Heterogeneity of hepatocellular malignant phenotype

Abstract

Hepatocellular carcinoma (HCC) is one of the most wide-spread carcinomas throughout the world - responsible for more than 600,000 deaths annually - and is strongly associated with several etiological factors; including aflatoxin B1, alcohol, and Hepatitis virus B and C. In HCC, many genes undergo somatic aberrations with a tendency to cluster at genes involved in cell cycle regulation, in the p53 and canonical Wnt signaling pathways, and in the TGF-β/IGF axis. Almost a third of HCCs display mutations affecting canonical Wnt signalling. However, the role of canonical Wnt signaling aberrations in HCC is not known in detail, since transgenic mice which express mutant β-catenin (an integral component of canonical Wnt signaling) do not develop liver tumours. To study the heterogeneity of hepatocellular malignancies, we have concentrated on canonical Wnt signaling in HCC cell lines. We have found that canonical Wnt signaling was active in 80% of well-differentiated, and 14% of poorly-differentiated cell lines respectively. Furthermore, ectopic expression of a mutant β-catenin resulted in strong canonical Wnt activity in well-differentiated, but not in poorly-differentiated HCC cells. Our findings suggested that heterogeneity in HCC exists even in the same pathway as exemplified by differential canonical Wnt signaling activity in well- and poorlydifferentiated HCC cell lines. During this study, we produced monoclonal antibodies against β-catenin to distinguish between the pools of nuclear/cytoplasmic and membraneassociated β-catenin in cells, since it is believed that the nuclear β-catenin pool is more potent in tumorigenesis. Monoclonal antibody (MAb), 4C9 recognised β-catenin out of adherens junctions, while another MAb, 9E10 recognised all β-catenin forms even though their epitopes were adjacent. Differential Wnt signaling activity in HCC cell lines prompted us to investigate the interactions between β-catenin and other molecules, which have important functions in hepatocytes and may affect β-catenin/TCF transcriptional activity. C/EBPα is a potent inhibitor of cell proliferation in HCC cell lines and is involved in liver-specific gene expression, and some somatic alterations of it have been observed in AML and HCC. We investigated the effect of C/EBPα on β-catenin signalling. We have found that C/EBPα inhibits mutant β-catenin-TCF transcriptional activity, and physically interacts with β- catenin in HCC cells. While we were analyzing some stably mock-transfected Huh7 clones to use as controls, we observed heterogeneity in their proliferation rates. Further analysis of these clones revealed that some clones ceased to proliferate when passaged extensively. One of these clones (C3) was not tumorigenic in immunodeficient mice. Based on these observations, we hypothesized that some cancer cells could produce senescencent progeny in cell culture. Indeed, we showed that breast- and liver-cancer-derived cells display senescencent phenotypes at variable ratios. By using our experimental system, we also showed that replicative senescence program may work independently of functional p53 and p16 pathways, and the SIP1 gene is partially responsible for replicative senescent phenotypes in our Huh7-derived senescent clone C3. Overexpression of mouse SIP1 in p53- and Rb-deficient Hep3B cells induced partial senescent phenotypes at early passages. However, stable Hep3B cells repressed mouse SIP1 expression by an unknown mechanism and escaped senescent arrest in late passages. Our results suggest that Wnt pathway may have a dual role in hepatocellular malignancy, as it is active/easily inducible in well-differentiated HCC cells and inactive/repressed in poorly-differentiated ones. The further study of β-catenin in tumor samples by using our monoclonal antibodies may reveal new aspects in β-catenin signaling. However, the mechanism of these phenomena and the inhibition of β-cateninTCF signaling by C/EBPα require more study to reach a more comprehensive conclusion. The study of reprogramming of replicative senescence in HCC-derived cells indicated that senescence program may work independent of p53 and p16 pathways and heterogeneity of hepatocellular malignancy exists even within the established HCC derived cell lines.