Origin of serum hepatitis B virus in acute exacerbation: Comparison with HBV in the liver and from other exacerbation

Authors

  • Chun-Jen Liu,

    1. Division of Gastroenterology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Jia-Horng Kao,

    1. Division of Gastroenterology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    3. Hepatitis Research Center, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Hurng-Yi Wang,

    1. Institute of Molecular Biology, Academia Sinica, Taipei, Taiwan
    Search for more papers by this author
  • Ming-Yang Lai,

    1. Division of Gastroenterology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Ting-Chih Chen,

    1. Division of Gastroenterology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author
  • Pei-Jer Chen,

    Corresponding author
    1. Division of Gastroenterology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    • Graduate Institute of Clinical Medicine, National Taiwan University College of Medicine, 1 Chang-Te St., Taipei 100, Taiwan
    Search for more papers by this author
    • fax: +886 2-23317624.

  • Ding-Shinn Chen

    1. Division of Gastroenterology, Department of Internal Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    2. Graduate Institute of Clinical Medicine, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
    Search for more papers by this author

Abstract

Acute exacerbation (AE) of chronic hepatitis B is usually preceded by reemergence or increase of hepatitis B virus (HBV) in the serum. To investigate the origin of the reemergence or increase, we compared the identity of the serum viral genome to that in the liver and in previous AE by full-length sequencing. The full-length viral genome and extent of quasispecies were obtained from serum and liver biopsy specimens at the same time from 9 subjects with hepatitis B exacerbation (group I). Composition of viral quasispecies was compared by the genetic diversity and the average number of nucleotide substitutions within and between different viral sources. Another 2 patients with repeated AEs (group II) were also enrolled, and their serial serum alanine aminotransferase, HBV DNA levels and full-length sequences were determined. In all group I patients, serum viral genome was identical to that in the liver. The genetic diversity and the average number of nucleotide difference were also comparable between serum and liver tissue. In 2 group II patients, the viral variant that emerged after previous AE was not identical to that caused by the subsequent AE. Dominant viral strains for serial AEs in a single patient did not show a sequential evolution, but presented as a horizontal selection of a minor population from the original viral pool. In conclusion, the findings suggest that viral strain in serum reflects the intrahepatic strain of the AE. Random reactivation of the original HBV pool, rather than a sequential evolution of one strain, also contributes to the onset of repeated AE. Supplementary material for this article can be found on the HEPATOLOGY website (http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). (HEPATOLOGY 2004;40:310–317.)

Acute exacerbation (AE) occurs frequently in hepatitis B surface antigen carriers.1 The exacerbation can accelerate the progression of liver disease.2 Hepatitis B virus (HBV) evolves rapidly in chronic hepatitis B patients, and the correlation of HBV variation with the onset of AE has been extensively studied.3–5 Previously, we investigated sequential full-length viral sequences in patients with AE6, 7 and found that most exacerbations were preceded by an upsurge of serum HBV identical to the preexisting HBV strain. After exacerbation, however, about half of the patients were repopulated by a different viral variant in the circulation.

The pathogenesis of AE actually takes place in the liver. Therefore, it is better to directly characterize the viral genome from the liver biopsy specimen. As noted previously, some potential pathogenic strains may replicate actively within the liver but not be released into the circulation.8 Thus, HBV genomic changes that are correlated with the development of hepatitis B exacerbation might occur and be retained within the liver but not be detected in the specimens obtained from the circulation. To resolve these questions, the representation of the serum viral genome to the corresponding intrahepatic one was first determined by comparing the dominant viral nucleotide sequence and genetic complexity of the viral quasispecies between both tissue specimens obtained simultaneously from individuals at different stages of hepatitis B exacerbation.

From the viral evolution point of view, we need to consider the relationship between individual HBV strains observed during consecutive AEs in a single patient. There are two possible models. The first suggests that HBV responsible for current AE is an offspring from the earlier one—that means a direct and sequential evolution resembling a parent-and-child relationship. There is plenty of evidence to support this model.9–11 The second model suggests that the HBV strain in each AE is independent from the other. Each one emerged from the patient's total viral pools irrespective of the previous one, thus strains from individual AEs would act like siblings. To address which model operates, we performed a longitudinal, full-length, viral genomic study in patients who developed consecutive episodes of AE as well as a phylogenetic analysis of their relationship.

