Expression of Wnt-5a and its clinicopathological significance in hepatocellular carcinoma

Results Compared to normal tissue, Wnt-5a mRNA expression was clearly increased in hepatocellular carcinoma, chronic hepatitis and cirrhosis. On immunohistochemistry, immunostaining of Wnt-5a showed a bell-shaped pattern: low to undetectable levels were present in normal tissue and in tumour samples, whereas strong immunostaining was seen in chronic hepatitis, cirrhosis and dysplastic liver cells. Reduction or loss of Wnt-5a protein expression was found in 80.7% of hepatocellular carcinoma cases ( n = 92) and was significantly associated with higher tumour stage ( p < 0.001), serum AFP level ( p = 0.025), low membranous expression of E-cadherin ( p < 0.0001) and β-catenin ( p = 0.036) and high Ki-67 labelling indices (LIs, p = 0.001). Conclusion Wnt-5a mRNA and protein levels are higher than normal in hepatitis and cirrhosis and appear to be related to the presence of hepatitis B virus infection. However, Wnt-5a protein expression is frequently lost in hepatocellular carcinoma; this supports the notion that this protein has a tumour suppressor function in hepatocellular carcinoma. Keywords β-Catenin E-cadherin Hepatocellular carcinoma Wnt-5a 1 Introduction Hepatocellular carcinoma (HCC) is endemic primarily in China, Taiwan, Korea and sub-Saharan Africa, and is a leading cause of cancer-related deaths in these countries. The major HCC risk factors include various chemicals and viruses. Among these, chronic HBV and HCV infections account for the development of more than 80% of HCC cases worldwide. Other known risk factors, including AFB1 uptake, cigarette smoking, and heavy alcohol consumption, are capable of inducing HCC alone, but they also have synergetic effects [1,2] . Although the major risk factors for the development of HCC are well recognized, the molecular basis of liver cell malignant transformation is still not completely understood. Recent advances have shown that apart from autocrine stimulation by growth factors such as insulin-like growth factor-II [3] and transforming growth factor-α [4] , the dysregulation of at least four different growth regulatory pathways is frequently involved in hepatocarcinogenesis. These include the RB-1, the transforming growth factor-β, the p53 and the Wnt/β-catenin signalling pathways; these pathways also interfere with each other at different levels [1,5,6] . Wnt proteins are secreted and exert their effects through activation of distinct intracellular signalling pathways. Based on their biological activities in specific assays, vertebrate Wnt proteins have been divided into those involved in canonical signalling with transforming activities in mammary epithelial cells and those involved in non-canonical pathways [7–9] . Members of the canonical Wnt signalling pathway act via frizzled (Fz) receptors to induce β-catenin stabilization by preventing its phosphorylation, which targets β-catenin for proteosomal degradation. In this canonical Wnt pathway, stabilized β-catenin enters the cell nucleus, associates with members of the T-cell factor and lymphoid enhancer factor (TCF/LEF-1) family of transcription factors, and stimulates the expression of Wnt/β-catenin target genes, including regulators of cell growth and proliferation, as well as modulators of cell death pathways and cell–cell communication [10–13] . The inappropriate activation of canonical signalling, which induces intracellular β-catenin accumulation, has recently been implicated in several different malignancies [14,15] . β-catenin is a multifunctional protein that integrates the E-cadherin–catenin intercellular adhesion system with Wnt signalling during embryonic development [16] . Furthermore, reduced expression of E-cadherin is a frequent finding in HCC; it is most often due to CpG-island methylation within the promoter region of the gene [17] . The loss of E-cadherin not only affects cell adhesion, but may also lead to the redistribution of submembranously sequestered β-catenin into the cytoplasm [18,19] . Currently, the signalling and physiological functions of the non-canonical group of Wnt-proteins are not completely understood. Some reports indicate that Wnt-5a has oncogenic