Host and Viral Genetic Variation in HBV-Related Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is the fifth most common cancer in men and the second leading cause of cancer deaths globally. The high prevalence of HCC is due in part to the high prevalence of chronic HBV infection and the high mortality rate is due to the lack of biomarkers for early detection and limited treatment options for late stage HCC. The observed individual variance in development of HCC is attributable to differences in HBV genotype and mutations, host predisposing germline genetic variations, the acquisition of tumor-specific somatic mutations, as well as environmental factors. HBV genotype C and mutations in the preS, basic core promoter (BCP) or HBx regions are associated with an increased risk of HCC. Genome-wide association studies have identified common polymorphisms in KIF1B, HLA-DQ, STAT4, and GRIK1 with altered risk of HBV-related HCC. HBV integration into growth control genes (such as TERT), pro-oncogenic genes, or tumor suppressor genes and the oncogenic activity of truncated HBx promote hepatocarcinogenesis. Somatic mutations in the TERT promoter and classic cancer signaling pathways, including Wnt (CTNNB1), cell cycle regulation (TP53), and epigenetic modification (ARID2 and MLL4) are frequently detected in hepatic tumor tissues. The identification of HBV and host variation associated with tumor initiation and progression has clinical utility for improving early diagnosis and prognosis; whereas the identification of somatic mutations driving tumorigenesis hold promise to inform precision treatment for HCC patients.


INTRODUCTION
Hepatocellular carcinoma (HCC) is the fifth most common cancer in men and the second leading cause of cancer deaths worldwide (El-Serag, 2011;Torre et al., 2015). HCC prevalence is highest in East and Southeast Asia and sub-Saharan Africa, but the incidence rates of HCC have increased in the United States and Western Europe over the past few decades (Lee, 2015;Njei et al., 2015). Early diagnosis and surgical resection remain the key to potential curative treatment; however, most HCC patients present with late stage tumors and have poor prognosis. HCC surveillance is mainly based on sonography and alpha-fetoprotein (AFP) measurement, both of which lack sufficient sensitivity and specificity (Yim and Lok, 2006;Bruix et al., 2011).
Hepatitis B virus (HBV) is one of the leading risk factors for HCC, especially in HBV endemic areas. Despite the existence of an effective vaccine, about 257 million people, or 3.5% of the global population, were afflicted with chronic HBV infection in 2015, WHO (2017). The clinical spectrum of chronic HBV infection ranges from asymptomatic carrier status to chronic hepatitis B (CHB), which may evolve to liver cirrhosis and HCC (EASL, 2012). An estimated 8-20% of untreated adults with CHB will develop liver cirrhosis within 5 years (Terrault et al., 2016) and 2-8% of those with cirrhosis develop HCC annually (Yim and Lok, 2006;Bruix et al., 2011;Hung et al., 2017).
Environmental and host risk factors for HCC include childhood acquisition of HBV infection, cirrhosis, aflatoxin B1 exposure, heavy alcohol use, smoking, male sex, advancing age, obesity, type 2 diabetes and an impaired immune response (Bruix et al., 2016). In addition to HBV, hepatitis C virus (HCV) chronic infection is a major cause of HCC, but co-infection by HBV and HCV do not seem to confer greater HCC risk than HBV or HCV monoinfection in a meta-analysis (Cho et al., 2011). Co-infection of HBV and hepatitis delta virus (HDV), occurring in ∼5% of those HBV infected, was associated with 3 to 6-fold higher incidence of HCC in longitudinal cohorts (Fattovich et al., 2000;Romeo et al., 2009;Ji et al., 2012;Kushner et al., 2015).
Family clustering and incidence differences among different ancestry groups suggests that inherited genetic factors may contribute to HCC risk (Shen et al., 1991). Common and rare single nucleotide polymorphisms (SNPs) and structural genomic changes may predispose or restrict HCC development. The identification of moderate to high penetrant genetic variants associated with HBV-related HCC might identify HCC susceptible patients that would lead to earlier intervention and better outcomes (Thomas et al., 2009;Brennan et al., 2011;Michailidou et al., 2015). For example, screening for BRCA1 and BRCA2 mutations is routinely used to identify women at risk for breast and ovarian cancer and is routinely used in precision treatment protocols (Brennan et al., 2011;Michailidou et al., 2015).
Somatic mutations occurring in hepatocytes as the consequence of exogenous and endogenous mutagenic factors, are likely involved in HCC initiation and progression, as they are in many other cancers. Identification of cancer diver mutations in HCC are essential for HCC prognostics and for precision therapy targeting perturbed pathways.
Here, we review recent advancements and identify knowledge gaps in HBV-related HCC genetics at three levels: the human host, the virus, and somatic mutations in liver tumors and their associations with clinical outcomes.
The P region encodes the polymerase/reverse transcriptase, which is involved in genome replication. The preC/C codes the nucleocapsid hepatitis B core antigen (HBcAg) or the hepatitis B e antigen (HBeAg) translated from initiated codons at the core or precore regions, respectively. HBcAg and HBeAg are biomarkers for HBV active infection or infectivity. The X ORF encodes a nonstructural protein (HBx) with multiple functions in viral replication and oncogenic activity (Liang, 2009).

