Indian Journal of Pathology and Microbiology

: 2012  |  Volume : 55  |  Issue : 4  |  Page : 443--449

Epithelial mesenchymal transition in urothelial carcinoma: Twist in the tale

Purnima Paliwal, Disha Arora, Ashwani K Mishra 
 Department of Histopathology, National Institute of Pathology, Safdarjang Hospital Campus, New Delhi, India

Correspondence Address:
Purnima Paliwal
Department of Histopathology, National Institute of Pathology, Safdarjang Hospital Campus, Post Box 4909, New Delhi


Context: Epithelial to mesenchymal transition (EMT) is a process involving conversion of cells from an epithelial to mesenchymal phenotype. The role of candidate genes promoting EMT and favoring a promigratory phenotype has been demonstrated in epithelial cancer. Existing scientific research has not yielded a clinically relevant biomarker with predictive capacity beyond grade and stage in bladder cancer. Aim: The purpose of this study is to evaluate the immunohistochemical expression pattern of a panel of epithelial and mesenchymal markers in paraffin-embedded archival material of primary urothelial carcinoma as evidence of EMT. Materials and Methods: Immunohistochemical expression of transcription factor twist, epithelial (E-cadherin, cytokeratin) and mesenchymal (vimentin, N-cadherin) markers was analyzed on archival paraffin-embedded tissue samples from 48 patients with diagnosis of primary urothelial carcinoma of bladder. Statistical Analysis: Karl Pearson«SQ»s χ2 test was used to evaluate the association between the expression of immunohistochemical markers and various clinico-pathologic variables. Non-parametric Kendall«SQ»s tau-b statistics was used to determine the correlation between categorical variables. Results and Conclusion: The study demonstrated statistically significant association of cytokeratin, E-cadherin, vimentin, and twist with stage and grade of bladder cancer. Since these markers form part of the spectrum of changes associated with EMT, the study establishes proof of concept of the existence of this process in vivo. A significant negative correlation was noted between the expression of twist and E-cadherin. Exploiting its role as a transcriptional repressor of E-cadherin, twist may prove to be a useful candidate for targeted therapy in urologic oncology.

How to cite this article:
Paliwal P, Arora D, Mishra AK. Epithelial mesenchymal transition in urothelial carcinoma: Twist in the tale.Indian J Pathol Microbiol 2012;55:443-449

How to cite this URL:
Paliwal P, Arora D, Mishra AK. Epithelial mesenchymal transition in urothelial carcinoma: Twist in the tale. Indian J Pathol Microbiol [serial online] 2012 [cited 2020 Oct 28 ];55:443-449
Available from:

Full Text


Bladder cancer is a heterogeneous disease because tumors with similar stage and grade have divergent clinical manifestations and outcomes. The predictive accuracy of tumor-node-metastasis (TNM) staging alone is limited, and there is a lack of indices that could allow prospective assessment of risk for individual patients.

The molecular genetics of the metastatic cascade is complex and is actively investigated for signaling pathways and molecular targets that could prevent tumor spread. The hunt for the "metastasis signature" and the "metastatic oncogene" has pointed in the direction of epithelial to mesenchymal transition (EMT).

Cell lines displaying EMT phenotype have increased invasiveness and are associated with poor patient outcome in various tumors like breast, gastrointestinal, and prostate cancer. [1],[2],[3] Baumgart et al.[4] used immunohistochemistry to demonstrate that bladder cancer cell lines displaying EMT phenotype have enhanced motility and invasive properties. EMT classically involves upregulation of mesenchymal and downregulation of epithelial markers mediated by nuclear transcription factors. Analyzing the expression of these molecules across various grades and stages of urothelial carcinoma may provide useful information on the invasive/metastatic ability of these tumors. Recently, twist, which is a basic helix-loop-helix transcription factor, was found to be a key factor inducing EMT and promoting cancer cell invasion. [5]

With this background, we analyzed immunohistochemical expression of transcription factor twist, epithelial (E-cadherin, cytokeratin) and mesenchymal (vimentin, N-cadherin) markers in archival paraffin-embedded tissue samples of urothelial carcinoma as evidence of EMT. The aim of the study was to evaluate potential biomarkers from the EMT pathway that can be objectively quantified and incorporated into routine pathologic analysis.

