Skip Navigation


Carcinogenesis Advance Access originally published online on July 8, 2006
Carcinogenesis 2007 28(2):246-258; doi:10.1093/carcin/bgl120
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
28/2/246    most recent
bgl120v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (11)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Boireau, S.
Right arrow Articles by Hollande, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Boireau, S.
Right arrow Articles by Hollande, F.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

DNA-methylation-dependent alterations of claudin-4 expression in human bladder carcinoma

Stéphanie Boireau1,{dagger}, Michael Buchert1,{dagger}, Michael S. Samuel2, Julie Pannequin1, Joanne L. Ryan1,8, Armelle Choquet1, Héliette Chapuis3, Xavier Rebillard5, Christophe Avancès4,6, Matthias Ernst2, Dominique Joubert1, Nicolas Mottet4,7 and Frédéric Hollande1,*

1 CNRS UMR5203, INSERM U661, Université Montpellier I, Université Montpellier II, Institut de Génomique Fonctionnelle, Département d'Oncologie Cellulaire et Moléculaire, 141 Rue de la Cardonille Montpellier F-34094 Cedex 5, France
2 Ludwig Institute for Cancer Research, Parkville Vic 3050, Australia
3 Service d'Anatomo-pathologie, CHU Groupe Hospitalisation Carémeau Nîmes, France
4 Service d'Urologie, CHU Groupe Hospitalisation Carémeau Nîmes, France
5 Service d'Urologie, Clinique Mutualiste Beausoleil 34000 Montpellier, France
6 Service d'Urologie, Polyclinique Kennedy 30000 Nîmes, France
7 Present address: Service d'Urologie, Clinique Mutualiste St Etienne, France
8 Present address: UMR CNRS 5810, Faculté de Pharmacie Montpellier, France

*To whom correspondence and requests for reprints should be addressed at: Institut de Génomique Fonctionnelle, Faculté de Pharmacie, Laboratoire de Biochimie, Bâtiment E, 15 Avenue C. Flahault, 34093 Montpellier Cedex 5, France. Fax: +33 0 4 67 66 81 49; Email: frederic.hollande{at}univ-montp1.fr


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
 References
 
The expression pattern of tight junction (TJ) proteins is frequently disrupted in epithelial tumors. In particular, isoform- and organ-specific alterations of claudins have been detected in human cancers, highlighting them as interesting tools for the prognosis or treatment of various carcinomas. However, the molecular mechanisms responsible for these alterations are seldom identified. Here, we analyzed the expression and localization of claudins 1, 4, and 7 in human bladder carcinoma. Claudin-4 expression was significantly altered in 26/39 tumors, contrasting with the rare modifications detected in the expression of claudins 1 and 7. Overexpression of claudin-4 in differentiated carcinomas was followed by a strong downregulation in invasive/high-grade tumors, and this expression pattern was associated to the 1-year survival of bladder tumor patients. A CpG island was identified within the coding sequence of the CLDN4 gene, and treatment with a methyl-transferase inhibitor restored expression of the protein in primary cultures prepared from high-grade human bladder tumors. In addition, claudin-4 expression correlated with its gene methylation profile in healthy and tumoral bladders from 20 patients, and downregulation of claudin-4 expression was detected in the urothelium of mice overexpressing DNA methyl transferase 3a (Dnmt3a). Delocalization of claudins 1 and 4 from TJs was observed in most human bladder tumors and in the bladder tumor cell line HT-1376. Although the CLDN4 gene was unmethylated in these cells, pharmacological inhibition of methyl transferases re-addressed the two proteins to TJs, resulting in an increase of cell polarization and transepithelial resistance. These biological effects were prevented by expression of claudin-4-specific siRNAs, demonstrating the important role played by claudin-4 in maintaining a functional regulation of homeostasis in urothelial cells. Results of this study indicate that the TJ barrier is disrupted from early stages of urothelial tumorigenesis. In addition, we identified hypermethylation as the mechanism leading to the alteration of claudin-4 expression, and maybe also localization, in bladder carcinoma.

Abbreviations: CPE, Clostridium Perfringens Enterotoxin; Dnmt3a, DNA methyl transferase 3a; MSP, methylation-specific PCR; SMA, smooth muscle actin; TER, Transepithelial resistance; TJ, tight junction


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 Supplementary material
 References
 
Tight junctions (TJs) are the most apical component of cell/cell complexes, constituting a continuous seal around the apical cell border and playing a critical role in regulating the balance between differentiation, proliferation and cell death (1,2). They are made up of at least three types of integral transmembrane proteins, which are responsible for the formation of a network of intramembrane fibrils, and by a growing number of cytoplasmic TJ plaque proteins connecting the transmembrane proteins to the cytoskeleton and serving as a platform for cell signaling and protein trafficking (3). Among the integral proteins cooperating to form TJ, the recent literature describes the modulation of claudins during oncogenic transformation, indicating that they could play an important role in influencing tumor progression and invasion (4,5). Recent reports suggest that TJ protein expression could be disrupted before the onset of epithelial/mesenchyme transition (6), but it remains unclear at what stage of tumor development alterations of claudin expression occur, and whether their disruption is associated with tumor progression. In the urothelial bladder, TJs are formed between adjacent umbrella cells, where they play an important role in the separation between the basolateral membrane and the apical membrane, where specific proteins named uroplakins create a barrier against transcellular ion permeability (7). Several TJ proteins have been detected in the urothelium, among them claudins 1, 4, 8 and 12 (8). Two of these proteins, claudins 1 and 4, are expressed at TJs and along the cell/cell membrane of umbrella cells, but are also detected in intermediate and basal cells (8–10), where their function remains unclear. In addition, preliminary results from our group indicated that urothelial cells also expressed significant levels of claudin-7, a ubiquitous claudin isoform previously shown to localize both at TJs and along the basolateral membrane of epithelial cells in other tissues (11). The expression of claudins 1, 4 and 7 was shown to be disrupted in other types of cancer (12–14). Indeed, overexpression of the CLDN-4 gene, sometimes coupled to cytoplasmic relocalization has been detected in ovarian or pancreatic carcinomas (15,16), and its interaction with the Clostridium Perfringens Enterotoxin (CPE) offers interesting prospects for cancer treatment (17,18). Furthermore, claudin-1 was recently shown to regulate cellular transformation and metastatic behavior in colon cancer (19) and to constitute a promising prognostic marker for stage II colon cancer (20). Claudin-7 overexpression was detected in hepatocellular carcinoma, as well as early gastric cancers (13,21), while methylation-dependent downregulation of this protein seems correlated with the progression of breast carcinoma (22). However, the contribution of these proteins to tumorigenesis of the bladder is still unknown. In fact, very little data is currently available on the molecular events related to TJ regulation in urothelial cancer, although a decreased number of TJ strands was detected by freeze fracture electron microscopy 20 years ago in transitional cell carcinomas of the bladder (23). In view of the critical role played by TJs in the protection against toxins and proinflammatory agents contained in urine, and since claudins 1, 4 and 7 have been identified as relevant targets in other types of carcinomas, the purpose of this study was to analyze the expression pattern of these three proteins in human bladder carcinoma. Our results demonstrate the occurrence of profound alterations in the expression and localization of claudin–4 in bladder carcinoma. In addition, the methylation profile of CLDN-4 gene coding sequence was found to regulate the expression level of claudin-4 in the bladder urothelium, and to be involved in the profound downregulation of the protein in advanced bladder cancer.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 Supplementary material
 References
 
Antibodies
Anti-claudins 1, 4 and 7 antibodies were from Zymed (San Francisco, CA), pan-cytokeratin, smooth muscle actin (SMA), and desmin from Sigma (St Quentin Fallavier, France). Alexa Fluor 488-labeled was from Molecular Probes, while Cy3-labeled and horseradish peroxidase (HRP)-conjugated antibodies were from Chemicon, (Mundolsheim, France). Antibodies for cis-Golgi and trans-Golgi markers (GM130 and p230) and endosomal vesicle proteins (EEA-1, Rab 7, CD63) were provided by Drs C. Brock, S. Roche and M. Vidal.

Primary cultures
Tumor samples were cut into 1 mm pieces and digested for 10 min at 37°C, in Dulbecco's Modified Eagle Medium (DMEM) containing 1 mg/ml collagenase P (Roche Diagnostics, Meylan, France), 1 mg/ml protease and 0.8 mg/ml type I soybean trypsin inhibitor (Sigma). Undigested tumor pieces and dissociated cells were resuspended in growth medium and seeded on collagen IV-coated dishes (Sigma). When needed, cells were lysed after 4-day treatments in growth medium with or without 10 µM 5-aza-deoxycytidine (5-aza-CdR; Sigma), before extracting proteins, genomic DNA and RNA, as described elsewhere.

