Hepatitis B virus as a gene delivery vector activating foreign antigenic T cell response that abrogates viral expression in mouse models


  • Qiang Deng,

    1. Laboratoire Pathogenèse des Virus de l'Hépatite B, Institut Pasteur, Paris, France
    2. Institut National de la Santé et de la Recherche Médicale U845, Centre de Recherche Croissance et Signalisation, Faculté de Médecine Paris Descartes, Paris, France
    Current affiliation:
    1. Unit of Tumor Virus, Institut Pasteur of Shanghai, Shanghai, China
    Search for more papers by this author
  • Maryline Mancini-Bourgine,

    1. Laboratoire Pathogenèse des Virus de l'Hépatite B, Institut Pasteur, Paris, France
    2. Institut National de la Santé et de la Recherche Médicale U845, Centre de Recherche Croissance et Signalisation, Faculté de Médecine Paris Descartes, Paris, France
    Search for more papers by this author
  • Xiaoming Zhang,

    1. Unité Régulation Immunitaire et Vaccinologie, Paris, France
    2. Institut National de la Santé et de la Recherche Médicale U883, Paris, France
    Search for more papers by this author
  • Marie-Christine Cumont,

    1. Unité de Recherche et d'Expertise Histotechnologie et Pathologie, Institut Pasteur, Paris, France
    Search for more papers by this author
  • Ren Zhu,

    1. Laboratoire Pathogenèse des Virus de l'Hépatite B, Institut Pasteur, Paris, France
    2. Institut National de la Santé et de la Recherche Médicale U845, Centre de Recherche Croissance et Signalisation, Faculté de Médecine Paris Descartes, Paris, France
    Search for more papers by this author
  • Yu-Chun Lone,

    1. Institut National de la Santé et de la Recherche Médicale U542, Hôpital Paul Brousse, Villejuif, France
    Search for more papers by this author
  • Marie-Louise Michel

    Corresponding author
    1. Laboratoire Pathogenèse des Virus de l'Hépatite B, Institut Pasteur, Paris, France
    2. Institut National de la Santé et de la Recherche Médicale U845, Centre de Recherche Croissance et Signalisation, Faculté de Médecine Paris Descartes, Paris, France
    • Bâtiment Lwoff, Institut Pasteur, 28 rue du Docteur Roux, 75724 Paris Cedex 15
    Search for more papers by this author
    • fax: (33)-1-40-61-38-41.

  • Potential conflict of interest: Nothing to report.


Chronic hepatitis B virus (HBV) infection is characterized by functionally impaired T cell responses. To ensure active immunotherapy, the immune response must be switched from exhausted T cells to functional effectors that can attain the liver and cure the viral infection. We thus designed a recombinant HBV (rHBV) containing a modified viral core gene that specifically delivers a foreign antigenic polyepitope to the liver. This recombinant virus could only be self-maintained in hepatocytes already infected by HBV through capsid complementation. A strong foreign epitope-specific T cell response was first primed in the periphery by way of DNA immunization in human leukocyte antigen (HLA)-A2/DR1 transgenic mice. After the hydrodynamic (hyd.) injection of rHBV, expression of the foreign antigenic polyepitope in hepatocytes attracted/reactivated a vigorous T cell response in situ. Most liver-infiltrating CD8+ T cells proved to be functional effectors. Following DNA priming and hyd. injection, the rHBV-based expression of hepatitis B surface antigen (HBsAg) in mouse liver was almost completely inhibited without causing major liver injury. Studies in HBsAg/HLA-A2/DR1 transgenic mice further validated our approach. Conclusion: For the first time, HBV was used as a gene delivery vector, which strongly triggered functional T cell response and subsequently controlled the viral expression in the liver of surrogate mouse models for HBV infection. It might represent an innovative and promising strategy of active immunotherapy during HBV persistent infection. This concept could even be more generally extended to other chronic viral diseases. (HEPATOLOGY 2009.)

Hepatitis B virus (HBV) remains a major cause of human disease worldwide. Functional impairment of HBV-specific T cell responses is believed to be the principal reason for host inability to eliminate the persistent infection.1 This reflects a balance between effector functions required to eliminate the pathogen and their potential to cause immunopathology.

In order to treat chronic HBV infection, intrahepatic T cell responses need to be switched from a state of dysfunction to fully efficient effector T cells. Previous studies have shown that HBV-based therapeutic vaccination might specifically activate T cell responses in chronic HBV carriers, although these appear to be transient and followed by a gradual decline.2, 3 For this reason, escape from built-up tolerance seems to be one of the most crucial points in active immunotherapy for persistent viral infection.4 To achieve this goal, we propose an innovative strategy based on a non-HBV antigen-mediated functional T cell response that is intended to specifically target HBV and HBV-infected liver cells. Briefly, the non-HBV antigen consists of a portion of a foreign antigenic polyepitope that has been engineered in the HBV genome to create a recombinant hepatitis B virus (rHBV). Thus, HBV is used as a gene therapy vector to deliver the foreign antigenic polyepitope to the liver.

