Introduction

Transitional cell carcinomas (TCC) of the bladder define a group of histologically and genetically diverse cancers that account for approximately 4% of all adult malignancies with an annual incidence of approximately 53 200 cases in the United States (Cancer Facts and Figures, 2000). Approximately 75% of these cases are superficial (TIS, Ta, T1), 20% muscle infiltrating (T2–T4) and 5% metastatic at the time of diagnosis. TCC has been classified into two groups with distinct clinical behavior and different molecular profiles: low-grade tumors (always papillary and usually superficial), and high-grade tumors (either papillary or nonpapillary, and often invasive) (Reuter and Melamed, 1989). Clinically, patients diagnosed with localized TCC have a 5-year relative survival rate of 93%. However, patients presenting with regional and distant disease spread have 5-year relative survival rates of 49 and 6%, respectively (Cancer Facts and Figures, 2000). Biologically, the inactivation of the key regulatory RB1 and TP53 pathway has been shown to be necessary for the transformation and immortalization of bladder cancer, both by in vitro and in vivo studies (Dalbagni et al., 1993; Reznikoff et al., 1994; Cordon-Cardo et al., 1997; Markl and Jones, 1998; Lu et al., 2002). However, the molecular events involved in bladder tumorigenesis and tumor progression are not completely characterized. The chances of superficial tumor progression are increased with increased pathological stage and grade, tumor size, presence of concomitant carcinoma in situ and multifocality (Reuter and Melamed, 1989; Cancer Facts and Figures, 2000). In addition, numerous markers have been identified which correlate to some extent with tumor stage and possibly with prognosis. However, the power of many of these markers in predicting the clinical outcome of individual tumors is still limited and better markers are demanded for detection and predictive purposes.

Myopodin is the second member of the synaptopodin gene family, a novel class of actin-binding proteins (Mundel et al., 1997), with no significant homology to any known protein. It is strongly expressed in skeletal and smooth muscle tissues (Weins et al., 2001). Myopodin is a dual compartment protein that displays actin-bundling activity and shuttles between nucleus and cytoplasm in a differentiation-dependent and stress-induced fashion. Invasive prostate tumors display complete or partial deletions in the myopodin gene, which suggests a potential tumor suppressor role for myopodin in prostate cancer (Lin et al., 2001). This recent study led us to explore the role of myopodin in bladder cancer and evaluate whether alterations in myopodin expression might be present and provide clinical utility in the management of patients with bladder tumors.

Materials and methods

Cell culture

Nine bladder cancer cell lines (T24, J82, 5637, HT-1376, RT4, SCaBER, TCCSUP, UMUC-3, and HT1197) were obtained from ATCC (Rockville, MD, USA) and maintained the following standard procedures.

Clinical evaluation of myopodin expression

(1) Tissue samples in tissue microarrays

In all, 25 paired bladder tumors and their respective normal urothelium were first evaluated. Three different bladder cancer microarrays were constructed for this study using a precision instrument (Beecher Instruments, Silver Spring, MD, USA). From each specimen, triplicate cores with diameters of 0.6 mm were punched and arrayed on the recipient paraffin block (Hoos et al., 2001). These tissue microarrays included a total of 173 bladder primary TCC tumors obtained under the Institutional Review Board-approved protocols. Tumor stage and grade were defined according to consensualized criteria (Mostofi, 1973; American Joint Committee on Cancer, 1988). A total of 40 superficial and 64 invasive TCC tumors were analysed in two microarrays. These tumors corresponded to 14 grade 1, eight grade 2 and 82 grade 3 lesions. The third tissue microarray comprised a cohort of 69 bladder primary TCC cases with known p53 and pRB status, and consisted of two superficial and 67 invasive lesions.

