Potential conflict of interest: Nothing to report.
The virion half-life of hepatitis B virus (HBV) is currently estimated at approximately 1 day. This estimate has been obtained from drug perturbation experiments with reverse transcriptase inhibitors. However, the analyses of those experiments have not considered the export of virions produced from preformed mature DNA-containing HBV capsids in infected cells. Data from 3 acutely infected chimpanzees indicates that there is approximately 10-fold more total intracellular HBV DNA than HBV DNA in blood, and therefore the half-life of virions for chimpanzees during acute infection is 10-fold shorter at 3.8 hours than the half-life associated with export of total intracellular HBV DNA. Mathematical model simulations duplicating the viral dynamics observed in drug perturbation experiments suggest a half-life of at most 4.4 hours for HBV virions in chronically infected humans, significantly shorter than current estimates, but consistent with the half-lives of virions for hepatitis C virus and HIV. This faster turnover of HBV in blood indicates a correspondingly higher replication rate and risk of mutation against hepatitis B antiviral therapy. In conclusion, we find the half-life of HBV virions is approximately 4 hours, significantly shorter than current estimates of 1 day. This new value is consistent with virion half-life estimates for HIV and hepatitis C virus. (HEPATOLOGY 2006;44:1117–1121.)
The infectious particle of hepatitis B virus (HBV) has a diameter of 42 nm, within the size range of 30 to 60 nm of virions for hepatitis C virus (HCV),1 and smaller than the 145 nm size of virions for human immunodeficiency virus (HIV).2 Each of these viruses has been the subject of dynamical analyses to determine the half-life of virions in infected individuals. It seems anomalous that the half-life of HBV virions is estimated to be approximately 1 day,3–6 whereas for HCV and HIV it is considerably shorter.
HCV virions have a half-life of less than 3 hours.7, 8 The half-life for HIV virions was 6 hours at its first estimate and now is considered to be approximately 30 minutes.8, 9 These estimates have been obtained through apheresis and drug perturbation experiments. The latter methodology involves administering therapy to inhibit new virion production and analyzing the decay rates of virus in blood.
The drugs used in HBV calculations were reverse transcriptase inhibitors such as lamivudine and adefovir dipivoxil, which provide a partial block in the maturation of capsids containing HBV RNA to those containing intracellular HBV DNA. Lamivudine inhibits elongation of the viral minus strand DNA through competitive inhibition of the natural substrate dCTP with the chain terminator 3TC-TP, and also inhibits polymerase activity for second-strand DNA synthesis.10 Adefovir interferes with the priming of reverse transcription as well as elongation of the viral minus strand DNA.11 The rate of inhibition of these drugs is concentration dependent,10 and as with HIV, the ability to block full maturation of capsids is dependent on the length of the remaining second-strand DNA piece required for production of mature capsids competent for envelopment and secretion.12 This explains why de novo production of HBV virions under lamivudine and adefovir is almost completely suppressed whereas prevention of initial cccDNA formation through repair of the 200-base gap is limited.13
However, these drugs do not inhibit export of virions produced from preformed mature intracellular HBV DNA-containing capsids competent for envelopment. Therefore, the dynamics of virus in blood after the introduction of therapy involves 2 processes: the export of newly enveloped intracellular HBV DNA-containing capsids to replenish the virion pool and the clearance of virus from blood. This first process was not considered in previous calculations, leading to the possibility that the viral decay rate in blood may not represent slow clearance, but rather a slow rate of export with a faster clearance rate hidden beneath this.
The difficulty in incorporating information on the intracellular HBV DNA pool in HBV half-life calculations lies in the requirement of liver biopsies at regular time points of the experiment. Whereas virion estimates in peripheral blood are easily obtained, repeated liver biopsies pose a significant health risk. However data are available on the relationship between intracellular HBV DNA and extracellular virions from analysis of acute HBV infection in chimpanzees.14 Mathematical modeling in those experiments predicted that the HBV virion half-life was significantly shorter than current estimates. Although HBV infection in chimpanzees may differ from that of humans, they remain the only animal model for this viral infection, as the closest relative to humans.
A shorter half-life for HBV has significant implications for the risk of chronically infected individuals developing resistance to antiviral therapy. A shorter half-life indicates higher viral replication and therefore increased risk of mutation.