HBV infects only humans and chimpanzees. In the present study, we used the hydrodynamic (hyd.) injection technique5, 6 to deliver rHBV directly to the liver, in an attempt to mimic viral infection and acute hepatitis in a mouse model. A human leukocyte antigen (HLA)-transgenic mouse lineage (HLA-A2/DR1) was used as a humanized model to assess immune response.7 In the presence of peripheral foreign epitope-specific T effector cells, the hyd. injection of rHBV attracted a strong intrahepatic immune response that quickly abrogated rHBV and hepatitis B surface antigen (HBsAg) expression without causing any marked liver injury. Because rHBV is expected to replicate in hepatocytes bearing the wild-type HBV (wtHBV) only and share the same mechanism in the viral cycle, the functional immune response elicited by the foreign polyepitope may bypass the exhausted T cell response present during persistent HBV infection.


ALT, alanine aminotransferase; EBV, Epstein-Barr virus; HBc, hepatitis B virus core protein; HBsAg, hepatitis B virus surface antigen; HBV, hepatitis B virus; HIV, human immunodeficiency virus; HLA, human leukocyte antigen; hyd., hydrodynamic; IFN-γ, interferon-γ; i.m., intramuscular injection; ORF, open reading frame; pCMV, cytomegalovirus plasmid; pgRNA, pregenomic RNA; prHBV1.3, recombinant hepatitis B virus plasmid; pwtHBV, wild-type hepatitis B virus plasmid; rHBc, recombinant hepatitis B virus core protein; rHBV, recombinant hepatitis B virus; SEM, standard error of the mean; TNF-α, tumor necrosis factor α; wtHBV, wild-type hepatitis B virus.

Materials and Methods


HLA-A2/DR1 mice have been described.7 The HBsAg/HLA-A2/DR1 triple transgenic lineage is endowed with an HLA-A2/HLA-DR1 background and produces HBsAg in mouse liver and sera following transgene expression (Refs.8, 9, and unpublished data). DNA immunization was performed by injecting 100 μg plasmid DNA as described.10 For hyd. injection, 25 μg plasmid DNA was injected through the tail veins in a volume of phosphate-buffered saline equivalent to 8% of the mouse body weight to 10- to 15-week-old female mice.5

DNA Constructs.

The recombinant hepatitis B virus plasmid (prHBV1.3) is constructed under an HBV D genotype background (No. V01460, GenBank) (Fig. 1 and Supporting Materials and Methods). The prHBV1.3/HBV core protein (HBc) plasmid had an additional expression cassette of HBc driven by an SV40 early promoter and inserted downstream of the rHBV genome in prHBV1.3 (Fig. 1D). The cytomegalovirus (CMV)-rHBc plasmid (pCMV-rHBc) encodes rHBc, as well as a full copy of rHBV genome driven by the cytomegalovirus promoter, which starts from the HBV e antigen open reading frame (ORF) and ends with the HBV polyA processing sequence (nt 1981) (Fig. 1D).

Figure 1.

Schematic representation of rHBV constructs (A). HBV pgRNA is represented by the capping site (CAP), encapsidation signal (ϵ), and polyadenylation site (An). The distance between AUG codons of core and polymerase ORFs is given in nucleotides (nts). (B) rHBV pgRNA. A short DNA sequence encoding the foreign antigenic polyepitope was inserted in-frame within the core ORF allowing the expression of a chimeric protein (rHBc). (C) Schematic representation of rHBc-encoding domain. Two in-frame ATG codons for HBV pre-core/core proteins are indicated. The polyepitope comprises a B cell epitope (FLAG), three HLA-A2–restricted CD8+ T cell epitopes derived from HIV gag, the Flu matrix and EBV BMLF-1 proteins, and a universal CD4+ T cell epitope PADRE. (D) rHBV-based plasmids used in this study. prHBV1.3 carries 1.3 copies of the rHBV genome. prHBV1.3/HBc bears an additional cassette for HBV core gene expression driven by the SV40 early gene promoter (P-SV40), with a polyadenylation site from bovine growth hormone (BGH pA). pCMV-rHBc allows the expression of rHBc as well as the rHBV genome driven by the cytomegalovirus early gene promoter (pCMV). The polyepitope insert and HBc in-frame ORFs are indicated by arrows. Nucleotide positions are given according to the HBV sequence of the ayw subtype.

Cell Preparation and Flow Cytometry Analysis.

Mice were perfused with phosphate-buffered saline and the liver was smashed in a 100-μm cell strainer (BD Biosciences). Cell pellets were suspended in 40% Percoll (Sigma) and then centrifuged to remove hepatocyte clumps from the top layer.11 Isolated cells were labeled with the following monoclonal antibodies: anti-CD3 (145-2C11), anti-CD4 (GK1.5), anti-CD8 (53–6.7), anti-CD69 (H1.2F3), anti-CD62L (MEL-14), anti-CD107a (1D4B), anti–interferon-γ (IFN-γ) (XMG1.2), and anti–tumor necrosis factor α (TNF-α) (MP6-XT22) (BD Biosciences). Cells were labeled with HLA-A2 tetramers conjugated with influenza (Flu) or Epstein-Barr virus (EBV) peptides or with pentamers conjugated with HBV S348–357 or S183–191 (ProImmune). HCV NS3 (KLVALGINAV) tetramer and pentamer were used as controls for the staining. For the degranulation assay,12 anti-CD107a was added during peptide stimulation. Intracellular cytokine detection was performed as described12 (Supplementary Materials and Methods). Cells were acquired on FACSCanto (BD Biosciences) and analyzed using Flowjo software (Tree Star).

