In this chapter, we present the founding framework of a novel theory termed as Cancer Evolution-Development (Cancer Evo-Dev), based on the current understanding of hepatitis B virus (HBV) induced hepatocarcinogenesis. The interactions of genetic predispositions and HBV infection is responsible for the maintenance of chronic non-resolving inflammation. Under the inflammatory microenvironment, pro-inflammatory factors trans-activate the expression of cytidine deaminases and suppress the expression of uracil DNA glycosylase. The imbalance between the mutagenic forces and mutation-correcting forces facilitates the generations of somatic mutations, viral mutations, and viral integrations into the host genomes. The majority of cells with genomic mutations and mutated viruses are eliminated in survival competition. Only a small percentage of the mutated cells adapted to the hostile environment can survive, retro-differentiate, and function as cancer-initiating cells, representing a process of “mutation-selection-adaptation”. Cancer Evo-Dev lays the theoretical foundation for understanding the mechanisms by which chronic infection of HBV promotes hepatocarcinogenesis. This theory also plays an important role in specific prophylaxis, prediction, early diagnosis, and targeted treatment of cancers.
- hepatitis B virus
Chronic infection of hepatitis B virus (HBV) is a major cause of hepatocellular carcinoma (HCC). Although the strong etiologic relationship between HBV infection and HCC has been supported by substantial evidence, the underlying mechanism is still elusive. There are more than nine thousand studies investigating HBV-induced hepatocarcinogenesis, which yield 143 genes function in 137 pathways . Most of the studies are one-sided investigations, only a few trying to provide a theoretical hypothesis and to promote the system-level understanding of HBV-induced HCC (HBV-HCC). In past decades, continuous attempts have been made to investigate carcinogenesis from an evolutionary point of view. In 1976, Dr. Nowell first proposed that most neoplasms originate from a single cell. Malignant cells are more genetically unstable than normal cells  In 2006, it was pointed out that cancer clone genetic diversification and sub-clonal selection occurs within the microenvironment, which is similar to the process of Darwinian natural selection . This viewpoint was put forward mainly based on morphological evidence and only a limited number of gene mutations and related signaling pathways were discussed. The widespread application of new generation sequencing promotes the investigation of genetic diversification and clonal selection within tissue ecosystems. It was found that the number of mutations in cancer range from 10 to hundreds of thousands. The majority of mutations are “passengers” and a small part are “drivers” . Cancer cells acquire a variety of critical phenotypes via driver mutations, which compound to enhance the capabilities of self-renewal, migration, and invasion. The mutational spectra in cancers can reflect the characteristics of the mutational process, including the error-prone repair and genotoxic exposure . Interestingly, the cytidine deaminase induced mutation is dominant in most cancers . Cytidine deaminase is upregulated during inflammation and defense against many viruses, including HBV. Epidemiological and experimental evidence identified the co-evolution of HBV and cancer cells during chronic inflammation. In turn, the mutant cells and viruses also affect the inflammatory microenvironment . Thus, there is a similarity between the process of carcinogenesis and Darwinian evolution. Furthermore, the investigation of cancer evolution can draw upon the understanding of developmental processes. Development is referred to the process that a fertilized egg develops into an individual. In humans, the fertilized diploid cell differentiates into various functional and/or structural cells to form different organs and tissues within 40 weeks. This process resembles the process of long-term organic evolution morphologically, from single cell creatures to multicellular creatures, and from aquatic creatures to terrestrial mammals. Some evolutionarily conserved molecules, like Hedgehog, HOX, and Myc are essential for the developmental process, suggesting evolution and development have similar inherent mechanisms [8, 9, 10, 11]. The integration of evolution and developmental biology was termed
2. Framework of
The synergetic effects of genetic predisposition and environmental factors contribute to the imbalance of the immune system, resulting in the activation and maintenance of non-resolving inflammation, that functions as the microenvironment for the
3. Chronic inflammation is indispensable for HBV-HCC evolution