Molecular Mechanisms of HBV-Related HCC
There are at least three prevailing mechanisms proposed for the development of HBV-related HCC (Kremsdorf et al., 2006;Block et al., 2007;Hai et al., 2014; Figure 1). First, chronic inflammation and regeneration of hepatocytes during chronic HBV infection may lead to the accumulation of genetic alterations that confer cell growth advantage. Second, the integration of HBV DNA into the host genome may activate the host genes controlling cell proliferation and cause genomic instability. Finally, HBV proteins, mainly HBx, may promote cell proliferation (Kremsdorf et al., 2006;Block et al., 2007;Hai et al., 2014). It is also likely that all three mechanisms contribute to HCC development.

HBV Viral Load
The general consensus is that persistent high-level HBV replication poses greater risk of developing liver cirrhosis and HCC (Sanchez-Tapias et al., 2002;Chen et al., 2006aChen et al., ,b, 2007Fattovich et al., 2008). A large prospective study, which followed 3653 HBsAg positive participants enrolled in the Taiwanese Reveal-HBV cohort for over a decade, found that HBV DNA levels at study entry were positively correlated with incidence of HCC in a dose-dependent manner. Individuals with HBV levels greater than 1 million copies /mL were 10-fold more likely to develop HCC than those with less than 300 copies/mL (Chen et al., 2006b). Serum HBV DNA viral load is also associated HCC tumor recurrence (Hung et al., 2008;Wu et al., 2009). A viral load of greater than 10,000 copies/mL (2000 IU/mL) was independently associated with HBV-related HCC recurrence in patients who underwent liver resection (Hung et al., 2008). Antiviral therapy in these patients decreased tumor recurrence (Li et al., 2010).

