Multiple roles for hepatitis B and C viruses and the host in the development of hepatocellular carcinoma

Abstract

Chronic hepatitis B and C viral infections are major risk factors for hepatocellular carcinoma (HCC) in the United States and worldwide. Direct and indirect mechanisms of viral infection lead to the development of HCC. Chronic viral infection leads to inflammation and liver damage, culminating in cirrhosis, the penultimate step in the progression toward HCC. Host, viral, and environmental factors likely interact to promote oncogenesis. Clinical considerations include recommendations for screening for HCC in persons at risk, treatment with antivirals, and an emerging role for immunotherapy in HCC. We pose unanswered questions regarding HCC susceptibility and pathogenesis in the setting of chronic hepatitis B and C.

I. Introduction

Hepatitis B virus (HBV) and hepatitis C virus (HCV) cause persistent liver injury and regenerative responses that can progress to liver cirrhosis and hepatocellular carcinoma (HCC). HCC is an aggressive form of cancer that arises from hepatocytes, the main parenchymal cell type of the liver (1, 2). Globally, HCC ranks 6^th in cancer incidence but 2^nd in estimated number of cancer deaths in 2018, according to the International Agency for Research of the Cancer in the World Health Organization. In the United States, the incidence of HCC has been rising in the past decades, with 5 year survival estimated at only 18% (3). At least 60% of HCC can be attributed to underlying chronic hepatitis B (CHB) or chronic hepatitis C (CHC) (4). Here, we review mechanisms of HCC pathogenesis in the setting of CHB and CHC, and their relevance to clinical management and therapy. Figure 1 provides a schematic overview.

Figure: Multiple factors are involved in hepatocarcinogenesis with chronic HBV and HCV infection.

The vast majority of hepatocellular carcinomas (HCCs) develop in the setting of cirrhosis, whereas a fraction of HCCs can develop without cirrhosis in association with HBV or AAV2 genomic viral integration events. Cirrhosis develops after many years of necroinflammatory responses to viral infections, and is accelerated by alcohol, viral co-infection, and fatty liver disease. Cirrhosis is associated with epigenetic changes that drive the development of HCC. The earliest somatic mutations linked to HCC include disruption of p53, activation of β-catenin, and/or promoter mutations that activate TERT. Progression is associated with increasing load of mutations in tumors. Additional factors associated with HCC include male sex, hereditary factors, viral factors, and environmental toxins such as aristolochic acids and aflatoxin.

II. Pathogenesis of HCC in the setting of viral infection

CHB and CHC share common pathways in HCC pathogenesis, related to chronic hepatocellular injury and turnover in the setting of inflammation. However, there are also important differences in mechanisms of infection, co-factors and host genetics.

A. Mechanisms for HCC common to HBV and HCV

Maladaptive immune activation promotes HCC:

Both HBV and HCV are non-cytopathic viruses. As such, hepatocellular injury during HBV or HCV infection is largely mediated by host immune responses (5). While both viruses circumvent innate immune responses in early infection, MHC class I-restricted CD8 T-cells mediate both hepatocellular injury and virus control by directly killing virus-infected hepatocytes and secreting antiviral cytokines such as IFN-gamma and TNF. In addition, B-cells control viral spread by secreting neutralizing antibodies, whereas MHC class II-restricted antiviral CD4 T-cells regulate the overall adaptive immune response that promotes the resolution of acute hepatitis and viremia.

In CHB and CHC, antiviral T-cells are tolerized or “exhausted” by persistent activation upon exposure to cognate viral antigens and ongoing inflammation with the induction of immune inhibitory receptors (e.g. PD-1 and CTLA-4), regulatory T-cells and cytokines (6, 7). CHB is also associated with virus-specific and global B-cell deficits (8, 9). In this setting, innate immune components including dendritic cells, Kupffer cells, and natural killer cells (NK-cells) can contribute to persistent inflammation and hepatocellular injury (10).

