Cancer Research Cell Death Mechanisms and Cancer Therapy Sign up for Cancer Research eTOC's
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
Cancer Research 68, 1451, March 1, 2008. doi: 10.1158/0008-5472.CAN-07-6013
© 2008 American Association for Cancer Research
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 Email this article to a friend
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 Yamashita, T.
Right arrow Articles by Wang, X. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yamashita, T.
Right arrow Articles by Wang, X. W.

Cell, Tumor, and Stem Cell Biology

EpCAM and {alpha}-Fetoprotein Expression Defines Novel Prognostic Subtypes of Hepatocellular Carcinoma

Taro Yamashita1, Marshonna Forgues1, Wei Wang1, Jin Woo Kim1, Qinghai Ye4, Huliang Jia4, Anuradha Budhu1, Krista A. Zanetti1,3, Yidong Chen2, Lun-Xiu Qin4, Zhao-You Tang4 and Xin Wei Wang1

1 Liver Carcinogenesis Section, Laboratory of Human Carcinogenesis, 2 Genetics Branch, Center for Cancer Research, and 3 Cancer Prevention Fellowship Program, Division of Cancer Prevention, National Cancer Institute, Bethesda, Maryland; and 4 Liver Cancer Institute and Zhongshan Hospital, Fudan University, Shanghai, China

Requests for reprints: Xin Wei Wang, Laboratory of Human Carcinogenesis, Center for Cancer Research, National Cancer Institute, 37 Convent Drive, Room 3044A, MSC 4258, Bethesda, MD 20892-4258. Phone: 301-496-2099; Fax: 301-496-0497; E-mail: xw3u{at}nih.gov.


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
The heterogeneous nature of hepatocellular carcinoma (HCC) and the lack of appropriate biomarkers have hampered patient prognosis and treatment stratification. Recently, we have identified that a hepatic stem cell marker, epithelial cell adhesion molecule (EpCAM), may serve as an early biomarker of HCC because its expression is highly elevated in premalignant hepatic tissues and in a subset of HCC. In this study, we aimed to identify novel HCC subtypes that resemble certain stages of liver lineages by searching for EpCAM-coexpressed genes. A unique signature of EpCAM-positive HCCs was identified by cDNA microarray analysis of 40 HCC cases and validated by oligonucleotide microarray analysis of 238 independent HCC cases, which was further confirmed by immunohistochemical analysis of an additional 101 HCC cases. EpCAM-positive HCC displayed a distinct molecular signature with features of hepatic progenitor cells including the presence of known stem/progenitor markers such as cytokeratin 19, c-Kit, EpCAM, and activated Wnt-β-catenin signaling, whereas EpCAM-negative HCC displayed genes with features of mature hepatocytes. Moreover, EpCAM-positive and EpCAM-negative HCC could be further subclassified into four groups with prognostic implication by determining the level of {alpha}-fetoprotein (AFP). These four subtypes displayed distinct gene expression patterns with features resembling certain stages of hepatic lineages. Taken together, we proposed an easy classification system defined by EpCAM and AFP to reveal HCC subtypes similar to hepatic cell maturation lineages, which may enable prognostic stratification and assessment of HCC patients with adjuvant therapy and provide new insights into the potential cellular origin of HCC and its activated molecular pathways. [Cancer Res 2008;68(5):1451–61]


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Cancer is heterogeneous in terms of its biological behavior and response to treatment (1, 2). Because heterogeneity can compromise treatment options, classification of cancer according to its biological nature is required to provide the best qualified therapy for each cancer patient. Biomarker-based cancer classification has become clinically relevant in breast cancer (human epidermal growth factor receptor-2 expression and trastuzumab/Herceptin response) and lung cancer (epidermal growth factor receptor mutation and gefitinib/Iressa response; refs. 3–5). Thus, identification of potential biomarkers and therapeutic targets in cancer to provide personalized therapy is of great interest (6).

Hepatocellular carcinoma (HCC) is the third leading cause of cancer death worldwide with a dismal outcome (7). Several staging systems are currently available for HCC classification, which are based on tumor morphology, tumor number and size, vascular invasion, metastasis status, serum {alpha}-fetoprotein (AFP) levels, and hepatic reserve markers (8). Although these systems can grade patients based on prognosis, they are unable to predict pharmacologic responses to chemotherapeutic agents, largely due to a lack of specific biomarkers and mechanistic targets. To overcome these limitations, new technologies based on genome-wide screening have been applied to HCC classification (9–14). For example, primary HCC tissues with a propensity to metastasize or recur have a significantly different gene expression profile when compared with relapse-free HCC tissues (9, 10, 15). In addition, gene expression profiling of rat fetal hepatoblasts and HCC has identified a novel subtype that may have features of hepatic progenitor cells (HPC; ref. 13). These studies indicate that profiling the transcriptional characteristics (transcriptome) of HCC can provide insight into the cellular origin of a tumor and uncover HCC subtypes.

Recently, we have identified TACSTD1, which encodes a pan-carcinoma antigen epithelial cell adhesion molecule (EpCAM), to be an early biomarker of HCC because its expression is highly elevated in premalignant hepatic tissues and in a subset of HCC (16). EpCAM (also known as CO17-1A, EGP, EGP40, GA733-2, KSA, Ly74, M1S2, M4S1, MIC18, MK-1, TROP1, and hEGP-2; ref. 17) is highly expressed in many human cancers with an epithelial origin (18). The function of EpCAM and the regulation of its expression are largely unknown (19). In the adult liver, hepatocytes are negative and bile duct epithelium is positive for EpCAM expression. However, in the embryonic liver, the majority of hepatocytes express EpCAM (20). In the cirrhotic liver, EpCAM is expressed in proliferating bile ductules that are thought to be derived from HPC (20). Recent studies indicate that both hepatic stem cells and fetal hepatoblast cells (a HPC lineage evolved from hepatic stem cells) express EpCAM (21). Of note, ~35% of HCC cases express EpCAM (16, 20, 22). Likewise, AFP, a known HCC prognostic factor, is expressed in embryonic liver and is silent in adult liver (23), but ~60% of HCC patients have elevated serum AFP (24). A subclass of AFP-positive (AFP+) HCCs have a unique gene expression signature and a poor survival outcome (10, 13). Taken together, these findings suggest that the status of EpCAM and AFP in HCC may reflect unique HCC subtypes and thus could serve as unique biomarkers to stratify HCC patients.

