Clinical Cancer Research The Science of Cancer Health Disparities Stand Up to Cancer
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wild, P. J.
Right arrow Articles by Hartmann, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wild, P. J.
Right arrow Articles by Hartmann, A.
Clinical Cancer Research Vol. 11, 4415-4429, June 15, 2005
© 2005 American Association for Cancer Research

Imaging, Diagnosis, Prognosis

Gene Expression Profiling of Progressive Papillary Noninvasive Carcinomas of the Urinary Bladder

Peter J. Wild1, Alexander Herr3, Christoph Wissmann5, Robert Stoehr2, Andre Rosenthal6, Dirk Zaak7, Ronald Simon8, Ruth Knuechel9, Christian Pilarsky4 and Arndt Hartmann1

Authors' Affiliations: 1 Institute of Pathology and 2 Department of Urology, University of Regensburg, Regensburg, Germany; 3 Institute of Clinical Genetics, Medical Faculty Carl Gustav Carus, University of Technology and 4 Department of Surgery, University Hospital Dresden, Dresden, Germany; 5 Department of Internal Medicine, Charite Campus Virchow-Klinikum, Humboldt University, Berlin, Germany; 6 Signature Diagnostics AG, Potsdam, Germany; 7 Department of Urology, Ludwig-Maximilian University, Munich, Germany; 8 Institute of Pathology, University of Basel, Basel, Switzerland; and 9 Institute of Pathology, University of Aachen, Aachen, Germany

Requests for reprints: Arndt Hartmann, Institute of Pathology, University of Regensburg, Franz-Josef-Strauss-Allee 11, D-93053 Regensburg, Germany. Phone: 49-0941-944-6605; Fax: 49-0941-944-6602; E-mail: arndt.hartmann{at}klinik.uni-r.de.


   Abstract
Top
 Abstract
Materials and Methods
 Results
 Discussion
 References
 
Purpose: The aim of the present study was to define gene expression profiles of noninvasive and invasive bladder cancer, to identify potential therapeutic or screening targets in bladder cancer, and to define genetic changes relevant for tumor progression of recurrent papillary bladder cancer (pTa).

Experimental Design: Overall, 67 bladder neoplasms (46 pTa, 3 pTis, 10 pT1, and 8 pT2) and eight normal bladder specimens were investigated by a combination of laser microdissection and gene expression profiling. Eight of 16 patients with recurrent noninvasive papillary bladder tumors developed carcinoma in situ (pTis) or invasive bladder cancer (≥pT1G2) in the course of time. RNA expression results of the putative progression marker cathepsin E (CTSE) were confirmed by immunohistochemistry using high-throughput tissue microarray analysis (n = 776). Univariate analysis of factors regarding overall survival, progression-free survival, and recurrence-free survival in patients with urothelial bladder cancer was done.

Results: Hierarchical cluster analyses revealed no differences between pTaG1 and pTaG2 tumors. However, distinct groups of invasive cancers with different gene expression profiles in papillary and solid tumors were found. Progression-associated gene profiles could be defined (e.g., FABP4 and CTSE) and were already present in the preceding noninvasive papillary tumors. CTSE expression (P = 0.003) and a high Ki-67 labeling index of at least 5% (P = 0.01) were the only factors that correlated significantly with progression-free survival of pTa tumors in our gene expression approach.

Conclusions: Gene expression profiling revealed novel genes with potential clinical utility to select patients that are more likely to develop aggressive disease.

Key Words: Bladder Neoplasms • Carcinoma • Papillary • Biological Markers • Disease Progression

At the time of first diagnosis, ~70% of bladder tumors are noninvasive papillary low-grade tumors (pTa). However, >60% of these tumors will recur at least once and progress to infiltrating or less differentiated neoplasms in 5% to 10% of cases (1). Despite the fact that the majority of superficial bladder tumors are clinically benign, regular cystoscopic follow-up at intervals is done in all patients with noninvasive bladder cancer after complete transurethral resection to detect recurrence and progression. A plethora of potential clinical and histopathologic factors indicative of tumor progression are currently discussed in the context of papillary bladder cancer: high tumor grade (2), tumor size of >5 cm (3), multifocality (4), adjacent carcinoma in situ (5), and high rate of recurrence. However, none of these markers reliably predicts a higher progression rate in papillary carcinoma of the bladder (pTa). It is important to identify the small subgroup of patients that will most likely benefit from close clinical follow-up and Bacillus Calmette-Guerin instillation therapy. New molecular prognostic markers for the prediction of tumor recurrence and progression are urgently needed.

Mutations of the tumor suppressor genes TP53 and RB1 are common and have predictive value in clinical studies of invasive bladder cancer (6–8). Although TP53 alterations have been suggested as prognostic marker in pTa tumors (9), the prognostic value of both TP53 and RB1 is restricted to invasive tumors. In superficial bladder cancer, homogeneous expression of cytokeratin 20 (10), lack of FGFR3 mutations (11, 12), and high nuclear Ki-67 labeling index (12) show promise in predicting recurrence. However, there is no prospectively evaluated set of molecular markers with sufficient predictive power to select patients for a differential therapeutic approach.

Molecular profiling may identify clinically useful molecular alterations (reviewed in ref. 13). Recent studies have offered preliminary data on gene expression profiles of urothelial carcinomas and derivative cell lines (14–22). These studies have provided useful insights into the molecular biology of bladder cancer, but exclusive expression analysis of bulk tissue without microdissection of tumor cells from surrounding normal tissue, pooling of samples of the same tumor grade and stage, and lack of longitudinal analysis of recurrent superficial bladder tumors hamper the interpretation of the data.

The aim of the present study was to investigate the gene expression of papillary superficial urothelial tumors in patients with a known clinical course using a combination of oligonucleotide and tissue microarray technologies.


   Materials and Methods
Top
 Abstract
 Materials and Methods
Results
 Discussion
 References
 
Chip design. The metg001A GeneChip, a custom-designed Affymetrix oligonucleotide array, consisted of 6,117 probe sets representing 2,800 human genes based on the annotation of the probe sets with the GoldenPath assembly (http://genome.ucsc.edu/). We selected two different sets of genes. First, genes that are known to be involved in the progression of human tumors including signal transduction pathway genes (e.g., TGFB, RAS, and WNT). Second, cDNA fragments differentially expressed in silico, which were selected by systematically screening whole EST libraries for genes that are differentially expressed in normal and tumor tissues (23).