(2) Immunohistochemistry

Protein expression patterns were assessed using both cytospin and tissue microarrays outlined above. Standard avidin–biotin immunoperoxidase procedures were used for immunohistochemistry. We used the following panel of antibodies: myopodin, SRIB2 at 1 : 20 (Weins et al., 2001); E-cadherin, mouse monoclonal clone 36 at 1 : 1000 (2.5 g/ml) (BD Transductions Labs, Lexington, KY, USA); zyxin, mouse monoclonal clone 21 at 1 : 25 (10 g/ml) with microwave pretreatment of the slides (Transduction Labs, Lexington, KY, USA); underphosphorylated RB, mouse monoclonal clone G99-549 at a final concentration of 10 g/ml (PharMingen, San Diego, CA, USA); total RB, mouse monoclonal clone 3C8 at a final concentration of 1.2 g/ml (QED Bioscience, San Diego, CA, USA) and mouse anti-human monoclonal antibodies to p53 (1 : 500, Ab-2, clone 1801; Calbiochem, Cambridge, MA, USA). P53 and p21 staining was evaluated taking the cutoffs of 20% (McShane et al., 2000) and 10% (Lu et al., 2002) respectively. Lacking a cutoff consensus for the other markers, they were analysed as continuous variables, or several cutoffs were taken when they were considered as categorical.

(3) Data analysis

All cTCC (n=173) were used for the analysis of association between myopodin and other markers. These cases were also utilized for evaluating marker expression versus histopathological stage and tumor grade, using the nonparametric Wilcoxon–Mann–Whitney and Kruskall–Wallis tests (Dawson-Saunders and Trapp, 1994). The consensus value of the three representative cores from each tumor sample arrayed was used for statistical analyses. Expression values are displayed as mean values accompanied of 95% confidence intervals and/or range.

The prognostic value of the clustering analysis was investigated. The relation of markers identified in DNA microarray analysis to the outcome was also evaluated using a subset of 69 cTCC cases for which follow-up was available. Overall survival time was defined as the months elapsed between transurethral resection (two superficial lesions) or cystectomy (rest of cases) and death from disease (or the last follow-up date). Patients who were alive at the last follow-up or lost to follow-up were censored. The association of marker expression levels with overall survival was analysed using the log-rank test. Survival curves were plotted using standard Kaplan–Meier methodology (Dawson-Saunders and Trapp, 1994). Associations between markers were analysed using Kendall's tau test and the SPSS statistical package (version 8.0).

Western blot analysis

Western blot analysis was performed as previously described (Weins et al., 2001). Protein extraction was carried out at 4°C homogenization buffer (20 mM Tris, 500 mM NaCl, pH 7.5) supplemented with 1% Triton X-100 (Sigma), 5 mM EDTA and protease inhibitors. Proteins were separated on 8% SDS–PAGE and transferred to Immobilon P membranes (Millipore). Primary antibodies against myopodin (SRIB1/2) were used at 1 : 250 and HRP-conjugated secondary antibodies (Promega) at 1 : 20 000. Blots probed for myopodin were reprobed for tubulin to evaluate equal loading. Blots were analysed by densitometry using Image Quant software, tubulin serving as an internal control to standardize all signals. Myopodin content of the cell lines were compared using T-test analysis.

Transient transfections

Myopodin-GFP (Weins et al., 2001) was transfected into 60–80% confluent bladder cancer cells using GenePORTER transfection reagent (Gene Therapy Systems) according to the manufacturer's instructions and analysed by direct fluorescence microscopy in living cells under a Nikon TE300 fluorescence microscopy with the appropriate filter. Pictures were captured with a Spot cooled CCD camera (Diagnostic Instruments) and processed with Adobe Photoshop 5.0 software.

Cell sorting

Cells were trypsinized at 80% confluency, washed with PBS and stained with Hoechst 33342 for 2 h at 37°C. Cell-cycle subpopulations G1, S and G2 M were sorted using a MoFlow sorter. Cytospins of each cell-cycle phase were prepared and fixed with methanol : acetone (50%) for 10 min at -20°C and kept at -80°C until used for immunolabelings.

Results

Experimental design

The present analysis was conducted under five major sets of experiments. First, the potential clinical significance of myopodin was evaluated using immunohistochemistry on clinical material. Initially, paired normal urothelium and bladder tumors were evaluated to assess the potential involvement of myopodin in progression of the disease. Second, a cohort of superficial and invasive bladder neoplasms contained in two tissue microarrays were used to evaluate the association of myopodin with histopathological stage and grade. A third tissue microarray, containing bladder tumors with characterized p53 and pRB alterations and annotated follow-up, was used to delineate associations between myopodin and these critical pathways in bladder cancer, as well as clinical outcome. Fourth, the expression pattern of myopodin was explored in vitro by transient transfections of bladder cancer cell lines. Finally, myopodin expression and sublocalization along cell cycle were analysed by immunocytochemistry of cytospins containing enriched cell-cycle populations of bladder cancer cell lines.

