Molecular alterations in hepatocellular carcinoma associated with hepatitis B and hepatitis C infections

Chronic infections with hepatitis B (HBV) and hepatitis C viruses (HCV) are the leading cause of cirrhosis and hepatocellular carcinoma (HCC) worldwide. Both viruses encode multifunctional regulatory proteins activating several oncogenic pathways, which induce accumulation of multiple genetic alterations in the infected hepatocytes. Gene mutations in HBV- and HCV-induced HCCs frequently impair the TP53, Wnt/b-catenin, RAS/RAF/MAPK kinase and AKT/mTOR pathways, which represent important anti-cancer targets. In this review, we highlight the molecular mechanisms underlying the pathogenesis of primary liver cancer, with particular emphasis on the host genetic variations identified by high-throughput technologies. In addition, we discuss the importance of genetic alterations, such as mutations in the telomerase reverse transcriptase (TERT) promoter, for the diagnosis, prognosis, and tumor stratification for development of more effective treatment approaches.


INTRODUCTION
Liver cancer is one of the most common malignancies in the world, ranking fifth in men and ninth in women in incidence, and second among both sexes in mortality [1]. In 2012, the estimated number of new cancer cases and deaths was 782,000 and 746,000, respectively [1]. The highest incidence has been reported in Eastern and South-Eastern Asia [age-standardized rates (ASR) of 20.9 and 12.3 per 100,000 population, respectively] and Western Africa (ASR 12.1 per 100,000 population), (Table 1). On the other hand, most developed countries have low (ASR <5 per 100,000) or intermediate (ASR 5-10 per 100,000) rates with some exceptions, such as the high incidence (ASR 34.8 cases per 100,000 men) of liver cancer reported in Southern Italy [2].
Hepatocellular carcinoma (HCC) is the most common histological type accounting for approximately 70-85% of primary liver tumors [3]. Chronic HBV and HCV infections represent the major cause of HCC, being associated with more than 80% of cases worldwide [4]. Indeed, pooled estimates of lifetime relative risk to develop HCC are 15 -20 fold higher in HBV or HCV positive patients compared to non-infected subjects [5]. Non-viral risk factors include alcoholic liver disease, nonalcoholic steatohepatitis, aflatoxin B1 dietary exposure, obesity, and diabetes [6][7][8]. The relative contribution of viral and non-viral factors to HCC development varies in different populations. The estimated prevalence of virusrelated HCC is lower in North America (42%) and Europe (48%), and higher in Africa (80%) and Asia (87%) [4,9]. A meta-analysis of hepatitis B surface antigen (HBsAg) and anti-HCV antibody prevalence among 27,881 HCC cases from 36 countries showed a large predominance of HBsAg in Asian, African and Latin American countries and a significant higher frequency of anti-HCV antibodies in Europe and United States [4]. The exception to these patterns is represented by the high rates of HCV-related HCC in Japan and Egypt [4].
The HBV-and HCV-related carcinogenesis initiates in the context of chronic hepatitis, and progresses to HCC in a multistep process lasting for as long as 30 years [10] (Figure 1). During HCC progression, several environmental factors (aflatoxin B1, alcohol consumption, cigarette smoking, hepatotoxic chemical agents) as well as host cowww.impactjournals.com/oncotarget factors (elevated serum androgen levels, genetic polymorphisms, DNA repair enzymes) may synergize and lead to progressive accumulation of multiple genomic changes in the hepatocytes [11,12]. Among these, non-synonymous mutations in TP53 and CTNNB1 genes are well known cancer drivers for HCC development with variable frequencies depending on the underlying etiology [13,14].
Over the last decade, massively parallel sequencing technologies allowed to further uncover the genomic diversity of HCC and to identify consistent gene alterations activating signaling pathways relevant to cell transformation [15,16]. Such analyses allowed to identify HCC subgroups characterized by definite genetic profiles that may be linked to specific oncogenic factors and are useful to further stratify HCCs for personalized medicine applications [17].
Here, we review the molecular pathogenesis of primary liver cancer with particular emphasis on the host genetic variations identified by high-throughput technologies in the context of HBV and HCV related HCC. We discuss the importance of genetic alterations in diagnosis, prognosis as well as in tumor stratification for more efficient treatment approaches.

