Prostate stem cell antigen predicts tumour recurrence in superficial transitional cell carcinoma of the urinary bladder


Tomoharu Fukumori, Department of Urology, The University of Tokushima Graduate School Institute of Health Biosciences, 3-18-15 Kuramoto, Tokushima 770–8503, Japan. e-mail:



To evaluate the relationship between prostate stem cell antigen (PSCA) expression level in transitional cell carcinoma (TCC) of the urinary bladder and various clinicopathological features, including stage and grade; and to determine whether PSCA mRNA expression predicts disease recurrence in superficial (not muscle-invasive) TCC of the bladder.


Real-time reverse transcriptase-polymerase chain reaction (RT-PCR) was performed on 97 TCC tissue samples and in 36 samples of normal bladder urothelium; the findings were analysed in relation to clinicopathological factors. Immunohistochemical expression was examined using light and confocal immunofluorescence microscopy to validate the RT-PCR data.


Twenty-seven patients developed disease recurrence, while the remaining 22 had no evidence of recurrence of superficial TCC of the bladder. There was significantly higher PSCA mRNA expression in TCC than in normal urothelium samples (P = 0.008). Superficial (TaT1) tumours had significantly higher PSCA expression than muscle-invasive (≥ pT2) tumours (P < 0.001). There was no significant difference between patients with G1–2 tumours and those with G3 tumours (P = 0.109). Immunohistochemical analysis showed markedly greater PSCA expression in superficial than invasive TCC. Notably, from a multivariate analysis, the expression level of PSCA was an independent predictor of disease recurrence in superficial TCC (P = 0.012).


These findings suggest that the PSCA expression level measured by real-time RT-PCR could be a valuable prognostic marker for tumour recurrence in superficial TCC of the bladder.


prostate stem cell antigen


glyceraldehyde-3-phosphate dehydrogenase


alkaline phosphatase


confocal laser scanning light microscopy.


TCC is the fifth most common solid tumour in man, with an overall estimated annual incidence of 63 210 new cases in the USA [1]; >70% of cases of TCC of the bladder are superficial at initial presentation, with <30% of patients being diagnosed with muscle-invasive disease. Despite treatment, about half of patients with muscle-invasive bladder cancer develop recurrence, leading to 12 100 deaths annually [2]. Superficial bladder cancer is defined as a tumour limited to the surface epithelium or invading the lamina propria [3]. Superficial bladder tumours that are removed by transurethral resection frequently recur [4]. Consequently, it is clinically relevant to predict the recurrence of superficial cancer before increased biological potential and invasion. The clinical and histopathological features and biomarkers that predict tumour recurrence and progression of bladder cancer were reported previously [5,6]. The identification of a diagnostic marker to predict tumour recurrence will facilitate more effective management of this cancer.

Prostate stem cell antigen (PSCA) is a member of the LY-6/Thy-1 family of glycosylphosphatidylinositol-anchored cell-surface proteins, and is a human cancer marker closely related to stem cell antigen-2 [7,8]. It was reported that PSCA is expressed by normal prostate and overexpression occurs in prostate cancer. Furthermore, it was previously reported to correlate with the progression of prostatic carcinoma [9,10]. In addition to the prostate, PSCA is expressed by bladder, oesophagus, stomach and pancreatic tumours [11–13]. Recently, low levels of PSCA expression were reported in the transitional epithelium of normal bladder, and there was greater expression in most superficial and muscle-invasive specimens of TCC [14].

The real-time RT-PCR method is sensitive enough to detect low-level gene expression and accurate enough to quantify the full range of expression [15]. The objective of the present study was to evaluate PSCA mRNA expression in TCC specimens of the urinary bladder, and to assess the prognostic value of PSCA expression, focusing on its correlation with tumour recurrence in superficial TCC.