Tumor samples. The expression profiles of 67 bladder tumor samples from 33 individuals were analyzed by DNA chip hybridization. The 67 tumors included 46 pTa, 3 pTis, 10 pT1, and 8 pT2 tumors that were obtained from the Department of Urology, Ludwig-Maximilian University of Munich, Munich, Germany. Estimated median progression-free survival time of patients (n = 16) with recurrent pTaG1/pTaG2 tumors was 53 months. All specimens were immediately frozen in liquid nitrogen in the operating room and shipped to the University of Regensburg on dry ice. Tumor grade and stage were assigned according to the 1998 consensus classification of urothelial neoplasms of the urinary bladder (WHO; ref. 2) and were evaluated by two urological pathologists (R.K. and A.H.). Specimens were histologically scored according the revised (2004) WHO classification (24). Growth pattern was determined for all invasive tumors. Papillary growth was defined by the presence of a papillary tumor component (≥20%) with a histologic grade identical to the invasive tumor. All other tumors were considered to have a solid growth pattern. After serial sectioning of tissue for routine pathology examination, samples were stored at –80°C. Eight normal urothelial specimens from eight individuals were used as reference samples. Three of those were obtained from patients with no history of bladder neoplasia. Tables 1 and 2 summarize clinicopathologic characteristics of the analyzed samples. The Institutional Review Board (IRB no. 19/99) of the Ludwig-Maximilian University approved analysis of tissues from human subjects.


View this table:
[in this window]
[in a new window]

  		Table 1. Clinical data on disease courses of recurrent pTa tumors with/without progression and results of class prediction based on gene expression

 

View this table:
[in this window]
[in a new window]

  		Table 2. Clinical data on disease courses of noninvasive and invasive urinary bladder neoplasms

 
Microdissection and RNA isolation. Total RNA was isolated from five to twenty 5-µm frozen sections using sterilized (3 hours, 250°C) glass slides and racks. To prevent RNase contamination, frozen slides (–80°C) were directly transferred into 100% ethanol (1 minute) and subsequently stained in 1% methylene blue, pretreated with 0.1% diethylpyrocarbonate. Pure populations of >90% normal or neoplastic urothelial cells were obtained using a PALM Robot-Microbeam laser microdissection device (Wolfratshausen, Germany) as described previously (25). At least 2,000 cells were isolated from each specimen. Poly(A)+ RNA was isolated from lysed cells by magnetic separation (PolyATract System 1000; Promega, Heidelberg, Germany) according to the manufacturer's specifications.

cRNA synthesis and gene expression profiling. Linear amplification (two rounds) was done as described previously (26). After priming with the Affymetrix T7-oligo-dT promoter-primer combination (5'-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGGT24-3' at 100 mmol/L), first- and second-strand synthesis and in vitro transcription, the amplified RNA was again amplified in two subsequent rounds of cDNA synthesis and in vitro transcription. The cDNA of each round of amplification was tested by Taqman PCR (see below) for its integrity, and cDNAs of low quality were excluded from further analysis. Within the last in vitro transcription, biotinylated nucleotides were incorporated into the amplified RNA. Hybridization and detection of the labeled amplified RNA on the metg001A Affymetrix GeneChip was done according to the manufacturer's instructions.

Taqman PCR tested integrity of cDNA of the two amplification rounds (first round: succinate dehydrogenase complex, subunit A (SDHA) assay and second round: glyceraldehyde-3-phosphate dehydrogenase assay). cDNA, reverse transcribed from 1 ng of amplified RNA, was used for a Taqman assay (Applied Biosystems, Weiterstadt, Germany). Genes were amplified with the Taqman Universal PCR Master Mix according to the manufacturer's conditions, using the ABI PRISM 5700 Sequence Detection System. The following oligonucleotides were used for real-time reverse transcription-PCR: glyceraldehyde-3-phosphate dehydrogenase (forward) 5'-GAAGGTGAAGGTCGGAGTC-3'; glyceraldehyde-3-phosphate dehydrogenase (reverse) 5'-GAAGATGGTGATGGGATTTC-3'; glyceraldehyde-3-phosphate dehydrogenase (probe) 5'-FAM-CAAGCTTCCCGTTCTCAGCC-Tamra-3'. SDHA (forward) 5'-TGTCATCGCACTGTGCATAGAG-3'; SDHA (reverse) 5'-CCGTAGCCTCCTGTGGCA-3'; SDHA (probe) 5'-CCATCCATCGCATAAGAGCAAAGAACACTG-Tamra-3'. At each round of amplification, CT values of <30 were required for further processing.

Data processing. GeneChips were scanned using an Agilent GeneArray Scanner (Agilent Technologies, Palo Alto, CA) and processed as described (26). In brief, raw intensity values were extracted from the Cel files. For background correction, the chip was partitioned into 16 tiles. For each tile, the mean of the 2% probes with the lowest intensities were determined and subtracted from each probe value, respectively. The background corrected probe intensity values were normalized by dividing them by the median value of all probes. A representative expression value for each probe set (PMQ value) was generated by using the 75% percentile of the perfect match intensities. For each probe set, a nonparametric Wilcoxon test was calculated by comparing the intensities of the perfect match and mismatch probes to test the probe sets for presence or absence of an expression signal.

To minimize technically caused data perturbation, a model-fitting algorithm was applied to the PMQ data. For this purpose, an ideal expression profile was constructed by determining the median PMQ value of all analyzed chips thus representing a theoretical reference chip. Expression data of each individual chip were compared with the reference chip. To reduce nonlinear effects, probe sets were separated into two groups by means of their predominant call (absence or presence; P < 0.05). Both groups were then fitted individually. A linear regression using "Robust Statistics" rules (27) was done, and the data were linear transformed so that a slope, m = 1 and an intercept, n = 0 was reached. Finally, resulting PMQ values were normalized again by dividing them by the chip median PMQ. For further analysis, the data were transformed into log space (ln). For genes represented by more than one probe set, the probe set with the highest average expression value was chosen, unless found saturated.

Analysis and presentation of expression data. Medians of PMQ values were used to generate a correlation matrix based on the Pearson linear regression coefficient. The 20 genes whose expression displayed the highest correlation coefficient from each subgroup were selected. Principal components analysis using GeneMaths (Applied Maths, v1.5) used the resulting 225 genes to group the patients.

The signal-to-noise ratio was used to compute gene-class correlations and to sort genes accordingly. Assessment of significance for gene markers was based on permutation tests (28). The top signal-to-noise scores for top marker genes were compared with corresponding scores for random permutation versions of the class labels. Typically, 1,000 random permutations were used, and the 5% and 1% significance levels were determined to compare them with the values obtained for the real data. The following gene-class correlations were computed, using the Genesis software package (29): tumor (pTa, pTis, pT1, and pT2) versus normal (Supplementary Data A); pTaG1 versus pTaG2; pTa low-grade versus pTa high-grade; pT1-pT2 papillary versus solid growth pattern (Supplementary Data B).