In this study, we determined whether an EpCAM-coexpressed signature in HCC could be used to reveal HCC subtypes. Using gene expression data from three independent HCC cohorts analyzed by cDNA microarray, oligonucleotide microarray, and immunohistochemistry, we successfully classified HCCs into EpCAM+ and EpCAM– groups. We revealed that EpCAM+ HCC had a unique gene expression signature. Functional network analyses indicated that genes associated with protein synthesis and cellular development, including Wnt/β-catenin signaling, were activated, whereas mature hepatocyte–specific genes were inactivated in EpCAM+ HCCs. Immunohistochemistry analysis confirmed that EpCAM+ HCCs had an elevated expression of HPC markers [e.g., c-Kit, cytokeratin 19 (CK19), and EpCAM] and an activation of β-catenin, which suggests that EpCAM+ HCCs are related to a HPC-like phenotype with an activated Wnt-β-catenin signaling pathway. Noticeably, EpCAM+ and EpCAM– HCC could be further subclassified into four subtypes based on patients' serum level of AFP. Each of these four subtypes had a unique expression pattern with features resembling various stages of hepatic lineages, such as hepatic stem cell–like HCC, bile duct epithelium–like HCC, hepatocytic progenitor–like HCC, and mature hepatocyte–like HCC. Importantly, hepatic stem cell–like and hepatocytic progenitor–like HCC subtypes had a poor prognosis. We suggest that such a convenient classification system based on the expression of EpCAM and AFP can enable subgrouping of HCCs that are linked to HCC prognosis and unique molecular features. This system may allow for accurate assessment of outcome and the development of personalized molecular targeted therapy in HCC patients.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Clinical specimens and RNA extraction. HCC samples were obtained with informed consent from hepatitis B virus (HBV)–positive Chinese patients who underwent radical resection at the Liver Cancer Institute and Zhongshan Hospital (Fudan University, Shanghai, China) and the study was approved by the Institutional Review Board of the Liver Cancer Institute. The 40 HCC cases (cohort 1) used in the cDNA microarray analysis were previously described (9). In addition, we used our in-house oligonucleotide microarray data of 238 independent HBV-positive HCC cases (cohort 2) for validation of the EpCAM-coexpressed gene signature. Total RNA was extracted using TRIzol (Invitrogen) according to the manufacturer's instructions. A total of 101 formalin-fixed and paraffin-embedded HCC samples (cohort 3) were used for immunohistochemical analysis, 7 of which were a part of cohort 1. Among the remaining 94 samples, 69 were also used in a recent study (12). All HCC samples were histologically confirmed by two independent pathologists and no fibrolamellar HCC was included.

Quantitative reverse transcription-PCR. TACSTD1, BAMBI, DKK1, UGT2B7, and APOC1 expression were measured in quadruplicate in 32 HCC and corresponding adjacent nontumor liver samples, for which a sufficient amount and quality of RNA was available using the Applied Biosystems 7700 Sequence Detection System (Applied Biosystems). Probes used for the analyses were TACSTD1, Hs00158980_m1; BAMBI, Hs00180818_m1; DKK1, Hs00183740_m1; UGT2B7, Hs00426592_m1; APOC1, Hs00155790_m1; and 18S, Hs99999901_s1 (Applied Biosystems). All procedures were done according to the manufacturer's instructions.

Immunohistochemical analysis. Immunohistochemistry was done with Envision+ kits (DAKO) according to the manufacturer's instructions. The primary antibodies were anti-p53 monoclonal antibody clone DO-7 (DAKO), anti–β-catenin monoclonal antibody clone 14 (BD Transduction Laboratories), anti-CK19 monoclonal antibody clone RCK108 (DAKO), anti–cyclooxygenase 2 (Cox-2) monoclonal antibody (Cayman Chemical), anti–c-Kit polyclonal antibodies (DAKO), and anti-EpCAM monoclonal antibody clone VU-1D9 (Oncogene Research Products). The staining area and intensities were evaluated in each sample and graded from 0 to 3 (0, 0–5%; 1, 5–25%; 2, 25–50%; 3, >50%) and 0 to 2 (0, negative; 1, weak; 2, strong), respectively. The sum of the area and intensity scores of each marker (immunohistochemistry score) were calculated and used for principal component analysis. Samples with >5% positive staining in a given area for a particular antibody were considered as positive cases (most of EpCAM+ cases had >25% positive cells). For indirect immunofluorescence assay and confocal analysis, the primary antibodies anti-CK19 monoclonal antibody clone RCK108, anti-EpCAM monoclonal antibody clone VU-1D9, and anti-AFP rabbit polyclonal antibody (DAKO) and secondary antibodies Alexa 488 FITC-conjugated antimouse/antirabbit IgG or Alexa 568 Texas red–conjugated antimouse/antirabbit IgG (Molecular Probes) were used. Confocal fluorescence microscopic analysis was done essentially as previously described (25).

Microarray studies and statistical analysis. The cDNA microarray (9) and the oligonucleotide microarray data sets are publicly available at the National Center for Biotechnology Information Gene Expression Omnibus database (GEO accession nos. GSE364 and GSE5975, respectively). The BRB-ArrayTools software (version 3.3) was used for class comparison and prediction analyses, as previously described (9, 12). Hierarchical clustering analysis was done with the GENESIS software (version 1.6) developed by Alexander Sturn (IBMT-TUG). Multidimensional scaling analysis was done with the Partek Genomics Suite software. The interaction networks of EpCAM coexpressed genes were generated using Ingenuity Pathways Analysis (version 3.1, Ingenuity Systems). The significance of gene enrichment with a particular biologically relevant function was determined by a one-sided Fisher's exact test.

