Antiviral treatment of patients with chronic hepatitis B (CHB) in the grey zone of treatment comands risk management in order to optimize the health outcome. In this sense, the identification of HBV mutants related with an increased risk of hepatocellular carcinoma (HCC) could be useful to identify subpopulations with potential indication of antiviral treatment.

To analyze the prevalence/persistence of hepatitis B virus (HBV) preS and basal core promoter (BCP)/precore/core variants associated to HCC development in CHB patients in the grey zone.

Work was designed as a longitudinal retrospective study, including 106 plasma samples from 31 patients with CHB in the grey zone of treatment: Hepatitis B e antigen negative, HBV-DNA levels between 12-20000 IU/mL, normal or discordant transaminase levels during follow up and mild/moderate necro-inflammatory activity in liver biopsy or Fibroscan (up to 9.5 kPa). Serum HBV-DNA was tested using the Abbott Real Time HBV Assay and the BCP/precore/core and the hepatitis B surface antigen (HBsAg) coding regions were analyzed in positive samples by PCR/bulk-sequencing to identify the HCC-related HBV mutants.

High-risk HCC related mutants were detected in 24 (77%) patients: 19 (61%) in the BCP/precore/core, and 7 (23%) in the HBsAg coding region (2 preS1 and 5 preS2 deletions). The prevalence of preS deletions was genotype-dependent: 3/5 (60%) patients with preS2 deletions and 1/2 with preS1 deletions were infected with the HBV-E genotype. Since HBV-E was the most prevalent in sub-Saharan patients, a correlation between preS deletions and ethnicity was also found: 6/8 (75%) sub-Saharan vs 1/19 (5%) Caucasian patients had preS deletions (P = 0.00016). Remarkably, this correlation was maintained in those patients infected with HBV-A, a minor genotype in sub-Saharan patients: 2/2 patients infected with HBV-A from West Africa vs 0/6 of Caucasian origin had preS deletions. The HCC related variants were the major strains and persisted over time (up to 48 mo). Patients with preS deletions had a significant higher prevalence of F2 fibrosis stage than the negatives (57% vs 10%, P = 0.0078).

HBV genetic analysis of selected populations, like sub-Saharans infected with HBV-E/A genotypes, will allow identification of subpopulations with risk of HCC development due to accumulation of high-risk HBV variants, thus commanding their increased clinical surveillance.

In most cases, patients with chronic hepatitis B (CHB) treated with nucleos(t)ide analogues (NAs) such as tenofovir or entecavir achieve a maintained virological suppression[1-3]. Although the inhibition of viral replication has an evident beneficial effect on the progression of the liver disease, the antiviral treatment does not achieve eradication of the hepatitis B virus (HBV) infection due to the persistence of intracellular covalently closed circular DNA in the nucleus of infected hepatocytes[4]. Thus, therapy is generally recommended for prolonged periods of time and the discontinuation of NAs is controversial because the early interruption of treatment is usually associated with a reactivation of viral replication that may induce severe flares if not controlled by the host immunity, requiring the re-start of the antiviral treatment[5]. According to EASL guidelines treatment may be discontinued in hepatitis B e antigen (HBeAg) negative patients only when hepatitis B surface antigen (HBsAg) loss is achieved or after 3 years of virological suppression[1]. Conversely, AASLD guidelines still recommends lifelong antiviral treatment in patients with HBeAg negative CHB[2].

The indications for CHB treatment are mainly based on the combination of three criteria: serum HBV-DNA levels, alanine aminotransferase (ALT) values and severity of liver disease. Patients should be considered for treatment when they have HBV-DNA levels above 2000 IU/mL, persistent or intermittent serum ALT abnormal levels and severity of the liver disease assessed by liver biopsy (or non-invasive markers validated in HBV-infected patients) showing moderate to severe active necroinflammation and/or at least moderate fibrosis using a standardised scoring system[1,2]. Conversely, there are HBeAg-negative patients with maintained minimal or mild fibrosis, for whom the treatment is not clearly indicated: A) patients with maintained normal ALT levels and low or undetectable serum HBV-DNA: B) patients with marginally elevated ALT and/or HBV-DNA titters ranging 2000-20000 IU/mL; and C) persistently normal ALT and HBV-DNA titters higher than 2000 IU/mL[6,7].

