004 and March 2013. Patients who had undergone early cystectomy were excluded from the study. The protocol for the research project was approved by the Institutional Review Board for Clinical Studies (Medical Ethics Committee ID: NMU-900), and informed consent was obtained from all the patients.

To examine the expression levels of KLα and KLβ, immunohistochemistry (IHC) was carried out. We used paraffin-embedded tissues obtained from all 161 patients in the study to examine the association between the KLα/KLβ expression levels and clinicopathological variables. The paraffin blocks were cut and placed on SuperFrost Plus Microscope Slides (Thermo Fisher Scientific, Yokohama, Japan). The sections were deparaffinized and antigen retrieval was carried out in citric acid buffer (pH 6.0) in an autoclave. IHC staining was performed with the Histofine ABC kit (Nichirei, Tokyo, Japan). Briefly, slides were treated with 1% hydrogen peroxide in methanol to block endogenous peroxidase activity. The slides were then incubated overnight at 4°C with anti-KLα antibody (sc-22220, rabbit polyclonal, dilution 1/500) and anti-KLβ antibody (sc-74343, rabbit polyclonal, dilution 1/200) (both from Santa Cruz Biotechnology, Santa Cruz, CA, USA), respectively. The slides were counterstained with hematoxylin, dehydrated, and mounted on a cover slide. We evaluated each slide using IHC scores (IHC score = intensity score + population score; intensity: none, 0; low, 1; intermediate, 2; and high, 3; population: none, 0; 0–25%, 1; 25–50%, 2; 50–75%, 3; and 75–100%, 4). The KLα/KLβ expression was categorized into low or high according to the IHC score as follows: low, IHC score ≤4; high, IHC score 5 or 6.

The human urothelial carcinoma cell lines MGH-U3, J82, and UM-UC-3 were used in this study. MGH-U3 was a gift from Dr H. LaRue (Laval University Cancer Research Centre, Quebec, Canada). J82 and UM-UC-3 cells were purchased from the American Type Culture Collection (ATCC; Manassas, VA, USA). The cell lines were maintained in RPMI-1640 medium (Nacalai Tesque, Kyoto, japan), supplemented with 10% fetal bovine serum (FBS; JRH, Tokyo, japan) and 1% penicillin/streptomycin (Thermo Fisher Scientific) in a standard humidified incubator at 37°C in an atmosphere of 5% CO₂.

Western blot analysis was performed using protein extracted from the cultured cells. All proteins were extracted using the RIPA lysis buffer (Sigma-Aldrich, St. Louis, MO, USA), which contained 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 2.5 mM sodium pyrophosphate, 1 mM Na₃VO₄, 1 mM phenylmethylsulfonyl fluoride, and 1 µg/ml leupeptin. Protein concentrations were quantified using Protein Quantification kit-Wide Range (Dojindo, Kumamoto, Japan). Total protein (20 µg) was diluted with sodium dodecyl sulfate (SDS) loading buffer containing 2.5% β-mercaptoethanol, boiled at 95°C for 5 min, and electrophoresed on 10% SDS-polyacrylamide gels using a Mini-PROTEAN Tetra electrophoresis cell at 200 V for 35 min. The gels were then transferred onto polyvinylidene difluoride membranes using a semi-dry transfer apparatus (both from Bio-Rad, Hercules, CA, USA) at 15 V for 45 min. After blocking overnight in Tris-buffered saline (pH 7.6) containing 5% skim milk and 0.1% Tween-20, the membranes were incubated for 1 h with the primary anti-KLα rabbit polyclonal antibody (dilution 1/200), anti-KLβ rabbit antibody (dilution 1/200), and anti-actin mouse monoclonal antibody (dilution 1/10,000) as an internal loading control, followed by 1 h with horseradish peroxidase-conjugated goat anti-rabbit IgG or anti-mouse IgG antibody. The bound secondary antibody was detected using a West Pico detection kit (Thermo Fisher Scientific). The secondary antibodies (Santa Cruz Biotechnology) were used at 1:5,000, 1:5,000, and 1:10,000 dilutions, respectively.