We formulated a method for estimating HBV virion half-life from the relationship between intracellular HBV DNA and extracellular virions in plasma. Data from 3 acutely infected chimpanzees provided values of both infection components over reasonable time periods, and corresponding half-life estimates. Model simulations based on estimates from these data duplicate the viral dynamics in humans observed in drug perturbation experiments and predict a shorter HBV virion half-life in humans than previously estimated.
HBV, hepatitis B virus; HCV, hepatitis C virus; HIV, human immunodeficiency virus; ssDNA, single-stranded DNA; dsDNA, double-stranded DNA.
Materials and Methods
Animals were handled according to humane use and care guidelines specified by the Animal Research Committees at the National Institutes of Health and the Scripps Research Institute. All animals were housed at Bioqual Laboratories (Rockville, MD), an institution accredited by the American Association for Accreditation of Laboratory Animal Care International and under contract to the National Institute of Allergy and Infectious Diseases.
Three healthy, young adult, HBV-seronegative chimpanzees (Ch1615, Ch1620, and Ch1627) were inoculated with 108 genome equivalents (GE) of a monoclonal HBV (genotype ayw) contained in pooled serum from HBV transgenic mice.15
HBV DNA Detection in Serum.
Total DNA was isolated from 800 μL serum aliquots by proteinase K digestion and precipitation of nucleic acids after organic extraction. Briefly, 200 μL 5× DNA lysis buffer (200 mmol/L Tris-HCl, pH = 8; 100 mmol/L EDTA; 5% SDS) and 50 μL proteinase K (20 mg/mL) were added to 800 μL plasma and incubated overnight at 37°C. Samples were then subjected to phenol, phenol/CHCl3, and CHCl3 extraction; nucleic acids were pelleted by centrifugation after precipitation with an equal volume of isopropanol, and the pelleted nucleic acids were resuspended in 50 μL TE (10 mmol/L Tris-HCl, pH = 7.5; 1 mmol/L EDTA). The HBV DNA content in 5 μL of each sample was quantified by real-time PCR using a Bio-Rad iCycler system exactly as described.16
HBV DNA Detection in Liver.
Total liver DNA was extracted from frozen liver biopsy samples as described,16, 17 and the levels of HBV DNA replicative intermediates were determined by quantitative real-time PCR using a Bio-Rad iCycler system exactly as described.16
Model of Chronic HBV Infection.
We consider a simplified version of the full model of HBV infection described in Murray et al.14 Here, we consider chronic infection where the number of infected cells has reached a steady state. Three separate components of infection are considered: I, the number of infected hepatocytes; D, the number of intracellular HBV DNA-containing capsids; and V, the number of virions in plasma. Virion production and clearance is described through the following mathematical model:
where H is the number of uninfected hepatocytes, k is the rate of infection, a represents the rate of production of intracellular HBV DNA-containing capsids per infected hepatocyte, β is the rate of export of these capsids to blood, c is the rate of clearance of virions in plasma, and δ is the rate of clearance of infected hepatocytes.
For conversion to numbers of virions per milliliter in humans, we assumed there were 6 × 1010 hepatocytes in a human liver, based on Mackay et al.,18 and that virions were distributed through 3 L of serum. For chimpanzees, the corresponding values, adjusted for weight, were 4 × 1010 hepatocytes and 2 L of serum. The number of uninfected hepatocytes H was determined by subtracting the number of infected hepatocytes I from the total.
Numerical simulations of the model in Equation 1 for Fig. 2 were performed with parameter values k = 1.67 × 10−12, δ = 0.053, a = 150, β = 0.87, c = 3.8, and initial values for each component of I0 = 6 × 1010, D0 = 9.8 × 1012, and V0 = 2.2 × 1012. Calculations were performed within Matlab version 6.5 (The MathWorks Inc., Natick, Mass.).
Mathematical modeling of the dynamics of acute HBV infection in 3 chimpanzees provided maximum likelihood estimates of a number of parameters representing rates of infection and clearance.14 Among these were the export rate β (0.46 day−1), and virion clearance rate c (mean 6.7 day−1). These provide half-life estimates of time to export of 1.5 days and for time to clearance of free virions of 2.5 hours. These maximum likelihood estimates were obtained from fitting several sets of data simultaneously with a model that contained many of the processes of infection and clearance. However, the relative values of the export rate and virion clearance rate β/c can be estimated through a much easier calculation, based on a simpler model in Equation 1. This model only includes variables representing the number of infected hepatocytes I, the number of intracellular HBV DNA-containing capsids D, and the number of virions in plasma V.