The molecular targeting of the oncogenic signaling elements involved in the EMT program may formulate strategies to counteract the progression from localized cancers into disseminated diseases.

 Materials and Methods

Patient Selection

The study included 48 consecutive patients with diagnosis of primary urothelial carcinoma of bladder who were registered with the Department of Urology of the hospital affiliated with the institute. These patients had undergone a transurethral resection or a partial/total cystectomy between 2007 and 2010. Archival formalin-fixed paraffin embedded tumor tissue samples of selected patients were taken from the files of the institute. Clinico-pathologic data are summarized in [Table 1]. The subjects were 32-80 years old with a median age of 62 years. Males constituted the majority in the sample set (79%), and the rest were females.{Table 1}

All histology slides were reviewed for histopathologic grading and staging. Tumor staging was determined using TNM (1997 UICC TNM) classification and grading was based on WHO 2004/ISUP 1998 guidelines. Representative blocks were chosen for immunohistochemistry. Trans-urethral-resection of bladder tumor (TURBT) samples with insufficient material for immunohistochemical analysis and those with highly necrotic tumors were excluded from the study. Two blocks were selected from the radical cystectomy specimens, one of which included the invasive edge of tumor. The cases were categorized into two main groups-muscle invasive (MIBC) and non-muscle invasive (NMIBC), with 25 cases (52.1%) of muscle invasive cancer and 23 cases (47.9%) of non-muscle invasive cancer (n = 48). Of the 48 samples, 18 (37.5%) were low grade and 30 (62.5%) were high grade. Information on nodal metastasis was obtained from radiology records. Regional lymph node dissection was performed in 11 of the 15 radical cystectomies and showed metastatic deposits in 4 cases.

Immunohistochemical Evaluation

Representative paraffin sections of 48 chosen primary urothelial cancer tissues were selected. Super-sensitive polymer HRP IHC detection system (BioGenex, San Ramon, CA, USA) was used. Tissue sections of 4 μm were used on poly-l-lysine coated slides for immunohistochemical analysis. Antigen retrieval was done with citrate buffer (pH 6, 0.01 mol/L) by using microwave technique. Incubation with primary antibody was carried out in a moist chamber at room temperature for 60 min for cytokeratin, vimentin, and E-cadherin; and overnight for N-cadherin and twist [Table 2]. Negative controls were treated identically with the primary antibodies omitted. Positive controls consisted of tissue known to contain the protein of interest [Table 2]. Colonic mucosa was the best positive control for twist, showing strong patchy staining in the crypt epithelium [Figure 1]. The staining was visualized by incubating with 3,3′-diaminobenzidine for 3-5 min, and then counterstained with hematoxylin.{Figure 1}{Table 2}

For evaluation of immunohistochemistry, 1000 cells were counted at magnification 400×. The tissue sections were scored semiquantitatively, assessing staining intensity and protein localization, including membrane, cytoplasmic, or nuclear localization. Scoring was performed by two independent observers and the discordant cases were reanalyzed to reach a consensus. The normal expression patterns, i.e., cytoplasmic staining for cytokeratin and vimentin, membranous for cadherins and nuclear for twist, were considered for statistical analysis.

A combined score was given, multiplying the percentage score of tumor cells expressing the marker and the intensity (score) with which the tumor cells expressed it. The percentage of stained cells was noted on scores of 0-3 as 0 (no detectable immunostain), 1 (1%-10% nuclei), 2 (11%-50% nuclei), and 3 (>50% nuclei). [6] The intensity was scored as 1-3 (mild/moderate/strong), respectively.