Southern Blot Analysis of Viral DNA Intermediates.

The Huh 7 cell line was used for transient transfection assays. rHBV-associated viral DNA was extracted 3 days after DNA transfection from Huh 7 cells lysed in 100 mM Tris/HCl (pH 8.0) containing 0.02% Nonidet P40. The lysates were treated with DNase I and RNase A, and then underwent protease K digestion in the presence of 1% sodium dodecyl sulfate for 2 hours at 55°C. The viral DNA was ethanol-precipitated after phenol extraction. Southern blot hybridization was performed as described elsewhere,13 using an HBV genome–specific 32P-labeled probe.

Histological Assays and Immunofluorescence Staining.

Five-micrometer-thick cryostat sections were mounted on superfrost plus slides. Before staining, the slides were fixed in ice-cold acetone for 5 to 10 minutes. Liver sections were stained with hematoxylin-eosin or immunostained with anti-Flag monoclonal antibody (Sigma-Aldrich) or fluorescein isothiocyanate–labeled anti-HBs antibody (Abcam) at 4°C overnight. A secondary Alexa 488–labeled antibody (Molecular Probes) was used for anti-Flag staining. After washing, the liver sections were mounted with anti-fade reagent containing 4′,6-diamidino-2-phenylindole and visualized directly under a fluorescence microscope. Transfected cells were fixed with 4% paraformaldehyde followed by permeablization with 0.1% Triton X 100 in phosphate-buffered saline before immunostaining.


Data are expressed as the mean ± standard error of the mean (SEM). Nonparametric unpaired comparisons were performed using Student unpaired t tests with GraphPad Prism software.


Construction of the rHBV Genome Bearing a Foreign Polyepitope.

rHBV was constructed by modifying the wtHBV genome. The only difference was a 325-nt fragment within the viral core gene that was substituted by an in-frame 180-nt sequence encoding a foreign antigenic polyepitope (Fig. 1A,B). As a result, the ORF of the viral polymerase gene (pol) was shifted forward by 135 nt in viral pregenomic RNA (pgRNA), enabling potentially easier access to the translational machinery. Furthermore, the translational start site of the pol gene was optimized according to Kozak's rules in order to facilitate potential ribosome entry. The foreign polyepitope was engineered with three CD8+ T cell epitopes combined with a promiscuous CD4+ T cell epitope (PADRE) that universally matched up with most prevalent major histocompatibility complex class II molecules.14 In view of their clinical relevance, three well-known HLA-A2–restricted epitopes derived from common human viruses (human immunodeficiency virus [HIV] gag p17 aa 76–84, Flu matrix protein aa 58–66, and EBV BML-F1 protein aa 259–267) were chosen in order to elicit a vigorous immune response in vivo. In the rHBV construct, the HLA-A2–restricted epitope HBV core (HBc) aa 18–2715 was preserved. A short B cell epitope (FLAG) was placed at the N-terminal part of the foreign sequence to act as a convenient marker of detection (Fig. 1C). A chimeric antigenic protein referred here to as rHBc could thus be generated, with the foreign polyepitope fused in-frame within the truncated core proteins (Fig. 1C).

Expression of the Recombinant Protein Carrying the Polyepitope.

The chimeric rHBc protein was expressed by prHBV1.3, a plasmid bearing 1.3 copies of the rHBV genome (Fig. 1D). After the transfection of the hepatoma cell line HepG2 with prHBV1.3, rHBc was identified in cell lysates with its expected size (15 kDa) by way of Western blot assay (Fig. 2A). Using the anti-Flag antibody, rHBc protein was further visualized in the cytoplasm of transfected HepG2 cells by way of immunofluorescence staining (Fig. 2B, right panel). HBsAg, which is carried by viral envelope proteins, was detected by staining with specific anti-HBs antibody (Fig. 2B, left panel). We also took advantage of the hyd. injection technique in order to deliver rHBV directly into mouse liver. Four days after the injection of prHBV1.3 via the tail vein, both rHBc and HBsAg were identified in liver sections by immunoflorescence staining with anti-FLAG and anti-HBs antibodies, respectively (Fig 2C). These experiments indicate that the rHBc carrying the foreign polyepitope could be expressed in vitro and in vivo in the context of the rHBV genome.

Figure 2.

Expression of rHBV and chimeric antigenic protein. (A) Western blot assay of HepG2 cell lysate after transfection with prHBV1.3. The rHBc product is indicated with its molecular weight (lane 2). FLAG-tagged proliferating cell nuclear antigen was used as a control (lane 1). (B) Immunofluorescence staining of HepG2 cells transfected with prHBV1.3, using anti-HBs (left panel) or anti-FLAG antibodies (right panel). The nucleus is stained with 4′,6-diamidino-2-phenylindole. Bars, 10 μm. (C) Staining of liver sections taken from mice 4 days after the hyd. injection of prHBV1.3, labeled as in (B). Bars, 200 μm.