Cell culture and transient transfections
HT-1376 cells were maintained at 37°C in DMEM supplemented with 10% fetal bovine serum (FBS; Eurobio, Les Ulis, France), 1% L-glutamine, 1% MEM non-essential amino acid and 1% MEM vitamin (Sigma) and penicillin/streptomycin. Cells were seeded in 12-well plates (1 x 105 cells/well), 2 days before transfection. Cells were transfected with 500 ng of claudin-4 or control siRNA, and 1.65 µl/well of PEI (Qbiogene, Illkirch, France) was used to transfect cells according to the manufacturer's instructions. After 72 h post-transfection, cells were lysed (for immunoblotting or PCR) or fixed in 2.5% paraformaldehyde for immunostaining. Control siRNA was directed against ß-galactosidase, and claudin-4-specific siRNA was a mixture of the three following siRNA oligonucleotides: Sense1: 5'-GCUGAACAAUGGCCUCCAUTT-3', Antisense1: 5'-AUGGAGGCCAUUGUUCAGCTT-3'; Sense2: 5'-CCUUACUCCGCCAAGUAUUTT-3', Antisense2: 5'-AAUACUUGGCGGAGUAAGGTT-3'; Sense3: 5'-CCACCCUCCUCUGGAUAUUTT-3', Antisense3: 5'-AAUAUCCAGAGGAGGGUGGTT-3'.

Transepithelial resistance (TER) measurement
TER was measured as described previously (24). Cells (1 x 105 cells/well) were seeded in the apical compartment of 0.4 µm pore size cell culture insert (BD, Le pont de Claix, France). Inserts were placed in 12-well tissue culture plates (BD), containing 1 ml of medium in the basolateral compartment and incubated at 37°C with 5% CO2. In all experiments, epithelial cells were grown to confluence on cells inserts. TER was monitored as an indication of TJ formation and epithelial monolayer integrity using a Millicell-ERS Voltohmmeter, (Millipore, Molsheim, France). Electrical resistance measurements were taken daily in two sites on each side of the transwell membrane, in order to ensure that measurements reflected as well as possible the resistance across the whole monolayer surface. The mean of these two measurements was calculated, and results were expressed as Ohms x cm2, after subtracting the values for the resistance of the membrane alone.

Human tumor samples
Specimens of papillary bladder carcinomas and macroscopically and histologically normal urothelium from 39 patients were obtained after transurethral bladder resection or cystectomy, according to French government regulations and local committee guidelines. Tumor samples were provided by the pathologist and represented mostly the exophytic component of both early and late stage tumors. Tumors were grouped into superficial (stage pTa) or invasive (stage ≥ pT1) according to the stages described in the TNM 2002 classification (25), and were graded following the 1998 update of the World Health Organization/International Society of Urologic Pathologists grading system (26) into: well-differentiated (grade G1), moderately differentiated (grade G2) or poorly differentiated (grade G3). Sections were prepared from frozen tumor samples, and immunohistochemical detection of claudins was performed on randomly chosen sections from three areas of each sample. The epithelial content was assessed by staining adjacent sections with hematoxylin/eosin. Other sections were lysed for western blot analysis and, when sufficient quantities were available, the homogeneity of urothelial/mesenchymal tissue proportions within the samples was confirmed by probing western blots with antibodies against stromal markers (desmin and SMA). Whenever possible, adjacent sections were also used for RNA extraction and genomic DNA preparations.

Immunoblotting
Protein lysates were prepared at 4°C in RIPA buffer [20 mM Tris (pH 7.5), 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 1% Na-deoxycholate, 50 mM NaF, 40 mM beta-glycerol phosphate, 10 mM HEPES (pH 7.3), 5 mM EDTA, 0.2 mM Na-orthovanadate, 1 mM DTT, 1% protease inhibitor cocktail (Sigma)] from ~50 tissue cryosections per patient (20 µl/mg tissue), or from confluent cell cultures. Proteins (20 µg) were separated on 12% SDS–PAGE, and transferred onto nitrocellulose membranes using a semi-dry western blotting apparatus from Bio-rad (3 mA/cm2, 40 min). After saturation in phosphate-buffered saline (PBS), 5% milk, 0.1% Tween-20, membranes were incubated overnight at 4°C with the primary antibodies [antibody dilutions: claudins 1, 4 and 7 (1/500); actin (1/5000); SMA and desmin (1/1000)]. After incubating with HRP-conjugated secondary antibodies (anti-rabbit or anti-mouse IgG, 1/5000 dilution), proteins were visualized using the ECL Plus system (Amersham, Orsay, France).

Immunofluorescence
Sections were fixed in 1% paraformaldehyde, permeabilized and incubated with 50 mM NH4Cl to reduce background staining. After saturation with PBS, 5% bovine serum albumin (BSA), 0.3% gelatin, primary antibodies were incubated overnight at 4°C (claudins 1 and 4 antibodies, 1/100 dilution; claudin-7 antibody, 1/50). Tissue sections were then washed with PBS and incubated with Alexa Fluor 488- and/or Cy3-coupled secondary antibodies (1/500 dilution). The samples were mounted in mowiol (Aldrich, Lyon, France). Slides were observed using a Biorad MRC 1024 confocal microscope. ‘Blind’ quantification of the immunofluorescence data was performed by two independent observers, and scoring was done in two ways: (i) in the scoring for staining intensity, 0 = weaker staining in the tumor than in the control urothelium, 1 represented similar intensity in both samples, and 2 represented higher intensity in the tumor sample; (ii) intracellular localization of the proteins, irrespective of the overall intensity, was determined with a 4-point scoring system for each cell compartment (membrane, cytoplasm and nucleus), where 0 = protein was absent from the compartment; 1 = <30% of cells expressed the protein in this compartment; 2 = 30–70% of cells express the protein in this compartment; and 3 = >70% of cells expressed the protein in the compartment of interest.

Transmission electron microscopy and immunogold detection of claudins
For the ultrastructural determination of healthy or tumoral bladder tissue structures, tissue samples were fixed overnight at 4°C in a solution of 3.5% glutaraldehyde in PBS (0.1 M, pH 7.4). After rinsing in PBS, post-fixation was performed for 1 h at room temperature in 2% osmic acid. After two rinses in PBS, tissues were dehydrated using a gradient of increasing ethanol solutions (30–100%), then embedded in Spurr. Thin sections (85 nm, Leica–Reichert Ultracut E) were collected at different levels of each block, and counterstained with uranyl acetate (1.5% in 70% ethanol) and lead citrate, and observed using a Hitachi 7100 transmission electron microscope.

For immunogold detection of claudins, tissues were fixed overnight at 4°C in PBS containing 3% paraformaldehyde/0.1% glutaraldehyde. Dehydration was performed using two consecutive 15 min incubations in 30 and 50% ethanol solutions, followed by 70% ethanol solution. Tissues were then incubated in 1.5% uranyl acetate in 70% ethanol for 30 min at room temperature, followed by three consecutive 15 min periods in 80% followed by 95% ethanol. After embedding in LR White resin, 85 nm sections were collected using an ultramicrotome on Nickel grids.

Sections were then incubated for 2 h at room temperature in PBS + 0.5% freshwater fish gelatin + 1% BSA (‘incubation buffer’), then overnight at 4°C with the relevant primary antibody (claudin-1 or 4, dilution 1/20; claudin 7, dilution 1/10). After 5 x 12 min washes, sections were incubated for 50 min at room temperature with secondary antibodies (anti-mouse or anti-rabbit IgG coupled to 15 nm gold beads, dilution 1/20, Aurion, France), washed again (5 x 12 min), incubated for 2 min in 2% glutaraldehyde, and rinsed briefly with filtered milliQ water. (All washes and antibody dilution were made in the incubation buffer described above). Observation was performed using a Hitachi 7100 transmission electron microscope.

Sequencing of CLDN4 coding sequence
The genomic DNA was isolated from the tumor samples of 12 patients using the QIAamp DNA Mini kit (QIAgen). The coding sequence of the CLDN4 gene was PCR amplified using AccuPrime DNA polymerase (Invitrogen) with the following primers: forward primer, 5'-CGGGATCCACCACCATGGGGCTACAGGTAATGGG-3'; reverse primer, 5'-CGAGATCTTACACGTAGTTGCTGGCAGC-3'. PCR conditions were : initial denaturation at 94°C for 4 min, followed by 35 cycles at 94°C for 1 min, 58°C for 30 s, 72°C for 1 min, final elongation was for 10 min at 72°C. The PCR fragments were gel purified and sent to GENOME express (Meylan, France) for automated sequencing using the following two primers for each PCR fragment: Sense sequencing primer, 5'-GCTCTGGGCGTGCTGCTGTC-3'; antisense primer, 5'- CCACCAGCGGATTGTAGAAG-3'. The sequences were checked for the presence of mutations by comparing them with the wild-type CLDN4 coding sequence using the MultAlign software (http://searchlauncher.bcm.tmc.edu/multi-align/multi-align.html).

RNA extraction and mRNA quantitation
Frozen tissue samples were first lysed at 4°C using a Mixer Mill MM 300 in RNeasy Lysis buffer (RLT buffer) (Qiagen, Hilden, Germany), using sterile steel beads. Total RNA was then extracted from tumor or normal samples using the RNeasy Mini kit (Qiagen, Hilden, Germany), and RNA quality was controlled using Agilent nanochips (Agilent Technologies, Palo Alto, USA). mRNAs were reversed transcribed using M-MLV-RT from Invitrogen (Cergy Pontoise, France), and quantitated by RT–QPCR (Roche Applied Science, Meylan, France). Homogeneity of samples was checked by quantitating the mRNA for a housekeeping gene (ß-actin), and the proportion of stroma was quantified analyzing the expression of desmin. Primers for claudin-4 PCR were: (Forward) 5'-CGG GAT CCA CCA CCA TGG GGC TAC AGG TAA TGG G-3', (Reverse) 5'-CCA CCA GCG GAT TGT AGA AG-3'; ß-actin primers were: (F) 5'-CGG GAA ATC TGC GTG ACA T-3', (R) 5'-AAG GAA GGC TGG AAG AGT GC-3'; and desmin primers were: (F) 5 -CAT CGC GGC TAA GAA CAT TT-3', (R) 5'-GCC TCA TCA GGG AAT CGT TA-3'.