Abbreviations

AE, acute exacerbation; HBV, hepatitis B virus; ALT, alanine aminotransferase; PCR, polymerase chain reaction; cccDNA, covalently closed circular DNA; kw, average number of nucleotide differences within different viral sources.

Materials and Methods

Collection of Patients and Sera

We identified 9 patients with hepatitis B exacerbation (group I) at the National Taiwan University Hospital. Liver biopsy was performed to examine the grade of necroinflammation and the stage of fibrosis as clinically indicated. Seven of the patients were sampled during the serum alanine aminotransferase (ALT) rising phase of AE (group Ia) (Table 1). The other 2 patients were sampled during the serum ALT declining phase of AE (group Ib). Serum specimens were also collected simultaneously from each patient for the comparative viral genomic study.

Table 1. Clinical, Virological, and Histological Features of 9 Chronic Hepatitis B Exacerbation Patients Whose Serum and Liver Biopsy Specimens Were Collected Simultaneously
CaseAge/SexHBeAg/Anti-HBeHBV GenotypeBilirubin (mg/dL)ALT (U/L)HBV DNA (pg/mL)Histology, Grade/Stage Score
B/ALT-S/AP (ALT-P)*B/DNA-S/AP (DNA-P)
  • Abbreviations: HBeAg, hepatitis B e antigen; B, baseline; ALT-S, serum ALT; AP, after exacerbation; ALT-P, peak of serum ALT; DNA-S, serum HBV DNA; DNA-P, peak of serum HBV DNA during exacerbation; M, male; Ba, Asia subtype of HBV genotype B; F, female; C, HBV genotype C; +, positive; −, negative.

  • *

    The serum ALT (ALT-S) and HBV DNA (DNA-S) levels when the sample was collected are expressed in bold.

  • Histological grade and stage scored according to Ishak et al.28

Group Ia Samples collected during ALT rising phase of exacerbation
Ia-135/M+/−Ba1.545/219/29 (343)1,510/18,000/234 (18,000)6/1
Ia-220/F+/−Ba1.631/157/69 (267)268/16,000/1,540 (16,000)6/0
Ia-347/M−/+Ba2.368/424/124 (860)5,800/11,000/3,640 (11,000)5/2
Ia-421/M+/−Ba1.265/220/45 (335)25/16,480/780 (16,480)6/1
Ia-545/F+/−Ba1.128/440/62 (1,017)<2.5/1,400/<2.5 (11,030)4/2
Ia-642/M+/−Ba0.665/252/77 (989)11,000/18,000/380 (18,000)7/1
Ia-726/M+/−Ba0.936/160/88 (502)6,580/12,440/690 (12,440)4/1
Group Ib Samples collected during ALT declining phase of exacerbation
Ib-142/M−/+C0.635/112/45 (647)440/<2.5/<2.5 (2,540)4/2
Ib-256/M−/+C0.545/95/46 (812)168/<2.5/8 (574)5/3

Another two hepatitis B surface antigen carriers who experienced consecutive AEs of chronic hepatitis B (group II) were also prospectively collected. After enrollment, we evaluated serum ALT and HBV DNA levels monthly throughout the course of hepatitis B exacerbation (flare) in each patient. The diagnosis of hepatitis B exacerbation was based upon an abrupt elevation of serum ALT level to beyond 200 U/L, or a greater than 3-fold increase in baseline serum ALT level.12 Sera obtained at 4 points of each episode of AE were collected: at baseline (immediately before the surge of serum HBV DNA, ALT within or nearly within normal range), at the peak of serum HBV DNA level (virological peak), at the peak of serum ALT level (ALT peak), and after ALT peak. Part of the sera was stored at −70°C to await further virological assay.

Two hepatitis B e antigen–positive carriers who had persistent serum ALT less than 1.5-fold the upper limit of normal for 1 year served as controls. To determine any change of the dominant viral sequence, 2 separate sera 1 year apart were obtained from each patient.

Hepatitis Virus Markers

The sera were tested for hepatitis markers with commercial kits (AxSYM System, Abbott Laboratories, North Chicago, IL). Serum HBV DNA was determined by a branched DNA assay (QUANTIPLEX HBV DNA Assay, Version 2, Bayer Corporation, Tarrytown, NY; detection limit: 2.5 pg/mL, or 700,000 genome equivalents/mL).