Rescue of rHBV Replication by Way of in trans Encapsidation.

rHBV is an HBV virus that is defective because of its disrupted core gene. However, it is expected to complete the replicative cycle and be encapsidated in hepatocytes with the help of core proteins produced in trans. rHBV replication was directly evidenced by the detection of viral DNA intermediates in the Huh 7 cell line after the transfection of prHBV1.3/HBc, a plasmid carrying the rHBV genome and an additional HBc-expressing cassette (Fig. 1D). In the presence of the core protein expressed by prHBV1.3/HBc, the intracellular replicative intermediates of rHBV were identified in both their typical relaxed circular and double-stranded linear forms, as detected by HBV-specific probe hybridization (Fig. 3A). These replicative forms are smaller in size than wtHBV, indicating successful packaging of the viral genome and maturation of the rHBV nucleocapsid.16 By contrast, in the absence of core protein, the replicative forms of rHBV could not be observed in prHBV1.3-transfected cells. Moreover, the core-rescued rHBV exhibited a more efficient replication than wtHBV. This was demonstrated by cotransfecting Huh 7 cells with prHBV1.3 and the plasmid carrying 1.2 copies of the wtHBV genome (pwtHBV).17 Transfection with prHBV1.3 alone did not result in viral replication (Fig. 3B). By contrast, the replicative intermediates of rHBV turned out to be the major form (relaxed circular bands of smaller size) after cotransfection with pwtHBV at varying ratios, whereas wild-type replicative forms were found at much lower levels or were even undetectable. In addition, the viral DNA resulting from prHBV1.3/HBc transfection exhibited a more than 10-fold yield compared with pwtHBV during normalized experiments (Fig. 3B). Overall, these findings indicate that core-rescued rHBV could be comaintained with wtHBV in the hepatic cell line and compete for viral replication.

Figure 3.

Rescue of rHBV replication and HBsAg particle production. (A) Viral DNA detected by way of southern blot assay with an HBV-specific probe in Huh 7 cells transfected with prHBV1.3 (lanes 1, 3) or prHBV1.3/HBc plasmids (lanes 2, 4). wtHBV DNA was extracted from the HepAD38 cell line30 as a control (lane 5). Bands corresponding to relaxed circular (RC), double-stranded linear (DSL), and single-stranded (SS) HBV DNA are indicated. M, molecular weight markers (Kb). HBc + indicates expression of the HBc by the vector used in transfection experiments. (B) Southern blot assay of viral DNA in Huh 7 cells 3 days after the cotransfection of pwtHBV and prHBV1.3 at different ratios or prHBV1.3/HBc, as indicated in the table.

Activation of a Peripheral Polyepitope-Specific T Cell Response.

A DNA plasmid encoding rHBc (pCMV-rHBc) (Fig. 1D) was constructed and used to immunize HLA-A2/DR1 transgenic mice.7 Two weeks after an intramuscular (i.m.) injection of DNA, all tested mice exhibited vigorous polyepitope-specific T cell responses as detected by an ex vivo IFN-γ/enzyme-linked immunosorbent spot assay (Fig. 4A).18 The Flu matrix–derived epitope was the most frequently recognized and the most powerful of the three foreign CD8+ T cell epitopes. The EBV BML-F1–derived epitope elicited a subdominant immune response. HIV-specific T cell response was only detectable after 1 week of in vitro stimulation of splenocytes with the cognate peptide (data not shown). Four HBsAg-derived, HLA-A2–restricted peptides (S183–191, S348–357, S370–378 [data not shown] and S170–181 [data not shown]) and pools of overlapping 15-mer peptides derived from HBV core protein (C1→30 and C136→185) were used for an ex vivo stimulation of spleen cells (Fig. 4A). Only the peptide pool C1→30 that included the HBc aa 18–27 epitope was reactive. No HBsAg-specific T cells were detectable. Both the Flu- and EBV-specific T cell responses even predominated over the response to the HLA-A2 epitope HBc aa 18–27. The hierarchy of activation of the T cell response might reflect competition among these T cell epitopes during antigen processing and presentation.19 The T cell response to the Flu matrix epitope was also quantified using an HLA-A2 tetramer carrying the Flu peptide in order to label splenocytes from DNA-immunized mice. Flu-specific T cells thus represented around 10% of CD8+ T cells from the spleen (Fig. 4B, right panel). Mice receiving a pCMV-rHBc injection also developed a T helper response against the major histocompatibility complex class II–restricted epitope PADRE, as demonstrated by peptide-specific proliferation assay (Fig. 4C) and by the IFN-γ enzyme-linked immunosorbent spot assay (Fig. 4A).

Figure 4.