HBV Genotype
Ten HBV genotypes (A to J) that diverge by >8% of their nucleotide sequences have been identified globally. HBV genotypes distribute within distinct geographic regions and ethnic populations (Lin and Kao, 2015). Phylogenetic analysis of HBV genotypes indicate that global distribution of HBV genotypes corresponds with the major prehistoric and modern human migration patterns, after HBV established infection in humans around 33,000 years ago (Paraskevis et al., 2013). Different genotypes prevail in the two regions with the highest HBV and HCC prevalence: genotypes B and C are prevalent in East Asia where vertical transmission is predominant, whereas genotypes A and D are prevalent in sub-Saharan Africa where FIGURE 1 | Viral and host genetic factors involved in the development of HBV-induced HCC. HBV infection, if not resolved, may develop to chronic hepatitis and progress to liver cirrhosis and subsequently HCC. Molecular mechanisms of HBV-related HCC involve (1) chronic inflammation and regeneration of hepatocytes; (2) accumulation of genetic alterations that confer cell growth advantage; (3) integration of HBV DNA into the host genome and activation of host genes controlling cell proliferation; (4) genomic instability; and (5) direct promotion of cell proliferation by viral proteins (mainly HBx). The development of HCC is the consequence of the interaction of environmental factors (e.g., aflatoxin), HBV viral factors (genotypes, HBV DNA levels and HBV mutants) and host genetic susceptible variants (e.g., HLA-DP variants affecting HBV clearance), along with somatic mutations (TP53, TERT, gene deletion/amplification) disrupting cell growth control. HBV genotype C and mutations in the preS, basic core promoter or HBx gene are associated with an increased risk of HCC. horizontal transmission is more common. HBV genotypes A and D are the main genotypes in low HBV prevalence regions including Europe and North America (Lin and Kao, 2015).
The risk of HCC appears to differ by HBV subtypes. In East Asia, HBV genotype C is associated with higher risk of HCC compared to genotype B (Chan et al., 2004;Yu et al., 2005;Yang et al., 2008;Tseng et al., 2012;Vutien et al., 2013;Lin and Kao, 2015;Raffetti et al., 2016). Serum HBV viral load is significantly higher in patients with HBV genotype C than in those infected with genotype B (Chu et al., 2002;Chan et al., 2004;Yu et al., 2005;Lin and Kao, 2015). The increased risk of HCC might be the consequence of longer exposure to high levels of HBV in patients with genotype C Ni et al., 2004). In contrast, HCC risk for genotype D and A appear to be similar, although there are reports that patients with HBV genotype A experience longer sustained remissions (Sanchez-Tapias et al., 2002) and are more responsive to interferon alpha (IFNα) treatment compared to those infected with HBV genotype D (Erhardt et al., 2005). There are at least 5 co-circulating HBV genotypes in Alaska Native people, which offers rare opportunity for direct comparisons of HCC risk among genotypes. A longitudinal cohort of Alaska Natives followed for over 30 years reported that carriers of genotypes A (OR 4), C (OR 16), and F (OR 14) are at higher risk compared with genotypes B or D (Livingston et al., 2007;Ching et al., 2016). A study of 1,000 Alaska Native children and young adults followed for over 23 years found that HBV genotype F1 was associated with the highest risk for HCC . These studies support a strong role for HBV genotypes in predicting HCC risk. In comparison, HCV genotype 3 was also associated with higher risk of HCC compared to other HCV genotypes (van der Meer et al., 2012;Kanwal et al., 2014).

HBV Mutations
HBV DNA replication errors occur at a much higher rate than for other DNA viruses because HBV reverse transcriptase lacks a proofreading function (Chotiyaputta and Lok, 2009). The estimated nucleotide substitution rate was 2.2 × 10 −6 substitutions/site/year, which is higher than other DNA viruses but lower than RNA viruses (Paraskevis et al., 2013). Mutations in the HBV genome commonly associated with HCC are summarized in Table 1, with the preC/C and basic core promoter (BCP) regions harboring the most frequent mutations.

Pres/S Region Mutants
PreS mutations, especially the preS1 deletion, preS2 deletion, and preS2 start codon mutations, may induce an unbalanced production of envelope proteins that accumulate in the endoplasmic reticulum (ER) of the hepatocytes, leading to ER stress, oxidative DNA damage and genomic instability (Pollicino et al., 2014). These cytotoxic effects cause liver cell damage and a Frequency ranges were extracted from studies included in the meta-analyses (Liu et al., 2009;Yang et al., 2015). b Pooled RR from reference Yang et al. (2015).
regeneration (Pollicino et al., 2014). A meta-analysis of 9 crosssectional studies including 388 HCC cases demonstrated that any one of the preS mutants (C1653T, T1753V, and A1762T/G1764A) was associated with a 3.8-fold increased risk of HCC (Liu et al., 2009), consistent with the result from a recent meta-analysis of prospective studies comprising 360 cases (RR 3.8) (Yang et al., 2015). PreS/S deletions and the Pre S2 start codon mutation had a RR 4.0 and 2.6, respectively (Yang et al., 2015). Specifically, C2964A, C3116T, and C7A were independently associated with an increased risk of HCC (Yin et al., 2010). Mutations introducing stop codons in the S genomic region have been proposed to enhance tumor development by encoding truncated proteins with transcriptional transactivation activity (Pollicino et al., 2014). The premature stop codon mutations at position 172 or 182 in the S genomic region are associated with higher risk for liver cirrhosis and HCC Pollicino et al., 2014).