Virological Assays

Extraction, Amplification, and Sequencing of Serum and Liver HBV DNA.

Serum DNA was extracted and collected as previously described.6, 7 Liver DNA was extracted from 10 to 25 mg liver biopsy tissue specimens stored in liquid nitrogen using a commercially available kit (QIAamp DNA Mini Kit, QIAGEN Inc., Valencia, CA) and collected in 200 μL ddH2O. We then performed polymerase chain reaction (PCR) and sequencing of the full-length genome of the dominant HBV strain, as previously described.6, 7, 13

Preparation and Sequencing of HBV Covalently Closed Circular DNA (cccDNA) From Liver Biopsy Specimens.

From the life-cycle point of view, all serum and intrahepatic HBVs originate from the cccDNA within the nucleus of the infected hepatocyte.14 To further clarify the representation of the serum viral genome to cccDNA HBV, we also compared the subgenomic nucleotide sequence and genetic complexity among the circulating viral genome, the intrahepatic one, and the cccDNA in case Ia-7. Two strategies were adopted in this study to enrich cccDNA, as previously described15, 16: (1) treatment of HBV DNA with a mung bean nuclease before amplification with a primer set across the nick region, and (2) a highly selective PCR assay for cccDNA, taking advantage of the gap and incomplete region in the partially double-stranded virion DNA (cccForward: nucleotide 1540-1559, 5′-GCGGWCTCCCCGTCTGTGCC-3′; cccReverse: nucleotide 1922-1902, 5′-CTTTATACGG-GTCAATGTCCA-3′).

Accordingly, in case Ia-7, a total of 2 μg of DNA extracted from liver biopsy tissue by a commercial kit was first treated with 10 U of mung bean nuclease (New England Biolabs, Beverly, MA) in 50 μL of reaction mixture containing 50 mmol/L sodium acetate (pH 5.0), 30 mmol/L sodium chloride, and 1 mmol/L zinc sulfate, as previously described.16 As a control, 5 μL DNA solution extracted from 200 μL serum obtained simultaneously from case Ia-7 was also subjected to digestion with mung bean nuclease under the same condition. After incubation at 30°C for 30 minutes, the mixture was purified by phenol/chloroform (1:1) extraction and precipitated with absolute ethanol. Finally, extracted DNA was dissolved in 10 μL ddH2O for further PCR specific for cccDNA. In each PCR mixture, about 105 viral genome equivalents were added. The amplicon was directly sequenced by using cccForward as a sequencing primer.

Determining the Similarity of Viral Nucleotide Sequence Among Serum, Liver Specimen, and Intrahepatic cccDNA.

The similarity between dominant serum and liver full-length viral genome was determined in all group I patients. The subgenomic viral nucleotide sequences among serum, liver biopsy specimen, and intrahepatic cccDNA were further compared in case Ia-7.

Analysis of HBV Quasispecies Obtained Simultaneously From Serum, Liver Specimens, and Intrahepatic cccDNA.

In addition to direct sequencing, the subcloning and sequencing of full-length or subgenomic PCR products obtained from serum, liver specimens and intrahepatic cccDNA were also performed in randomly selected cases as previously described.6, 7, 13

Determination of the Nucleotide and Derived Amino Acid Changes of the Dominant HBV Strains in Patients With Consecutive Episodes of AE.

The nucleotides and their derived amino acid sequences of the dominant HBV strains obtained at different time-points in each AE were compared with the corresponding one at enrollment (baseline of first AE) for each patient. The changes of amino acids encoded from 4 different open reading frames for polymerase, core, surface, and X proteins were analyzed simultaneously. The relationships of all the HBV strains obtained during repeated AE were further analyzed in a phylogenetic tree.