Myopodin is differentially expressed between bladder tumors and their respective normal urothelium

We evaluated the phenotype of myopodin in normal urothelium and tested whether any variation of its expression might be taking place in bladder tumors. In all, 25 paired normal urothelium and respective bladder tumors were analysed. Myopodin was expressed in both the nucleus and the cytoplasm of normal urothelium samples (Figure 1a). Cytoplasmic expression of myopodin was similar between bladder tumors and respective normal urothelium, whereas nuclear myopodin expression was lost depending on the stage of bladder tumors. Tumors with higher stage showed lower nuclear expression of myopodin. Superficial tumors displayed certain nuclear staining, while most of the invasive tumors showed a complete loss of the nuclear myopodin (Table 1).

Figure 1.

Representative expression patterns of myopodin in normal urothelium (a), transitional superficial bladder tumors (b) and invasive bladder neoplasms (c). Myopodin is expressed in both the cytoplasm and nucleus of normal urothelium and bladder tumors. However, its nuclear expression is significantly decreased in invasive bladder neoplasms (P<0.001). (original magnifications: (a)–(c), 400)

Full figure and legend (614K)

Table 1 - Immunohistochemical analysis of nuclear myopodin in paired normal urothelium (NU) and bladder tumors (TM).

Full table

Myopodin expression is associated with histopathological stage and tumor grade in primary bladder tumors

We studied the potential utility of myopodin expression to identify patients with superficial and invasive disease in a larger cohort of patients with bladder cancer. Myopodin expression was analysed by immunohistochemistry in superficial and invasive primary tumors spotted onto tissue microarrays. We observed cytoplasmic and nuclear localization of myopodin in superficial bladder tumors (Figure 1b). Nuclear expression of myopodin was significantly lower in invasive and high-grade tumors (Figure 1c) and segregated with superficial and invasive bladder lesions. Moreover, it reached a significant association with stage and grade in our series (Table 2).

Table 2 - Descriptive analysis of associations between nuclear myopodin expression levels (%) and stage and grade of 173 bladder tumors.

Full table

Expression of myopodin is associated with expression of p53 and RB related markers

We evaluated the association of myopodin expression with a selected group of genes previously reported to be altered during bladder cancer progression, including p53, total and underphosphorylated retinoblastoma (underPRB), cyclin E, p21, moesin, zyxin and E-cadherin. Immunohistochemical expression levels of these proteins were characterized in the bladder tumors contained in our tissue microarrays. Loss of myopodin expression was significantly associated with p21 (P<0.0005, n=146), underPRB (P<0.0005, n=149), E-cadherin (P<0.0005, n=149) and zyxin expression (P=0.023, n=145). Thus, expression of myopodin was associated with the p53 effector, the phosphorylation status of pRB and altered expression of adhesion molecules such as E-cadherin and zyxin. No significant association was found between nuclear expression patterns of myopodin and p53, cyclin D1 or cyclin E.

Expression of myopodin is associated with overall survival

The association of the nuclear myopodin expression with overall survival was evaluated in a tissue microarray with a subset of 69 bladder tumors whose follow-up was available. Patients displaying nuclear expression of myopodin showed a longer disease-specific survival than those who had lost this nuclear staining. The nuclear expression of myopodin provided good prognostic information and reached significant association with overall survival (Figure 2a). Based on this identified association of nuclear myopodin expression with underPRB, in the patients with bladder cancer contained in our tissue microarrays, we also evaluated the potential impact of a combined loss of retinoblastoma and myopodin expression on overall survival. We observed that nuclear myopodin expression remained a good prognostic biomarker for overall survival independent of the status of retinoblastoma expression (Figure 2b).

Figure 2.