Polyepitope-specific T cell response in mice injected with pCMV-rHBc. (A) Enzyme-linked immunosorbent spot assay of splenocytes from six HLA-A2/DR1 transgenic mice 15 days after an intramuscular injection of pCMV-rHBc. Each bar represents the number of IFN-γ–secreting T cells per million splenocytes for each individual mouse, and the mean value is indicated by black bars. Peptides or peptide pools used to stimulate splenocytes ex vivo are derived from HIV gag (HIV), influenza matrix (FLU), EBV-BMLF1 (EBV), HBcAg (C18-27; C1→30; C136→185), HBsAg (S348–357 and S183–191), and PADRE. IFN-γ spots >50 per million cells are considered positive (dotted line). (B) FACS analysis of Flu-specific T cells. Splenocytes were stained with anti-CD8 and HLA-A2 tetramer carrying the Flu peptide. Left, nonimmunized mouse. Right, a representative mouse as in Fig. 4A. (C) Proliferative response of splenocytes from pCMV-rHBc–immunized HLA-A2/DR1 transgenic mice following in vitro stimulation with PADRE peptide. Responses are expressed as a proliferation index. The dotted line corresponds to the cutoff value of the stimulation index.

Targeting of Polyepitope-Specific T Cell Responses to the Liver.

We then designed a protocol of rHBV-based active immunization in HLA-A2/DR1 transgenic mice (Fig. 5A). In the absence of a mouse model susceptible to HBV infection, we used hyd. injection to ensure that rHBV would be directly expressed in mouse livers.5, 6 HLA-A2/DR1 mice were immunized by way of an i.m. injection of pCMV-rHBc at day 0 in order to prime a polyepitope-specific T cell response in the periphery. At day 15, prHBV1.3 was injected hydrodynamically to mimic viral infection. In this way, rHBV and the foreign antigenic products would be processed and presented in situ, providing intrahepatic targets for the CD8+ T cell response.

Figure 5.

T cell responses in liver and spleen following hyd. injection of rHBV. (A) Protocol for active immunization in HLA-A2/DR1 transgenic mice. T cell responses were primed at day 0 (D0) through i.m. injection of pCMV-rHBc. At D15, mice received hyd. injections of either prHBV1.3 or pCMV-βGal, pwtHBV control plasmids. A group of mice received a second i.m. injection of pCMV-rHBc at D15. Lymphocytes from the spleen and liver were isolated at D22 for FACS analysis. (B) FACS analysis of intrahepatic lymphocytes from a representative mouse receiving either pCMV-rHBc priming/hyd. pCMV-βGal injection (left panel), pCMV-rHBc priming/hyd. prHBV1.3 (middle panel), or pCMV-rHBc priming/hyd. pwtHBV (right panel) after staining with anti-CD3 and anti-CD8. Proportions of CD8+ (circle) and CD4+ (square) T cells among lymphocytes are indicated. (C) FACS analysis of intrahepatic lymphocytes from a representative mouse receiving pCMV-rHBc priming/hyd. prHBV1.3 (upper panels) and a mouse receiving pCMV-rHBc priming/hyd. pwtHBV (lower panels). CD3+ CD8+ T cells were stained with tetramers conjugated with Flu or EBV peptides and pentamers conjugated with C18–27, S348–357, or S183–191 peptides. (D) Analysis of intrahepatic (left panel) and splenic lymphocytes (right panel) from four groups of mice: mice receiving pCMV-rHBc priming/hyd. prHBV1.3 (white bars [n = 8]); mice receiving pCMV-rHBc priming/hyd. pwtHBV (dotted bars [n = 4]); mice receiving pCMV-rHBc priming/hyd. pCMV-βGal (gray bars [n = 6]); and mice injected with i.m. pCMV-rHBc two times (striped bars [n = 3]). Results are expressed as the mean ± SEM of percentage of CD8+, CD4+, and Flu-specific T cells in total lymphocytes. P values for comparison of each group are indicated **P < 0.01; ***P < 0.001.

Following this immunization protocol, the mice mounted a vigorous T cell response, with a large number of CD8+ T lymphocytes infiltrating the liver, as revealed by fluorescence-activated cell sorting (FACS) at day 7 after the hyd. injection. As shown in Fig. 5B, the percentage of infiltrating CD8+ T cells represented up to 37.4% of total hepatic lymphocytes in the mouse receiving prHBV1.3. In comparison, 7.31% and 16.6% of CD8+ lymphocytes were found in the liver of mice that received pCMV-βGal or pwtHBV as controls, respectively. Specific tetramer staining indicated that around 52% of hepatic CD8+ T cells were Flu-specific. The subdominant EBV-specific T cell population accounted for less than 1.62% (Fig. 5C, upper panels). In contrast, only 5% of Flu-specific CD8+ T cells were detected in the liver of a representative mouse receiving pwtHBV as control (Fig. 5C, lower panels). Therefore, Flu-specific T cells represented the major CD8+ T cells population in the intrahepatic infiltrate after rHBV-based active immunization.