Class prediction. Analysis of expression data was done using GeneCluster 2.0 software (http://www-genome.wi.mit.edu/cancer/software/genecluster2/gc2.html; ref. 28). To build a class predictor for progressive or nonprogressive pTa bladder cancer, genes were filtered using linear Pearson's regression. Genes were ranked by the absolute values of the correlation coefficient and the 500 top ranked genes were used for further processing. A predictor was built using all samples with the implemented weighed voting algorithm allowing 31 genes at maximum. The predictor was tested using the leave one out cross-validation procedure. Progression of recurrent superficial papillary bladder cancer (pTaG1 and pTaG2) was defined if pTis or ≥pT1 occurred as relapse.

Quantitative reverse transcription-PCR. The following oligonucleotides were used for quantitative reverse transcription-PCR (RT-PCR): fatty acid binding protein 4, FABP4 (forward) 5'-AACCTTAGATGGGGGTGTCC-3'; FABP4 (reverse) 5'-ATGCGAACTTCAGTCCAGGT-3'. cDNA reverse transcribed from 1 ng of amplified RNA of the second round was used for a Taqman assay (Applied Biosystems). Genes were amplified with the Taqman Universal PCR Master Mix according to manufacturer's conditions, using the ABI PRISM 5700 Sequence Detection System. {Delta}CT values [i.e., CT(SDHA) – CT(FABP4)] were calculated and compared between pTa tumors with and without progression using a two-sided Mann-Whitney U test.

Bladder cancer tissue microarray. A tissue microarray was constructed as described previously (30) and contained 874 formalin-fixed, paraffin-embedded bladder cancer tissues. Only the initial biopsy (n = 776) of patients with multiple subsequent tumors was included in the analysis. An experienced pathologist (G.S.) evaluated H&E-stained slides of all tumors. Tumor stage and grade were assigned according to International Union Against Cancer (UICC) and WHO criteria. Stage pT1 was defined by unequivocal tumor invasion of the suburothelial stroma and tumor-free muscularis propria. Cases with stromal invasion but no muscular bladder wall in the biopsy were classified as pT1 or higher (pT1–). These cases were included only for statistical analyses of age, gender, and grading but were excluded for all other analyses. Clinical follow-up data were available for 651 bladder cancer patients with a median follow-up period of 39 months (1-205 months). Time to recurrence and time to progression were selected as end points in patients with pTa or pT1 urothelial carcinomas. Recurrences were defined as cystoscopically visible tumors with histologic verification, and tumor progression was defined as the presence of muscularis propria invasion in a recurrent lesion. The University of Basel Institutional Review Board granted approval for the study.

Immunohistochemistry. Immunohistochemical studies used an avidin-biotin peroxidase method with a diaminobenzidine chromatogen. After antigen retrieval, (microwave oven for 30 minutes at 250 W) immunohistochemistry was carried out in a NEXES immunostainer (Ventana, Tucson, AZ) following the manufacturer's instructions. The following primary antibodies were used: anti-TP53 (monoclonal, clone DO-7, DAKO A/S, Glostrup, Denmark); anti-CTSE (goat polyclonal N-19, Santa Cruz Biotechnology, Inc., Santa Cruz, CA; dilution 1:100), and anti-MIB1 (rabbit monoclonal, Dianova, Hamburg, Germany; 1:800). One surgical pathologist (A.H.) did a blinded evaluation of the slides without knowledge of the molecular or clinical data. TP53 positivity was defined as strong nuclear staining in at least 10% of the tumor cells. Cathepsin E (CTSE) staining intensity and the number of positive cells were estimated using a four-tired scoring system: 0 (negative), no CTSE staining or weak staining in ≤20% tumor cells; 1+ (weak), weak staining in >20% and ≤80% tumor cells or moderate staining in ≤20% tumor cells; 2+ (moderate), weak staining in >80% tumor cells or moderate staining in >20% and ≤80% tumor cells or strong staining in ≤20% tumor cells; 3+ (strong), moderate staining in >80% tumor cells or strong staining in >20% tumor cells. The percentage of Ki-67-positive cells of each specimen was determined as described previously (31). High Ki-67 labeling index was defined if at least 5% of the tumor cells were positive.

Statistical analysis of tissue microarray data. Statistical analyses were completed using SPSS version 10.0 (SPSS, Chicago, IL). Contingency table analysis and two-sided Fisher's exact tests were used to study the statistical association among clinicopathologic, immunohistochemical, and molecular variables. Recurrence-free survival, progression-free survival, and overall survival curves comparing patients with or without any of the factors were calculated using the Kaplan-Meier method, with significance evaluated by two-sided log-rank statistics. For the analysis of recurrence or progression of pTa and pT1 urothelial bladder cancers, patients were censored at the date when cystectomy was done or at the time of their last tumor-free clinical follow-up appointment. For overall survival analysis, patients were censored at the time of their last tumor-free clinical follow-up appointment or at their date of death not related to the tumor.


   Results
Top
 Abstract
 Materials and Methods
 Results
Discussion
 References
 
Bladder cancer gene expression reflecting morphology and biological behavior. We analyzed the expression profiles of 67 bladder cancer tissues from 33 patients plus eight normal urothelial reference samples from eight individuals, including three specimens of patients with no history of bladder neoplasia. Gene expression profiles were examined using custom-designed Affymetrix oligonucleotide microarrays. To provide high-quality starting material for our chip-based RNA expression analysis bladder cancer specimens were laser microdissected (PALM, Wolfratshausen, Germany). Microdissection is required to reduce tumor heterogeneity and enrich tumor cells up to 90%. Similarly, normal urothelium was obtained by laser microdissection (25). With RNA integrity being tested during the first two rounds of linear amplification, at least 2,000 cells were needed. In concordance with others (32), we found the bias introduced by a linear amplification protocol to be minimal and tolerable. To verify the reproducibility, double determinations were made, showing a good correlation (data not shown). Good concordance of different probe sets for a single gene was found (e.g., FABP4).