The maladaptive immune responses in the inflamed liver microenvironment can compromise effective anti-tumor immunity. For example, anti-tumor T and/or NK-cell responses are tamped down by myeloid-derived suppressor cells, regulatory T-cells, inhibitory receptors, and non-immune cells such as liver sinusoidal endothelial cells and hepatic stellate cells, all leading to tumor evasion of the immune system (11). NK-cells infiltrating human HCC can be metabolically deranged with mitochondrial fragmentation, reduced cytotoxicity and reduced survival (12). Furthermore, a critical role for platelets in hepatic inflammation and injury as well as HCC development was shown in a mouse model of CHB in which anti-platelet therapy reduced hepatic accumulation of HBV-specific CD8 T-cells and hepatocellular injury, prevented HCC and improved survival (13, 14).

Cytidine deaminases:

The cytidine deaminase family of enzymes is induced by interferons and inflammatory cytokines in response to viral infection as an innate defense mechanism (15). Among their members, APOBEC3A and APOBEC3B were shown to degrade HBV covalently closed circular DNA (cccDNA) (16). Cytidine deaminases have been implicated in promoting HBV mutations that enable immune evasion and associated with somatic mutations in human cancers (15). However, APOBEC mutational signature was not significantly correlated with HCC in a recent genetic analysis of different cancer types (17).

Hepatic regeneration and mutagenesis in the setting of chronic viral hepatitis:

Hepatocytes are not highly proliferative under normal circumstances, dividing every 100 days or less (18). However, chronic immune-mediated hepatocellular injury can lead to increased hepatocyte turnover by 3–5 times (19) with accumulation of hepatocytes with DNA damage (20, 21). Along these lines, the risk for HCC development is greater among patients with immune-active CHB with flares of hepatitis and high viremia as compared to inactive carriers without active inflammatory activity or cirrhosis (22), reflecting increased hepatocyte turnover and oxidative stress with the generation of reactive oxygen species (ROS) that promote mutagenesis (23).

Role of ER stress with unfolded protein response:

Viral assembly and maturation for HBV and HCV occur in the endoplasmic reticulum (ER). The production of large amounts of viral proteins can cause ER stress and activate the unfolded protein response (UPR) pathway. For example, HBsAg made in abundant quantities can accumulate in the ER of infected hepatocytes, cause their “ground-glass” appearance on H&E stain and activate the UPR, leading to DNA damage and genomic instability (24, 25). Indeed, overexpression of large HBsAg in the hepatocytes resulted in HCC in an HBV transgenic mouse model, with ER retention of large HBsAg with “ground glass” hepatocytes and increased hepatic inflammation, regenerative hyperplasia, transcriptional deregulation and aneuploidy (26, 27).

Similarly, HCV can activate the UPR due to HCV RNA replication in the ER and anchoring of viral factors on the ER membrane, with increased ROS and TGFβ1 expression that can contribute to liver fibrosis and HCC (28, 29).

Role of cirrhosis in transformation:

Chronic inflammation, necrosis, and liver regenerative responses from CHB and CHC result in cirrhosis—likely a critical step in the development of HCC as it is present in over 90% of incident cases. Cirrhosis is the end stage of hepatic fibrogenesis in which activated hepatic stellate cells transdifferentiate into myofibroblasts and deposit extracellular matrix factors, particularly collagen (30). While cirrhosis is associated with hepatocyte senescence with shortened telomeres and lack of telomerase activity, HCC development is associated with telomerase reactivation (31) and differential methylation patterns of hepatocyte genomic DNA including at TERT and TP53 loci (32). The epigenetic changes may promote tumorigenesis by blocking the expression of tumor suppressors and enabling expression of genes that allow cells to escape senescence (32, 33). Liver stiffness caused by fibrosis may directly impact hepatocyte function and cause a permissive microenvironment for early transformation (34).

