| |
Persistence of cccDNA during the natural history of chronic hepatitis B and decline during adefovir dipivoxil therapy
|
| |
| |
Gastroenterology; June 2004, Volume 126, Number 7
Bettina Werle—lapostolle * Scott Bowden ‡ Stephen Locarnini ‡ Karsten Wursthorn § Jorg Petersen § George Lau || Christian Trepo * Patrick Marcellin ¶ Zachary Goodman # William E. Delaney IV ** Shelly Xiong ** Carol L. Brosgart ** Shan—Shan Chen ** Craig S. Gibbs ** Fabien Zoulim * *
*INSERM Unit 271, Lyon, France, ‡Victorian Infectious Diseases Reference Laboratory, Melbourne, Australia, §Universitaetsklinikum Hamburg-Eppendorf and Heinrich-Pette-Institute for Experimental Virology and Immunology, Hamburg, Germany, ||Queen Mary Hospital, Hong Kong, Hong Kong, ¶INSERM Unité 481 and Hôpital Beaujon, Clichy, France, #Armed Forces Institute of Pathology, Washington, DC, USA, **Gilead Sciences, Foster City, California, USA
In this report, we describe the development of a novel and selective real-time PCR assay for quantifying cccDNA and total HBV DNA in liver biopsy specimens from CH-B patients. This technique allowed, for the first time, an analysis of cccDNA levels in significant numbers of patients in different phases of the natural history of chronic hepatitis B as well as a longitudinal analysis of patients undergoing antiviral therapy during a placebo-controlled study of adefovir for chronic hepatitis B (ADV).
"...Chronic HBV infection is believed to be maintained by a replicative form of HBV DNA termed cccDNA. During infection, HBV cccDNA accumulates in cell nuclei in which it persists... Median changes in serum HBV DNA were --4.7 vs. --0.6 log10 copies/mL in ADV-treated and placebo patients, respectively (P < 0.001)...the decay rate of HBV replicative forms during ADV therapy can be ranked as follows: serum HBV DNA > total intracellular HBV DNA > cccDNA... After 48 weeks, patients receiving placebo had a median --0.09 log10 ng/mL change in HBsAg titer (P = 0.85 vs. baseline). However, patients receiving ADV therapy demonstrated a median 73% reduction in serum HBsAg (median change of --0.58 log10 ng/mL); this change was significant compared with baseline (P < 0.001) and compared with changes in the placebo group (P < 0.001). Changes in HBsAg were significantly and positively correlated with changes in intrahepatic cccDNA (P = 0.008), total intracellular HBV DNA (P = 0.002), and serum HBV DNA (P < 0.001)... Changes in cccDNA were highly correlated with changes in total intrahepatic HBV DNA (P < 0.001) as well as with changes in serum HBV DNA (P = 0.008). We also examined whether changes in cccDNA correlated with changes in ALT levels, liver histology, or the number of cells staining positive for either HBcAg, or HBsAg; none of these correlations were significant... our study of HBV cccDNA in the livers of chronically infected patients provides new insight on the persistence of cccDNA throughout the natural history of CH-B and on the kinetics of cccDNA clearance during potent antiviral therapy. Our observations, together with previously published clinical studies, further support the argument for long-term antiviral therapy using agents that are both potent and have a low incidence of viral resistance. Additional clinical studies focusing on virological outcomes are ongoing. The development of rational combination regimens may enhance antiviral response and further minimize resistance, thus fostering the progressive clearance of cccDNA reservoirs and enhanced anti-HBV immune responses."
INTRODUCTION
Chronic hepatitis B virus (HBV) infection remains a major health problem, affecting approximately 400 million people worldwide. Although HBV replication is only mildly cytopathic, cellular immune responses directed against the virus can produce substantial liver damage and result in chronic hepatitis, cirrhosis, and hepatocellular carcinoma. Chronic hepatitis B (CH-B) encompasses a broad spectrum of disease and can be divided into several phases of natural history, which are distinguishable by serologic and virologic markers. One goal of antiviral therapy is to accelerate patients with chronic active disease into later natural history phases, wherein viral replication abates and the incidence of serious liver disease is significantly reduced.
