Epithelial-mesenchymal transitions and hepatocarcinogenesis

Abstract

Epithelial-mesenchymal transitions (EMTs) are believed to play a role in invasion and metastasis of many types of tumors. In this issue of the JCI, Chen et al. show that a gene that has been associated with aggressive biology in hepatocellular carcinomas initiates a molecular cascade that results in EMT.

Hepatocellular carcinoma (HCC) is the fifth most commonly diagnosed, and the third most deadly, cancer worldwide (1). In the United States between 1975 and 2006, HCC was the only cancer with increasing mortality in men and women (2). Surveillance in patients at increased risk for developing HCC has been recommended for detection of early HCC (3); however, this is underutilized, and many patients present with locally advanced or metastatic disease. These patients are not liver transplantation candidates and have limited therapeutic options. Thus, there is an urgent need for identification of patients at risk for HCC and further risk stratification of these patients to improve the outcomes. Therefore, groups like Chen et al. (4) are applying gene profiling to identify better prognostic and/or therapeutic targets. Their present work, in this issue of the JCI, reveals that the oncogene chromodomain helicase/ATPase DNA binding protein 1–like gene (CDH1L) is commonly amplified in HCC (4).

Chen et al. demonstrated why CDH1L amplification is important in HCC (Figure 1 and ref. 4). CDH1L functions as a transcription factor that induces expression of the guanine nucleotide exchange factor (GEF) ARHGEF9. GEFs catalyze exchange of GDP for GTP on RhoGTPases, a family of GTPases that includes Rho and Rac proteins as well as Cdc42 (5). Chen et al. demonstrated that Cdc42 was the target of ARHGEF9 and that increased ARHGEF9 enhanced formation of Cdc42-GTP (i.e., activated Cdc42) in HCC cells (4). Cdc42 collaborates with other Rho kinases to modulate assembly and disassembly of the actin cytoskeleton and has been shown to regulate establishment of cell polarity, filopodia formation, and centrosome reorientation in migrating cells. Protein kinase cascades that control cell viability and proliferative activity, including JNK, p38MAPK, and p21-associated protein kinases (PAK), are also downstream targets of Cdc42-GTP (5). Activating Cdc42 induced HCC cells to undergo changes in the actin cytoskeleton, filopodia formation, adherens junction disruption, and cell migration, which suggests that generating invasive cancer cells in damaged livers involves induction of an epithelial-mesenchymal transition (EMT).

Figure 1

Hepatocarcinogenesis and EMT. In healthy livers, hepatocytes exhibit features of epithelial cells because they are polarized and adherent to each other. However, unlike epithelial cells in other tissues, healthy hepatocytes lack a basement membrane. Hence, healthy, mature hepatocytes are not physically separated from the adjacent stroma and hepatic stellate cells that reside in the space of Disse, sandwiched between the hepatic epithelia and fenestrated endothelial cells that line hepatic sinusoids. Indeed, because the sinusoids themselves also lack a typical basement membrane, endothelial cells, hepatic stellate cells, and hepatocytes all coexist within the same mesenchyme. During hepatocarcinogenesis, genetic events that enhance the migratory capabilities of hepatocytes may occur, and the normal absence of membrane barriers to physically separate hepatocytes from the hepatic sinusoids permits easy entry of motile neoplastic hepatocytes into the vascular system. The current study by Chen et al. (4) delineates a mechanism by which malignant hepatocytes become more motile and invasive. Namely, amplification of CDH1L results in induction of ARHGEF9 and resultant activation of the small GTP-binding protein Cdc42. Cdc42-GTP orchestrates various events, including reorganization of cytoskeletal elements that permit neoplastic hepatocytes to migrate away from neighboring epithelial cells, move through the space of Disse, invade neighboring vascular spaces, and be swept along with the blood flow to other parts of the hepatic lobule and/or out of the liver to extrahepatic sites.

EMT and HCC

EMT is a complex process through which epithelial cells gradually dissolve connections to adjacent cells, rearrange their cytoskeletons, upregulate production of matrix remodeling factors, and migrate out of epithelial sheets into adjacent stroma. EMT occurs routinely during development and contributes to fetal tissue formation. It also occurs during some wound healing responses in adults and is believed to be a major contributing mechanism to cancer invasion and metastasis (6–8). Consistent with this information, the current work demonstrated that CDH1L-driven expression of AFHGEF9, and resultant activation of Cdc42 and EMT, was associated with venous invasion, formation of microsatellite tumors and metastases, and reduced survival from HCC in mice and with increased vascular invasion and reduced survival in HCC patients (4). Thus, CDH1L is one of the factors that promote EMT in HCC.