As a defense mechanism responding to exogenous infection and injury, acute inflammation is beneficial to humans. However, chronic inflammation, also termed non-resolving inflammation is essential for carcinogenesis. The weak immunity, HBV mutation, and HBV genotype contribute to the chronicity of inflammation. During the development of HBV-induced HCC evolution, non-resolving inflammation is evident. By relieving hepatic inflammation, antiviral therapy can significantly lower the risks of HCC occurrence and postoperative recurrence [15, 16]. Interestingly, the risk of HCC is still significantly higher in the complete responder group of oral-administered antiviral therapy, compared with the subjects with inactive chronic hepatitis B (CHB) . The active inflammation on chronic infection background also indicated postoperative recurrence [18, 19]. The close association between chronic inflammation and the risk of HCC can be explained from the perspective of
3.1 Chronicity of HBV infection and hepatic inflammation
The oncogenic capability of HBV is closely related to its capacity to induce and maintain chronic inflammation. The chronicity of HBV infection is dependent on 3 aspects: infection occasion, HBV genotypes, and genetic predisposition of the key immune molecules. HBV infection in early childhood is generally believed to be one of the major causes of chronic HBV infection in adulthood. The perinatal infection occurred in 8.7% and 84.2% of infants born to hepatitis B e-antigen (HBeAg)-positive mothers who did and did not receive immunoprophylaxis, respectively. The infection rates were 0.4% and 6.7% for infants born to HBeAg-negative mothers and HBeAg-positive mothers, respectively. Furthermore, the chronicity of HBV infection acquired perinatally was 28.2% and 64.5% for infants born to HBeAg-negative mothers and HBeAg-positive mothers, respectively . This vulnerability of infants may due to the immaturity of the immune system. Although perinatal HBV infection is an important cause of chronic HBV infection, the chronic transformation of acute hepatitis B is the predominant cause of chronic HBV infection in adults. In China, 8.5% of patients with acute hepatitis B develop into chronic HBV infection 6 months after acute infection . The HBV genotype and genetic predisposition of immune molecules contribute to this transformation.
According to sequence divergence of 8% in the whole viral genome, HBV can be classified into eight genotypes (A to H) . Variant genotypes are distributed unevenly around the world, and the predominant one in mainland China is genotype C (68.3%), followed by genotype B (25.5%) . Under selection pressure from the inflammatory microenvironment, the fates of different HBV genotypes are distinct. Genotype B HBV is prone to causing acute infection, whereas genotype C HBV is associated with chronic infection and contributes independently to the development of HCC [22, 25, 26].
The genetic predisposition of immune molecules is the third major cause of chronic HBV infection. The single nucleotide polymorphisms (SNPs) in the loci encoding human leukocyte antigen class II (HLA-II) are significantly associated with vaccine response as well as the risk of CHB, HBV-induced liver cirrhosis, and HBV-HCC [27, 28, 29, 30, 31, 32]. Interestingly, the allele frequencies of SNPs affecting the expression of HLA-DP and HLA-DQ are variant in different human races. The polymorphic genotypes that are more frequent in the Han Chinese than in European populations are significantly associated with the increased risk of chronicity of HBV infection as well as the immune selection of HBV mutations related to end-stage liver diseases . These data suggest that the Han Chinese are inherently more apt to progress into chronic infection once exposed to HBV infection than Europeans. This might be partly responsible for the fact that chronic HBV infection, HBV-induced liver cirrhosis, and HBV-HCC are more frequent in Chinese than in European populations. The genetic polymorphisms of HLA-II may facilitate the progression of CHB into HCC through predisposing immune imbalance and maintain HBV infection. Due to the chronicity of inflammation, the mutagenic force that serves as antiviral immunity is prone to injury the human genome, thus induce
3.2 HBV promote the generation of inflammatory mutations
The APOBECs are powerful endogenous mutagenic factors that can catalyze irreversible cytidine and deoxycytidine deamination to convert bases from cytosine to uracil, creating a cytosine-to-uracil mismatch in minus-strand and reverse-transcript G-to-A (guanosine-to-adenosine) transitions in plus-stranded DNA. APOBEC3s play important roles in the innate immune system . Mutagenesis mediated by APOBEC3s can increase the viral mutation load to a level that exceeds the threshold for viral viability. Accordingly, APOBEC3s can similarly increase the number of somatic mutations to a threshold that exceeds the host’s repair ability and starts the
3.3 HBV affects the selection pressure of the inflammatory microenvironment