Relative Sizes of Intracellular and Extracellular HBV DNA Compartments in Chimpanzees.
In the steady state, export of new virus from infected cells will match clearance of free virus in blood so that the last equation in (Equation 1) will be 0. Hence βD = cV and we obtain
This says that the relative quantity of HBV DNA in blood to the quantity of HBV DNA within infected cells gives the ratio of export to clearance. Therefore, if export is slower than clearance β < c, there will be less HBV DNA in blood than within infected cells: V < D
Ratios of extracellular to intracellular HBV DNA V/D, for 3 acutely HBV infected chimpanzees, are shown in Fig. 1. These ratios are significantly less than 1 for each animal (P < .0001, Student t test). Calculating mean log10V/D over all chimpanzees gives an estimated V/D ratio of 0.105. Using this value determines that β = 0.105 × c through Equation 2, and therefore clearance of virions in plasma is approximately 10 times faster than export of HBV DNA from infected cells. Given that the maximum likelihood estimate of HBV DNA export in these chimpanzees was 0.46,14 this implies c = 0.46/0.105 = 4.4, determining a half-life of HBV in blood for acutely infected chimpanzees of 3.8 hours. The discrepancy between this estimate for c and the previous value of 6.7,14 can be attributed mostly to that latter methodology simultaneously fitting 5 different viral components whereas this method need only be consistent with the ratio of 2 of these components.
Intracellular HBV DNA quantification in Murray et al.14 did not distinguish between capsids containing single-stranded DNA (ssDNA) and double-stranded DNA (dsDNA). At peak infection in the 3 chimpanzees, the ratio of dsDNA to ssDNA was 0.92 for Ch1615, 0.69 for Ch1620, and 0.94 for Ch1627, giving a mean of 0.85. Hence, at peak infection the component of mature dsDNA-containing capsids that are ready to be exported as new virions, forms approximately 46% [0.85/(1 + 0.85) × 100%] of the total HBV DNA. Subsequent to peak infection, noncytolytic immune mechanisms are thought to prevent the formation of pregenomic RNA-containing capsids while maturation of preformed capsids and virus secretion continue.14, 19, 20 Therefore as ssDNA capsids mature and are not replaced, the dsDNA content of total intracellular HBV DNA will increase from 46% to approximately 100% before returning to 46% as the imperfect block in new ssDNA formation reaches a new steady state.
The Kinetics of First Phase HBV DNA Decay After Commencement of Therapy Is Consistent With the Export Rate of New Virions From Infected Cells.
The export rate of 0.46 day−1 of HBV DNA-containing capsids from acutely infected chimpanzees14 was obtained using intracellular measurements of total HBV DNA. Adjusting this value to be representative of the export rate only of dsDNA-containing capsids, which can vary between 46% and roughly 100% of the total HBV DNA content for these chimpanzees, implies a value between 0.46 and 1 day−1 in acutely infected chimpanzees. Although viral dynamics of acute infection in chimpanzees might differ from chronic infection in humans, this export rate range is consistent with the first phase loss of HBV DNA in plasma after commencement of therapy in humans, at a rate of 0.63 day−1 (mean half-life of 1.1 days).5
Ratios of extracellular to intracellular HBV DNA calculated for chimpanzees suggest that the clearance rate of free virions should be between 4.4-fold [given that dsDNA may form only 46% of total intracellular HBV DNA and total HBV DNA is 9.5-fold (1/0.105) higher than extracellular HBV DNA] and 9.5-fold faster than the export rate of dsDNA-containing capsids. On the basis of the chimpanzee calculations, the lowest value for the clearance rate c of free virions should be approximately 2 day−1 (0.46/0.105 × 46/100), much higher than the first phase loss of HBV DNA in humans. Hence, the rate of loss of HBV DNA during the first phase after commencement of therapy in HBV-infected humans is representative of the export rate of new virions from infected cells and not of the clearance rate of free virions.
Simulation of HBV DNA Dynamics in Chronically Infected Humans Commencing Antiviral Therapy With a Reverse Transcriptase Inhibitor.