Detection of CLDN-4 gene methylation
The CLDN-4 genomic sequence was searched for the presence of CpG islands using the CpG Island Searcher (http://www.cpgislands.com) (27). Genomic DNA was extracted from tissues or primary tumor cells using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany) and 0.5–1 µg of genomic DNA was modified with sodium bisulfite using the CpGenome DNA Modification kit (Chemicon, Temecula, USA). Two methylation-specific PCR (MSP) reactions were then performed on each sodium bisulfite-treated genomic DNA sample, using primers specific for the methylated [5'-TTATGGGGTTATAGGTAATGGGTATC-3' (MF), 5'-ATAAACCCTCCCAAATAATCTAGC-3' (MR)), or the unmethylated DNA (5'-TATGGGGTTATAGGTAATGGGTATT-3' (UMF), 5'-TAAACCCTCCCAAATAATCTACAAA -3' (UMR)]. Primers were designed using MethPrimer (http://itsa.ucsf.edu/~urolab/methprimer) (28). MSP conditions were: 95°C for 5 min, followed by addition of 2.5 U GoldStar DNA polymerase (Eurogentec, Seraing, Belgium); 40 cycles of 95°C/30 s, 55°C/30 s, 72°C/30 s; final extension at 72°C/4 min. The PCR products were then run on ethidium bromide-stained agarose gels.

Mouse tissues and histochemistry
DNA methyl transferase 3a (Dnmt3a) overexpressing mice were generated by a knock-in gene targeting method to introduce a cDNA fragment encoding the full-length Dnmt3a protein into the endogenous A33 antigen locus, in view of the tissue-specific nature of this gene. The A33 antigen is a transmembrane glycoprotein prominently expressed in adults by intestinal epithelial cells, as well as in the bladder urothelium (29,30). In the mice used here, transcription at the targeted locus produces a bi-cistronic mRNA, which encodes both A33 antigen and Dnmt3a. Translation of the Dnmt3a cistron is initiated at an internal ribosome entry site (IRES) (Samuel, M.S. and Ernst, M., Manuscript in preparation). The resulting mice overexpress Dnmt3a within the intestinal and bladder epithelium. After bladder dissection, paraffin embedding, tissue sectioning and hematoxylin/eosin staining were performed as described in ref. 31. Following dewaxing and hydration, paraffin tissue sections were pretreated with peroxidase for 20 min at room temperature. Antigen was retrieved by boiling samples for 20 min in 10 mM Tris–1 mM EDTA, pH 9. Slides were allowed to cool down to room temperature, and incubated overnight at 4°C with antibodies in PBS + 0.05% BSA. In all cases, the Envision+ kit (DAKO) was used as a secondary reagent. Stainings were developed using DAB (Sigma) and slides were counterstained with hematoxylin and mounted.

Statistical analysis
Optimal exposure times of membranes were used for each patient, and protein expression was quantified using NIH Image 1.62, and adjusted for background noise and protein loading. The ratio of expression in tumor versus healthy samples was calculated for each patient from the results of three experiments. For each patient, a one-sample t-test was used to determine whether the mean ratio was significantly different from the theoretical value of 1 (µ) (which reflects a similar expression in tumor and healthy samples). The formulation of this test was: Formula (X bar represents the mean ratio, S the standard error and n the number of replicates for each sample, n = 3). The ratio was considered as significantly different from 1 when P < 0.05. Associations between claudin expression or localization and clinical or pathological data were analyzed using Fischer's exact test, using SAS 9.1 for Windows (Cary, NC, USA). Student's t-test was used in order to determine the statistical significance of differences between primary cell cultures transfected or not with claudin-4 siRNA.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 Supplementary material
 References
 
Alterations of CLDN protein expression in bladder tumor samples
Post-transcriptional and post-translational events are essential to the regulation of claudin expression and functions (32,33). Here, we first quantified the expression of claudins 1, 4 and 7 proteins in whole lysates from healthy urothelium and bladder tumor samples of 39 patients using western blotting. Significant variations of claudin expression were detected between control and tumor samples (Figure 1A), independent of variations in the proportion of urothelium and stroma within each sample, which were assessed by probing the same membranes with antibodies against SMA and desmin (Figure 1A). Overall, claudins 1, 4 and 7 expression results were found to match well with the staining intensity for these proteins in the immunohistochemical analysis of control and tumoral tissue sections (compare for example Patient 6 on Figures 1A, 6a–c and 6g–i).


Figure 1
View larger version (21K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1 Variations of claudins 1, 4 and 7 expression in human bladder tumors in comparison with control urothelium from the same patients. (A) Western blot analysis of claudins 1 and 4 expression in healthy (H) and tumor (T) samples from six patients with bladder carcinoma. Actin expression was quantified to standardize protein loading, and SMA and desmin expressions were used to assess the proportion of stroma versus urothelium. (B) Averaged ratios and SEM of claudins 1, 4 and 7 expression in tumor/control sample from 39 patients, in relation with tumor stage and grade. Expression levels and statistical significance were determined as described in Materials and methods (Figure 1: overexpression in tumor; Figure 1: downregulation in tumor; {leftrightarrow}: expression not significantly different from control). (C) Table representing P-values of statistical analysis between the results of claudins 1, 4 and 7 expression, and tumor pathology or one-year survival of patients (Fischer's exact test, n = 39).

 
Claudin-4 expression was significantly modified in 26 out of 39 tumor samples (Figure 1B), with a significant trend towards overexpression in superficial and low/medium grade (G1/G2) tumors, and a pronounced downregulation in ~60% of invasive and/or high-grade (G3) tumors (Figure 1B). We then tried to assess whether the apparent reduction of claudin-4 expression in invasive and high-grade tumors could be due to the presence of a truncating mutation within the coding sequence of its gene, which could make it undetectable by currently available claudin-4 antibodies, since they are all directed against the C-terminal end of the protein. Upon sequencing, no mutations were found within the coding sequence or the 1200 bp upstream of the single exon of the CLDN-4 gene, in genomic DNA extracted from control samples and tumors with downregulated claudin-4 expression (data not shown). Finally, no lower molecular weight bands were ever detected with the claudin-4 antibody during western blotting experiments, arguing against a tumor-specific degradation of claudin-4.

The expression pattern of claudins 1 and 7 were also found to be modified in bladder tumors, albeit in fewer patient samples than in the case of claudin-4 (15/39 for claudin-1, 12/39 for claudin-7; Figure 1B), and without correlation with the pathological status of the tumors (Figure 1C).

Finally, variations of claudin-4 expression were tightly associated with the tumor stage (P < 0.001) and grade (P = 0.029), and with the 1-year survival of patients (P = 0.040) (Figure 1C). Indeed, of the nine patients who deceased within 1 year of surgery, seven displayed downregulated claudin-4 expression within their bladder tumor and none of these tumors overexpressed the protein. Finally, no other associations were found between the variation of claudin 1 or 4 expression or localization and the clinical data, including tumor recurrence or metastasis within 1 year of surgery (data not shown).

Alteration of claudin-4 mRNA expression in bladder tumor samples
We initially sought to determine whether the alterations of claudin-4 expression detected in bladder carcinoma could be due to the mutations within the promoter region or the coding sequence of the CLDN-4 gene. We therefore amplified by PCR a genomic DNA region starting 1 kb upstream of the transcription initiation site and including the coding sequence of this single-exon gene. Sequencing of these regions was performed from the genomic DNA of 12 patients with bladder carcinoma (7 displaying claudin-4 downregulation and 5 with overexpression of the protein), but sequence alignment showed that no mutations were present in any of the samples analyzed (data not shown). We then quantified the expression of claudin-4 mRNA from healthy and tumor bladder samples where enough remaining material was available, in order to confirm whether claudin-4 downregulation was due to a transcriptional mechanism. Total RNA was extracted from nine patients' samples, and expression of claudin-4 mRNA was quantified using Real-Time Quantitative PCR. Our results demonstrate that expression levels of claudin-4 protein (Figure 2A) and mRNA (Figure 2B) varied similarly between tumor and healthy samples, suggesting that regulation of claudin-4 in bladder tumor samples occurs at the transcriptional level. Quantification of desmin mRNA confirmed that the modifications of claudin-4 mRNA levels were not due to variations of urothelium/mesenchyme proportion (data not shown). In most patients, we found that the amplitude of variations in claudin-4 mRNA expression between the tumor and the control samples was different from those detected at protein level (Patients 19, 22, 33, 35 and 39 in Figure 2C). This could indicate that claudin-4 expression is further regulated at the post-translational level. Furthermore, claudin-4 protein and mRNA levels varied between control samples (compare samples 19H, 39H, 12H and 22H in Figure 2A and B), which could be partly due to the potentially different degree of inflammation between control samples of various patients, generally collected close to the tumor area. In any case, for each individual patient (except in the case of Patient 33), all variations of claudin-4 mRNA expression between tumor and healthy samples were found to corroborate variations detected at the protein level (Figure 2C).