HBV and hepatocellular carcinoma
HBV is a partially double-stranded hepatotropic DNA virus containing four partial overlapping open reading frames (ORFs) encoding the reverse transcriptase/ polymerase (Pol), the capsid protein (core antigen HBcAg), three envelope proteins (L, M, and S) and the transactivating protein x [18].
HBV infection contributes to hepatocarcinogenesis by different mechanisms including 1) expression of HBx protein; 2) integration of viral DNA into the host genome; and 3) accumulation of somatic mutations  [10,19,20].

HBV HBx protein
The HBV protein HBx transactivates viral and cellular genes by interacting with nuclear transcription factors, such as cyclic adenosine monophosphate(cAMP) response element-binding protein (CREB), activating protein 1 (AP-1), nuclear factor kappa B (NF-kB), and specificity protein 1 (Sp-1). HBx affects also several cellular pathways including DNA repair, cell proliferation, differentiation and apoptosis [20][21][22][23][24]. In addition, HBx protein trans-activates DNA methyltransferase 1 (DNMT1) and DNMT3A genes in the HBV infected hepatocytes, resulting in the suppression of cell cycle regulators P16INK4A and p21 Cip1/CDKN1A, cell-adhesion molecule E-cadherin as well as SFRP1 and SFRP5 genes, which inhibit Wnt signaling pathway [25][26][27][28][29][30]. Moreover, Wnt/β-catenin pathway is directly activated by HBx protein, which interferes with proteasomal degradation of β-catenin [31, 32]. More recently, HBx has been shown to activate the Yes-associated protein (YAP) oncogene, a downstream effector of the Hippo-signaling pathway, which represents a key element in HCC development [33]. The HBx protein can also bind to the p53 oncosuppressor, leading to the disruption of the p53/XPB/XPD complex of the transcriptional factor II H and compromising the nucleotide excision repair mechanism [34]. Recent studies showed that HBx is able to activate AKT, favoring persistent, non-cytopathic HBV replication and inhibition

HBV and aflatoxin B1 interaction
In HBV-associated HCC, there is a strong overrepresentation of TP53 mutations, particularly in geographic regions endemic for HBV and with dietary exposure to aflatoxin B1 (AFB1) [49]. Specifically, AFB1 induces a non-synonymous mutation (G to T transversion) changing arginine to serine at codon 249 of TP53 gene in up to 50% of HCCs. The mutated p53, together with chronic HBV infection, synergistically increase the risk to develop HCC [49,50]. Indeed, the p53 R249S is able to bind the HBx protein and to promote hepatocyte transformation [51].

HCV and hepatocellular carcinoma
HCV is a single-stranded RNA virus encoding a large polyprotein of 3,000 amino acids. The HCV polyprotein can be cleaved by viral and cellular proteases into four structural proteins (capsid protein C, envelope glycoproteins E1 and E2, and protein P7), and six nonstructural proteins (NS2,NS3, NS4A, NS4B, NS5A, and NS5B) [52].
HCV causes chronic hepatitis in more than 80% of infected subjects, versus the 10% in HBV infected patients, and is up to 20 fold more efficient than HBV in promoting liver cirrhosis [53]. Pathogenesis of HCV-related HCC mainly relies on the ability of the virus to cause chronic inflammation, immune-mediated hepatocyte death, tissue damage, fibrosis and evolution to cirrhosis [54-56]. The HCV core protein C as well as the non-structural proteins NS3, NS5A, and NS5B induce hepatocarcinogenesis through their ability to perturb several cellular pathways, such as DNA repair, proliferation and apoptosis [57-59].

HCV core protein
The HCV core protein binds to numerous transcription factors, thus regulating expression of several host genes [60, 61]. In addition, it promotes cell growth and survival by activation of mitogen-activated protein kinase (MAPK) signaling cascade, including MEK1, ERK1/2, JNK, p38 MAP kinases, and MKP1 Map kinases [62][63][64]. HCV core enhances cell proliferation by inhibiting the synthesis of p53, p21 CDK inhibitor, and E2F-1 as well as the phosphorylation of pRb [65]. Moreover, it is able to suppress immune-mediated apoptosis by inhibiting caspase-8 via over-expression of the cellular FADD-like interleukin-1 converting enzyme (c-FLIP) [66]. In addition, it enhances angiogenesis by triggering the production of TGF-β2 and VEGF proteins, and stabilizing the hypoxia-inducible factor 1 (HIF-1a) [67]. HCV core protein induces IL-6, gp130, leptin receptor, and STAT3 over expression, which in turn may deregulate c-Myc and cyclin D1 downstream the STAT3 signaling pathway [68]. HCV core protein also activates the Wnt/b-catenin cascade, which is known to play a significant role in the HCC development [69].