Cytokeratin and E-cadherin are markers of epithelial phenotype; therefore, strong diffuse expression is expected in urothelial cancer. For statistical analysis, expression of these markers was reported as strong or reduced based on a cut-off taken as the first quartile of the combined score, which was 4 for both the antibodies. Score of >4 was reported as strong expression and ≤4 was reported as reduced expression.

Any immunohistochemical expression of vimentin and N-cadherin was noted as positive because the presence of either of these mesenchymal markers in urothelial carcinoma cells represents novel expression. In benign urologic disease, only few epithelial cells are twist positive; a threshold of 2% has been defined beyond which immunohistochemical positivity for novel twist expression was found sensitive. [3]

Statistical Analysis

Karl Pearson's χ2 test was used to evaluate the association between the expression of immunohistochemical markers and various clinico-pathologic variables. Non-parametric Kendall's tau-b statistics was used to determine the correlation between categorical variables. Two-sided P < 0.05 was considered statistically significant. Data analysis was performed using SPSS 19.0 version software.


On histopathologic review, there was agreement on diagnosis and staging in all cases. Majority of the cases were conventional urothelial carcinoma with a few histologic variants. These included two cases of sarcomatoid variant and one micropapillary variant of urothelial carcinoma.

Immunohistochemical Expression of Cytokeratin, E-cadherin, Vimentin, N-cadherin, and twist

Results of immunohistochemistry in primary tumor tissue and their correlation with stage and grade of tumor are summarized in [Table 3] and [Table 4].{Table 3}{Table 4}

Cytokeratin Expression

Cytokeratin immunostaining was diffuse and strong in the normal bladder epithelium. In the urothelial carcinoma cases, the staining pattern was uniform throughout the tumor, with staining scores ranging from 1 to 9 with a median score of 6. A reduction or loss of cytokeratin expression was significantly correlated with pathologic tumor stage (P < 0.004) and histologic grade (P < 0.001) of tumor.

Vimentin Expression

Novel cytoplasmic expression of vimentin was seen in 20% of MIBC. Three cases showed strong diffuse cytoplasmic positivity in more than 70% of tumor cells, two of which were sarcomatoid variants of urothelial carcinoma. In the category of sarcomatoid carcinomas, neighboring sections stained for pan-cytokeratin showed focal immunoreactivity suggesting the epithelial origin of the vimentin-positive cells. In the remaining two cases, vimentin was expressed in 10%-20% of tumor cells. None of the NMIBC expressed vimentin. Association between vimentin expression and pathologic stage of bladder cancer was significant (P = 0.023). No difference in staining patterns was observed between the center of tumor and invasive edge, However, focal cytoplasmic vimentin immunoreactivity was noted at the invasive edge in a single case of muscle invasive high-grade urothelial cancer.

E-cadherin Expression

Staining pattern was uniform throughout the tumor with scores ranging from 1 to 9 with a median score of 6. A reduction or loss of E-cadherin expression was significantly correlated with pathologic tumor stage (P < 0.001) and histologic grade (P < 0.01) of tumor [Figure 2]a, b. Aberrant localization of E-cadherin in the cytoplasm or nucleus was not observed in any of our cases. In some of the cases showing poor tissue preservation or prominent cautery artifact, the staining appeared to be cytoplasmic. However, better preserved areas in the same tissue demonstrated membranous staining.{Figure 2}

N-cadherin Expression

Novel membranous expression was noted in 20% of MIBC and 13% of NMIBC. Most cases showed focal incomplete membranous staining with mild to moderate intensity and scores ranging from 1 to 4 [Figure 3]a. Cytoplasmic staining was noted in two cases of NMIBC with one case showing moderate cytoplasmic positivity in 30%-40% of tumor cells [Figure 3]b. However, as for E-cadherin, only membranous staining was considered for statistical analysis. Focal N-cadherin expression was noted in the tumor-associated stroma in five cases even in the absence of epithelial expression. Association of N-cadherin expression with pathologic stage or grade of bladder cancer was not statistically significant (P = 0.424, 0.518).{Figure 3}