PreC/C Region Mutants
Mounting evidence supports the importance of the naturally occurring basal core promoter (BCP) double mutation, A1762T/G1764A, as a risk factor for HCC (Fang et al., 2008;Yang et al., 2008Yang et al., , 2015Yang Z. et al., 2016). In a metaanalysis of 15 prospective studies comprising 1336 HCC cases, individuals with the double mutation had a 3-fold higher risk of developing HCC (Yang et al., 2015). In a more recent meta-analysis that included 3729 HCC cases, double mutation carriers had a 5-fold higher HCC risk (Yang Z. et al., 2016). Other prospective long-term follow-up studies found that the incidence of HCC in double mutation carriers was 4 to 5-fold higher (Fang et al., 2008;Yang et al., 2008). The presence of the BCP double mutation reduces HBeAg production and viral load before anti-HBe seroconversion (Buckwold et al., 1996;Fang et al., 2009), suggesting that its oncogenic mechanism is not directly attributable to viral replication level. However, the preC stop codon G1896A, which prevents the production of HBeAg, does not appear to affect HCC risk as demonstrated in a recent meta-analysis with 600 HCC cases (Yang et al., 2015).

Polymerase/Reverse Transcriptase Region Mutants
Anti-HBV nucleos(t)ide analogs (NAs) are effective in lowering HBV DNA levels through their inhibition of HBV polymerase/reverse transcriptase (Terrault et al., 2016). However, the emergence of and the selection for resistance mutations in the polymerase/reverse transcriptase region confer a virial survival advantage and is a major barrier to the success of NAs treatment. In a longitudinal study of lamivudine (LMV)-resistant CHB patients, emergence of the rtA181T/sW172X mutant in LMV-resistant patients increased the risk of HCC Yeh et al., 2011). NIH3T3 cells expressing this mutant showed greater oncogenic potential in nude mice (Lai and Yeh, 2008;Lai et al., 2009). It is an open question whether resistance to other nucelos(t)ide analogs also increases the risk of HCC.

X Region Mutants
HBx protein is a transcriptional activator of various host cellular genes involved in growth control, DNA repair, and epigenetic modification (Xu et al., 2014). It activates the Ras/Raf/mitogenactivated protein (MAP) kinase pathway, which is involved in hepatocarcinogenesis, and interacts with the tumor protein 53 gene (TP53), interfering with its function as a tumor suppressor (Di Bisceglie, 2009).
Certain point mutations in the HBx gene, in particular, the K130M and V131I double substitutions, are more frequent in patients with liver cirrhosis and/or HCC than in patients with chronic hepatitis B (Baptista et al., 1999). Due to the partial overlap of BCP with Hbx, the same nucleotide substitutions result in both HBx K130M and V131I and precore BCP A1762T/G1764A changes. Increasing number of HBx mutations is correlated with increased risk of HCC. The double xK130M+xV131I and triple xK130M+xV131I+ xV5M mutations are associated with a 4-5-fold increased the risk of HCC . Substitutions at position 10, 30, 38, 88, 94, and 144 have also been reported to be associated with HCC (Muroyama et al., 2006;Lee et al., 2011;Wang et al., 2012;Shi et al., 2016; Table 1). Even though the precise mechanism remains to be elucidated, HBx plays a central role in hepatocarcinogenesis and could be an attractive therapeutic target for HCC suppression.
In summary, major HBV viral factors associated with increased risk of HCC include genotype C and F, higher HBV-DNA levels, mutations in the preS/S region, the double mutation A1762T/G1764A in the basic core promoter, and the double or triple mutation K130M, V131I xV5M in the HBx gene. Different genotypes of the virus (which are due to the accumulation of viral mutations over time) have differing propensities for the development of HCC. Epidemiological association data and some functional evidence suggest the likely causal role of viral mutations, particularly for genotype differences. However, the temporal and causal relationship of HBV mutations with HCC remains to be firmly established, as most results were obtained from case-control and cross-sectional studies and rarely from prospective longitudinal HBV cohorts. HBV mutations may also increase risk of HCC indirectly by the acquisition of mutations that leads to immune escape and increased viral replication. Higher HBV burden would contribute to chronic inflammation promoting the destruction and regeneration of HBV infected hepatocytes mediated by host cytotoxic T lymphocyte (CTL) immune response and increasing the opportunity for replication errors during cell mitosis (Block et al., 2007). The emergence and accumulation of genetic and epigenetic alterations with cell growth advantage eventually lead to hepatocarcinogenesis (Block et al., 2007). The exact molecular mechanisms by which HBV mutations promote hepatocarcinogenesis warrants further investigation.

