Liver Development and Cancer Formation in
Zebrafish
Jeng-Wei Lu,
1,2Yu Hsia,
1Hsiao-Chen Tu,
1,3Yung-Chun Hsiao,
1,3Wan-Yu Yang,
1Horng-Dar Wang
3and Chiou-Hwa Yuh
1,4,5*
Liver is the largest organ in the human body, and it regulates many physi-ological processes. Many studies on liver development in different model organisms have demonstrated that the mechanism of hepatogenesis is conserved in vertebrates. The identification of the genes and regulatory pathways involved in liver formation provides a basis for the diagnosis of liver diseases and therapeutic interventions. Hepatocellular carcinoma is the third leading cause of mortality worldwide. In the last decade, genetic alterations, which include the gain and loss of DNA, as well as mutations and epigenomic changes, have been identified as important factors in liver cancer. Many genetic pathways are dysregulated during carcinogenesis. Here, we review the gene regulatory networks that underlie liver organo-genesis and the dysregulation of these pathways in liver cancer. The genes and pathways involved in hepatogenesis and liver cancer are largely conserved between zebrafish and humans, making this an ideal model organism for the study of this disease. A better understanding of liver development may aid in the development of new diagnostic and ther-apeutic approaches to liver cancer. Birth Defects Research (Part C) 93:157–172, 2011.VC 2011 Wiley-Liss, Inc.
INTRODUCTION
Liver is an important organ that metabolizes dietary molecules and urea, detoxifies toxic compounds, stores glycogen, and exhibits both endocrine and exocrine properties (Lemaigre and Zaret, 2004; Zorn, 2008; Si-Tayeb et al., 2010). As part of its endocrine functions, the liver secretes many hormones, including insulin-like growth fac-tors, angiotensinogen, and throm-bopoietin, as well as serum pro-teins, such as albumin and
apoli-poproteins. The liver secretes bile to aid digestion as part of its exo-crine function. In addition, the liver contains vasculature in the form of a portal vein, hepatic ar-tery, venuoles, and arterioles, which control blood flow and transport molecules to the circula-tory system. The liver also exhibits a regenerative response to injury and an immune response against foreign materials (Si-Tayeb et al., 2010). In the past two decades, scientists have identified the
mechanisms by which the liver performs these functions.
These complicated tasks are performed by different liver cell types that include hepatocytes, cholangiocytes (bile duct cells), endothelial cells, liver sinusoidal endothelial cells, pit cells (natural killer cells), Kupffer cells (macro-phages), and hepatic satellite cells. Although hepatocytes account for 78% of the liver vol-ume, cooperation between differ-ent cell types contributes to liver function (Zorn, 2008; Si-Tayeb et al., 2010). The extremely com-plex liver tissue architecture is crucial for normal liver function. For the past decade, developmen-tal biologists have been studying how the liver differentiates from the endoderm into such a compli-cated organ and how its cells arrange to form its three-dimen-sional architecture.
During the past decade, scien-tists have been studying the em-bryonic development of the liver (hepatogenesis) in the mouse, chick, Xenopus, and zebrafish (Zorn, 2008; Chu and Sadler, 2009; Zorn and Wells, 2009; Si-Tayeb et al., 2010). The collected
REVIEW
VC 2011 Wiley-Liss, Inc.
Birth Defects Research (Part C) 93:157–172 (2011)
1Institute of Molecular and Genomic Medicine, National Health Research Institutes, Taiwan, Republic of China 2Department of Life Sciences, National Central University, Taiwan, Republic of China
3Institute of Biotechnology, National Tsing-Hua University, Taiwan, Republic of China
4Institute of Bioinformatics and Structural Biology, National Tsing-Hua University, Taiwan, Republic of China 5Department of Biological Science and Technology, National Chiao Tung University, Taiwan, Republic of China
Supported by grants from the National Science Council, Taiwan (NSC 97-3112-B-400-008 and NSC 99-2321-B-400-001) and the National Health Research Institutes, Taiwan (MG-094-PP-14, MG-095-PP-08, MG-096-PP-05, and MG-097-PP-07 ) to Dr. Yuh, CH, and the NSC grant (NSC 97-2311-B-007-004-MY3) and NTHU Booster Grant (99N2903E1) to Dr. Wang, HD.
*Correspondence to: Chiou-Hwa Yuh, R2-3033, #35 Keyan Road, Zhunan Town, Miaoli County 350, Taiwan, Republic of China. E-mail: [email protected]
knowledge about hepatogenesis is immense, and the results from these studies demonstrate that there are evolutionarily conserved networks that underlie liver devel-opment in vertebrates. Although there are still some missing pieces in our understanding of liver orga-nogenesis, animal models have become a promising tool for eluci-dating the complete gene regula-tory networks (GRNs) that direct hepatogenesis.
Because the liver is such a vital organ, liver failure is a life-threat-ening condition. Liver diseases include hepatic fibrosis, cirrhosis, hepatitis, hepatocellular carcinoma (HCC), and cholangiocarcinoma. HCC is the fifth most common can-cer and ranks as the third leading cause of mortality worldwide (Rob-erts and Gores, 2005; El-Serag and Rudolph, 2007). Many lines of evi-dence suggest that hepatocarcino-genesis partially recapitulates fetal liver development; both adult can-cer cells and fetal liver cells have the capacity for the hallmarks of cancer: self-renewal, sustaining proliferative signaling, enabling replication immortality, resisting cell death, and creating the micro-environment (Kung et al., 2010; Hanahan and Weinberg, 2011). Many differentiated adult HCCs present a less differentiated pheno-type than normal liver, and similar to fetal liver supports the progeni-tor cell differentiation arrest model (Sell and Leffert, 2008). Many sig-naling pathways, such as the trans-forming growth factor beta (TGF-b) and Wingless (Wnt)/b-catenin sig-naling pathways, play important roles in both liver development and HCC (Tang et al., 2008; Ikegami, 2009). In a normal liver, the pre-cise identities and sources of pro-liferative signals, as well as the mechanisms that control the release of mitogenic signals, remain poorly understood because of their complexity. In contrast, the source of cell survival and proliferative sig-nals during liver development and mitogenic signaling in cancer cells are more fully understood (Lemmon and Schlessinger, 2010; Si-Tayeb et al., 2010; Witsch et al., 2010). Research on liver cell
speci-fication, budding, and differentia-tion during embryogenesis should improve our knowledge and under-standing of pathologic liver condi-tions. Most importantly, a greater understanding of the GRNs involved in liver development should help determine the GRNs of the multi-ple, distinct cell types present dur-ing carcinogenesis.
Zebrafish (Danio rerio) is a popu-lar research model for genetics and developmental biology and is used in a variety of biomedical research fields, including angiogenesis, neu-rogenesis, organogenesis, human diseases, aging, toxicity, pathology, behavior, cancer studies, and drug screening (Spitsbergen and Kent, 2003; Zon and Peterson, 2005; Lieschke and Currie, 2007). Recently, more advanced technolo-gies as discussed below have been developed and applied in zebrafish research to develop a zebrafish HCC model.
In the first section of this review, the basic mechanisms that control liver organogenesis are summar-ized, and subjects of special inter-est (demarcated by subheadings) that illustrate our understanding of these complicated networks follow the first section. In subsequent sec-tions, we address the progress and discoveries made on liver disease and hepatocarcinogenesis over the past decade, with emphasis on the activation of signal transduction pathways during hepatocarcinogen-esis (demarcated by subheadings) due to genomic instability and mutations. Finally, recent advances using zebrafish to study the mecha-nisms of cancer formation and as a drug-screening platform are re-ported. New techniques that have been developed in zebrafish, which provide both an excellent model for liver disease and a bridge between basic science and translational research, are discussed.
GENE REGULATORY NETWORKS UNDERLIE LIVER DEVELOPMENT
In the past decade, studies of liver development in various model organisms have revealed that an
evolutionarily conserved mecha-nism that includes cell origins, transcription factors, and signaling pathways directs the majority of hepatogenesis. The induction sig-nals from the adjacent mesoderm are also conserved (Zorn and Wells, 2009; Si-Tayeb et al., 2010). With this knowledge, he-patic-like tissue can be induced from embryonic stem cells in vitro using the proper signal ligands (Zorn, 2008).
The Developmental Events and Timeline of Liver Organogenesis in Mice and Zebrafish
Endoderm-derived hepatocytes and cholangiocytes constitute 73% of the liver cell population. The major developmental events of liver organogenesis consist of endoderm formation, hepatic specification, liver bud growth, and hepatocyte/ biliary differentiation (left panel of Fig. 1). In mice, liver development proceeds from endoderm pattern-ing durpattern-ing gastrulation and the early somite stages, when differen-tial Wnt and fibroblast growth fac-tor (FGF) signaling is required to induce the formation of the foregut, midgut, and hindgut along the anterior-posterior (A-P) axis. Fate map studies have revealed that the mouse embryonic liver originates from the ventral foregut endoderm at embryonic day 8.0 (e8.0) (Trem-blay and Zaret, 2005). The homeo-domain factor Hhex becomes enriched in the hepatic endoderm by e8.5 in mice. At e9.5, the he-patic endodermal cells delaminate from the epithelium and invade the septum transversum mesenchyme (STM) to form the liver bud. Between e9.5 and e15, the liver bud undergoes tremendous growth with the help of mesenchymal sig-nals, which include FGF, bone mor-phogenetic protein (BMP), hepato-cyte growth factor (HGF), Wnt, TGF-b, and retinoic acid (Zorn, 2008; Nakamura and Nishina, 2009; Zorn and Wells, 2009; Si-Tayeb et al., 2010).
