Virion half-life in chronic hepatitis B infection is strongly correlated with levels of viremia


  • Potential conflict of interest: Nothing to report.


Analysis of hepatitis B virus (HBV) kinetics with mathematical models may disclose new aspects of HBV infection and host response mechanisms. To determine the kinetics of virion decay from the blood of patients in different phases of chronic infection, we applied mathematical modeling to real-time polymerase chain reaction assays, which enable quantification of viremia and intrahepatic HBV productivity by measuring both copy number and activity of covalently closed circular DNA (relaxed circular DNA/covalently closed circular DNA) in the liver of 80 untreated chronically active HBV carriers (38 hepatitis B e antigen [HBeAg]-positive and 42 HBeAg-negative individuals). We found that the half-life of circulating virions is very fast (median 46 and 2.5 minutes in HBeAg-positive and HBeAg-negative individuals, respectively) and strongly related to viremia, with clearance rates significantly accelerating as viral loads decrease. To investigate whether immune components can influence the kinetics of virion decay, we analyzed viral dynamics in immunodeficient urokinase-type plasminogen activator chimera mice. Virion half-life in mice (range, 44 minutes to >4 hours) was comparable to estimates determined in high viremic carriers, implying that clearance rates in these patients are mostly determined by common nonspecific mechanisms. Notably, the lack of correlation between virion half-life and viremia in mice indicated that immune components significantly accelerate virion clearance rates in individuals with low titers. Conclusion: Our analyses suggest that both host defense mechanisms and levels of circulating virions affect the kinetics of HBV decay assessed in the serum of chronic carriers. Identification of the factors affecting clearance rates will be important for future antiviral drug developments and it may give insights into the mechanisms involved in clearance of other chronic infections, such as human immunodeficiency virus and hepatitis C virus. (HEPATOLOGY 2008.)

Chronic infection with hepatitis B virus (HBV) can cause various degrees of liver damage and is strongly associated with the development of liver cirrhosis and hepatocellular carcinoma. Furthermore, high levels of HBV replicative activity, in the absence of transaminase elevation, have been shown to dramatically increase the incidence of hepatocellular carcinoma development.1, 2 It is known that viral loads change with time and in the natural course of chronic HBV infection.3 Fluctuations in viremia are thought to depend on diverse counteracting virological and immune-mediated mechanisms. Yet the factors affecting rates of viral production and clearance from the blood stream remain to be elucidated. Analysis of viral dynamics may shed light on these issues.

The method for determining viral replication rates is usually to interfere with the production process by applying inhibitory drugs, analyze the rates of viral decay in blood, and from this infer the clearance rates of virus and infected cells by the immune system. By performing such calculations the free virus half-life of human immunodeficiency virus (HIV) was first estimated at 6 hours.4 However, later analysis of HIV through apheresis revealed that HIV half-life could be as low as 28 minutes.5 The virion half-life of hepatitis C virus (HCV) is within this range, at approximately 3 hours.5, 6 In the case of HBV, analysis made on individuals under antiviral therapy revealed a surprising heterogeneity in the kinetics of viral decay.7 Recent modeling of HBV kinetics showed that clearance rates of circulating virions can be determined by using the relationship between the intracellular HBV DNA pool in the liver and extracellular virions in plasma.8, 9 Using this approach, the half-life of HBV in serum of acutely infected chimpanzees was estimated at approximately 4 hours.8 Up to now, a similar assessment of HBV dynamics in chronically infected individuals was not available due to the lack of intrahepatic HBV DNA data.

By applying mathematical modeling to sensitive molecular assays that enable accurate determination of both viremia and intrahepatic viral productivity, we determined clearance kinetics of HBV virions from the bloodstream of 80 treatment-naive HBV active carriers in the various phases of chronic infection. Furthermore, we performed analysis of viral dynamics in immunodeficient urokinase-type plasminogen activator (uPA) chimera mice repopulated with primary tupaia hepatocytes (PTH) and infected with woolly monkey HBV (WMHBV) to investigate whether immune components and viremia levels can influence the kinetics of virion decay in vivo.


ALT, alanine aminotransferase; anti-HBe, anti-hepatitis B e antibody; cccDNA, covalently closed circular DNA; CK-18, cytokeratin 18; HBV, hepatitis B virus; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen; HCV, hepatitis C virus; HDV, hepatitis D virus; HIV, human immunodeficiency virus; NK, natural killer; PTH, primary tupaia hepatocytes; rcDNA, relaxed circular DNA; SVP, subviral particle; uPA, urokinase-type plasminogen activator; WMHBV, woolly monkey hepatitis B virus.