HOST GENETIC VARIATION AND HBV-RELATED HCC
Although the incidence of HCC among HBV carriers is much higher (223-fold) than among non-carriers (Beasley et al., 1981), only a fraction of patients with chronic HBV infection develop HCC. Results from twin studies, family clustering studies and incidence differences between continental ancestry groups suggest that host genetic factors contribute to HCC susceptibility (Shen et al., 1991). Genome-wide association studies (GWAS) provide an agnostic (hypothesis-free) method to identify susceptibility and resistance loci for HBV-related HCC. The ability to securely identify genetic susceptibility loci associated with HCC is critically dependent on sample size for sufficient power to detect associations with small to moderate effect sizes in GWAS. This is due to the need to correct for multiple testing that tends to inflate false positive signals; therefore, the human genetics community has determined that a genome-wide significance threshold of p < 5 × 10 −8 and independent replication association is required to firmly establish a genotype-phenotype association. The top associations between host genetic variants with HCC are summarized in Table 2.
Only associations in the HLA class II region have been found to replicate in at least two studies. These studies, performed primarily in East Asians, suggest that no common, moderate to high penetrant alleles contribute to HCC development.
The HCC association of intergenic SNPs near HLA-DQA1 and HLA-DQA2, and HLA-DQB1, encoding the alpha and beta chains of the HLA-DQ protein, were identified by GWAS in two large studies Jiang et al., 2013). HLA class II proteins, comprising HLA-DP, DQ, and DR proteins, play a central role in extracellular antigen presentation to CD4 T cells stimulating B cell that lead to antibody production against foreign pathogens, including HBV and HCV.
Outside of the HLA region, genetic associations with HBVrelated HCC have been reported for several genes. A two-stage GWAS composed of five cohorts from East Asian identified the intronic rs17401966 SNP in the KIF1B gene (OR = 0.61) with genome-wide significance in a meta-analysis . KIF1B encodes a kinesin superfamily member involved in the transport of organelles and vesicles. Both germline and somatic loss-of-function mutations in the KIF1B isoform have been detected in multiple cancers . However, this association failed to replicate in multiple independent studies, several of which were well-powered ( Table 2; Hu et al., 2012;Li et al., 2012;Sawai et al., 2012;Jiang et al., 2013;Sopipong et al., 2013). The latest meta-analysis of 12 cohorts found statistical significance only when the original 5 cohorts reporting positive associations were included, but when these 5 cohorts were removed the association was abrogated (Su et al., 2017). Further validation and functional studies are warranted to support or not a role of KIF1b variation in HCC.
Other candidate genes implicated by GWAS include STAT4, and GRIK1 ( Table 2; Li et al., 2012;Jiang et al., 2013). STAT4 is a transcription factor involved in development of Th1 cells and production of IFN-γ, a cytokine with antiviral and antitumor activities. STAT4 rs7574865 is an expression quantitative trait locus (eQTL) with dosage effect in HCC tissues (Jiang et al., 2013). Glutamate receptor GRIK1, is involved in cancer development , though its functional role in HCC has not been experimentally demonstrated. These associations have not yet been replicated or validated by functional assessment. Recently, a functional genomic approach revealed a c-Myc binding SNP that regulates a putative tumor suppressor gene EPB41 and was associated with predisposition to HCC (Yang et al., 2016b). A 2011 experts ′ commentary highlighted the need for further GWAS studies of HCC with larger subject enrollment, clearly delineated phenotypes, and replication . The lack of replication for HCC GWAS studies is most likely due to relatively small sample sizes of the study although genetic and phenotypic heterogeneity may have also contributed. True causal variants should have similar effects across all populations while marker SNPs may have different effects in different ancestry or ethnic populations. These differences in effect size or statistical significance in haplotype structure among different populations, the inclusion of trans-ethnic or continental racial populations can be used to refine position of the causal locus by finemapping (Franceschini et al., 2012). The newer population specific genotyping arrays, in combination with larger casecontrol groups representing diverse populations, should reveal additional germline variants that contribute to the observed variance in HBV-HCC susceptibility. Finally, whole genome or exome sequencing might also identify causal rare variants associated with HBV-HCC. These studies should identify critical pathophysiological pathways for HBV-related liver cirrhosis and HCC.