properties, based on the findings that the Wnt-5a mRNA level is upregulated in lung, prostate, gastric, and breast cancers [20,21] . Wnt-5a expression is correlated with increased cell motility and invasiveness of melanoma and breast cancer cell tumour-associated macrophages [22,23] . On the other hand, it has also been suggested that Wnt-5a acts as a tumour suppressor, since Wnt-5a can inhibit the β-catenin pathway. For instance, Wnt-5a has been shown to induce the downregulation of β-catenin through Siah2 [24] and to inhibit the transcriptional activity of TCF/LEF [25,26] . Antisense Wnt-5a mimics Wnt-1-mediated C57MG cell transformation [27] . Wnt-5a negatively regulates B-cell proliferation, and Wnt-5a heterozygous mice develop B-cell lymphoma [28] . Furthermore, Wnt-5a inhibits the proliferation, migration, and invasiveness of thyroid tumour and colorectal cancer cell lines [29,30] . Thus, the functions of Wnt-5a in human cancers are controversial and still unclear. The present study investigated the clinical significance of Wnt-5a expression in hepatocellular carcinoma. Since Wnt-5a is considered to be multifunctional, the expressions of β-catenin and E-cadherin, as well as the tumour proliferation rate using the Ki-67 index, were also evaluated. 2 Materials and methods 2.1 Patients and specimens We consecutively collected tumours from 114 patients (mean age, 52; range, 12–78 years; male:female, 106:8) who had undergone surgery for HCC at the Jinling Hospital or the Nanjing 81 hospital from January 2002 to September 2006. Among the 114 patients, 75 had serum alpha-fetoprotein (AFP) ≥30 μg/l, 98 were serum positive for hepatitis B surface antigen (HBsAg), which HCV infection is negative by serum assay, and the remainder were negative for HBV or HCV infection by serum and immunohistochemistry assay. On gross examination, 11 cases had tumour sizes that were ≤2 cm and 103 had tumour sizes >2 cm (median tumour size, 5.6 cm; range, 1.0–17 cm). The histopathological diagnosis was made according to the WHO classification (2000) [31] 16 cases were well differentiated, 63 cases were moderately differentiated, and 35 cases were poorly differentiated. The tumour T classification was according to the tumour-node-metastasis classification of the International Union against Cancer. In total, 69 HCC cases had liver cirrhosis, 37 cases had chronic hepatitis, and 8 cases had basically normal liver tissue. Furthermore, normal liver tissues obtained during surgery for liver cholelithiasis ( n = 3) and HBV-infected liver biopsy tissues ( n = 10) were studied. Of the 114 cases, fresh tissue, including tumour and adjacent non-tumourous liver tissues of HCC, which included chronic hepatitis ( n = 5) and cirrhosis ( n = 17), were obtained immediately after resection of the tumours in 22 cases. Normal liver tissues ( n = 3) with no HBV or HCV infection were obtained during surgery for liver cholelithiasis. In these 25 cases, one part of the fresh tissue was snap-frozen in liquid nitrogen immediately and stored at −80 °C. The other part was fixed in 10% buffered formalin and embedded in paraffin. 2.2 Extraction of RNA and real-time RT-PCR Total RNA was extracted from the tumour, adjacent non-tumourous liver tissues ( n = 22) and normal liver tissues ( n = 3) using 10-mm frozen sections. To isolate the RNA from defined areas containing at least 80% tumour cells, all tumours were manually microdissected under direct visual control through a dissecting microscope. Total RNA in the frozen tissues was extracted using TRIZOL (Invitrogen), following the manufacturer's recommendations. Total RNA was digested with DNase I (Invitrogen) and was used for the first-strand cDNA reaction. The reaction mixture consisted of 5 μg of DNase I-treated RNA, 1×reverse transcriptase buffer, 2.5 mM dNTP mix, 3.5 μM oligo primer, and 2.5 U/ml multiScribe reverse transcriptase (PE Applied Biosystems). Each sample was handled using the same protocol, with the exception that reverse transcriptase was added to exclude the presence of interference from genomic DNA. Real-time RT PCR was carried out using SYBER green dye in a Rotor Gene 3000 Detection System (Corbett Research, Sydney, Australia). Each SYBER green reaction (25 