Recombinant human KLβ was purchased from R&D Systems (Minneapolis, MN, USA). The cell proliferation assay was used to examine the effects of the exogenous KLβ. Each of the cell lines was incubated with two different concentrations of KLβ (10 or 50 ng/ml) for 48 h in serum-free medium at 37°C in an atmosphere of 5% CO₂.

We evaluated whether the cell migration ability would increase with exogenous KLβ. At first, we performed the Matrigel invasion assay using BioCoat Matrigel Invasion Chambers (BD Biosciences, Piscataway, NJ, USA). Each cell line was incubated at 37°C in an atmosphere of 5% CO₂ with or without 50 ng/ml of KLβ. After 48 h of incubation, non-invading cells on the upper chambers were removed and the invading cells in the lower chambers were stained with calcein AM (PromoKine, Heidelberg, Germany) and the cells were immediately examined under a fluorescence microscope (Leica DMI 4000B).

We examined whether exogenous KLβ could increase the ability of the cancer cells to invade the endothelial cell layer, using human umbilical vascular endothelial cells (HUVECs; Lonza, Tokyo, Japan). Using a FluoroBlok insert system, the insert membrane chambers were coated with 30,000 HUVECs on fibronectin (Wako, Osaka, Japan) for HUVEC adhesion and incubated at 37°C in an atmosphere of 5% CO₂. After 8 h of incubation, low-concentrated colcemid (Nacalai Tesque) was added to the insert membrane to inhibit the migration of endothelial cells. Then, each cancer cell line was sprinkled onto the membrane. After a 24-h incubation period, non-invading cells in the upper chambers were removed and images of the invading cells on the membrane of the lower chambers were captured and visualized under a fluorescence microscope. The cells that attached to the bottom of the membrane were stained and examined as aforementioned.

To examine for anchorage-independent growth, we performed a soft agar colony formation assay. The base agar layer was prepared using 1.2% agar solution (Difco, Franklin Lakes, NJ, USA) mixed with an equal volume of 2X Dulbecco's modified Eagle's medium (DMEM; Sigma-Aldrich)/20% FBS in a 24-well culture plate. Cell suspensions (30,000 cells/ml) were prepared and mixed with both the 1.2% agar solution and 2X DMEM/20% FBS in the same manner as described above. Exogenous KLβ (50 ng/ml) or PBS as a control was added into each well and incubation was carried out at 37°C in an atmosphere of 5% CO₂. A week after seeding, the number of growing colonies was counted under a microscope.

Voided urine samples from 59 NMIBC patients and 10 MIBC patients were collected prior to surgery and stored at −80°C. Urine was also obtained from four healthy volunteers as assay controls. All 73 urine samples were thawed just before use and analyzed for KLβ concentration with an enzyme-linked immunosorbent assay (ELISA) kit (Cloud-Clone Corp., Houston, TX, USA), using a Tecan microplate reader (Tecan Systems, Inc., San Jose, CA, USA).

Statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software, Inc., San Diego, CA, USA). The figures were also constructed using GraphPad Prism 5.0. The comparison between high and low KLβ expression was calculated using the Mann-Whitney U test or the Student's t-test. The survival curve was obtained using the Kaplan-Meier method and compared by the log-rank test for each prognostic variable. A multivariate analysis was performed using the Statistical Package for the Social Sciences, version 19.0 (SPSS, Inc., Chicago, IL, USA). P<0.05 was considered statistically significant.

To investigate the association between KLα/KLβ and various clinicopathological variables, we performed IHC analysis of the KLα/KLβ expression levels. Fig. 1A shows representative images of weak, intermediate, and strong expression of KLα (left panels) and KLβ (right panels). KLα expression was higher in NMIBC than that in MIBC (P=0.0061; Fig. 1B, left), whereas the opposite was true for KLβ expression (P=0.0028; Fig. 1B, right). Table I lists the clinicopathological background of the cohorts of 155 NMIBC patients and comparisons of the variables with high and low KLα/KLβ expression. At initial TURBT, the median age for this cohort was 71 years [interquartile range (IQR) 34–94] and the median follow-up period was 53 months (IQR 1–126). The pathological data, such as the tumor category, tumor grade, and concomitant carcinoma in situ (CIS), were significantly different between patients with low and high KLα expression. In the case of KLβ, patients with high expression were significantly of high stage and high grade, and had a high rate of concomitant CIS and high presence of lymphovascular invasion (LVI) compared with patients with low KLβ expression. KLα and KLβ showed opposite trends for both invasiveness and tumor grade.