Chronic infection is believed to be maintained by a replicative form of HBV DNA termed cccDNA. During infection, HBV cccDNA accumulates in cell nuclei in which it persists as a stable episome and acts as a template for the transcription of viral genes. Viral transcripts are translated into capsid and polymerase proteins and subsequently encapsidated and retrotranscribed into new partially double-stranded viral genomes. DNA-containing nucleocapsids can be enveloped and secreted from the cell as mature virus or cycled back to the nucleus to maintain the cccDNA pool. Considering the long half-life of hepatocytes, the limiting factor in eliminating infection is thought to be the clearance of cccDNA reservoirs from infected cells. Two immune mechanisms have been proposed to mediate cccDNA clearance: (1) a cytolytic mechanism by which infected cells are eliminated and replaced by cells from an uninfected lineage and (2) a noncytolytic, cytokine-induced "curing" of infected cells. There is evidence from animal models supporting each mechanism, and, although these mechanisms are not mutually exclusive, it is currently unclear how infection is resolved in patients.
Despite the crucial role of cccDNA during persistent infection and the importance of understanding clearance mechanisms, few data have been collected from patients. Indeed, our current understanding of cccDNA has been obtained primarily through studies of the woodchuck and duck hepatitis B virus models. Historic obstacles to the study of HBV cccDNA have been (1) the requirement for liver biopsy specimens, which are difficult to collect, especially from patients in quiescent natural history phases; and (2) the lack of sensitive, specific, and quantitative methods for detection of cccDNA from biopsy specimens. In this study, we report the development of a novel real-time PCR assay that allowed us to quantify levels of cccDNA in biopsy specimens collected from CH-B patients. To gain insight on the persistence of cccDNA and clearance mechanisms, we studied patients in different phases of the natural history of CH-B and a group of patients participating in a placebo-controlled trial of the recently approved nucleotide analog adefovir dipivoxil (ADV). These studies have provided novel insight into cccDNA persistence and have major implications for the monitoring of CH-B and antiviral therapy.
ABSTRACT
Background & Aims: Hepatitis B virus (HBV) covalently closed circular DNA (cccDNA) is a unique episomal replicative intermediate responsible for persistent infection of hepatocytes. Technical constraints have hampered the direct study of cccDNA maintenance and clearance mechanisms in patients. The aim of this study was to develop a sensitive and specific assay for quantifying cccDNA in biopsy samples from chronic hepatitis B patients during different natural history phases and in patients undergoing antiviral therapy.
Methods: Intrahepatic cccDNA levels were quantified by a specific real-time PCR assay. Ninety-eight liver biopsy samples from patients in the major phases of the natural history of chronic hepatitis B and 32 pairs of samples from patients receiving adefovir dipivoxil (ADV) therapy were assessed.
Results: cccDNA was detected, at levels ranging over 3 orders of magnitude, in patients in different phases of the natural history of chronic hepatitis B. cccDNA levels were strongly correlated with levels of total intracellular HBV DNA and serum HBV DNA. Forty-eight weeks of ADV therapy resulted in a significant 0.8 log decrease in cccDNA copies/cell. Changes in cccDNA were correlated with a similar reduction in serum HBsAg titer but not with a decrease in the number of HBV antigen-positive cells during ADV treatment.
Conclusions: cccDNA persists throughout the natural history of chronic hepatitis B, even in patients with serologic evidence of viral clearance. Long-term ADV therapy significantly decreased cccDNA levels by a primarily noncytolytic mechanism.
Patients
Biopsy specimens were collected from a total of 98 CH-B patients corresponding to the relevant phases of the natural history including 63 patients with hepatitis B e antigen—positive (HBeAg+) active CH-B, 18 patients with HBeAg-negative (HBeAg--) CH-B (presumed precore mutant), 10 inactive carriers (hepatitis B surface antigen—positive [HBsAg+], HBeAg--, antiHBeAb+), and 7 HBsAg seroconverters (serologically resolved patients with HBsAg--, anti-HBsAb+, HBeAg--, anti-HBeAb+, and anti—hepatitis B core antibody positive [anti-HBcAb+]). Thirty-two patients were also part of a phase III clinical trial evaluating the antiviral efficacy of ADV (10 patients received placebo, 22 received 10 mg or 30 mg of ADV daily); baseline biopsy specimens were included in the natural history study. Biopsy specimens were stored at --80°C until experimental analysis.
