Potential conflict of interest: Nothing to report.
Occult HBV infection is characterized by the persistence of HBV DNA in the liver of individuals negative for HBV surface antigen (HBsAg). Occult HBV may exist in the hepatocytes as a free genome, although the factors responsible for the very low viral replication and gene expression usually observed in this peculiar kind of infection are mostly unknown. Aims of this study were to investigate whether the viral genomic variability might account for the HBsAg negativity and the inhibition of the viral replication in occult HBV carriers, and to verify in vitro the replication capability of occult HBV strains. We studied liver viral isolates from 17 HBV patients, 13 with occult infection and 4 HBsAg-positive. Full-length HBV genomes from each case were amplified and directly sequenced. Additionally, full-length HBV DNA from eight occult-HBV and two HBsAg-positive cases were cloned and sequenced. Finally, three entire, linear HBV genomes from occult cases were transiently transfected in HuH7 cells. Direct sequencing showed the absence of mutations capable of interfering with viral replication and gene expression in the major viral population of each case. Cloning experiments showed highly divergent HBV strains both in HBsAg-positive and HBsAg-negative individual cases (range of divergence 1.4%-7.1%). All of the 3 transfected full-length HBV isolates showed normal patterns of replication in vitro. Conclusion: Multiple viral variants accumulate in the liver of occult HBV-infected patients. Occult HBV strains are replication-competent in vitro, suggesting that host, rather than viral factors are responsible for cryptic HBV infection. (HEPATOLOGY 2007;45:277–285.)
Occult HBV infection is characterized by the persistence of HBV DNA in the liver tissue (±serum) of hepatitis B surface antigen (HBsAg)-negative individuals with or without serological markers of previous infection (namely, antibodies to HBsAg and/or to HB-core antigen).1–3 Occult HBV infection is unexpectedly frequent—in particular in hepatitis C virus (HCV) carriers—and, besides its biological interest, it is of potential clinical relevance because it (1) can be transmitted by blood transfusion or liver transplantation inducing typical hepatitis B in the recipients; (2) may reactivate in conditions of immunosuppression causing severe forms of acute hepatitis B; (3) might favor the progression of the liver disease toward cirrhosis in patients with HBsAg-negative chronic hepatitis; and (4) represents an important risk factor for the development of HCC.2–9 Despite its biological and clinical implications, however, the virological features of occult HBV have been insufficiently highlighted. One main question concerns the factor(s) involved in the lack of circulating HBsAg despite the persistence of viral genomes. A first hypothesis is that this event might be attributable to rearrangements in the S gene leading to the synthesis of an antigenically modified S protein undetectable by the commercially available HBsAg assays.10–12 However, the studies on HBV variability have generated conflicting results so far.4 Moreover, sequence analysis has been rarely extended to the full-length viral genome to detect mutations outside the S-gene capable of interfering with its expression and/or with viral replication,1, 13, 14 and, importantly, most of the available data rely on HBV isolates from the serum of the patients, whereas it must be stressed that in occult HBV infection viral genomes persist in the liver in the presence of very low or undetectable levels of serum HBV DNA.1–4 Consequently, we do believe that entire HBV strains isolated from liver tissues must be examined to investigate viral heterogeneity in the case of occult HBV infection.
Aims of this study were to analyze the genomic variability of viral populations infecting the liver of occult HBV carriers, and to explore in vitro the replication capacity and gene expression of different occult HBV isolates.
BCP, basal core promoter; CTL, cytotoxic T lymphocyte; HBeAg, hepatitis B e antigenl HBsAg, hepatitis B surface antigen; RT, reverse transcriptase.
Patients and Methods
A detailed description of patients and experimental procedures, as the GenBank accession numbers of HBV isolates, are provided as supplemental material.
We studied frozen liver specimens from 17 HBV-infected individuals, 13 HBsAg-negative with occult infection, and 4 HBsAg-positive included as a control group (Table 1). The study protocol was performed according to the principles of the Declaration of Helsinki, and informed consent was obtained from all the patients
Table 1. Demographic, Clinical, and Virological Characteristics of the Patients with Occult or Overt HBV Infection
Analysis of the Entire PreS-S Genomic Region of HBV DNA by PCR and Cloning Procedures
DNA was extracted from frozen liver specimens by standardized procedures,1, 13 and amplified by PCR techniques using the Expand High Fidelity PCR System (Roche Diagnostics, Mannheim, Germany) and oligonucleotide primers specific for HBV DNA sequences flanking the entire S genomic region (Supplementary Table 1). Amplicons were cloned using the TOPO TA cloning kit (Invitrogen, Paisley, UK).