μl) contained 1 μl diluted cDNA and 10.5 μl SYBR Green PCR Master Mix, as well as 5 pmol forward and reverse primer (Wnt-5a: forward 5′ ATTCTTGGTGGTCGCTAGG 3′, reverse 5′ CTGTCCTTGAGAAAGTCCTG 3′). Samples were activated by incubation at 94 °C for 5 min and denatured at 94 °C for 20 s. This was followed by annealing at 60 °C for 20 s and extension at 72 °C for 20 s for 40 cycles. In all of the cDNA samples, gene expressions of Wnt-5a and beta-actin (forward 5′ CCTGTACGCCAACACAGTGC 3′; reverse 5′ ATACTCCTGCTTGCTGATCC 3′), an internal quantitative control, were determined by SYBR green fluorescence using the Rotor-Gene 3000; the ratios of Wnt-5a and beta-actin gene expressions represented the normalized relative levels of Wnt-5a expression. A no-template negative control was also included in each experiment. Analyses of all tumour samples were carried out at least twice, and the mean value was calculated. 2.3 Protein preparation and Western blot analysis Western blot analysis of Wnt-5a expression was done to confirm the specificity of the anti-Wnt-5a antibody. Briefly, tumour and adjacent non-tumourous liver cells from eight cases were lysed, and their protein content was determined before separation by gel electrophoresis, as described previously [32] . The proteins were subsequently transferred to nitrocellulose membranes (Bio-Rad Laboratories, Hercules, CA, USA). The membranes were incubated, first for 1 h with the anti-Wnt-5a antibody (sc-23698, Santa Cruz Biotechnology, Santa Cruz, USA, 1:4000) and then for 30 min with a donkey anti-goat IgG-HRP (sc-2020, Santa Cruz Biotechnology). The antibody–antigen complex was detected using an enhanced chemiluminescence detection system (sc-2048, Santa Cruz Biotechnology). 2.4 Immunohistochemistry Immunohistochemical staining was performed on thin sections (4 μm) of paraffin-embedded archival tissue. The samples were dewaxed with xylene/ethanol before antigen retrieval (pressure cooking for 1 min at full pressure, 15 psi, in 0.001 M EDTA buffer, pH 8.0). The primary antibodies used were Wnt-5a (1:2000), β-catenin (C19220, BD Transduction Laboratories, San Diego, CA, USA, 1:400), E-catenin (4A2C7, ZYMED, San Francisco, California, USA, 1:200), and Ki-67 (MIB-1, Dako, Copenhagen, Denmark, 1:100). Immunohistochemical staining of Wnt-5a was done using the goat ABC Staining System (sc-2023, Santa Cruz Biotechnology) and the staining of other antibodies was done using the Dako Envision Plus System (K5007, Dako). The antibody binding was visualized with 3,3′-diaminobenzidine tetrahydrochloride before brief counterstaining with Mayer's hematoxylin. For monoclonal antibodies of mouse origin, negative controls were obtained using isotypic mouse immunoglobulin (IgG1) in the same dilution as the primary antibody of concern, whereas non-immune goat serum was used for the Wnt-5a antibody. All control experiments gave negative results. 2.5 Evaluation of immunostaining All of the immunostained sections were reviewed by two authors (P. MH and L. ZH) who had no knowledge of the patients’ clinical status. Cases with discrepant results were re-evaluated jointly, and agreement was reached. For expression of Wnt-5a protein, beta-catenin, and E-cadherin, five areas were selected at random and scored in cases with multiple areas of low intensity that occurred during evaluation of immunostaining. The degree of immunohistochemical staining was recorded using a semiquantitative and subjective grading system that considered both the intensity of staining and the proportion of tumour cells with an unequivocal positive reaction (Intensity: 0, no staining; 1, weak staining; 2, positive staining; and 3, strong staining) (area: 0, no staining; 1, positive staining in <10% of tumour cells; 2, positive staining in 10–50% of tumour cells; 3, positive staining in >50% of tumour cells). The staining index was determined as the staining intensity times the positive area. Ki-67-positive cells were counted by monitoring at least 200 HCC cells from at least ten randomly selected fields. The percentage of antigen-positive nuclei among the total number of nuclei counted was calculated to obtain the nuclear labelling index (LI). In the subsequent statistical analysis, the cut-off points for the staining index categories were mainly based on median values as well as each marker's frequency distribution curve and the size of the subgroups. Therefore, cytoplasmic Wnt-5a, membranous β-catenin, and membranous E-cadherin staining indices were categorized by their median value as high (>4) or low (0–4), and the cytoplasmic β-catenin staining index was categorized being high (>3) or low (0–3). However, nuclear β-catenin expression was categorized based on the absence (staining index = 0) or presence (staining index ≥ 1) of staining. The Ki-67 labelling indices were divided into two groups (LI < 10% and LI ≥ 10%). 