Phylogenetic Analyses of Viral Sequences

Viral nucleotide sequences were aligned by using the default parameter of CLUSTAL W.17 Phylogenetic relationships were constructed by neighbor-joining method18 based on the Kimura 2-parameter model.19 To evaluate the statistical support, 1,000 bootstrap replications were performed. The tree construction, bootstrap replications, and determination of genetic distance were conducted by Molecular Evolution Genetics Analysis software package version 2.1.20

Mantel's test was carried out to determine if the sequences from a given source were genetically closer to each other than to sequences from other sources as described previously,21, 22 with the help of software written by Philippe Casgrain (http://www.fas.umontreal.ca/biol/casgrain/fr/labo/R/index.html). In our analysis, 1,000 permutations were performed to get the exact P value of the observed correlation.

DNA polymorphism, including number of polymorphic sites, nucleotide diversity, and average number of nucleotide differences within (kw) and between different viral sources, were estimated by DnaSP.23 If the viral clones of different tissues were actually derived from the same origin, the kw will be close to the average number of nucleotide differences between different viral sources. On the contrary, if viral clones collected from different tissues indeed have different origin, the average number of nucleotide differences between different viral sources will be greater than the kw.

The nucleotides were numbered from the EcoRI site.24 HBV genotype B and C reference strains obtained from GenBank (accession numbers D00331 and D50520, respectively) serve as outgroups. All the dominant HBV full-length sequences were deposited in the GenBank database (accession numbers AY596102 to AY596112).

Ethical Considerations

The study was approved by the Ethical Committee of the National Taiwan University Hospital, and the sera were sampled after obtaining written informed consent from the patients.

Statistics

The mean number of nucleotide variation per genome at each time-point was compared by using the paired Student t test. Because of the correlated nature of the observed samples, generalized estimating equation model was used to correct the statistical inference. In all analyses, a one-tailed P < .05 was considered statistically significant.

Results

Characteristics of Patients

The sequential clinical, virological, and histological data in the 9 patients with hepatitis B exacerbation (7 males, 2 females; age, 21–56 years) are shown in Table 1. The clinical and virological characteristics of the 2 group II patients with repeated AE are shown in Table 2. Case II-1 developed 3 episodes of AE and case II-2 developed 2 episodes of AE. The serial serum ALT and HBV DNA levels and the time-points to sample serum specimens for viral genomic analyses are shown in Figure 1.

Table 2. Clinical and Virological Data of 2 Hepatitis B Surface Antigen Carriers Who Developed Repeated Acute Exacerbation
CaseAge/SexHBeAg/Anti-HBe/HBV GenotypeInterval Between 1st and 2nd AE (Month)Interval Between 2nd and 3rd AE (Month)First AESecond AEThird AE
ALT (U/L)HBV DNA (pg/mL)ALT (U/L)HBV DNA (pg/mL)ALT (U/L)HBV DNA (pg/mL)
  • Abbreviations: HBeAg, hepatitis B e antigen; B, baseline; DNA-P, peak of serum HBV DNA; ALT-P, peak of serum ALT; AP, after peak of serum ALT; M, male; +, positive reaction; −, negative reaction.

  • *

    Case II-2 did not develop a third episode of AE.

     B/DNA-P/ALT-P/APB/DNA-P/ALT-P/APB/DNA-P/ALT-P/APB/DNA-P/ALT-P/APB/DNA-P/ALT-P/APB/DNA-P/ALT-P/AP
II-135/M+/−/Ba5775/269/1490/78250/3600/1248/35936/256/1480/53120/1,694/41/<2.514/186/259/77<2.5/>18000/1954/200
II-241/M+/−/Ba5*31/62/1148/3115/9393/2.8/<2.527/354/1278/54<2.5/>18000/352/<2.5**
Figure 1.

Clinical and virological profiles through the course of repeated acute exacerbation of chronic hepatitis B: (A) in case II-1 with 3 acute exacerbations and (B) in case II-2 with 2 acute exacerbations. The X axis represents the months after enrollment. The serum ALT levels are expressed as ALT over ULN ratios. The upper limit of normal, which equals 41 U/L for serum ALT at National Taiwan University Hospital. The unit of serum HBV DNA level before applying logarithm is pg/mL. The arrows indicate the time-points when sera were obtained for full-length amplification of HBV genome. AE, acute exacerbation; HBV, hepatitis B virus; ALT, alanine aminotransferase; ULN, upper limit of normal.

Comparison of the Viral Genomes Between Serum and Liver Specimens in Patients With Hepatitis B Exacerbation

Comparison of the Dominant Viral Nucleotide Sequences.