The indication of treatment in these patients, in the so named “grey zone” of treatment, requires to balance the risks and benefits for health outcomes. Although low levels of HBV replication and the absence of significant liver damage are good prognostic factors in the natural history of CHB, it cannot be ruled out that some of these patients may be infected with viral strains with mutations associated with high-risk of hepatocellular carcinoma (HCC) development, which are an additional risk of progression for liver disease[8,9]. There are several studies that associate certain mutations in the HBV genome with an increased risk of HCC development, but the results are controversial because the HBV genotypes or even different HBeAg status have different mutation patterns. One of the mechanisms by which HBV can promote direct carcinogenesis is the ability of wild-type and mutated/truncated viral proteins like hepatitis B x-protein, hepatitis B core-protein (HBc) and preS region (preS) to affect cell functions, activate oncogenic pathways and sensitize liver cells to mutagens[10]. The most common HBV specific mutations are at the preS2 start codon, preS deletions, and point mutations in the basal core promoter (BCP) and in the precore/core coding region[11-13].

For these reasons, the aim of this work is to analyse the prevalence and persistence over time of HBV mutants that predispose to the development of HCC in patients with CHB but without clear indication of treatment, and to analyze their role as a tool in the selection of suitable subpopulations for antiviral treatment.

In this work we performed a longitudinal retrospective study to analyze the prevalence and persistence over time of HBV mutants related with risk of HCC development in HBeAg-negative/anti-HBe-positive patients within the grey zone of treatment, in order to identify subpopulations susceptible to be treated.

Our results confirm the presence of high-risk HCC mutations in BCP/precore or preS regions, in 19 (61%) patients, using a statistical analysis of ITT, with a higher prevalence in the BCP/precore that in the preS region. The most frequently detected high-risk mutations in the BCP/precore region were the G1896A and G1899A. Indeed, finding these mutations in this group of patients is common and frequently related with the Mediterranean variants in patients with HBeAg-negative serological profile[14]. Thus, the G1896A mutation introduces a stop codon at position 28 of the HBc protein that prevents the synthesis of HBeAg[15,16]. The simultaneous presence of A1762T and G1764A mutations that are related to a decrease in HBeAg levels[17-19] and an increase of the HBV-DNA titters, was also confirmed in 10% of the patients[20]. The prevalence of these variants in our cohort of patients was genotype-dependent: BCP variants (C1653T, T1753V, A1762T/G1764A and C1773T) were more frequent in HBV-A genotype, as previously described, except for the C1653T mutation, which was more frequent in HBV-D. Precore/core variant G1899A was more frequent in HBV-E genotype in our cohort, tallying with the published data[21], but also found in a high proportion of patients infected with HBV-A genotype. The prevalence of the G1896A mutation in our patients is consistent with other studies carried out in Spanish population, in which this variant was found with similar frequency in genotypes A and D[22], contrasting with other works conducted in different geographical regions, in which no correlation with the HBV-A genotype was found[23,24]. These discrepancies probably reflect that the HBeAg-negative serological profile may present a different genetic basis in at least some HBV genotypes, with a greater representativeness than in other populations of the G1896A mutation in Mediterranean population, in which this serological pattern appears be more related to the simultaneous presence of A1762T/G1764A double mutation[22].

Although the BCP and precore/core variants has been usually associated to HCC development in meta-analysis studies, a recent work performed in HBeAg-negative Spanish Caucasian patients in the grey zone without antiviral treatment reflects that, in the long term, their clinical evolution is favorable, without progression of liver disease or HCC development in the vast majority of them[25]. Considering that the main objective of antiviral treatment is to prevent the fibrosis progression, these data support the idea that treatment does not provide substantial benefits in this group of patients. In this case, the low progression of HBV disease can be related with the low viral replication levels.

HBV-DNA titters have been classically considered as a prognostic marker of HCC development risk[26]. The maintenance of HBV-DNA levels undetectable or below 2,000 IU/mL can compensate the risk of harbouring HCC-related variants. However, it cannot be ruled that the accumulation of mutations in both BCP/precore/core region and other viral regions like HBsAg, maintained over time, in combination with persistence of viral replication, even at low or moderate levels are risk factors for HCC development[9]. For this reason, we have focused our work on the analysis of the prevalence and persistence over time of the BCP/precore/core and HBsAg regions mutants, present as main strains of HBV detectable by means of bulk sequencing techniques.