Of the 155 NMIBC patients, 55 (35.5%) had an intravesical recurrence at median 33.5 months after initial TURBT and 18 patients (11.6%) progressed to muscle invasive tumors at median 51 months after initial TURBT. With regard to the intravesical progression- and recurrence-free survival in patients with high or low KLα expression, there was no significant difference between the two groups [P=0.68 (Fig. 2A) and P=0.81 (Fig. 2B), respectively]. In contrast, the progression-free survival for the patients with high KLβ expression was significantly shorter than that for the patients with low expression (P<0.0001; Fig. 2C). The intravesical recurrence-free survival was not significantly different between the two groups (P=0.31; Fig. 2d).

Table II shows the univariate analysis of prognostic factors for both progression- and intravesical recurrence-free survival. The analysis revealed that high KLβ expression was a poor prognostic factor for progression-free survival [hazards ratio (HR) =13, 95% CI 4.2–38; P<0.0001] but not for intravesical recurrence-free survival (HR=1.4, 95% CI 0.73–2.6; P= 0.3). On the other hand, LVI was a predictive factor for progression-free survival (HR=11, 3.1–37; P=0.0002) but not for intravesical recurrence-free survival (HR=1.8, 95% CI 0.82–3.7; P=0.14). The other factors such as T category, tumor grade, and concomitant with CIS showed similar results. Table III shows the multivariate analysis of prognostic factors for progression-free survival. The Cox proportional hazard model analysis identified high KLβ expression (HR=6.9, 95% CI 2.6–18; P<0.0001) as an independent prognostic factor of progression to muscle invasive disease.

To verify the expression level of KLβ in the three urothelial cancer cell lines, western blot analysis was performed. There were variations in the endogenous KLβ expression levels (Fig. 3A), where UM-UC-3 expressed the lowest level among the three cell lines. Exogenous KLβ treatment at a concentration of 50 ng/ml promoted urothelial cancer cell proliferation by 120–140% in the MGH-U3 and UM-UC-3 cells, whereas no effect was observed in J82 cells (Fig. 3B). With regard to the Matrigel invasion assay, KLβ treatment enhanced the invasiveness of all three cell lines (Fig. 3C).

We assessed the association between KLβ expression and LVI status. Tumors with LVI had significantly higher KLβ expression than LVI-negative tumors in the IHC analysis (P<0.0001; Table I and Fig. 4A). To confirm this result with in vitro experiments, we used the transendothelial migration assay (Fig. 4B). Cells that broke through the layer of the endothelial cells and migrated to the bottom of the insert were quantified with a fluorescence-based spectrophotometer and microscopy (Fig. 4C). Exogenous KLβ treatment enhanced the transendothelial migration ability of all three cell lines. These findings suggest that KLβ is highly related to disease progression via enhanced tumor invasion through the vessel wall.

Cells were suspended in soft agar and incubated with or without KLβ. The number of colonies was counted 7 days after seeding. Evaluation on day 7 showed a notable increase in the colony formation ability of cells treated with KLβ in all three cell lines (Fig. 5). The results suggested that stimulation with KLβ enhanced the cell anchorage-independent growth capability in vitro.

The preoperative voided urine KLβ concentration in MIBC patients was significantly higher than that in NMIBC patients (Fig. 6). However, there was no significant difference between the healthy volunteers and the MIBC patients and NMIBC patients, respectively (Fig. 6). To ascertain the association between urine KLβ concentration and the result of urine cytology, we categorized the studied patients into three groups: negative, class I or II; suspicious, class III; and positive, class IV or V. There were no significant differences in urine KLβ concentration among the groups (data not shown).