(a) Kaplan–Meier curve of the survival analysis of patients with bladder tumors stratified by myopodin expression. Nuclear myopodin is significantly associated with overall survival in the subset of 69 bladder tumors (median follow-up time: 36 months) (P=0.04). (b) Kaplan–Meier curve of survival analysis of patients with bladder tumors stratified by a combination of underphosphorylated retinoblastoma (UnderPRB) and nuclear myopodin expression. The presence of nuclear myopodin remained as a good prognostic biomarker independent of UnderPRB

Full figure and legend (149K)

Myopodin expression in bladder cancer cell lines derived from superficial, invasive and metastatic TCC

In order to evaluate expression levels of myopodin in bladder cancer cell lines derived from tumors of different stage and grade, we performed immunocytochemistry, immunohistochemistry labelings and Western blot analysis of total and nuclear cell extracts (Figure 3a and b). Interestingly, we could not detect different expression levels of myopodin within these bladder cancer cell lines. Transient overexpression studies using a full-length GFP–myopodin fusion protein (Weins et al., 2001), showed not only preferential nuclear localization of GFP–myopodin, but also cytoplasmic distribution along the actin filaments (Figure 3b). These results confirmed the in vivo observed pattern in normal urothelium (Figure 1a). They suggest the relevance of the cellular myopodin distribution and point to a potential structural function of myopodin in the cytoplasm and a signaling-related role in the nucleus.

Figure 3.

(a) Densitometry analysis of relative myopodin expression in studied bladder cancer cell lines. Error bars indicate the standard error. There is no significant (P>0.1) difference in myopodin content among the cell lines. (b) Representative Western blot used for densitometry analysis. (c) Endogenous and myopodin–GFP expression in representative bladder cancer cell lines

Full figure and legend (374K)

Expression of myopodin during cell-cycle

Since nuclear myopodin expression redistributes between the nucleus and the cytoplasm in the muscle (Weins et al., 2001), and nuclear myopodin was associated with critical G1/S checkpoint regulators in our primary bladder tumors, we evaluated the potential association of myopodin expression with the cell cycle. We performed sorting analysis for cell-cycle phase of cell lines derived from bladder tumors of different stages and grades in order to obtain enriched populations of each cell-cycle phase (Figure 4a). Bladder cancer cell lines derived from superficial (RT4) and low-grade tumors (5637) showed nuclear localization of myopodin in G1 and S but not in G2/M (Figure 4b). However, cells derived from invasive tumors (T24) and from metastatic tumors (TCCSUP) showed loss of nuclear myopodin expression (Figure 4b) in G1 and partially in S with a shift of nuclear localization to G2/M. The same myopodin distribution was noted in TCCSUP and T24 cells and was also observed in the invasive transitional cell carcinoma lines HT1197, HT1376, T24 and J82 (data not shown).

Figure 4.

(a) Cell-cycle phase distribution of sorted bladder cancer cells. (b) Myopodin distribution along the cell-cycle on cytospins of bladder cancer cells sorted in (a). Nuclear localization was noted in G1/S for RT4 and 5637 (papillary and low grade), whereas in TCCSUP and T24 nuclear myopodin localization shifted to G2/M

Full figure and legend (347K)

Discussion

Data from this study provide new insights regarding critical issues about myopodin redistribution and its involvement in the progression of bladder cancer in a cell-cycle-dependent manner. Myopodin has been identified as a novel molecular target in bladder cancer with a potential tumor suppressor role. Loss of nuclear myopodin expression was related to bladder cancer histopathology both in vitro and in vivo in clinical material. More importantly, myopodin is a good prognosticator for disease-specific outcome. This is the first report where the loss of myopodin expression was shown to be of prognostic clinical relevance to establish the outcome of patients with bladder cancer.

A recent study has shown that deletions of the myopodin gene, which is located on chromosome 4q, may play a role in prostate cancer progression (Lin et al., 2001). This is of particular interest since comparative genomic hybridizations have shown that deletions in chromosome 4q may also occur in approximately 30% of invasive bladder tumors (Simon et al., 2000). Furthermore, loss of heterozygosity in 4q has been described in 52% of carcinoma in situ lesions, which display many of the alterations reported in invasive TCC (Rosin et al., 1995). These observations, together with other studies describing chromosomal alterations of this locus in bladder cancer (Simon et al., 1998; Koo et al., 1999; Shaw et al., 1999; El-Rifai et al., 2000), led us to evaluate the potential role of myopodin as a target of interest in bladder cancer. The availability of specific antibodies against myopodin allowed us to explore its clinical value in bladder cancer progression by characterizing the expression of myopodin in clinical material and bladder cancer cell lines. The presence of cytoplasmic staining in normal urothelium and neoplastic lesions suggests that complete deletions of this locus might not be the mechanism by which the expression of myopodin in the nuclei is lost along the progression of the disease.