B. HBV-specific features associated with HCC

HBV is a hepatotropic, partially double-stranded DNA virus within the Hepadnaviridae family, with overlapping open reading frames that encode HBV envelope, nucleocapsid, polymerase and X proteins. HBV enters hepatocytes through its receptor NTCP, a hepatocyte-specific bile acid transporter, and reaches the nucleus where viral replication and transcription take place with the cccDNA as the template for all HBV gene products (35). In patients with CHB, high HBV DNA levels, sero-positivity for HBV e antigen (HBeAg), and basal core promoter mutations have been associated with increased risk for HCC (36, 37). We provide select highlights regarding HBV-specific mechanisms of HCC development.

Integration into host genome:

HBV DNA sequences can integrate into the hepatocyte genome, becoming a permanent part of the cell and its progenitors (38), with potential contribution to tumorigenesis by inducing genomic instability and disrupting cancer related genes. HBV integration is more frequently detected in tumor tissues (86.4%) as compared to adjacent liver tissues (30.7%), and at multiple loci involving cancer-associated genes such as CCNA2, TERT, MLL4, KMT2B and CCNE1 (39, 40). The role for HBV integration in HCC pathogenesis is becoming clearer with advancements in sequencing technology, which has enabled analysis of integration of HBV in the genome. When HCC tissues were sampled multiple times and sequenced deeply, integration events were found in cancer-associated genes both in early “truncal” events, and in later subclonal events (32), suggesting a mechanistic role for HBV integration both in early tumor formation and its progression. In addition, integration of HBV into the genome has been linked to gene fusion events such as with LINE1, which was found in 23% of HBV-associated HCCs and is suspected to drive HCC (41). HBV integration can occur early in HBV infection, and with minimal associated liver inflammation (37), suggesting that early therapeutic intervention might minimize accumulation of genetic damage (42).

Interestingly, AAV2 viral integrations have also been detected in some of the same driver genes as HBV, suggesting that this virus may also play a mechanistic role in HCC formation (43). However, the absolute contribution to HCC by AAV is unclear.

HBV gene products with oncogenic potential:

There is some evidence for direct oncogenic potential for HBV. First, HBV infection of primary cells upregulates host factors involved in the cell cycle including PPARA, RXRA, and CEBPB, (44). Second, several HBV proteins have been implicated as potential “viral oncoproteins” that can directly drive HCC, including HBsAg, HBcAg, and X protein (HBx). In addition to HBsAg mentioned above, HBcAg variants have been found in populations at high risk of HCC, with increased viral replication and hepatocyte turnover with upregulation of genes associated with fibrosis and cancer, including MYC, when used to infect a mouse model repopulated with human hepatocytes (45). Furthermore, HBV X protein (HBx) is known for its generalized transcriptional activating properties and may drive tumorigenesis directly, although with variable results depending on the genetic background (reviewed in (35)). A truncated form of HBx is also seen in 46% HBV-derived HCC and can directly transform cell lines (46, 47). Further studies are needed to clarify the role of HBV proteins in directly promoting HCC.

C. HCV-specific features that are associated with HCC

Viral and host factors:

HCV is a hepatotropic positive-strand RNA virus with high sequence heterogeneity due to error-prone viral RNA polymerase (48). As an RNA virus, HCV cannot integrate into the genome. However, CHC can promote hepatocarcinogenesis by activating inflammatory and toxin-responsive pathways. The HCV gene products including HCV core protein, Core-E1-E2, NS3 and NS5 proteins have all been shown to generate ROS or alter miRNA expression, leading to hepatic steatohepatitis, injury, and/or turnover (reviewed in (49)). Of interest, miRNA122 is the most abundant miRNA in the liver and is required for the HCV life cycle (diverse roles of miRNA122 reviewed in (50)). An important role for miRNA122 as a tumor suppressor of HCC was shown in a mouse knock-out model deleted for the miRNA122 locus (51). Finally, HCC development was reported in transgenic mice with hepatic expression of HCV NS3 protein (52), in which steatohepatitis precedes adenoma and HCC development consistent with a progression from injury to carcinogenesis.