The study included 97 specimens of TCC of the bladder, 49 with superficial (not muscle-invasive) TaT1 tumours, 48 with muscle-invasive (≥pT2) tumours, and 36 of normal urothelium. TCC tumours were obtained from patients undergoing transurethral resection, partial cystectomy or radical cystectomy for bladder cancer at our institutions. Informed consent for this study was obtained from all patients, and the Research Ethics Committee of the Tokushima University, School of Medicine approved the study. Primary TCC tissue samples were obtained from tissue that grossly and clearly appeared to be a tumour of the urinary bladder. Tissue from the same areas of tumour specimens and normal mucosa samples were fixed in 10% formalin and confirmed by histopathological examination to assess the relative proportion of tumour cells, benign epithelium, stroma and lymphocytes. Tumour samples with marked inflammation or necroses were excluded from further analysis. None of the patients had received previous chemotherapy, intravesical instillation therapy, or radiation. We took a random cold-cup biopsy at the same time as transurethral resection to exclude cases of carcinoma in situ. A recurrence was defined as any evidence of tumour in a retained bladder at least 3 months after surgery. Tumours were staged according to International Union against Cancer [16] and graded histologically according to the criteria of the WHO classification system [17]. Patient and tumour characteristics, and the clinical outcome, are summarized in Table 1.

Table 1.  Patient characteristics and clinical outcome
VariablesN (%) of patients or mean/median (range)
Total number97
Mean (range) age at surgery, years69.2 (46–93)
 Male77 (79.4)
 Female20 (20.6)
Tumour stage
 pTa21 (21.6)
 pT128 (28.9)
 ≥pT248 (49.5)
Histological grade
 G1–254 (55.7)
 G343 (44.3)
Surgical intervention
 Transurethral resection53 (54.7)
 Partial cystectomy 4 (4.1)
 Radical cystectomy40 (41.2)
Median (range) follow-up, months37.6 (2.8–136.0)

Tissue samples for RNA extraction were immediately submerged in RNALaterTM (Takara Bio Inc., Japan), then incubated at 4 °C overnight before storage at − 80 °C until use for RNA extraction. Total RNA was extracted using the RNeasy kit (Qiagen Inc., Valencia, CA, USA) following the manufacturer’s instructions. To avoid contamination by DNA during RNA preparation, RNA was digested with RNase-free DNase set protocol (Qiagen), as recommended by the supplier. Total RNA was quantified using an ultraviolet spectrophotometer, and the quality and integrity were assessed on a 1.5% agarose gel. Total RNA (1 µg) was reverse transcribed using a First-Standard cDNA Synthesis Kit (Amersham Pharmacia Biotech Inc., Germany) according to the protocol of the manufacturer. To examine the expression levels of the target gene (PSCA) and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH), real-time quantitative RT-PCR was performed on the LightCycler® System (Roche, Diagnostic GmbH, Mannheim, Germany) using SYBR Green I dye. Selected sequences of sense and antisense primers used for amplification of PSCA were: PSCA sense 5′-CAGGTGAGCAACGAGGAC-3′; PSCA antisense 5′-GTTCTTCTTGCCCACGTAGT-3′. For GAPDH, the PCR reaction mixture was carried out according to the LightCycler-Primer Set, ready-to-use amplification primer mix for RT-PCR. Each PSCA reaction (20 µL) contained 5 µL of cDNA, primers at 0.5 µm, and MgCl2 at 4 mm. The PCR thermal cycling conditions were an initial denaturation step at 95 °C for 10 min, 40 cycles consisting of denaturation at 95 °C for 10 s, annealing at 56 °C for 10 s and extension at 72 °C for 12 s. PSCA and GAPDH sequences were amplified in duplicate from the tissue samples. To ensure that the correct product was amplified in the reaction, all samples were separated by 2% agarose gel electrophoresis. For each sample, the amounts of the target (PSCA) and of an endogenous control (GAPDH) were determined using a calibration curve. The amount of the target is calculated as a ratio of target gene copies to housekeeping gene copies to obtain a normalized value. We set several thresholds arbitrarily to select the best value preliminarily with respect to postoperative recurrence. A separate standard curve for PSCA was constructed using serial dilutions of a positive control. PCR products starting at a concentration of 105 copies down to 102 were used as standards (Fig. 1A). The positive control and target samples were simultaneously amplified using the same reaction mixture. To verify the melting curve results (Fig. 1B), representative PCR products were sequenced. LightCycler data were analysed quantitatively using LightCycler software (LightCycler Operator’s Manual, version 3.5, Roche Diagnostics).