Amplification and Cloning of Full-Length HBV Genomes
HBV genomes were full-length amplified using two different PCR methods. HBV sequences from HBsAg-positive patients were obtained by a one-step full-length genome amplification method,15 whereas full-length HBV DNAs from patients with occult HBV infection were amplified by a nested PCR approach using 2 different primer pairs (Supplementary Table 1) to obtain the amplification of 2 overlapping subgenomic fragments, namely “fragment A” and “fragment B” (Supplementary Fig. 1), covering the entire HBV genome.13 The amplicons were cloned into the pCR2.1 plasmid of the TOPO TA cloning kit (Invitrogen).
Nucleotide and Amino Acid Sequences Analysis
Nucleotide and amino acid sequences of HBV DNA isolates from occult HBV-infected patients and HBsAg-positive controls were aligned using the CLUSTAL W program.16 Genotyping of HBV DNA sequences was performed by comparison with the representative genotypes along the entire genome.
HBV DNA Transfection
Three different constructs (LT3.1, LT4.2, and LT7.4) containing full-length HBV genomes isolated from occult HBV cases 3occ, 4occ and 7occ, respectively, were used to perform transient transfections. Each isolate represented the dominant strain of the HBV population infecting the liver of each patient. Linear HBV monomers were released from LT3.1, LT4.2, and LT7.4 constructs and from the wild-type (WT) control plasmid pUC-HBV (genotype D, subtype ayw) by cleavage with SapI restriction enzyme (New England Biolabs), gel purified, and transiently transfected into HuH7 and HepG2 human hepatoma cells using the FuGENE transfection reagent (Roche).
HBV DNA and RNA Analyses
HBV DNA was purified from intracellular and extracellular core particles and detected by Southern blot analysis as described.17 Total RNA was extracted from HuH7 cells after HBV DNA transfection and analyzed by Northern blot as described.17
Hepatitis B Antigens in the Medium
HBsAg in the cultured medium and HBV “e” antigen (HBeAg) in culture supernatants were measured by ELISA assays (DiaSorin S.pA., Saluggia, Italy).
Genomic Variability of Intrahepatic Occult HBV
Analysis of the PreS-S Genomic Region (by Cloning and Sequencing).
The genomic variability of HBV strains isolated from the liver of 17 patients (13 occult HBV-infected and 4 HBsAg-positive controls), was studied by amplifying and cloning the entire HBV preS-S genomic region. From 5 to 10 clones for each patient were sequenced. Sixteen patients were infected with genotype D HBV strains, whereas the virus infecting one occult HBV case belonged to genotype A. A considerable intraindividual genetic diversity was noted among HBV sequences derived from each patient, irrespective of the occult or overt status of the infection (Table 2). In particular, viral nucleotide sequences differed from each other with evolutionary distances up to 0.044 by the Kimura's 2-parameter method. Within the preS1 region, a deletion of codon 108 was observed in all of the clones from one patient with occult HBV infection (patient 7occ) (Table 3). In addition, in all of the HBV strains from the same patient the preS2 start codon was abolished by a point mutation (ATG→ATA). Whereas no important mutation was found in the preS1 region of HBV clones from the remaining 16 patients, the preS2 start codon was mutated in another two patients with occult HBV infection (Table 3). Within the preS2 region, in-frame deletions were present in HBV strains from 1 patient (patient 10occ) with occult infection (Δ16-23 amino acid, aa) and from 2 patients (2HBs and 3HBs) of the control group (Δ15-23 aa and Δ18-20 aa, respectively). The analysis of the S region showed that the HBV strains from 12 of the 13 occult HBV-infected patients and from 1 of the 4 control patients had from 1 to 3 aa changes within the putative human leukocyte antigen class I–restricted cytotoxic T lymphocyte (CTL) epitope (aa28-51) of the HBsAg.18 Moreover, substitutions within the major hydrophilic region (MHR) of HBsAg (aa104-172) were found in occult (1-10 aa/patient) and overt (4-7 aa/patient) HBV infection (Table 3). Only 4 aa changes involved positions located inside the immunodominant “a” determinant of the HBsAg (aa 124-147). Most substitutions (M125T, T127L, T131I, Y134F/N) were within the first hydrophilic loop of the “a” determinant in both occult and overt HBV cases (Table 3). Neither the G145R substitution nor other mutations reported to abolish the double-loop structure of the “a” determinant19 and to interfere with the HBsAg detection were found in any of the HBV isolates. Finally, the recently described point mutation causing a G→S aa substitution at position 265 and abolishing the HBsAg expression20 was not found in any HBV clone from the 17 cases.