Myopodin-reported tissue distribution included skeletal muscle, heart, colon, stomach, uterus and lung by Western blot, where its expression was restricted to muscle cell layers (Weins et al., 2001). In situ hybridization studies have recently reported the loss of myopodin in prostate cancer (Lin et al., 2001). Here we show that myopodin is also expressed in uroepithelial tumors. Myopodin is a dual compartment protein. It is strongly expressed in the nucleus of undifferentiated myoblast and redistributes during myoblast differentiation to the Z-disc (Weins et al., 2001). Myopodin is not found in the nucleus of mytotubes. However, it can re-enter the nucleus of these fully differentiated cells under conditions of cellular stress (Weins et al., 2001). These observations suggest that myopodin may also shuttle between the nucleus and the cytoplasm in other cellular types, for example, urothelial cells, and in other cellular stress conditions, for example, cancer.

Owing to the association that we found between myopodin and p21 and retinoblastoma, intrinsic proteins of the G1 checkpoint (together with p16/INK4a and cyclin D1) (Cordon-Cardo et al., 1997; Lu et al., 2002), we tested the hypothesis that nuclear myopodin localization is cell-cycle regulated. This observation prompted us to design experiments to identify potential implications of myopodin in the normal phenotype of the urothelium as well as in the progression of the disease. The immunostaining of cytospins obtained after cell-sorting analysis revealed a cell-cycle-dependent subcellular distribution of myopodin. Most remarkably, we observed that nuclei of low-stage and low-grade bladder cancer cells were positive for myopodin in G1 and S, whereas in invasive cancer cells this nuclear sublocalization was lost and shifted towards G2 and M. Hence, the cell-cycle-dependent subcellular myopodin localization in G1/S could classify bladder cancer cells based on the stage and grade of the tumors they were derived from. In addition, this observation in vitro suggests that the cell-cycle-dependent subcellular myopodin distribution might be critical for normal urothelial function and that alterations in this distribution may be involved in pathways triggering bladder cells to uncontrolled growth.

This expression pattern was also observed in vivo. In normal urothelium, myopodin was expressed in the nucleus as well as in the cytoplasm. The dual distribution was confirmed by transient overexpression of GFP–myopodin in the bladder cancer cell lines. The proportion of cells displaying nuclear staining in the normal urothelium may account for the standard percentages of cells under G1 phase. However, in the bladder tumors, especially when derived from invasive cancers, the nuclear myopodin expression was lost, suggesting the involvement of this protein in the progression of the disease. Furthermore, using tissue microarrays containing large series of samples obtained from tumors belonging to different patients with bladder cancer, we observed that the loss of nuclear myopodin expression was strongly associated with the stage and grade of the tumor. Myopodin could stratify patients with superficial and invasive disease, an observation of diagnostic utility. The more aggressive and advanced bladder tumors showed complete loss of or low nuclear myopodin expression. The concept of bladder cancer progression includes the development of invasive tumors from superficial tumors. Since the expression levels of myopodin were lower in invasive tumors as compared to superficial lesions, this event can be considered related to bladder cancer progression. Moreover, myopodin expression was also highly correlated with expression profiles of other molecular targets known to be involved in the progression of bladder cancer (Dalbagni et al., 1993; Reznikoff et al., 1994; Cordon-Cardo et al., 1997; Markl and Jones, 1998). Patients with preserved nuclear myopodin expression had a longer survival than those who showed a complete loss of nuclear staining. Most of the patients included in the series of patients with follow-up had invasive disease. Thus, myopodin could identify which patients subject to cystoprostatectomy would display a milder aggressive outcome. This finding is of clinical relevance because nuclear myopodin expression could stratify which patients should be subjected to adjuvant chemotherapy after surgical treatment. Further investigation with randomized studies would determine whether the loss of myopodin expression could predict chemotherapy response in invasive patients. Since there is lack of consensus in the use of prognosticators in clinical practice to predict the clinical outcome of individual tumors, the identification of novel molecular targets of predictive utility warrants detailed research to target specific clinical needs. An independent prognostic utility of myopodin in invasive disease can be suggested since nuclear myopodin expression was revealed as a good prognosticator independent of the loss of retinoblastoma expression, an early event in the progression of the disease (Cordon-Cardo et al., 1997; Lu et al., 2002).