The relative distribution of T cells was assessed in mice receiving either rHBV-based active immunization or controls (Fig. 5D). The percentage of CD8+ T cells observed in the livers of mice receiving pCMV-rHBc priming/hyd. prHBV1.3 was significantly higher than in mice receiving pCMV-rHBc priming/hyd. pwtHBV (P < 0.01) and in mice receiving pCMV-rHBc priming/hyd. pCMV-βGal (P < 0.001), and in those receiving two i.m. injections of pCMV-rHBc (P < 0.01). A similar but relatively lower increase of CD8+ T cells was observed in the spleen. As for hepatic lymphocytes, Flu-specific CD8+ T cells were in a very great majority, but their levels were significantly lower in the spleen (P < 0.0001). Overall, these experiments strongly suggest that the CD8+ T cell response (mostly Flu-specific) activated in the periphery was attracted to the liver and/or underwent an in situ expansion following rHBV-based active immunization.

An Acute Intrahepatic T Cell Response Without Marked Liver Injury.

To further characterize the intrahepatic infiltrates, liver sections collected 4 days after hyd. injection of prHBV1.3 underwent histological staining (Fig. 6A). A remarkable infiltration of inflammatory cells in the liver was observed, mostly accumulating in cell clusters of varying sizes, indicating a rapidly developing and localized inflammatory response (Fig. 6A, lower panels). The infiltrates appeared to be rHBc priming–dependent, because few such inflammatory foci were found in liver sections from mice receiving hyd. injection of prHBV1.3 only (Fig. 6A, upper right panel); nor did mice receiving pCMV-rHBc priming but with a pCMV-βGal hyd. injection exhibit any cell clusters (Fig. 6A, upper left panel).

Figure 6.

Analysis of liver infiltrating lymphocytes. (A) Histological analysis of liver sections taken 4 days after hyd. injection. Hematoxylin-eosin staining of liver sections from a representative HLA-A2/DR1 mouse receiving pCMV-rHBc priming/hyd. pCMV-βGal (upper left panel), a mouse receiving hyd. prHBV1.3 without priming (upper right panel), and a mouse receiving pCMV-rHBc priming/hyd. prHBV1.3 (lower panels). Lower right panel is a higher magnification of an area from the lower left panel. Cell clusters of inflammatory foci are indicated with frames. Arrows indicate cells undergoing degeneration. Bars, 200 μm. ALT values of sera collected at day 4 are indicated. (B) Phenotype of intrahepatic lymphocytes taken from a representative mouse 4 days after pCMV-rHBc priming/hyd. prHBV1.3 injection. CD8+ T cells were gated for analysis of Flu+ or Flu cells following Flu-tetramer labeling (left panel). Quantification of CD69+ and CD62L+ cells was done on tetramer positive (middle panel) and tetramer negative (right panel) CD8+ T cells. (C) Functional profile of intrahepatic CD3+ CD8+ T cells. Cells were analyzed for CD107a surface marker, intracellular TNF-α (upper panels), and IFN-γ (lower panels) following ex vivo stimulation with peptides derived from the polyepitope (HIV, Flu, EBV), HBcAg (C18-27), and HBsAg (S348–357/183–171) or without stimulation.

CD8+ T cells were the principal lymphocyte population in liver infiltrates. Most of these cells, regardless of Flu specificity, were phenotyped as CD62Llow, CD69high (Fig. 6B), and CD44+, CD27+, CD45RA (data not shown), corresponding to activated or memory effector T cells. Following ex vivo stimulation by peptides derived from the polyepitope, HBcAg or HBsAg, mostly Flu-specific T effectors expanded rapidly and mainly produced IFN-γ (Fig. 6C, lower panels) and some TNF-α (Fig. 6C, upper panels). Upon Flu peptide stimulation, around 59% of CD8+ T cells were positive after surface staining with CD107a, a marker of cellular degranulation12 (Fig. 6C). We then tried to determine whether these functional T effectors could be responsible for liver injury. At day 4 after the hyd. injection, the mean alanine aminotransferase (ALT) level was 94.18 mU/mL (n = 11) in the sera of mice receiving rHBV-based immunization. By comparison, ALT levels mostly remained normal in mice receiving pCMV-rHBc priming/hyd. pCMV-βGal (mean 38.00 mU/mL [n = 4]). In the absence of peripherally primed T cells, mice exhibited a strong HBsAg expression after prHBV1.3 hyd. injection, as examined in both the liver and sera (Fig. 7A, right panel, and Fig. 7B). In contrast, HBsAg was undetectable in liver sections from mice receiving rHBV-based immunization (Fig. 7A, left panel). A dramatic drop in HBsAg levels in sera (>100-fold) was also observed in the corresponding mice (Fig. 7B). Overall, these experiments clearly suggested a rapid noncytolytic control of rHBV gene expression by activated CD8+ T effectors without marked cytolytic injury.

Figure 7.

Noncytolytic control of rHBV-based HBsAg expression. (A) Immunostaining of HBsAg on liver sections taken 4 days after hyd. injection with prHBV1.3 from mice that were either primed with a pCMV-rHBc intramuscular injection (left panel) or unprimed (right panel). Bars, 100 μm. (B) Mean level of HBsAg (μg/mL) in sera of mice as in (A) with pCMV-rHBc priming (white bar) or without priming (gray bar) before hyd. injection of prHBV1.3.