The prevalence of preS mutations, which have been correlated with HCC in previous meta-analysis, showed a strong correlation in our cohort with the geographical origin and, in consequence, with the HBV genotypes: 75% of our sub-Saharan patients, the majority infected with HBV-E genotype, had deletions in the preS1 or preS2 regions. In addition, the only 2 sub-Saharan patients infected with HBV-A had also deletions in preS2, while this change was not found in any of the remaining patients infected with HBV-A, all of them with Caucasian origin. HBV-E genotype has not been extensively studied until now. Previous studies performed in Germany confirm our results showing that patients in the grey zone of treatment infected with the HBV-E genotype have a higher prevalence of preS1 and preS2 deletions in comparison with HBV-A and D genotypes[27]. However, the absolute prevalence rate of these mutations in HBeAg negative patients differs between different studies, ranging from 16% in the German series to 75% in our cohort. The phylogenetic analysis of HBV-E strains from different regions of Africa, indicates that the spread of HBV-E genotype in West Africa is a relatively recent event, as reflected by the conspicuously low genetic diversity of this genotype despite the higher prevalence in this geographical region[28,29]. This fact together with the high degree of isolation and scattering of human populations in these regions favours the appearance of viral strains relatively stable over time, limited to specific geographical locations. In this sense, it must be noted that the country origin of the sub-Saharan patients included in our study is located in the South/Central regions of West Africa, corresponding to the area in which the genetic variability of HBV-E is lower, among the countries with high prevalence of this viral genotype[30,31]. In this context, the striking high proportion of preS deletions found in sub-Saharan patients in our series could reflect the existence of isolated viral strains of HBV-E in this region with a higher prevalence than that described for neighbouring countries. It may constitute a subpopulation of patients with increased risk of developing HCC by simultaneously sharing high-risk mutations in different regions of HBV genome.

Especially significant was the persistence of the high-risk mutations over time, as major strains of the viral quasispecies, as demonstrated by its identification with bulk sequencing techniques. As regards amino acid substitutions, these were generally very stable, although changes in the composition were observed. Thus, G1899A mutation remained continuously detectable during a follow-up period of up to 48 mo, with sporadic disappearances of the mutant strain in 3 patients and a definitive loss in only 1 patient. This persistence of point mutations in the BCP/precore/core region can be related to the low level of viral replication, which hinders the appearance and selection of mutations that restore the wild-type genotype in the population. Regarding deletion mutations in the preS region, these were maintained without sporadic eliminations during the follow-up in all the patients who carried them, with proven periods of up to 40 mo. The persistence, without fluctuations, of these deletions seems to influence the presence in the geographical origin of the patients in which these strains are majority. Under these conditions, exclusive infection with deleted strains would make the emergence of wild-type strains impossible due to the absence of a genetic template on which to act. The persistence over time of these mutations favours their accumulation, as major strains, in chronic infected patients. Thus, it was possible to confirm in our patients, not only the coexistence of the mutations A1762T/G1764A and G1896A/G1899A widely documented and associated with a high-risk of carcinogenesis, but also the accumulation of 3-5 risk mutations in other 5 patients, including a case that presented 3 risk mutations in the BCP/precore/core in combination with a deletion in the preS region. In any case, the persistence and accumulation over time of these mutations may have clinical relevance, since it perpetuates the molecular processes by which these viral variants favour the development of HCC.

The selection in this work of patients in the grey zone of treatment, with low-moderate fibrosis levels and normal or marginally elevated transaminase levels, makes it difficult to evaluate unless through a long-term follow-up, the possible harmful effect of the presence of these mutations on the progression of liver disease. However, and even recognizing the limitations of transient elastometry techniques for discriminate patients with low or moderate stages of fibrosis, our data showed a higher prevalence of F2 stage in patients with preS2 deletion than in those in which this mutation was not detected. On the other hand, the behaviour of AST and ALT was quite different between both types of patients, with is a higher percentage of patients with fluctuant AST or ALT levels during follow up in patients harbouring preS deletions.

These results focus an increased risk of HCC development in sub-Saharan patients. Africa represents one of the most HBV endemic regions in the world and HCC mainly due to CHB infection is a major cause of premature death, suggesting that viral strains located in Africa, which mostly belong to HBV-A, are more likely to cause HCC[32]. In this way, occult HBV infection was shown to be present in 75% of Black Africans with HCC, and genotype A was shown to be 4.5 times more likely than other genotypes to cause HCC in Black Africans, whilst tumours occurred at a significantly younger age[33]. Our findings also correlate a higher risk of HCC development with HBV-E, which almost exclusively occurs in African people. Moreover, it has been linked the distribution of HBV genotype E infection with African countries with high incidences of HCC[34].

In conclusion our data indicate that the presence of preS mutations should be assessed in patients with sub-Saharan origin, especially if they are infected with HBV-E and HBV-A genotype, in order to identify subpopulations of patients in which the antiviral treatment can be indicated to minimize the risk of HCC due to accumulation of high-risk HBV genetic variants. In this sense, the development of detection systems for deletion mutations in preS region is technically feasible with conventional techniques of Molecular Biology available for most centers. Its use, especially if it is restricted to preselected subpopulations, can help the treatment decision in patients in grey area of treatment.

The most important limitations of this study are: (1) The number of patients. It must be increased to elucidate the significance of prevalence data of some mutations; (2) The necessity to perform longitudinal studies with longer follow-up periods to clarify the progression of liver disease in this kind of patients. However, our data confirm the accumulation and persistence of HBV mutants related with HCC in different regions of HBV genome especially in the subpopulation of sub-Saharan origin.

The authors thank Dr. Pablo Castán for his careful language assistance.