The dominant full-length viral genome obtained from the serum specimen was identical to that obtained from the liver biopsy specimen in all group I patients.

Comparison of the Genetic Complexity of the Viral Quasispecies.

The complexity of the viral quasispecies was also comparable between serum and liver viral clones in 2 group Ia patients (cases Ia-1 and Ia-2). Eight to 12 full-length HBV clones were obtained after PCR from either the serum or liver biopsy specimen in each case. The full-length viral nucleotide consensus sequence of all the serum and liver HBV clones was first determined by multiple alignments. Then individual serum or liver clonal genomic difference from the viral nucleotide consensus sequence was determined. Overall, the number of nucleotide differences in each serum or liver viral clone from the consensus sequence ranged between 1 and 26 in case Ia-1, and between 1 and 24 in case Ia-2 (Supplementary Table 1).

Levels of viral DNA polymorphism, in terms of number of mutations and nucleotide diversity, are shown in Table 3. In case Ia-1, viral clones from the liver biopsy specimen exhibited a higher level of nucleotide polymorphism than those from serum (nucleotide diversity, 0.0056 vs. 0.0041; kw, 16.14 vs. 10.57), but the opposite was observed in case Ia-2 (nucleotide diversity, 0.0033 vs. 0.0042; kw, 10.45 vs. 13.50). Nevertheless, the differences were not statistically significant (P > .05). In both cases, the average number of nucleotide differences between viral clones from different tissues was not greater than those from the same tissue (kw), which indicated that all viral clones may actually have the same tissue origin. Results from Mantel's test further supported this point. The correlation coefficient between genetic distance and the origin of viral clones was 0.025 (P > .3) and 0.024 (P > .3) for case Ia-1 and case Ia-2, respectively. To sum up, viral clones from the same tissue were not genetically closer to each other than to sequences from other tissues.

Table 3. Comparison of Viral Quasispecies Composition Between Serum and Liver Biopsy Tissue Specimen
  LiverSerum
  • *

    P > .05.

Case Ia-1 n = 8n = 8
 Number of mutations5334
 Nucleotide diversity0.0056*0.0041*
 Average number of nucleotide difference within different viral sources16.1410.57
 Average number of nucleotide difference between different viral sources13.55
Case Ia-2 n = 12n = 8
 Number of mutations3942
 Nucleotide diversity0.0033*0.0042*
 Average number of nucleotide difference within different viral sources10.4513.50
 Average number of nucleotide difference between different viral sources11.50

Phylogenetic Analysis.

The phylogenetic relationships of all serum and liver viral clones in each case revealed that the viral sequences derived from different tissues were clustered together, which indicated that there was no evidence of genetic differentiation among viral clones from different origins (data not shown). In case Ia-7, viral clones derived from serum, liver biopsy specimen, and intrahepatic cccDNA were also sequenced and compared. Again, sequences from serum and liver intermingled with cccDNA and showed no specific clusters (Fig. 2A).

Figure 2.

(A) Phylogenetic analysis of serum hepatitis B virus (HBV) clones (S-C), liver HBV clones (L-C), and cccDNA clones (cccDNA-C) obtained from case Ia-7, based on the subgenomes (nt 1540-1922) encompassing gap and incomplete region in the partially double-stranded virion DNA. S-C1 denotes the viral nucleotide sequence of serum clone-1. The rest may be deduced by analogy. Phylogenetic analysis of (B) serum HBV strains obtained from case II-1 and of (C) serum HBV strains obtained from case II-2, based on full-length hepatitis B viral nucleotide sequences collected through repeated episodes of hepatitis B exacerbation. 1-B denotes full-length viral nucleotide sequence obtained at baseline of the first acute exacerbation. The rest may be deduced by analogy. The phylogenetic tree was constructed by the program of MEGA version 2.1 and is rooted. HBV genotype B and C reference strains obtained from GenBank (accession numbers: D00331 and D50520, respectively) serve as outgroups. cccDNA, covalently closed circular DNA; DNA/P, peak of serum HBV DNA level; B, at baseline; AP, after alanine aminotransferase peak; ALT, alanine aminotransferase; ALT/P, ALT peak.