Patients and Methods

Patient Characteristics.

The study included 80 treatment-naive adults with chronic HBV infection that were seen in the outpatient clinic at the University of Hamburg, Germany. Patients underwent needle liver biopsy to determine grading and staging of liver disease. A small piece of the same needle liver specimens was stored at −80°C until analysis was performed. Informed consent was obtained from all patients and the protocol for the study was performed according to the principles of the Declaration of Helsinki and approved by the Ethical Committee of the city and state of Hamburg. A total of 38 individuals (30 men) were hepatitis B e antigen (HBeAg)-positive and 42 (29 men) were HBeAg-negative and anti-HBe positive. All patients were hepatitis B surface antigen (HBsAg)-positive and were negative for serological markers of HCV, HIV, and hepatitis D virus (HDV) infection. Patient age ranged from 19 to 64 years (median, 38 years). Only carriers with clearly detectable viral titers (>200 HBV-DNA copies/mL), HBsAg levels (>0.5 μg/mL) in serum and intrahepatic viral productivity (total HBV-DNA/cell > covalently closed circular DNA [cccDNA]/cell) were included in this study.

Biochemical and Serological Analyses.

Serum HBsAg levels were measured using the quantitative assay of electroimmunodiffusion (Laurell method).10, 11 Serum HBV DNA was quantified using real-time fluorescent-probe polymerase chain reaction (PCR) (TaqMan; Roche, Mannheim, Germany), with a lower limit of quantification of 200 HBV DNA genomes/mL.11 HBV DNA genotyping was performed using the Inno-Lipa assay (Innogenetics, Belgium).

Extraction and Quantification of Cellular and Viral DNA from Needle Liver Biopsies.

DNA was extracted from human liver needle biopsies and mouse liver specimens using the MasterPure DNA purification kit (Epicentre, Biozym, Germany) and amplification was performed in the Light Cycler System (Roche Diagnostics, Mannheim, Germany) using HBV-specific primers and fluorescence hybridization probes as described.11–13 Intrahepatic HBV-DNA values (lower limit of quantification ≥0.002 HBV-DNA copies/cell) were normalized for cellular DNA content using the beta-globin gene kit (Roche DNA Control Kit; Roche Diagnostics), while cccDNA copies were determined after treatment with plasmid-safe deoxyribonuclease.13 The median cell number used to determine viral loads in liver biopsy samples was 7,075 genome equivalents) (range, 3,500–14,000 genome equivalents). To estimate intrahepatic amounts of replicating virus (relaxed circular DNA [rcDNA]), total intracellular HBV-DNA values were modified to exclude cccDNA.13

Determination of Liver Repopulation and Viral Loads in Mice.

In uPA transgenic mice, expression of the uPA transgene is driven by the albumin promoter and induces death of the transgene-carrying hepatocytes, thus providing a growth advantage for transplanted hepatocytes.14 To perform xenotransplantation experiments, uPA mice (Jackson Laboratories, ME) were crossed with RAG2//Pfp double knockout mice (KO) (Taconic Farms, Denmark), which lack mature T-cell, B-cell, and natural killer (NK)-cell activity. Animals were maintained under specific pathogen-free conditions in accordance with institutional guidelines under approved protocols. Mouse genotypes were determined by PCR as reported.15 The Asian tree shrew (Tupaia belangeri) is phylogenetically related to primates, and isolated hepatocytes (PTH) can be efficiently infected with WMHBV. Because of the high levels of repopulation (up to 90%) achievable with PTH, we determined virion half-life in PTH-chimera mice chronically infected with WMHBV.16 Tupaias were obtained from the German Primate Centre in Göttingen, Germany, and maintained in the animal facility of the University of Freiburg in accordance with institutionally-approved protocols. For PTH preparation, animals were anesthetized by intramuscular injection of ketamine (5 mg/100 g body weight) and xylazine (1 mg/100 g body weight) and the liver was perfused in situ by collagenase perfusion.16 A total of 1 × 106 viable freshly isolated or cryopreserved PTH were injected intrasplenically into 15-25-day-old uPA+/−/RAG2/Pfp KO mice anesthetized with isoflurane, according to protocols approved by the Animal Care Committee of Hamburg.