This discovery compliments and extends growing evidence that implicates EMT in hepatocarcinogenesis. For example, tetraspanin TM4SF5-mediated EMT causes loss of contact inhibition and uncontrollable growth of cultured human hepatocarcinoma cells, as well as tumor formation and/or metastasis in mice (9, 10). The pro-EMT transcription factors Twist and Snail are expressed in about 40%–70% of HCCs and correlate with evidence of adherens junction disruption and a worse prognosis (11, 12). More than 60% of HCCs express eukaryotic initiation factor 5A2 (EIF5A2), with the highest levels occurring in metastatic cancers. EIF5A2 activates Rho kinases to stimulate formation of stress fibers, lamellipodia, EMT, and HCC cell migration (13). The hepatitis B X gene (HBx; ref. 14) and hepatitis C core proteins (15) induce EMT in cultured liver cells, consistent with evidence that chronic viral hepatitis increases the risk for HCC. HBx promotes EMT by activating STAT5b, and the level of STAT5b activation in HCC correlates with advanced tumor stage, venous infiltration, and microsatellite formation (14). Hepatoma cell lines also differ somewhat in their endogenous expression of putative EMT regulators, with the greatest expression occurring in lines expressing more mesenchymal proteins, such as Hep3B (16, 17). The Hedgehog pathway is active in Hep3B cells and is required for growth (18). Hedgehog signaling promotes EMT during development and in many other types of cancers (19). Moreover, pathway activation is common in HCC and correlates with aggressive tumor biology (20). Elevated expression of another EMT mediator, integrin-linked kinase (ILK), identifies hepatoma cell lines that are resistant to epidermal growth factor receptor–targeted therapies, and knocking down ILK restores sensitivity to these agents (21). EMT is also induced in neoplastic hepatocytes by TGF-β, IL-like EMT inducer (ILE1), and oncogenic Ras interactions that promote distant metastasis (22). Similarly, ectopic expression of Twist or Snail in hepatoma cell lines with low endogenous expression of these factors enhances cancer cell motility and invasiveness (12, 17). Thus, several mechanisms appear capable of stimulating EMT in HCC, and, regardless of the specific factor responsible for EMT induction, HCCs with features of EMT consistently demonstrate more vascular invasion, metastases, and a poorer prognosis than do HCCs lacking EMT characteristics.

Evidence demonstrating that EMT correlates with aggressive biology in HCC suggests that EMT conveys certain survival advantages to tumor cells. These include escape from apoptotic stimuli that are prevalent within the primary tumor (TGF-β, hypoxia) and improved access to nutrients and/or growth factors (by migrating to sites remote from the competing forces that are operative in the primary tumor). Cells that have undergone EMT also acquire the ability to generate factors that stimulate production of vasculature and stroma. This property sustains the outgrowth of new epithelial tumor nodules within or near the primary tumor and after the subpopulations of the metastatic (mesenchymal-appearing) tumor cells have arrived in distant sites and undergone mesenchymal-epithelial transition (MET) to reacquire their epithelial phenotype (7). It is therefore tempting to speculate that situations that promote hepatocarcinogenesis stimulate EMT and that liver cells that have acquired pro-EMT modifications, such as CDH1L amplifications, might have a growth advantage in such environments.

Cirrhosis, EMT, and HCC

The single greatest risk factor for HCC in humans is cirrhosis (3). Cirrhosis is a conserved response to various types of chronic liver injury that kill liver epithelial cells (i.e., hepatocytes and cholangiocytes). It is the result of suboptimal repair that is unable to fully regenerate or replace dead liver epithelial cells, thereby maintaining stimuli for tissue remodeling that result in chronic expansion of liver epithelial progenitor populations, ongoing inflammation, angiogenesis, and fibrogenesis. Hence, cirrhosis is a consequence of unsuccessful wound healing. In many tissues (e.g., kidney, skin, and lung), chronic wound healing responses that lead to fibrosis provide a stimulus for EMT (6). However, whether EMT occurs during chronic fibrosing liver injury remains controversial (23, 24).

As recently reviewed elsewhere (25), there is consistent evidence that profibrogenic factors that increase during cirrhosis, such as TGF-β, can induce hepatocytes and cholangiocytes to undergo EMT in culture. Populations of cells that typically accumulate in cirrhotic livers, such as hepatic stellate cells and progenitors, also have features of cells that are undergoing EMT. In addition, EMT-inducing mechanisms become activated during cirrhosis, and various markers of EMT have been well documented in liver tissues of rodents and patients with various types of chronic liver disease. However, rigorous lineage-tracing proof that EMT actually occurs in situ during cirrhosis, and which cells are involved, is still lacking. With regard to the latter point, it is important to emphasize that lineage-tracing studies that failed to demonstrate EMT in certain types of liver cells (26, 27) do not prove that EMT is impossible in other liver cell types, despite recent speculation that a role for EMT in liver repair has been largely refuted (24). Rather, additional research is needed to resolve the debate about non–cancer cell EMT in chronically damaged livers, particularly given the compelling evidence from many groups — including Chen et al. — that malignant liver cells are capable of undergoing EMT in this context.