RESULTS
CccDNA levels during different phase of the natural history of chronic hepatitis B
We developed and validated a real-time PCR assay for the specific detection of cccDNA and total intracellular HBV DNA in liver biopsy specimens. Using this assay, we analyzed a panel of 98 patients in different phases of the natural history of CH-B. cccDNA and total intracellular HBV DNA were measured and normalized to per cell values (using the cellular -globin gene) at 3 independent laboratories, and the results were integrated.
Median intrahepatic cccDNA levels in patients with HBeAg+ CH-B (1.4 copies/cell) were significantly higher than in HBeAg-- patients (0.01 copies/cell; P < 0.001) or inactive carriers (0.02 copies/cell; P < 0.001). Patients with evidence of HBsAg clearance had extremely low levels of cccDNA (0.002 copies/cell; P < 0.001, compared with HBeAg+ patients). Similarly, median total intracellular HBV DNA levels in HBeAg+ patients (155 copies/cell) were significantly higher compared with HBeAg-- patients (0.6 copies/cell; P < 0.001), inactive carriers (2 copies/cell) (P < 0.001), and HBsAg-- patients (0.19 copies/cell; P < 0.001). Levels of cccDNA and total intracellular HBV DNA were both highly correlated with viral load (HBV DNA) in serum of these patients (P < 0.001, data not shown).
Effect of 48 weeks of ADV therapy on cccDNA levels in patients
We next determined the effect of 48 weeks of ADV therapy on cccDNA and total intrahepatic HBV DNA in 32 patients (22 treated, 10 placebo) with HBeAg+ CH-B. Our integrated analysis revealed median changes in cccDNA and total intrahepatic HBV DNA in ADV-treated patients of --0.80 and --1.63 log10 copies/cell, respectively. This equates to an 84% reduction of cccDNA and 98% reduction of total intracellular HBV DNA by the end of treatment. Reductions in cccDNA and total intrahepatic HBV DNA were highly significant compared with placebo patients who had median changes of +0.32 and --0.0001 log10 copies/cell in cccDNA and total intrahepatic HBV DNA, respectively (P = 0.002, cccDNA, treated vs. placebo; P < 0.001 total intracellular HBV DNA, treated vs. placebo). To ensure that the integrated results were not biased by the results from a single laboratory, we analyzed each site's data independently. Although absolute values for cccDNA and total intracellular HBV DNA varied between sites, the magnitudes of DNA loss were very similar and supported the integrated findings. Median changes in serum HBV DNA were --4.7 vs. --0.6 log10 copies/mL in ADV-treated and placebo patients, respectively (P < 0.001). Therefore, the decay rate of HBV replicative forms during ADV therapy can be ranked as follows: serum HBV DNA > total intracellular HBV DNA > cccDNA.
We also measured changes in HBsAg titers in the serum of ADV-treated patients. At baseline, patients in the ADV and placebo groups had similar HBsAg titers (medians of 332,599 and 217,378 ng/mL, respectively, P = 0.48). After 48 weeks, patients receiving placebo had a median --0.09 log10 ng/mL change in HBsAg titer (P = 0.85 vs. baseline). However, patients receiving ADV therapy demonstrated a median 73% reduction in serum HBsAg (median change of --0.58 log10 ng/mL); this change was significant compared with baseline (P < 0.001) and compared with changes in the placebo group (P < 0.001). Changes in HBsAg were significantly and positively correlated with changes in intrahepatic cccDNA (P = 0.008), total intracellular HBV DNA (P = 0.002), and serum HBV DNA (P < 0.001).
Interestingly, patients who underwent HBeAg seroconversion had significantly lower baseline levels of intracellular cccDNA compared with patients who remained HBeAg+ (P = 0.034). However, patients receiving ADV underwent similar declines in cccDNA, regardless of whether they underwent HBeAg seroconversion (n = 5, --0.94 log10 copies/cell) or not (n = 17, --0.75 log10 copies/cell) (P = 0.91).