The zebrafish embryonic fate map at the 50% epiboly stage (Kimmel et al., 1990; Woo and Fraser, 1995; Warga and
Nus-slein-Volhard, 1999; Gritsman et al., 2000; Dougan et al., 2003; Keegan et al., 2004) indicates that the formation of the mesendoder-mal lineage is established by the
end of gastrulation, i.e., 10 hr postfertilization (hpf) (Kimmel et al., 1995). Multiple signals influ-ence the specification of the endo-derm, ectoendo-derm, and mesoderm.
Signals that include maternal nu-clear b-catenin, BMP, FGF, Nodal, and Wnt all affect cell fates. The endoderm develops from the four most marginal blastoderm tiers of the late blastula-stage embryo under the primary influence of Nodal signaling (Chan et al., 2009). According to the fate map, endodermal cell lineages are pri-marily derived from the more dor-sal and lateral cells of the blasto-derm margin (Warga and Nus-slein-Volhard, 1999). The endo-derm further differentiates into the pharynx, stomach/intestine, and liver. Liver-specific transcrip-tion factors, such as hhes and prox1, are initially expressed at 24 hpf, and liver development is com-plete by 120 hpf.
Induction Signals and
Transcription Factors Essential for Liver Development
Signals generated by neighboring cells activate signaling pathways and ultimately trigger transcription factors that activate downstream targets. These events further sub-divided the territory of the liver into the specialized cell types present in this complicated organ. GRNs con-trol the exact spatial and temporal expression patterns of all genes and the architecture of the system. In Figure 1, the molecular events that occur during hepatogenesis from fertilization to liver maturation are summarized from previous studies.
Many endoderm-specific tran-scription factors are induced and activated by maternal Nodal and Wnt/b-catenin signals as well as unknown signals emitted from the zebrafish yolk syncytial layer (Chan et al., 2009). Studies on mice, Xenopus, and zebrafish have described a generalized model that shows that a high Nodal level is required to induce the formation of the anterior mes-endoderm and mes-endoderm, whereas lower Nodal levels are sufficient for inducing the mesoderm and posterior tissue.
During gastrulation and somite formation, the endoderm cells elongate and become the gut
Figure 1. Gene regulatory networks for hepatogenesis. On the left panel, events for liver formation and the timeline from mouse or zebrafish are indicated. One the right, from top to bottom shows the gene interactions require for each territory highlighted with a different background color. Genes are indicated as lines with arrows, whereas proteins are illustrated as bubbles. Activation of genes is achieved by signal transduc-tion pathways or other genes.
LIVER DEVELOPMENT AND CANCER FORMATION IN ZEBRAFISH 159
tube. The foregut endoderm receives FGF signals from the neighboring cardiac mesoderm and BMP signals from the STM. Several FGFs (FGF1, FGF2, FGF8, and FGF10) are expressed in the mouse cardiac mesoderm, and knockouts of these FGFs affect liver formation. Although the source of FGF signals in zebrafish is still unclear, the requirement for FGF signaling in hepatic specifica-tion is evolutionarily conserved (Si-Tayeb et al., 2010). In mice, BMP4 is under the control of Gata4 and is expressed in the STM to regulate early hepatic develop-ment; however, other BMP family members, such as BMP2, are also present in the immediate vicinity. The requirement of BMP signals for the induction of hepatic specifi-cation is conserved in zebrafish; nevertheless, the source of the BMP signal originates in the lateral plate mesoderm rather than the STM. The Wnt antagonist (sFRP5) is required to release Wnt repres-sion of Hhex expresrepres-sion in the an-terior endoderm at the hepatic specification stage. Subsequently, the Wnt signal (wnt2bb) expressed in the lateral plate mes-oderm is essential for the onset of the differentiation of hepatic pro-genitor cells in zebrafish, and the requirement for Wnt in promoting hepatogenesis is evolutionarily conserved (Si-Tayeb et al., 2010).
Many evolutionarily conserved transcription factors are expressed in the anterior endoderm, includ-ing gata4, gata6, foxa1, foxa2, hnf1b, prox1, and hhex. An analy-sis of Hhex knockout mouse embryos found that Hhex is required for early bud morphogen-esis and differentiation into hepa-tocytes (Keng et al., 2000; Hunter et al., 2007). Previous studies in mice have shown that GATA4 and GATA6 may positively regulate the expression of Hhex by binding to its promoter (Denson et al., 2000; Zhao et al., 2005). In zebrafish, hhex is expressed in the hepatic bud from 22 to 50 hpf, and knock-downs of hhex with morpholinos perturbed liver development in a dose-dependent manner (Wallace et al., 2001). Overexpression of
the dominant-negative FGF recep-tor between 18 and 26 hpf has been shown to decrease the expression of gata4, gata6, prox1, and ceruloplasmin in zebrafish (Shin et al., 2007). Mutations in lost-a-fin or expression of a domi-nant negative form of the BMP/ activin receptor has been shown to result in reduced expression of hhex and prox1 in the liver regions of zebrafish embryos (Shin et al., 2007).
Following the specification stage, the liver bud undergoes tre-mendous growth in all of the spe-cies in vertebrate. The liver then becomes the major site of fetal hematopoiesis in mice; however, the hepatic vasculature and hema-topoiesis are not essential for zebrafish liver development. Many ligands are expressed in the sur-rounding mesoderm tissue (STM and hepatic mesenchyme in mice), including BMP, FGF (via PI3 ki-nase), HGF, Wnt, TGF-b, and reti-noic acid. The receptors and com-ponents downstream of the signal-ing pathways, including BMPR, FGFR, c-MET (tyrosine kinase re-ceptor of HGF), TGFR, PI3K, JNK, Arf6, Raf1, Smad2/3, b-catenin, c-Jun, Tbx3, and NF-jB, are expressed in hepatocytes (Zorn, 2008). Tumor necrosis factor-a (TNF-a) binds to TNFR and acti-vates three separate signaling pathways: the cell death, cell sur-vival, and cell proliferation signals (Nakamura and Nishina, 2009). In promoting cell proliferation, TNF-a/TNFR1 and HGF/c-MET can both activate MKK4 and MKK7, which then activate JNK; JNK activation promotes c-JUN phosphorylation and cdc2 gene expression (Naka-mura and Nishina, 2009). Simulta-neously, TNF-a/TNFR1 activates NF-jB signaling for cell survival that contradicts the cell death signal.
In mouse embryos, bipotential hepatoblasts differentiate into he-patocytes or biliary epithelial cells (BECs) on approximately day e13. The hepatoblasts that are sur-rounded by the portal mesen-chyme become BECs in response to the TGF-b and Wnt signals sent from the portal mesenchyme.
These signals downregulate the expression of prohepatic transcrip-tion factors, such as NHF4a, Tbx3, and C/ebp. These mesenchyme signals also upregulate the expression of biliary epithelial cell-specific transcription factors, such as Onecut1 (Oc1), Onecut2 (Oc2), hhex, and hnf1b. In addition, foxm1b mutant embryos lack BECs. Continuous signals (Notch, epidermal growth factor (EGF), and HGF) from the portal mesen-chyme are essential for ductal plate remodeling, whereas other signals (OSM, Dex, HGF, and Wnt) promote hepatocyte maturation (Zorn, 2008). Figure 1 illustrates the GRNs that direct liver organo-genesis from endoderm specifica-tion to hepatocyte/biliary differen-tiation. There are many missing links in this representation, such as the intramodular links, and the direct binding; regulation among these transcription factors and sig-naling pathways is not well eluci-dated and requires further study. Epigenomic Changes
Associated with Liver Development in Zebrafish
Epigenetic regulation of gene expression also plays an important role in liver development in zebra-fish. The two prevailing forms of epigenetic control over gene expression are DNA methylation and histone acetylation (Chu and Sadler, 2009; Hamilton, 2010). Transcriptional inactivation can be achieved via DNA methyltransfer-ase (DNMT), which adds a methyl group at the 50 position of a cyto-sine base, or histone deacetylases (HDACs), which remove an acetyl group from an e-N-acetyl lysine amino acid on a histone (Okano et al., 1999; Bird, 2002; Choudh-ary et al., 2009). Analysis of dnmt or hdac mutant embryos has dem-onstrated that epigenetic mecha-nisms control both hepatic specifi-cation and liver bud outgrowth in zebrafish.
In zebrafish, the knockdown of both hdac1 and hdac3 has been found to result in severe embry-onic development defects, in which aberrant hdac3 produced
greater effects on the liver than hdac1; hdac1 depletion affected liver size and caused the formation of ectopic endocrine tissue (Farooq et al., 2008; Noel et al., 2008). Dnmt1 mutant embryos undergo normal hepatic patterning but dif-ferentiate with increased apopto-sis; therefore, the liver cannot grow larger. The uhrf1 gene is essential for maintaining DNA methylation via the recruitment of Dnmt1 to hemimethylated DNA in mammalian cells. Abnormalities in either of these genes contribute to epigenetic changes that cause he-patocytes to undergo apoptosis (Bostick et al., 2007; Sharif et al., 2007; Chu and Sadler, 2009). The results summarized above indicate that epigenetic changes signifi-cantly influence liver develop-ment; however, this represents a new avenue of liver development research.