Figure 2
View larger version (20K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2 Claudin-4 expression is regulated at the transcriptional level in human bladder tumors. (A) Protein expression of claudin-4 and actin in healthy (H) and tumor (T) samples derived from the bladder of four patients with bladder carcinoma. (B) Histogram summarizing the relative expression of claudin-4 mRNA in the healthy or tumor samples from the same four patients. Data in A and B represent results from single experiments, each one representative of three similar experiments. (C) Table summarizing the level of claudin-4 protein expression in relation to claudin-4 mRNA expression for the nine patients where sufficient amounts of clean RNA were obtained from both control and tumor samples. Results are expressed as averaged ratios of expression in the tumor/expression in the control sample, for both protein and mRNA levels (n = 3).

 
Inhibition of methyl transferases restores claudin-4 expression in primary cultures of human bladder tumor cells
We decided to assess whether CLDN-4 gene promoter hypermethylation could be responsible for the downregulation of claudin-4 expression in invasive and high-grade bladder tumors, similar to what was recently demonstrated in the case of claudin-7 expression in high-grade breast carcinomas (22). While the 4 kb genomic DNA sequence covering the putative promoter and the 5' non-coding region of the CLDN-4 gene lack CpG islands, we identified a CpG island starting 227nt upstream of, and stretching along, the entire coding sequence of this gene (Figure 3A). Based on the stringent criteria for CpG islands put forward by Takai and Jones (34), this region is likely to constitute a functional CpG island.


Figure 3
View larger version (15K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3 Hypermethylation of the CLDN-4 gene is responsible for claudin-4 downregulation in bladder tumor cell primary cultures. (A) Localization of the CpG island and position of primers specific for the methylated (MF: methylated forward, MR: methylated reverse) or unmethylated (UMF: unmethylated forward, UMR: unmethylated reverse) sequence on the CLDN-4 gene. (B) Cytokeratin immunostaining of three independent primary tumor cell cultures (PC): PC-A, PC-B and PC-C. (C) Western blot analysis of claudin-4 and actin expression, and (D) MSP analysis of the CLDN-4 gene methylation status [amplification with primers specific for unmethylated (UM) or methylated (M) sequence], in the same three primary cultures prepared from bladder tumors, without (–) or with (+) treatment with 10 µM 5-aza-CdR. (E) Table summarizing the expression profile of claudin-4 [not expressed (–) or expressed (+)], as well as the methylation status of the CLDN-4 gene [unmethylated (–) or methylated (+)], in five primary cultures (PC-A–PC-E) prepared from urothelial bladder carcinoma, with or without treatment with 10 µM 5-aza-CdR.

 
In order to assess whether increased methylation could be responsible for the downregulation of claudin-4 expression in invasive and high-grade tumors, we first decided to analyze directly the impact of CLDN-4 gene hypermethylation in vitro on primary cell cultures prepared from five different human bladder carcinomas. All cells within the cultures displayed positive staining for cytokeratins, confirming their epithelial nature and the homogeneity of the cell populations used (Figure 3B). Protein lysates and genomic DNA were prepared from these cells, and we analyzed claudin-4 protein expression as well as the methylation of a region located within the CpG island (+7 to +151 in the CLDN-4 exon), using MSP. Three of the five primary cell cultures (PC-B, C, and D) did not express detectable levels of claudin-4, and their genomic DNA was found to contain almost exclusively hypermethylated CLDN-4. Strikingly, a 4-day treatment with the demethylating agent 5-aza-CdR (10 µM) induced de novo expression of claudin-4 in these three primary cultures (for example PC-B and C in Figure 3C and D). The two other primary cell cultures (PC-A and E) did express high levels of claudin-4, and the CpG island within their CLDN-4 gene was found to be completely unmethylated before and after treatment with 5-aza-CdR (Figure 3E). These results suggest that, once the urothelial content within bladder tumors has been isolated through primary culture, there is a causal relationship in tumor cells between the methylation of the CLDN-4 gene coding sequence and the downregulation of claudin-4 expression.

Hypermethylation of the CLDN-4 gene in vivo downregulates claudin-4 expression in the urothelial bladder of mice overexpressing DNA methyl transferase 3a
We further examined whether the causal relationship between increased methylation of the CLDN-4 gene and reduced expression of the protein was relevant in vivo, and whether it could have a role to play in the physiological regulation of this gene outside of a tumor context. In order to do so, we made use of a mouse knock-in model expressing DNA methyl transferase 3a (Dnmt3a), originally generated in order to study the impact of methylation processes on the intestinal epithelium (Samuel,M.S. and Ernst,M., Manuscript in preparation). Indeed, in these A33D3a knock-in mice, Dnmt3a is expressed from the gene locus of the endogenous A33 antigen, a transmembrane protein specifically expressed by epithelial cells not only in the intestine, but also in the bladder (29).

This de novo DNA methyl transferase is highly expressed during embryogenesis and has been previously shown to be overexpressed in bladder tumors (35). Recent studies have shown that Dnmt3a is sensitive to 5-aza-CdR (36), and its activity can be detected in bladder tumor cells, independently from that of the more broadly expressed Dnmt1 isoform (37). We therefore examined whether overexpression of this DNA methyl transferase in the bladder epithelium was capable of affecting the expression of claudin-4 in vivo.

Although the overall morphology of the bladder urothelium was conserved in A33D3a mice, umbrella cells appeared less polarized and differentiated than in control mice, as assessed by hematoxylin/eosin staining (Figure 4A). These cells displayed characteristically large cytoplasms in the bladder of A33wt mice, but their overall size was markedly reduced in A33D3a mice, where they are harder to distinguish from the underlying cell layers (Figure 4A). Maximal expression of the transgene was detected in the uppermost layer of the bladder epithelium in A33D3a mice, corresponding to the umbrella cell layer (Figure 4B). In the bladder of control A33wt mice, regular claudin-4 expression was detected at the TJ of umbrella cells, as well as around the cell membrane of underlying cell layers (Figure 4C), as described previously (8). In A33D3a mice, although claudin-4 expression appeared normal in basal and intermediate cells, the protein was completely undetectable in umbrella cells, where expression of Dnmt3a was the highest (Figure 4C). This effect was not due to a general downregulation of cell/cell junctions between neighboring umbrella cells, since expression of occludin was still detected at TJ in the bladder of A33D3a mice. The localization of other junctional proteins such as claudin-7 and E-cadherin was slightly altered in mice overexpressing Dnmt3a but, unlike claudin-4, they were still expressed by umbrella cells in A33D3a mice (Figure 4C).


Figure 4
View larger version (77K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4 Downregulation of claudin-4 expression in vivo in bladder cells overexpressing Dnmt3a. (A) Hematoxylin/eosin staining of the bladder urothelium from A33wt control mice (A1) and mice overexpressing DNA methyl transferase 3a in the A33 gene locus (A33D3a, A2). Arrows point to umbrella cells. (B) Expression of Dnmt3a (arrowheads) in the bladder urothelium of A33D3a mice but not in A33wt controls. (C) Immunofluorescent detection of claudins 4, 7, Occludin and E-cadherin in serial sections of bladder urothelium from A33wt (top) and A33D3a (bottom) mice. Claudin-4 and occludin staining are completely restricted to the TJ of umbrella cells in A33wt mice (white arrows). In A33D3a mice, occludin staining remains unchanged, while claudin-4 expression is undetectable (arrowheads). Claudin-7 and E-cadherin staining were found all along the basolateral membrane in both control and A33D3a mice. No reproducible staining could be obtained in mouse tissues with the polyclonal claudin-1 antibodies in our possession (71-7800 and 51-9000 from Zymed). Photographs depict representative fields from the bladder of three separate mice per group. Bars represent 100 µm (A and B) or 20 µm (C).

 
The methylation profile of the CLDN-4 gene coding sequence is directly correlated to the expression of the protein in human bladder tumors
Results presented above indicated that increased methylation within the coding sequence of the CLDN-4 gene induced a downregulation of this protein in primary cultures of human bladder tumor cells and in transgenic mice overexpressing Dnmt3a. In order to assess whether this type of regulation could explain the pronounced downregulation of claudin-4 expression in human bladder tumors, we analyzed the proportion of methylated and unmethylated CLDN-4 gene sequence in healthy and tumoral bladder samples from 20 patients, in correlation with their respective expression of the protein.

In genomic DNA extracted from healthy samples, this region was amplified both with primers specific for the unmethylated DNA and with primers targeting the methylated sequence (Figure 5A). Interestingly, the level of amplification with methylation-specific primers was inversely proportional to levels of claudin-4 protein expression in healthy samples (see samples 19H, 39H, 12H and 22H in Figure 5A), suggesting that variations of methylation levels could play a role in the physiological regulation of claudin-4 expression. Similarly, the relative level of amplification with methylation-specific primers was decreased in the few tumor samples where claudin-4 expression was enhanced (e.g. Patient 39), indicating that modulation of CLDN-4 gene methylation could operate as a general mechanism to regulate the expression of this protein. In agreement with this hypothesis, a low level of methylation within the CLDN-4 gene was detected in claudin-4-expressing urothelial cells isolated from healthy bladder urothelium (Figure 5B).