The association of EpCAM expression and clinicopathologic characteristics was examined with either Mann-Whitney U tests or {chi}2 tests. The comparison of gene expression data between two groups was examined by Mann-Whitney U tests or among four groups by the Kruskal-Wallis test. Correlation of gene expression data was examined by the Spearman correlation coefficient. The above analyses were done with GraphPad Prism software 4.0 (GraphPad Software). Univariate and multivariate logistic regression analyses were done using STATA software 9.0 (STATACorp LP). In both analyses, EpCAM expression was the independent binary variable (EpCAM+ and EpCAM–), whereas expression of CK19, c-Kit, nuclear β-catenin, Cox-2, and p53 were covariates in the model. The covariates were categorized into two groups, 0 and 1, based on whether the sum of the area and intensity scores of the marker for immunohistochemistry equaled 0 or 1 to 5, respectively. The Kaplan-Meier survival analysis with {chi}2 test was done to compare patient survival using Excel-based WinSTAT software 2001.1.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Gene signature of EpCAM+ HCC. To identify EpCAM-coexpressed genes in EpCAM+ HCC, we first analyzed a cDNA microarray data set of 40 primary HCC tissues (cohort 1; clinicopathologic characteristics of the patients are available in Supplementary Table S1; ref. 9). Fourteen (35%) HCC cases were considered EpCAM+ because they had ≥2-fold increase in the level of TACSTD1 in tumor tissues compared with nontumor tissues. EpCAM-positive expression was confirmed both by quantitative reverse transcription-PCR (RT-PCR) of 31 HCC samples with available RNA and by immunohistochemistry of 16 HCC samples with available formalin-fixed tissues. Using the given criteria for defining positive and negative cases, 26 of 31 (84%) cases from quantitative RT-PCR-array or 14 of 16 (88%) cases from immunohistochemistry array comparison were concordant. Further analysis revealed a significant positive correlation between quantitative RT-PCR or immunohistochemistry and cDNA microarray results (r = 0.75, P < 0.0001 or r = 0.73, P = 0.0012, respectively; Supplementary Fig. S1).

To search for differentially expressed genes between EpCAM+ and EpCAM– HCCs, we applied a class comparison analysis with univariate t tests and a global permutation test of the class labels (x1,000) using a supervised strategy with BRB-ArrayTools (9, 12, 16). A comparison of EpCAM+ and EpCAM– HCCs yielded a total of 71 differentially expressed genes (P < 0.005) with a ≥2-fold difference between the two classes (Fig. 1A ). To further validate EpCAM-coexpressed genes, we tested the 71 significant genes in an independent oligonucleotide microarray data set consisting of 238 additional HCC cases (cohort 2; Supplementary Table S2). Among the 71 significant genes, 59 overlapped in this array platform and were thus selected for further analysis. Hierarchical cluster analysis of these 59 genes resulted in a separation of two major subtypes among the 238 HCC cases, one containing mainly EpCAM+ HCCs and the other containing mostly EpCAM– HCCs (Fig. 1B). To further test if these 59 genes could be used as a signature to predict HCC subtypes based on EpCAM expression, we applied six different multivariate class prediction algorithms with 10-fold cross-validation and 1,000 random permutations (Fig. 1C). These analyses resulted in a statistically significant prediction of EpCAM+ and EpCAM– HCC cases with accuracy ranging from 82% to 97% (P < 0.001). Thus, we concluded that most of the 70 genes identified by microarray were associated with EpCAM expression.


Figure 1
View larger version (77K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Gene signatures of EpCAM+ HCC. A, hierarchical cluster analysis of HCC. A dendrogram of the two-way hierarchical cluster analysis of gene expression data from 40 HCC cases (cohort 1) using TACSTD1 and 70 EpCAM-coregulated genes (P < 0.005, 2-fold cutoff, 1,000 univariate permutation tests) is shown. Each cell in the matrix represents the expression level of a gene in an individual sample. Red and green cells depict high and low expression levels, respectively, as indicated by the scale bar. B, hierarchical cluster analysis of HCC. An independent data set consisting of 238 HCC cases was analyzed using EpCAM-coregulated genes. EpCAM+ and EpCAM– HCC cases are clearly separated into two major branches. C, class prediction analysis of HCC. Six different class prediction algorithms were used to classify 238 HCC cases using EpCAM-coregulated genes with 10-fold cross-validation. All classifier models predicted the EpCAM+ HCC class with 82% to 97% accuracy with statistical significance (P < 0.001). CCP, compound covariate predictor; LDA, diagonal linear discriminant analysis; 1NN, 1-nearest neighbor; 3NN, 3-nearest neighbors; NC, nearest centroid; SVM, support vector machines.

 
Functional networks activated in EpCAM+ HCC. To explore the functional relation of EpCAM coexpressed genes, we carried out an interaction network analysis. Among the 52 up-regulated genes, 35 mapped to seven relevant interaction networks with statistical significance (P < 0.01; Supplementary Fig. S2). The top five most statistically significant functions of the identified interaction networks of up-regulated or down-regulated genes are included in Supplementary Table S3. We then removed redundant nodes that were not directly connected with EpCAM-coexpressed genes. The final set of 22 EpCAM coexpressed genes (red labels) seemed to be functionally linked to the signaling networks of nine genes (CALD1, CTNNB1, EGF, Ins1, MYCN, MYOD1, PPARA, TGFB1, and TP53; Fig. 2A ). Similarly, among the 18 down-regulated genes, 7 (green labels) seemed to be functionally linked to the signaling networks of five genes (NFE2L2, UGT1A6, TCF1, HNF4A, and PLG; Fig. 2B). It seemed that genes associated with organ development and protein synthesis were selectively activated in EpCAM+ HCC, possibly through interactions of β-catenin (CTNNB1), transforming growth factor β1, insulin, epidermal growth factor, and p53 signaling. In contrast, a decrease of mature hepatocyte functions such as lipid metabolism and drug metabolism occurred in EpCAM+ HCC, possibly through inactivation of hepatic nuclear factors and the Nrf2 signaling cascade.


Figure 2
View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Interaction network and functional analysis of EpCAM coregulated genes. Twenty-two up-regulated genes (red nodes; A) and seven down-regulated genes (green nodes; B) in EpCAM+ HCC were clustered and are shown with their potential regulators (white nodes; A and B). Abbreviations of Edge Labels, Nodes, and Edges used in the analysis are indicated in Supplementary Fig. S2. C, quantitative RT-PCR analysis of representative genes involved in the functional networks (BAMBI, DKK1, UGT2B7, and APOC1). Relative fold increases or decreases in HCCs compared with adjacent nontumor liver tissues are indicated in 32 HCC cases (cohort 1) with statistical significance (Mann-Whitney U test). D, scatter plot analysis of quantitative RT-PCR data. Gene expression levels of BAMBI and DKK1 were positively correlated with those of TACSTD1, whereas gene expression levels of UGT2B7 were negatively correlated with those of TACSTD1 in HCC with statistical significance (Spearman's correlation coefficient).