Treatment of chronically HBV-infected humans by a reverse transcriptase inhibitor produces a biphasic decay in HBV DNA per milliliter of serum.3–6 In one of these experiments, viral load decreased from a baseline value of 4.7 × 108 HBV DNA copies per milliliter with first and second phase half-lives of 1.1 and 18.2 days, respectively, where the extent of the first phase decline was determined by the efficacy of the drug estimated at 0.993.5 This mean behavior is represented by the data points in Fig. 2C. In these and other experiments,3–6 the first phase was interpreted as representing the half-life of virus in blood, with the second phase determined by the loss of infected hepatocytes. The previous analysis suggests, however, that this first phase loss represents export of the preformed intracellular HBV DNA pool as mature virions, where therapy has inhibited replacement of this intracellular compartment. We estimate the clearance rate of free virions c through the relationship between the export rate, as measured by the first phase decay, and the relative quantity of intracellular mature (double stranded) HBV DNA compared to free virions, a value of at least 4.4. This is determined through c ≥ 4.4β. Numerical simulations of our mathematical model in Equation 1 show that the data are consistent with a virion half-life of at most log(2)/c = log(2)/3.8 = 4.4 hours, where the first phase represents export of preformed HBV DNA-containing capsids from infected hepatocytes (Fig. 2B). The second phase in the simulations represents loss of infected hepatocytes (Fig. 2A), in agreement with literature observations.
The virion production pathway for HBV has complicated calculations for the half-life of free virus for this disease. Previous calculations have not considered the store of HBV DNA-containing capsids within infected cells, so that the slow clearance of virus after initiation of drug therapy has been attributed to a long half-life of free virus, rather than slow export of virus from infected hepatocytes. Here, we show using data from acutely HBV-infected chimpanzees, that the slow half-life of approximately 1 day attributed to clearance of virus,3–6 is instead the half-life for export of new virus from infected cells. The export rate of HBV DNA from liver in those animals during immune clearance was estimated to produce a half-life between 17 hours and 1.5 days, in line with the 15-hour to 1.1-day half-life of first phase dynamics after antiviral therapy. The export rate of HBV DNA from transgenically infected mice was calculated at 16 hours using bromodeoxyuridine labeling.19 Both these rates match the 15-hour to 1.1-day half-life of first phase viral decay after antiviral therapy. Moreover the decay observed in the quantity of HBV DNA in blood during immune clearance in chimpanzees follows the pattern of decay detected in intracellular HBV DNA rather than being due to any extracellular process.14
Our calculations suggest that the half-life of HBV in blood for humans is at most 4.4 hours (Fig. 2), a duration significantly less than previous estimates of 1 day. This value is consistent with estimates of the half-life of HCV virions (2.7 hours,7 1.7–3 hours8) and also for estimates of the half-life of HIV virions (6 hours,9 0.5–2 hours8).
Chronically HBV-infected individuals exhibit approximately 109 virions per milliliter of serum. Assuming these are distributed through 3 L of serum gives a total viral load of 3 × 1012. However, this does not include the already formed double-stranded HBV DNA capsids within infected cells. Hence, if reverse transcriptase inhibitor therapy is commenced, there is a total pool of approximately 1013 virions available to infect new cells (those in blood plus the 4-fold higher population within cells), assuming a total blockage in new production. The HBV genome has 3,200 base pairs and assuming HBV reverse transcription exhibits a mutation rate of at least 10−6 base substitutions per site per replication, an order of magnitude less than HIV,21 then any of the 3,200 base substitutions from the consensus sequence will be achieved with a probability of 0.0032 per virion. Thus, of the 1013 virions present after the introduction of therapy, there will be 3.2 × 1010 single mutation virions covering the entire spectrum of all possible single mutations. Similar calculations lead to the conclusion that there will also be two-thirds of all double mutations already present.
Traditional therapy for chronically HBV-infected individuals involves monotherapy with either interferon-α or a reverse transcriptase inhibitor such as lamivudine. A single mutation in the polymerase gene involving a substitution of isoleucine for methionine at position 552 incurs a 10,000-fold increase in the IC50 for lamivudine.22 The double mutations L528M/M552V also provide this level of resistance. Previous calculations suggest that the single mutation clone will be present at a population size of 3.3 × 106 [the number of virions with single mutations (3.2 × 1010), divided by the number of different possible single mutations (3 × 3,200)]. Similarly, there will be approximately 1 copy of each double mutation. Hence monotherapy with drugs susceptible to single and double mutational resistance would seem doomed to failure, especially given that infection at the cellular level is slowly removed.