Clinical factors contributing to the therapeutic challenge of hepatocellular carcinoma (HCC) are manifold: tumors arise often in patients with compromised liver function, therefore limiting therapeutic options; symptoms develop only at later stages of tumor progression, and tumors tend to invade normal structures or occur in multiple locations simultaneously. Ninety percent of patients with tumors larger than 5 cm will have synchronous intrahepatic metastases at the time of presentation. In the majority of patients undergoing partial hepatectomy for HCC, intra-hepatic or distant metastases will occur. Likewise, in about fifty percent of the patients who die within five years after liver transplantation for HCC, intra- and extrahepatic recurrences are the cause of death. Tumor characteristics predicting recurrence of HCC following resection include portalvein invasion, intrahepatic metastasis, extratumoral spread, high mitotic index [3-5], and a sarcomatous phenotype. Such observations demonstrate that tumor cells that are easily invading normal tissues and achieve access to the circulation harbor the greatest risk of tumor recurrence. The molecular events promoting invasiveness of HCC cells are still widely unknown. Further understanding of these processes is urgently needed for the development of rational strategies for prevention and treatment of metastatic disease in HCC.
Investigations in different tumor types and murine models of cancer depict a characteristic cascade of events necessary for a tumor to achieve successful dissemination of metastases from a primary tumor. After malignant transformation and initial tumor growth, the first step towards metastasis formation is the initiation of vascularization of the tumor, a process known as the angiogenic switch. While the tumor progresses, new genetic changes occur leading to ① a reduced cell-cell adhesion, and, ② alterations in the interaction of the tumor cell with the extracellular matrix, allowing invasion into surrounding tissues including blood vessels (which are mediated by changes in integrin expression and activation of proteolytic enzymes). Tumor cells reaching the circulation must have developed mechanisms to suppress anoikis, which is programmed cell death caused by disruption of cell-substrate adhesion. At distant sites, tumor cells need to leave the circulation and migrate to sites with favorable conditions. Here, angiogenesis is again initiated and cell progression of the metastasis continues. It is obvious that this sequence of events will be dependent on multiple molecular factors in malignant and normal (e.g. stroma and immune competent) cells. Some of these factors have been characterized in HCC. So far, and in agreement with data from other tumor types, there is evidence that a high proliferative index (as measured by PCNA or Ki-67 expression) is associated with reduced tumor differentiation and might be predictive of recurrence after tumor resection[7-9]. In HCC, several signalingpathways are candidates for driving this proliferation, mainly those related to growth factor receptors, including Met, the receptor for the hepatocyte growth factor/scatter facter (HGF/SF) and epidermal growth factor receptor (EGFR)[10,11]. While enhanced proliferation is an important step during tumorigenesis, increases in invasiveness and cell mobility are necessary for the generation of metastases. It has been demonstrated that signaling through the Ras/MAPK pathway (which confers signals from growth factor receptors, including EGFR and Met), could cooperate with the TGF-β signal transduction pathway in promoting the switch from a epithelial to mesenchymal phenotype in cancer cells, rendering them significantly more mobile and invasive. In vitro data from HCC cell lines support this possibility, as it has been demonstrated that TGF-β signaling enhances proliferation of HCC cell lines. Disassembly of protein complexes involved in cell-cell adhesion has been demonstrated as a consequence of TGF-β and Ras/MAPK signaling. Accordingly, loss of expression of the cell-adhesion molecule E-cadherin and nuclear expression of its binding partner β-catenin, occurs frequently in high-grade HCC and correlates with a poor prognosis. There is also strong evidence that HCC indeed activates angiogenesis. For example, vascular endothelial growth factor (VEGF), a promoter of angiogenesis, was found to be upregulated in HCC. A recent prospective study in 100 patients undergoing resection of HCC demonstrated a strong correlation between serum concentration of vascular endothelial growth factor (VEGF) and microscopic vascular invasion. In the same study, high serum levels were associated also with the presence of intrahepatic metastases and reduced disease-free survival. As indicated, successful invasion into surrounding tissues can also be facilitated by proteases that allow malignant cells to migrate through physiologic barriers, e.g. basement membranes. Indeed, experiments using HCC cell lines revealed that membrane-type 1 matrix metal loproteinase confer invasiveness and promote intrahepatic metastases. In addition, expression of urokinase-type plasminogen activator (uPA), uPA receptor (uPAR) and plasminogen activator inhibitor type-1 (PAI-1) was found to be associated with tumor invasiveness in a mouse model of HCC and human HCC. Overall, these data suggest that mechanisms of metastasis in HCC that are consistent with data in other types of solid tumors. However, most of the evidence is circumstantial and correlative, reflecting a lack of appropriate model systems.
As Li et al describe in this issue of the World Journal of Gastroenterology, new model systems might be available that could lend insight into the mechanisms underlying the metastatic process in HCC. The authors generated two clonally related cell line derivatives from human HCC cell line MHCC97 that show dramatic differences in their metastatic potential. Notably, the authors used an in vivo selection process that included the orthotopic transplantation of xenografts into the livers of mice, thus simulating the clinical situation as closely as possible. The differences in metastatic potential were mirrored by significant phenotypical differences. These included morphologic differences as well as faster proliferation, enhanced in vitro invasiveness, and increased alpha-fetoprotein (AFP) production of the highly malignant derivative compared to its less metastatic counterpart. These findings are consistent with the expected metastatic phenotype of H CC and it will be most exciting to see results from molecular analyses of these cells. Particularly, analyses of differential gene expression, e.g. using cDNA expression arrays, will be revealing and could be compared to results from the analysis expression profiles in human tumors.
Identification of differentially expressed genes represents an important step toward understanding metastasis in this disease. However, genetic evidence will be necessary to pinpoint the culprits in this process. So far, transgenic mouse models of HCC have been created in which liver-specific expression of potential oncogenes, including Cyclin D1, c-myc, TGF-a, and Met leads consistently to the development of HCC-like tumors[21,11,22]. Interestingly, tumors occurring in these models lack a highly invasive or metastatic phenotype. A next step is to rebuild the full phenotype of HCC in mice using modified transgenic models that include combinations of growth and metastasis promoting genes. Based on these models, rational therapeutic approaches could be developed and tested.