LIVER DISEASE AND HEPATOCARCINOGENESIS Because the liver is such a vital organ, liver failure is a life-threat-ening condition. Many etiological factors, such as chronic exposure to aflatoxin B1, alcohol consump-tion, and chronic viral infection with the hepatitis B virus (HBV) or hepatitis C virus (HCV), can cause liver damage (Morgan et al., 2004; McGlynn and London, 2005; Seeff and Hoofnagle, 2006; El-Serag and Rudolph, 2007; Marrero and Marrero, 2007). The develop-ment of HCC involves multiple steps that include steatosis, fibro-sis, cirrhofibro-sis, adenoma, and carci-noma (Tarantino et al., 2007). In fact, it is believed that more than 80% of all HCC cases are the result of infection by either HBV or HCV (Chen et al., 1997). Two bil-lion people worldwide have been infected with HBV; of these cases, 360 million suffer from chronic infection, and 600,000 die each year from HBV-related liver dis-ease or HCC (Shepard et al., 2006). HCC is one of the deadliest cancers, and there is still no effec-tive therapy available. Thus, an understanding of the molecular mechanisms involved in
hepato-carcinogenesis and the develop-ment of therapeutic approaches to treating liver cancer have become important.
GENETIC ALTERATIONS IN HEPATOCELLULAR
CARCINOMA
The pathogenesis of HBV-associ-ated HCC has been extensively studied, and molecular changes that occur during malignant trans-formation have been identified. It has been postulated that the insertion of HBV DNA into the human genome results in chromo-somal instability that causes can-cer formation by several different mechanisms. Chronic HBV infec-tion may trigger specific oncogenic pathways and cause the accumu-lation of genetic and epigenetic alterations in regulatory genes (Cougot et al., 2005; Tsai and Chung, 2010) that promote HCC. Transactivation of oncogenes, inactivation of tumor suppressor genes (TSGs), and alteration of the cell cycle by HBV proteins are all involved in the progression of hepatocellular carcinogenesis.
Hepatocarcinogenesis is a multi-step process that involves genetic alterations, including gain or loss of DNA, mutation of oncogenes and tumor suppressors, dysregu-lation of signaling pathways, epi-genomic changes, and changes in the expression of microRNA (Hoshida et al., 2010; Zucman-Rossi, 2010). The development of new high-throughput genomic technologies has promoted the classification of the molecular di-versity in human liver cancer and allowed us to understand the mul-tiple steps of hepatocarcinogenesis (Hoshida et al., 2009, 2010; Ung et al., 2009).
Analysis of 137 tumors using high-density allelotyping revealed that a b-catenin mutation associ-ated with chromosome 8p losses was related to a chromosome sta-bility group (Laurent-Puig et al., 2001). Similarly, analysis of 60 tumors discovered that b-catenin mutations were associated with a chromosome stability group.
How-ever, losses in chromosome 8q were not found in this study (Zucman-Rossi, 2010). In addi-tion, an HNF1A mutaaddi-tion, CDH1 methylation, and Wnt pathway activation were associated with a genomic stability group during hepatocarcinogenesis (Boyault et al., 2007). In a chromosome instability group, many of the chromosome areas exhibited the most frequent allelic losses, and axis inhibition protein 1 (AXIN1) and p53 were frequently mutated (Laurent-Puig et al., 2001). Simi-larly, analysis of 60 tumors found that AXIN1 and p53 mutations were associated with a chromo-some instability group (Boyault et al., 2007). Some chromosomal loss of heterozygosity (LOH) is associated with a chromosome instability group of the HCC that includes 4q, 13q, 16p, 16q, and 17p from both studies; other genomic losses were found in one study but not in the other. These chromosomal regions contain key players in HCC, such as p53 (chro-mosome 17p), Rb (chro(chro-mosome 13q), AXIN1, and cyclin-depend-ent kinase inhibitor 2A (CDKN2A) (Laurent-Puig and Zucman-Rossi, 2006). Other genetic alterations in the genomic instability group include mutations in PI3K2CA and the methylation of CDKN2A, as well as the activation of the mitotic cell cycle, the AKT pathway, and developmental and imprinting genes, e.g., insulin-like growth factor 2 (IGF-2) (Boyault et al., 2007; Zucman-Rossi, 2010). The signatures of 16 genes to classify the HCC would be clinically useful for determining the dysregulated pathways and predicting drug response (Boyault et al., 2007). Here, we have summarized the most important genetic alterations in human HCC. Table 1 lists all of the signaling pathways and down-stream cascades responsible for hepatogenesis and hepatocarcino-genesis.
Alterations of the Wnt/b-Catenin Signaling Pathway
Inappropriate reactivation of the Wnt pathway that results from
LIVER DEVELOPMENT AND CANCER FORMATION IN ZEBRAFISH 161
alterations in the b-catenin gene (CTNNB1) has been implicated in liver oncogenesis (Buendia, 2000). b-Catenin is the most frequently observed mutation-activating oncogene in HCC; alterations in this gene are found in 20 to 50% of HCC patients (de La Coste et al., 1998; Miyoshi et al., 1998). b-Catenin has dual functions in ad-hesion and Wnt signaling. N-ter-minal mutations of b-catenin trig-ger dominant oncogenic activity (Morin et al., 1997), and the loss of consensus phosphorylation sites onb-catenin has been identified in many mutations, which suggests that b-catenin is negatively regu-lated by GSK3b/APC/axin via phosphorylation (Laurent-Puig and Zucman-Rossi, 2006). The muta-tion rates of the tumor suppres-sors AXIN1 and AXIN2 in human HCC are5 to 25% and 3 to 10%, respectively (Ishizaki et al., 2004; Zucman-Rossi et al., 2007; Zuc-man-Rossi, 2010).
A previous study that investi-gated HCC cell lines, HB (hepato-blastoma), and primary HCC showed that mutations in the CTNNB1, AXIN1, and AXIN2 genes caused Wnt/b-catenin pathway dysregulation (Taniguchi et al., 2002). All of the CTNNB1 tions in HCC were missense muta-tions, minor delemuta-tions, or small insertions (Jeng et al., 2000). The AXIN1 mutations included trunca-tion mutatrunca-tions due to either small deletions or nonsense mutations, which encoded a truncated protein in the cytoplasmic GSK3b complex that inhibited the Wnt pathway; this suggests that AXIN1 functions as a tumor suppressor. Inactiva-tion of AXIN1 prevents phospho-rylation of b-catenin by GSK3b, which leads to an accumulation of b-catenin and the activation of Wnt target genes (Laurent-Puig and Zucman-Rossi, 2006). Previ-ous studies have shown that AXIN2 is a transcriptional target of the TCF/LEF transcription factor complex downstream of activated b-catenin (Jho et al., 2002). AXIN2, which is mutated in 3 to 10% of HCC cases, functions as an antagonist by promotingb-catenin degradation (Ishizaki et al., 2004;
Zucman-Rossi, 2010). These results may explain the presence of tumors with mutations in both the CTNNB1 and AXIN1 or AXIN2 genes, which contribute to the activation of the Wnt signaling pathway (Taniguchi et al., 2002).
In the canonical Wnt signaling pathway, Wnt binds to its cell sur-face receptor, causing dissociation ofb-catenin from the APC complex and preventing degradation (Chu and Sadler, 2009). Wnt/b-catenin signaling is activated relatively early during development and regeneration. When activated, Wnt/b-catenin signaling switches on the expression of target genes. These downstream target genes are important in cell cycle progres-sion and contribute to the initia-tion of the regenerainitia-tion process (Nejak-Bowen and Monga, 2010). Dysregulation of Wnt/b-catenin signaling was found in zebrafish liver tumors (Lam et al., 2006); dysregulation of ctnnb, wif1, wnt2, ctnnbip1, and ccnd1 also suggested the presence of deregu-lation in the Wnt/b-catenin signal-ing pathway in zebrafish liver tumors (Lam and Gong, 2006).
