**Mechanisms of HBx Mediated Liver Cancer: Multiple Pathways and Opportunities**

Mark A. Feitelson et al.\*

*Department of Biology, Temple University, Philadelphia, PA USA* 

#### **1. Introduction**

286 Hepatocellular Carcinoma – Basic Research

Traub O, Look J, Paul D, Willecke K. Cyclic adenosine monophosphate stimulates

Vinken M, Decrock E, Leybaert L, Bultynck G, Himpens B, Vanhaecke T, Rogiers V. Non-

Zhang M, Thorgeirsson SS. Modulation of Connexins during Differentiation of Oval Cell

mouse hepatocytes. Eur J Cell Biol43:48-54, 1987.

into Hepatocytes. Exp Cell Res 213:37-42, 1994.

Acta. 2011 (on line).

biosynthesis and phosphorylation of the 26 kDa gap junction protein in cultured

channel functions of connexins in cell growth and cell death. Biochim Biophys

Chronic hepatitis B virus (HBV) infection is associated with a high risk for the development of chronic liver diseases (CLDs) which include hepatitis, cirrhosis and hepatocellular carcinoma (HCC). HCC is among the top five most prevalent tumor types worldwide, has few effective treatment options, and is highly lethal. The pathogenesis of CLD and HCC is immune mediated, and the virus has developed a number of defense mechanisms that essentially prevent infected cells from being effectively eliminated by the immune system. This, in part, involves the sustained, high level expression of the virus encoded protein, hepatitis B x antigen (HBx). Recent work has shown that HBx blocks pathways of innate immunity (Kumar et al., 2011; Wei et al., 2010), thereby blunting the development of adaptive immunity that is central to virus elimination. In addition, HBx inhibits immune mediated apoptosis by multiple pathways, including those mediated by Fas and tumor necrosis factor alpha (TNF). In this context, HBx has been shown to up-regulate TNF expression (Lara-Pezzi et al., 1998), which is thought to kill uninfected hepatocytes more readily than infected cells, thereby promoting expansion of the virus within the liver, since virus infected hepatocytes would preferentially regenerate following a bout of chronic hepatitis. HBx also switches the growth signals mediated by elevated transforming growth factor beta 1 (TGF1) from that of negative growth regulation to that of positive growth regulation. TGF1 is a transcriptional target of HBx (Yoo et al., 1996), suggesting that HBx expression in the liver promotes fibrogenesis and the development of cirrhosis. Within the infected hepatocyte, HBx blocks the action of tumor suppressors, such as p53 and Rb (Feitelson et al., 2008), and up-regulates the expression of selected host genes that strongly promote hepatocarcinogenesis even in the absence of HBx (see below). Recent work has also

*1Department of Biology, Temple University, Philadelphia, PA, USA* 

*2Department of Medical Biology, Pamukkale University School of Medicine, Kinikli Denizli, Turkey* 

<sup>\*</sup> Alla Arzumanyan1, Tiffany Friedman1, N. Lale Tufan2, Zhaorui Lian1, Marcia M. Clayton1, Joyce Kang3, Helena M. G. P. V. Reis4, Jingbo Pan5, Jie Liu6, Patrick Arbuthnot7 and Michael Kew7

*<sup>3</sup>Division of Medical Microbiology, Guiyang Medical College, Guizhou Province, People's Republic of China 4MIT Portugal Program, Av. Antonio Jose de Almeida, Lisboa, Portugal 5Department of Pathology, Anatomy and Cell Biology, Thomas Jefferson University, Philadelphia, PA, USA 6Department of Digestive Diseases of Huashan Hospital, Fudan University, Shanghai, P. R. China 7Molecular Hepatology Research Unit, Department of Medicine, University of the Witwatersrand, Johannesburg,* 

*South Africa* 

Mechanisms of HBx Mediated Liver Cancer: Multiple Pathways and Opportunities 289

have been observed in cirrhotic nodules (Wang et al., 1991a, 1991b). As indicated above, HBx *trans*-activates HBV enhancers and promoters, thereby promoting long term virus replication. However, it is proposed that when the levels of intracellular HBx increase with time among patients with CLD, it *trans*-regulates the expression of many cellular genes as well (Balsano et al., 1994; Twu & Schloemer, 1987) by a variety of mechanisms. It is postulated that these changes in cellular gene expression help to make cells more permissive to continued virus replication, but also protect the cells from immune responses aimed at removal of infected hepatocytes. This is accomplished by triggering EMT (Du et al., 2010; Yang et al., 2009), by promoting up-regulated expression of selected oncogene associated pathways, and by turning off tumor suppressor, senescence and apoptotic pathways (Kew, 2011; Oishi et al., 2007; Park et al., 2011; Xu et al., 2010) that are often activated by immune responses against virus infected cells. The fact that HBx promotes cell cycle progression and cell growth (Feitelson et al., 2005), means that when this happens in normal hepatocytes, negative growth regulatory (senescence and tumor suppressor) pathways are triggered to reestablish homeostasis. The latter may underlie the putative "proapoptotic" properties of HBx observed in cell lines and *in vivo*, even though there is a considerable literature showing that HBx is also "anti-apoptotic" (Assrir et al., 2010). In this model, it is proposed that apoptosis is a cellular response to inappropriate growth stimulatory signals in the liver mediated by HBx during chronic infection and not due to an inherent property of HBx. Although there is considerable literature suggesting that HBx inhibits several DNA repair systems (e.g., Cheng et al., 2010; Martin-Lluesma et al., 2008; Mathonnet et al., 2004; Qadri et al, 2011), which would promote the development of mutations in the liver prior to the appearance of tumors, it appears that a major contribution of HBx to the pathogenesis of CLD is epigenetic. This is because many natural effectors of HBx correlate with HBx expression in chronically infected human livers and because mutations are not widespread in preneoplastic hepatocytes (Feitelson et al., 2002). The finding that HBx and its natural effectors (target genes) correlate in nontumor liver, but are mostly absent from adjacent tumor tissues, suggests that HBx and its target genes drive pathogenesis prior to the appearance of tumor, but are no longer rate limiting once tumors appear. In the latter case, it is proposed that epigenetic mechanisms mediated by HBx are replaced by genetic mechanisms that are independent of HBx. If so, then HBx may play a predominant role in

the pathogenesis of CLD, but a more modest role in tumor progression.