Iron and CHC:

CHC is associated with increased iron deposition in the liver, which increases ROS and promote HCC (53). Iron absorption into the body is controlled by hepcidin production by hepatocytes, and HCV was shown to down-regulate hepcidin, leading to iron accumulation in the liver (54). Surprisingly, hepatic iron overload and C282Y mutation was associated with increased HCC risk in patients with cirrhosis due to alcohol but not CHC (55).

D. Co-infections and Co-factors in HCC pathogenesis related to HBV and HCV infection:

Viral Coinfections:

Viral co-infections can impact the clinical course of CHB and CHC. Chronic co-infection with HBV and hepatitis delta virus (HDV) is associated with a more rapid progression to cirrhosis and HCC, with HDV large delta antigen proposed to induce an inflammatory response with increased endoplasmic reticulum stress and ROS (Reviewed in (56)). Due to similar routes of transmission, there is a greater prevalence of HCV and HBV coinfection among HIV-infected patients, with accelerated liver disease progression and HCC in the setting of HIV-associated immune dysregulation (57).

Dietary carcinogens and toxins:

Aflatoxin is produced by a mold found on peanuts and other crops, with high levels detected in the diet in Africa and China. A cooperative effect between aflatoxin and HBV has been suggested by epidemiological data in which geographic regions with high rates of HCC also show high rates of aflatoxin and CHB (58) and directly shown in HBV transgenic mice (59). HCCs from these regions have frequent mutations in TP53, particularly at codon 249 (58, 60). Aflatoxin is also linked to a specific mutational signature within HCC (33). Notably, African migrants to Europe with HCC were found to have a strong Aflatoxin mutational signature in their primary HCC tumors that dissipated in HCC subclones that developed in Europe, suggesting that the aflatoxin-specific risk diminishes when the exposure is removed (80).

Additional risk factors associated with worse outcomes in patients with HCC include aristolochic acids (61), alcohol use and smoking (62).

E. Host genetic factors

Germline associations:

A family history of liver cancer is an important risk factor for the development of HCC. Multiple independent groups have found that in patients with CHB or CHC, having a first degree relative with HCC doubles the risk of incident HCC as compared to those without family history of HCC (63–65). This suggests that host genetic factors contribute to the risk for HCC, as well as potential shared environmental exposures among family members. Therefore, the American Association for the Study of Liver Diseases (AASLD) HBV guidelines recommend regular screening for HCC based on first-degree relative with HCC (66).

However, it is not clear whether germline variants in genes associated with hereditary cancer syndromes are linked to an increased risk of HCC in patients with CHB or CHC. An analysis of 363 HCCs in The Cancer Genome Atlas (TCGA) database, a collective at the National Cancer Institute, discovered one patient with CHB and a germline TERT promoter activating mutation (67). A few other germline pathogenic variants in cancer-associated genes were found in a separate analysis of the same TCGA cohort (68), but these did not reach statistically significant enrichment in HCC. Genome-wide association studies (GWAS) of HBV-associated HCC in different populations identified various loci associated with immunity and inflammation including a STAT4 variant with known association with autoimmunity, HLA class I/II, and CTLA-4, in addition to tumor suppressor genes and DNA repair genes, but with small odds ratios and inconsistent validation across studies (69–72).

Somatic alterations:

HCC tissues from patients with CHB and CHC have been examined for changes in gene expression and for mutations (Reviewed in (33)). These analyses have been especially useful in improving our understanding of the mechanisms of hepatocarcinogenesis and in predicting outcomes. The most commonly mutated gene region in HCC is the promoter of TERT, which encodes telomerase, and is found in up to 90% of HCC (73). TERT mutations have been found in pre-neoplastic lesions in cirrhotic and non-cirrhotic livers (31). The next most common mutations occur in TP53 (inactivating) and CTNNB1 (activating) (33). Interestingly, the prevalent types of mutations have patterns of co-occurrence and mutual exclusivity. For example, mutations in CTNNB1 are usually not linked to mutations in TP53 or AXIN1. In contrast, CTNNB1 and TERT mutations are often found together. TP53 mutation usually coincides with mutations in KEAP1, TSC2, or amplification in CCND1.