Twist Expression

Twist, being a transcription factor, only nuclear expression was taken into account for statistical analysis. Staining was focal and heterogeneous in distribution with scores ranging from 1 to 4. Novel nuclear expression of twist showed significant association with pathologic stage and grade of bladder cancer (P = 0.047, 0.044). Nuclear twist expression was noted in 32% of the MIBC as against 8.7% of NMIBC [Figure 4]. Of the high-grade tumors, 30% showed nuclear expression of twist versus 5.6% of low-grade tumors. Focal cytoplasmic expression of twist with mild-moderate intensity was noted in 8 (16.7%) cases. Tumor-associated stroma did not show immunoreactivity to twist.{Figure 4}

Adjoining normal/dysplastic bladder epithelium showed diffuse strong staining for cytokeratin and E-cadherin. Immunoreactivity for twist/vimentin/N-cadherin was not observed in this tissue. Immunohistochemical evaluation of the four nodal metastatic deposits revealed novel cytoplasmic/nuclear twist expression in three cases, reduced E-cadherin in all, reduced cytokeratin in three, and N-cadherin expression in one case. None of these deposits showed vimentin expression.

Associations Among Immunohistochemical Markers

All the cases which expressed twist had reduced expression of cytokeratin and E-cadherin. Thus, novel expression of twist demonstrated a statistically significant negative correlation with cytokeratin and E-cadherin (<0.001, P = 0.001). Using the non-parametric Kendall's Tau-b statistics, cytokeratin and E-cadherin expression correlated positively (P < 0.01). Reduced E-cadherin expression correlated with reduced cytokeratin (P < 0.01), novel vimentin (P < 0.05), and novel nuclear twist (P < 0.01) expression. Also, cytokeratin and E-cadherin expression showed negative correlation with vimentin (P < 0.05) and twist (P < 0.01) expression. No significant associations of N-cadherin were found with any of the other markers.

Association of Epithelial Markers (cytokeratin and E-cadherin) with Stage and Grade of Urothelial Cancer

When the expression of epithelial markers (cytokeratin and E-cadherin) was analyzed as a group, their association with stage and grade of urothelial cancers was significant (P = 0.002, 0.003).

Association of Novel Expression of EMT Markers (vimentin, N-cadherin, twist) with Stage and Grade of Urothelial Cancer

When considered as a group, association of these markers and pathologic stage of urothelial cancer was significant (P = 0.010). However, the same could not be demonstrated with grade (P = 0.160).


Cytokeratin and vimentin form an integral part of the EMT process and were chosen based on their robust immunohistochemical staining results. The "cadherin switch" represents one of the defining features of EMT and its prognostic value has been reported in bladder tumorigenesis. [6],[7] Twist was chosen based on recent reports highlighting its inverse relationship with E-cadherin and its potential role as a prognostic factor in bladder cancer. [3],[8]

Cytokeratins are one of the hallmarks of the epithelial phenotype. In accordance with the concept of EMT, reduced expression of cytokeratin correlated significantly with the invasiveness and histologic grade of tumor. However, alteration in expression pattern of cytokeratins with stage/grade of urothelial carcinoma has been previously reported and is not specific to EMT. [9]

In this study, reduced cytokeratin expression correlated with novel vimentin expression. Vimentin expression has been linked to high invasive ability of the tumor due to motile mesenchymal phenotype. [10],[11] Accordingly, in this study, none of the non-muscle invasive tumors demonstrated vimentin expression. This is in contrast to the findings of Baumgart et al. who reported vimentin expression in 6.8% of superficial tumors. [4] Vimentin expression in this study was confined to high-grade and high-stage tumors. This may imply that novel expression of vimentin is a late event in the pathway of EMT, and therefore its utility as a biomarker is questionable.