2.6 Statistical analysis All statistical calculations were carried out using SPSS.11 software. The differences between Wnt-5a mRNA expression in HCC and non-tumourous liver tissue were statistically analysed using the paired t -test or the unpaired Student's t -test for single comparisons, and one-way analysis of variance (ANOVA) for multiple comparisons. The differences between the Wnt-5a, β-catenin, and E-cadherin protein expressions and the clinicopathological features of hepatocellular carcinoma were statistically analysed using the χ 2 -test or Fisher's exact test. Pearson's correlation was used to determine whether there was a positive or negative correlation between mRNA and protein expression, as well as between the expressions of different proteins. A difference was defined as significant at p < 0.05. 3 Results 3.1 Wnt-5a gene expression in HCC The relative Wnt-5a and beta-actin gene expression levels in normal liver, hepatocellular carcinoma, and adjacent non-tumourous liver tissue determined using real-time RT-PCR are shown in Fig. 1 . The expression level was lowest in normal liver tissue (1.573 ± 0.266, n = 3). It was 10.165 ± 1.858 in adjacent non-tumourous chronic hepatitis, compared to 9.667 ± 3.088 in HCC from the same liver sample ( n = 5), 4.875 ± 1.528 in adjacent non-tumourous cirrhosis samples compared to 6.379 ± 1.392 in HCC tissue from the same sample ( n = 17). Based on the t -test, statistically significant differences were found between normal liver tissue and hepatitis, cirrhosis, or HCC (7.126 ± 1.282, n = 22, p < 0.05), but not between HCC and adjacent non-tumourous liver tissue. The Wnt5a mRNA level in moderately differentiated tumour tissues ( n = 15) was upregulated by two-fold ( p < 0.05) compared to the level in poorly differentiated tumour tissues ( n = 6); however, there was no significant difference in the level between a well differentiated tumour ( n = 1) and tumours with other degrees of differentiation, as there was only one well differentiated tumour. Furthermore, no significant differences between Wnt-5a gene expression levels and other clinicopathological findings such as age, serum AFP concentration, tumour size, and HCC tumour stage, were found. 3.2 Wnt-5a protein expression in HCC Western blotting and immunohistochemistry were performed to evaluate Wnt-5a protein expression in tumour and non-tumourous liver cells, respectively. The Wnt-5a protein expression had a cytoplasmic staining pattern in non-tumourous liver cells and in tumour cells, but stromal cells did not express Wnt-5a protein. Normal liver cells were negative or showed weakly focal expression for Wnt-5a ( Fig. 2 A ), whereas all chronic hepatitis, cirrhosis cells, and dysplastic liver cells exhibited strong positive immunostaining for Wnt-5a (Fig. 2B, C). In contrast, in 92/114 (80.7%) of the HCCs, Wnt-5a immunostaining was reduced or lost ( Fig. 2 D,E). Western blot analysis showed the same Wnt-5a protein expression results in hepatocellular carcinoma and in the adjacent non-tumourous liver tissues ( Fig. 3 ). In addition, a significant correlation was seen between the normalized Wnt-5a gene mRNA expression ratio and the protein expression level in normal liver tissues, chronic hepatitis, and cirrhosis ( r = 0.356, p = 0.024); no significant correlation was seen in tumour tissue ( r = −0.291, p = 0.178), though there appeared to be an inverse correlation. Furthermore, statistical analysis of the relationship between Wnt-5a protein expression and the clinicopathological features showed a significant association between negative Wnt-5a expression and a worse tumour stage ( p < 0.001) or a higher serum AFP level ( p = 0.025). However, there were no significant differences between Wnt-5a protein expression and the other clinicopathological findings in hepatocellular carcinoma ( Table 1 ). 