In an inflammatory microenvironment, continuous necrosis and proliferation can help to accumulate somatic mutations, and tumor-initiating cell clones with strong viability are selected. HBV replication directly reflects the selective stress and influences the evolution of HCC. It has been revealed by various studies that HBV DNA load increases the risk of HCC in CHB patients [25, 36]. A high level of HBV DNA load either in serum or liver tissue predicts poor postoperative prognosis in HCC . Meanwhile, HBV in turn affecting the selective pressure of the inflammatory microenvironment. The innate immune and adaptive immune against HBV are both participate in the selection of malignant cells. During the chronic infection of HBV, APOBEC3B is stimulated and reduces the occupancy of H3K27me3 on the promoter of CC-chemokine ligand 2 (CCL2). By this mechanism, APOBEC3B upregulates the CCL2 to enhance the recruitment of tumor-associated macrophages (TAMs) and myeloid-derived suppressor cells (MDSCs). TAMs and MDSCs suppress the function of CD8+T cells and are associated with a poor prognosis of HCC . HBV can transmit into natural killer (NK) cells through exosomes, thereby inducing the disfunction of NK cells, which promotes the HCC evolution [39, 40]. HBV also induced the exhaustion of HBV-specific CD8+ T cells through impairing the mitochondrial functions, including electron transport, membrane transport, and the transcription of mitochondrial DNA . The glucose metabolism of T cells is reprogramed by HBV, which leads to increased lactate production and decreased migration of T cells . During chronic infection, HBV promotes the recruitment of regulatory T cells (Tregs) through activating the growth factor-beta (TGF-β)/miR-34a/ CC-motif chemokine ligand 22 (CCL22) axis . The increased Tregs suppress HCC antigen-specific immune responses and HBV antigen-specific immune response at the same time . Thus, the HBV that survives the survival competition can in turn affect the inflammatory microenvironment.
4. Roles of HBV mutation during the process of HCC evolution
During HBV-induced hepatocarcinogenesis, viruses also experience the process of evolution. Viral evolution serves as a valuable clue to investigate the mechanism underlying the HBV-HCC . The generation and accumulation of HBV mutations abide by the Darwinian model: mutation-selection–adaptation. In the inflammatory microenvironment, most HBV mutants are eliminated by the antiviral immune response. Only a tiny fraction of mutant viruses that facilitate the regeneration of hepatocytes can survive and gradually develop into the HCC-promoting clones.
4.1 The generation of HBV mutation
Two major mechanisms are responsible for the generation of HBV mutation. The first pattern is the replicative errors. During viral replication, the partially double-stranded HBV DNA is generated from an intermediate RNA through the reverse transcription activity of the viral polymerase. Due to lack of proofreading capacity, the HBV genome has a higher mutation rate than other DNA viruses, which is in the range of 1.5 × 10−5 to 5 × 10−5 nucleotide substitutions per site per year, which can increase after HBeAg seroconversion . The second viral mutation pattern is induced by host cytidine deaminases . The APOBEC family has a dual effect on HBV: reduction of HBV and induction of HBV mutations . The expression levels of APOBEC3s are positively correlated with the quasispecies complexity of HBV . The genetic polymorphisms predisposing the IL-6 induced APOBEC3B-UNG imbalance significantly promote the generation of HCC related HBV mutations . Although many HBV genome fragments, including the Enhancer II (EnhII) /basal core promoter (BCP)/ precore region and the S region, are generally sensitive to editing by members of APOBEC3 [49, 50, 51, 52, 53], the sequence encoding HBV X protein (HBx) is more vulnerable. APOBEC3 prefers the HBx region as its editing target and generates carboxylic acid–terminal truncated HBx (Ct-HBx). Although most HBV mutations are random, the directional evolution of HBV occurs under the selective pressures of chronic inflammation. In the immune tolerance phase of chronic infection, the immune pressure is weak, and most of the individual viruses are wild-type. Immune pressure increases with the progression of chronic inflammation, which facilitates the gradual occurrence of viral mutations, especially in HBeAg-negative individuals [54, 55]. HCC-related HBV mutations are selected by the immune microenvironment before the occurrence of HCC and can be used as predictive markers. Single-nucleotide polymorphisms of the inflammatory signaling pathway genes, including STAT3, NF-κB, HLA-DP, and HLA-DQ have been demonstrated to maintain the chronic infection and to facilitate the selection of HCC-related HBV mutations that contribute to the risk of liver cancer [32, 56, 57]. However, those viral mutants that affect the pre-cancer hepatocytes are less infectious to normal liver cells, which leads to a process of “dead-end” evolution.