Figure 1.

Real-time PCR analysis of PSCA mRNA levels. A. Standard curve generated by the LightCycler software. Solid lines indicate PSCA diluted from 105 to 102 copy numbers, assayed in duplicate. B. Melting curves for all PCR products after PSCA amplification. There was only one sharp peak for each sample, indicating that only the specific PCR product was amplified.

Tissue obtained from patients with TCC of the bladder was frozen in OCT compound (Tissue-Tek, Sakura Fine Chemical, Tokyo) and stored at − 80 °C. Serial sections of 10 µm were cut using a cryotome and mounted on glass slides. To compare different tumour types under the same conditions, sections from the various specimens were mounted together and immunostained at the same time. After this, the sections were dried and fixed with 4% paraformaldehyde in 0.1 m PBS for 10 min at 20 °C. The sections were then incubated with 1% BSA in PBS for 1 h at 20 °C, and incubated with rabbit polyclonal antibody against PSCA (anti-PSCA; 1 : 500; Zymed Laboratory Inc., South San Francisco, CA, USA) for 2–3 days at 20 °C. After being rinsed in PBS, the sections were further incubated with alkaline phosphatase (ALP)-conjugated goat anti-rabbit IgG (1 : 200, Vector Laboratories, Burlingame, CA, USA) in 1% BSA in PBS for 2 h at 20 °C. The sections were then rinsed and reacted with an ALP-substrate (Vector Red, Vector), according to the manufacture’s protocol, after which they were counter-stained with haematoxylin, and flat-mounted with Vectashield (Vector). After this, the PSCA-immunostained tissue could be examined with Nomarski optics. Some sections were also fluorostained; after incubation with anti-PSCA antibody, the sections were incubated with fluorescein isothiocyanate-labelled donkey anti-mouse IgG (1 : 200; Jackson) in 1% BSA in PBS for 2 h at 20 °C. In addition, propidium iodide (2 mg/mL; Sigma, St Louis, MO, USA) was applied for 10 min. After this, the sections were mounted with Vectashield and examined with a confocal laser scanning light microscope (CLSM; Radiance 2000, Bio-Rad, Hercules, CA, USA). As a control, some sections were incubated without anti-PSCA antibody, and these showed no staining after ALP- or fluorescent-labelling. The present immunohistochemistry results are consistent with those in a previous study [18].

The difference between PSCA expression and several clinicopathological variables, such as stage and grade, were assessed using Student’s t-test. The prognostic significance of some factors (PSCA expression level, tumour stage, tumour grade, age, and gender) was assessed according to the Cox proportional hazards regression model. Values are presented as the mean (sem), with statistical significance indicated at P < 0.05.


PSCA mRNA values in bladder cancer and normal urothelium tissues were normalized to GAPDH mRNA. The quantification of PSCA gene expression by real-time PCR showed a significantly greater mean ratio of PSCA to GAPDH expression in TCC specimens than in normal urothelium, at 1274.9 (275.4) vs 37.3 (8.7) (P = 0.008). There was a greater mean ratio of PSCA to GAPDH expression in superficial (TaT1) than in invasive (≥pT2) tumours of the bladder, at 2357.2 (497.0) vs 148.4 (39.4), respectively (P < 0.001). As shown in Fig. 2A, patients with ≥ pT2 tumours had significantly lower PSCA expression levels than patients with Ta or T1 tumours, at 148.4 (39.4) vs 2142.5 (750.2) or 2518.2 (673.4), respectively (both P < 0.001). The difference between Ta and T1 tumours was not significant (P = 0.713). There was no significant difference in PSCA expression level in patients with G1–2 and G3 tumours (Fig. 2B; P = 0.109).