Sarcomatoid variants of urothelial carcinoma are highly aggressive tumors and are resistant to current standard treatment protocols. The transition to mesenchymal phenotype (EMT) precedes the progression to microscopically recognizable sarcomatoid transformation and occurs in a subset of conventional carcinomas. [12] Thus, the presence of strong diffuse vimentin immunoexpression may be extrapolated to signify an aggressive subtype of urothelial carcinoma.

E-(epithelial) cadherin has a significant function in intercellular adhesion of epithelial cells, the establishment of epithelial polarization, glandular differentiation, and stratification. A reduction or loss in expression of E-cadherin is a hallmark of EMT. [3] Our results correlated with those of other authors in that E-cadherin expression was significantly correlated with stage and histologic grade of urothelial cancer. [13],[14] The correlation of reduced E-cadherin expression with reduced cytokeratin, novel vimentin, and novel nuclear twist expression was consistent with mesenchymal phenotype of urothelial carcinoma.

It has been reported that a change in the location of E-cadherin (from membrane to cytoplasm) correlates with higher aggressiveness of the tumor. [15] Disassembly of adherence junctions, which are arranged by cadherins, is the primary event in acquisition of the mesenchymal phenotype. Since these are located in the cell membrane, membranous staining should be considered as evidence of epithelial phenotype, and both absence of membranous staining or cytoplasmic staining have similar implications. It has been postulated that cytoplasmic staining is the result of disturbance in E-cadherin-cytoskeleton interaction. [16] In the present study, none of the cases in either category demonstrated cytoplasmic staining. It is essential to ensure that poor tissue preservation/cautery artifact/high nuclear-cytoplasmic ratio does not create a false impression of cytoplasmic staining.

Loss of E-cadherin and novel expression of N-cadherin/P-cadherin (cadherin switch) is one of the defining features of EMT, and has been reported in bladder tumorigenesis. [6],[17] Although E-cadherin downregulation is an absolute prerequisite for EMT, this may or may not be accompanied by simultaneous expression of N-cadherin. No significant correlation was observed between E-cadherin expression and N-cadherin expression in our study. Occasionally, N-cadherin immunoreactivity was also noted in the tumor stroma. Some authors have related the stromal positivity to the role of this molecule in cancer cell motility. [3]

Correlation of N-cadherin was not found to be significant with either clinico-pathologic variables or with other immunohistochemical markers. Results of other authors on urothelial carcinoma have not been encouraging. [4] Adding to the controversy, recently, N-cadherin-negative muscle-invasive urothelial tumors have been reported to have poor prognosis. [18] It is desirable to study this marker with a larger number of samples to conclusively prove or disprove its association with the invasiveness of tumor.

Twist is a basic helix-loop-helix transcription factor which is one of the key molecules regulating EMT. It is considered to be a potential oncogene promoting proliferation and inhibiting apoptosis. [19] In the development of metastatic cancer cells by type 3 EMT, twist can act independently to repress E-cadherin and to upregulate fibronectin and N-cadherin. [20] Expression of twist is sufficient to induce in vitro EMT in breast cells and its inactivation inhibits the development of metastasis in vivo. [5]

Living up to expectations, novel nuclear twist expression significantly correlated with the stage of urothelial cancer and the histologic grade of tumor. However, it is worth mentioning that this molecule requires prolonged and meticulous antigen retrieval. Also, the staining in tumor tissue is focal and heterogeneous, which would make evaluation in small biopsies/tissue microarray cores unreliable. Mild focal cytoplasmic expression of twist was also noted in 16.7% cases. Twist, being a transcription factor, only nuclear expression was taken into account for statistical analysis. This issue is debatable since some studies have included both cytoplasmic and nuclear immunoreactivity for the purpose of EMT. [21] The basis of aberrant localization is unclear and its correlation to EMT is speculative.