PTH contents in mouse livers were determined by real-time PCR using tupaia-specific β-actin primers: F: aac gag atg aga ttg gca; R: caa tcc aag tcc tcg gc and SYBR-green (Roche Diagnostics, Mannheim, Germany). Serial dilutions of PTH genome equivalents extracted from defined amounts of isolated PTH were used as the standard. Viral loads were determined both in serum and in the liver of infected mice by real-time PCR with WMHBV-specific primers.16 A total of 30 ng of genomic DNA extracted from each chimera mouse liver was used to determine PTH amounts and intrahepatic viral loads. Total WMHBV DNA and cccDNA copy number were normalized for PTH contents (median, 1942; range, 413–2,450 PTH) present in each given mouse liver specimen analyzed.


Levels of PTH repopulation in mouse livers were confirmed by immunohistochemistry using cytokeratin 18 (CK-18) monoclonal antibody (1:200), which specifically recognizes tupaia and human CK-18 but not mouse CK-18, or with a rabbit hepatitis B core antigen antiserum (1:200), (both from DAKO Diagnostika, Hamburg, Germany). Specific signals were detected using Envision anti-mouse or anti-rabbit horseradish peroxidase (Dako, Hamburg, Germany) immunofluorescence staining (Invitrogen).


The Wilcoxon rank sum test was used for nonparametric comparisons. The Spearman rank correlation was used to calculate r2 and P values for correlations, with values and slopes of regression lines determined by least-squares regression. The chi-square test was used for comparison of categorical variables. P values <0.05 were considered significant. The mathematical model used to determine free-virion half-life during HBV chronic infection was previously reported.8


Determination of Intrahepatic Virion Productivity and Circulating Viral Loads.

We calculated steady-state levels of viral loads both in needle liver biopsies and corresponding sera from 80 treatment-naive HBV chronic carriers (38 HBeAg-positive and 42 HBeAg-negative individuals). As shown in Table 1, these individuals displayed an extensive viremia range (from 300 to 4 × 109 HBV-DNA copies/mL) with confirmed intrahepatic viral productivity (from 0.09 to 2 × 103 rcDNA/cell), since levels of intrahepatic total HBV-DNA were always greater than the cccDNA copy number.

Table 1. Comparison of Patients With or Without Detectable HBeAg
  1. Abbreviations: HBcAg, hepatitis B core antigen; n.s., not significant; ULN, upper limit of normal.

Median age (years)32 (19–64)39 (20–60)0.003
 HBV-DNA (copies/mL,median)4.5 × 107 (103–4 × 109)6.5 × 104 (300 × 4 × 108)<0.0001
 ALT (ULN, median)2.1 (0.4–15)1.77 (0.3–6.7)n.s.
 HBsAg (μg/mL, median)42 (2–324)13.5 (2.4–62)<0.0001
 GenotypeD = 21; A = 6; A + D = 3; B = 3; C = 2; E = 2; G = 1D = 22; A = 10; A + D = 9; E = 1n.s.
 rcDNA (copies/cell, median)98 (0.6–2 × 103)1.2 (0.09–104)<0.0001
 cccDNA (copies/cell, median)1.8 (0.03–54)0.12 (0.001–15)<0.0001
 Median grading (Desmet score)1.75 (1–3)2 (1–4)0.048
 Median staging (Desmet score)1 (0–4)1 (0–4)n.s.
 Median HBsAg staining (% of cells)30 (90–0)5 (90–0)0.001
 Median HBcAg staining (% of cells)20 (80–0)0 (15–0)<0.0001

In agreement with previous analysis,13 we found that variations in term of intrahepatic viral loads between treatment-naive HBeAg-positive and HBeAg-negative individuals were due to both different cccDNA content (Δ1 − log) and significantly lower (P < 0.0001) virion productivity (rcDNA/cell; Δ2− log) in most HBeAg-negative patients. Nevertheless, a greater difference in terms of viral loads was still observed at the serological level. By assuming 3,000 mL as the average amount of human plasma, median steady-state levels of HBV-DNA in serum were 1.4 × 1011 versus 2 × 108 (P < 0.0001; Fig. 1A) in HBeAg-positive and HBeAg-negative individuals, respectively.

Figure 1.