Indeed, if further research were to prove that EMT occurred more generally during fibrogenic wound healing responses induced by chronic liver injury, this might provide an explanation for the strong association between cirrhosis and primary liver cancer (3). Namely, situations that result in cirrhosis apply chronic selection pressure that confers advantage to the outgrowth of cells maximally successful at undergoing EMT. These may include healthy liver epithelial progenitors as well as progenitors that have acquired chromosomal alterations (such as CDH1L amplifications) that enhance their capacity for EMT. If other growth-dysregulatory mutations develop, the latter cells then become cancer stem/initiating cells that fuel aggressive biology in HCC. Several lines of historical evidence support the concept that EMT is the biological bridge that links cirrhosis and HCC. First, markers of EMT have been detected in the blood of rodents and humans with cirrhosis (28, 29). Second, various events that enhance the efficiency of cancer cell EMT predict cancer invasion, metastasis, and reduced survival in HCC patients. Third, circulating EMT markers are most elevated in HCC patients with metastatic cancer (29, 30). The present study by Chen et al. advances this field by demonstrating that CDH1L-ARHGEF9-Cdc42 signaling plays an important role in regulating liver cell EMT (4). Their data suggest that malignant liver cells that acquire the ability to activate this pathway in a cell-autonomous fashion evade mechanisms that normally control EMT, permitting them to repopulate and potentially overgrow damaged livers. This knowledge, in turn, identifies molecular targets that might be exploited to improve HCC screening, allocation of existing HCC therapies, and development of novel HCC treatments for cirrhotic patients.

Acknowledgments

The authors’ work is supported by NIDDK/NIAAA, NIH, grants RO1 DK077794 and RO1 AA01054.

Address correspondence to: Anna Mae Diehl, Gastroenterology Division, Duke University, Snyderman Building (GSRB1), Suite 1073, LaSalle Street, Durham, NC 27710. Phone: 919.684.2366; Fax: 919.684.4183; E-mail: diehl004@mc.duke.edu.

Footnotes

Conflict of interest: The authors have declared that no conflict of interest exists.

Citation for this article: J Clin Invest. 2010;120(4):1031–1034. doi:10.1172/JCI42615.

See the related article at CHD1L promotes hepatocellular carcinoma progression and metastasis in mice and is associated with these processes in human patients.