The transcriptional repression of E-cadherin is one of the main mechanisms in gene silencing, and twist plays a pivotal role in this process. [19] In accordance with this data, all the cases with novel nuclear twist expression had reduced expression of E-cadherin and cytokeratin. Some authors have stated that twist overexpression induces neoexpression of mesenchymal markers such as N-cadherin, vimentin, and fibronectin. [22] However, we found no significant association with vimentin or N-cadherin expression.

It was noteworthy that the characteristic EMT profile of reduced E-cadherin + cytokeratin and novel expression of vimentin + N-cadherin + twist was rarely observed concurrently in our study. EMT is accompanied by altered expression of a wide variety of epithelial/mesenchymal markers. However, the loss or gain of these markers at a given point of time may not be complete. It is a dynamic process with a continuum of changes occurring in sequential fashion. Acquisition of mesenchymal phenotype may be transitory, therefore demonstrating the same in tissue samples obtained at one point of time may not be feasible. This continuum has been well demonstrated by various authors in studies using gene expression on tumor cell lines of varying degrees of invasiveness. [4]

Other than predicting tumor aggressiveness, the recognition of EMT may have an impact on the choice of therapeutic regimen. It has been demonstrated that tumor cells with mesenchymal phenotype upregulate alternate receptor tyrosine kinase pathways as mechanisms for survival signaling. [23] Results of chemotherapy have been disappointing in improving survival of urothelial carcinoma. The inhibition of twist through small interfering RNAs (siRNAs) has been demonstrated to restore chemosensitivity in urothelial tumors. [3] Being a transcriptional repressor of E-cadherin, twist is a potential candidate for targeted therapy in urologic oncology.

When the epithelial markers were considered as a group, significant association was noted between the number of markers altered and stage/grade. The same was true for novel expression of markers (vimentin, N-cadherin, twist) when correlated with stage. The use of multiple markers was complimentary in supporting the hypothesis of EMT.

The present study demonstrates altered immunoexpression and statistically significant association of cytokeratin, E-cadherin, vimentin, and twist with stage and grade of bladder cancer. Since these markers form part of the spectrum of changes associated with EMT, the study establishes proof of concept of the existence of this process in vivo. However, the transitory nature and heterogeneous distribution of this process may hamper the utility of immunohistochemistry on archival tumor tissue as a prognostic tool. No single marker could conclusively confirm/exclude the occurrence of EMT; these markers need further evaluation as independent prognostic variables in larger cohorts with survival data analysis.