Loads of circulating and intrahepatic virions differ significantly between HBeAg-positive and HBeAg-negative individuals. (A) Total HBV-DNA levels calculated for 3 L of serum. Each dot represents a single patient measurement: 38 HBeAg-positive individuals (filled dots) and 42 HBeAg-negative individuals (open dots). HBV-DNA copies were determined by real-time PCR using HBV-specific primers and fluorescence resonance energy transfer (FRET) probes (lower limit of quantification [LLoD] ≥200 HBV-DNA copies/mL). (B) The daily production of HBV-DNA from the liver of each patient was estimated by scaling steady-state amounts of rcDNA copies determined per cell (LLoD ≥ 0.002 HBV-DNA copies/cell) by the 1 × 1011 cells present in a typical human liver.

The steady-state levels of total HBV DNA in serum V are determined by the rate of clearance c (day−1) of free virions from plasma and export of new virions at rate β (day−1) from the intrahepatic pool of HBV DNA-containing capsids D. Within nucleocapsids, the pregenomic RNA transcript is first converted to single-stranded DNA and then to double-stranded HBV-rcDNA,17 which is finally exported at a daily rate β to produce free virions in plasma. Since we could not determine the rate of maturation from single-stranded DNA to rcDNA within infected hepatocytes, in agreement with previous studies we assumed that rcDNA forms approximately 50% of total intracellular HBV-DNA containing capsids detected in our assay.8 Thus we considered rcDNA contents as D × 0.5, where the value of D was obtained by multiplying the amount of rcDNA per cell by the 1 × 1011 cells present in the liver, and the daily production rate of new virions from the liver was given by 0.5βD. The value of the export rate, β = equation image was chosen consistent with an experimentally estimated export half-life of newly synthesized virions of 1 day8 and was used to convert the number of rcDNA per cell to the expected number of rcDNA that leave the cell daily to replenish virion levels in plasma.

To determine HBV virion half-life in treatment-naive patients, we assumed that newly produced virions (from the pool of rcDNA) are exported at the same rate β, regardless of the levels of infection and HBeAg status. The good correlation found between intrahepatic amounts of progeny virions (rcDNA/cell) and serum HBV-DNA titers in all patients analyzed here (r = 0.74; P < 0.0001), supports the hypothesis that export rates do not differ significantly between HBeAg-positive and HBeAg-negative individuals.13

Under these assumptions, median virion production per day was 5 × 1012 virions for HBeAg-positive individuals, while it was 83-fold lower for HBeAg-negative patients (median 6 × 1010 virions, P < 0.0001; Fig. 1B). As a consequence, median viral loads in the plasma of HBeAg-positive individuals were 316-fold less than production in the liver, whereas in HBeAg-negative individuals median viral loads in plasma were 1,720-fold lower than production (P < 0.0001). This implies a faster clearance of virions in individuals with low viral titers.

Estimation of HBV half-life in serum.

During the chronic phase of HBV infection, the processes of virion export and clearance will be approximately in balance so that 0.5βD = cV, and therefore c=β(0.5D/V). Free-virion half-life in plasma is then calculated as t1/2 = equation image = equation image and with β = equation image we find t1/2 = equation image.

Applying these calculations to these treatment naive patients uncovers a strong correlation between kinetics of viral decay and viral load, which is highly significant (P < 0.0001; r = 0.82; Fig. 2). The median virion half-life for HBeAg-positive individuals is 46 minutes (interquartile range, 4–224 minutes), whereas it is much shorter (median, 150 seconds) for the HBeAg-negative individuals (interquartile range, 24 seconds to 13 minutes). The slope of the regression line for log half-life compared to log viral load is 0.59. This implies the relationship t1/2 = A × V0.59 ≈ A equation image, so that as viral load decreases the speed at which virions are cleared accelerates at an increasing rate.

Figure 2.

The half-life of circulating virions strongly correlates with viremia in individuals chronically infected with HBV. Correlation between log half-life (in minutes) of free virions and log viremia levels (HBV-DNA copies/mL) in serum of 80 patients chronically infected with HBV. Dots represent measurements performed in each patient (filled dots, HBeAg-positive; open dots, HBeAg-negative). Since the slope of the regression line is 0.59, clearance rates accelerate as viral load in serum decreases.