4.2 The “dead-end” evolution of HBV
Hepatitis B virus belongs to the Hepadnaviridae family and is evolutionarily conservative in the long-term evolution of species . However, the evolution of the HBV genome is evident in infected individuals during chronic infection. Previous research by our group established the wild-type HBV sequences of HBV subgenotypes B2 and C2, based on the whole HBV genome sequenced using 1000 asymptomatic carriers of the HBV surface antigen from community-based epidemiologic surveys. Based on the wild-type HBV sequences, HCC-related mutations and their development patterns were subsequently identified. We also observed that HBV mutations posing a significant HCC risk are located mainly within the BCP and preS regions [59, 60, 61]. During the HBV-induced carcinogenic “trilogy” (chronic hepatitis, liver cirrhosis, HCC), the species and frequencies of those mutations often accumulate consecutively and can be used to predict the occurrence and development of liver cirrhosis and HCC [15, 32, 56, 62]. Retrospective and prospective cohort studies have both identified a combination of HBV mutations (C1653T, A1762T/G1764A, and T1753V) that have significant predictive value [32, 63, 64]. Among them, the A1762T/G1764A mutation usually appears in the early stage; other mutations, including T1753V, C1653T, preS deletion, are evident only in a late stage of the evolution . Reaction to chronic HBV infection (characterized by the immune response–induced hepatocyte injury and release of transaminase) is usually accompanied by HBeAg seroconversion and an increase in HBV mutations, indicating the selective effect of immune cells on viral mutants. The deficiency of CD8+ T cell epitopes is one of the main features of HBV mutations. The mutant virus with a low density of CD8+ T cells epitopes can evade immune eradication [66, 67]. The proportion of mutant preS/S region is higher in patients with occult HBV infection than in CHB patients [67, 68]. Therefore, CD+8 T cell is essential for the immune selection of HCC-related HBV mutants.
Hepatitis B virus acquired during infancy or early childhood, or at the early infection stage in adults, is usually the wild type [15, 32, 56]. During the chronic inflammation process, especially after an HBeAg shift from HBeAg-positive to HBeAg-negative, mutant HBV subgroups gradually increase. Although the HCC-related HBV mutants are present in fetal cord blood, neonatal infection is usually caused by wild-type HBV rather than by mutant subgroups. At 1–15 years in HBV-infected children, the frequencies of HCC-related mutations increase with increasing age. However, compared with their mothers, who have been exposed to chronic infection for at least about 25 years, the children have fewer HCC-related HBV mutations . The foregoing results are based on analyses of serum HBV. In individuals with chronic HBV infection, most all HBV is synthesized in hepatocytes and released into the circulation at a pace of up to 1011 viral particles daily . The immune microenvironment of circulation, tumor tissue, and tumor-adjacent liver tissue are all necessary for the HBV evolution . Interestingly, HBV evolves more advanced in the sera than in the tumors of HCC patients. The evolutionary similarity between the sera-derived HBV strains and adjacent tissue-derived ones is significantly stronger than that between sera-derived HBV strains and tumor-derived ones . Although tumor-adjacent tissues are pathologically categorized as “normal,” they are typical precancerous lesions and have already entered the middle stage of the cancer evolutionary process. The HCCs that relapse more than 2 years after resection are considered to be recurrent HCC and not a result of the initial HCC cell diffusion into remnant liver tissue . The species and frequencies of certain HBV mutations in adjacent tissues are distinct in the different populations. Together with immune markers and expression levels of inflammatory genes, they can therefore be used to predict prognosis in HCC patients receiving curative surgery. For example, HBV mutations in the EnhII/BCP/PreC region, such as A1762T/G1764A, can serve as predictive markers for survival and recurrence , indicating that HBV evolution in adjacent tissues continues until the patient dies. Antiviral therapy can block HBV evolution in adjacent tissues by easing inflammation and notably prolongs survival in HCC patients .