Figure 5
View larger version (20K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5 Correlation between the methylation profile of the CLDN-4 gene and claudin-4 expression in human bladder tumors. (A) Top panel, claudin-4 protein expression in healthy bladder (H) and tumor (T) samples from four patients with bladder carcinoma (similar to Figure 2A). Lower panel, methylation profile of the CLDN-4 gene in the genomic DNA of healthy and tumoral bladder samples from the same four patients. The PCR was performed using primers specific for the unmethylated (UM) or methylated (M) sequence. (B) Left, bright field photograph and cytokeratin staining of a representative primary culture prepared from healthy human urothelium. Right, expression of claudin-4 and actin, and methylation profile of the CLDN-4 gene, in the same primary cell culture. (C) Table summarizing the level of claudin-4 protein expression, and the relative amplification with primers specific for the unmethylated (UM) or the methylated (M) sequence of the CLDN-4 gene, in 20 tumor samples (–: no amplification; +: weak amplification; ++: maximum amplification).

 
In bladder tumors, the CLDN-4 genomic DNA region was mostly amplified with methylation-specific primers in samples where claudin-4 expression was found to be downregulated (samples 19T, 12T and 22T, Figure 5A), similarly to what was found in primary cultures. Analyses performed on samples coming from 20 patients (Figure 5C) demonstrated a tight association between claudin-4 downregulation and increased CLDN-4 gene amplification with methylation-specific primers. In contrast, such an increase in methylation was not detected in tumors with normal levels of claudin-4, while unmethylated DNA was predominant in the CLDN-4 gene within most tumors overexpressing the protein (e.g. Patient 39 in Figure 5A). These results strongly suggested that the methylation level of the CLDN-4 gene coding sequence is correlated to the expression of the protein in the healthy human bladder and in human bladder tumors.

Subcellular distribution of claudins within healthy and tumoral bladder urothelium
In order to establish whether claudins 1, 4 and 7 were still targeted to TJs in human bladder tumors, we assessed the localization of these proteins in a panel of tumor samples in comparison with healthy bladder urothelium. The three claudin isoforms were detected at the TJ, as well as along the basolateral membrane of most cell layers across the urothelium of all macroscopically healthy bladder samples (Figure 6a–f), as already described in several species (8–10). Claudin-1 immunostaining was stronger in the intermediate and basal urothelial cell layers (Figure 6a and d), while claudin-4 was prominently expressed in umbrella cells (Figure 6b and e). Claudin-7 was found to be distributed throughout the bladder epithelium (Figure 6c and f). No significant staining for these claudins was found within the cytoplasm of cells in healthy bladder samples. The presence of claudins 1, 4 and 7 at the membrane of umbrella cells, but also of intermediate and basal cells, was confirmed using immunogold staining followed by electron microscopy. In umbrella cells, claudins 1 and 4 were expressed along the lateral membrane, including at TJs, while claudin-7 was detected predominantly along the lower lateral membrane and the basal membrane (Supplementary Figure S1).


Figure 6
View larger version (73K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6 Immunohistochemical analysis of claudins 1 and 4 localization in healthy bladder and tumor urothelium. Immunostaining for claudins 1 (green) and 4 (red) on transverse (a–c) and longitudinal sections (d–f) from healthy urothelium, or from tumor urothelium at various stages and grades (g–u). White arrows indicate membrane staining, black arrows nuclear staining and grey arrows point to speckled cytoplasmic staining in tumor samples. White arrowheads point to areas of irregular staining for claudin-7 in G3 tumors, and grey arrowheads indicate regular claudin-7 staining at the tumor/stroma interface. Bars indicate 40 or 12 µm (inset).

 
In contrast, a loss of membrane localization of claudins 1 and 4 was detected in virtually all bladder tumor samples, albeit to different degrees. Disappearance from the cell membrane was found to be complete in a large proportion of tumors, while residual membrane staining for claudins 1 and 4 was detected in 54 and 38% of tumor samples, respectively (white arrows in Figure 6j, k, m and q), with a decreasing proportion of cells displaying membrane claudin expression in invasive and high grade tumors. In addition, cytoplasmic immunoreactivity for claudins 1 and 4 was detected in all but three bladder tumor samples, including most superficial and low grade tumors (Figure 6g, h, j, k, n, p, q and s). Cytoplasmic staining was prominant in tumors displaying increased staining intensity when compared with the matching control tissue (Figure 6g and h compared with a and b). However, weak cytoplasmic staining was still detected in tumors where the overall staining intensity was decreased (Figure 6n and t). Finally, the cytoplasmic staining for claudins 1 and 4 was often found to have a speckled pattern (grey arrows in Figure 6n, s and t), but we were not able to detect any co-localization between these proteins and endosomial and Golgi markers (data not shown). In contrast, fewer disruptions were detected in the distribution pattern of claudin-7. The protein was still localized at the membrane in all the tumor samples analyzed, although the membrane staining appeared speckled in high-grade tumors (cf. white arrowheads in Figure 6o and u), reflecting an irregular and fragmented expression of claudin-7. Yet, continuous staining was generally detected in areas of contact between the tumors and the surrounding stroma (grey arrowheads, Figure 6r).

Overall, these results imply that claudins 1 and 4 targeting is disrupted in virtually all tumor specimens and suggest that important alterations of their localization represent a hallmark of bladder carcinogenesis.

In the process of studying the impact of hypermethylation on claudin-4 expression (see above), we had attempted to study the effect of a 5-aza-CdR treatment on the human bladder cell line HT-1376, originally developed from a grade-3 tumor. This approach did not prove useful to study claudin-4 expression since untreated HT-1376 cells already expressed high levels of claudin-4 (not shown). However, confocal analysis indicated that this protein was mislocalized away from the TJ, into the cytoplasm of confluent HT-1376 cells (Figure 7A, top panel). Treatment of these cells with 5-aza-CdR for >48 h resulted in the restoration of proper claudin-4 targeting to the TJ, correlating with an increased polarization of the cell monolayer (Figure 7A, bottom panel). Claudin-4 expression was unaffected under these conditions, somewhat unsurprisingly since the CLDN-4 gene was already unmethylated in HT-1376 cells prior to 5-aza-CdR treatment (see Figure 8B). A similar result was observed when cells were stained with a claudin-1-specific antibody (Figure 7A, bottom panel). This result indicated that hypermethylation regulates the expression of one or several genes encoding essential proteins for claudins 1 and 4 targeting in HT-1376 cells, suggesting that a similar mechanism could be responsible for their mislocalization in human bladder tumors. Since the establishment of adherens junctions is known to facilitate TJ protein recruitment, and since methylation-induced E-cadherin downregulation has been described previously in bladder tumors (38), we analyzed E-cadherin localization in HT-1376 cells. We found that E-cadherin was already present at the plasma membrane of HT-1376 cells without treatment with 5-aza-CdR, suggesting that, in HT-1376 cells at least, E-cadherin expression is not downregulated by hypermethylation and is properly targeted to the membrane (Figure 7B). In addition, this methylation-driven regulation of claudin localization did not reflect a general effect on TJ proteins, since occludin (Figure 7B) and ZO-1 (not shown) were properly addressed to the membrane in untreated HT-1376 cells.


Figure 7
View larger version (63K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 7 Hypermethylation is responsible for claudins 1and 4 mislocalization in a bladder tumor cell line. (A) Immunostaining for claudins 1 and 4 on XY and XZ confocal sections of HT-1376 cells, without or with treatment with 5-aza-CdR, as indicated. (B) Immunostaining for E-cadherin and occludin on XY and XZ confocal sections of untreated HT-1376 cells. Bar indicates 50 µm.

 
Functional consequences of claudin-4 downregulation in bladder tumors
We then sought to determine the functional consequences of claudin-4 mislocalization and/or downregulation in bladder tumors. We initially tried to grow primary bladder tumor cell cultures on suspended wells or in semi-solid medium in order to study their capacity to form a functional paracellular barrier and to assess their tumorigenicity in vitro. Since our attempts to reproducibly grow these cells under such conditions were unsuccessful, we analyzed the biological impact of claudin-4 mislocalization in the human bladder tumor cell line HT-1376. In these cells, induction of claudin-4 targeting to TJs after treatment with 5-aza-CdR occured concomitantly with a significant increase in TER (Figure 8A), indicating that the paracellular barrier was reinforced under these conditions. We analyzed the functional role of claudin-4 in this pathway by preventing the restoration of its proper targeting to the TJ induced by 5-aza-CdR treatment. To this effect, HT-1376 cells were transiently transfected with siRNA directed against claudin-4 mRNA, during treatment of the cells with 5-aza-CdR. The high expression level of claudin-4 detected by western blotting in HT-1376 cells was not significantly affected by 5-aza-CdR treatment, but strong downregulation of the protein was induced by transfection with a selective claudin-4 siRNA (Figure 8B). The claudin-4 gene was unmethylated under all conditions, in agreement with the high expression of the protein. Immunofluorescent detection confirmed that claudin-4 levels were reduced in siRNA transfected cells (Figure 8D). Levels of ZO-1 remained unchanged and the protein was still targeted to TJs, although the proportion of cytoplasmic ZO-1 appeared slightly higher in cells transfected with claudin-4 siRNA (Figure 8C). Furthermore, we found that the downregulation of claudin-4 prevented the increased polarisation of HT-1376 cells normally seen after 5-aza-CdR treatment (see XZ representations in Figures 7A and 8C). It is noteworthy that pronounced modifications of cellular size and polarization were also detected in umbrella cells of A33D3a mice, where claudin-4 expression, but not that of claudin-7 or occludin, was downregulated (see Figure 4A and C). In addition, our results demonstrate that incubation with claudin-4 siRNA largely prevented the increase in TER normally observed in 5-aza-CdR-treated HT-1376 cells, while a control siRNA had no effect (Figure 8C).