3.3 Tumour cell proliferation in relation to HCC Wnt-5a expression To investigate the biological functions of Wnt-5a in HCC, the Ki-67 proliferation index was assessed in relation to the Wnt-5a status ( Fig. 2 F). A strong correlation between a high Ki-67 proliferation index and the loss of Wnt-5a protein ( r = −0.324, p = 0.001) was found in HCC. However, the Ki-67 proliferation index was higher in low Wnt-5a protein expression tumours than in high Wnt-5a expression tumours ( p = 0.030). 3.4 Altered β-catenin expression in relation to lost Wnt-5a expression in HCC In non-neoplastic liver tissue, a thin membranous β-catenin signal delineated the hepatocytes, and strong membranous and pale cytoplasmic staining of bile ductules was observed. Altered expressions patterns of β-catenin were found in 55.3% (63/114) of HCC cases, and included decreased membranous and/or increased cytoplasmic expression of β-catenin in 54 cases (47.4%, Fig. 4 A,B ), as well as focal or generalized nuclear accumulation of β-catenin in 9 of 114 (7.9%) tumour specimens ( Fig. 4 C) which reflect the mutation of β-catenin gene exon3. In contrast, no evidence of altered β-catenin expression was found in cirrhotic nodules or dysplastic liver cells in adjacent non-cancerous liver tissue. In tumour tissues, reduced membranous expression but not nuclear accumulation of β-catenin was positively correlated with low cytoplasmic Wnt-5a expression ( r = 0.157, p = 0.036). Moreover, altered β-catenin expression was significantly associated with a worse histopathological tumour grade ( p = 0.003) and was not significantly associated with the other clinicopathological parameters. 3.5 Reduced membranous expression of E-cadherin in relation to lost Wnt-5a expression in HCC Moderately positive E-cadherin expression was frequently observed in the membranes of non-cancerous liver cells, and strong positive expression was noted in the intrahepatic bile ducts. Decreased E-cadherin expression was observed in 69 of 114 (60.5%) tumours ( Fig. 4 D), and was positively correlated with low cytoplasmic Wnt-5a expression ( r = 0.333, p < 0.001) and reduced membranous expression of β-catenin ( r = 0.530, p < 0.001) in tumour tissues. There were statistically significant differences in the degree of the decrease in E-cadherin expression in HCC associated with a worse histological grade ( p = 0.042, well differentiated vs. moderately and poorly differentiated) and the T4 tumour stage ( p = 0.007, T1, T2, and T3 vs. T4 tumour stage). Furthermore, no association between altered E-cadherin expression and the other clinicopathological parameters was detected in HCC. 4 Discussion The present study demonstrates increased expression of steady-state Wnt-5a mRNA in HCC, which is correlated with a worse histological grade, while Wnt-5a expression was low in normal liver tissue. This pattern fit with observations in other epithelial type tumours, such as carcinomas of lung, gastric, and neck, where Wnt-5a mRNA is overexpressed [20,21,33] . However, not all epithelial-type tumours exhibit high Wnt-5a expression levels, since endometrial carcinomas show low Wnt-5a mRNA levels [34] . It is notable that Wnt-5a mRNA expression in non-tumourous liver tissue, including chronic hepatitis and cirrhosis, was significantly upregulated compared to expression in normal liver tissue. It thus appears that HBV infection induces Wnt-5a gene upregulation at the transcriptional level in the liver where Wnt-5a could play a tumour suppressor role. As seen by immunohistochemistry, Wnt-5a protein expression is restricted to liver cells. A bell-shaped expression pattern was observed on immunostaining: there was little to no staining in normal liver, strongly positive staining for Wnt-5a in chronic hepatitis, cirrhosis, and dysplastic liver cells, and a reduction or loss of staining in tumours (80.7%). This reduction in staining in tumours may be attributed to the fact that our study has been conducted using quite advanced cases of