We considered the possibility that other processes may also have affected estimated virion half-lives without being related to clearance of circulating virions from serum. One possible explanation for the loss of HBV DNA in serum relative to rcDNA within cells could be that rcDNA is exported at slower rates as intracellular virion levels decrease. Although HBeAg-negative individuals harbored significantly lower rcDNA levels per cccDNA template (P = 0.017), their steady-state ratios were still greater than unity (median 13.7 versus 49 rcDNA copies/cccDNA in HBeAg-negative versus HBeAg-positive individuals, respectively), suggesting sufficient steady-state levels of rcDNA containing nucleocapsids for export.13 Hence, even though a fraction of rcDNA may be recruited to the cell's nucleus to amplify numbers of the cccDNA viral template, this amount would not be expected to produce the 6-log decrease in half-life calculations shown in Fig. 2.

Clearance of infected cells is also unlikely to contribute significantly to the different virion half-lives estimated in serum of these patients, since median transaminase values, as well as inflammation and fibrosis status did not differ significantly between HBeAg-positive and HBeAg-negative individuals (Table 1).

Determination of Virion Half-Life in the Bloodstream of Immunodeficient uPA Mice.

Repopulation of the liver of immunodeficient uPA mice with hepadnavirus-permissive primary hepatocytes, opens new opportunities to investigate mechanisms of viral infection and clearance. To start unraveling the mechanisms responsible for the different viral clearance kinetics observed in chronically infected HBV patients, virion half-lives were determined in immunodeficient uPA+/− transgenic mice repopulated with tupaia hepatocytes and infected with WMHBV.15, 16 Since these mice lack mature B cells, T cells, and NK cells, half-lives of circulating virions can be estimated without assistance from these immune components. Viral loads were determined both in serum and in the liver of chronically infected mice using real-time PCR and WMHBV-DNA–specific primers. In this model of hepadnavirus infection, viremia levels in mice (N = 7) ranged from 5 × 104 to 9.5 × 108 WMHBV-DNA copies per milliliter of serum. Intrahepatic amounts of total WMHBV-DNA and cccDNA were normalized for the number of PTH estimated in each mouse liver specimen analyzed, and intrahepatic viral loads (D) were calculated by applying the same formula used for estimation in patients and by multiplying rcDNA/PTH values by the total number of PTH estimated in each mouse liver. Considering that hepatocytes constitute approximately 60% of liver cells and that a typical mouse liver contains 100 million cells, the percentage of PTH repopulation ranged from 8% to 82% (range, 5 × 106 to 5 × 107 PTH/mouse liver). Infection and percentage of PTH repopulation were also confirmed by immunohistochemistry (Fig. 3A,B). Virion half-lives in high viremic mice (WMHBV-DNA/mL ≥108) were comparable (mean, 2.6 hours; N = 3) with estimates assessed in patients with similar viremia, suggesting that common nonspecific mechanisms determine the kinetics of virion clearance in patients with high viral loads (Fig. 3C).

Figure 3.

Correlation between virion half-life and viremia in immunodeficient infected mice. (A) Repopulation of the murine liver with PTH is shown by specific Cyt-18 staining. (B) Core antigen (HBcAg) staining of infected PTH in a liver section of a chronically-infected mouse. (C) Free virion half-life was determined in seven immunodeficient uPA mice (black squares) repopulated with PTH and chronically infected with WMHBV. Virion half-life in mice ranged from 44 minutes to more than 4 hours and viremia ranged from 4.6 × 104 to 9.5 × 108 WMHBV-DNA copies/mL. Values calculated in patients are shown as gray dots for comparison.

Contrary to the situation observed in humans chronically infected with HBV having functional immune systems, virion half-lives in low viremic mice (median, 3.3 × 105 HBV-DNA/mL serum; N = 4) were not significantly different (median t1/2 = 67 minutes) from the mice with high viral titers (P = 0.4) (Fig. 3C), although for all mice virion half-lives were correlated with WMHBV-DNA levels in plasma (r = 0.82; P = 0.03). However, the decrease in virion half-life for the mice over the range of viral levels observed in patients was only 1 log compared to the 6-log decrease for individuals chronically infected with HBV. The finding that clearance rates are not as affected by viral loads in immunodeficient mice suggests that additional host defense mechanisms contribute to clear the virus from blood in chronically infected individuals with low HBV titers.

Relationship Between Viral Titers and HBsAg Concentrations.