 
To further validate the gene signatures associated with EpCAM expression, we carried out quantitative RT-PCR of several representative genes in the pathway. We selected the up-regulated BAMBI and DKK1 (representing classic Wnt-β-catenin signaling targets refs. 26, 27) and the down-regulated UGT2B7 and APOC1 genes (markers of mature hepatocytes). Consistent with the array data, the expression levels of BAMBI and DKK1 were significantly increased and the expression levels of UGT2B7 and APOC1 were significantly decreased in EpCAM+ HCC (Fig. 2C). There was a positive correlation between BAMBI and TACSTD1 (r = 0.73, P < 0.0001) or DKK1 and TACSTD1 (r = 0.72, P < 0.0001) and an inverse correlation between UGT2B7 and TACSTD1 (r = –0.61, P = 0.0002) or APOC1 and TACSTD1 (r = –0.34, P = 0.055; Fig. 2D).

Validation of functional networks activated in EpCAM+ HCC by immunohistochemical analysis. The above data suggest that EpCAM+ HCC may potentially reflect an HPC origin, which includes an elevated expression of HPC markers such as activated Wnt-β-catenin signaling, a key player during embryogenesis. To further validate our findings, we carried out immunohistochemical analysis on an additional 101 paraffin-embedded HCC tissues (cohort 3; Supplementary Table S4). We also selected β-catenin and p53 because they were identified as two molecular nodes in EpCAM+ HCC by interaction network analysis (Fig. 2A), CK19 and c-Kit as markers for HPC, and Cox-2 as a control.

Representative staining of each marker on serial sections of EpCAM+ and EpCAM– HCC is shown in Fig. 3A . The typical expression pattern of each marker in adjacent nontumor liver tissues is depicted in Supplementary Fig. S3. In EpCAM+ HCC, there was strong positive cytoplasmic and nuclear immunoreactivity to β-catenin, accompanied by immunoreactivity to CK19 and c-Kit (Fig. 3A). In contrast, in EpCAM– HCC, there was only weak immunoreactivity to β-catenin in the cell membrane, and neither CK19 nor c-Kit expression was detected. The frequency of expression of CK19, c-Kit, and nuclear β-catenin was significantly higher in EpCAM+ HCC when compared with EpCAM– HCC (P = 0.0007, P < 0.0001, and P = 0.0007, respectively; Fig. 3B). No significant difference in the frequency of p53 and Cox-2 expression was observed between EpCAM+ and EpCAM– HCC. Most of the EpCAM+ cases had >25% positively stained cells whereas most of the CK19- or c-Kit-positive cases had ~5% to 10% positively stained cells. To explore the association of EpCAM expression with the expression of other markers in HCC, we carried out a logistic regression analysis using EpCAM expression as the independent binary variable (Supplementary Table S5). Both univariate and multivariate analyses revealed that CK19, c-Kit, and β-catenin expression were strongly associated with an increased odds of having EpCAM expression in HCC tissues, whereas p53 and Cox-2 expression were not associated with EpCAM expression. Multidimensional scaling based on EpCAM, β-catenin, CK19, and c-Kit expression revealed a clear separation between EpCAM+ and EpCAM– HCC (Fig. 3C), indicating a robust and simple HCC subtype classification system based on EpCAM expression and associated HPC markers.


Figure 3
View larger version (82K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3. Immunohistochemical analyses of EpCAM+ and EpCAM– HCC cases. A, representative photomicrographs of EpCAM+ HCC and EpCAM– HCC stained with anti-EpCAM, anti–β-catenin, anti-CK19, anti–c-Kit, anti-p53, and anti–Cox-2 antibodies (bar, 50 µm). B, summary of immunohistochemical analysis of EpCAM+ and EpCAM– HCC. The frequency of expression of CK19, c-Kit, and nuclear β-catenin was significantly higher in EpCAM+ HCC when compared with EpCAM– HCC. C, principal component analysis of immunohistochemistry expression data. The immunohistochemistry score of each marker was calculated as described in Materials and Methods. A clear separation between EpCAM+ and EpCAM– HCC is visualized.

 
HCC subtypes defined by EpCAM and AFP and human liver cell lineages. The studies above indicate that EpCAM+ and EpCAM– HCCs are distinct tumor types with a unique activation of certain molecular signaling pathways and stem cell markers. These results suggest that difference of EpCAM expression in HCC may reflect the difference of cellular origins in liver lineages. However, when we examined the expression of EpCAM in adjacent noncancerous liver tissues, it was noted that EpCAM was abundantly expressed in both small bile ductules and a cluster of hepatocytes near the portal triad where suspected hepatic stem cell niches were located (Supplementary Fig. S3). It is possible that EpCAM expression may be confined to early stages of hepatic and/or biliary lineages. Recent findings by Schmelzer et al. (21, 28) indicate that another known HPC marker, AFP, is differentially expressed between hepatic stem cells and hepatoblasts. Consistently, in the cirrhotic liver tissues, whereas some hepatocytes were EpCAM+ and AFP+, other hepatocytes were only EpCAM+ or AFP+ (Fig. 4A, left ). Many EpCAM+ AFP– cells had a morphology resembling ductular cells (image 1) whereas EpCAM+ AFP+ cells resembled small hepatocyte-like cells (image 2). Although EpCAM– AFP+ cells (image 3) and EpCAM– AFP– cells (image 4) bore a resemblance to mature hepatocytes, EpCAM– AFP+ cells were always detected in the periportal area. These results are consistent with the findings by Schmelzer et al. (21, 28), suggesting that cells with differing EpCAM and AFP status reflect particular human liver cell lineages. Thus, we hypothesized that additional HCC subtypes may exist within EpCAM+ HCCs that resemble small bile ductules or HPC/stem cells and that EpCAM+ or EpCAM– HCC can be further subgrouped based on the expression of AFP.


Figure 4
View larger version (57K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4. Molecular features of HCC subtypes defined by EpCAM and AFP expression. A, immunofluorescence of a cirrhotic liver section stained with anti-EpCAM (green), anti-AFP (red), and 4',6-diamidino-2-phenylindole (DAPI; blue). EpCAM/AFP double-positive cells were only detected in periportal area (left). Four distinct cell types were detected based on the expression of EpCAM and AFP (right; 1–4). Dotted lines, cirrhotic nodules. B, multidimensional scaling analysis of microarray data of 40 HCC samples (cohort 1) and 238 HCC samples (cohort 2) based on the expression of 71 EpCAM-coregulated genes. Axes, first three principal components of these genes. Orange, type A HCC; yellow, type B HCC; light blue, type C HCC; blue, type D HCC. C, expression patterns of EpCAM coregulated genes and well-known HPC markers in each HCC subtype. Gene expression levels of KRT19, KRT7, and TACSTD1 (EpCAM; stem/epithelial marker), AFP and DLK1, PROM1 (CD133; stem/hepatoblast marker), BAMBI and DKK1 (Wnt/β-catenin signal marker), and CYP3A4 (mature hepatocyte marker) were measured in 238 HCC cases analyzed by oligonucleotide microarray. Each HCC subtype represents a distinct molecular portrait of EpCAM coregulated genes and HPC marker expression.