Figure 8
View larger version (65K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 8 Loss of claudin-4 at the TJ is responsible for decreased polarization and TER in HT-1376 cells. (A) Time-course of TER level in HT-1376 cells with or without treatment with 10 µM 5-aza-CdR. (B) Claudin-4 expression and CLDN-4 gene methylation profile in HT-1376 cells treated (plus) or not (–) with 5-aza-CdR, with or without transfection with control (CT) or claudin-4-specific siRNA, as indicated. Histogram represents the mean ± SEM from three similar experiments, after correction for protein loading using actin expression. (C) Immunostaining for claudin-4 (green) and ZO-1 (red) on XY and XZ confocal sections in HT-1376 cells, treated with 5-aza-CdR and transfected with a control siRNA (CT) or with a claudin-4-specific siRNA. (D) Quantification of the TER in untreated HT-1376 cells (black bars) and cells treated with 10 µM 5-aza-CdR for several days as indicated, and transfected with control (CT) siRNA (white bars) or with claudin-4-specific (C4) siRNA (grey bars). *P < 0.05, Student's t-test.

 
In order to gain insight into the functional consequences of claudin alterations in human bladder tumors in situ, we used transmission electron microscopy to analyze whether the structural integrity of TJ complexes was maintained in samples of early stage/grade and advanced bladder tumors, in comparison with healthy bladder urothelium. Regular alignments of desmosomes, adherens junctions and TJs were detected along umbrella cell/cell contacts in the healthy bladder. In contrast, TJ structures were undetectable in several areas of superficial and low/medium grade tumors, whereas desmosomes and adherens junctions appeared completely similar to those found in the healthy urothelium. In high-grade and invasive carcinomas, cell/cell contacts were markedly altered. TJ complexes were virtually absent, and the number of desmosomes and adherens junctions was greatly reduced (Supplementary Figure S2 online). These results corroborate the fact that the function of the TJ barrier is affected from early stages of tumorigenesis in the urothelial bladder.


   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
Supplementary material
 References
 
The first important finding of this study concerned the significant changes of claudin-4 expression detected in a large proportion of the tumor samples analyzed, while the expressions of claudins 1 and 7 were modified in a much smaller number of samples. In addition, pronounced mislocalizations of claudins 1 and 4 were detected in a majority of bladder tumors, prompting us to analyze the mechanism underlying the disruptions of the claudin-4 expression pattern and to assess the functional consequences of claudin-4 alterations in bladder tumor cells.

Modifications of claudin-4 expression enabled discrimination between superficial and low-grade bladder tumors, where the protein was mostly overexpressed, from invasive and high-grade tumors, where it was significantly downregulated. In addition, this expression pattern of claudin-4 was significantly associated with 1 year post-surgery patient survival, but no correlation was found at this stage between claudin-4 expression and tumor recurrence or appearance of metastases. The nature of such correlation is still a matter of debate, since claudin-4 expression was shown to decrease the invasiveness and metastatic potential of pancreatic cancer (39), whereas overexpression of claudins 3 and 4 promoted ovarian tumor metastasis through increased invasion and survival (40). Larger patient numbers and a longer follow-up should help to properly assess the evolution of low-stage and low-grade tumors, and to conclude whether claudin-4 has a role to play as a prognostic marker for bladder cancer, similar to that of claudin-1 in stage II colorectal carcinomas (20).

The pattern of claudin-4 expression in bladder carcinomas identified here is markedly different from that observed in pancreatic and ovarian cancers, where overexpression was the prominent feature (14,15,41). Early stage bladder tumors were found in the present study to overexpress claudin-4, suggesting that these tumors could be interesting targets for treatment with the CPE. CPE binds selectively to claudins 3 and 4, thereby mediating cytolysis of cancer cells (16,17), and the potential of this therapeutic approach was recently vindicated in vivo on chemotherapy-resistant mammary tumor xenografts (18). However, we found that all bladder tumors overexpressing claudin-4 display very little membrane staining for this protein, implying that the protein is probably largely inaccessible to the toxin and thereby limiting the potential use of CPE for treatment of superficial bladder cancer.

The sharp downregulation of claudin-4, detected only in invasive and high grade tumors, is consistent with the recent demonstration that claudin-4 expression decreases the invasive potential of pancreatic cancer cells (39). Downregulation of other claudins (2,10,41) was detected in gastric or breast cancers, and does not seem to involve genetic alterations (22,42,43). Similarly, our results showed the absence of mutations within the coding sequence of the CLDN-4 gene in all tumors analyzed. In view of the well-documented effects of promoter hypermethylation on gene silencing in cancer (44), and since hypermethylation of the promoter region was previously found to induce a downregulation of the CLDN-7 gene in breast carcinoma (22), we sought to determine whether a methylation process could be responsible for the variations of claudin-4 expression in bladder tumors. While no CpG island was detected within the promoter region of the CLDN-4 gene, a large CpG island was found to span the whole coding sequence of this single-exon gene. Despite their frequent occurrence (34), information concerning the role of CpG islands within the coding sequence of genes is scarce. Hypermethylation of the first exon has been linked to a downregulation of the gene encoding the estrogen receptor-alpha in BRCA1-linked breast cancers (45). Here, we show that variations in the proportions of methylated DNA within the CLDN-4 coding sequence were inversely correlated to the expression level of the protein in bladder tumors, as well as in the healthy urothelium. We found an association in tumors between hypermethylation of the CLDN-4 coding sequence and claudin-4 downregulation, and variations of methylation levels induced in human tumor cells in culture and in Dnmt3a overexpressing mice induced significant modulation of claudin-4 expression. In particular, downregulation of claudin-4 expression in umbrella cells, induced by overexpression of Dnmt3a in vivo, further argues in favor of the fact that the CLDN-4 gene can indeed be regulated through methylation of its coding sequence, even outside of a tumor context. In addition, partial methylation of the CLDN-4 gene was detected in healthy urothelial mucosa and in primary cultures derived thereof, and the relative amount of methylated CLDN-4 in these samples was also inversely correlated with their respective level of claudin-4 expression. We therefore propose that methylation of the CLDN-4 gene coding sequence is highly likely to play a role in the physiological regulation of claudin-4 expression. Such regulation could be involved in the modulation of claudin-4 expression during the differentiation of urothelial cells, since the different bladder cell layers do not seem to express similar levels of claudin-4. Although we cannot completely rule out that the presence within the samples of methylated CLDN-4 DNA originates from stromal tissue contamination, it is highly unlikely to play a role in these results. Indeed, genomic DNA samples were obtained from tissue sections directly adjacent to those used for RNA and protein lysates, where the minor variations of stromal tissue observed did not correlate at all with variations of claudin-4 expression. Finally, it is noteworthy that, although fewer modifications of claudin-7 expression were detected, most of them involved a downregulation of the protein. Expression of this protein was previously shown to be downregulated in breast carcinoma, mostly due to gene promoter methylation (22). Results of the present study therefore suggest that a similar modulation could be responsible for the downregulation of claudin-7 expression in bladder tumors. However, the populations of tumors displaying claudins 4 and 7 downregulations were found to overlap only partially, suggesting that the CLDN-4 and CLDN-7 genes could be sensitive to different isoforms of DNA methyl-transferases or, alternatively, that different molecular processes are involved in modulating the sensitivity of these two genes.

In the present work, we also showed that claudins 1 and 4 were mislocalized in 90% of bladder carcinoma samples, including superficial and low-grade tumors. In contrast, the localization of claudin-7 was largely restricted to the membrane in all carcinoma samples, albeit with a fragmented appearance in high-grade tumors. A pattern of overexpression coupled to abnormal targeting of claudin-1 was identified previously in colon carcinoma (46), and partial cytoplasmic localization of claudin-4 was previously described, primarily in ovarian cancer (16,47). Increased internalization of claudins has recently been demonstrated on MDCK cells (48) and, since both claudins analyzed in this work displayed a speckled cytoplasmic staining profile in some tumor samples, we tried to assess whether they could be colocalized with intracellular vesicles. However, claudin-1 or claudin-4 immunoreactivity did not colocalize with early and late endosomes, or with cis- and trans-Golgi markers (data not shown).

The reasons underlying claudin-4 mislocalization remain unclear. A mutation within the PDZ-binding domain of claudin-16 was previously shown to impair its ability to localize at the TJ, resulting in its accumulation in lysosomes (49). However, we did not detect any mutation within the coding sequence of CLDN-4 mRNAs in tumor samples displaying profound mislocalization of the protein. Results obtained on the HT-1376 bladder tumor cell line suggest that hypermethylation could also be involved in this process within bladder tumors, perhaps by controlling the expression of an essential modulator of claudin-4 targeting to the TJ. Although the nature of such a modulator remains to be clarified, this hypothesis would be consistent with previous results demonstrating the role of methylation processes on polarity and differentiation (50).