Taken together, the Hepadnaviridae family members are highly conservative across species . Wild-type HBV has the advantage of infecting hepatocytes, facilitating viral spread from one individual to another, and contributing to the maintenance of its viral species. The HCC-related mutants can cause malignant transformation but have lost the advantage of person-to-person infection. Those mutants are therefore usually eliminated at the death of the carriers, which is termed “dead-end” evolution.
4.3 High-risk HBV mutations promote the
During hepatocarcinogenesis, high-risk HBV mutations are selected by the immune microenvironment. Because of overlapping open reading frames, HBV mutations altering the genes necessary for viral replication are unlikely transferred into their progeny viruses. Natural selection ensures only the fittest survive to pass their genes on to the next generation. Thus, the random natural mutations are therefore constrained to special regions of the HBV genome, especially in the fragment of HBx gene and large envelope protein gene fragment (preS1/preS2/S). These HBV mutations that survive the selective pressure can promote the evolution of HCC, which is supported by many pieces of evidence from epidemiology studies and mechanism studies.
Previous longitudinal studies, especially cohort studies, support that combo HBV mutations including A1762T/G1764A, C1653T, and T1753V in HBx gene in sera can predict the occurrence of HCC [64, 70]. The mutations in the HBV preS fragment, including the preS deletion, accumulate during the process of inflammation-HCC transformation, which is significantly associated with increased risk of HCC [62, 71, 72]. Epidemiological evidence identified the interaction effect between HBV mutations and genetic polymorphisms of immune molecules. For the population with the infection of genotype B HBV, the SNPs of HLA-DP, including rs3077 (T allele), rs2281388 (T allele), rs3135021 (G allele), and rs9277535 (G allele) can promote the HBV persistence and are associated with a higher prevalence of HBV mutation increasing HCC risk. Moreover, the effects of HBV mutations on HCC risk are selectively significant in subjects with these HLA-DP SNPs that promote HBV persistence . For the population with the infection of genotype C HBV, the HLA-DQ SNP, rs9275319 (GG genotype), is significantly associated with an increased prevalence of preS1 start codon mutation, an HCC-risk mutation . The SNPs of STAT3 SNPs appear to promote HCC evolution in the host with HBV mutations . The interaction effect of STAT3 rs1053004 with T1674C/G and the interaction effect of STAT3 rs4796793 with preS2 start codon mutation are both significantly associated with an increased HCC risk. The T allele of rs223406 impairs the promoter activity of NFKBIA, a key molecule of the NF-κB signaling pathway. The interaction of rs223406 T allele with A1762T/G1764A is significantly associated with an increased risk of HCC . The genetic polymorphisms predisposing the imbalance of APOBEC3B and UNG increase the risk of HCC through through facilitate the generation of APOBEC3B-signature HBV mutations. Furthermore, the positive rate of APOBEC-signature HBV mutations consecutively increased from asymptomatic HBsAg carrier (ASC) to HCC in HBV-infected subjects . This line of evidence highlights the important role of HBV mutation in the process of HCC evolution.
Experimental evidence also confirms that HBV mutation can endow the hepatocytes with a survival advantage. The HBx with A1762T/G1764A–based combo mutations can upregulate the expression of S-phase kinase-associated protein 2 (SKP2) by activating E2F1, a transcription factor, downregulate cell cycle inhibitors, and facilitate the ubiquitin-mediated proteasomal degradation of p21, thereby enhancing the proliferation of HCC cells [73, 74]. Moreover, HBx with A1762T/G1764A–based combo mutations also enhance the cell migration through activating the Wnt/β-catenin signaling pathway . Ct-HBx mutation can promote cell metastasis and invasiveness by activating the C-Jun/matrix metalloproteinase protein 10 signaling pathway [15, 76]. The HBx gene with K130M/V131I mutations enhances HCC evolution by activating the arachidonic acid metabolism and the hypoxia-inducible factor-1a [77, 78]. Besides the mutated HBx gene, the mutated preS1, preS2, and S regions also notably facilitate carcinogenesis [18, 61]. The preS2 region with F141L can significantly downregulate the expression of the p53 pathway and upregulate the expression of cyclin-dependent kinase 4 and cyclin A, thereby promoting proliferation and colony-forming rates . The accumulation of mutant envelop protein in the endoplasmic reticulum (ER) leads to the activation of ER stress signaling pathway . ER stress promotes HCC evolution through generating reactive oxygen species (ROS), inducing oxidative DNA damage, and ultimately increasing genomic instability [81, 82]. Although HBV mutation plays important role in hepatocarcinogenesis, somatic mutation of the human genome is the direct cause of cell evolution.