Early work characterized HBx as a *trans*-regulatory protein that was initially shown to upregulate the expression of almost every target gene that was evaluated using mostly reporter gene assays in transient transfected cell lines (Rossner, 1992). It seemed that in order to better understand what HBx was doing *in vivo*, the natural effectors and targets of HBx in the infected liver had to be identified and characterized. HBx targets that were up- or downregulated were identified by microarray analysis, miRNA arrays, chromatin immunoprecipitation, and by other techniques (e.g., Hu et al., 2006; Sung et al., 2009; Wu et al., 2001, 2002). Some of the targets include telomerase (Liu et al, 2010), the ras pathway signaling molecule, RASSF1A (Yang, et al, 2010), the metastasis associated protein, MTA (Bui-Nguyen et al, 2010), -catenin (Lian et al., 2006; Pan et al., 2007), E-cadherin (Liu et al., 2006), c-myc (Wu et al., 2001), and DNA methyltransferase 1 (Zheng et al., 2009). HBx is a protein binding protein that also regulates gene expression by activating a number of signal

**3. Natural targets of HBx** 

shown that HBx promotes phenotypic changes in hepatocytes characteristic of epithelial-tomesenchymal transition (EMT). One of the molecular hallmarks of EMT, down-regulated expression of the cell adhesion molecule, E-cadherin, is blocked by sustained HBx expression via several mechanisms (Feitelson et al., 2009). HBx also overrides immune mediated apoptotic signals by constitutively activating key signaling pathways, such as nuclear factor kappa B (NF-B), which is known to be hepatoprotective (Beg et al., 1995, 1996), and phosphatidylinositol 3-kinase (PI3K)/Akt, which is known to promote growth in many tumor types (Chung et al., 2004). The finding that HBx stabilizes -catenin by a variety of mechanisms, and up-regulates ErbB2 (Liu et al., 2009), further underscores the importance of these actions in maintaining hepatocellular growth and survival required for virus propagation during the many years and decades that span chronic infection. Unfortunately, these same pathways are also those that contribute centrally to the development of HCC. This body of work provides many opportunities for the development of diagnostic markers that form a fingerprint of those chronically infected patients who are most likely to go on and develop HCC. These markers will serve as therapeutic targets for the repositioning of known drugs for this new indication, and/or the discovery of new drugs that will target rate limiting pathways during multi-step carcinogenesis. In doing so, this work proposes that the chemoprevention of cancer, instead of the treatment of tumor bearing patients, is worth pursuing, and could likely reduce or eliminate the morbidity and mortality associated with chronic HBV infection long before tumors appear. This represents an important challenge, since the knowledge gained will identify cause and effect relationships important for the identification of definitive biomarkers and pharmacological targets that participate decisively in tumorigenesis.

#### **2. Relationship between HBx expression and the pathogenesis of CLD and (HCC): A model**

HBx is one of four genes expressed by HBV during infection, and is known to have gene regulatory functions. Truncated envelope polypeptides that appear during chronic infection may also regulate gene expression and contribute to the pathogenesis of CLD and liver cancer (Chen et al., 2006; Lauer et al., 1992), but their contributions are less well characterized. HBx has been defined as a *trans*-activating protein that promotes virus gene expression and replication during infection (Belloni et al., 2009; Spandau & Lee, 1988; Tsuge et al., 2010). Experimental infection of newborn woodchucks with the related woodchuck hepatitis virus (WHV) results in the development of carriers in nearly 100% of cases, and most of these go on to develop severe chronic hepatitis and HCC (Tennant & Gerin, 2001). However, infection of neonatal woodchucks with an X protein negative clone of WHV failed to establish the chronic carrier state (Chen et al., 1993; Zoulim et al., 1994). This suggests that X protein promotes viremia. The impact of X protein on virus gene expression and replication is also supported by considerable *in vitro* data (Benhenda et al., 2009; Keasler et al., 2009; Tsuge et al., 2010). During the course of CLD, bouts of hepatitis are associated with hepatocellular destruction and regeneration. Among infected cells, the X open reading frame (ORF), which is at the end of the virus genome, becomes repeatedly integrated into host DNA at the replication forks that exist in host DNA during regeneration. This suggests that the intracellular levels of HBx increase with the severity and progression of CLD, and there is now considerable experimental evidence to support this hypothesis (Feitelson et al., 1993a; Jin et al, 2001; Wang et al., 1991a, 1991b). In fact, the highest levels of HBx expression have been observed in cirrhotic nodules (Wang et al., 1991a, 1991b). As indicated above, HBx *trans*-activates HBV enhancers and promoters, thereby promoting long term virus replication. However, it is proposed that when the levels of intracellular HBx increase with time among patients with CLD, it *trans*-regulates the expression of many cellular genes as well (Balsano et al., 1994; Twu & Schloemer, 1987) by a variety of mechanisms. It is postulated that these changes in cellular gene expression help to make cells more permissive to continued virus replication, but also protect the cells from immune responses aimed at removal of infected hepatocytes. This is accomplished by triggering EMT (Du et al., 2010; Yang et al., 2009), by promoting up-regulated expression of selected oncogene associated pathways, and by turning off tumor suppressor, senescence and apoptotic pathways (Kew, 2011; Oishi et al., 2007; Park et al., 2011; Xu et al., 2010) that are often activated by immune responses against virus infected cells. The fact that HBx promotes cell cycle progression and cell growth (Feitelson et al., 2005), means that when this happens in normal hepatocytes, negative growth regulatory (senescence and tumor suppressor) pathways are triggered to reestablish homeostasis. The latter may underlie the putative "proapoptotic" properties of HBx observed in cell lines and *in vivo*, even though there is a considerable literature showing that HBx is also "anti-apoptotic" (Assrir et al., 2010). In this model, it is proposed that apoptosis is a cellular response to inappropriate growth stimulatory signals in the liver mediated by HBx during chronic infection and not due to an inherent property of HBx. Although there is considerable literature suggesting that HBx inhibits several DNA repair systems (e.g., Cheng et al., 2010; Martin-Lluesma et al., 2008; Mathonnet et al., 2004; Qadri et al, 2011), which would promote the development of mutations in the liver prior to the appearance of tumors, it appears that a major contribution of HBx to the pathogenesis of CLD is epigenetic. This is because many natural effectors of HBx correlate with HBx expression in chronically infected human livers and because mutations are not widespread in preneoplastic hepatocytes (Feitelson et al., 2002). The finding that HBx and its natural effectors (target genes) correlate in nontumor liver, but are mostly absent from adjacent tumor tissues, suggests that HBx and its target genes drive pathogenesis prior to the appearance of tumor, but are no longer rate limiting once tumors appear. In the latter case, it is proposed that epigenetic mechanisms mediated by HBx are replaced by genetic mechanisms that are independent of HBx. If so, then HBx may play a predominant role in the pathogenesis of CLD, but a more modest role in tumor progression.