Table 2. Comparison Among HBV Isolates from Patients with Occult or Overt HBV Infection Within the Entire PreS-S Genomic Region
Analysis of HBV Core, pol, X Genes, and Transcription Regulatory Regions (by Direct Sequencing).
HBV variability was further investigated by PCR amplification and direct sequencing of the entire HBV genomes isolated from all of the 17 cases included in the study. The nucleotide mutation G1896A, introducing a stop signal at codon 28 within the precore region and preventing the HBeAg synthesis, was found in 8 of the 13 cases with occult HBV and in all 4 HBsAg-positive cases (Table 3). This stop codon was also present in the genotype A strain infecting patient LT3, which also carried the C1858T mutation required to stabilize the hairpin structure of the genotype A encapsidation signal in the presence of the G1896A substitution.21
A number of missense mutations within different core antigen immunogenic epitopes were observed in HBV isolates from both patients with occult infection and control cases. In particular, a high number of mutations was found within the CD4+ T-cell epitope located between aa 48 and 69 in the core protein.22 Eight different aa changes (S49T, I59V, L60V, V63G, E64D, M66L, T67N, A69S) were detected within this epitope, although only a single aa substitution was present in most of the HBV isolates, with the exception of the occult HBV patient 10occ showing four aa substitutions (E64D, M66L, T67N, A69S) in epitope 48-69. Interestingly, 10occ genomes carried five additional point mutations (V74G, E77Q, P79Q, A80T, S87T) known to reduce both HBe and HBc antigenicity (Table 3).23, 24 Moreover, all patients except one with occult infection showed 1 to 5 point mutations within the HBcAg hot-spot mutational domain (aa 80 to 120) previously reported to be associated with severe forms of liver disease.23, 25, 26 Concerning the core protein CD8+ CTL-epitopes, we found 1 to 2 aa substitutions (including L19S, S21A/G, D22E, F24Y/L, and V27A) within the epitope 18-27 of HBV isolates from 8 cases with occult infection and 2 HBsAg-positive cases, whereas the CTL-epitope 141-151 was conserved in the viral isolates from 14 patients and showed a single point mutation in isolates from each of the other 3 cases (Table 3).
Pol gene analysis showed that no viral strain carried the T→P substitution in the Pol terminal protein that potentially impairs the pregenome encapsidation and induces occult infection.27 The M204I substitution inside the conserved YMDD motif of the reverse transcriptase (RT) domain of the Pol-gene was found in HBV isolates from 5 patients with occult HBV infection and in no control case. The M204I substitution was associated with the RT L80V/I mutation in 2 of those 5 cases (Table 3). Both of these mutations are usually selected during lamivudine treatment,28 whereas none of our patients had been previously treated with antivirals. Finally, the S317P aa mutation within the RT domain, shown to affect HBV replication efficiency,29 was not found in any of the HBV strains from our patients. Moreover, none of the isolates carried any of the mutations in the RNase-H coding region reported to block mature HBV DNA synthesis because of the abnormal persistence of the RNA:DNA heteroduplex generated during reverse transcriptase.30, 31 Finally, all CTL-epitopes in the Pol-gene32 were well conserved in both occult and overt HBV isolates (Table 3).