A heat map using the expression profiles of the 225 genes with the highest correlation coefficient was used to visually verify gene correlation with classes (Fig. 1). We were able to identify not only genes correlating with distinct histopathologic subtypes but also genes representing mixed types. The analysis revealed that pTa tumors expressed a unique pattern of genes clearly discriminating them from pTis, pT1, and pT2 tumors. The expression profile of the single pTaG3 tumor (sample 18-1) mimicked the profile of pT1 tumors with papillary growth pattern. In general, expression patterns of pT1 tumors shared more similarity with pT2 than with pTa tumors.



View larger version (90K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 1. Heatmap representing the individual gene expression profiles of 75 bladder tissue samples. Samples are organized in columns and grouped from left to right in respect to their tumor stage. Genes are ordered in rows. Each horizontal block contains the 20 most discriminating genes for an individual tumor stage or groups and subsets of them. Definition of the examined group is shown on the right margin. Yellow bars under groups of samples are these definitions graphically. We were able to identify not only genes correlated with distinct histopathologic subtypes but also genes representing mixed types: N, normal; pTis; pTa, pTa tumors with and without progression; prog., only pTa tumors with recorded progression; pTa + 1, combination of pTa and pT1 (versus normal, pTis, and pT2); pT1; pT1 + 2, combination of pT1 and pT2 (versus normal, pTis, and pTa); pT2; sol, invasive tumors with solid growth pattern (versus papillary growth pattern). Expression of a gene (red) and absence of expression (blue). Due to the uneven sample size distribution, samples were analyzed in groups represented by the mean of the expression values. The top 20 genes overexpressed in the examined sample group were selected. Median centered expression values were normalized per patient and per gene (Genesis standard procedure).

 
We used principal component analysis to visualize the distinct groups identified by the selected genes (Fig. 2A and B). Using principal component analysis, we found that pTa tumors were strongly discriminated from all other samples. In detail, principal component analysis revealed a clear distinction between superficial (pTa) and invasive tumors (pT1 and pT2), whereas the transition from normal to invasive tumor samples was less apparent. The carcinomas in situ showed significant overlap with normal samples that could be explained by contamination of pTis with normal cells.



View larger version (13K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 2. A and B, three-dimensional graphical representation of the computed principal component analysis using a subset of 225 informative genes. The distance in space between the colored spots are the degree of relatedness between the samples. Close proximity of invasive lesions (pT1 and pT2) is noted, with separation from pTa specimens and normal samples. Color code: normal (green), pTis (red), pTa (blue), pT1 (yellow), pT2 (turquoise).

 
Comparing bladder tumors (pTis, pTa, pT1, and pT2) with normal urothelial samples, significantly expressed genes were selected using a probability value of 5% (Supplementary Data A). Furthermore, genes representative of an invasive solid growth pattern were identified, distinguishing these tumors from invasive papillary neoplasia (e.g., CTSE and FGFR3; Supplementary Data B). In contrast, no significantly regulated genes were found when pTaG1 were compared with pTaG2 samples (data not shown). In consideration of the recently revised WHO classification of pTa tumors (24), no genes passed permutation testing when pTa low-grade tumors were compared with pTa high-grade tumors (data not shown).

Prediction of progression in pTa samples. Of the 16 investigated patients with recurrent noninvasive papillary bladder tumors, eight developed carcinoma in situ or invasive bladder cancer in the course of time. Median follow-up of these patients was 36 months. Of note, progression-free survival time in patients (n = 3) with subsequent invasive papillary bladder cancer was significantly shorter than in that with subsequent carcinoma in situ (n = 5; P = 0.004). The expression data from 42 pTa samples (Table 1) with information on progression was used to build a predictor to separate progressing from nonprogressing tumors. The expression values were filtered to contain only data points of the 500 genes with the highest correlation coefficient.

A predictor was built using 42 recurrent pTa samples as training set for the implemented weighed voting algorithm, allowing 31 features. A leave-one-out cross-validation approach resulted in a predictor that correctly classified 33 samples but failed to predict nine, six of which being false positive and three false negative (sensitivity, 85.7%; specificity, 71.4%).

Signal-to-noise analysis revealed that the expression of only a limited number of genes was significantly correlated with progression (i.e., the signal-to-noise ratio approached the level of 1.0). Based on this observation, we chose to build a predictor using 10 genes. Testing the predictor using leave-one-out cross-validation on the same data set resulted in 31 correct classifications, two wrong calls, and nine predictions not exceeding the arbitrarily set threshold of 0.3.

Twenty-two different genes were used to build up the 42 individual predictors for the cross-validation. Eleven of them were selected more frequently, whereas 11 others appeared only sporadically among the predictor genes. Among the former ones, seven genes were up-regulated in the fraction of patients with progression: FABP4, GSTM4, SERPINA1, HDAC1, C20ORF1, DNLC2A, and PTK6. At the same time, four genes were found with decreased expression in these patients: UBC, MGMT, ITGB3BP, and PAIP2 (Fig. 3A).



View larger version (26K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 3. A, graphical representation of the predictor. The heatmap represents the log10-scaled, mean centered, and SD normalized expression values. Overexpression (red) and underexpression (blue). Genes are ordered horizontally with regard to their overexpression in the fraction of patients with tumor progression using the signal to noise ratio as measure. Gene symbols are added at the right margin. Genes overexpressed in patients with progression (red) and genes underexpressed in that group (blue). Clinical samples are ordered vertically in respect to their progress prediction: samples (left) were predicted to show progress as indicated by the black bar (bottom). Samples (right) were predicted not to progress as indicated by the gray bar (bottom). Samples without a significant call are found in the middle of the array. The absence of a respective bar indicates that no call exceeding a significance level of 0.3 could be obtained. The corresponding observed clinical findings are depicted in the lowermost bar: patients with reported progression (black) and progression-free survival (gray). B, scatterplot of the difference between CT values of quantitative RT-PCR for SDHA and FABP4. pTaG1/2 tumors without progression (blue) and pTaG1/2 tumors with progression (red); *, outlier.

 
Validation of FABP4 expression by quantitative reverse transcription-PCR. In expression profiling experiments, high FABP4 expression was associated with pTa bladder cancer with subsequent progression. FABP4 was also expressed in normal bladder epithelium. In contrast, loss of FABP4 expression was observed in samples with invasive bladder cancer (Fig. 4A). Gene expression results were confirmed by quantitative RT-PCR, showing a correlation between FABP4 array expression data (PMQ values) and FABP4 RT-PCR results (CT values; Fig. 3B).



View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 4. Analysis of the expression patterns of FABP4 (A) and CTSE (B) within the different tumor groups. Prog., progression; pap., papillary growth pattern; solid, solid growth pattern.