 
We carried out multidimensional scaling analysis based on the expression of EpCAM-coexpressed genes in both HCC cohort 1 and cohort 2 (Fig. 4B) stratified by EpCAM and AFP status, which resulted in four subgroups: A, B, C, and D. When a well-defined serum AFP cutoff (10, 13, 29) was added to the stratification, EpCAM+ (types A and B) and EpCAM– (types C and D) HCC were largely separated by EpCAM-coexpressed genes in three dimensional space, consistent with the results of hierarchical clustering using the same gene set (Fig. 1A and B). Noticeably, the AFP-high (type B and C) and AFP-low (type A and D) HCC subtypes were also separated in both cohorts, indicating that each HCC subtype had a distinct expression pattern. Furthermore, we used six different multivariate Class Prediction algorithms with 10-fold cross-validation and 1,000 random permutations described above to compare type A with type B, as well as type C with type D. These analyses showed that type A was significantly different from type B and type C was significantly different from type D, based on the expression of EpCAM-coexpressed genes (Supplementary Table S6).

Next, we closely examined the expression pattern of well-known HPC and mature hepatocyte markers in each HCC subtype using cohort 2 (Fig. 4C; refs. 21, 30). Stem/biliary epithelial markers KRT19 (CK19) and KRT7 (CK7) were more abundantly expressed in EpCAM+ AFP– HCC, whereas stem/hepatoblast markers DLK1 and PROM1 (CD133) as well as Wnt/β-catenin signaling target genes DKK1 and BAMBI were more abundantly expressed in EpCAM+ AFP+ HCC. In contrast, the classic mature hepatocyte marker CYP3A4 was weakly expressed in EpCAM+ AFP+ HCC and mostly expressed in EpCAM– AFP– HCC. Taken together, it seemed that there are distinct differences in the transcriptional characteristics of these four HCC subtypes stratified by EpCAM and AFP status, which may correlate with certain stages of human liver lineages.

Prognosis of HCC subtypes defined by EpCAM and AFP. To investigate the clinical outcomes of each HCC subtypes identified above, we carried out Kaplan-Meier survival analysis of HCC cases in cohort 1 (n = 40), cohort 2 (n = 238), and cohort 3 (n = 94; seven cases used for immunohistochemistry were excluded because they were included in cohort 1). When Kaplan-Meier analysis was done according to these classifications, EpCAM+ AFP+ (type B) and EpCAM– AFP+ (type C) HCC correlated with poor prognosis, whereas EpCAM– AFP– (type D) HCC correlated with an intermediate prognosis (Fig. 5A ). Interestingly, EpCAM+ AFP– (type A) HCC correlated with a good prognosis, and these results were validated in all three independent cohorts with statistical or borderline significance (P = 0.06, P = 0.05, and P = 0.0004 in cohorts 1, 2, and 3, respectively).


Figure 5
View larger version (23K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5. A, prognostic outcomes of HCC subtypes. Kaplan-Meier survival analyses of four different types of HCCs stratified by EpCAM and AFP status in cohort 1 analyzed by cDNA microarray (left), in cohort 2 analyzed by oligo microarray (middle), and in cohort 3 analyzed by immunohistochemistry (right). AFP+ and AFP– HCCs were defined as those with >300 or ≤300 ng/mL of serum AFP, respectively. The statistical P value was generated by the Cox-Mantel log-rank test. The survival curve of each HCC subtype is indicated as red (EpCAM+AFP–), blue (EpCAM+AFP+), green (EpCAM–AFP+), or black (EpCAM–AFP–). B, a model depicting the parallels between proposed HCC subtypes and normal human liver cell lineage. HCC subtypes and liver lineages are characterized by expression of EpCAM and AFP. Red and blue letters show positive and negative expression, respectively. Different clinical outcomes associated with each HCC subtype are indicated. β-Catenin signaling is activated in HpSC-HCC, implying the utility of Wnt/β-catenin signaling inhibitors for eradication of this HCC subtype.

 
Next, we investigated the clinical characteristics of each HCC subtype in cohort 2 and cohort 3 (Table 1 ). Serum AFP values, onset ages, frequencies of portal vein invasion, and tumor-node-metastasis (TNM) stages were different among four HCC subtypes with statistical significance (P < 0.05). EpCAM+ AFP+ HCC developed in young patients with advanced TNM stages and portal vein invasion, whereas EpCAM– AFP– HCC developed in more elder patients with early TNM stages. EpCAM+ AFP– HCC also developed in young patients but was associated with early TNM stages and low frequencies of portal vein invasion. These data suggested that HCC subtypes similar to certain maturation lineages represent distinct prognosis and clinical stages, which may correlate with various cell proliferative and invasive tumor cells properties in each HCC subtype.


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

  		Table 1. Clinicopathologic characteristics of HCC subtypes defined by EpCAM and AFP