Alterations in the p53 Gene The most frequently mutated TSG in HCC is p53 (Hsu et al., 1991; Zucman-Rossi, 2010), which is activated in response to DNA damage and either promotes apoptosis or induces cell cycle arrest to permit DNA repair (Levine et al., 1991). The p53 TSG is located on chromosome 17p13.1 (Isobe et al., 1986) and plays a major role in HCC, irre-spective of the etiology (Edamoto et al., 2003). In more than 50% of HCC tumors, a G? T transversion at codon 249 of p53 was found af-ter high aflatoxin B1 (AFB1) expo-sure (Bressac et al., 1991). In contrast, patients who were not exposed to AFB1 had lower rates of p53 gene mutation (20%) without specific codon hotspots (Laurent-Puig and Zucman-Rossi, 2006). Moreover, analysis of intra-tumoral nodular lesions within HCC samples has found genetic heterogeneity in p53, and the p53
mutation has been proposed to correlate with shortened survival and a poor prognosis (Honda et al., 1998; Buendia, 2000). Alterations in the
Retinoblastoma Protein, CDKN2A, and Gankyrin
The tumor suppressor retino-blastoma protein (Rb) is critical for the development of several cancer types. In normal cell signaling, Rb prevents cell division and cell cycle progression. Frequent allelic dele-tions on chromosome 13q that cause inactivation of the tumor-suppressor Rb gene located at 13q14 have been observed in HCCs (Friend et al., 1986). LOH at the Rb locus has been observed in 25 to 48% of cases of HCC (Kuroki et al., 1995), and pRb expression has been shown to be strongly downre-gulated in 30 to 50% of tumors, which correlates with genetic alter-ations in the p53 gene (Buendia, 2000). However, an Rb mutation alone is found in less than 11% of HCC cases (Zhang et al., 1994). This result is indicative of the het-erogeneity of human HCC.
There are many different ways of inactivating pRb, including dele-terious mutations in the gene itself, loss of TGF-b responsive-ness, and the inactivation of cyclin D-dependent kinase inhibitor 2A (CDKN2A). CDKN2A functions as a tumor suppressor in the retino-blastoma pathway (Hickman et al., 2002), and it is mutated in 10 to 60% of HCC patients (Zucman-Rossi, 2010). In HCC, LOH at 9p occurs in 20% of cases, and homozygous deletions at 9p21 (where CDKN2A is located) have been detected (Liew et al., 1999). In the majority of tumors, inactivation of the gene was achieved by de novo methyla-tion of the CDKN2A promoter, which led to the absence of protein expression in 30 to 70% of cases (Matsuda et al., 1999).
Overexpression of gankyrin, an oncoprotein that contains seven ankyrin repeats, has been found in HCC patients (Higashitsuji et al., 2000). Gankyrin downregulates p53 protein levels via
ubiquityla-TABLE 1. Sig nalin g Path ways an d Down str eam Casca des Res pon sible for Hepat ogen esis and Hepat ocarcinog enesis Ligand Rece ptor Transdu cer Transc riptio n factor Target Liv er deve lopment Hepato carcinog enesis TGF-b BMPs TGFR Activin Smad 2/3 gat a6, gata4, hnf1b, fox a1, foxa 2, prox1, and hhex b 1-integ rin 1. Spec ifica tion of thr ee germ la yers 2. Me senchyma l signals for hep atic sp ecificat ion 3. Liv er bud grow th 1. BMP s overe xpres sion 2. TGF -b signa l pathw ay activa ted by HB x HGF c-Met SEK K1,MK K4, MKK7 , GRB2, GAB1, phosp holipa se C, PI3K, and ERK P38, c-ju n, ATF2/ 7, and b -ca tenin hnf 1b, hnf6, ck19, hnf4a , c/e bp, hnf 1a, alb, b 1-integ rin, and pho sphatases 1. Liv er bud grow th 2. Duct al plate remo deling 3. Prom oting hepato cyte matu ration 1. HGF overexp ression 2. Me t overe xpres sion 3. Me t mut ations 4. Me t d u plication Wnt Frizzled AXIN /APC / GSK3 b Wnt antago nist sFRPs AXIN s are the negat ive regulat ors b -Cat enin (CTNNB 1) hnf 4a, c/ebp, hnf1a, al b, Sur vivin, Bcl-XL, Cyclins, CDKs, Rho / Rock, an d FAK 1. Spec ifica tion of thr ee germ la yers 2. Inh ibited at hepat ic specif ication stag e 3. Prom oting liver bu d grow th 4. Prom oting hepato cyte matu ration 1. b -Catenin mut ation lead s to its stabi lization 2. Muta tions in the CTNN B1 , AX IN1, an d AXIN 2 gene s 3. Inac tiva tion of AX IN1 thr ough geno mic delet ion 4. Sile ncing sFRPs thr ough hype rmet hylatio n b y H B x 5. Sile ncing CDKN2A throu gh m ethylatio n FGF FGFR, AL K6 b -Cat enin, id3 gat a6, gata4, hnf1b , foxa1, foxa 2, prox1, and hhex 1. Spec ifica tion of thr ee germ la yers 2. Me senchyma l signals for hepatic sp ecificat ion 3. Liv er bud growt h b -Cat enin mutation lea ds to its stabilization EGF, PD GF, an d VEG F EGFR , PDGF R, and VEG FR Ra s/Raf/ MAP2K/MA PK id3 hnf 1b, hnf6, and ck19 1. EGF is require d for ductal plate remodeling 2. Proli fera tion, m igration, an d surv ival 1. Ra s m uta tions 2. Ra f hype ractiv ated 3. MEK 1/2 overexp ression 4. ERK1/ 2 overe xpres sion 5. EGF overe xpres sion EGF, PD GF, an d VEG F EGFR , PDGF R, and VEG FR PI3K/ PDK1/ Akt/m TOR/ HIF1 a, HIF1 b PTEN is antagonism 1. Proli fera tion,mig ration, and surv ivial 2. Antia popto sis 1. PTEN inact ivat ion thr ough gene delet ion 2. AKT overe xpres sion 3. mTOR ov erexpr ession IGF IGF-1 R IGF-2R Grow th and prolife ration 1. Abnorm alities in IGF and IGF-1 R 2. Ove rexpre ssion of IGF and IGF-1 R Foxm 1b Prox1 G2/ M phase Prolifer ation Prolife ration TNF a TNFR IKK c, IKK b , p50/p6 5-Re lA Apoptos is FasL FasR Apoptos is Xbp 1 E R stress Apoptos is
tion and degradation (Kim et al., 2009). Gankyrin also binds to Rb, accelerating the degradation of Rb in vivo and in vitro (Higashitsuji et al., 2000). Collectively, frequent alterations in Rb, CDKN2A, and gankyrin play an important role in hepatocarcinogenesis.
DYSREGULATED
SIGNALING PATHWAYS IN HEPATOCARCINOGENESIS Previous studies have reported the occurrence of aberrant activation of signaling pathways to sustain proliferative signaling (e.g., the EGF and RAS/mitogen-activated protein kinase pathways), to resist cell death (e.g., Akt, the mecha-nistic target of the rapamycin pathway), to enable replicative immortality (e.g., the Wnt and Hedgehog pathways), and to induce angiogenesis (e.g., vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF) pathways), which are capabilities acquired during the multistep development of human tumors (Hoshida et al., 2010). The first fundamental hallmark of can-cer is its ability to sustain chronic proliferation. In the past decade, many signaling pathways have been found to be dysregulated in cancer, contributing to tumor for-mation and progression. These pathways include the Wnt/ b-cate-nin (as described earlier), HGF/c-MET, EGFR, IGF, MAPK, and PI3K/ AKT/mTOR pathways (Nejak-Bowen and Monga, 2010; Whit-taker et al., 2010; Zender et al., 2010). These pathways play signif-icant roles in liver organogenesis. Human and zebrafish liver tumors share a molecular framework that is deregulated during tumorigene-sis, which is indicative of the evo-lutionarily conserved properties of these pathways and their down-stream transducers (Lam and Gong, 2006; Lam et al., 2006). The HGF/c-MET, EGFR, and IGF Signaling Pathway
c-MET, the HGF tyrosine kinase receptor, is predominantly ex-pressed on the surface of epithelial
and endothelial cells. The HGF ligand exerts its effects by binding to c-MET, which regulates many important events during embryo-genesis, including cell prolifera-tion, migraprolifera-tion, survival, branch-ing morphogenesis, and angiogen-esis. Upon HGF binding to c-MET, the signal cascade occurs via phosphorylation of the adaptor proteins growth factor receptor-bound protein 2 (GRB2) and GRB2-associated-binding protein 1 (GAB1), which then activate downstream effectors, such as phospholipase C, PI3K, and ERK (Whittaker et al., 2010).
The receptor for IGF1, IGF1R, is a key regulator of anchorage-inde-pendent growth (Whittaker et al., 2010). Following liver damage or viral transactivation, the IGF-2 re-ceptor is upregulated by altered methylation of the IGF-2 promoter (Feitelson et al., 2004; Whittaker et al., 2010). In zebrafish HCC, several IGF-binding proteins (IGFBPs), such as igfbp2b, were significantly hypomethylated and may have upregulated the expres-sion of IGF-2 (Mirbahai et al., 2011). In the early stages of tu-morigenesis in highly proliferating tumor cells, the lack of a vascular supply results in hypoxia (Kelly et al., 2008; Mirbahai et al., 2011). The anaerobic conditions and the presence of IGF result in increased expression of hypoxia-inducible factor 1 (HIF-1) (Kelly et al., 2008).