The analysis of the X gene showed that the dominant HBV populations infecting the 17 studied patients had no mutation able to impair the X protein production. In 6 patients with occult infection and in 3 HBsAg-positive cases, we found HBV genomes carrying the I127T, K130M/Q, V131I/L/G aa substitutions (Table 3) previously identified in isolates from patients with severe acute and chronic liver diseases33–36 and with HCC.37, 38
The regulatory elements scattered over the entire HBV genome were also analyzed in all viral isolates from the 17 studied patients to detect mutations potentially implicated in the inhibition of viral replication and/or gene expression. The preS1 promoter had a WT sequence in all but 2 patients (Supplementary Fig. 2), whereas the S promoter showed no mutations capable of impairing preS2/S expression (Supplementary Fig. 3). The analysis of the enhancer I/X promoter revealed several point mutations in HBV genomes from patients with both occult and overt HBV infection. Case 9occ showed a number of unique mutations located within the central core domain of enhancer I, which contains binding sites for transcription factors and plays a major role in regulating HBV gene expression39, 40 (Supplementary Fig. 4). Enhancer II analysis showed the presence of the point mutations C1665A and C1678T (located in the HNF4- and HNF3-binding sites, respectively) in the HBV genomes from most of the patients with either occult or overt infection. On the contrary, the A1625C and A1679C mutations in the SP1 binding site and in the HNF3 binding motif, respectively, were found in three patients with occult infection (Supplementary Fig. 5). Analysis of the basal core promoter (BCP) showed the presence of one to 4 point mutations in HBV isolates from almost all of the 17 patients. In particular, the double A1762T and G1764A nucleotide exchange, which has been reported to lead to a decrease of HBeAg expression and an enhanced viral replication,35, 36 was found in 4 of 13 patients with occult HBV infection and in 3 of 4 HBsAg-positive cases. Two occult HBV patients also carried the C1766T mutation together with the A1762T and G1764A substitutions. Such triple mutations was reported to confer significantly higher viral replication than A1762T/G1764 alone.41
Phylogenetic Analysis of Occult HBV Isolates
We next performed, based on material availability, cloning and subsequent phylogenetic analysis of the full-length HBV DNA from eight cases with occult (1occ-8occ) and two with overt (1HBs and 2HBs) HBV infection (Fig. 1). From 2 to 8 clones (median, 5) were studied for each case. Nucleotide sequences alignment confirmed the presence of quasispecies diversity in individual patients with occult HBV infection and the degree of heterogeneity observed in the preS-S region. In particular, we did not detect additional mutations capable of interfering with the viral activities, with the exception of important rearrangements in C-terminal end of the X protein found in three of five, two of five, and one of eight clones from patients 2occ, 6occ, and 5occ, respectively. In patient 2occ, HBV clones LT2.1 and LT2.2 had, respectively, an in-frame deletion of 60 aa (Δ 69-129) and 61 aa (Δ 72-131), whereas clone LT2.4 showed 2 different in-frame deletions of 61 and 6 aa, respectively (Δ72-131 and Δ139-144). In patient 6occ, HBV clones LT6.2 and LT6.4 showed a deeply rearranged sequence from aa 77 to 144. Finally, 1 of the 8 clones from patient 5occ showed an in-frame deletion of 69 aa (Δ77–144 aa).
Functional Analysis of Occult HBV Genomes
Three cloned, full-length occult HBV genomes (LT7.4, LT3.1, and LT4.2), isolated from patients 7occ, 3occ, and 4occ, respectively, were used in transient transfection experiments of HuH7 cells. A standard wild-type HBV genome of genotype D (WT HBV) was used as a control. Among the three HBV genomes, LT3.1 was genotype A and carried the precore stop codon that prevents the HBeAg synthesis. LT4.2 carried both the preS2 start codon and the precore stop codon mutations. LT7.4 had both the deletion of the last codon of the preS1 and the mutated preS2 start codon. Two days after transfection, HBV DNA from intracellular replicative intermediates and extracellular viral particles were analyzed by Southern blotting. As shown in Fig. 2 A,B, all of the 3 HBV genomes were replication competent and were able to release viral particles into the cell culture medium. Northern blot analysis showed pregenomic/pre-C messenger RNA signals of LT7.4, LT3.1, and LT4.2 comparable to those of the control (Fig. 2C). The amount of HBsAg secreted from cells transfected with LT3.1 genome was slightly higher, as compared with that from WT HBV-transfected cells. In contrast, HBsAg was detected at low concentrations in the medium of cells transfected with LT7.4 and LT4.2 isolates (Fig. 2D). HBeAg was detected only in the medium of the LT4.2-transfected cells because the LT7.4 and LT3.1 genomes carried the precore stop codon (Fig. 2E).