Diverse virus-host interactions may be responsible for the different kinetics of virion clearance observed in patients. Neutralizing antibodies directed against HBsAg play a key role in the recovery from HBV infection by facilitating the removal of viral particles from the bloodstream. Thus, the dependence of virion half-life on viral load could be due to a relatively fixed amount of neutralizing antibodies. In this scenario, for high viral loads there is much more virus than antibody and so there is very little effect on antibody-mediated viral clearance, leading to a longer virion half-life. As viral load decreases, the relative proportion of antibody to antigen increases, allowing a greater relative effect, thereby decreasing virion half-life. Intriguingly, quantitative analysis of HBsAg concentrations in serum revealed that the ratios between HBsAg and HBV-DNA genomes are not constant among patients with different viral loads, being inversely correlated with viral titers (Fig. 4). Indeed, our data show that as viral loads decrease, the relative amount of HBsAg, and hence of empty subviral particles (SVPs) per virion, increases drastically (5-log increase; P < 0.0001; r = 0.97). Thus, it is unlikely that the faster kinetics of viral clearance observed in low viremic patients can be caused by a shift in the ratio of HBsAg antibodies to virions, since such antibodies are expected to be rapidly absorbed by the overwhelming excess of SVPs.

Figure 4.

The proportion of HBsAg present per virion in the serum of individuals chronically infected with HBV increases dramatically as viral loads decrease. HBsAg concentrations were quantified by using the Laurell test. HBsAg levels are expressed in picograms per HBV-DNA genome and correlated with viremia. Full dots, HBeAg-positive; empty dots, HBeAg-negative. P and r values are indicated.


We found that clearance kinetics of HBV virions from the bloodstream of chronically infected individuals are faster than previous estimates obtained by monitoring viral decay in serum of patients after administration of drugs interfering with the pathway of viral replication.7 In this study, the half-life of circulating virions was determined by taking into account both viremia and intrahepatic virological profiles. Both copy number and activity of the cccDNA (rcDNA/cccDNA), which may vary depending on the occurrence of HBV genotype mutations and the phase of chronic hepatitis B infection,13 were measured in liver biopsy samples of 80 untreated patients. We found that the kinetics of free virion decay determined in high viremic chronic carriers are consistent with HBV half-life estimates recently performed in chimpanzees. Indeed, the 4-hour virion half-life previously determined in the early phase of acute infection, before the immune system takes control of the infection,8 corresponds to the upper-quartile of half-life values represented in Fig. 2, where viremia is similar to levels observed at acute stage disease. Such a short half-life for circulating HBV virions is not surprising considering the small size of the HBV particles and corresponds to estimates obtained for other viruses, such as HCV and HIV.5

By analyzing the dynamics of viral decline during antiviral therapy, a strong variation in HBV clearance rates was observed among chronically infected individuals, though the reasons for this discrepancy remained unclear.7 In this report, our analysis uncovers for the first time the existence of a strong correlation between virion half-life and viremia levels. Our data show that clearance rates of circulating virions significantly accelerate as viral loads in blood decrease.

Complex host-factor interplay is likely to affect HBV clearance kinetics in the various phases of chronic infection, and HBV-specific inhibition of host defenses may play an important role. At the same time, however, a broad range of innate and adaptive immune responses are expected to interact in an attempt to clear the infection. Thus, it is reasonable to suppose that HBV half-life will change in accordance with the strength of these combined effects. Though viremia levels are influenced by the rates of virion productivity, our study reveals that additional factors may contribute to the shorter virion half-life observed in plasma of most HBeAg-negative patients. For instance, inhibition of virion secretion cannot be entirely ruled out, since export rates could not be experimentally measured in patients. We reasoned that if secretion of newly synthesized viral particles was specifically impaired in some patients, accumulation of intrahepatic rcDNA levels may be expected. However, HBV replication was significantly reduced in the majority of HBeAg-negative individuals analyzed, and the good correlation found between steady-state levels of pregenomic RNA transcript and rcDNA13 and between rcDNA and viremia provides evidence that virion productivity but not export of produced virions was hindered in low-viremic HBeAg-negative patients.