Since claudins 1 and 4 play an important role in the regulation and the selectivity of epithelial and endothelial permeability (51–53), their mislocalization, although limited to the tumor area, could have important consequences for urothelial bladder homeostasis. In the present study, the essential role of claudin-4 in the maintenance of the paracellular barrier was confirmed in the bladder tumor cell line HT-1376. Furthermore, we observed that downregulation of claudin-4 was correlated with a partial loss of urothelial cell polarization in the same cell line and in the umbrella cells of mice overexpressing Dnmt3a. Since virtually all tumor samples displayed altered localization of claudins 1 and 4, disruptions of the TJ-mediated paracellular permeability barrier are likely to represent a hallmark of early urothelial carcinogenesis. In addition, mislocalization of this protein may induce a transurothelial leakage of ions or macromolecules from the urine, including biologically active growth factors such as epithelial growth factor (EGF) (54).

In summary, this study describes for the first time a complex disruption pattern of claudin-4 expression and localization in urothelial carcinoma. Hypermethylation of the gene coding sequence appears responsible for claudin-4 downregulation in advanced tumors, and our results suggest that variations of methylation may act as a regulatory mechanism for claudin-4 expression in the bladder. Further studies will be necessary to clarify the potential of claudin-4 as a biomarker for the diagnosis of early bladder cancers.


   Supplementary material
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
References
 
Supplementary material is available at: http://www.carcin.oupjournals.org/


   Footnotes
 
{dagger}These authors contributed equally to the present work. Back


   Acknowledgments
 
The authors are greatly indebted to prof. J.P. Bali for his continuous support throughout this project. The authors thank Prof. Pierre Costa for his help in setting up the project, Nicole Lautredou-Auduy and Chantal Cazevieille (Centre Régional d'Imagerie Cellulaire, Montpellier Rio Imaging facility, Montpellier) for their technical assistance and their help in the interpretation of confocal and electron microscopy data, and Dr R. Daniel for his help in the assessment of tumor sample pathology. The authors are also grateful to Dr J.K. Heath for critical reading of the manuscript, and to Drs Serge Roche, Michel Vidal and Carsten Brock for providing various antibodies. This work was supported by the Ardèche, the Loire and the Hérault sections of the Ligue contre le cancer, Association pour la Recherche sur le Cancer (ARC), INSERM (CRES no. 4CR04G), GEFLUC, and PHRC CHU Nîmes (no. 7035).

Conflict of Interest Statement: None declared.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Supplementary material
 References
 