HCV NS3 protein
The HCV NS3 protein is a multifunctional protein with protease, RNA helicase, and NTPase activity. NS3 can promote hepatocarcinogenesis by its binding with certain cellular proteins, such as p21 and p53 [70]. Recently, HCV NS3/4A protease was demonstrated to activate the EGFR signaling pathway through the proteolytic cleavage of tyrosine phosphatase T-cell protein (TC-PTP), resulting in increased EGFR activity and the downstream PI3K/Akt pathway [71]. MAP kinase signaling, through activation of JNK, was also implicated in HCV NS3 protein-mediated cell growth in infected cells [65].

HCV NS5A protein
NS5A has been shown to bind a wide range of cellular proteins controlling signal transduction and host microenvironment [72]. Particularly, the truncated HCV NS5A protein localizes to the nucleus and acts as a transcriptional activator. NS5A can bind cellular signaling components and regulatory protein kinases, leading to the suppression of the host immune response and inhibition of apoptosis [73]. NS5A binds and stabilizes β-catenin, inducing activation of the c-Myc promoter and increased c-Myc expression, which increases production of reactive oxygen species, DNA damage, and cell-cycle deregulation [74,75]. NS5A also stabilizes poly(ADPribose) polymerase 1 (PARP-1), which is involved in DNA repair and apoptosis, thus contributing to genetic instability and accumulation of mutations in HCV-infected hepatocytes [59, 76, 77].

Gene expression profiling in HCC
Early studies on gene expression profiling highlighted the wide heterogeneity of global gene expression patterns in liver tumors [78,79]. Hierarchical clustering analysis of tumor-specific genes contributed to classify HCC subtypes, unravel the complex pathogenesis of HCC and stratify tumors according to their etiological factor, clinical stage, recurrence rate, and prognosis [80][81][82]. Several reports showed strong expression signatures in genes regulating cell proliferation and antiapoptotic pathways (i.e., PNCA and cell cycle regulators CDK4, CCNB1, CCNA2, and CKS2), ubiquitination mechanisms [83,84], as well as molecular markers of tumor progression like HSP70, CAP2, GPC3, and GS [85]. A class-comparison analysis performed in our lab (HCV-related HCC, HCV-related non HCC and metastatic liver tissue vs. normal control; HCV-related HCC vs. autologous HCV-related non HCC liver tissue) identified a gene-set that distinguish the different types of liver disease [86]. In particular, the time course analysis allowed to identify several candidate genes as progression markers (e.g., GPC3, CXCL12, SPINK1, GLUL, UBD, TM4SF5, DPT, SCD, MAL2, TRIM55, COL4A2) [86]. Altogether, these data are useful for developing a specific gene-chip including those genes showing the highest fold increase.

Somatic mutations in HBV and HCV-related HCC
Genomic instability of viral-related HCCs is characterized by high frequency of somatic mutations. Several studies showed that TP53 oncosuppressor and CTNNB1 oncogene are the most frequently mutated genes in primary liver cancer, being identified in about 25% and 30% of HCCs, respectively [13]. Up to 75% of missense TP53 mutations, other than the R249S induced by AFB1, are scattered over 200 codons of the TP53 region encoding for the DNA-binding domain [123][124][125][126], and show similar frequencies in HCCs with different etiologies [13]. Such a finding suggests that chronic inflammation, reactive oxygen species, and oxidative DNA damage, which are common effects of cancer causing factors, may be responsible for such variations. TP53 mutations may cause several pathway deregulations in HCCs. Okada et al. identified 83 genes differentially expressed in TP53 mutant compared to wild type TP53 liver tumors [127]. The genes differentially expressed in TP53 mutant tumors include cell cycle regulators (CCNG2, BZAP45) and cell proliferation-related genes (SSR1, ANXA2, S100A10, and PTMA) [127]. These data support the hypothesis that p53 mutant tumors have higher malignant potentials compared with wild type p53 [128,129].
CTNNB1 gene, expressing β-catenin, and AXIN1 and AXIN2 genes, encoding for components of β-catenin degradation complex, are frequently mutated in liver cancers [130,131,[131][132][133]. Interestingly, CTNNB1 mutations have been shown to occur mainly in alcohol and HCV-related tumors [16,134,135]. Guichard et al. reported CTNNB1 mutations in 11.4% and 33.3% of HBV and HCV-related HCCs, respectively, and in 41.8% of alcohol-related HCCs [16]. In addition, mutations in CTNNB1 and TP53 genes appear to be mutually exclusive, suggesting that inactivation of either pathway is sufficient to induce cell transformation [16].
The frequency of mutations in different genes seems related to the cancer etiology. TP53 gene was mostly mutated in HBV-related HCC, while CTNNB1, TERT, CDKN2A, SMARCA2, and HGF genes were mainly mutated in alcohol-related HCCs, and IL6ST was mutated in HCCs with no known etiology. Conversely, no specific gene mutation was associated with HCV infection, metabolic syndrome and hemochromatosis [138][139][140].
Somatic mutations in the TERT promoter have been identified as the first recurrent genetic alteration in 25% of dysplastic cirrhotic nodules [141,142]. These mutations create a consensus binding sequence for a ternary complex factor and induce expression of telomerase reverse transcriptase [148]. Conversely, several other genes known to be recurrently mutated in HCC, including CTNNB1, TP53, ARID1A, ARID2, RPS6KA3, NFE2L2 and KEAP1 were not mutated in dysplastic nodules [141]. TERT promoter mutations may be considered as biomarkers for the identification of premalignant lesions developed in cirrhosis patients with a high risk of progression to HCC.