HCC. These results indicate the presence of post-translational dysregulation, which should be a late event in HCC. The same results were obtained with Western blot analysis. Moreover, in HCC, low to absent Wnt-5a protein expression was significantly associated with high serum AFP levels, high Ki-67 LIs, and a worse tumour stage. The finding of downregulation of Wnt-5a protein expression in the present study is compatible with the notion that Wnt-5a has tumour suppressor activity in HCC, in which the cellular defence mechanism may subsequently be lost. Interestingly, we found that in non-tumourous liver tissue, Wnt-5a protein expression was parallel to that of Wnt-5a mRNA expression, while in HCC, there appeared to be an inverse correlation that was not statistically significant. This discrepancy between transcriptional and translational control is congruent with observations made in other tumour types, such as aggressive invasive breast ductal tumours, where a selective loss of protein expression was found in the dedifferentiated state that paralleled the tumour's aggressive state [35] . Previous studies indicated that the actions of Wnt might be dependent upon the Fz species expressed in cells or tissues [26,36,37] . Thus, we investigated the related gene changes in five cases of HCC, which Wnt-5a protein lost expression, using the Wnt signalling pathway microarray. The results show that gene expression of Fr family members, such as Fz4 or Fz5, and LRP gene expression undergo no significant changes, but that Fz2, which is not a specific receptor for Wnt-5a, is upregulated in HCC (data not shown). The mechanism underlying the discrepancy between the transcript and protein expression of Wnt-5a in HCC warrants further study. Furthermore, the decreased membranous and/or increased cytoplasmic expression of β-catenin was detected in 47.4% of HCCs, which is significantly correlated with low Wnt-5a expression and a worse histological grade; this suggests that the aberrant expression of β-catenin might be due to the post-translational dysregulation of the upstream Wnt-5a gene in the tumour cells. Moreover, Wang et al. [38] and Joo et al. [39] stressed that the non-nuclear overexpression of β-catenin appeared to be a frequent finding in the predominantly HBV-associated HCC, and that this contributes to tumour progression by stimulating tumour cell proliferation. In agreement with this notion, our immunohistochemical staining results indicate that the nuclear accumulation of β-catenin (7.9%) was not the major event in HCC. E-cadherin is a major β -catenin-binding protein that may also act as a tumour suppressor. Previous studies have indicated that downregulation of E-cadherin is associated with chronic HBV infection [40] , and is especially related to the action of HBxAg protein [41–43] . Our data show that reduced membranous immunostaining of E-cadherin is frequent in HCC tissue (60.5%), which is associated with aberrant β-catenin expression, reduced or lost expression of Wnt-5a, a worse histological grade, and a T4 tumour stage. Based on previously reported data [17,19,44] , it is likely that E-cadherin and Wnt-5a have an upstream regulatory function over β-catenin in HCC; in hepatocarcinogenesis, reduced E-cadherin and Wnt-5a expressions frequently induce β-catenin redistribution through the Wnt signing pathway and the cell adhesion molecule pathway. In summary, the Wnt-5a mRNA and protein levels are overexpressed in hepatitis and cirrhosis and should be related to HBV infection. In HCC, Wnt-5a protein expression is frequently lost, which supports the notion that this protein has a tumour suppressor function in HCC. The lost expression of Wnt-5a may be associated with the aberrant expression of β-catenin and may represent a rescue mechanism for the control of β-catenin dependent pathways in HCC. Practice points • The functions of Wnt-5a in human cancers are controversial and remain to be defined. • Wnt-5a protein might have a tumour suppressor function in hepatocellular carcinoma. • Detection of Wnt-5a protein could serve as a biomarker for early diagnosis of hepatocellular carcinoma. Research agenda • A future study with a large sample size of fresh hepatocellular carcinoma tissues