INTERACTION OF HOST GENES AND HBV: HBV DNA INTEGRATION, SOMATIC MUTATIONS AND EPIGENETIC MODIFICATION IN HCC ONCOGENESIS
Environmental factors, HBV genotypes and genetic variation, and host genetic variation are involved in the development and progression of HCC. The interplay between these factors might lead to the initiation of HCC (Figure 1).

HBV Integration and HCC
HBV DNA integration into the host genome is one of the proposed molecular mechanisms of hepatocarcinogenesis. HBV-DNA integration occurs during both the acute and chronic stages and integrated HBV-DNA is detectable in 75-90% of HCC tissues (Murakami et al., 2005;Zhao et al., 2016). HBV integration events may cause direct gene disruption, HBV promoter-driven transcription of host genes, viral-host transcript fusion and induce genome instability (Tao et al., 2011;Jiang et al., 2012;Sung et al., 2012), which may lead to activation of proto-oncogenes or inactivation of tumor suppressor genes. Oncogenic activity of the cellular and viral genes resulting from the HBV integration confers a selective growth advantage to cells with accumulation of genetic defects, leading to hepatocarcinogenesis (Tao et al., 2011;Jiang et al., 2012;Sung et al., 2012).
Quantification of HBV integration sites may have prognostic value in predicting HCC survival. In a genome-wide survey of 88 HCC patients, increasing number of HBV integrations was correlated with HCC occurrence at younger age and shorter survival (Sung et al., 2012). A viral-human chimeric HBx-LINE1 RNA hybrid, detected in 23% of HCC tumors, was also predictive of poorer survival (Lau et al., 2014).

Somatic Mutations and Copy Number Alteration in HCC
Most somatic mutations are harmless passenger mutations accumulated in the process of tumor growth, which occur at random and confer no selective advantage for tumor cells. Driver mutations confer selective growth advantage and cause transformation of a normal cell to a cancer cell, which are clinically relevant. Recurrent coding changes in multiple tumor 3 | HBV integration into host genes in HCC tissues from next generation sequencing studies.

Telomere Maintenance
The TERT gene encodes a catalytic subunit of telomerase that maintains genomic integrity. TERT expression is repressed in somatic cells, but not in proliferative cells in self-renewing tissues and cancers. Somatic mutations in the TERT promoter are found across multiple cancer types. Immortality associated with cancer cells has been attributed to telomerase over-expression. TERT promoter mutations create a potential binding site for E-twenty-six/ternary complex factors (ETS/TCF) transcription factors and are associated with increased promoter activity, increased expression of TERT and increased telomerase activity (Heidenreich et al., 2014;Bell et al., 2015;Borah et al., 2015).
TERT promoter mutations are the most frequent somatic genetic alterations observed in HCC, with an overall prevalence of approximately 60%, with ranges from 30 to 40% for HBVrelated HCC to 60 to 80% for HCV-related HCC (Nault et al., 2013;Totoki et al., 2014;Fujimoto et al., 2015;Schulze et al., 2015;Kawai-Kitahata et al., 2016;Yang et al., 2016a). HCC with HCV infection and alcohol intake more often harbor TERT promoter mutations than those with HBV infection (Nault et al., 2013). The lower rate of TERT promoter mutations in HBV-related HCC could be partially explained by the frequent insertion of HBV DNA in the TERT promoter serving as additional mechanism inducing telomerase transcription (Nault et al., 2013).