To investigate whether rates of viral clearance also depend on viremia levels in animals unable to mount a specific immune response against circulating virions, we determined virion half-life in uPA transgenic mice lacking mature T-cells, B-cells, and NK-cells (RAG-2−/−/pfp−/). Immunodeficient uPA mice were transplanted with PTH and chronically infected with the hepadnavirus WMHBV.6 Though mice were infected with a HBV-closely related virus, the finding that virion clearance rates in high viremic mice were comparable with estimates assessed in patients with similar viremia suggests that common nonspecific mechanisms determine kinetics of viral decay in patients with high viral loads. However, virion half-life in mice was not significantly affected by viremia, supporting the hypothesis that additional processes contribute to clear circulating virions in naturally HBeAg seroconverted individuals with low titers.

Kupffer cells and dendritic cells play a major role in modulating the immune response. Emerging data provide evidence that several viruses have evolved mechanisms to directly suppress the host defense by altering different pathways of the innate immune response.18 Notably, chronic hepatitis B infection has been associated with a significant reduction of dendritic cell functions and impairment of the innate immune response.19 Furthermore, downregulation of Toll-like receptor signaling was recently demonstrated in HBeAg-positive chronic carriers, revealing the occurrence of potentially important interactions between virions, HBeAg, and the innate immune response.20 Under this scenario, the direct correlation found between viremia and kinetics of viral decay in blood would support the hypothesis that viral components can suppress activation of innate immune surveillance and hence maturation and expansion of effector T cells and B cells, which, in turn, may affect clearance kinetics.

Antibodies to HBsAg have virus-neutralizing activity and are synthesized early in infection, though they are not detectable because of the huge excess of circulating SVPs. Though it is not known what effect SVPs have on the course of infection, it may be speculated that they act as immune system decoys. Our quantitative measurements of HBsAg concentrations in serum of patients displaying a broad viremia range show that as viremia decreases, the relative amount of envelope antigens, and hence the number of empty SVPs per HBV-DNA–containing virion, increases dramatically. Considering that SVP production is not impaired in HBeAg-negative patients13 and that acquisition of viral DNA integrations may also contribute to HBsAg titers, antibodies to HBsAg are expected to be rapidly absorbed by the still-overwhelming excess of SVPs present in low-viremic carriers. It is therefore surprising that the relative increase in the ratio of HBsAg concentrations to virions does not extend the half-life of circulating infectious particles. To shorten the half-life of virions in low-viremic naturally-seroconverted HBeAg-negative individuals, additional mechanisms or the improvement of immune responsiveness may be involved. Since the humoral immune response includes a variety of antibodies reacting with different epitopes on the surface of viral particles, the involvement of antibodies or other immune-mediated mechanisms preferentially directed against virions cannot be excluded.21

Although the molecular mechanisms behind the kinetics of viral decline need to be experimentally identified, the shorter HBV half-life determined in low-viremic patients is important because it may strengthen the immune response under antiviral treatment. Indeed, suppression of viremia induced by lamivudine treatment has been shown to restore T-cell responsiveness in chronic HBV infection,22 and a significant inverse correlation between detection of an HBV-specific T-cell response and HBV viral load has been reported.23, 24 Consistent with this hypothesis is the observation that individuals with reduced viral loads, and probably exhibiting faster virion half-live values, are more likely to have HBeAg seroconversion, which may be followed by clearance of HBsAg and detection of neutralizing antibodies. It will therefore be interesting to investigate clearance kinetics of circulating virions in patients undergoing antiviral treatment by taking into account intrahepatic viral productivity and applying the same mathematic modeling.

Immunostaining procedures and estimation of cccDNA copy number based on molecular assays clearly show that only a minority of the liver cells are infected in the HBeAg-negative low phase of replication.13 Considering the huge amount of hepatocytes (1 × 1011) present in the human liver, which may also be less prone to infection,25 and the lower steady-state amounts of circulating virions in these HBeAg-negative individuals, productive infection of hepatocytes may be a relatively rare event. It is therefore conceivable that an accelerated clearance of virions from the bloodstream may hinder de novo infection of the hepatocytes. Therefore, better understanding of the interactions between viral elements and components of the immune system may provide new clues about the mechanisms determining clearance of HBV infection, and hence, can be helpful in designing new strategies aimed at eradicating the virus in individuals chronically infected with HBV. Finally, it will be important to investigate whether similar relationships between viral loads and clearance kinetics can be induced in other chronic viral infections such as HCV and HIV.


We thank Fritz von Weizsäcker for providing the tupaia hepatocytes, Alexander Quaas for histological analysis of biopsies, and Roswitha Reusch for her continuous excellent assistance with animal care.