#### **3. Natural targets of HBx**

288 Hepatocellular Carcinoma – Basic Research

shown that HBx promotes phenotypic changes in hepatocytes characteristic of epithelial-tomesenchymal transition (EMT). One of the molecular hallmarks of EMT, down-regulated expression of the cell adhesion molecule, E-cadherin, is blocked by sustained HBx expression via several mechanisms (Feitelson et al., 2009). HBx also overrides immune mediated apoptotic signals by constitutively activating key signaling pathways, such as nuclear factor kappa B (NF-B), which is known to be hepatoprotective (Beg et al., 1995, 1996), and phosphatidylinositol 3-kinase (PI3K)/Akt, which is known to promote growth in many tumor types (Chung et al., 2004). The finding that HBx stabilizes -catenin by a variety of mechanisms, and up-regulates ErbB2 (Liu et al., 2009), further underscores the importance of these actions in maintaining hepatocellular growth and survival required for virus propagation during the many years and decades that span chronic infection. Unfortunately, these same pathways are also those that contribute centrally to the development of HCC. This body of work provides many opportunities for the development of diagnostic markers that form a fingerprint of those chronically infected patients who are most likely to go on and develop HCC. These markers will serve as therapeutic targets for the repositioning of known drugs for this new indication, and/or the discovery of new drugs that will target rate limiting pathways during multi-step carcinogenesis. In doing so, this work proposes that the chemoprevention of cancer, instead of the treatment of tumor bearing patients, is worth pursuing, and could likely reduce or eliminate the morbidity and mortality associated with chronic HBV infection long before tumors appear. This represents an important challenge, since the knowledge gained will identify cause and effect relationships important for the identification of definitive biomarkers and pharmacological

**2. Relationship between HBx expression and the pathogenesis of CLD and** 

HBx is one of four genes expressed by HBV during infection, and is known to have gene regulatory functions. Truncated envelope polypeptides that appear during chronic infection may also regulate gene expression and contribute to the pathogenesis of CLD and liver cancer (Chen et al., 2006; Lauer et al., 1992), but their contributions are less well characterized. HBx has been defined as a *trans*-activating protein that promotes virus gene expression and replication during infection (Belloni et al., 2009; Spandau & Lee, 1988; Tsuge et al., 2010). Experimental infection of newborn woodchucks with the related woodchuck hepatitis virus (WHV) results in the development of carriers in nearly 100% of cases, and most of these go on to develop severe chronic hepatitis and HCC (Tennant & Gerin, 2001). However, infection of neonatal woodchucks with an X protein negative clone of WHV failed to establish the chronic carrier state (Chen et al., 1993; Zoulim et al., 1994). This suggests that X protein promotes viremia. The impact of X protein on virus gene expression and replication is also supported by considerable *in vitro* data (Benhenda et al., 2009; Keasler et al., 2009; Tsuge et al., 2010). During the course of CLD, bouts of hepatitis are associated with hepatocellular destruction and regeneration. Among infected cells, the X open reading frame (ORF), which is at the end of the virus genome, becomes repeatedly integrated into host DNA at the replication forks that exist in host DNA during regeneration. This suggests that the intracellular levels of HBx increase with the severity and progression of CLD, and there is now considerable experimental evidence to support this hypothesis (Feitelson et al., 1993a; Jin et al, 2001; Wang et al., 1991a, 1991b). In fact, the highest levels of HBx expression

targets that participate decisively in tumorigenesis.