Immunohistochemical staining of biopsy specimens for HBV antigens
For patients in the ADV study, immunohistochemical staining was performed to determine the number of hepatitis B core antigen (HBcAg) and HBsAg positive cells in each biopsy specimen (data not shown). In most biopsy specimens, only a minority of the cells stained positive for HBcAg or HBsAg, and positive cells tended to cluster together. There was a significant correlation between cccDNA copy number and the number of HBcAg+ cells (P = 0.015, data not shown). A stronger correlation existed between the level of total intrahepatic HBV DNA and number of HBcAg+ cells (P < 0.001, data not shown). Normalization of cccDNA levels to the number of infected cells using HBcAg as a marker (instead of total cell number using b-globin) yielded a median copy number of 33 copies per infected cell, which is consistent with copy numbers derived from animal hepadnavirus models. Although 48 weeks of antiviral therapy significantly decreased cccDNA levels, the number of infected (positive staining) cells did not change significantly (data not shown), indicating a decrease in cccDNA copy number per cell rather than a loss of a proportion of infected cells.
Correlation of cccDNA changes with baseline and posttherapy parameters
Using the cccDNA and clinical data available for patients in the ADV study, we examined whether baseline parameters were predictive of cccDNA loss and whether changes in cccDNA during therapy correlated with changes in other clinical parameters. Baseline alanine aminotransferase (ALT), liver histology (histologic activity index), and viral load, using linear regression models, were not predictive for cccDNA loss. Changes in cccDNA were highly correlated with changes in total intrahepatic HBV DNA (P < 0.001) as well as with changes in serum HBV DNA (P = 0.008). We also examined whether changes in cccDNA correlated with changes in ALT levels, liver histology, or the number of cells staining positive for either HBcAg, or HBsAg; none of these correlations were significant.
AUTHOR DISCUSSION
In this report, we describe the development of a novel and selective real-time PCR assay for quantifying cccDNA and total HBV DNA in liver biopsy specimens from CH-B patients. This technique allowed, for the first time, an analysis of cccDNA levels in significant numbers of patients in different phases of the natural history of CH-B as well as a longitudinal analysis of patients undergoing antiviral therapy during a placebo-controlled study. This is noteworthy in that these findings are likely derived from the last placebo-controlled trial in CH-B patients.
Our cross-sectional natural history study revealed that cccDNA levels were significantly higher in HBeAg+ patients compared with HBeAg-- patients, inactive carriers, and patients who underwent HBsAg clearance. This result is consistent with our observations of greater levels of intrahepatic HBV replicative intermediates and serum HBV DNA in HBeAg+ patients (compared with the later groups) and with previous reports of viral load in HBeAg+ and HBeAg-- patients. Surprisingly, cccDNA levels in HBeAg-- patients, who are infected with precore mutant HBV and have active liver disease, were not significantly different (P = 0.68), than those in inactive carriers who have no evidence of ongoing liver disease. This is potentially attributable to the natural variation in HBV replication levels observed in patients infected with precore mutant HBV and the relatively small sample size for the inactive carrier group. However, these data could also indicate that the primary difference between a patient with HBeAg-- active hepatitis and an inactive carrier is due to immunologic rather than virologic factors.
The observation that cccDNA remained detectable in all HBeAg-- patients explains why viral reactivation has been observed, to varying frequencies, in all these patient groups. Indeed, fluctuating levels of viral replication and hepatitis are common in patients with precore mutant infection. Viral reactivation and hepatitis flares have also been observed, albeit less frequently, in inactive carriers during long-term natural history studies. Even patients with serologic evidence of HBV clearance can develop active disease during periods of severe immune suppression such as transplantation, cancer chemotherapy, or steroid use. The observed low-level persistence of cccDNA also explains the occurrence of de novo HBV infection in liver transplantation patients who received grafts from donors who previously resolved infection. Overall, our observations from the natural history study are consistent with the theory that resolution of hepadnaviral infection does not occur through "sterilizing immunity" but rather by consistent control of viral replication by the host immune system.
Clinical trials of nucleoside analogs have indicated that short-term therapy is followed by viral recrudescence in most CH-B patients. Studies performed in animal hepadnavirus models suggested that the long half-life of viral cccDNA is responsible for viral persistence and reactivation following drug cessation. Here, we show, for the first time, that antiviral therapy with a potent HBV polymerase inhibitor significantly reduces cccDNA in CH-B patients. Importantly, reductions in cccDNA were correlated with, and similar in magnitude to, reductions in serum HBsAg titer. The parallel change in HBsAg provides further evidence that transcriptionally active cccDNA is being depleted during therapy.