1Blick T, Widodo E, Hugo H, Waltham M, Lenburg ME, Neve RM, et al. Epithelial mesenchymal transition traits in human breast cancer cell lines. Clin Exp Metastasis 2008;25:629-42.
2Rosivatz E, Becker KF, Kremmer E, Schott C, Blechschmidt K, Hofler H, et al. Expression and nuclear localization of Snail, an E-cadherin repressor, in adenocarcinomas of the upper gastrointestinal tract. Virchows Arch 2006;448:277-87.
3Fondrevelle M, Kantelip B, Reiter RE, Chopin DK, Thiery JP, Monnien F, et al. The expression of Twist has an impact on survival in human bladder cancer and is influenced by the smoking status. Urol Oncol 2009;27:268-76.
4Baumgart E, Cohen MS, Neto BS, Jacobs MA, Wotkowicz C, Rieger-Christ KM, et al. Identification and prognostic significance of an epithelial-mesenchymal transition expression profile in human bladder tumors. Clin Cancer Res 2007;13:1685-94.
5Yang J, Mani SA, Donaher JL, Ramaswamy S, Itzykson RA, Come C, et al. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 2004;117:927-39.
6Gravdal K, Halvorsen OJ, Haukaas SA, Akslen LA. A switch from E-cadherin to N-cadherin expression indicates epithelial to mesenchymal transition and is of strong and independent importance for the progress of prostate cancer. Clin Cancer Res 2007;13:7003-11.
7Wheelock MJ, Shintani Y, Maeda M, Fukumoto Y, Johnson KR. Cadherin switching. J Cell Sci 2008;121:727-35.
8Peinado H, Olmeda D, Cano A. Snail, Zeb and bHLH factors in tumour progression: An alliance against the epithelial phenotype? Nat Rev Cancer 2007;7:415-28.
9Schaafsma HE, Ramaekers FC, van Muijen GN, Lane EB, Leigh M, Robben H, et al. Distribution of cytokeratin polypeptides in human transitional cell carcinomas, with special emphasis on changing expression patterns during tumor progression. Am J Pathol 1990;136:329-43.
10Ivaska J, Pallari HM, Nevo J, Eriksson JE. Novel functions of vimentin in cell adhesion, migration, and signalling. Exp Cell Res 2007;313:2050-62.
11Lang SH, Hyde C, Reid IN, Hitchcock IS, Hart CA, Gordon Bryden AA, et al. Enhanced expression of vimentin in motile prostate cell lines and in poorly differentiated and metastatic prostate carcinoma. Prostate 2002;52:253-63.
12McConkey D, Lee S, Choi W, Tran M, Majewski T, Lee S, et al. Molecular genetics of bladder cancer: Emerging mechanisms of tumor initiation and progression. Urol Oncol 2010;28:429-40.
13Frixen UH, Behrens J, Sachs M, Eberle G, Voss B, Warda A, et al. E-cadherin-mediated cell-cell adhesion prevents invasiveness of human carcinoma cells. J Cell Biol 1991;113:173-85.
14Byrne RR, Shariat SF, Brown R, Kattan MW, Morton RA Jr, Wheeler TM, et al. E-cadherin immunostaining of bladder transitional cell carcinoma, carcinoma in situ and lymph node metastases with long-term followup. J Urol 2001;165:1473-9.
15Stanczak A, Stec R, Bodnar L, Olszewski W, Cichowicz M, Kowzlowski W, et al. Prognostic significance of Wnt-1, β-catenin and E-cadherin expression in advanced colorectal carcinoma. Pathol Oncol Res 2011;17:955-63.
16Bringuier PP, Umbas R, Schaafsma HE, Karthaus HF, Debruyne FM, Schalken JA. Decreased E-cadherin immunoreactivity correlates with poor survival in patients with bladder tumors. Cancer Res 1993;53:3241-5.
17Bryan RT, Atherfold PA, Yeo Y, Jones LJ, Harrison RF, Wallace DM, et al. Cadherin switching dictates the biology of transitional cell carcinoma of the bladder: Ex vivo and in vitro studies. J Pathol 2008;215:184-94.
18Jäger T, Becker M, Eisenhardt A, Tilki D, Totsch M, Schmid KW, et al. The prognostic value of cadherin switch in bladder cancer. Oncol Rep 2010;23:1125-32.
19Wallerand H, Robert G, Pasticier G, Ravaud A, Ballanger P, Reiter RE, et al. The epithelial-mesenchymal transition-inducing factor TWIST is an attractive target in advanced and/or metastatic bladder and prostate cancers. Urol Oncol 2010;28:473-9.
20Zeisberg M, Neilson EG. Biomarkers for epithelial-mesenchymal transitions. J Clin Invest 2009;119:1429-37.
21Sasaki K, Natsugoe S, Ishigami S, Matsumoto M, Okumura H, Setoyama T, et al. Significance of Twist expression and its association with E-cadherin in esophageal squamous cell carcinoma. J Exp Clin Cancer Res 2009;28:158-66.
22Kang Y, Massague J. Epithelial-mesenchymal transitions: TWIST in development and metastasis. Cell 2004;118:277-9.
23Barr S, Thomson S, Buck E, Russo S, Petti F, Sujka-Kwok I, et al. Bypassing cellular EGF receptor dependence through epithelial-to-mesenchymal-like transitions. Clin Exp Metastasis 2008;25:685-93.