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Prognostic variability and resistance against therapeutic agents in cancer are likely governed by the heterogeneity of activated gene expression patterns and molecular pathways, which may be related to differences in cellular origin (2, 31). The concept that some HCC may originate from a HPC has recently emerged (32, 33). The liver, an organ capable of self-regenerating, contains several cell types with longevity (30). Experimental evidence indicates that hepatic stem cells, rich in fetal and neonatal livers, can give rise to hepatoblasts, which further differentiate toward either the biliary lineage or hepatocytic progenitor cells and then mature hepatocytes (21, 30, 34). Recent studies indicate that whereas certain HPCs express EpCAM and hepatoblasts express EpCAM and AFP, mature hepatocytes are negative for both markers (21, 28). Consistently, human fetal liver multipotent progenitor cells and rat hepatic progenitor/oval cells express EpCAM but not AFP (35, 36). It should be noted that biliary epithelial cells or malignant cells of biliary lineage are EpCAM+ but AFP–. In this study, we revealed that HCC can be stratified by EpCAM and AFP status into four subtypes that may reflect different hepatic lineages and prognosis. Pathway analysis indicates that the underlying molecular biological activities differ significantly between the two poor prognostic HCC subtypes (i.e., EpCAM+ AFP+ and EpCAM– AFP+). Whereas activation of Wnt-β-catenin signaling is mainly associated with EpCAM+ AFP+ HCC, mature hepatocyte–specific genes are highly expressed in EpCAM– AFP+ HCC. These findings suggest that these two types of aggressive HCC are distinct and thus different therapeutic interventions should be considered. It should be emphasized that such an easy classification has been reproduced in three independent cohorts that have been analyzed by different technical platforms. Importantly, this classification can be reproduced in another cohort with a mixed etiologic background including hepatitis C cases, indicating a common applicability of this approach.5 It should be noted that the classification system is only applicable at the present time for surgically eligible HCC patients because only ~20% of HCC patients are currently qualified for resection. It would be interesting to determine whether serum EpCAM and AFP can be used as a prognostic classifier. In addition, a majority of the cases analyzed in this study are still HBV+. A further validation study is clearly required to reveal the utility of this classification system in HCC with various etiologic backgrounds. Moreover, it is interesting to determine in the future study whether EpCAM can serve as a biomarker to detect early lesions including dysplastic nodules.

Based on the unique gene expression patterns, which may be related to different liver cell lineages (30), we propose the following nomenclatures for the four subtypes described above: EpCAM+ AFP+ HCC as hepatic stem cell–like HCC (HpSC-HCC), EpCAM+ AFP– HCC as bile duct epithelium–like HCC (BDE-HCC), EpCAM– AFP+ HCC as hepatocytic progenitor–like HCC (HP-HCC), and EpCAM– AFP– HCC as mature hepatocyte–like HCC (MH-HCC) (Fig. 5B). We have shown that several known HPC markers such as CK19 and c-Kit are more abundantly expressed in HpSC-HCC whereas mature hepatocyte–specific genes such as CYP3A4 are more abundantly expressed in mature hepatocyte–like HCC. Moreover, HpSC-HCC has elevated Wnt-β-catenin signaling, implying that activation of this pathway may be a critical event in the molecular pathogenesis of this HCC subtype. Interestingly, bile duct epithelium–like HCC correlated with a good prognosis and the molecular details should be clarified in future studies. Taken together, we suggest that these four HCC subtypes, classified by EpCAM and AFP, represent different molecular portraits of HCC, which may reflect different prognosis and specific activated pathways depending on tumor cell origin.

Several studies have discussed HCC with putative HPC origins (33, 37). In particular, a novel poor prognostic HCC subtype that shares gene expression patterns with fetal hepatoblasts was recently identified (13). CK19, CK7, and vimentin were more abundantly expressed in the poorly prognostic hepatoblastoma subtype than in the hepatocyte subtype of HCC, which may correspond to our HpSC-HCC and MH-HCC, respectively. Although conceptually similar, several differences clearly exist between our study and that by Lee et al. For example, Lee et al. have shown that activation of activator protein 1 (AP-1) transcription factors may play a key role in the development of the hepatoblastoma subtype. In contrast, we have found that HpSC-HCC is associated with activation of Wnt-β-catenin signaling and the Myc transcription factor. These differences may be due to mutations in different molecular nodes caused by different etiologic factors acting within the same pathway. Importantly, however, a functional merge of the AP-1, Myc, and Wnt-β-catenin signaling molecular networks is clearly evident in many published studies and highlights the importance of identifying a common molecular node to achieve targeted therapies. It is also possible that heterogeneity, a feature of cancer stem cells (38), may exist in these types of tumors because both hepatoblastoma subtype HCC and HpSC-HCC are considered to be evolved from hepatic stem cells. Further detailed analyses of these HCC subtypes may provide molecular insight into liver cancer stem cells.

Our gene expression profiling, network pathway analysis, and immunohistochemical analysis reveal a close correlation between Wnt-β-catenin signaling and EpCAM+ HCC. Wnt-β-catenin signaling plays a pivotal role in embryogenesis and stem cell maintenance (39) and is known to be activated during liver development (40, 41). Importantly, Wnt-β-catenin signaling is also significantly activated in HCC (42). Furthermore, β-catenin mutations have been identified in a majority of hepatoblastoma, the pediatric tumor believed to originate from hepatoblasts (43). Although the importance of this signaling pathway in hepatocarcinogenesis is well established, a method to measure its activity remains controversial. Current assays measure nuclear or cytoplasmic β-catenin or frizzled-7 receptor overexpression as markers of Wnt-β-catenin activation (42–44). However, different technical platforms and measurement of different targets result in a differential proportion of HCCs with activated Wnt-β-catenin signaling. In addition, the prognostic value of Wnt-β-catenin signal activation is contentious and divisible, which may reflect the tentative criteria of Wnt-β-catenin signaling activation (45–49). Our novel HCC classification model based on EpCAM and AFP expression clearly stratified HCC subtypes with features of certain liver lineages, and Wnt-β-catenin signaling was more likely activated in HpSC-HCC with poor prognosis. Encouragingly, we have recently shown that EpCAM is a direct transcriptional target of the Wnt-β-catenin signaling pathway, further emphasizing the functional link between these two molecular nodes (50). Furthermore, either silencing of EpCAM expression by RNA interference or specific inhibition of β-catenin resulted in cell killing of EpCAM+ HCC cell lines (50). It is plausible that both EpCAM and β-catenin may serve as unique molecular targets for HpSC-HCC. Further studies using animal models would be needed to establish the utility of such stratification and targeted therapy approach.


   Acknowledgments
 
Grant support: Intramural Research Program of the U.S. National Cancer Institute, China National Natural Science Foundation for Distinguished Young Scholars grant 30325041 and Key Program Project grant 30430720, China National "863" R&D High-Tech Key Project grant 2002BA711A02-4, State Key Basic Research Program of China grant G1998051210, and Key Program Project of the Shanghai Science and Technology Committee grant 04JC14028 (H.L. Jia, Q.H. Ye, L-X. Qin, and Z-Y. Tang).

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.

We thank Curtis C. Harris and Snorri S. Thorgeirsson for useful comments, Zhihua Geng for technical assistance, Jennifer Wang for artworks, and NIH Fellows Editorial Board, Dorothea Dudek, and Karen MacPherson for editorial assistance.