Active Immunotherapy in HBsAg Transgenic Mice.

As a surrogate model for HBV chronic infection, we applied the rHBV-based active immunotherapy protocol to a transgenic mouse lineage expressing HBV envelope proteins in the liver and secreting HBsAg particles in sera.8, 20 This lineage was further back-crossed with HLA-A2/DR1 transgenic mice and was devoid of mouse major histocompatibility complex class I and class II molecules. Three groups of mice were used as controls (Fig. 8A). Following priming and hyd. injection, the mice were bled weekly to monitor serum HBsAg and ALT levels. An initial decrease in serum HBsAg levels was observed 2 weeks after priming (Fig. 8A), which may be related to DNA immunization. A second marked decrease was revealed in all pCMV-rHBc–primed mice 1 week after hyd. injection of prHBV1.3. This decrease in HBsAg could even reach around 90% of the initial level. Although HBsAg clearance was not complete (antigen levels fluctuated around 20% of baseline during 2 months of follow-up), this effect remained strong and sustained. In contrast, for mice that received either pCDNA3.1 i.m./prHBV1.3 hyd. or pCMV-rHBc i.m./pwtHBV hyd. injection, a strong increase in HBsAg level was observed. Expression of HBsAg by prHBV1.3 in the absence of pCMV-rHBc priming (Fig. 7B) or by pwtHBV plasmid accounted for this increase (Supporting Fig. 1). The rHBV-based active immunization protocol was therefore the only procedure able to suppress HBsAg at low levels after hyd. injection.

Figure 8.

Control of HBsAg expression in HBsAg/HLA-A2/DR1 transgenic mice. (A) Follow-up of HBsAg levels in sera of HBsAg/HLA-A2/DR1 transgenic mice after either an i.m. injection of pCDNA3.1 plasmid at week 0 followed by a hyd. injection of prHBV1.3 (dotted bars [n = 8]) or priming by i.m. injection of pCMV-rHBc at week 0 followed by hyd. injection of prHBV1.3 (black bars [n = 7]), pwtHBV (striped bars [n = 9]), or pCDNA3.1 (white bars, n = 7) 2 weeks later. Mice were bled weekly and HBsAg levels were quantified using the Monolisa HBsAg detection kit. Results were expressed as percentage HBsAg decrease or increase at different time points. The mean HBsAg levels at week 0 (1,001 ± 138 ng/mL; range, 136-3,164 ng/mL) were considered to be 100% in order to calculate the percentage change. Results are expressed as the mean ± SEM. P values for comparison of each group are indicated. *P < 0.05. **P < 0.01. ***P < 0.001. (B) Mean levels (mU/mL) of serum ALT (sALT) in the four groups of mice as in (A), examined at weeks 3 and 4 (7 and 14 days after hyd. injection).

Serum ALT levels were also evaluated 1 and 2 weeks after hyd. injection (Fig. 8B). A slight but significant transient increase in serum ALT levels was observed in HBsAg-transgenic mice following hyd. injection of prHBV1.3 in the absence of pCMV-rHBc priming compared with the three other groups of mice, for which serum ALT levels remained within normal values. These experiments suggest that rHBV-derived functional T cell response not only controls the recombinant virus-derived HBsAg expression as shown in Fig. 7, but also controls HBV transgene expression in the liver without major pathological implications.


It is noteworthy that patients with chronic HBV infection do not exhibit global immune deficiency. Indeed, their antiviral immunity to other invading pathogens usually remains intact, although some functional skewing was reported recently.21 We designed a novel therapeutic approach based on activating a non–HBV-specific CD8+ T cell response that could substitute for the deficient HBV-specific antiviral immunity and target infected liver cells and would not be subjected to functional exhaustion during chronic hepatitis B.

A recombinant HBV virus was designed as both gene delivery vector and antigenic carrier, enabling the intrahepatic expression of a foreign antigenic polyepitope. An ideal vector for gene therapy is one that is likely to target abnormal cells without being harmful to healthy neighboring cells. We created the rHBV virus with viral core gene disrupted by insertion of the foreign antigenic sequence. This core-deficient virus is not competent for replication unless the host hepatocyte affords the wild type viral capsid protein in trans. One meaningful finding of the study was that core-rescued rHBV exhibited even stronger and more competitive intracellular replication than wtHBV. This probably resulted from an increase in the yields of the pol gene product. rHBV will probably infects human hepatocytes specifically, as does the wild-type virus. However, rHBV cannot replicate in healthy recipients, but is comaintained in chronic patients, and only in hepatocytes with persistent wtHBV infection. Moreover, once the natural HBV infection is eliminated, the rHBV will subsequently terminate its pseudo viral life. Interestingly, the rHBV in our study was similar to a naturally occurring HBV variant (ΔC-144) identified by Will and collegues.22 This variant produced 2- to 4.5-fold more progeny DNA than wtHBV when sufficiently complemented with wild-type core protein. In addition, rHBV has a short viral genome that favors pgRNA packaging. It could thus be expected that rHBV would dominate within the covalently closed circular DNA pool in the cell nucleus and compete for the replication of wtHBV.