Immunotherapeutic approaches and gene mutations in HCC
In several cancer types, the activity of tumor antigen-specific T-cells is tightly regulated by the balanced expression of stimulatory and inhibitory molecules defined as "immune checkpoints" [144,145]. Therapies targeting these checkpoints, such as those directed against cytotoxic T-lymphocyte antigen 4 (CTLA-4) and programmed death 1 receptor (PD-1), have shown to be more effective in cancers characterized by high rates of somatic mutations [146]. Recent studies have indicated that a high tumor mutation burden increases responsiveness to CTLA-4 inhibition in melanoma, to PD-1 inhibition in non-small cell lung cancer and in mismatch repair-deficient colorectal cancers [147]. The hypothesis is that the higher number of genetic variations leads to a greater number of mutated epitopes in tumor proteins (neoantigens). Such neoantigens may be characterized by an improved MHCbinding profile, resulting in superior presentation to T cells for eliciting a stronger cytotoxic response [148,149]. Very recently, this has been experimentally proven in animal models [150,151] as well as in melanoma patients treated with the anti-CTLA-4 monoclonal antibody, ipilimumab [152,153].
Liver sinusoidal endothelial cells express high levels of the inhibitory molecule program death receptor ligand 1 (PD-L1) and low levels of the co-stimulatory molecules CD80 and CD86, thereby limiting their ability to effectively activate CD4-positive (CD41) and CD8 1 T lymphocytes [154,155]. Immune checkpoint inhibitors have been recently evaluated in HCC patients. The anti-CTLA-4 monoclonal antibody tremelimumab showed a safe profile and antitumor activity in HCC patients with chronic HCV infection [156]. Very recently, results from a phase I/II clinical trial (ClinicalTrials.gov Identifier: NCT01658878) presented at the last 2015 ASCO Meeting showed that nivolumab, a fully humanized IgG4 monoclonal antibody to PD-1, may be a promising treatment for patients with advanced HCC [157]. Indeed, the overall survival at 1 year was 62% and overall objective responses rate was 19%, including complete response (CR) in 5% and partial response in 14% of enrolled patients. Such responses are significantly higher compared to responses to the kinase inhibitor sorafenib, the current standard of care for late stage HCC.
Further studies are needed to evaluate whether the immune responses elicited by mutated epitopes could lead to an increased efficacy of anti immune-checkpoint therapies also in liver cancer.

CONCLUSIONS
Classification of liver cancers in homogeneous subgroups characterized by specific molecular alterations is an important tool for the application of personalized therapies. Several commonly altered pathways have emerged following the integration of data obtained with multiple high-throughput analyses. Common oncogenic drivers, differentially represented in HCCs with different etiologies, include genetic alterations affecting TERT, Wnt/beta-catenin, JAK/STAT and PI3K-AKT-mTOR pathways. Drugs targeting these pathways are now available and have been approved in clinical trials.

CONFLICTS OF INTEREST
The authors have no conflicts of interest to disclose.