Relationships Between Mutated Genes
Certain subsets of altered genes share the same pathways or interact, contributing to the complexity and heterogeneity of hepatocarcinogenesis. Activated mutations of CTNNB1 are significantly associated with mutations in the TERT promoter (Nault et al., 2013;Totoki et al., 2014). It has been proposed that TERT might be a direct target of CTNNB1 (Hoffmeyer et al., 2012;Zhang et al., 2012). Alterations in RPS6KA3 are frequently associated with AXIN1 mutations, suggesting cooperation between RPS6KA3 inactivation and Wnt/β-catenin activation in tumorigenesis (Guichard et al., 2012). On the other hand, a number of mutations appear to be mutually exclusive and rarely appear together in the same tumor, such as CTNNB1 mutations with TP53 mutations (Guichard et al., 2012;Ahn et al., 2014) and AXIN1 mutations with CTNNB1 mutations (Satoh et al., 2000). ARID1A/ARID2 mutations are negatively associated with mutations in TP53 (Lee, 2015). The consequences of network interactions between driver mutations may offer deeper insight into tumorigenesis.

HCC GENETICS AND PRECISION MEDICINE Somatic Mutations and Therapeutic Targets of HCC
Targeted therapy based on genomic alterations is a core tenet of precision treatment. Exome sequencing analysis of 243 liver tumors found that 28% of patients harbor at least one alteration potentially targetable by an FDA-approved drug, and 86% harbored a mutation targetable by a drug studied in phase I to phase III clinical trials (Schulze et al., 2015). β-catenin reduction in CTNNB1-mutated HCCs in a murine model led to complete tumor response, showing a clear benefit of therapeutic targeting of this molecule (Delgado et al., 2015). Adenovirus mediated gene transfer of wild-type AXIN1 induced apoptosis in hepatocellular cancer cells that had accumulated β-catenin as a consequence of mutations in APC, CTNNB1, or AXIN1 genes, suggesting that AXIN1 may be an effective therapeutic molecule for suppressing HCC growth (Satoh et al., 2000).

Somatic Mutations and HCC Prognosis
Somatic mutations of driver genes may be predictive of HCC prognosis. In a study of over 300 HCC Chinese patients, TP53 hotspot mutations (R249S and V157F were strongly associated with decreased overall survival, indicating that these mutations can be used as prognostic markers in HCC in patients at risk for high aflatoxin exposures (Woo et al., 2011). Mechanistically, a synergistic interaction of aflatoxin B1 induced TP53 mutations together with HBV chronic inflammation may advance the development of HCC. In a multivariate analysis of 231 Korean HCC patients, the RB1 somatic mutation was the only independent prognostic factor for reduced cancer-specific survival and accelerated recurrence (Ahn et al., 2014). In a cohort of resected HCCs, CDKN2A inactivation was associated with poor prognosis (Schulze et al., 2015). These predictive and prognostic molecular markers may have clinical utility for personalized treatment plans.

CONCLUSION, CHALLENGES AND FUTURE DIRECTIONS
The development of HCC is multifactorial with viral, host and environmental factors contributing to chronic inflammation, cirrhosis and ultimately, hepatocarcinogenesis (Figure 1). Interplay between host germline variants, persistent high HBV viral load, viral genotypes and mutations, HBV integration into host chromosomes, the oncogenic potential of the HBx protein, and the occurrence of somatic cancer driver mutations may contribute independently or jointly to the oncogenesis of HBV-related HCC (Figure 1). We hope that eventually genetic profiling of the virus and host will identify the individuals who are at higher risk of HCC and those who will benefit most from HCC treatment options. Since early HCC is largely asymptomatic and biopsies are rarely performed; there is a paucity of pre-tumor and early-stage tumor tissue available, a major impediment for genetic interrogation. This is in contrast to breast, colon, and prostate cancers where biopsies are standard of care for patients with abnormal findings during routine screening. Validated genetic biomarkers would have utility for early diagnosis, molecular classification of tumors, and prognostics as well as identify new targets for drug interventions. With the advent of ever larger HCC cohorts, denser, population-specific genotyping arrays, and next generation whole exome/genome sequencing of HCC family clusters, it may be possible to provide earlier diagnosis of HCC and to develop bespoke treatment for persons with HCC to improve outcomes.

AUTHOR CONTRIBUTIONS
PA and CW: conceived and wrote the paper; JX: wrote and revision of the manuscript; YY: revision of the manuscript.

FUNDING
This project has been funded in whole or in part with federal funds from the National Cancer Institute, National Institutes of health, under contract HHSN26120080001E. This Research was supported in part by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

ACKNOWLEDGMENTS
We thank George Nelson and Victor David for proofreading.