References

1. Parkin DM. Global cancer statistics in the year. 2000. Lancet Oncol. 2001;2(9):533–543.
   View this article via: PubMed CrossRef Google Scholar
2. Edwards BK, et al. Annual report to the nation on the status of cancer, 1975-2006, featuring colorectal cancer trends and impact of interventions (risk factors, screeingin, and treatment) to reduce future rates. Cancer. 2010;116(3):544–573.
   View this article via: PubMed CrossRef Google Scholar
3. Bruix J, Sherman M. Management of hepatocellular carcinoma. Hepatology. 2005;42(5):1208–1236.
   View this article via: PubMed CrossRef Google Scholar
4. Chen L, et al. CHD1L promotes hepatocellular carcinoma progression and metastasis in mice and is associated with these processes in human patients. J Clin Invest. 2010;120(4):1178–1191.
   View this article via: JCI Google Scholar
5. Hall A. Rho GTPases and the control of cell behaviour. Biochem Soc Trans. 2005;33(Pt5):891–895.
   View this article via: PubMed CrossRef Google Scholar
6. Kalluri R, Weinberg RA. The basics of epithelial-mesenchymal transition. J Clin Invest. 2009;119(6):1420–1428.
   View this article via: JCI PubMed Google Scholar
7. Acloque H, Adams MS, Fishwick K, Bronner-Fraser M, Nieto MA. Epithelial-mesenchymal transitions: the importance of changing cell state in development and disease. J Clin Invest. 2009;119(6):1438–1449.
   View this article via: JCI PubMed CrossRef Google Scholar
8. Zeisberg M, Neilson EG. Biomarkers for epithelial-mesenchymal transitions. J Clin Invest. 2009;119(6):1429–1437.
   View this article via: JCI PubMed Google Scholar
9. Lee SA, et al. Tetraspanin TM4SF5 mediates loss of contact inhibition through epithelial-mesenchymal transition in human hepatocarcinoma. J Clin Invest. 2008;118(4):1354–1366.
   View this article via: JCI PubMed CrossRef Google Scholar
10. Lee SA, et al. Blockade of four-transmembrane L6 family member 5 (TM4SF5)-mediated tumorigenicity in hepatocytes by a synthetic chalcone derivative. Hepatology. 2009;49(4):1316–1325.
    View this article via: PubMed CrossRef Google Scholar
11. Sun T, et al. Expression and functional significance of Twist1 in hepatocellular carcinoma: Its role in vasculogenic mimicry. Hepatology. 2009;51(2):545–556.
    View this article via: PubMed Google Scholar
12. Yang MH, et al. Comprehensive analysis of the indpeendent effect of twist and shanil in promoting metastasis of hepatocellular carcinoma. Hepatology. 2009;50(5):1464–1474.
    View this article via: PubMed CrossRef Google Scholar
13. Tang DJ, et al. Overexpression of eukaryotic initiation factor 5A2 enhances cell motility and promotes tumor metastasis in hepatocellular carcinoma . [published online ahead of print November 30, 2009]. Hepatology. doi:10.1002/hep.2345.
    View this article via: PubMed CrossRef Google Scholar
14. Lee TK, et al. Signal transducers and activators of transcription 5b activation enhances hepatocellular carcinoma aggressieness through induction of epithelial-mesenchymal transition. Cancer Res. 2006;66(20):9948–9956.
    View this article via: PubMed Google Scholar
15. Battaglia S, et al. Liver cacncer-derived hepatitis C virus core proteins shift TGF-beta responses from tumor suppression to epithelial-mesenchymal transition. PLoS One. 2009;4(2):e4355.
    View this article via: PubMed CrossRef Google Scholar
16. Slany A, et al. Cell characterization by proteom profiling applied to primary hepatoctyes and hepatocyte cell lines Hep-G2 and Hep-3B. J Proteome Res. 2010;9(1):6–21.
    View this article via: PubMed CrossRef Google Scholar
17. Matsuo N, et al. Twist expression promotes migration and invasion in hepatocellular carcinoma. BMC Cancer. 2009;9:240.
    View this article via: PubMed CrossRef Google Scholar
18. Sicklick JK, et al. Dysregulation of the Hedgehog pathway in human hepatocarcinogenesis. Carcinogenesis. 2006;27(4):748–757.
    View this article via: PubMed CrossRef Google Scholar
19. Katoh Y, Katoh M. Hedgehog signaling, epithelial-to-mesenchymal transition and miRNA (review). Int J Mol Med. 2008;22(3):271–275.
    View this article via: PubMed Google Scholar
20. Katoh Y, Katoh M. Hedgehog target genes: mechanisms of carcinogenesis induced by aberrant hedgehog signaling activation. Curr Mol Med. 2009;9(7):873–886.
    View this article via: PubMed CrossRef Google Scholar
21. Fuchs BC, et al. Epithelial-to-mesenchymal transition and integrin-linked kinase mediate sensitivity to epidermal growth factor receptor inhibition in human hepatoma cells. Cancer Res. 2008;68(7):2391–2399.
    View this article via: PubMed CrossRef Google Scholar
22. Lahsnig C, et al. ILE1 requires oncogenic Ras for the epithelial to mesecnhymal transition of hepatocytes and liver carcinoma progression. Oncogene. 2009;28(5):638–650.
    View this article via: PubMed CrossRef Google Scholar
23. Parola M, Pinzani M. Hepatic wound repair. Fibrogenesis Tissue Repair. 2009;2(1):4.
    View this article via: PubMed CrossRef Google Scholar
24. Wells RG. The epithelial-to-mesenchymal transition in liver fibrosis: here today, gone tomorrow? [published online ahead of print January 27, 2010]. Hepatology. doi:10.1002/hep.23529.
    View this article via: PubMed CrossRef Google Scholar
25. Choi SS, Diehl AM. Epithelial-to-mesenchymal transitions in the liver. Hepatology. 2009;50(6):2007–2013.
    View this article via: PubMed CrossRef Google Scholar
26. Sackett SD, et al. Foxl1 is a marker of bipotential hepatic progenitor cells in mice. Hepatology. 2009;49(3):920–929.
    View this article via: PubMed CrossRef Google Scholar
27. Taura K, et al. Hepatocytes do not undergo epithelial-mesenchymal transition in liver fibrosis in mice. [published online ahead of print October 13, 2009]. Hepatology. doi:10.1002/hep.23368.
    View this article via: PubMed CrossRef Google Scholar
28. Witek RP, et al. Liver cell-derived microparticles activate hedgehog signaling and alter gene expression in hepatic endothelial cells. Gastroenterology. 2009;136:320–330.
    View this article via: PubMed CrossRef Google Scholar
29. Min AL, et al. High expression of Snail mRNA in blood from hepatocellular carcinoma patients with extra-hepatic metastasis. Clin Exp Metastasis. 2009;26(7):759–767.
    View this article via: PubMed Google Scholar
30. Murata K, Suzuki H, Okano H, Oyamada T, Yasuda T, Sakamoto A. Hypoxia-induced des-gamma-carboxy prothrombin production in hepatocellular carcinoma. Int J Oncol. 2010;36(1):161–170.
    View this article via: PubMed Google Scholar