Nature of HBV Genome and the Development of AE in Patients With Repeated Episodes of AE

In both patients, when compared with the corresponding viral genome at enrollment, the nucleotide substitutions and derived amino acid changes in individual viral genome obtained at different time-point are shown in Supplementary Tables 2 and 3 and summarized in Table 4. Particularly, in case II-1, viral strain emerged after AE was not identical to the offending virus in subsequent AE. In addition, the viral strains responsible for the viral surge in each AE were quite different. The viral strains responsible for the viral surge during the first, second, and third AE were 0, 12, and 27 nucleotides, respectively, different from the one at enrollment. Similar findings could be observed in case II-2.

Table 4. Summary of Hepatitis B Viral Genomic Evolution in 2 Patients With Repeated AEs in Comparison With The One at Enrollment (Baseline of First AE)
 First AESecond AEThird AE
BaselineDNA/PALT/PAPBaselineDNA/PALT/PAPBaselineDNA/PALT/PAP
  1. NOTE. Bar indicates no changes when compared with the viral genome at baseline of first AE. Third AE does not occur in Patient II-2.

  2. Abbreviations: DNA/P, HBV DNA peak; ALT/P, peak of ALT during exacerbation; AP, after ALT peak.

Case II-1            
No. of variation0NA271212226827271820
 No. of variations in C gene (% total)0 (0)NA13 (48)8 (67)8 (67)11 (50)19 (28)4 (15)4 (15)7 (39)9 (45)
  No. of derived amino acid changes0NA8558154445
 No. of variations in S gene (% total)0 (0)NA4 (15)1 (8)1 (8)2 (9)8 (12)19 (70)19 (70)2 (11)3 (15)
  No. of derived amino acid changes0NA21127151422
Case II-2            
No. of variation0161313121632    
 No. of variations in C gene (% total)08 (50)6 (46)8 (62)8 (67)8 (50)8 (25)    
  No. of derived amino acid changes0868888    
 No. of variations in S gene (% total)02 (12)1 (7)1 (7)1 (7)2 (13)1 (3)    
  No. of derived amino acid changes0211120    

HBV Strains Surged During AE Might Exist in the Viral Pools at Baseline.

The relationships of the viral strains obtained from different time-points are shown in Figures 2B and 2C. We found that the viral strains responsible for the viral surge in each AE were not clustered in the same lineage. In addition, the viral strain surged in each AE was not directly originated from the one emerged after preceding AE. Instead, these surged viruses seemed to be derived from the one unrelated to the preceding AE. This observation indicated that the viral strain surged in each AE might have been present as a minor population within the baseline viral pools. To test this theory, we designed strain-specific PCR assays by using primer pairs specially matching the strain surged during the subsequent AE. For example, in case II-1, the viral strain surged during 3rd AE was 27 nucleotides different from the one at baseline of the first AE. To clarify whether the viral strain was already present within the baseline viral pools, we designed a primer pair (forward: nt 2640-2662, 5′-CCTGCCAGGTTTTATCCCAATG-3′; reverse: nt 3017-2996, 5′-TTGTTGGCGTCCGGCCAGTTGT-3′) specially matching the strain surged during 3rd AE. After PCR and direct sequencing, we found that this viral strain was indeed detectable at the baseline serum viral pools. Similar results can be obtained with regard to the viral strains surged during other AEs in both patients (data not shown).

Evolution of Viral Genome in Hepatitis B e Antigen Carriers With Persistently Normal or Mildly Elevated Serum ALT

The serum HBV DNA level fluctuated slightly but always exceeded 100 pg+mL in these 2 carriers. The 2 full-length viral genomes obtained at 1 year apart were identical in both hepatitis B e antigen–positive HBV carriers. Neither insertion nor deletion was observed in each viral genomic pair.

Discussion

In this study, we compared the serum viral genome to the intrahepatic genome. We found the serum genome was identical to the dominant intrahepatic one. This finding implied the viral variant that emerged after AE actually originated from the liver. At the same time, by investigating the sequential full-length viral nucleotide sequences in patients with repeated episodes of AE, we clearly demonstrated that the viral variant, if emerged after AE, existed transiently and did not become the precursor of the viral strain in subsequent AE. On the contrary, the viral strain responsible for each AE could be detected as a minor population at baseline serum viral pools. In contrast, the serum dominant viral genome remained unchanged during one-year follow-up in 2 hepatitis B e antigen–positive chronic hepatitis B patients with persistently mildly elevated serum ALT. Complementary to our previous findings,7 we suggest that reactivation of the original HBV pool contributes to the onset of repeated AE.