   Footnotes
 
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Current address for W. Wang: Center for Biologics Evaluation and Research, Food and Drug Administration, Bethesda, MD 20892. Current address for J.W. Kim: Center for Human Genomics, Wake Forest University, Winston-Salem, NC 27157.

5 Snorri Thorgeirsson, personal communication. Back

Received 10/29/07. Revised 12/17/07. Accepted 12/27/07.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Vogelstein B, Kinzler KW. Cancer genes and the pathways they control. Nat Med 2004;10:789–99.[CrossRef][Medline]
  2. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57–70.[CrossRef][Medline]
  3. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344:783–92.[Abstract/Free Full Text]
  4. Paez JG, Janne PA, Lee JC, et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science 2004;304:1497–500.[Abstract/Free Full Text]
  5. Lynch TJ, Bell DW, Sordella R, et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 2004;350:2129–39.[Abstract/Free Full Text]
  6. Ludwig JA, Weinstein JN. Biomarkers in cancer staging, prognosis and treatment selection. Nat Rev Cancer 2005;5:845–56.[CrossRef][Medline]
  7. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA Cancer J Clin 2005;55:74–108.[Abstract/Free Full Text]
  8. Sala M, Forner A, Varela M, Bruix J. Prognostic prediction in patients with hepatocellular carcinoma. Semin Liver Dis 2005;25:171–80.[CrossRef][Medline]
  9. Ye QH, Qin LX, Forgues M, et al. Predicting hepatitis B virus-positive metastatic hepatocellular carcinomas using gene expression profiling and supervised machine learning. Nat Med 2003;9:416–23.[CrossRef][Medline]
  10. Lee JS, Chu IS, Heo J, et al. Classification and prediction of survival in hepatocellular carcinoma by gene expression profiling. Hepatology 2004;40:667–76.[CrossRef][Medline]
  11. Ding SJ, Li Y, Tan YX, et al. From proteomic analysis to clinical significance: overexpression of cytokeratin 19 correlates with hepatocellular carcinoma metastasis. Mol Cell Proteomics 2004;3:73–81.[Abstract/Free Full Text]
  12. Budhu A, Forgues M, Ye QH, et al. Prediction of venous metastases, recurrence and prognosis in hepatocellular carcinoma based on a unique immune response signature of the liver microenvironment. Cancer Cell 2006;10:99–111.[CrossRef][Medline]
  13. Lee JS, Heo J, Libbrecht L, et al. A novel prognostic subtype of human hepatocellular carcinoma derived from hepatic progenitor cells. Nat Med 2006;12:410–6.[CrossRef][Medline]
  14. Wurmbach E, Chen YB, Khitrov G, et al. Genome-wide molecular profiles of HCV-induced dysplasia and hepatocellular carcinoma. Hepatology 2007;45:938–47.[CrossRef][Medline]
  15. Iizuka N, Oka M, Yamada-Okabe H, et al. Oligonucleotide microarray for prediction of early intrahepatic recurrence of hepatocellular carcinoma after curative resection. Lancet 2003;361:923–9.[CrossRef][Medline]
  16. Kim JW, Ye Q, Forgues M, et al. Cancer-associated molecular signature in the tissue samples of patients with cirrhosis. Hepatology 2004;39:518–27.[CrossRef][Medline]
  17. Herlyn M, Steplewski Z, Herlyn D, Koprowski H. Colorectal carcinoma-specific antigen: detection by means of monoclonal antibodies. Proc Natl Acad Sci U S A 1979;76:1438–42.[Abstract/Free Full Text]
  18. Balzar M, Winter MJ, de Boer CJ, Litvinov SV. The biology of the 17-1A antigen (Ep-CAM). J Mol Med 1999;77:699–712.[CrossRef][Medline]
  19. Winter MJ, Nagelkerken B, Mertens AE, Rees-Bakker HA, Briaire-de Bruijn IH, Litvinov SV. Expression of Ep-CAM shifts the state of cadherin-mediated adhesions from strong to weak. Exp Cell Res 2003;285:50–8.[CrossRef][Medline]
  20. de Boer CJ, van Krieken JH, Janssen-van Rhijn CM, Litvinov SV. Expression of Ep-CAM in normal, regenerating, metaplastic, and neoplastic liver. J Pathol 1999;188:201–6.[CrossRef][Medline]
  21. Schmelzer E, Wauthier E, Reid LM. The phenotypes of pluripotent human hepatic progenitors. Stem Cells 2006;24:1852–8.[CrossRef][Medline]
  22. Ruck P, Wichert G, Handgretinger R, Kaiserling E. Ep-CAM in malignant liver tumours. J Pathol 2000;191:102–3.[CrossRef][Medline]
  23. Takayasu K, Arii S, Ikai I et al. Prospective cohort study of transarterial chemoembolization for unresectable hepatocellular carcinoma in 8510 patients. Gastroenterology 2006;131:461–9.[CrossRef][Medline]
  24. Taketa K. {alpha}-Fetoprotein: reevaluation in hepatology. Hepatology 1990;12:1420–32.[Medline]
  25. Wang W, Budhu A, Forgues M, Wang XW. Temporal and spatial control of nucleophosmin by the Ran-Crm1 complex in centrosome duplication. Nat Cell Biol 2005;7:823–30.[CrossRef][Medline]
  26. Chamorro MN, Schwartz DR, Vonica A, Brivanlou AH, Cho KR, Varmus HE. FGF-20 and DKK1 are transcriptional targets of β-catenin and FGF-20 is implicated in cancer and development. EMBO J 2005;24:73–84.[CrossRef][Medline]
  27. Sekiya T, Adachi S, Kohu K et al. Identification of BMP and activin membrane-bound inhibitor (BAMBI), an inhibitor of transforming growth factor-β signaling, as a target of the β-catenin pathway in colorectal tumor cells. J Biol Chem 2004;279:6840–6.[Abstract/Free Full Text]
  28. Schmelzer E, Zhang L, Bruce A, et al. Human hepatic stem cells from fetal and postnatal donors. J Exp Med 2007;204:1973–87.[Abstract/Free Full Text]
  29. Johnson Pj. The role of serum {alpha}-fetoprotein estimation in the diagnosis and management of hepatocellular carcinoma. Clin Liver Dis 2001;5:145–59.[CrossRef][Medline]
  30. Fausto N. Liver regeneration and repair: hepatocytes, progenitor cells, and stem cells. Hepatology 2004;39:1477–87.[CrossRef][Medline]