  1. Matter K. and Balda M.S. (2003) Signalling to and from tight junctions. Nat. Rev. Mol. Cell Biol. 4:225–236.[CrossRef][Web of Science][Medline]
  2. Schneeberger E.E. and Lynch R.D. (2004) The tight junction: a multifunctional complex. Am. J. Physiol. Cell Physiol. 286:C1213–C1228.[Abstract/Free Full Text]
  3. Kohler K. and Zahraoui A. (2005) Tight junction: a co-ordinator of cell signalling and membrane trafficking. Biol. Cell 97:659–665.[CrossRef][Web of Science][Medline]
  4. Cereijido M., Shoshani L., Contreras R.G. (2000) Molecular physiology and pathophysiology of tight junctions. I. Biogenesis of tight junctions and epithelial polarity. Am. J. Physiol. Gastrointest. Liver Physiol. 279:G477–G482.[Abstract/Free Full Text]
  5. Tsukita S. and Furuse M. (2000) Pores in the wall: claudins constitute tight junction strands containing aqueous pores. J. Cell. Biol. 149:13–16.[Abstract/Free Full Text]
  6. Blanco D., Vicent S., Elizegi E., Pino I., Fraga M.F., Esteller M., Saffiotti U., Lecanda F., Montuenga L.M. (2004) Altered expression of adhesion molecules and epithelial-mesenchymal transition in silica-induced rat lung carcinogenesis. Lab. Invest. 84:999–1012.[CrossRef][Web of Science][Medline]
  7. Haynes M.D., Martin T.A., Jenkins S.A., Kynaston H.G., Matthews P.N., Jiang W.G. (2005) Tight junctions and bladder cancer (review). Int. J. Mol. Med. 16:3–9.[Web of Science][Medline]
  8. Acharya P., Beckel J., Ruiz W.G., Wang E., Rojas R., Birder L., Apodaca G. (2004) Distribution of the tight junction proteins ZO-1, occludin, and claudin-4, -8, and -12 in bladder epithelium. Am. J. Physiol. Renal. Physiol. 287:F305–F318.[Abstract/Free Full Text]
  9. Langbein L., Grund C., Kuhn C., Praetzel S., Kartenbeck J., Brandner J.M., Moll I., Franke W.W. (2002) Tight junctions and compositionally related junctional structures in mammalian stratified epithelia and cell cultures derived therefrom. Eur. J. Cell Biol. 81:419–435.[CrossRef][Web of Science][Medline]
  10. Mysorekar I.U., Mulvey M.A., Hultgren S.J., Gordon J.I. (2002) Molecular regulation of urothelial renewal and host defenses during infection with uropathogenic Escherichia coli. J. Biol. Chem. 277:7412–7419.[Abstract/Free Full Text]
  11. Ladwein M., Pape U.F., Schmidt D.S., Schnolzer M., Fiedler S., Langbein L., Franke W.W., Moldenhauer G., Zoller M. (2005) The cell–cell adhesion molecule EpCAM interacts directly with the tight junction protein claudin-7. Exp. Cell. Res. 309:345–357.[CrossRef][Web of Science][Medline]
  12. Hough C.D., Sherman-Baust C.A., Pizer E.S., Montz F.J., Im D.D., Rosenshein N.B., Cho K.R., Riggins G.J., Morin P.J. (2000) Large-scale serial analysis of gene expression reveals genes differentially expressed in ovarian cancer. Cancer. Res. 60:6281–6287.[Abstract/Free Full Text]
  13. Johnson A.H., Frierson H.F., Zaika A., Powell S.M., Roche J., Crowe S., Moskaluk C.A., El-Rifai W. (2005) Expression of tight-junction protein claudin-7 is an early event in gastric tumorigenesis. Am. J. Pathol. 167:577–584.[Abstract/Free Full Text]
  14. Nichols L.S., Ashfaq R., Iacobuzio-Donahue C.A. (2004) Claudin 4 protein expression in primary and metastatic pancreatic cancer: support for use as a therapeutic target. Am. J. Clin. Pathol. 121:226–230.[CrossRef][Web of Science][Medline]
  15. Hibbs K., Skubitz K.M., Pambuccian S.E., Casey R.C., Burleson K.M., Oegema T.R. Jr, Thiele J.J., Grindle S.M., Bliss R.L., Skubitz A.P. (2004) Differential gene expression in ovarian carcinoma: identification of potential biomarkers. Am. J. Pathol. 165:397–414.[Abstract/Free Full Text]
  16. Kominsky S.L., Vali M., Korz D., Gabig T.G., Weitzman S.A., Argani P., Sukumar S. (2004) Clostridium perfringens enterotoxin elicits rapid and specific cytolysis of breast carcinoma cells mediated through tight junction proteins claudin 3 and 4. Am. J. Pathol. 164:1627–1633.[Abstract/Free Full Text]
  17. Michl P., Buchholz M., Rolke M., Kunsch S., Lohr M., McClane B., Tsukita S., Leder G., Adler G., Gress T.M. (2001) Claudin-4: a new target for pancreatic cancer treatment using Clostridium perfringens enterotoxin. Gastroenterology 121:678–684.[CrossRef][Web of Science][Medline]
  18. Santin A.D., Cane S., Bellone S., Palmieri M., Siegel E.R., Thomas M., Roman J.J., Burnett A., Cannon M.J., Pecorelli S. (2005) Treatment of chemotherapy-resistant human ovarian cancer xenografts in C.B-17/SCID mice by intraperitoneal administration of Clostridium perfringens enterotoxin. Cancer. Res. 65:4334–4342.[Abstract/Free Full Text]
  19. Dhawan P., Singh A.B., Deane N.G., No Y., Shiou S.R., Schmidt C., Neff J., Washington M.K., Beauchamp R.D. (2005) Claudin-1 regulates cellular transformation and metastatic behavior in colon cancer. J. Clin. Invest. 115:1765–1776.[CrossRef][Web of Science][Medline]
  20. Resnick M.B., Konkin T., Routhier J., Sabo E., Pricolo V.E. (2005) Claudin-1 is a strong prognostic indicator in stage II colonic cancer: a tissue microarray study. Mod. Pathol. 18:511–518.[CrossRef][Web of Science][Medline]
  21. Borlak J., Meier T., Halter R., Spanel R., Spanel-Borowski K. (2005) Epidermal growth factor-induced hepatocellular carcinoma: gene expression profiles in precursor lesions, early stage and solitary tumours. Oncogene. 24:1809–1819.[CrossRef][Web of Science][Medline]
  22. Kominsky S.L., Argani P., Korz D., Evron E., Raman V., Garrett E., Rein A., Sauter G., Kallioniemi O.P., Sukumar S. (2003) Loss of the tight junction protein claudin-7 correlates with histological grade in both ductal carcinoma in situ and invasive ductal carcinoma of the breast. Oncogene 22:2021–2033.[CrossRef][Web of Science][Medline]
  23. Saito T. (1984) Ultrastructural changes on the junctional complexes in the human urinary bladder carcinoma by thin sectioning and freeze fracture. J. Clin. Electron Micros. 17:201–209.
  24. Hollande F., Lee D.J., Choquet A., Roche S., Baldwin G.S. (2003) Adherens junctions and tight junctions are regulated via different pathways by progastrin in epithelial cells. J. Cell. Sci. 116:1187–1197.[Abstract/Free Full Text]
  25. In Sobin L.H. and Wittekind C. (Eds.). TNM Classification of Malignant Tumours (2002) (John Wiley & Sons, New York).
  26. Epstein J.I., Amin M.B., Reuter V.R., Mostofi F.K. (1998) The World Health Organization/International Society of Urological Pathology consensus classification of urothelial (transitional cell) neoplasms of the urinary bladder. Bladder Consensus Conference Committee. Am. J. Surg. Pathol. 22:1435–1448.[CrossRef][Web of Science][Medline]
  27. Takai D. and Jones P.A. (2003) The CpG island searcher: a new WWW resource. In. Silico. Biol. 3:235–240.[Medline]
  28. Li L.C. and Dahiya R. (2002) MethPrimer: designing primers for methylation PCRs. Bioinformatics 18:1427–1431.[Abstract/Free Full Text]
  29. Johnstone C.N., Tebbutt N.C., Abud H.E., et al. (2000) Characterization of mouse A33 antigen, a definitive marker for basolateral surfaces of intestinal epithelial cells. Am. J. Physiol. Gastrointest. Liver Physiol. 279:G500–G510.[Abstract/Free Full Text]
  30. Johnstone C.N., White S.J., Tebbutt N.C., Clay F.J., Ernst M., Biggs W.H., Viars C.S., Czekay S., Arden K.C., Heath J.K. (2002) Analysis of the regulation of the A33 antigen gene reveals intestine-specific mechanisms of gene expression. J. Biol. Chem. 277:34531–34539.[Abstract/Free Full Text]
  31. Tebbutt N.C., Giraud A.S., Inglese M., et al. (2002) Reciprocal regulation of gastrointestinal homeostasis by SHP2 and STAT-mediated trefoil gene activation in gp130 mutant mice. Nat. Med. 8:1089–1097.[CrossRef][Web of Science][Medline]
  32. Swisshelm K., Macek R., Kubbies M. (2005) Role of claudins in tumorigenesis. Adv. Drug. Deliv. Rev. 57:919–928.[CrossRef][Web of Science][Medline]
  33. Mullin J.M., Agostino N., Rendon-Huerta E., et al. (2005) Keynote review: epithelial and endothelial barriers in human disease. Drug. Discov. Today 10:395–408.[CrossRef][Web of Science][Medline]
  34. Takai D. and Jones P.A. (2002) Comprehensive analysis of CpG islands in human chromosomes 21 and 22. Proc. Natl Acad. Sci. USA 99:3740–3745.[Abstract/Free Full Text]
  35. Robertson K.D., Uzvolgyi E., Liang G., Talmadge C., Sumegi J., Gonzales F.A., Jones P.A. (1999) The human DNA methyltransferases (DNMTs) 1, 3a and 3b: coordinate mRNA expression in normal tissues and overexpression in tumors. Nucleic Acids Res. 27:2291–2298.[Abstract/Free Full Text]
  36. Oka M., Meacham A.M., Hamazaki T., Rodic N., Chang L.J., Terada N. (2005) De novo DNA methyltransferases Dnmt3a and Dnmt3b primarily mediate the cytotoxic effect of 5-aza-2'-deoxycytidine. Oncogene 24:3091–3099.[CrossRef][Web of Science][Medline]
  37. Ting A.H., Jair K.W., Schuebel K.E., Baylin S.B. (2006) Differential requirement for DNA methyltransferase 1 in maintaining human cancer cell gene promoter hypermethylation. Cancer Res. 66:729–735.[Abstract/Free Full Text]
  38. Ribeiro-Filho L.A., Franks J., Sasaki M., et al. (2002) CpG hypermethylation of promoter region and inactivation of E-cadherin gene in human bladder cancer. Mol. Carcinog. 34:187–198.[CrossRef][Web of Science][Medline]
  39. Michl P., Barth C., Buchholz M., et al. (2003) Claudin-4 expression decreases invasiveness and metastatic potential of pancreatic cancer. Cancer Res. 63:6265–6271.[Abstract/Free Full Text]
  40. Agarwal R., D'Souza T., Morin P.J. (2005) Claudin-3 and claudin-4 expression in ovarian epithelial cells enhances invasion and is associated with increased matrix metalloproteinase-2 activity. Cancer Res. 65:7378–7385.[Abstract/Free Full Text]
  41. Lu K.H., Patterson A.P., Wang L., et al. (2004) Selection of potential markers for epithelial ovarian cancer with gene expression arrays and recursive descent partition analysis. Clin. Cancer Res. 10:3291–3300.[Abstract/Free Full Text]
  42. Katoh M. (2003) CLDN23 gene, frequently down-regulated in intestinal-type gastric cancer, is a novel member of CLAUDIN gene family. Int. J. Mol. Med. 11:683–689.[Web of Science][Medline]
  43. Kramer F., White K., Kubbies M., Swisshelm K., Weber B.H. (2000) Genomic organization of claudin-1 and its assessment in hereditary and sporadic breast cancer. Hum. Genet. 107:249–256.[CrossRef][Web of Science][Medline]
  44. Herman J.G. and Baylin S.B. (2003) Gene silencing in cancer in association with promoter hypermethylation. N. Engl. J. Med. 349:2042–2054.[Free Full Text]
  45. Archey W.B., McEachern K.A., Robson M., Offit K., Vaziri S.A., Casey G., Borg A., Arrick B.A. (2002) Increased CpG methylation of the estrogen receptor gene in BRCA1-linked estrogen receptor-negative breast cancers. Oncogene. 21:7034–7041.[CrossRef][Web of Science][Medline]
  46. Miwa N., Furuse M., Tsukita S., Niikawa N., Nakamura Y., Furukawa Y. (2000) Involvement of claudin-1 in the beta-catenin/Tcf signaling pathway and its frequent upregulation in human colorectal cancers. Oncol. Res. 12:469–476.
  47. Rangel L.B., Agarwal R., D'Souza T., Pizer E.S., Alo P.L., Lancaster W.D., Gregoire L., Schwartz D.R., Cho K.R., Morin P.J. (2003) Tight junction proteins claudin-3 and claudin-4 are frequently overexpressed in ovarian cancer but not in ovarian cystadenomas. Clin. Cancer Res. 9:2567–2575.[Abstract/Free Full Text]
  48. Matsuda M., Kubo A., Furuse M., Tsukita S. (2004) A peculiar internalization of claudins, tight junction-specific adhesion molecules, during the intercellular movement of epithelial cells. J. Cell Sci. 117:1247–1257.[Abstract/Free Full Text]
  49. Muller D., Kausalya P.J., Claverie-Martin F., Meij I.C., Eggert P., Garcia-Nieto V., Hunziker W. (2003) A novel claudin 16 mutation associated with childhood hypercalciuria abolishes binding to ZO-1 and results in lysosomal mistargeting. Am. J. Hum. Genet. 73:1293–1301.[CrossRef][Web of Science][Medline]
  50. Plachot C. and Lelievre S.A. (2004) DNA methylation control of tissue polarity and cellular differentiation in the mammary epithelium. Exp. Cell Res. 298:122–132.[CrossRef][Web of Science][Medline]
  51. Heiskala M., Peterson P.A., Yang Y. (2001) The roles of claudin superfamily proteins in paracellular transport. Traffic 2:93–98.[Medline]
  52. Colegio O.R., Van Itallie C., Rahner C., Anderson J.M. (2003) Claudin extracellular domains determine paracellular charge selectivity and resistance but not tight junction fibril architecture. Am. J. Physiol. Cell Physiol. 284:C1346–C1354.[Abstract/Free Full Text]
  53. Van Itallie C., Rahner C., Anderson J.M. (2001) Regulated expression of claudin-4 decreases paracellular conductance through a selective decrease in sodium permeability. J. Clin. Invest. 107:1319–1327.[Web of Science][Medline]
  54. Mroczkowski B. and Reich M. (1993) Identification of biologically active epidermal growth factor precursor in human fluids and secretions. Endocrinology 132:417–425.[Abstract/Free Full Text]
Received March 11, 2006; revised June 4, 2006; accepted June 30, 2006.


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Clin. Cancer Res.Home page
F. R. Fritzsche, B. Oelrich, M. Johannsen, I. Kristiansen, H. Moch, K. Jung, and G. Kristiansen
Claudin-1 Protein Expression is a Prognostic Marker of Patient Survival in Renal Cell Carcinomas
Clin. Cancer Res., November 1, 2008; 14(21): 7035 - 7042.
[Abstract] [Full Text] [PDF]

Home page
 Molecular Cancer TherapeuticsHome page
Y. C. Ip, S. T. Cheung, Y. T. Lee, J. C. Ho, and S. T. Fan
Inhibition of hepatocellular carcinoma invasion by suppression of claudin-10 in HLE cells
Mol. Cancer Ther., November 1, 2007; 6(11): 2858 - 2867.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
28/2/246    most recent
bgl120v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (11)
Right arrowRequest Permissions
Google Scholar
Right arrow Articles by Boireau, S.
Right arrow Articles by Hollande, F.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Boireau, S.
Right arrow Articles by Hollande, F.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?