will validate the current findings. • Understanding the molecular mechanism by which the downregulated expression of Wnt-5a protein contributes to hepatocellular carcinoma carcinogenesis. Conflict of interest statement None declared. Acknowledgements This project was granted financial support from China Postdoctoral Science Foundation (No. 20060390941) and Jiangsu Planned Projects for Postdoctoral Research Foundation (No. 0601024B). References [1] F. Staib S.P. Hussain L.J. Hofseth X.W. Wang C.C. Harris TP53 and liver carcinogenesis Hum Mutat 21 2003 201 206 [2] H.B. El-Serag Hepatocellular carcinoma: an epidemiologic view J Clin Gastroenterol 35 2002 S72 S78 [3] K. Breuhahn S. Vreden R. Haddad S. Beckebaum D. Stippel P. Flemming Molecular profiling of human hepatocellular carcinoma defines mutually exclusive interferon regulation and insulin-like growth factor II overexpression Cancer Res 64 2004 6058 6064 [4] J. Zhang W.L. Wang Q. Li Q. Qiao Expression of transforming growth factor-alpha and hepatitis B surface antigen in human hepatocellular carcinoma tissues and its significance World J Gastroenterol 10 2004 830 833 [5] K. Breuhahn T. Longerich P. Schirmacher Dysregulation of growth factor signaling in human hepatocellular carcinoma Oncogene 25 2006 3787 3800 [6] Y. Idobe Y. Murawaki Y. Kitamura H. Kawasaki Expression of transforming growth factor-beta 1 in hepatocellular carcinoma in comparison with the non-tumour tissue Hepatogastroenterology 50 2003 54 59 [7] T. Reya H. Clevers Wnt signalling in stem cells and cancer Nature 434 2005 843 850 [8] F. Rijsewijk M. Schuermann E. Wagenaar P. Parren D. Weigel R. Nusse The Drosophila homolog of the mouse mammary oncogene int-1 is identical to the segment polarity gene wingless Cell 50 1987 649 657 [9] M.T. Veeman J.D. Axelrod R.T. Moon A second canon. Functions and mechanisms of beta-catenin-independent Wnt signaling Dev Cell 5 2003 367 377 [10] T.C. He A.B. Sparks C. Rago H. Hermeking L. Zawel L.T. da Costa Identification of c-MYC as a target of the APC pathway Science 281 1998 1509 1512 [11] B. Mann M. Gelos A. Siedow M.L. Hanski A. Gratchev M. Ilyas Target genes of beta-catenin-T cell-factor/lymphoid-enhancer-factor signaling in human colorectal carcinomas Proc Natl Acad Sci U S A 96 1999 1603 1608 [12] T. Yamada A.S. Takaoka Y. Naishiro R. Hayashi K. Maruyama C. Maesawa Transactivation of the multidrug resistance 1 gene by T-cell factor 4/beta-catenin complex in early colorectal carcinogenesis Cancer Res 60 2000 4761 4766 [13] T. Zhang T. Otevrel Z. Gao Z. Gao S.M. Ehrlich J.Z. Fields Evidence that APC regulates survivin expression: a possible mechanism contributing to the stem cell origin of colon cancer Cancer Res 61 2001 8664 8667 [14] R. Karim G. Tse T. Putti R. Scolyer S. Lee The significance of the Wnt pathway in the pathology of human cancers Pathology 36 2004 120 128 [15] J. Taipale P.A. Beachy The Hedgehog and Wnt signalling pathways in cancer Nature 411 2001 349 354 [16] V. Korinek N. Barker K. Willert M. Molenaar J. Roose G. Wagenaar Two members of the Tcf family implicated in Wnt/beta-catenin signaling during embryogenesis in the mouse Mol Cell Biol 18 1998 1248 1256 [17] Y. Kanai S. Ushijima A.M. Hui A. Ochiai H. Tsuda M. Sakamoto The E-cadherin gene is silenced by CpG methylation in human hepatocellular carcinomas Int J Cancer 71 1997 355 359 [18] G. Christofori Changing neighbours, changing behaviour: cell adhesion molecule-mediated signalling during tumour progression EMBO J 22 2003 2318 2323 [19] Y. Hayashida K. Honda M. Idogawa Y. Ino M. Ono A. Tsuchida E-cadherin regulates the association between beta-catenin and actinin-4 Cancer Res 65 2005 8836 8845 [20] M. Kurayoshi N. Oue H. Yamamoto M. Kishida A. Inoue T. Asahara Expression of Wnt-5a is correlated with aggressiveness of gastric cancer by stimulating cell migration and invasion Cancer Res 66 2006 10439 10448 [21] R.V. Iozzo I. Eichstetter K.G. Danielson Aberrant expression of the growth factor Wnt-5A in human malignancy Cancer Res 55 1995 3495 3499 [22] A.T. Weeraratna Y. Jiang G. Hostetter K. Rosenblatt P. Duray M. Bittner Wnt5a signaling directly affects cell motility and invasion of metastatic melanoma Cancer Cell 1 2002 279 288 [23] T. Pukrop F. Klemm