  31. 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]
  32. Thorgeirsson SS, Grisham JW. Hepatic stem cells. Semin Liver Dis 2003;23:301.[CrossRef]
  33. Libbrecht L, Desmet V, Van Damme B, Roskams T. The immunohistochemical phenotype of dysplastic foci in human liver: correlation with putative progenitor cells. J Hepatol 2000;33:76–84.[CrossRef][Medline]
  34. Roskams T. Liver stem cells and their implication in hepatocellular and cholangiocarcinoma. Oncogene 2006;25:3818–22.[CrossRef][Medline]
  35. Dan YY, Riehle KJ, Lazaro C, et al. Isolation of multipotent progenitor cells from human fetal liver capable of differentiating into liver and mesenchymal lineages. Proc Natl Acad Sci U S A 2006;103:9912–7.[Abstract/Free Full Text]
  36. Yovchev MI, Grozdanov PN, Joseph B, Gupta S, Dabeva MD. Novel hepatic progenitor cell surface markers in the adult rat liver. Hepatology 2007;45:139–49.[CrossRef][Medline]
  37. Robrechts C, De Vos R, Van den HM, et al. Primary liver tumour of intermediate (hepatocyte-bile duct cell) phenotype: a progenitor cell tumour? Liver 1998;18:288–93.[Medline]
  38. Morrison SJ, Kimble J. Asymmetric and symmetric stem-cell divisions in development and cancer. Nature 2006;441:1068–74.[CrossRef][Medline]
  39. Reya T, Clevers H. Wnt signalling in stem cells and cancer. Nature 2005;434:843–50.[CrossRef][Medline]
  40. Micsenyi A, Tan X, Sneddon T, Luo JH, Michalopoulos GK, Monga SP. β-Catenin is temporally regulated during normal liver development. Gastroenterology 2004;126:1134–46.[CrossRef][Medline]
  41. Monga SP, Monga HK, Tan X, Mule K, Pediaditakis P, Michalopoulos GK. β-Catenin antisense studies in embryonic liver cultures: role in proliferation, apoptosis, and lineage specification. Gastroenterology 2003;124:202–16.[CrossRef][Medline]
  42. Branda M, Wands JR. Signal transduction cascades and hepatitis B and C related hepatocellular carcinoma. Hepatology 2006;43:891–902.[CrossRef][Medline]
  43. Giles RH, van Es JH, Clevers H. Caught up in a Wnt storm: Wnt signaling in cancer. Biochim Biophys Acta 2003;1653:1–24.[Medline]
  44. Merle P, de la MS, Kim M, et al. Functional consequences of frizzled-7 receptor overexpression in human hepatocellular carcinoma. Gastroenterology 2004;127:1110–22.[CrossRef]
  45. Inagawa S, Itabashi M, Adachi S, et al. Expression and prognostic roles of β-catenin in hepatocellular carcinoma: correlation with tumor progression and postoperative survival. Clin Cancer Res 2002;8:450–6.[Abstract/Free Full Text]
  46. Mao TL, Chu JS, Jeng YM, Lai PL, Hsu HC. Expression of mutant nuclear β-catenin correlates with non-invasive hepatocellular carcinoma, absence of portal vein spread, and good prognosis. J Pathol 2001;193:95–101.[CrossRef][Medline]
  47. Hsu HC, Jeng YM, Mao TL, Chu JS, Lai PL, Peng SY. β-Catenin mutations are associated with a subset of low-stage hepatocellular carcinoma negative for hepatitis B virus and with favorable prognosis. Am J Pathol 2000;157:763–70.[Abstract/Free Full Text]
  48. Endo K, Ueda T, Ueyama J, Ohta T, Terada T. Immunoreactive E-cadherin, {alpha}-catenin, β-catenin, and {gamma}-catenin proteins in hepatocellular carcinoma: relationships with tumor grade, clinicopathologic parameters, and patients' survival. Hum Pathol 2000;31:558–65.[CrossRef][Medline]
  49. Wong CM, Fan ST, Ng IO. β-Catenin mutation and overexpression in hepatocellular carcinoma: clinicopathologic and prognostic significance. Cancer 2001;92:136–45.[CrossRef][Medline]
  50. Yamashita T, Budhu A, Forgues M, Wang XW. Activation of hepatic stem cell marker EpCAM by Wnt-β-catenin signaling in hepatocellular carcinoma. Cancer Res 2007;67:10831–9.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 NEJMHome page
J. Ji, J. Shi, A. Budhu, Z. Yu, M. Forgues, S. Roessler, S. Ambs, Y. Chen, P. S. Meltzer, C. M. Croce, et al.
MicroRNA Expression, Survival, and Response to Interferon in Liver Cancer
N. Engl. J. Med., October 8, 2009; 361(15): 1437 - 1447.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
Y. Hoshida, S. M.B. Nijman, M. Kobayashi, J. A. Chan, J.-P. Brunet, D. Y. Chiang, A. Villanueva, P. Newell, K. Ikeda, M. Hashimoto, et al.
Integrative Transcriptome Analysis Reveals Common Molecular Subclasses of Human Hepatocellular Carcinoma
Cancer Res., September 15, 2009; 69(18): 7385 - 7392.
[Abstract] [Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
F. Dong, A. S. Budhu, and X. W. Wang
Translating the Metastasis Paradigm from Scientific Theory to Clinical Oncology
Clin. Cancer Res., April 15, 2009; 15(8): 2588 - 2593.
[Abstract] [Full Text] [PDF]

Home page
 Acta Biochim Biophys SinHome page
Y. Pei, T. Zhang, V. Renault, and X. Zhang
An overview of hepatocellular carcinoma study by omics-based methods
Acta Biochim Biophys Sin, January 1, 2009; 41(1): 1 - 15.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Res.Home page
D. Y. Chiang, A. Villanueva, Y. Hoshida, J. Peix, P. Newell, B. Minguez, A. C. LeBlanc, D. J. Donovan, S. N. Thung, M. Sole, et al.
Focal Gains of VEGFA and Molecular Classification of Hepatocellular Carcinoma
Cancer Res., August 15, 2008; 68(16): 6779 - 6788.
[Abstract] [Full Text] [PDF]

Home page
 JCOHome page
S. Sell and H. L. Leffert
Liver Cancer Stem Cells
J. Clin. Oncol., June 10, 2008; 26(17): 2800 - 2805.
[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 Email this article to a friend
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 Yamashita, T.
Right arrow Articles by Wang, X. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yamashita, T.
Right arrow Articles by Wang, X. W.

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