In this study, we explored the possible role of HBV genomic variability as the cause of 2 major biological aspects that characterize the occult HBV status: the lack of HBsAg production (or detection) and the inhibition of viral replication. We therefore isolated and examined entire viral genomes from a congruous number of occult HBV carriers, and compared them with isolates from HBsAg-positive individuals. In particular, both coding and transcription regulatory regions of each HBV isolate were analyzed in detail. Moreover, the capacity of a number of occult isolates to be transcribed and to replicate was verified in vitro. Our results clearly demonstrate the absence in all cases of relevant mutations at the level of the genomic HBsAg coding region, thus indicating an impaired synthesis rather than a failure in antigen detection. Concerning the preS region, most of the isolates did not show any significant mutation. Only 1 case had a small in-frame deletion in the preS1 region, and 2 more cases had a mutated preS2 start codon with a consequent defect in the preS2 protein synthesis. These viruses appeared to normally replicate once transfected in HuH7 cells and to be competent in HBsAg production, even if the amount of HBsAg in the cell culture medium was lower when compared with preS wild-type strains. Indeed, the preS2 protein is known to be dispensable for HBV replication and transmission, and the preS2-defective HBVs are frequently detected in chronic HBsAg carriers, where, interestingly, they appear to be significantly associated with cirrhosis and HCC.42, 43 All of our occult HBV cases had HCC, and some of them carried viral variants (preS2-defective and/or core-promoter mutants and/or X-deleted strains) strongly suspected to have pro-oncogenic activities that might be maintained also when the infection is cryptic.13 Recently, genomic alterations (mainly deletions) of the preS-S regulatory regions have been reported to impair HBsAg production, becoming a major factor in determining occult infection.14 However, we have carefully examined these regions in our viral isolates, and we did not detect any relevant rearrangement. Moreover, we did not find any of the mutations located in different parts of the viral genome previously reported to be associated with—and possibly responsible for—individual cases of occult HBV infection.2
Detection of YMDD mutated strains in 5 of the 13 (38%) occult HBV cases examined is of particular interest. These mutants are known to usually emerge under lamivudine therapy, whereas none of our patients had previously been treated with anti-virals. Interestingly, a similar observation was made in 2 previous studies on occult HBV carriers.14, 44 Why the presence of the YMDD variants is quite frequent in occult HBV infection, whereas it has only anecdotally been reported in untreated patients with overt infection, is unknown. The hypothesis that this mutant could be transmitted from HBsAg-positive individuals who developed lamivudine resistance is not plausible, considering that all of our patients underwent surgical resection between 1999 and 2000,13 before lamivudine was approved for the treatment of chronic hepatitis B in Italy. In vitro evidence suggests that YMDD variants are less efficient than wild-type strains in replicating.17, 45, 46 However, when they emerge in lamivudine-treated individuals, they usually induce a productive infection characterized by fairly high viremia levels and clinical reactivation that, in some cases, evolves into a severe acute hepatitis.47, 48 Altogether, that the selection of these mutants may play a role in the induction of the occult status is unlikely. On the contrary, this observation might be an indirect proof that the occult status is the consequence of a strong suppression of HBV transcriptional and/or replicative activity. In this perspective, the very few intrahepatic YMDD mutant strains that are likely present in most of the untreated HBV-infected patients may emerge only when the replication of the wild-type virus is strongly inhibited, thus generating the “replication space” needed by the less “fit” YMDD mutant.49 Obviously, this speculation might explain the high frequency of detection of YMDD mutants in occult HBV cases but cannot help to reveal the mechanisms causing the virus suppression that, in any case, also involves these mutants. These mechanisms have to exert a very strong inhibitory effect, because several of our occult isolates carried genomic mutations—such as those at basal core promoter level—that are known to confer an enhanced replicative capability to HBV.50 From the clinical point of view, the intrahepatic presence of YMDD mutants may acquire relevance in case of acute reactivation of the occult HBV infection because these variants are resistant to lamivudine treatment.
Another intriguing result of our study was the detection in almost all cases—regardless of the overt or occult status of the infection—of viral strains carrying mutations that modify different CTL epitopes in the envelope, the nucleocapsid, and the polymerase proteins. A negative selection of CTL escape mutants has been reported to be common during chronic HBV infection and to contribute to viral persistence.32 Although the emergence and selection of viral escape mutants and HBV quasispecies is not surprising in cases with “overt” infection and active viral replication, explaining how they can occur in cases with occult HBV infection where viral replication is suppressed is not easy. One may hypothesize that occult HBV status is generally preceded by a period of productive infection of variable length during which viral variants are selected. These variants subsequently persist also when the mechanisms strongly suppressing viral replication and causing the establishment of the occult “status” of the infection are overcome. However, the hypothesis that occult HBV carriers might be primarily infected by a pool of variant viruses already present in the infecting inoculum cannot be ruled out.
In conclusion, occult HBV populations show a large intra-individual genetic heterogeneity, which is comparable to that observed in HBsAg-positive subjects. In most cases, however, the viral genomic variability does not appear to play a fundamental role in inducing the occult HBV status, which is dependent on a strong suppression of the virus replication and gene expression. The mechanisms responsible for this suppression are, however, mostly unknown. The host immune system likely plays a critical role; however, epigenetic factors, which recently have been implicated in the regulation of the HBV replication,17 also might be involved in inducing—or contributing to—the establishment and/or maintenance of the occult status. To clarify the molecular bases of the occult HBV status is of pivotal importance both for understanding the HBV biology and the virus/host interplay and for the identification of new pharmacological approaches to the cure of HBV infection.