Figure 2.

PSCA expression level in TCC specimens of different stages (A) and grades (B). A. Patients with ≥ pT2 tumours had significantly lower PSCA expression level than patients with Ta or T1 tumours (both P < 0.001). The difference between Ta and T1 tumours was not significant (P = 0.713). B. There was no significant difference in patients with G1–2 tumours or G3 tumours (P = 0.109).

To validate the real-time PCR data, we stained representative samples of both superficial and invasive TCC with a specific antibody to PSCA using fluorescent immunohistochemistry for CLSM. As seen in Fig. 3b superficial TCC samples had strong homogenous staining for PSCA, that was mainly expressed in the cytoplasm of the superficial TCC cells. By contrast, there was less pronounced immunostaining for PSCA in invasive TCC tissue (Fig. 3d). The fluorescent staining results were confirmed by ALP and haematoxylin staining (Fig. 3e–h), which also showed strong staining for PSCA in superficial, but not in invasive TCC.

Figure 3.

Immunostaining of PSCA in superficial and invasive bladder cancer specimens. Representative images of immunofluorescent (a–d) and ALP (e–h) staining of bladder cancer epithelium with polyclonal anti-PSCA antibody are shown. a: Superficial bladder cancer tissue (negative control) with no primary antibody showed no staining. b: There was overexpression of PSCA in the same sample (green). The nuclei were then counterstained with propidium iodide (red). c: Invasive bladder cancer tissue (negative control) showed no staining. d: Minimal green staining of the same sample indicated reduced PSCA expression. e: Superficial bladder cancer tissue (negative control) with no primary antibody showed no staining. f: There was overexpression of PSCA in the same sample (pink within the cytoplasm). g: Invasive bladder cancer tissue (negative control) showed no staining. h: There was limited PSCA expression in the same sample. Specimens were then counterstained with haematoxylin (e–h). Bars = 100 µm.

We evaluated PSCA expression in superficial TCC with or with no tumour recurrence; there was tumour recurrence in superficial TaT1 tumours in 27 of 49 patients (55%) after a median follow-up of 37.6 months. We divided the 49 cases of superficial TCC into subgroups of high and low expression using a threshold value of 1576. Those with no recurrence had greater PSCA expression than those with recurrence, at 3509.2 (951.8) vs 1418.5 (394.7), respectively (P = 0.035) (Fig. 4). To evaluate the predictive value of each variable for disease recurrence in superficial TCC we used a multivariate analysis (Cox proportional hazards model) considering age, gender, tumour stage, tumour grade and PSCA mRNA expression level. Interestingly, the multivariate and univariate analysis showed that PSCA expression alone was an independent predictive factor of tumour recurrence in superficial TaT1 tumours (Table 2). There was no significant relationship between recurrence in superficial TCC and other clinicopathological variables such as tumour stage (P = 0.258), tumour grade (P = 0.149), age at surgery (P = 0.113) or gender (P = 0.228).

Figure 4.

PSCA expression in patients with and without recurrent superficial TCC. There was significantly greater PSCA expression in patients without than in patients with recurrent TCC (P = 0.035).

Table 2.  Univariate and multivariate analysis of prognostic factors of recurrence in patients with superficial TCC of the urinary bladder
VariablesHazard ratio (95% CI)
  • *

    P < 0.05.