The rate of cccDNA loss was nearly an order of magnitude lower than that of intracellular HBV replicative intermediates and several orders of magnitude lower than the rate of serum HBV DNA loss. Factors contributing to the relatively slow loss of cccDNA during therapy potentially include (1) the existence of cccDNA as a chromatinized episome; (2) the asymmetric nature of hepadnaviral replication, which protects cccDNA from being directly depleted by chain-terminating polymerase inhibitors; (3) the preferential use of mature viral nucleocapsids to replenish cccDNA pools during periods of reduced replication; and (4) the potential for hepatocyte reinfection by residual circulating virus.
Results from our study potentially explain those of kinetic analyses of viral load decay during potent antiviral therapy. Two main phases of viral load decline have previously been identified: a first phase of rapid viral decline (attributed to the inhibition of viral production in infected cells) and a second phase of slow decline (hypothesized to represent the clearance of infected cells). The slow rate of cccDNA clearance that we observed is likely to contribute to the slower second phase kinetic. The observed longevity of cccDNA also reinforces the prediction that long-term antiviral therapy will be required to control HBV replication unless a vigorous host immune response is mounted. In this respect, our results are reminiscent of studies of HIV proviral DNA. Fortunately, treatment of CH-B patients with potent nucleos/tides such as lamivudine and adefovir dipivoxil can result in partial restoration of immune responses, which are necessary for the durable host-mediated control of infection.
Mechanisms for clearance of viral cccDNA are being actively debated. Based on results of acute hepadnaviral infections in chimpanzees and woodchucks, it is unclear whether the primary clearance mechanism of hepadnavirus infection is mediated by a noncytolytic TH-1 immune response or by immune-mediated cell killing, followed by hepatocyte cell division. Although both mechanisms are likely to factor into viral clearance, our data support the former mechanism, at least in the setting of antiviral therapy.
Immunohistochemical staining did not identify a significant correlation between loss of cccDNA and the disappearance of viral antigen expressing cells, suggesting that cccDNA loss during antiviral therapy occurred at a similar rate in most cells, rather than by the clearance or destruction of a subset of infected cells. We hypothesize that decreases in intrahepatic cccDNA were primarily due to the potent suppression of viral DNA synthesis, which would effectively deplete the pool of mature cytoplasmic nucleocapsids available for conversion into cccDNA. Our analyses also indicated that cccDNA clearance during antiviral therapy was not significantly correlated with baseline indicators of host-mediated hepatocyte lysis (e.g., high baseline ALT levels and high baseline histologic activity index). This also suggests that the clearance of cccDNA during antiviral therapy may be occurring primarily through a mechanism other than infected cell killing. Based on a mathematic modeling using cccDNA data derived from patients in the ADV substudy, it would take approximately 14.5 years to completely clear a chronically HBV-infected human liver of intracellular cccDNA. However, because intracellular cccDNA remained detectable in patients who resolved chronic hepatitis B (HBsAg seroconverters), it may not be necessary to eradicate completely cccDNA to establish long-term control of HBV infection.
Despite the widespread availability of quantitative PCR technology, it is not currently practical to measure cccDNA in biopsy specimens during routine clinical practice. However, monitoring cccDNA may provide an end point if a threshold can be determined, below which host-mediated control is likely to establish control of viral replication. Further longitudinal studies are ongoing to investigate such end points.
In conclusion, our study of HBV cccDNA in the livers of chronically infected patients provides new insight on the persistence of cccDNA throughout the natural history of CH-B and on the kinetics of cccDNA clearance during potent antiviral therapy. Our observations, together with previously published clinical studies, further support the argument for long-term antiviral therapy using agents that are both potent and have a low incidence of viral resistance. In this regard, additional clinical studies focusing on virological outcomes are ongoing. The development of rational combination regimens may enhance antiviral response and further minimize resistance, thus fostering the progressive clearance of cccDNA reservoirs and enhanced anti-HBV immune responses.
|
|
| |
| |
|
|
|