Somatic Associations with HBV:

HCC associated with HBV infection is enriched in certain genetic changes that are also seen in HCCs of other etiologies (reviewed in (73)). HBV tends to have the most aggressive of HCC molecular subclasses. Non-cirrhotic HBV integration events are thought to contribute to the etiology for this subclass, as are TP53 mutations and chromosomal instability (74). HCCs arising from HBV tend to have activation of the AKT pathway (75), and to be negative for CTNNB1 mutation (74).

There are also specific epigenetic changes associated with HCCs that arise from HBV. Whole genome bisulfite sequencing demonstrated differential methylation of a number of genes including genes involved in glycolysis, and genes implicated in cancers including MDM2, FGF4, FGF19, HSP90AA1 (76). The HBx protein has been implicated in altering the methylation of genes by directly activating DNA methyltransferases, leading to global changes in DNA methylation (77, 78).

A recent study performed an integrated proteomic and genomic analysis of HCCs associated with CHB (79). This revealed distinct subtypes of HCCs in CHB, with differential mortality that can be identified based on differential expression of proteins PYCR2 and ADH1A, suggesting that these proteins could be used as biomarkers in HCC.

Somatic Associations with HCV:

Mutations in TERT and CTNNB1 are enriched in HCV-associated HCC, though these mutations are present in HCCs of all etiologies (31, 80). In fact, no mutation signature emerging from genomic studies has been found to be specific to HCV (33).

Epigenetic studies related to HCV infection have examined the effect of viral infection on gene transcriptional regulation. HCV induces the expression of PPP2CA, which leads to epigenetic changes stemming from deregulation of histone modifications (81). HCV infection also leads to GaDD45B promoter hypermethylation and silencing, which inhibits DNA repair responses (82).

III. Clinically relevant considerations in viral hepatitis associated HCC.

Anti-viral treatment:

Therapeutic virus suppression can enhance virus-specific T-cell function and improve hepatic inflammation, liver enzyme levels and regulatory pathways (83, 84), albeit without necessarily providing protective immunity. CHC is now readily cured with direct-acting antiviral (DAA) therapy, but patients can be re-infected. HBV therapy reduces the risk of HCC by more than 50% (85). Current options for HBV therapy are well tolerated with a high barrier to resistance, although without providing a “cure” and may therefore require prolonged therapy (reviewed in (66)). Hence, there is strong rationale to improve antiviral therapy to achieve functional cure of CHB (86). Notably, differential effects between tenofovir and entecavir in reducing the risk of HCC have been reported, and further prospective examination is warrented (87–89).

Surveillance for HCC:

Patients at risk of HCC are recommended to enter into a surveillance program, typically with an ultrasound of the liver and/or alpha-fetoprotein level every six months (22, 66). The major determinant of risk is the presence of cirrhosis, with the risk of incident HCC in the setting of HBV or HCV-associated cirrhosis approaching levels as high as 3–8% per year. In geographic regions of Asia and Africa HCC occurs at a younger age and independently of cirrhosis, likely due to vertical transmission of HBV and environmental exposure to aflatoxin. Therefore, AASLD guidelines recommend screening all Asian and black men with HBV infection starting at age 40, and women at age 50 (66). Patients with HBV/HDV coinfection are also recommended to undergo regular surveillance for HCC regardless of whether or not they have cirrhosis.