 
Validation of CTSE expression by immunohistochemistry. From the 225-gene set, we selected the intracellular proteinase CTSE for further validation, because a primary antibody was commercially available. CTSE mRNA was predominantly expressed in the superficial types of bladder cancer (pTa) and in pT1 and pT2 tumors with papillary growth pattern. Highest expression was found in the pTaG2 subtype. Elevated mRNA levels could be detected in the progressing in contrast to the nonprogressing fraction of pTa tumors (Fig. 4B). Investigation of CTSE protein expression in a large series of bladder cancers using the tissue microarray technology was informative in 86.3% of cases (670 of 776). CTSE expression was detected in 60.1% (403 of 670) of bladder cancers (pTa, pT1, and pT2-4). In detail, weak cytoplasmic staining (1+) was seen in 27.6% (185 of 670), moderate staining (2+) in 26.9% (180 of 670), and strong staining (3+) in only 5.7% (38 of 670) of cases. Representative CTSE immunostaining patterns in urothelial carcinomas are shown in Fig. 5A-H.



View larger version (77K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 5. Immunohistochemical CTSE staining of primary bladder cancers on the tissue microarray. Semiquantitative scoring (0 to 3+) based on cytoplasmic staining intensity and the percentage of positive cells. A-H, representative examples of papillary low-grade noninvasive urothelial bladder cancers (pTa low) with negative (0; A, x 100; B, x 630), weak (2+; C, x 100; D, x 630), moderate (2+; E, x 100; F, x 630), and strong (3+; G, x 100; H, x 630) immunoreactivity for CTSE.

 
For descriptive data analysis, all relevant variables were correlated with CTSE immunohistochemistry. Table 3 summarizes clinicopathologic and molecular characteristics of the analyzed samples relative to CTSE staining intensity. Positive CTSE staining of any intensity was observed in 42.8% (72 of 166) of muscle invasive urothelial bladder cancers, compared with 18.8% (3 of 16) of muscle invasive squamous cell carcinomas (P = 0.4). CTSE expression was detected in 35.3% (6 of 17) of small cell bladder cancers, in 50.0% (5 of 10) of sarcomatoid bladder cancers, and in 60.0% (3 of 5) of primary adenocarcinomas. CTSE expression was significantly associated with early tumor stage (P < 0.0001) but not with age at diagnosis, gender, histologic grade, and histologic growth pattern of muscle invasive bladder cancer. Thirty-six percent (104 of 287) of pTa tumors and 46.5% (66 of 142) of pT1 urothelial carcinomas revealed moderate (2+) or strong (3+) CTSE immunoreactivity compared with 19.9% (33 of 166) of muscle invasive bladder cancers (both P < 0.0001). If tumors of all stages were jointly analyzed, neither Ki-67 labeling index nor TP53 immunohistochemistry was significantly associated with CTSE expression in urothelial bladder cancer.


View this table:
[in this window]
[in a new window]

  		Table 3. CTSE expression analysis of 776 bladder tumors using tissue microarrays

 
Overall survival, progression-free survival, and recurrence-free survival were compared between CTSE-negative (0 to 1+) and CTSE-positive (2+ to 3+) cases by univariate log-rank statistics (Table 4). CTSE expression was strongly associated with longer overall survival (pTa, pT1, and pT2-4). Bladder cancer patients with CTSE-positive tumors (2+ to 3+) had an estimated mean overall survival time of 178 months (95% confidence interval, 165-191) compared with 140 months (95% confidence interval, 131-148) in patients with no/weak CTSE staining (P = 0.003). However, CTSE expression was not associated with overall survival in clinically relevant subgroups (pTa, pT1, and pT2-4). Including all tumor stages of urothelial bladder cancer, high tumor stage and grade, TP53 overexpression, and high Ki-67 labeling index were associated with shorter overall survival (P < 0.0001).


View this table:
[in this window]
[in a new window]

  		Table 4. Univariate analysis of factors regarding overall survival, progression-free survival, and recurrence-free survival in patients with urothelial carcinomas of the bladder

 
Moderate or strong CTSE expression (P = 0.003; Fig. 6) and a high Ki-67 labeling index of at least 5% (P = 0.01) were the only factors that correlated significantly with progression-free survival of patients with pTa bladder cancer. The sensitivity and specificity of predicting progression of pTa tumors based only on CTSE immunohistochemistry was 87.5% and 67.6%, respectively. Of the investigated clinicopathologic or molecular variables, only a percentage of at least 5% Ki-67 positive cells was significantly associated with shorter recurrence-free survival time in patients with pTa tumors (P = 0.04). Specifically, there was no association with progression-free survival (P = 0.6) or recurrence-free survival (P = 0.7) in pT1 urothelial tumors.



View larger version (26K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 6. Distribution of time (months) to tumor progression among patients with low (0 to 1+) and high (2+ to 3+) CTSE immunoreactivity of pTa urothelial neoplasias as estimated by the method of Kaplan and Meier. Numbers of bladder cancer patients with pTa tumors at risk over time with negative (0 to 1+) and positive (2+ to 3+) CTSE expression.

 

   Discussion
Top
 Abstract
 Materials and Methods
 Results
 Discussion
References
 
Bladder cancer has been categorized into superficial (pTa, pT1, and CIS) and invasive (pT2-4) cancer depending on whether or not tumor infiltration extended to the muscular bladder wall (33). Based on genetic data, the revised WHO classification suggests another subdivision of urinary bladder neoplasia. Two genetic subtypes with marked difference in their degree of genetic instability correspond to morphologically defined entities. The genetically stable category includes low-grade noninvasive papillary tumors. The genetically unstable category contains high-grade (including pTaG3 and pTis) and invasive carcinomas (≥pT1; reviewed in ref. 24). Our gene expression analysis revealed that pTa tumors express a unique pattern of genes, discriminating them from pTis, pT1, and pT2 tumors. The expression profile of the pTaG3 tumor mimicked the profile of the papillary pT1 tumors. In general, expression patterns of pT1 tumors shared more similarity with pT2 tumors than with pTa tumors, linking available genetic information to the current WHO classification. Of note, no significantly regulated genes were found when pTaG1 were compared with pTaG2 samples. No genes passed permutation testing when pTa low-grade tumors were compared with high-grade tumors.