The true pathogenic HBV strains involved in the pathogenesis of chronic hepatitis B were the ones within the liver. Therefore, to clarify whether the serum viral genome could represent the intrahepatic one, we compared the viral genomes of both tissue specimens obtained simultaneously from patients with hepatitis B exacerbation. Our results revealed that the circulating viral genome was identical to the intrahepatic one in terms of both dominant viral nucleotide sequence and genetic complexity of the viral quasispecies, whether the specimens were collected at early or late phase of AE, and whether the specimens were collected at high or low viremic phase. We thus suggest that the serum viral genome adequately represents and reflects the intrahepatic one. From another perspective, since the serum viral genome obtained even at the late phase of AE was also identical to the intrahepatic one, we therefore speculate that the viral variant that emerged during AE most likely originated from the liver. The results were compatible with the rapid turnover of HBV in the circulation, more than 1011 virions/day,14 that are most likely produced by and released from the liver.

To clarify the representation of the serum viral genome to cccDNA HBV, we preliminarily compared the subgenomic nucleotide sequences and genetic complexity among the circulating viral genome, the intrahepatic one, and the cccDNA in 1 patient with hepatitis B exacerbation. Our finding again suggested that the circulating viral genome could represent the dominant one at the cccDNA level. However, viral genome and replication has also been documented in extrahepatic tissues, such as peripheral blood mononuclear cells.25 The contribution of such viral strains to the serum viral pools and their identity to the intrahepatic and circulating ones should be clarified in the future. Finally, different clones of cccDNA could be identified in the liver. The dynamics and evolution of cccDNA and their correlation to the development of AE should be addressed in the future.

Previous studies suggested that the upsurge of viral load before AE was less likely due to the development of new viral variants and more likely due to reactivation of a minor one from the original preexisting HBV pool (or quasispecies).6, 7 In this prospective, long-term, follow-up study, we further demonstrated that the viral variant, if emerged after AE, appeared transiently and was not related to the viral surge in subsequent AE. This finding strongly implied that the break of tolerance between virus and host leading to the onset of exacerbation was more likely related to the change of host immune surveillance over continuously evolving HBV pools.

In this study, a new viral strain was selected and reactivated before each AE. Viral variants emerging after each AE also differed. The evolution of HBV during chronic hepatitis B has been studied before.9–11, 27 The findings of those studies suggested a model of a sequential evolution from the preexisting dominant strain (Fig. 3A). In this model, the dominant HBVs responsible for each AE in the same individual are closely related and belong to the same clade. Nevertheless, in this study we suggested another model in which a horizontal selection and reactivation of different minor strains in the preexisting viral pool was the cause for the individual AE (Fig. 3B). However, during the natural course of chronic hepatitis B, it is likely that both models will operate together, depending upon the wax and waning of various selection pressures.11 A very important issue is to elucidate the characteristics of these selection factors.

Figure 3.

Two models for the origin and evolution of hepatitis B virus (HBV) quasispecies during repeated acute exacerbations (AEs) of chronic hepatitis B. (A) Sequential evolution model: HBV responsible for current AE (such as V1-2) is an offspring from an earlier one (V1)—that means a direct and sequential evolution relationship. (B) Horizontal selection model: Horizontal selection and reactivation of a minor preexisting viral strain (such as V1 and V3) from baseline viral pools (V1 + V2 + V3 + … + Vn) were responsible for the surge of viral load in each hepatitis B exacerbation. Viral variants (such as V2 and V4), if emerged after AE, existed only transiently and did not become the origin of the viral surge in subsequent AE. HBV, hepatitis B virus; V, HBV.

In conclusion, in single AE, we have already shown that variation of HBV genome rarely occurred before the onset of AE. In this long-term prospective follow-up study, we further demonstrated the insignificant impact of viral genomic variations on the development of AE. We thus suggest that future studies should focus on the host immune surveillance dysfunction leading to reactivation of preexisting virus and resultant hepatitis B exacerbation.

Ancillary