The MAPK Pathway
The mitogen-activated protein kinase (MAPK) pathways regulate crucial cellular processes during development, including prolifera-tion, differentiaprolifera-tion, angiogenesis, and survival (Whittaker et al., 2010). The MAPK signaling path-ways play vital roles in embryo-genesis and are often deregulated in various types of human cancer, including HCC. There are at least four subfamilies of MAPKs: extrac-ellular signal-regulated kinases 1 and 2 (ERK1/2), c-Jun N-terminal kinases (JNKs), p38 MAPKs, and ERK5. The activation of the Ras-MAPK pathway in zebrafish liver
tumors is indicated by upregula-tion of shc1, mapk1, dusp4, dusp6, and genes associated with Ras GTPases, e.g., rhoc, rhogap1, cdc42, rac1, g3bp2, and gef10 (Lam and Gong, 2006). Constitu-tive activation of a MAPK pathway has been found in 40% of mela-noma patients due to mutations of the B-Raf protein (Davies and Samuels, 2010).
The PI3K/AKT/mTOR Signaling Pathway
In normal tissue, the PI3K/AKT/ mTOR signaling pathway targets the lipid products of PI3K for de-phosphorylation. The PI3K/AKT/ mTOR pathway is negatively regu-lated by the tumor suppressor phosphatase and tensin homolog (PTEN) (Roberts and Gores, 2005; Whittaker et al., 2010). Binding of IGFs or EGF to their receptors acti-vates PI3K (Avila et al., 2006; Whittaker et al., 2010) and the downstream signal pathway. The PTEN gene is mutated in 5 to 10% of human HCC cases (Bamford et al., 2004; Whittaker et al., 2010), which results in the consti-tutive activation of the PI3K/AKT/ mTOR pathway (Hu et al., 2003; Whittaker et al., 2010). The PI3K/ AKT/mTOR pathway plays a criti-cal role in the pathogenesis of HCC (Whittaker et al., 2010).
Epigenomic Changes Associated with HCC
Several other genes from fami-lies that include ABCA, CHST, DHX, KCTD, MEGF, MYO, NPY, RNF, and TBCID have been found to be hypermethylated in both zebrafish and human HCCs. The genes with altered methylation in zebrafish HCC are associated with biological functions, such as cell death, cell morphology, inflam-matory response, DNA repair, and replication, and induced mol-ecules involved in cancer forma-tion, such as c-jun, shc, and pka (Mirbahai et al., 2010). Many HCCs exhibit methylation of at least one TSG promoter, particu-larly SOGS-1, APC, E-cadherin, and p15. Several epigenetically
silenced or aberrantly methylated putative TSGs found in HCCs were also inactivated in the non-tumorous liver. However, the fre-quency of methylation in tumor-ous liver is much higher than in nontumorous liver. Moreover, methylation is observed more fre-quently in HCV carriers than in HBV/HCV-double-negative HCC cases (Yang et al., 2003; Calvisi et al., 2007).
Hepatocarcinogenesis is associ-ated with increased levels of DNMT1, DNMT3a, and DNMT3b mRNA and a progressive increase in the number of epigenetically silenced genes compared to nor-mal liver and chronic hepatitis-, cirrhosis-, and HCC-affected livers (Oh et al., 2007). These combined epigenetic events facilitate the estimation of prognosis and risk of recurrence of HCC. Epigenetic changes could be used to custom-ize therapy and predict future sur-vival.
HEPATOCARCINOGENESIS RELATED TO THE HBV X ANTIGEN
One of the proteins encoded by HBV, the hepatitis B virus X protein (HBx), caused enhanced colony formation or transformation of cells in vitro in various cell lines (Koike et al., 1989; Shirakata et al., 1989; Seifer et al., 1991; Zhang et al., 2009). Several HBx trans-genic mouse models that develop HCC have also been created (Kim et al., 1991; Ullrich et al., 1994; Wu et al., 2006). A transgenic mouse model, in which the albumin promoter drives the expression of HBx, has been shown to develop HCC at 14 to 16 months of age without chemical treatment (Wu et al., 2006, 2008). We have used this mouse model to identify bio-markers for the early stages of HCC formation. Here, we have summarized the effect of HBx on carcinogenesis (Figure 2).
HBx Enhances Genomic Instability
As in most cancers, gross chro-mosomal abnormalities have been credited with contributing to cellu-lar transformation and tumor pro-gression in HBV-associated HCC (Gatza et al., 2005). Genomic instability has been explicitly linked to the expression of HBx. HBx increases the levels of several cellular oncogenes through the amplification of chromosome regions that encode these genes (Levy et al., 2002). HBx also decreases the expression of TSGs, p53 and Rb (Murakami et al., 1991), and AXIN1 (Satoh et al., 2000), via the deletion of the genomic portions containing those genes. Several studies have reported that HBx is associated with increased mutation frequen-cies within the cellular genome that are probably due to the inhi-bition of hepatocyte DNA repair by
Figure 2. HBx-induced hepatocarcinogenesis. HBx is involved in hepatocarcinogenesis, primarily via genomic instability and changes in epigenetic status, as well as gene expression. A normal hepatocyte must proceed through early events such as hyper-methylation and germline mutations to become a histologically distinguishable HCC. In later life stages, many oncogenic activa-tions, inactivaactiva-tions, and/or mutations of the tumor suppressor gene are involved in the transition of an early, benign tumor to a localized, malignant cancer. Further genomic changes, such as the amplification of chromosome regions that contain oncogenes and the deletion of genomic regions that encode tumor suppressor genes, cause HCC to become metastatic and incurable.
LIVER DEVELOPMENT AND CANCER FORMATION IN ZEBRAFISH 165
HBx (Slagle et al., 1996). The rate of chromosomal alterations is sig-nificantly increased in HBV-related tumors compared to tumors asso-ciated with other risk factors. HBV may therefore play a role in enhancing genomic instability (Cougot et al., 2005).
Signal Transduction Pathways Affected by HBx
HBx may contribute to the devel-opment of HCC via the activation of signaling pathways, such as NF-jB; these pathways affect TGF-b1 expression. This contribution was demonstrated by comparing HBx-positive to HBx-negative HepG2 cells (Pan et al., 2004). HBx can also shift TGF-b signaling from the tumor-suppressive to the oncogenic pathway in the early carcinogenic process by the activation of JNK ki-nase, which phosphorylates Smad3 at a linker region instead of the C-terminus (Murata et al., 2009). TGF-b participates in many different stages of liver organogenesis, from endoderm formation to hepatic specification and liver bud growth (Zorn and Wells, 2009). In the fetal liver, TNF-a engages TNF receptor 1 (TNFR1) and activates NF-jB sig-naling to transmit the cell survival signal (Nakamura and Nishina, 2009).
HBx may upregulate the expres-sion of multidrug resistance protein (MDR1) and inhibit apoptosis through the activation of the Src and PI3 kinase pathways (Kang-Park et al., 2006). HBx acts as a tumor inducer and stimulates the neoplastic transformation of normal cells, but HBx shifts its function to induce apoptosis in association with Ras (Wei et al., 2006). HBx can be phosphorylated by ERK, and the phosphorylated form has been found to repress the transcription of p21(WAF1/Cip1) and to translo-cate from the cytoplasm to the nu-cleus (Noh et al., 2004). HBx also induced centrosome hyperamplifi-cation and mitotic aberration via the activation of the Ras-MEK-MAPK pathway. These findings may provide a possible mechanism by which HBx causes genomic instabil-ity in an HBV-infected liver (Yun
et al., 2004). The protein kinase C pathway (Luber et al., 1993), RAS/ RAF/MAP kinase cascade (Benn and Schneider, 1994), c-Jun N-terminal activating protein kinase (Benn et al., 1996), and JAK1/STAT path-ways have been found to be sub-ject to activation by HBx (Lee and Yun, 1998). Based on these stud-ies, HBx interacts with many signal transduction pathways to induce HCC, and HBx may have different functions that depend on its associ-ation with different transduction pathways at early or late stages of viral infection.
Transactivation of Cellular Genes
HBx has been found to increase the levels of fibronectin (FN) mRNA and protein via the HBx-mediated transactivation of the FN promoter, which is NF-jB dependent. HBx also antagonized the repression of the FN promoter by the p53 tumor suppressor. Hence, the FN gene may be a natural target for HBx transactivation, perhaps through the activation of NF-jB and the inactivation of p53, thereby con-tributing to the accumulation of FN in the liver over the course of chronic HBV infection (Norton et al., 2004). It has been reported that HBx can activate all of the Pol I, Pol II, and Pol III genes via acti-vation of expression of the TATA-binding protein, which is a compo-nent of the basal transcription ap-paratus (Wang et al., 1995). Physical Binding and Functional Inactivation of the p53 Cellular Tumor Suppressor Protein
Tumor development precisely correlates with p53 binding to HBx in the cytoplasm and the complete blockage of p53 entry into the nu-cleus (Ueda et al., 1995). An anal-ysis of tumor cell DNA showed no evidence of p53 mutation except in advanced tumors, where a small proportion of cells may have acquired specific base substitu-tions. These results suggest that genetic changes in p53 are late events that may contribute to tu-mor progression (Ueda et al.,
1995). In addition to aberrations in the p53 gene, loss of the Rb gene or LOH at chromosome 13q was observed in six of seven in-formative cases of eight tumors that carried a mutated p53 gene (Murakami et al., 1991).
Epigenomic Changes Associated with HBx
HBx has been reported to repress E-cadherin expression via activation of the DNA methyltrans-ferase-mediated hypermethylation of the E-cadherin promoter (Lee et al., 2005). In a separate study, glutathione S-transferases P1 (GSTP1), enzymes that defend cells against damage mediated by oxidants and electrophilic carcino-gens, were suppressed through the hypermethylation of their pro-moter regions. These data indicate that epigenetic silencing of GSTP1 gene expression via CpG island DNA hypermethylation is common in human HBV-associated HCC (Zhong et al., 2002).
A ZEBRAFISH ANIMAL MODEL FOR THE STUDY OF LIVER DISEASES AND HCC Although the main focus of zebra-fish research has generally been on developmental biology, labora-tory observations of zebrafish have resulted in the identification of diseases that are similar to those found in humans, such as cancer. Thus, zebrafish became an animal model for human disease, and dozens of studies that used zebrafish as a cancer model have been published in the last decade (Feitsma and Cuppen, 2008).
Increasing rates of HCV infection have been associated with an increase in the incidence of HCC in the United States (El-Serag et al., 2003). HCV-induced hepatocarci-nogenesis is widely reported to be due to the HCV core protein, which inhibits p21 expression through in-hibition of the TGF-b pathway (Lee et al., 2002). In a transgenic zebrafish model, the HCV core protein induced HCC with or with-out treatment with thioacetamide, which is a hepatotoxin. However,
thioacetamide treatment can accelerate HCC development by twofold to yield fully developed HCC in 6 weeks (Rekha et al., 2008).
As previously mentioned, endemic HBV infection is strongly correlated with the high preva-lence of HCC in Asian countries (Parkin et al., 2001). Moreover, the HBx viral protein has been shown to modulate cell prolifera-tion and induce HCC. In a trans-genic zebrafish model, HBx under the control of a liver-specific pro-moter resulted in hepatic fat accu-mulation during the progression of hepatitis. Transgenic fish that express viral genes could be an excellent animal model for HCC for studying accelerated cancer for-mation, induction of fatty liver, and the synergistic effects of dif-ferent risk factors, including hepa-totoxin.
One of the advantages of using zebrafish as a high-throughput screening method for carcinogens is the ease of manipulation. Most of the carcinogens added to embryos have induced neoplasms derived from many tissues, such as epithelial, mesenchymal, neu-ral, and neural crest. The liver is a primary target organ for most car-cinogens, regardless of the devel-opmental stage of the fish at ex-posure. Low ppb concentrations of AFB1 are usually used for dietary carcinogenesis studies in rainbow trout to produce a high incidence of liver neoplasia (Bailey et al., 1996).
Zebrafish larvae are also an attractive model for studying alco-holic liver disease (ALD). In humans, acute alcohol abuse can result in steatosis, which may pro-gress to more severe hepatic dis-ease. The alcohol metabolism pathways in zebrafish are similar to those in humans, and the zebrafish liver is mature in larvae by 4 days postfertilization (dpf). Moreover, zebrafish larvae de-velop steatosis, which is a sign of ALD, as a result of alcohol being added to the water. The activation of Srebp is required for steatosis in the zebrafish ALD model. Deci-phering the molecular
pathogene-sis of the zebrafish ALD model became possible following the almost complete sequencing of its genome and by using a genetics tool (Passeri et al., 2009).
Several zebrafish strains have become powerful models for eluci-dating the mechanisms of carcino-genesis and are superior in vivo systems for rapid screening of anticancer genetic or chemical fac-tors (Rekha et al., 2008). In con-clusion, the zebrafish is an ex-cellent model to delineate the mechanisms that underlie hepato-carcinogenesis and as a therapeu-tic drug-screening platform. New Transgenic Technology for Studying HCC Using Zebrafish
Genetic screening that identifies genes required for developmental processes has been successfully performed in zebrafish. Previously, forward genetic screening in zebrafish identified mutants that developed hepatomegaly, which is a symptom of many liver disor-ders. Several new genes that play important roles in liver develop-ment, physiology, and pathology have been identified using forward genetic screening (Sadler et al., 2005).
The transgenic technologies available in zebrafish have improved over the last 2 decades (Stuart et al., 1988, 1990). Differ-ent delivery systems for the trans-gene, such as the injection of lin-ear DNA (Stuart et al., 1988), supercoiled plasmid DNA (Stuart et al., 1990; Culp et al., 1991), or recombinant bacterial artificial chromosomes into early-stage embryos (Culp et al., 1991) have been developed in zebrafish. Recently, a new transgenic tech-nology, Tol2-mediated transgene-sis, has been established. The Tol2 element is a naturally arising, active transposable element dis-covered in fish genomes. The Tol2 transposon system is considered to be a useful gene transfer vector in vertebrates ranging from fish to mammals (Urasaki et al., 2006). Using coinjection of in vitro-transcribed Tol2 RNA, the DNA fragment surrounded by the Tol2
element transposon can be effi-ciently excised and integrated into the genome (Kawakami et al., 2000). Tol2-mediated transgene-sis is an excellent method for cre-ating transgenic zebrafish because of the high transposition efficiency and the capacity to transfer a large DNA fragment (Urasaki et al., 2006).
Initially, many zebrafish labora-tories created green fluorescent protein (GFP) reporter transgenic lines as a marker for cells express-ing the gene of interest. Recently, as transgenesis has become com-mon in zebrafish laboratories, researchers have tested mamma-lian promoters, promoters from other fish species, and endoge-nous tissue-specific promoters to drive transgene expression (Dei-ters and Yoder, 2006). One of the most useful GFP transgenic fish lines came from a liver-specific promoter, liver fatty acid-binding protein (L-FABP) (Andre et al., 2000; Denovan-Wright et al., 2000). In the liver, L-FABP plays an important role in the intracellu-lar binding and trafficking of long-chain fatty acids. Isolation of the zebrafish L-FABP promoter and construction of GFP fish lines have been previously performed (Her et al., 2003a,b). Oncogenes driven by the L-FABP promoter have become a useful system for study-ing HCC in the zebrafish model. Studying Metastasis in Zebrafish Using the
Xenotransplantation Method The zebrafish is a vertebrate with a complex circulatory system and genetics that are similar to humans, which makes related experiments feasible (Weinstein, 2002). The zebrafish is an excel-lent model for cancer research. There are many advantages of the zebrafish compared to the mouse (Lam et al., 2006), including ease of experimental handling, drug treatment, and high-throughput screening, as well as the optical transparency of the vascular sys-tem and the feasibility of forward and reverse genetic approaches (Thisse and Zon, 2002). Recent
LIVER DEVELOPMENT AND CANCER FORMATION IN ZEBRAFISH 167
studies have demonstrated the possibility of injecting human can-cer cell lines into zebrafish (Wein-stein, 2002). In these studies, 2-day-old embryos were injected with 50 to 1000 cells in the yolk sac or near the vascular system. However, in the blastula stage, 1 to 100 cells were sufficient for tu-mor mass engraftment (Nicoli et al., 2007). Studies of the recipi-ent animals posttransplant are critical for determining tumor engraftment or metastasis (Taylor and Zon, 2009). In a successful xenotransplantation, the migration of CM-DiI or cancer cells labeled with another florescent dye can be traced in living embryos (Marques et al., 2009). Tumor angiogenesis induced by the cancer cells can also be investigated in an embryo 3 dpf. Cell invasion and angiogen-esis are dynamic processes; com-pared to a fluorescent microscope, a high-resolution confocal micro-scope can provide high quality, dynamic, and three-dimensional section images (Stoletov et al., 2007). When mice are used as a tumor transplantation model, only two to three animals can be cre-ated for one experiment; in con-trast, many more zebrafish than mouse embryos can be injected. In published transgenic line experiments, more than 500 one-cell-stage embryos were injected in a single day, generating 100 transplanted embryos, which is more prolific than the mouse model (Taylor and Zon, 2009). Currently, the tumor transplanta-tion assay is a popular and much easier method to test carcinogene-sis and screen cancer stem cells (White et al., 2008). In conclusion, zebrafish represents a promising animal model for tumor xeno-transplantation and carcinogenesis research.
CONCLUSION
The basic mechanisms that control liver formation are presented in the first section of this review. Evolutio-narily conserved GRNs that direct hepatogenesis are described. Many signaling pathways and genes are activated during the developmental
process. Although there are many reviews on liver organogenesis and development, to our knowledge, ours is the first review in which the relationships between different events have been connected within the network architecture at the mo-lecular level.
In the second section of this review, we have summarized known pathological liver condi-tions, primarily focusing on HBx-induced genomic instability and activation of signaling pathways and transcription as well as epige-netic status. The pathways that are deregulated in HCC are also discussed. There are common fea-tures shared by embryonic liver development and liver carcinogen-esis. The signaling pathways, tran-scription factors, and molecular machinery that dictate these events are used in both situations; however, mutations and changes at the genomic or epigenomic level, respectively, occur in the case of tumorigenesis. Although we have a good understanding of organogenesis, establishing how GRNs underlie hepatocarcinogene-sis remains a challenge.
Lastly, we have highlighted here examples of exciting research that have utilized zebrafish as a human disease model, especially for liver cancer. The zebrafish has been used by developmental biologists to decode developmental GRNs, and cancer biologists have used the zebrafish to create models for liver cancer via transgenesis and xeno-transplantation. The most signifi-cant advantages of using zebrafish in cancer studies are the low cost of high-throughput drug screening and the ease of toxicity screening. Zebrafish thus constitutes a bridge between basic science (i.e., liver development) and translational research (i.e., liver cancer).
FUTURE PERSPECTIVES The study of liver development has demonstrated that many tran-scription factors and signaling pathways and their components, as well as epigenetic changes, are essential for the specification, growth, and differentiation of the
liver. Gene mutations, gains or losses of DNA, and changes in methylation status of transcription factors and signaling pathways contribute to liver disorders and cancer formation. With the help of the zebrafish model and new tech-nologies, deciphering the GRNs that underlie hepatocarcinogenesis and finding a cure for liver cancer are possible.
In the future, there are many areas that will require intensive exploration. How do different cell types interact during develop-ment? What is the circuit in the diagrams and what are the nodes of interaction? Do the pathways that regulate cell proliferation and survival in the embryo also control regeneration and cancer formation in the adult? What are the cancer GRNs and what are the differences between normal embryonic devel-opment and cancer formation? Can we design drugs that target specific points in the network to correct disorders and reverse can-cer formation? The zebrafish is a well-established model for addressing these questions.
Because of its many unique advantages, including transparent embryos and mutant adults, rapid embryonic development, short sex maturation time, large numbers of progeny, and well-developed gene transfer technology, zebrafish has become a popular research model for genetic and developmental biology. In the past 3 decades, several large-scale genetic screens and the nearly complete zebrafish whole-genome sequencing have considerably increased the use of zebrafish in diverse research areas. Thus, the zebrafish model has become an established and evolving system for liver cancer. Several important technologies have been established in zebra-fish, including Tol2 and MultiSite gateway-based construction for gene analysis, as well as zebrafish xenotransplantation methods. Tu-mor xenografting has been recently developed in zebrafish to complement and overcome the deficiencies of other model sys-tems. As an established animal disease model, the zebrafish
pro-vides a great opportunity to test the functions of disease markers in vivo and can be an effective and efficient system for drug screen-ing. Research using zebrafish can integrate basic research with animal disease models and clinical research for successful imple-mentation in biopharmaceutical industries.
REFERENCES
Andre M, Ando S, Ballagny C, et al. 2000. Intestinal fatty acid binding protein gene expression reveals the cephalocaudal patterning during zebrafish gut morphogenesis. Int J Dev Biol 44:249–252.
Avila MA, Berasain C, Sangro B, et al. 2006. New therapies for hepa-tocellular carcinoma. Oncogene 25:3866–3884.
Bailey GS, Williams DE, Hendricks JD. 1996. Fish models for environ-mental carcinogenesis: the rainbow trout. Environ Health Perspect 104 (Suppl 1):5–21.
Bamford S, Dawson E, Forbes S, et al. 2004. The COSMIC (Catalogue of Somatic Mutations in Cancer) database and website. Br J Cancer 91:355–358.
Benn J, Schneider RJ. 1994. Hepati-tis B virus HBx protein activates Ras-GTP complex formation and estab-lishes a Ras, Raf, MAP kinase signal-ing cascade. Proc Natl Acad Sci USA 91:10350–10354.
Benn J, Su F, Doria M, et al. 1996. Hepatitis B virus HBx protein induces transcription factor AP-1 by activa-tion of extracellular signal-regulated and c-Jun N-terminal mitogen-acti-vated protein kinases. J Virol 70:4978–4985.
Bird A. 2002. DNA methylation pat-terns and epigenetic memory. Genes Dev 16:6–21.
Bostick M, Kim JK, Esteve PO, et al. 2007. UHRF1 plays a role in main-taining DNA methylation in mamma-lian cells. Science 317:1760–1764. Boyault S, Rickman DS, de Reynies
A, et al. 2007. Transcriptome classifi-cation of HCC is related to gene alter-ations and to new therapeutic targets. Hepatology 45:42–52.
Bressac B, Kew M, Wands J, et al. 1991. Selective G to T mutations of p53 gene in hepatocellular carcinoma from southern Africa. Nature 350:429–431.
Buendia MA. 2000. Genetics of hepa-tocellular carcinoma. Semin Cancer Biol 10:185–200.
Calvisi DF, Ladu S, Gorden A, et al. 2007. Mechanistic and prognostic sig-nificance of aberrant methylation in the molecular pathogenesis of human
hepatocellular carcinoma. J Clin Invest 117:2713–2722.
Chan TM, Longabaugh W, Bolouri H, et al. 2009. Developmental gene regulatory networks in the zebrafish embryo. Biochim Biophys Acta 1789:279–298.
Chen CJ, Yu MW, Liaw YF. 1997. Epidemiological characteristics and risk factors of hepatocellular carci-noma. J Gastroenterol Hepatol 12:S294–S308.
Choudhary C, Kumar C, Gnad F, et al. 2009. Lysine acetylation tar-gets protein complexes and co-regu-lates major cellular functions. Sci-ence 325:834–840.
Chu J, Sadler KC. 2009. New school in liver development: lessons from zebra-fish. Hepatology 50:1656–1663. Cougot D, Neuveut C, Buendia MA.
2005. HBV induced carcinogenesis. J Clin Virol 34 (Suppl 1):S75–S78. Culp P, Nusslein-Volhard C, Hopkins
N. 1991. High-frequency germ-line transmission of plasmid DNA sequen-ces injected into fertilized zebrafish eggs. Proc Natl Acad Sci USA 88:7953–7957.
Davies MA, Samuels Y. 2010. Analy-sis of the genome to personalize therapy for melanoma. Oncogene 29:5545–5555.
de La Coste A, Romagnolo B, Billuart P, et al. 1998. Somatic mutations of the beta-catenin gene are frequent in mouse and human hepatocellular carcinomas. Proc Natl Acad Sci USA 95:8847–8851.
Deiters A, Yoder JA. 2006. Condi-tional transgene and gene targeting methodologies in zebrafish. Zebrafish 3:415–429.
Denovan-Wright EM, Pierce M, Wright JM. 2000. Nucleotide sequence of cDNA clones coding for a brain-type fatty acid binding protein and its tissue-specific expression in adult zebrafish (Danio rerio). Biochim Biophys Acta 1492:221–226.
Denson LA, McClure MH, Bogue CW, et al. 2000. HNF3beta and GATA-4 transactivate the liver-enriched homeobox gene, Hex. Gene 246:311–320.
Dougan ST, Warga RM, Kane DA, et al. 2003. The role of the zebrafish nodal-related genes squint and cyclops in patterning of mesendo-derm. Development 130:1837–1851. Edamoto Y, Hara A, Biernat W, et al.
2003. Alterations of RB1, p53 and Wnt pathways in hepatocellular carcinomas associated with hepatitis C, hepatitis B and alcoholic liver cirrhosis. Int J Cancer 106:334–341.
El-Serag HB, Davila JA, Petersen NJ, et al. 2003. The continuing increase in the incidence of hepato-cellular carcinoma in the United States: an update. Ann Intern Med 139:817–823.
El-Serag HB, Rudolph KL. 2007. He-patocellular carcinoma: epidemiology
and molecular carcinogenesis. Gas-troenterology 132:2557–2576. Farooq M, Sulochana KN, Pan X,
et al. 2008. Histone deacetylase 3 (hdac3) is specifically required for liver development in zebrafish. Dev Biol 317:336–353.
Feitelson MA, Pan J, Lian Z. 2004. Early molecular and genetic determi-nants of primary liver malignancy. Surg Clin North Am 84:339–354. Feitsma H, Cuppen E. 2008.
Zebra-fish as a cancer model. Mol Cancer Res 6:685–694.
Friend SH, Bernards R, Rogelj S, et al. 1986. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 323:643–646.
Gatza ML, Chandhasin C, Ducu RI, et al. 2005. Impact of transforming viruses on cellular mutagenesis, ge-nome stability, and cellular transfor-mation. Environ Mol Mutagen 45:304–325.
Gritsman K, Talbot WS, Schier AF. 2000. Nodal signaling patterns the organizer. Development 127:921– 932.
Hamilton JP. 2010. Epigenetic mecha-nisms involved in the pathogenesis of hepatobiliary malignancies. Epige-nomics 2:233–243.
Hanahan D, Weinberg RA. 2011. Hallmarks of cancer: the next gener-ation. Cell 144:646–674.
Her GM, Chiang CC, Chen WY, et al. 2003a. In vivo studies of liver-type fatty acid binding protein (L-FABP) gene expression in liver of transgenic zebrafish (Danio rerio). FEBS Lett 538:125–133.
Her GM, Yeh YH, Wu JL. 2003b. 435-bp liver regulatory sequence in the liver fatty acid binding protein (L-FABP) gene is sufficient to modulate liver regional expression in trans-genic zebrafish. Dev Dyn 227:347– 356.
Hickman ES, Moroni MC, Helin K. 2002. The role of p53 and pRB in ap-optosis and cancer. Curr Opin Genet Dev 12:60–66.
Higashitsuji H, Itoh K, Nagao T, et al. 2000. Reduced stability of reti-noblastoma protein by gankyrin, an oncogenic ankyrin-repeat protein overexpressed in hepatomas. Nat Med 6:96–99.
Honda K, Sbisa E, Tullo A, et al. 1998. p53 mutation is a poor prognos-tic indicator for survival in patients with hepatocellular carcinoma under-going surgical tumour ablation. Br J Cancer 77:776–782.
Hoshida Y, Nijman SM, Kobayashi M, et al. 2009. Integrative transcrip-tome analysis reveals common mo-lecular subclasses of human hepato-cellular carcinoma. Cancer Res 69:7385–7392.
Hoshida Y, Toffanin S, Lachenmayer A, et al. 2010. Molecular classification and novel targets in hepatocellular LIVER DEVELOPMENT AND CANCER FORMATION IN ZEBRAFISH 169
carcinoma: recent advancements. Semin Liver Dis 30:35–51.
Hsu IC, Metcalf RA, Sun T, et al. 1991. Mutational hotspot in the p53 gene in human hepatocellular carcinomas. Na-ture 350:427–428.
Hu TH, Huang CC, Lin PR, et al. 2003. Expression and prognostic role of tumor suppressor gene PTEN/ MMAC1/TEP1 in hepatocellular carci-noma. Cancer 97:1929–1940. Hunter MP, Wilson CM, Jiang X,
et al. 2007. The homeobox gene Hhex is essential for proper hepatoblast dif-ferentiation and bile duct morphogen-esis. Dev Biol 308:355–367.
Ikegami T. 2009. Transforming growth factor-beta signaling and liver cancer stem cell. Hepatol Res 39:847–849. Ishizaki Y, Ikeda S, Fujimori M,
et al. 2004. Immunohistochemical analysis and mutational analyses of beta-catenin, Axin family and APC genes in hepatocellular carcinomas. Int J Oncol 24:1077–1083.
Isobe M, Emanuel BS, Givol D, et al. 1986. Localization of gene for human p53 tumour antigen to band 17p13. Nature 320:84–85.
Jeng YM, Wu MZ, Mao TL, et al. 2000. Somatic mutations of beta-catenin play a crucial role in the tu-morigenesis of sporadic hepatoblas-toma. Cancer Lett 152:45–51. Jho EH, Zhang T, Domon C, et al.
2002. Wnt/beta-catenin/Tcf signaling induces the transcription of Axin2, a negative regulator of the signaling pathway. Mol Cell Biol 22:1172– 1183.
Kang-Park S, Im JH, Lee JH, et al. 2006. PTEN modulates hepatitis B vi-rus-X protein induced survival signal-ing in Chang liver cells. Virus Res 122:53–60.
Kawakami K, Shima A, Kawakami N. 2000. Identification of a functional transposase of the Tol2 element, an Ac-like element from the Japanese medaka fish, and its transposition in the zebrafish germ lineage. Proc Natl Acad Sci USA 97:11403–11408. Keegan BR, Meyer D, Yelon D.
2004. Organization of cardiac cham-ber progenitors in the zebrafish blas-tula. Development 131:3081–3091. Kelly C, Smallbone K, Brady M.
2008. Tumour glycolysis: the many faces of HIF. J Theor Biol 254:508– 513.
Keng VW, Yagi H, Ikawa M, et al. 2000. Homeobox gene Hex is essen-tial for onset of mouse embryonic liver development and differentiation of the monocyte lineage. Biochem Biophys Res Commun 276:1155–1161. Kim CM, Koike K, Saito I, et al. 1991.
HBx gene of hepatitis B virus induces liver cancer in transgenic mice. Nature 351:317–320.
Kim SY, Hur W, Choi JE, et al. 2009. Functional characterization of human oncoprotein gankyrin in Zebrafish. Exp Mol Med 41:8–16.
Kimmel CB, Ballard WW, Kimmel SR, et al. 1995. Stages of embryonic development of the zebrafish. Dev Dyn 203:253–310.
Kimmel CB, Warga RM, Schilling TF. 1990. Origin and organization of the zebrafish fate map. Development 108:581–594.
Koike K, Shirakata Y, Yaginuma K, et al. 1989. Oncogenic potential of hepatitis B virus. Mol Biol Med 6:151–160.
Kung JW, Currie IS, Forbes SJ, et al. 2010. Liver development, regenera-tion, and carcinogenesis. J Biomed Bio-technol 2010:984248.
Kuroki T, Fujiwara Y, Nakamori S, et al. 1995. Evidence for the pres-ence of two tumour-suppressor genes for hepatocellular carcinoma on chromosome 13q. Br J Cancer 72:383–385.
Lam SH, Gong Z. 2006. Modeling liver cancer using zebrafish: a com-parative oncogenomics approach. Cell Cycle 5:573–577.
Lam SH, Wu YL, Vega VB, et al. 2006. Conservation of gene expres-sion signatures between zebrafish and human liver tumors and tumor progression. Nat Biotechnol 24:73– 75.
Laurent-Puig P, Legoix P, Bluteau O, et al. 2001. Genetic alterations associated with hepatocellular carci-nomas define distinct pathways of hepatocarcinogenesis. Gastroenterol-ogy 120:1763–1773.
Laurent-Puig P, Zucman-Rossi J. 2006. Genetics of hepatocellular tumors. Oncogene 25:3778–3786. Lee JO, Kwun HJ, Jung JK, et al. 2005.
Hepatitis B virus X protein represses E-cadherin expression via activation of DNA methyltransferase 1. Oncogene 24:6617–6625.
Lee MN, Jung EY, Kwun HJ, et al. 2002. Hepatitis C virus core protein represses the p21 promoter through inhibition of a TGF-beta pathway. J Gen Virol 83:2145–2151.
Lee YH, Yun Y. 1998. HBx protein of hepatitis B virus activates Jak1-STAT signaling. J Biol Chem 273:25510– 25515.
Lemaigre F, Zaret KS. 2004. Liver development update: new embryo models, cell lineage control, and morphogenesis. Curr Opin Genet Dev 14:582–590.
Lemmon MA, Schlessinger J. 2010. Cell signaling by receptor tyrosine ki-nases. Cell 141:1117–1134.
Levine AJ, Momand J, Finlay CA. 1991. The p53 tumour suppressor gene. Na-ture 351:453–456.
Levy L, Renard CA, Wei Y, et al. 2002. Genetic alterations and onco-genic pathways in hepatocellular car-cinoma. Ann N Y Acad Sci 963:21–36. Lieschke GJ, Currie PD. 2007. Animal models of human disease: zebrafish swim into view. Nat Rev Genet 8:353–367.
Liew CT, Li HM, Lo KW, et al. 1999. Frequent allelic loss on chromosome 9 in hepatocellular carcinoma. Int J Can-cer 81:319–324.
Luber B, Lauer U, Weiss L, et al. 1993. The hepatitis B virus transactivator HBx causes elevation of diacylglycerol and activation of protein kinase C. Res Virol 144:311–321.
Marques IJ, Weiss FU, Vlecken DH, et al. 2009. Metastatic behaviour of primary human tumours in a zebra-fish xenotransplantation model. BMC Cancer 9:128.
Marrero CR, Marrero JA. 2007. Viral hepatitis and hepatocellular carci-noma. Arch Med Res 38:612–620. Matsuda Y, Ichida T, Matsuzawa J,
et al. 1999. p16(INK4) is inactivated by extensive CpG methylation in human hepatocellular carcinoma. Gastroenter-ology 116:394–400.
McGlynn KA, London WT. 2005. Epi-demiology and natural history of he-patocellular carcinoma. Best Pract Res Clin Gastroenterol 19:3–23. Mirbahai L, Williams TD, Zhan H,
et al. 2010. Comprehensive profiling of zebrafish hepatic proximal pro-moter CpG island methylation and its modification during chemical carcino-genesis. BMC Genomics 12:3. Mirbahai L, Williams TD, Zhan H,
et al. 2011. Comprehensive profiling of zebrafish hepatic proximal pro-moter CpG island methylation and its modification during chemical carcino-genesis. BMC Genomics 12:3. Miyoshi Y, Iwao K, Nagasawa Y,
et al. 1998. Activation of the beta-catenin gene in primary hepatocellu-lar carcinomas by somatic alterations involving exon 3. Cancer Res 58:2524–2527.
Morgan TR, Mandayam S, Jamal MM. 2004. Alcohol and hepatocellular carci-noma. Gastroenterology 127:S87–S96. Morin PJ, Sparks AB, Korinek V, et al. 1997. Activation of beta-cate-nin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 275:1787–1790.
Murakami Y, Hayashi K, Hirohashi S, et al. 1991. Aberrations of the tu-mor suppressor p53 and retinoblas-toma genes in human hepatocellular carcinomas. Cancer Res 51:5520– 5525.
Murata M, Matsuzaki K, Yoshida K, et al. 2009. Hepatitis B virus X pro-tein shifts human hepatic transform-ing growth factor (TGF)-beta signal-ing from tumor suppression to onco-genesis in early chronic hepatitis B. Hepatology 49:1203–1217.
Nakamura T, Nishina H. 2009. Liver development: lessons from knockout mice and mutant fish. Hepatol Res 39:633–644.
Nejak-Bowen KN, Monga SP. 2010. Beta-catenin signaling, liver regener-ation and hepatocellular cancer: sorting the good from the bad. Semin Cancer Biol 21:44–58.