5. Roles of somatic mutation during the process of HCC evolution
The spontaneous rate of somatic mutations is not high enough to trigger the evolution process. HBV participates in the alteration of the host genome, both directly and indirectly. First, HBV can cause somatic mutations by directly integrating into the human genome. Second, mutant HBV contributes to the maintenance of non-resolving inflammation, that induces long-term up-regulation of APOBECs . Somatic mutations can be classified according to their effects on
5.1 Somatic mutations alter “stem-ness” signaling pathways
Based on the investigations of whole-exome sequencing, it is found that the somatic mutation in HCC evolution mainly altering six cancer related pathways: signaling pathway related with telomere maintenance, Wnt/b-catenin pathway, P53 and cell cycle pathway, oxidative stress pathway, epigenome modifiers, RAS/RAF/mitogen-activated protein kinase pathway, and PI3K/AKT/mTOR pathways . Among them, the somatic mutation related to telomeres pathway is most frequent. Telomerase is activated in more than 90% HCC patients. Somatic mutation within the promoter of telomerase reverse transcriptase (
5.2 Somatic mutations affect HCC evolution through regulating EMT
APOBECs can promote gene demethylation and remove epigenetic memory to stabilize the pluripotent state in embryonic stem cells through deaminating 5-methylcytosine (5mC) or 5-hydroxymethylcytosine (5hmC) [87, 88]. EMT is a landmark event of
5.3 Somatic mutations reprogram energy metabolism
To support the rapid growth of malignant cells, tumor tissues prefer to use glycolysis for energy production, even in the presence of oxygen. Glucose is more easily to be metabolized to lactate in tumor tissues than in normal tissues. This pattern of energy metabolism was identified in 1920 and was termed as Warburg effect . Warburg effect in TAMs promotes vascular network formation, augments extravasation of tumor cells out of blood vessels, and induces higher levels of EMT at inflammatory foci within the tumor . In the microenvironment with both hypoxia and hypoglycemia, stem cell-, angiogenic-, and EMT-biomarkers, as well as glycoprotein-P content and invasiveness of cancer cells are enhanced . Thus, we believe that the Warburg effect promotes the evolutionary process of cancer under both hypoxia and hypoglycemia conditions. The Warburg effect can provide essential energy for cell survival in a hostile microenvironment, furthermore, glycolysis generates the raw material for DNA synthesis of progeny cells. HBV infection and somatic mutation are both the possible origin of Warburg phenotype. In HBV-HCC, the major pattern of single nucleotide variants in mitochondrial DNA (mtDNA) is C > T, that is the character of APOBEC induced mutation. This kind of mutation mainly occurs in the D-loop region of mtDNA and promotes the proliferation, invasion, and metastasis of HCC cells . Pyruvate kinase M2 (PKM2), an alternatively spliced variant of the pyruvate kinase gene that is preferentially expressed during embryonic development and in cancer cells, alters the final rate-limiting step of glycolysis, resulting in the cancer-specific Warburg effect . Besides the Warburg effect, HCC cells also enhance other patterns of energy metabolism during evolution. For example, the inactivating mutation of ribosomal S6 kinase 2 (RSK2) can support cholesterol metabolism in HCC .
5.4 HBV integration
HBV integration is a kind of somatic mutation that is specific to the HBV-induced
As mentioned above, hepatocarcinogenesis involves the co-evolution of HBV and transformed cells. The interaction between somatic mutation and HBV mutation occurs during this process. The deletion, duplication, and translocation are observed near the insertion site of integrated HBV fragments . The frequency of HBV mutation is positive associated with the level of HBV integration. The prevalence of HBx mutation is significantly higher in patients with HBV integration in
Based on studies of HBV-induced hepatocarcinogenesis (a typical evolutionary process), we put forward the theory of
This work was supported by grant 2015CB554006 from the National Key Basic Research Program of China, grants 91529305, 81520108021, and 81673250, and 81521091 from the National Natural Science Foundation of China.
Conflict of interest
The authors declare no conflict of interest.