This study has identified myopodin as a key protein involved in bladder cancer progression using clinical material and in vitro studies. It remains to be elucidated what the exact function of myopodin in normal urothelium is and how myopodin is redistributed in the progression of the disease. Biologically, inactivations of the key regulatory RB1 and TP53 pathways are necessary for the transformation and immortalization of bladder cancer (Dalbagni et al., 1993; Reznikoff et al., 1994; Cordon-Cardo et al., 1997; Markl and Jones, 1998; Lu et al., 2002). The known involvement of these pathways in bladder cancer progression was used as a reference to contrast the identified loss of nuclear myopodin. We observed that the majority of patients with invasive tumors under study showed low retinoblastoma and low nuclear myopodin expression simultaneously, making loss of nuclear myopodin expression a relative early event in the progression of the disease (Reznikoff et al., 1994; Markl and Jones, 1998; Lu et al., 2002). Interestingly, we also found an association between the nuclear localizations of myopodin and zyxin and the expression of E-cadherin, other well-characterized biomarkers that are lost in invasive disease (Nix and Beckerle, 1997; Giroldi et al., 1999; Popov et al., 2000). Like myopodin, zyxin can be present in the cytoplasm and the nucleus (Nix and Beckerle, 1997; Wang and Gilmore, 2000; Weins et al., 2001). Both zyxin and myopodin interact with -actinin in the cytoplasm but not in the nucleus (Crawford et al., 1992; Weins et al., 2001). Based on its speckled pattern reminiscent of actin snRNP aggregates, zyxin was suggested to be involved in an actin-based mRNA transport pathway (Sahlas et al., 1993).

The use of tissue microarrays has allowed us to address the relevance of myopodin in bladder cancer. This high-throughput approach has shown the potential to accelerate molecular studies that seek associations between molecular changes and clinicopathologic features of cancer. Nuclear myopodin has been identified as a novel target in bladder cancer progression in a cell-cycle-dependent manner. Loss of nuclear myopodin expression was related to bladder cancer histopathology both in vitro and in vivo in clinical material. It was also shown to be associated with pRB, adhesion pathways in a cell-cycle-dependent manner, suggesting a mechanistic role of myopodin in the progression of the disease. The potential tumor suppressor role of myopodin in bladder cancer suggested by our results requires further investigation.

In summary, nuclear myopodin expression provided clinical value to stratify superficial and invasive disease. More importantly, myopodin was revealed as a good prognosticator for disease-specific outcome. Myopodin was shown to be of clinical relevance by the predictive nature of its loss of expression with the outcome of patients with bladder cancer.