HCC surveillance after achieving sustained virological response (SVR) in HCV:

When highly effective treatment for HCV became available with the arrival of DAAs, early studies reported higher rates of HCC than expected based on historical controls (reviewed in (90)), raising the hypothesis that anti-HCV inflammatory responses were also anti-tumorigenic and that its resolutionl facilitated HCC progression. However, subsequent studies showed reduced risk of all-cause mortality and incident HCC after HCV clearance, although patients with cirrhosis or advanced fibrosis may continue to have increased HCC risk for 10 years or more and need continued surveillance (91). Liver stiffness and harmful antiviral immune responses are reduced with HCV clearance, suggesting at least partial reversal of cirrhosis (91). As more patients who have achieved SVR are followed over time, it will be possible to define if/when HCC screening can stop, or whether any biomarkers can identify patients at persistently high risk for HCC.

Role of aspirin in chemoprevention:

Patients with CHB or CHC infection who take aspirin have been found to have a lower adjusted hazard ratio for the development of HCC (92–94). The mechanism of the cancer-protective effect may involve its disruption of pathogenic platelet aggregations in the liver vasculature, as mentioned earlier (13). The use of aspirin as a chemopreventative agent has not yet been studied prospectively and may warrant further study.

Immunotherapy with Checkpoint Inhibitors:

Blockade of immune regulatory pathways (e.g. PD-1, CTLA-4, Tim-3) has been shown to be safe and effective in reactivating immunity against a number of tumor types, with some impressive clinical results (95, 96). PD-1 blockade in patients with HCC has shown efficacy in 15–20%, with similar efficacy across HBV, HCV, and non-viral etiologies, and with a manageable adverse effects profile (97). However, blocking these regulatory pathways could also activate antiviral immune responses and induce hepatocellular injury (98), as evidenced by ALT elevations following PD-1 blockade in several virally suppressed HBeAg-negative CHB patients (including one who achieved HBsAg loss) (99) and a subset of CHC patients (100). A recent phase III study combining anti-PD-L1 with VEGF inhibitor showed significantly better overall and progression-free survival compared to sorafenib, which represents the most promising advance in HCC therapy in over a decade (101). Nevertheless, grade 3 or 4 side effects occurred in over 50% of patients in both treatment arms, indicating a need for careful consideration torwards the underlying hepatic functional reserve and reducing hepatic viral burden with antiviral therapy.

Immunotherapy with Engineered T Cells:

An exciting breakthrough in cancer immunotherapy has been the development of engineered autologous ‘CAR T cells’ that express chimeric antigen receptor (CAR) that target tumor antigens (102). This approach can be extended to chronic viral infections, given the similarities in T-cell ‘exhaustion’ and metabolic derangements (103). In fact, limited adoptive transfer of HBsAg-specific CAR T-cells into a liver transplant recipient with metastatic HBV-associated HCC demonstrated the approach could be done safely, although clinical efficacy was not demonstrated (104). Further studies are needed before considering generalized use in HBV-infected patients given the potential for immune attack of hepatocytes.

IV. Closing statement and unanswered questions:

It is clear that viral and host factors contribute to the development of HCC, but a number of questions remain.

For one, it is generally accepted that multiple genomic mutations are needed to result in HCC. However, it is not clear at which point the genetic changes irreversibly lead to cancer, with no chance of return to quiescence. Furthermore, although cirrhosis is a readily identifiable major risk factor, approaches are needed to prevent HCC. In addition to anti-viral drug treatment, future treatments could target early cancer driver genes, reduce smoldering inflammatory responses, perhaps by aspirin, or reverse fibrosis.

An approach to identify targets genes in early HCC that is being explored using animal models of liver cancer development is to express HCC driver genes such as MYC or MET or HBx, and then to identify factors that are required for cancer progression in this setting. Such “synthetic lethal” factors could be targeted with drugs to prevent cancer progression (105).

Another area for future study is to perform risk prediction models in HCC. Larger GWAS and other types of genomics or proteomics studies are needed to determine whether genetic tests or biomarkers can identify patients at high risk for HCC. Such patients could benefit from more aggressive screening programs.