Age, years; <65 vs ≥ 650.824 (0.361–1.884)0.6470.462 (0.177–1.200)0.113
Gender, male vs female0.974 (0.368–2.579)0.9582.000 (0.649–6.166) 0.228
Tumour stage, Ta vs T11.564 (0.715–3.422)0.2621.684 (0.683–4.151)0.258
Tumour grade, G1 + 2 vs G30.606 (0.243–1.508)0.2810.450 (0.152–1.332)0.149
PSCA expression, ≥ 1576 vs <15760.399 (0.160–0.994)0.048*0.293 (0.113–0.763)0.012*


Several studies reported the importance of PSCA as a cell-surface antigen that has potential use in the diagnosis and treatment of prostate, bladder, gastric and pancreatic cancers [9–14]. Bahrenberg et al.[11] reported that PSCA could serve as a potential marker for early carcinogenesis in urothelial tissues and that its expression is specific for epithelial cells. Furthermore, Amara et al.[14] reported that PSCA was overexpressed in most superficial tumours, and a significant percentage of invasive TCC. Recently, Cheng et al.[19] measured PSCA expression in urine samples, and showed its usefulness as a cytological marker of urothelial carcinoma via immunocytochemical analysis of urine. However, to date no studies have determined whether the expression level of PSCA could be used as a prognostic predictor in patients with recurrent superficial bladder tumours.

The main scope of the present study was to examine the prognostic relevance of PSCA as a predictive factor for recurrence in superficial TCC of the bladder. We evaluated the expression level of PSCA in TCC of the bladder and analysed the relationship between the expression of PSCA and several clinicopathological factors. There was a low level of PSCA expression in normal urothelium, and a high level of expression in urothelial carcinoma. Controversially, Bahrenberg et al.[11] reported that PSCA expression level was high in the normal urothelium. In addition, they found that PSCA expression was greater in confluent RT112 cells than in non-confluent cells. In agreement with this view, we found that the expression level of PSCA in confluent HT1376 TCC cells was three times higher than that of the scattered (non-confluent) HT1376 cells (data not shown). Furthermore, they reported that PSCA expression in RT112 cells depends on cell–cell contact and attachment to a surface, so that its function might be related to the adhesion of cells. In agreement with the findings of Amara et al.[14], the expression level of PSCA in the present samples was inversely correlated with tumour stage. The greatest mean level of PSCA expression was in cases of superficial TaT1 tumours, while there was lower expression in cases of muscle invasive (≥pT2) tumours. The results from CLSM supported the real time RT-PCR findings, confirming the greater expression of PSCA in superficial than in muscle-invasive TCC of the bladder.

Finally, we investigated the clinical significance of PSCA expression for its usefulness as a prognostic variable in TCC of the bladder, assessing, with other prognostic factors (tumour stage, tumour grade, age at surgery and gender) its effect on tumour recurrence, using multivariate analysis. Lower PSCA expression was associated with tumour recurrence and it became a strong prognostic determinant in cases of superficial TCC. Interestingly, there were no significant differences between Ta and T1 tumours (Fig. 2A), or between G1–2 and G3 tumours (Fig. 2B). These findings strongly suggest that PSCA expression level is an independent predictor of disease recurrence regardless of tumour stage or grade in patients with superficial TCC of the urinary bladder, as assessed by multivariate analysis. Furthermore, recurrent tumour cells, which expressed low levels of PSCA, have similar characteristics to those of scattered (non-confluent) cells with reduced cell–cell contacts. These tumour cells have a high tendency to recur, but the exact mechanism of PSCA down-regulation in recurrent superficial TCC of the urinary bladder has yet to be elucidated. Further studies are required to clarify the different factors regulating recurrence in superficial TCC of the urinary bladder.

In conclusion, we show for the first time that PSCA mRNA was expressed less in superficial TCC of the bladder in patients with disease recurrence than in those with no recurrence. Furthermore, our findings strongly suggest the role of PSCA as an independent predictor of tumour recurrence in cases of superficial TCC of the urinary bladder, and it could be used as a novel prognostic marker for superficial bladder cancer.


We thank Dr K. Ishimura, Ms. Toshiko suzuki- Yamamoto and Ms. E. Kiyokage (Department of Anatomy and Cell Biology) for their technical assistance to this study.


None declared.