For anti-HBV immunotherapy to be efficient, the activated T cell response must specifically attain the liver. Because mice are not susceptible to HBV infection, we used a hyd. injection of the plasmid that enabled the direct liver expression of rHBV in mouse hepatocytes as a surrogate for rHBV infection. Although it may represent an artificial procedure compared to natural infection, this method was previously used to set up a model of acute HBV infection and was well characterized regarding the kinetics of viral replication, potential liver damage and immune induction.6 In the presence of HBV liver-specific promoters, rHBV genes and the foreign antigenic polyepitope could be expressed, processed and presented as peptides on mouse liver cells, subsequently causing a vigorous intrahepatic T cell response. In view of the peculiar features of liver immunity, we developed a protocol of active immunization in the mouse, which combined peripheral priming and rHBV-based hyd. injection. DNA immunization by an intramuscular injection ensured that the mouse would have a strong functional peripheral CD8+ T cell response against the foreign polyepitope. The subsequent hyd. injection of rHBV could both attract and strengthen the T cell response to the liver. And indeed, most cell infiltrates proved to be activated functional CD8+ T cell effectors with specificity for the Flu-epitope as early as 4 days after hyd. injection. T cells of other specificities, including HBV core, might have reached the liver and thus participated in viral clearance as well. However, we did not detected any HBsAg-specific T cells in the liver. In the presence of well-developed specific immunity in vivo, memory or activated effector T cells could be reactivated rapidly by a second encounter with the invading pathogen in the liver. Cytokine-mediated noncytolytic mechanisms play a major antiviral role, as has been demonstrated elegantly in animal models.23 Although the invading pathogen is rapidly restricted, it cannot cause more widespread cytolytic damage in the liver. In our study, despite the influx of CD8+ T cells in the liver, most mice remained without any symptoms suggestive of severe liver damage.

There is currently no mouse model for chronic HBV infection. During our study, we used HBsAg/HLA-A2/DR1 triple-transgenic mice to mimic the persistent antigen production. These mice have an HLA-A2/HLA-DR1 background and express a high level of HBsAg in the liver that is secreted and remains present in sera throughout the animal's life. Injection of a DNA plasmid encoding HBsAg induced a high frequency of CD8+ T cells secreting IFN-γ in the periphery, with in vitro cytolytic activity and specificity to HLA-A2–restricted epitopes within the viral envelope antigen. However, the HBs-specific CD4+ T cell responses were tolerized in this mouse model, and HBsAg-based DNA immunization was shown to be inefficient in inhibiting HBsAg expression.9 In contrast, during the current study rHBV-based active immunization triggered functional foreign antigenic T cell responses that remarkably abrogated HBsAg expression in this mouse model.

The liver is a peculiar organ from an immunologic point of view.24, 25 Nevertheless, an increasing number of studies have demonstrated the presence of fully functional effector T cells in the liver.26, 27 Virus-specific T cells may be retained in the liver even though the virus is not known to infect this organ, and they can cause resulting bystander hepatitis even when hepatocytes do not express the antigen recognized by the T cells.27–29 This was also evidenced in our study of HBsAg transgenic mice. First, the peripherally activated T cell response, which mostly comprised Flu-specific T cells but no HBsAg-specific T cells, may have partially reduced HBsAg production. Because this HBsAg transgenic mouse cannot achieve complete viral replication due to the absence of the core gene, the inhibitory effect is probably due largely to cytokine-mediated bystander effects, through transcriptional regulation, as suggested by previous studies.8, 20 Secondly, after the hyd. injection of rHBV, a strong inhibitory effect on viral transgene expression was observed. This effect was long-lasting in mice that received rHBV-based active immunization only but not in mice in which the wtHBV was injected. This underlines the importance of expression of the polyepitope in the liver. In addition, pCMV-rHBc priming contributed to HBsAg decrease at later time points. This might be related to the liver influx of polyepitope-specific effector T cells through the bloodstream after pCMV-rHBc immunization. Several mechanisms are involved in cytokine-mediated anti-HBV effects, functioning at hierarchical levels during the viral replication.23 In patients with chronic infection, rHBV will infect hepatocytes and comaintain with wtHBV in the liver. Because rHBV and wtHBV share the same replicative cycle, we could reasonably expect that rHBV-based T cell responses might exert the anti-HBV functions in a broader and even more efficient way. However, we cannot exclude that rHBV could be maintained for a long time if introduced into a chronic HBV patient. This may theoretically result in exhaustion of polyepitope-specific T cells.

Overall, rHBV-based active immunization represents a novel strategy for the treatment of persistent HBV infection, and according to our findings could be efficient and feasible. The current concept of a foreign antigenic T cell response could even be more generally extended to other chronic viral diseases.


We thank D. Cougot, P. Soussan, and L. Fiette for technical advice on southern blot hybridization and histological assays, Yu Wei for critical discussions, and V. Deubel for support. pwtHBV was kindly provided by Dr. Margherita Melegari. Flu and EBV tetramers were provided by Dr. Immanuel Luescher. HCV tetramer and pentamer were provided by Dr. Yves Rivière.