Gene expression has been used for molecular classification of several cancers. Diffuse large B-cell lymphoma can be divided into important prognostic subgroups with germinal center B cell–like, activated B cell–like, and type 3 gene expression profile (34). Results of Alizadeh were confirmed using a tissue microarray (35) and have already become part of routine diagnostics. The transition of gene expression results to diagnostic applications with clinical effect (e.g., immunohistochemistry) has not been shown yet in bladder cancer. Microarray analysis was used to identify distinct gene expression profiles of carcinoma in situ, superficial, invasive, and metastasizing urothelial bladder cancer (13–22). Dyrskjot et al. (19) applied hierarchical cluster analysis to pooled suspensions of bladder cancer and identified gene clusters covarying with stage and grade. Using a cross-validation approach, a 32-gene molecular classifier was proposed, predicting disease progression and recurrence of pTa tumors.

In consideration of the revised WHO classification (24), a predictor was built based on gene expression data of 42 pTa samples, which were from two different classes: 21 samples (n = 12) experienced a progression to pTis or at least pT1 in the course of time, whereas another 21 specimens (n = 12) did not. After leave-one-out cross-validation, the predictor correctly classified 33 samples but did not predict nine. Six of the nine false positive samples are derived from patients who have not experienced a progression of their superficial bladder cancer thus far, thus adding new information to clinical and pathologic classification. This leaves the opportunity for a possible future progression. Elimination or class label change of these patients (patients 5, 6, 7, and 8) before building a predictor may lead to improved classification results (data not shown). Three specimens were derived from patients (patients 11, 12, and 13) with already recorded progression but having a negative signature (false negative). Unfortunately, no subsequent pTa tumor from these patients was available.

FABP4 was the strongest predictor that was up-regulated in the fraction of patients with progression. FABP4 expression was up-regulated in 17 of 21 pTa samples with progression and down-regulated in 21 of 24 pTa tumors without progression (Fig. 3A). The link between fatty acid transport and tumor progression may be considered as a bystander effect mediated by peroxisome proliferator-activated receptor activation. Activated peroxisome proliferator-activated receptors are known to induce FABP transcription (36) and bladder tumor progression (37). FABP4 is dynamically regulated in the course of bladder cancer progression (38, 39). Human adipocyte FABP (FABP4) is part of a family of highly homologous cytosolic proteins and was first purified from normal s.c. adipose tissue by Baxa et al. (40). The 15-kDa protein comprises about 1% of total cytosolic protein in human adipose tissue. FABP4 synthesis was only detected in lipoblasts of lipoblastoma and liposarcoma but not in other benign adipose tissue tumors and malignant connective tissue or epithelial tumors (41). In contrast, FABP4 expression by normal bladder urothelium was lost at various stages of carcinoma progression (39). In addition, low-grade tumors contained more FABP4 than their high-grade counterparts. Protein abundance and mRNA levels of FABP4 correlated in noninvasive and invasive bladder cancer. Loss of FABP4 was not compensated by an increase in epidermal FABP, as is the case in FABP4 knockout mice (38).

From the list of significantly expressed genes in the group of pTa tumors with progression (Fig. 1), anti-CTSE antibody was commercially available and could be used for further validation on the tissue microarray. CTSE was localized to 1q31 by in situ hybridization studies. CTSE is an intracellular aspartic protease that can be found at the gastric surface (42), in neuronal cells, and is also expressed at the infiltration zone of gastric carcinomas (43), cervical adenocarcinomas (44), and pancreatic ductal carcinomas (45). CTSE is suspected to be involved in the execution of neuronal death pathways (46) and parenchymal remodeling that occurs in fibrosing lung disease (47). It also has a role in antigen processing by hydrolyzing antigenic peptides and the invariant chain (CD74) which influences the expression and peptide loading of MHC II molecules in professional and nonprofessional antigen-presenting cells (48). In a rat model of urinary bladder tumor, CTSE was overexpressed in uracil-induced papillomatosis rather than in neoplastic lesions (49). According to Yamamoto et al. (49), CTSE immunohistochemistry may serve as an early biomarker for urinary bladder carcinogenesis. Mor et al. have analyzed gene expression profiles of urothelial bladder cancer using cDNA microarrays (21). The epithelial cell markers Keratin 7, S100P, and CTSE distinguished between pTa low-grade tumors and tumors of higher stage and differentiated between normal urothelium and noninvasive urothelial bladder cancer. These genes might be used as diagnostic tools for both population screening and follow-up of patients with previously identified pTa low-grade tumors (21). Orntoft et al. have reported a good correlation between transcript alterations and protein levels (50). Similar to Thykjaer et al. (17), we found a significant correlation of mRNA and protein expression for CTSE in pTa samples with and without progression. This is the first study investigating CTSE expression and clinical outcome of bladder cancer using tissue microarray technology. We clearly showed that CTSE expression is more common in noninvasive papillary bladder cancers with high risk of progression to invasive disease. However, CTSE expression did not seem to be an independent prognostic factor for overall survival in urothelial bladder cancer. We have shown that tissue microarrays are feasible tools for anticipating biological behavior of bladder cancer. Improving this approach by adding other immunohistochemical markers will further increase the predictive potential of this approach. Combined immunohistochemical analysis of a few key proteins may predict disease outcome in papillary noninvasive carcinomas of the bladder.

Normal urothelium, superficial, and invasive bladder cancers showed distinct gene expression profiles as revealed by hierarchical clustering and principal component analysis. The class prediction algorithm provided a genomic basis for diagnosis of pTa tumors with subsequent progressive behavior. Significantly regulated genes usable as potential early progression markers for superficial pTa tumors were identified. The role of CTSE expression in progression of bladder cancer was confirmed by immunohistochemistry. Gene expression profiling is a useful tool to diagnose bladder cancer subtypes and further characterize the molecular events associated with tumorigenesis and bladder cancer progression. Novel genes with clinical utility to select patients more likely to develop aggressive disease have been identified.


   Acknowledgments
 
We thank Andrea Lieschke, Monika Kerscher, Nina Nießl, and Rudolf Jung for excellent technical assistance and Stefanie Meyer for carefully revising the article.


   Footnotes
 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). P.J. Wild, A. Herr, and C. Wissmann contributed equally to this work. C. Pilarsky and A. Hartmann share the senior authorship for this work.

Received 2/10/05; accepted 3/30/05.