T. Hagemann D. Gradl M. Schulz S. Siemes Wnt5a signaling is critical for acrophage-induced invasion of breast cancer cell lines Proc Natl Acad Sci U S A 103 2006 5454 5459 [24] L. Topol X. Jiang H. Choi L. Garrett-Beal P.J. Carolan Y. Yang Wnt-5a inhibits the canonical Wnt pathway by promoting GSK-3-independent β-catenin degradation J Cell Biol 162 2003 899 908 [25] T. Ishitani S. Kishida J. Hyodo-Miura N. Ueno J. Yasuda M. Waterman The TAK1-NLK mitogen-activated protein kinase cascade functions in the Wnt-5a/Ca 2+ pathway to antagonize Wnt/b-catenin signaling Mol Cell Biol 23 2003 131 139 [26] A.J. Mikels R. Nusse Purified Wnt5a protein activates or inhibits b-catenin-TCF signaling depending on receptor context PLoS Biol 4 2006 570 582 [27] D.J. Olson D.M. Gibo Antisense wnt-5a mimics wnt-1 mediated C57MG mammary epithelial cell transformation Exp Cell Res 241 1998 134 141 [28] H. Liang Q. Chen A.H. Coles S.J. Anderson G. Pihan A. Bradley Wnt5a inhibits B cell proliferation and functions as a tumour suppressor in hematopoietic tissue Cancer Cell 4 2003 349 360 [29] N. Kremenevskaja R. von Wasielewski A.S. Rao C. Schöfl T. Andersson G. Brabant Wnt-5a has tumour suppressor activity in thyroid carcinoma Oncogene 24 2005 2144 2154 [30] J. Dejmek A. Dejmek A. Säfholm A. Sjölander T. Andersson Wnt-5a protein expression in primary dukes B colon cancers identifies a subgroup of patients with good prognosis Cancer Res 65 2005 9142 9146 [31] S. Hirohashi H.E. Blum K.G. Ishak Tumours of the liver and intrahepatic bile ducts S.R. Hamilton L.A. Aaltonen Pathology and genetics. Tumours of the digestive system. World Health Organisation Classification of Tumours 2000 IARC Press Lyon 157 202 [32] M. Jonsson T. Andersson Repression of Wnt-5a impairs DDR1 phosphorylation and modifies adhesion and migration of mammary cells J Cell Sci 114 2001 2043 2053 [33] C.S. Rhee M. Sen D. Lu C. Wu L. Leoni J. Rubin Wnt and frizzled receptors as potential targets for immunotherapy in head and neck squamous cell carcinomas Oncogene 21 2002 6598 6605 [34] T.D. Bui G. Tortora F. Ciardiello A.L. Harris Expression of Wnt5a is downregulated by extracellular matrix and mutated c-Ha-ras in the human mammary epithelial cell line MCF-10A Biochem Biophys Res Commun 239 1997 911 917 [35] M. Jonsson J. Dejmek P.O. Bendahl T. Andersson Loss of Wnt-5a protein is associated with early relapse in invasive ductal breast carcinomas Cancer Res 62 2002 409 416 [36] G. Liu A. Bafico S.A. Aaronson The mechanism of endogenous receptor activation functionally distinguishes prototype canonical and noncanonical Wnts Mol Cell Biol 25 2005 3475 3482 [37] X. He M. Semenov K. Tamai X. Zeng LDL receptor related proteins 5 and 6 in Wnt/b-catenin signaling: arrows point the way Development 131 2004 1663 1677 [38] C.M. Wong S.T. Fan I.O. Ng Beta-Catenin mutation and overexpression in hepatocellular carcinoma: clinicopathologic and prognostic significance Cancer 92 2001 136 145 [39] M. Joo H.K. Lee Y.K. Kang Expression of beta-catenin in hepatocellular carcinoma in relation to tumour cell proliferation and cyclin D1 expression J Korean Med Sci 18 2003 211 217 [40] Y. Wei J.T. Van Nhieu S. Prigent P. Srivatanakul P. Tiollais M.A. Buendia Altered expression of E-cadherin in hepatocellular carcinoma: correlations with genetic alterations, beta-catenin expression, and clinical features Hepatology 36 2002 692 701 [41] E. Lara-Pezzi P.L. Majano M. Yáñez-Mó M. Gómez-Gonzalo M. Carretero R. Moreno-Otero Effect of the hepatitis B virus HBx protein on integrin-mediated adhesion to and migration on extracellular matrix J Hepatol 34 2001 409 415 [42] E. Lara-Pezzi J.M. Serrador M.C. Montoya D. Zamora M. Yáñez-Mó M. Carretero The hepatitis B virus X protein (HBx) induces a migratory phenotype in a CD44-dependent manner: possible role of HBx in invasion and metastasis Hepatology 33 2001 1270 1281 [43] Z. Lian J. Liu L. Li X. Li N.L. Tufan M. Clayton Upregulated expression of a unique gene by hepatitis B x antigen promotes hepatocellular growth and tumourigenesis Neoplasia 5 2003 229 244 [44] J. Liu Z. Lian S. Han M.M.Y. Waye H. Wang M.C. Wu Downregulation of E-cadherin by hepatitis B virus X antigen in hepatocellullar carcinoma Oncogene 25 2006 1008 1017