**(HCC): A model** 

Early work characterized HBx as a *trans*-regulatory protein that was initially shown to upregulate the expression of almost every target gene that was evaluated using mostly reporter gene assays in transient transfected cell lines (Rossner, 1992). It seemed that in order to better understand what HBx was doing *in vivo*, the natural effectors and targets of HBx in the infected liver had to be identified and characterized. HBx targets that were up- or downregulated were identified by microarray analysis, miRNA arrays, chromatin immunoprecipitation, and by other techniques (e.g., Hu et al., 2006; Sung et al., 2009; Wu et al., 2001, 2002). Some of the targets include telomerase (Liu et al, 2010), the ras pathway signaling molecule, RASSF1A (Yang, et al, 2010), the metastasis associated protein, MTA (Bui-Nguyen et al, 2010), -catenin (Lian et al., 2006; Pan et al., 2007), E-cadherin (Liu et al., 2006), c-myc (Wu et al., 2001), and DNA methyltransferase 1 (Zheng et al., 2009). HBx is a protein binding protein that also regulates gene expression by activating a number of signal

Mechanisms of HBx Mediated Liver Cancer: Multiple Pathways and Opportunities 291

relationship between HBx, inflammation, and fibrogenesis seen in earlier studies (Wang et al., 1991a, 1991b). In this context, hepatic inflammation, fibrosis and cell death were demonstrated in TGF1 transgenic mice (Sanderson et al., 1995), underscoring the contribution of elevated TGF1 expression to CLD. Interestingly, HBx also blocks TGF1 mediated growth inhibition and apoptosis, in part, through the up-regulation of PI3K (Shih et al., 2000), suggesting that HBx may confer resistance to TGF1 mediated growth inhibition, while uninfected cells remain sensitive, thereby favoring survival of virus infected hepatocytes. These observations are consistent with the strong correlation between HBx staining and the progression of CLD among HBV infected carriers (Jin et al, 2001; Wang

Fig. 1. Proposed model of how HBx may contribute to the development of cirrhosis. See the

This close relationship is exemplified by the observations that HBx activation of NF-B resulted in the stimulation of the fibronectin (FN) promoter (Figure 1), and that liver tissue samples from chronically infected patients showed a strong correlation between HBx and FN mRNA in hepatocytes from fibrotic and cirrhotic livers (Norton et al., 2004). In this context, the fact that HBx binds to and inactivates the tumor suppressor protein, p53, both *in vitro* and *in vivo* (Feitelson et al., 1993b; Ueda et al., 1995), and that p53 normally suppresses the FN promoter, suggest that inactivation of p53 also results in increased FN production. Interestingly, up-regulation of FN in HBx expressing cells also showed a modest (50%) decrease in adherence to FN (Lara-Pezzi et al., 2001a, 2001b) and depressed expression of the FN receptor, 51 integrin. There was also an observed decrease in the levels of collagen/laminin receptor 1 subunit in HBx positive compared to negative cells

et al., 1991a, 1991b).

text for details.

**3.1.2 Fibronectin (FN)** 

transduction pathways in the cytoplasm (e.g., NF-B, PI3K/Akt, JAK/STAT, PKC, AP-1, ras, src, Wnt and others) (Feitelson & Duan, 1997; Henkler & Koshy, 1996; Kew, 2011). Constitutive activation of these signaling pathways has been identified with up-regulated expression of specific target genes. For example, HBx mediated activation of the mitogenactivated protein kinase (MAPK) pathway has been shown to up-regulate the expression of hypoxia-inducible factor-1 alpha (HIF-1) (Yoo et al., 2003), which promotes the survival of hepatocytes in cirrhotic nodules, where a hypoxic environment is known to exist during CLD. Further, HBx mediated constitutive activation of Wnt signaling is associated with upregulated expression of c-myc and cyclin-D1, both of which promote hepatocellular growth. In the nucleus, HBx interacts with the basal transcription machinery (Haviv, et al., 1995, 1996), binds to the transcriptional scaffolds CBP/p300 (Cougot et al., 2007) and mSin3a (Arzumanyan et al., 2011), and alters the extent of DNA methylation and histone acetylation (Zheng et al., 2009). Further, there is increasing evidence that HBx alters the expression of host gene expression by up- or down-regulating selected miRNAs (Kong et al., 2011; Wu, et al., 2011). In many cases, the natural targets of these epigenetic changes have not been identified. The importance of doing so will provide both prognostic markers and therapeutic targets relevant to the pathogenesis of CLD and HCC, thus providing opportunities for earlier intervention.

#### **3.1 HBx and fibrogenesis**

#### **3.1.1 Transforming growth factor beta 1 (TGF1)**

The close association between intrahepatic expression of HBx and the severity of CLD suggests that HBx may take a part in driving pathogenesis. TGF1 is an important mediator of fibrosis and apoptosis in carriers with CLD (Castilla et al., 1991; Liu et al., 1999), as indicated by the direct correlation between serum TGF1 levels, elevated aminotransferases, and fibrosis scored in liver biopsy specimens (Flisiak et al., 2004). HBx has been shown to transcriptionally up-regulate the expression of TGF1 both in cell cultures and in HBx transgenic mice (Martin-Vilchez et al., 2008; Norton et al., 2004; Yoo et al., 1996). In liver tissue with HBx protein expression, phospho-Smad2 was detectable, suggesting a functional link between viral protein expression and TGF-1 signaling. Phospho-Smad2 staining correlated significantly with fibrotic stage in patients with HBV infection and steatosis/steatohepatitis (Weng et al., 2009). HBx mediated up-regulation of TGF1 was further potentiated by suppressed expression of the natural inhibitor of TGF1, alpha-2 macroglobulin (2M, Figure 1) (Pan et al., 2004). HBx may suppress 2M gene expression by either activation of NF-B, which then blocks the activation of the 2M gene by STAT3, and/or by the HBx activation of PI3K, which then blocks 2M expression. Independent work showed that HBx also shifted TGF1 signaling from tumor suppression to tumor promotion in the livers of patients with chronic hepatitis B, and that this involved differential phosphorylation of smad3 *in vivo* (Murata et al., 2009). HBx was also shown to enhance TGF signaling by stabilizing a protein complex consisting of smad4 and components of the basic transcriptional machinery (Lee et al., 2001). The fact that HBx stimulates multiple signal transduction pathways (e.g., NF-B, PI3K, MAPK, Wnt, ras, src, etc), combined with altered smad signaling, also appear to override the homeostatic and growth inhibitory properties of TGF1. This results in the development of a strong profibrogenic environment in the liver (Akhurst, 2002) which may underlie the close relationship between HBx, inflammation, and fibrogenesis seen in earlier studies (Wang et al., 1991a, 1991b). In this context, hepatic inflammation, fibrosis and cell death were demonstrated in TGF1 transgenic mice (Sanderson et al., 1995), underscoring the contribution of elevated TGF1 expression to CLD. Interestingly, HBx also blocks TGF1 mediated growth inhibition and apoptosis, in part, through the up-regulation of PI3K (Shih et al., 2000), suggesting that HBx may confer resistance to TGF1 mediated growth inhibition, while uninfected cells remain sensitive, thereby favoring survival of virus infected hepatocytes. These observations are consistent with the strong correlation between HBx staining and the progression of CLD among HBV infected carriers (Jin et al, 2001; Wang et al., 1991a, 1991b).

Fig. 1. Proposed model of how HBx may contribute to the development of cirrhosis. See the text for details.

#### **3.1.2 Fibronectin (FN)**

290 Hepatocellular Carcinoma – Basic Research

transduction pathways in the cytoplasm (e.g., NF-B, PI3K/Akt, JAK/STAT, PKC, AP-1, ras, src, Wnt and others) (Feitelson & Duan, 1997; Henkler & Koshy, 1996; Kew, 2011). Constitutive activation of these signaling pathways has been identified with up-regulated expression of specific target genes. For example, HBx mediated activation of the mitogenactivated protein kinase (MAPK) pathway has been shown to up-regulate the expression of hypoxia-inducible factor-1 alpha (HIF-1) (Yoo et al., 2003), which promotes the survival of hepatocytes in cirrhotic nodules, where a hypoxic environment is known to exist during CLD. Further, HBx mediated constitutive activation of Wnt signaling is associated with upregulated expression of c-myc and cyclin-D1, both of which promote hepatocellular growth. In the nucleus, HBx interacts with the basal transcription machinery (Haviv, et al., 1995, 1996), binds to the transcriptional scaffolds CBP/p300 (Cougot et al., 2007) and mSin3a (Arzumanyan et al., 2011), and alters the extent of DNA methylation and histone acetylation (Zheng et al., 2009). Further, there is increasing evidence that HBx alters the expression of host gene expression by up- or down-regulating selected miRNAs (Kong et al., 2011; Wu, et al., 2011). In many cases, the natural targets of these epigenetic changes have not been identified. The importance of doing so will provide both prognostic markers and therapeutic targets relevant to the pathogenesis of CLD and HCC, thus providing

The close association between intrahepatic expression of HBx and the severity of CLD suggests that HBx may take a part in driving pathogenesis. TGF1 is an important mediator of fibrosis and apoptosis in carriers with CLD (Castilla et al., 1991; Liu et al., 1999), as indicated by the direct correlation between serum TGF1 levels, elevated aminotransferases, and fibrosis scored in liver biopsy specimens (Flisiak et al., 2004). HBx has been shown to transcriptionally up-regulate the expression of TGF1 both in cell cultures and in HBx transgenic mice (Martin-Vilchez et al., 2008; Norton et al., 2004; Yoo et al., 1996). In liver tissue with HBx protein expression, phospho-Smad2 was detectable, suggesting a functional link between viral protein expression and TGF-1 signaling. Phospho-Smad2 staining correlated significantly with fibrotic stage in patients with HBV infection and steatosis/steatohepatitis (Weng et al., 2009). HBx mediated up-regulation of TGF1 was further potentiated by suppressed expression of the natural inhibitor of TGF1, alpha-2 macroglobulin (2M, Figure 1) (Pan et al., 2004). HBx may suppress 2M gene expression by either activation of NF-B, which then blocks the activation of the 2M gene by STAT3, and/or by the HBx activation of PI3K, which then blocks 2M expression. Independent work showed that HBx also shifted TGF1 signaling from tumor suppression to tumor promotion in the livers of patients with chronic hepatitis B, and that this involved differential phosphorylation of smad3 *in vivo* (Murata et al., 2009). HBx was also shown to enhance TGF signaling by stabilizing a protein complex consisting of smad4 and components of the basic transcriptional machinery (Lee et al., 2001). The fact that HBx stimulates multiple signal transduction pathways (e.g., NF-B, PI3K, MAPK, Wnt, ras, src, etc), combined with altered smad signaling, also appear to override the homeostatic and growth inhibitory properties of TGF1. This results in the development of a strong profibrogenic environment in the liver (Akhurst, 2002) which may underlie the close

opportunities for earlier intervention.

**3.1.1 Transforming growth factor beta 1 (TGF1)** 

**3.1 HBx and fibrogenesis** 

This close relationship is exemplified by the observations that HBx activation of NF-B resulted in the stimulation of the fibronectin (FN) promoter (Figure 1), and that liver tissue samples from chronically infected patients showed a strong correlation between HBx and FN mRNA in hepatocytes from fibrotic and cirrhotic livers (Norton et al., 2004). In this context, the fact that HBx binds to and inactivates the tumor suppressor protein, p53, both *in vitro* and *in vivo* (Feitelson et al., 1993b; Ueda et al., 1995), and that p53 normally suppresses the FN promoter, suggest that inactivation of p53 also results in increased FN production. Interestingly, up-regulation of FN in HBx expressing cells also showed a modest (50%) decrease in adherence to FN (Lara-Pezzi et al., 2001a, 2001b) and depressed expression of the FN receptor, 51 integrin. There was also an observed decrease in the levels of collagen/laminin receptor 1 subunit in HBx positive compared to negative cells

Mechanisms of HBx Mediated Liver Cancer: Multiple Pathways and Opportunities 293

also known to promote tissue remodeling in the liver (Omenetti & Diehl, 2008), it is possible that this may contribute to the progression and formation of cirrhotic nodules in the liver of chronically infected patients. Preliminary data also suggests that HBx activates hedgehog signaling in liver cancer cells (Kim et al., 2011), although the role of this activation in hepatocarcinogenesis remains to be studied. Further, it is not clear whether the upregulation of hedgehog ligands is activated by HBx, and whether this in some way

Subtractive hybridization of mRNAs from HBx positive compared to negative human hepatoblastoma (HepG2) cells yielded a set of differentially expressed mRNAs that revealed additional mechanisms whereby HBx contributes to the pathogenesis of HCC. Several unique mRNAs were identified by subtractive hybridization, and among them were a number of previously uncharacterized transcripts. One of them, URG7, encoded a 99 amino acid polypeptide with no distinguishing functional motifs (Lian et al., 2001), was found to down-regulate the expression of the TGF1 inhibitor, 2M (Figure 1), suggesting that it contributes to the development and progression of fibrosis. It appears to do so by activation of PI3K, by stabilization of -catenin, and by blocking the activities of caspase 8 and 3 (Pan et al., 2007) (Figures 1 and 2). Among its many activities, HBx also activates PI3K (Lee et al., 2001), stabilizes -catenin (Lian et al., 2006), and blocks caspase 3 (Gottlob et al., 1998), suggesting that these functions may be carried out by URG7. Further data showed that both HBx and URG7 activated fragments of the β-catenin promoter, and also promoted expression of β-catenin target genes. These include c-myc (Terradillos et al., 1997), multidrug resistance gene 1 (MDR1) (Doong et al., 1998) and cyclin D1 (Park et al., 2006). While the activation of -catenin target genes by URG7 suggests that the latter promotes tumor formation, URG7 did not promote growth of HepG2 cells in soft agar, nor did it accelerate the outgrowth of HepG2 based tumors in SCID mice (Lian et al., 2001). Its role in blocking apoptosis, however, is shared with that of -catenin (Chen et al., 2001). Importantly, one of the major characteristics of tumor cells is resistance to immune mediated apoptosis. The finding that URG7 is over-expressed in infected liver, but not in HCC cells from clinical specimens, suggests that resistance to apoptosis precedes the development of tumor, and that it probably protects HBV infected cells from immune damage and elimination. On the molecular level, caspase 8, which is just up-stream of caspase 3, transmits death signals from Fas (T cell) and from TNF signaling (Figure 2). In this context, it had previously been shown that HBx blocks Fas mediated killing in primary human hepatocytes (Diao et al., 2001), which may actually be mediated by URG7. Further, the finding that HBx activates NF-B (Su & Schneider, 1996), that activated NF-B protects hepatocytes from cell death (Beg et al., 1995, 1996), and that NF-B transcriptionally activates URG7 (Pan et al., 2001), suggest a pathway that promotes persistence of the carrier state (and sustained HBV replication) even in the presence of recurring immune responses spanning many years. The findings of elevated TNF production in human hepatocytes infected with HBV, and that HBx targets this up-regulation (Lara-Pezzi et al, 1998), not only suggests that TNF is a target for HBx, but is also consistent with the strong correlation between HBx expression and inflammatory liver disease (Jin et al., 2001; Wang et al., 1991a, 1991b).

contributes to fibrogenesis.

**3.2 HBx up-regulated genes in chronically infected liver** 

**3.2.1 Up-regulated gene, clone 7 (URG7)** 

(Lara-Pezzi et al., 2001a), suggesting that HBx promotes the detachment of infected cells from the extracellular matrix (ECM). This detachment was associated with increased cell migration, indicating that changes in the ECM-cell relationship probably also contributed to alterations in tissue morphology that accompany the development of cirrhosis. Since activated ras and src signaling depress 51 expression (Varner et al., 1995), that HBx stimulates ras and src signaling (Klein & Schneider, 1997), and that HBx disrupts adherens junctions in a src dependent manner (Lara-Pezzi et al., 2001b), it is likely that the activation of these signaling pathways by HBx contribute importantly to decreased integrin expression, decreased cell adhesion, and an increased propensity for cell migration and loss of tissue morphology in the infected liver, and to metastasis in already established tumors.

#### **3.1.3 Lysyl hydroxylase (LH3)**

As indicated above, the accumulation and remodeling of ECM is central to the development of fibrosis and cirrhosis. In this context, the finding that HBx up-regulates the expression of the enzyme, lysyl hydroxylase 3 (LH3) in liver cells, and that LH3 co-stains with HBx in livers of HBV infected patients (unpublished observations), suggests another mechanism whereby an HBx target gene may contribute to fibrosis (Figure 1). LH3 mediates the chemical cross-linking of several collagen and collagen-like molecules (Myullyla et al., 2007). This may promote stabilization of the ECM during chronic infection. Given that LH3 knockout mice with disrupted formation of basement membranes during embryogenesis resulted in embryonic lethality (Myullyla et al., 2007), the over-expression of LH3 during chronic HBV infection may promote the development and persistence of basement membranes that are characteristic of fibrosis. This would sever the intimate relationship between hepatocytes and the bloodstream observed in normal livers. Although LH3 is associated with the endoplasmic reticulum, it has also been found in the extracellular space and in serum (Salo et al., 2006), implying that LH3 serum levels may be elevated in the blood prior to the development of HCC.

#### **3.1.4 Does HBx activate stellate cells?**

It is also possible that HBx expression promotes stellate cell activation. Although there is little evidence that HBV infects stellate cells, when HBx was transfected into a human stellate cell line, it promoted proliferation and up-regulated expression of fibrosis related molecules (Guo et al., 2009). Independent work showed that HBx expressing hepatocytes induced paracrine activation of human and rat hepatic stellate cells. When these cells were exposed to conditioned medium from HBx-expressing hepatocytes, they showed increased expression of collagen I, connective tissue growth factor, alpha smooth muscle actin, matrix metalloproteinase-2, and TGF, together with an enhanced proliferation rate (Martin-Vilchez et al., 2008). More recently, hedgehog signaling and ligand production have been demonstrated to be activated in clinical samples from HBV (and hepatitis C virus) infected patients These ligands promoted the *in vitro* expansion of liver myofibroblasts, activated endothelial cells, and progenitors expressing markers of tumor stem/initiating cells (Pereira et al., 2010). Independent data has shown that hedgehog signaling is profibrogenic, in that it promotes activation and EMT in quiescent hepatic stellate cells (Choi et al., 2009), and in the context of cholestatic liver injury (Omenetti et al., 2011). Given that hedgehog signaling is also known to promote tissue remodeling in the liver (Omenetti & Diehl, 2008), it is possible that this may contribute to the progression and formation of cirrhotic nodules in the liver of chronically infected patients. Preliminary data also suggests that HBx activates hedgehog signaling in liver cancer cells (Kim et al., 2011), although the role of this activation in hepatocarcinogenesis remains to be studied. Further, it is not clear whether the upregulation of hedgehog ligands is activated by HBx, and whether this in some way contributes to fibrogenesis.

#### **3.2 HBx up-regulated genes in chronically infected liver**

#### **3.2.1 Up-regulated gene, clone 7 (URG7)**

292 Hepatocellular Carcinoma – Basic Research

(Lara-Pezzi et al., 2001a), suggesting that HBx promotes the detachment of infected cells from the extracellular matrix (ECM). This detachment was associated with increased cell migration, indicating that changes in the ECM-cell relationship probably also contributed to alterations in tissue morphology that accompany the development of cirrhosis. Since activated ras and src signaling depress 51 expression (Varner et al., 1995), that HBx stimulates ras and src signaling (Klein & Schneider, 1997), and that HBx disrupts adherens junctions in a src dependent manner (Lara-Pezzi et al., 2001b), it is likely that the activation of these signaling pathways by HBx contribute importantly to decreased integrin expression, decreased cell adhesion, and an increased propensity for cell migration and loss of tissue morphology in the infected liver, and to metastasis in already

As indicated above, the accumulation and remodeling of ECM is central to the development of fibrosis and cirrhosis. In this context, the finding that HBx up-regulates the expression of the enzyme, lysyl hydroxylase 3 (LH3) in liver cells, and that LH3 co-stains with HBx in livers of HBV infected patients (unpublished observations), suggests another mechanism whereby an HBx target gene may contribute to fibrosis (Figure 1). LH3 mediates the chemical cross-linking of several collagen and collagen-like molecules (Myullyla et al., 2007). This may promote stabilization of the ECM during chronic infection. Given that LH3 knockout mice with disrupted formation of basement membranes during embryogenesis resulted in embryonic lethality (Myullyla et al., 2007), the over-expression of LH3 during chronic HBV infection may promote the development and persistence of basement membranes that are characteristic of fibrosis. This would sever the intimate relationship between hepatocytes and the bloodstream observed in normal livers. Although LH3 is associated with the endoplasmic reticulum, it has also been found in the extracellular space and in serum (Salo et al., 2006), implying that LH3 serum levels may

It is also possible that HBx expression promotes stellate cell activation. Although there is little evidence that HBV infects stellate cells, when HBx was transfected into a human stellate cell line, it promoted proliferation and up-regulated expression of fibrosis related molecules (Guo et al., 2009). Independent work showed that HBx expressing hepatocytes induced paracrine activation of human and rat hepatic stellate cells. When these cells were exposed to conditioned medium from HBx-expressing hepatocytes, they showed increased expression of collagen I, connective tissue growth factor, alpha smooth muscle actin, matrix metalloproteinase-2, and TGF, together with an enhanced proliferation rate (Martin-Vilchez et al., 2008). More recently, hedgehog signaling and ligand production have been demonstrated to be activated in clinical samples from HBV (and hepatitis C virus) infected patients These ligands promoted the *in vitro* expansion of liver myofibroblasts, activated endothelial cells, and progenitors expressing markers of tumor stem/initiating cells (Pereira et al., 2010). Independent data has shown that hedgehog signaling is profibrogenic, in that it promotes activation and EMT in quiescent hepatic stellate cells (Choi et al., 2009), and in the context of cholestatic liver injury (Omenetti et al., 2011). Given that hedgehog signaling is

established tumors.

**3.1.3 Lysyl hydroxylase (LH3)** 

be elevated in the blood prior to the development of HCC.

**3.1.4 Does HBx activate stellate cells?** 

Subtractive hybridization of mRNAs from HBx positive compared to negative human hepatoblastoma (HepG2) cells yielded a set of differentially expressed mRNAs that revealed additional mechanisms whereby HBx contributes to the pathogenesis of HCC. Several unique mRNAs were identified by subtractive hybridization, and among them were a number of previously uncharacterized transcripts. One of them, URG7, encoded a 99 amino acid polypeptide with no distinguishing functional motifs (Lian et al., 2001), was found to down-regulate the expression of the TGF1 inhibitor, 2M (Figure 1), suggesting that it contributes to the development and progression of fibrosis. It appears to do so by activation of PI3K, by stabilization of -catenin, and by blocking the activities of caspase 8 and 3 (Pan et al., 2007) (Figures 1 and 2). Among its many activities, HBx also activates PI3K (Lee et al., 2001), stabilizes -catenin (Lian et al., 2006), and blocks caspase 3 (Gottlob et al., 1998), suggesting that these functions may be carried out by URG7. Further data showed that both HBx and URG7 activated fragments of the β-catenin promoter, and also promoted expression of β-catenin target genes. These include c-myc (Terradillos et al., 1997), multidrug resistance gene 1 (MDR1) (Doong et al., 1998) and cyclin D1 (Park et al., 2006). While the activation of -catenin target genes by URG7 suggests that the latter promotes tumor formation, URG7 did not promote growth of HepG2 cells in soft agar, nor did it accelerate the outgrowth of HepG2 based tumors in SCID mice (Lian et al., 2001). Its role in blocking apoptosis, however, is shared with that of -catenin (Chen et al., 2001). Importantly, one of the major characteristics of tumor cells is resistance to immune mediated apoptosis. The finding that URG7 is over-expressed in infected liver, but not in HCC cells from clinical specimens, suggests that resistance to apoptosis precedes the development of tumor, and that it probably protects HBV infected cells from immune damage and elimination. On the molecular level, caspase 8, which is just up-stream of caspase 3, transmits death signals from Fas (T cell) and from TNF signaling (Figure 2). In this context, it had previously been shown that HBx blocks Fas mediated killing in primary human hepatocytes (Diao et al., 2001), which may actually be mediated by URG7. Further, the finding that HBx activates NF-B (Su & Schneider, 1996), that activated NF-B protects hepatocytes from cell death (Beg et al., 1995, 1996), and that NF-B transcriptionally activates URG7 (Pan et al., 2001), suggest a pathway that promotes persistence of the carrier state (and sustained HBV replication) even in the presence of recurring immune responses spanning many years. The findings of elevated TNF production in human hepatocytes infected with HBV, and that HBx targets this up-regulation (Lara-Pezzi et al, 1998), not only suggests that TNF is a target for HBx, but is also consistent with the strong correlation between HBx expression and inflammatory liver disease (Jin et al., 2001; Wang et al., 1991a, 1991b).

Mechanisms of HBx Mediated Liver Cancer: Multiple Pathways and Opportunities 295

with URG7, there was extensive co-staining between HBx and URG11 in chronically infected liver (Lian et al., 2006) but not in tumor. This suggests that URG11 promotes hepatocellular growth prior to the appearance of HCC. The ability of URG11 specific siRNA to block the growth of liver tumor cells both *in vitro* and *in vivo*, not only underscores the importance of elevated URG11 to cell growth, but also suggests that it may be a novel target for the development of specific therapeutics against HCC (Fan et al., 2011). Independent work has recently shown that URG11 was induced under hypoxic conditions in human kidney tubule cells (Du et al., 2010). The latter was associated with increased levels of HIF-1, which is also known to be a target of HBx (Holotnakova et al., 2010). Importantly, HIF-1 is known to *trans*-activate VEGF *in vivo* (Yoo et al., 2003), suggesting that neovascularization may occur in cirrhotic nodules prior to the appearance of HCC. If this occurs during the pathogenesis of chronic hepatitis B, it would most likely be observed in cirrhotic nodules, since this represents a hypoxic environment characterized by high levels of HBx expression (Wang et al., 1991a, 1991b). Interestingly, elevated expression of URG11 in kidney tubule cells was also associated with suppression of E-cadherin, and upregulation of the mesenchymal markers vimentin and alpha-SMA, suggesting that URG11 is associated with EMT. In chronic HBV infection, the development of cirrhosis is accompanied by considerable alterations in the tissue architecture within the liver, implying that URG11 may also play a significant role in tissue remodeling during the

**3.2.3 Elevated vascular endothelial growth factor receptor 3 (VEGFR-3)** 

Vascular endothelial growth factor receptor 3 (VEGFR-3), which is associated with angiogenesis, is a receptor tyrosine kinase that is expressed in lymphatic endothelial cells (Iljin et al., 2001). Binding of VEGFR-3 to the ligands VEGF-C or VEGF-D stimulate lymphangiogenesis (Alitalo & Carmeliet, 2002), while in carcinogenesis, the production of VEGFs by tumors promote metastases and result in decreased survival (Su et al., 2006). Elevated VEGF has been found in patients with HCC (Dahr et al., 2002, Poon et al., 2003). VEGFR-3 is also expressed in tumor cells from several tumor types (Bando et al., 2004, Su et al., 2006), including HCC (Dahr et al., 2002), implying the existence of an autocrine/paracrine loop that promotes tumor development independent of lymphangiogenesis (Su et al., 2006). In HCC, elevated VEGFR-3 is associated with portal vein invasion of tumors, increased hepatic tumor recurrence, and shorter survival (Dhar et al., 2002), suggesting that VEGFR-3 is important in the pathogenesis of HCC. In this context, differential display of HBx positive compared to negative cells showed that HBx upregulated the expression of an mRNA which encoded a splice variant of VEGFR-3 (Lian et al., 1997). This was verified at the mRNA and protein levels in HBx positive compared to negative HepG2 cells. In infected liver, expression of VEGFR-3 was prominent in nodules of HCC and correlated with HBx expression. VEGFR-3 stimulated cell cycle in culture, anchorage independent growth in soft agar, and accelerated tumor formation and larger tumor size in SCID mice injected with HepG2 cells over-expressing VEGFR-3. Further work showed that over-expression of VEGFR-3 in the absence of HBx resulted in activation of PI3K/Akt, which then activated -catenin gene expression (Figure 2), and with inactivation of the tumor suppressor, PTEN. Interestingly, HBx also mediates these changes, suggesting that they may be actually carried out by up-regulation of VEGFR-3. These findings also suggest that in addition to lymphangiogenesis, VEGFR-3 may promote tumorigenesis in

pathogenesis of chronic infection.

HBx associated HCC.

Additionally, the observation that HBx activates the expression of Fas ligand in HCC cell lines (Shin et al., 1999), may provide a way for virus infected cells to escape direct T cell killing by inducing apoptosis in such T cells. This would not only promote chronicity, but in tumor cells, an escape from immune elimination.