   References
Top
 Abstract
 Materials and Methods
 Results
 Discussion
 References
 

  1. Dalesio O, Schulman CC, Sylvester R, et al. Prognostic factors in superficial bladder tumors. A study of the European Organization for Research on Treatment of Cancer: Genitourinary Tract Cancer Cooperative Group. J Urol 1983;129:730–3.[Medline]
  2. Epstein JI, Amin MB, Reuter VR, Mostofi FK. 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 1998;22:1435–48.[CrossRef][Medline]
  3. Heney NM, Ahmed S, Flanagan MJ, et al. Superficial bladder cancer: progression and recurrence. J Urol 1983;130:1083–6.[Medline]
  4. Kiemeney LA, Witjes JA, Heijbroek RP, Verbeek AL, Debruyne FM. Predictability of recurrent and progressive disease in individual patients with primary superficial bladder cancer. J Urol 1993;150:60–4.[Medline]
  5. Althausen AF, Prout GR Jr, Daly JJ. Non-invasive papillary carcinoma of the bladder associated with carcinoma in situ. J Urol 1976;116:575–80.[Medline]
  6. Cordon-Cardo C, Zhang ZF, Dalbagni G, et al. Cooperative effects of p53 and pRB alterations in primary superficial bladder tumors. Cancer Res 1997;57:1217–21.[Abstract/Free Full Text]
  7. Cote RJ, Dunn MD, Chatterjee SJ, et al. Elevated and absent pRb expression is associated with bladder cancer progression and has cooperative effects with p53. Cancer Res 1998;58:1090–4.[Abstract/Free Full Text]
  8. Masters JR, Vani UD, Grigor KM, et al. Can p53 staining be used to identify patients with aggressive superficial bladder cancer? J Pathol 2003;200:74–81.[CrossRef][Medline]
  9. Sarkis AS, Dalbagni G, Cordon-Cardo C, et al. Nuclear overexpression of p53 protein in transitional cell bladder carcinoma: a marker for disease progression. J Natl Cancer Inst 1993;85:53–9.[Abstract/Free Full Text]
  10. Harnden P, Mahmood N, Southgate J. Expression of cytokeratin 20 redefines urothelial papillomas of the bladder. Lancet 1999;353:974–7.[CrossRef][Medline]
  11. van Rhijn BW, Lurkin I, Radvanyi F, Kirkels WJ, van der Kwast TH, Zwarthoff EC. The fibroblast growth factor receptor 3 (FGFR3) mutation is a strong indicator of superficial bladder cancer with low recurrence rate. Cancer Res 2001;61:1265–8.[Abstract/Free Full Text]
  12. van Rhijn BW, Vis AN, van der Kwast TH, et al. Molecular grading of urothelial cell carcinoma with fibroblast growth factor receptor 3 and MIB-1 is superior to pathologic grade for the prediction of clinical outcome. J Clin Oncol 2003;21:1912–21.[Abstract/Free Full Text]
  13. Dyrskjot L. Classification of bladder cancer by microarray expression profiling: towards a general clinical use of microarrays in cancer diagnostics. Expert Rev Mol Diagn 2003;3:635–47.[CrossRef][Medline]
  14. Sanchez-Carbayo M, Socci ND, Lozano JJ, et al. Gene discovery in bladder cancer progression using cDNA microarrays. Am J Pathol 2003;163:505–16.[Abstract/Free Full Text]
  15. Sanchez-Carbayo M, Socci ND, Charytonowicz E, et al. Molecular profiling of bladder cancer using cDNA microarrays: defining histogenesis and biological phenotypes. Cancer Res 2002;62:6973–80.[Abstract/Free Full Text]
  16. Duggan BJ, McKnight JJ, Williamson KE, et al. The need to embrace molecular profiling of tumor cells in prostate and bladder cancer. Clin Cancer Res 2003;9:1240–7.[Abstract/Free Full Text]
  17. Thykjaer T, Workman C, Kruhoffer M, et al. Identification of gene expression patterns in superficial and invasive human bladder cancer. Cancer Res 2001;61:2492–9.[Abstract/Free Full Text]
  18. Ying-Hao S, Qing Y, Lin-Hui W, et al. Monitoring gene expression profile changes in bladder transitional cell carcinoma using cDNA microarray. Urol Oncol 2002;7:207–12.[CrossRef][Medline]
  19. Dyrskjot L, Thykjaer T, Kruhoffer M, et al. Identifying distinct classes of bladder carcinoma using microarrays. Nat Genet 2003;33:90–6.[CrossRef][Medline]
  20. Dyrskjot L, Kruhoffer M, Thykjaer T, et al. Gene expression in the urinary bladder: a common carcinoma in situ gene expression signature exists disregarding histopathological classification. Cancer Res 2004;64:4040–8.[Abstract/Free Full Text]
  21. Mor O, Nativ O, Stein A, et al. Molecular analysis of transitional cell carcinoma using cDNA microarray. Oncogene 2003;22:7702–10.[CrossRef][Medline]
  22. Modlich O, Prisack HB, Pitschke G, et al. Identifying superficial, muscle-invasive, and metastasizing transitional cell carcinoma of the bladder: use of cDNA array analysis of gene expression profiles. Clin Cancer Res 2004;10:3410–21.[Abstract/Free Full Text]
  23. Schmitt AO, Specht T, Beckmann G, et al. Exhaustive mining of EST libraries for genes differentially expressed in normal and tumour tissues. Nucleic Acids Res 1999;27:4251–60.[Abstract/Free Full Text]
  24. Eble JN, Sauter G, Epstein JI, Sesterhenn IA, editors. World Health Organization classification of tumours. Pathology and genetics of tumours of the urinary system and male genital organs. Lyon: IARC Press; 2004.
  25. Wild P, Knuechel R, Dietmaier W, Hofstaedter F, Hartmann A. Laser microdissection and microsatellite analyses of breast cancer reveal a high degree of tumor heterogeneity. Pathobiology 2000;68:180–90.[CrossRef][Medline]
  26. Grutzmann R, Foerder M, Alldinger I, et al. Gene expression profiles of microdissected pancreatic ductal adenocarcinoma. Virchows Arch 2003.
  27. Huber PJ. Robust statistics. New York: Wiley; 1981.
  28. Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 1999;286:531–7.[Abstract/Free Full Text]
  29. Sturn A, Quackenbush J, Trajanoski Z. Genesis: cluster analysis of microarray data. Bioinformatics 2002;18:207–8.[Abstract/Free Full Text]
  30. Simon R, Atefy R, Wagner U, et al. HER-2 and TOP2A coamplification in urinary bladder cancer. Int J Cancer 2003;107:764–72.[CrossRef][Medline]
  31. Nocito A, Bubendorf L, Maria Tinner E, et al. Microarrays of bladder cancer tissue are highly representative of proliferation index and histological grade. J Pathol 2001;194:349–57.[CrossRef][Medline]
  32. Wang E, Miller LD, Ohnmacht GA, Liu ET, Marincola FM. High-fidelity mRNA amplification for gene profiling. Nat Biotechnol 2000;18:457–9.[CrossRef][Medline]
  33. Raghavan D, Shipley WU, Garnick MB, Russell PJ, Richie JP. Biology and management of bladder cancer. N Engl J Med 1990;322:1129–38.[Medline]
  34. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 2000;403:503–11.[CrossRef][Medline]
  35. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood 2004;103:275–82.[Abstract/Free Full Text]
  36. Tan NS, Shaw NS, Vinckenbosch N, et al. Selective cooperation between fatty acid binding proteins and peroxisome proliferator-activated receptors in regulating transcription. Mol Cell Biol 2002;22:5114–27.[Abstract/Free Full Text]
  37. Yoshimura R, Matsuyama M, Segawa Y, et al. Expression of peroxisome proliferator-activated receptors (PPARs) in human urinary bladder carcinoma and growth inhibition by its agonists. Int J Cancer 2003;104:597–602.[CrossRef][Medline]
  38. Gromova I, Gromov P, Wolf H, Celis JE. Protein abundancy and mRNA levels of the adipocyte-type fatty acid binding protein correlate in non-invasive and invasive bladder transitional cell carcinomas. Int J Oncol 1998;13:379–83.[Medline]
  39. Celis JE, Ostergaard M, Basse B, et al. Loss of adipocyte-type fatty acid binding protein and other protein biomarkers is associated with progression of human bladder transitional cell carcinomas. Cancer Res 1996;56:4782–90.[Abstract/Free Full Text]
  40. Baxa CA, Sha RS, Buelt MK, et al. Human adipocyte lipid-binding protein: purification of the protein and cloning of its complementary DNA. Biochemistry 1989;28:8683–90.[CrossRef][Medline]
  41. Bennett JH, Shousha S, Puddle B, Athanasou NA. Immunohistochemical identification of tumours of adipocytic differentiation using an antibody to aP2 protein. J Clin Pathol 1995;48:950–4.[Abstract/Free Full Text]
  42. Samloff IM, Taggart RT, Shiraishi T, et al. Slow moving proteinase. Isolation, characterization, and immunohistochemical localization in gastric mucosa. Gastroenterology 1987;93:77–84.[Medline]
  43. Matsuo K, Kobayashi I, Tsukuba T, et al. Immunohistochemical localization of cathepsins D and E in human gastric cancer: a possible correlation with local invasive and metastatic activities of carcinoma cells. Hum Pathol 1996;27:184–90.[CrossRef][Medline]
  44. Tenti P, Romagnoli S, Silini E, et al. Cervical adenocarcinomas express markers common to gastric, intestinal, and pancreatobiliary epithelial cells. Pathol Res Pract 1994;190:342–9.[Medline]
  45. Uno K, Azuma T, Nakajima M, et al. Clinical significance of cathepsin E in pancreatic juice in the diagnosis of pancreatic ductal adenocarcinoma. J Gastroenterol Hepatol 2000;15:1333–8.[CrossRef][Medline]
  46. Tsukuba T, Okamoto K, Yasuda Y, Morikawa W, Nakanishi H, Yamamoto K. New functional aspects of cathepsin D and cathepsin E. Mol Cells 2000;10:601–11.[CrossRef][Medline]
  47. Bosi F, Silini E, Luisetti M, et al. Aspartic proteinases in normal lung and interstitial pulmonary diseases. Am J Respir Cell Mol Biol 1993;8:626–32.
  48. Medd PG, Chain BM. Protein degradation in MHC class II antigen presentation: opportunities for immunomodulation. Semin Cell Dev Biol 2000;11:203–10.[CrossRef][Medline]
  49. Yamamoto S, Yonezawa S, Ichinose M, et al. Immunohistochemically demonstrated variation in expression of cathepsin E between uracil-induced papillomatosis and N-butyl-N-(4-hydroxybutyl)nitrosamine-induced preneoplastic and neoplastic changes in rat urinary bladder. Virchows Arch 1996;427:589–94.[Medline]
  50. Orntoft TF, Thykjaer T, Waldman FM, Wolf H, Celis JE. Genome-wide study of gene copy numbers, transcripts, and protein levels in pairs of non-invasive and invasive human transitional cell carcinomas. Mol Cell Proteomics 2002;1:37–45.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 GutHome page
C Hellerbrand, T Amann, J Schlegel, P Wild, F Bataille, T Spruss, A Hartmann, and A-K Bosserhoff
The novel gene MIA2 acts as a tumour suppressor in hepatocellular carcinoma
Gut, February 1, 2008; 57(2): 243 - 251.
[Abstract] [Full Text] [PDF]