References

1. American Joint Committee on Cancer (eds) (1988). 3rd edn Manual for Staging of Cancer. JB Lippincott Co: Philadelphia, pp. 194–195.
2. Cancer Facts and Figures (2000). American Cancer Society.
3. Crawford AW, Michelsen JW and Beckerle MC. (1992). J. Cell. Biol., 116, 1381–1393. | Article | PubMed | ISI | ChemPort |
4. Cordon-Cardo C, Zhang ZF, Dalbagni G, Drobnjak M, Charytonowicz E, Hu SX, Xu HJ, Reuter VE and Benedict WF. (1997). Cancer Res., 57, 1217–1221. | PubMed | ChemPort |
5. Dalbagni G, Presti J, Reuter V, Fair WR and Cordon-Cardo C. (1993). Lancet, 324, 469–471. | Article |
6. Dawson-Saunders B and Trapp RG. (1994). Basic & Clinical Biostatistics 2nd edn Appleton & Lange: Norwalk, CT.
7. El-Rifai W, Kamel D, Larramendy ML, Shoman S, Gad Y, Baithun S, El-Awady M, Eissa S, Khaled H, Soloneski S, Sheaff M and Knuutila S. (2000). Am. J. Pathol., 156, 871–878. | PubMed |
8. Giroldi LA, Bringuier PP, Shimazui T, Jansen K and Schalken JA. (1999). Int. J. Cancer, 82, 70–76. | Article | PubMed | ChemPort |
9. Hoos A, Urist MJ, Stojadinovic A, Mastorides S, Dudas ME, Leung DHI, Kuo D, Brennan MF, Lewis JJ and Cordon-Cardo C. (2001). Am. J. Pathol., 158, 1245–1251. | PubMed | ISI | ChemPort |
10. Koo SH, Kwon KC, Ihm CH, Jeon YM, Park JW and Sul CK. (1999). Cancer Genet. Cytogenet., 110, 87–93. | Article | PubMed | ISI | ChemPort |
11. Lin F, Yu YP, Woods J, Cieply K, Gooding B, Finkelstein P, Dhir R, Krill D, Becich MJ, Michalopoulos G, Finkelstein S and Luo JH. (2001). Am. J. Pathol., 159, 1603–1612.
12. Lu ML, Wikman F, Orntoft TF, Charytonowicz E, Rabbani F, Zhang Z, Dalbagni G, Pohar KS, Yu G and Cordon-Cardo C. (2002). Clin. Cancer Res., 8, 171–179. | PubMed | ISI | ChemPort |
13. Markl ID and Jones PA. (1998). Cancer Res., 58, 5348–5353. | PubMed | ISI | ChemPort |
14. McShane LM, Aamodt R, Cordon-Cardo C, Cote R, Faraggi D, Fradet Y, Grossman HB, Peng A, Taube SE, Waldman FM (2000). Clin. Cancer Res., 6, 1854–1864. | PubMed | ISI | ChemPort |
15. Mostofi FK. (1973). Histological Typing of Urinary Bladder Tumors. Geneva: World Health Organization.
16. Mundel P, Heid HW, Mundel TM, Kruger M, Reiser J and Kriz W. (1997). J. Cell. Biol., 139, 193–204. | Article | PubMed | ISI | ChemPort |
17. Nix DA and Beckerle MC. (1997). J. Cell. Biol., 138, 1139–1147. | Article | PubMed | ISI | ChemPort |
18. Popov Z, Gil-Diez de Medina S, Lefrere-Belda MA, Hoznek A, Bastuji-Garin S, Abbou CC, Thiery JP, Radvanyi F and Chopin DK. (2000). Br. J. Cancer, 83, 209–214. | Article | PubMed | ISI | ChemPort |
19. Reuter VE and Melamed MR. (1989). Diagnostic Surgical Pathology Sternberg SS (ed). Raven Press: New York, pp. 1355–1392.
20. Reznikoff CA, Belair C, Savelieva E, Zhai Y, Pfeifer K, Yeager T, Thompson KJ, DeVries S, Bindley C and Newton MA et al. (1994). Genes Dev., 8, 2227–2240. | Article | PubMed | ISI | ChemPort |
21. Rosin MP, Cairns P, Epstein JI, Schoenberg MP and Sidransky D. (1995). Cancer Res., 15, 5213–5216.
22. Sahlas DJ, Milankov K, Park PC and De Boni U. (1993). J. Cell Sci., 105, 347–357. | PubMed | ISI | ChemPort |
23. Shaw ME, Elder PA, Abbas A and Knowles MA. (1999). Int. J. Cancer, 80, 656–661.
24. Simon R, Burger H, Semjonov A, Hertle L, Terpe HJ and Becker W. (2000). Int. J. Oncol., 17, 1025–1029. | PubMed | ISI | ChemPort |
25. Simon R, Burger H, Brinkschmidt C, Bocker W, Hertle L and Terpe HJ. (1998). J. Pathol., 185, 345–351. | Article | PubMed | ISI | ChemPort |
26. Wang Y and Gilmore TD. (2000). Biochim. Biophys. Acta, 1538, 260–272.
27. Weins A, Schwarz K, Faul C, Barisoni L, Linke WA and Mundel P. (2001). J. Cell. Biol., 155, 393–404. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank Gero von Gersdorff for his critical review of this manuscript.