Home page
 J EndocrinolHome page
R. Moral, R. Wang, I. H Russo, C. A Lamartiniere, J. Pereira, and J. Russo
Effect of prenatal exposure to the endocrine disruptor bisphenol A on mammary gland morphology and gene expression signature
J. Endocrinol., January 1, 2008; 196(1): 101 - 112.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
T. Kawakubo, K. Okamoto, J.-i. Iwata, M. Shin, Y. Okamoto, A. Yasukochi, K. I. Nakayama, T. Kadowaki, T. Tsukuba, and K. Yamamoto
Cathepsin E Prevents Tumor Growth and Metastasis by Catalyzing the Proteolytic Release of Soluble TRAIL from Tumor Cell Surface
Cancer Res., November 15, 2007; 67(22): 10869 - 10878.
[Abstract] [Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
L. Dyrskjot, K. Zieger, F. X. Real, N. Malats, A. Carrato, C. Hurst, S. Kotwal, M. Knowles, P.-U. Malmstrom, M. de la Torre, et al.
Gene Expression Signatures Predict Outcome in Non-Muscle-Invasive Bladder Carcinoma: A Multicenter Validation Study
Clin. Cancer Res., June 15, 2007; 13(12): 3545 - 3551.
[Abstract] [Full Text] [PDF]

Home page
 J. Clin. Pathol.Home page
R Bauer, P J Wild, S Meyer, F Bataille, A Pauer, M Klinkhammer-Schalke, F Hofstaedter, and A K Bosserhoff
Prognostic relevance of P-cadherin expression in melanocytic skin tumours analysed by high-throughput tissue microarrays
J. Clin. Pathol., July 1, 2006; 59(7): 699 - 705.
[Abstract] [Full Text] [PDF]

Home page
 CarcinogenesisHome page
M. A. Knowles
Molecular subtypes of bladder cancer: Jekyll and Hyde or chalk and cheese?
Carcinogenesis, March 1, 2006; 27(3): 361 - 373.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplementary Data
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wild, P. J.
Right arrow Articles by Hartmann, A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation