Induction or expansion of T-cell responses by a hepatitis B DNA vaccine administered to chronic HBV carriers


  • Maryline Mancini-Bourgine,

    1. Carcinogénèse Hépatique et Virologie Moléculaire/Institut National de la Santé et de la Recherche Médicale Unité 370, Institut Pasteur, Paris, France
    Search for more papers by this author
  • Hélène Fontaine,

    1. Service d'Hépatologie, Hôpital Necker Enfants Malades, Paris, France
    Search for more papers by this author
  • Daniel Scott-Algara,

    1. Unité de Biologie des Rétrovirus, Institut Pasteur, Paris, France
    Search for more papers by this author
  • Stanislas Pol,

    1. Service d'Hépatologie, Hôpital Necker Enfants Malades, Paris, France
    Search for more papers by this author
  • Christian Bréchot,

    1. Carcinogénèse Hépatique et Virologie Moléculaire/Institut National de la Santé et de la Recherche Médicale Unité 370, Institut Pasteur, Paris, France
    2. Service d'Hépatologie, Hôpital Necker Enfants Malades, Paris, France
    Search for more papers by this author
  • Marie-Louise Michel

    Corresponding author
    1. Carcinogénèse Hépatique et Virologie Moléculaire/Institut National de la Santé et de la Recherche Médicale Unité 370, Institut Pasteur, Paris, France
    • CHVM/INSERM U 370, Département de Médecine Moléculaire, Institut Pasteur, 25-28 rue du Dr. Roux, 75015 Paris, France
    Search for more papers by this author
    • fax: (33) 1-45-68-89-43


Despite the availability of effective hepatitis B vaccines for many years, over 370 million people remain persistently infected with hepatitis B virus (HBV). Viral persistence is thought to be related to poor HBV-specific T-cell responses. A phase I clinical trial was performed in chronic HBV carriers to investigate whether HBV DNA vaccination could restore T-cell responsiveness. Ten patients with chronic active hepatitis B nonresponder to approved treatments for HBV infection were given 4 intramuscular injections of 1 mg of a DNA vaccine encoding HBV envelope proteins. HBV-specific T-cell responses were assessed by proliferation, ELISpot assays, and tetramer staining. Secondary end points included safety and the monitoring of HBV viraemia and serological markers. Proliferative responses to hepatitis B surface antigen were detected in two patients after DNA injections. Few HBV-specific interferon γ–secreting T cells were detectable before immunization, but the frequency of such responses was significantly increased by 3 DNA injections. Immunization was well tolerated. Serum HBV DNA levels decreased in 5 patients after 3 vaccine injections, and complete clearance was observed in 1 patient. In conclusion, this study provides evidence that HBV DNA vaccination is safe and immunologically effective. We demonstrate that DNA vaccination can specifically but transiently activate T-cell responses in some chronic HBV carriers who do not respond to current antiviral therapies. Supplementary material for this article can be found on the HEPATOLOGYwebsite ( (HEPATOLOGY 2004;40:874–882.)

Despite the availability of safe, effective vaccines, hepatitis B virus (HBV) remains a major cause of human disease worldwide, with an estimated 370 million chronic carriers representing the sole reservoir of the virus.1 Chronic HBV carriage may result in cirrhosis and hepatocellular carcinoma, with high mortality rates.

Interferon (IFN)-α treatment significantly decreases HBV replication in only one third of patients with hepatitis B e antigen (HBeAg)-positive chronic active hepatitis B.2 Nucleoside analogues, such as lamivudine and adefovir dipivoxyl, inhibit HBV replication and improve histological signs of liver disease,3, 4 but their use is limited by the risk of relapse after treatment discontinuation and the emergence of drug-resistant viral variants.5, 6 These potentially useful treatments are nonetheless only virostatic, and the ultimate goal is to abolish viral replication and to eliminate residual infected hepatocytes.

Patients with acute self-limited hepatitis B display detectable polyclonal and multispecific cytotoxic T lymphocyte (CTL) and T helper (Th) responses to viral antigens, whereas these responses are weak or absent in chronic HBV carriers.7, 8 Moreover, Th1 cytokine production following antigen stimulation of T cells in periphery is reduced or not detectable in chronic hepatitis in contrast to what is seen in acute hepatitis B.9–11 Strong, multispecific T-cell responses have been reported in patients recovering after IFN-α treatment12 and in some lamivudine-treated patients displaying a decrease in viral load.13, 14 Increasing the strength of HBV-specific T-cell responses to the levels found in patients recovering from infection is therefore a goal in the treatment of patients with chronic hepatitis. Therapeutic vaccine strategies aimed at inducing or increasing HBV-specific CD8+ and CD4+ T-cell responses have been tested in chronic HBV carriers. Immunization with a lipopeptide-based vaccine containing a CTL epitope derived from the HBV core antigen induced strong CTL responses in healthy patients but was not associated with viral clearance in chronic carriers.15, 16 Vaccine therapy by standard anti-HBV vaccination has been shown to be both effective and limited.17, 18

Immunization with a nucleic acid vaccine usually elicits antibody responses and T lymphocytes with a Th1 cytokine profile.19 In animal models of chronic hepatitis B infection, including nonhuman primates, intramuscular injection of a plasmid encoding HBV envelope proteins induces rapid, strong, and sustained humoral and cell-mediated immune responses.20–22 Clinical trials of DNA vaccines for hepatitis B conducted in healthy adult volunteers using a plasmid encoding hepatitis B surface antigen (HBsAg) and the gene gun as a delivery system showed good tolerance.23, 24 Although the potential immunogenicity of a DNA-based vaccine has been demonstrated in asymptomatic patients infected with human immunodeficiency virus,25 DNA vaccines have not been used therapeutically in chronic HBV carriers. We therefore carried out a phase I trial of a HBV DNA vaccine in patients with chronic active viral hepatitis, aiming to restore HBV-specific immune responses and to assess safety regarding liver disease.


HBV, hepatitis B virus; IFN, interferon; HBeAg, hepatitis B e antigen; CTL, cytotoxic T lymphocyte; Th, T helper; HBsAg, hepatitis B surface antigen; S, small protein of the HBV envelope; preS2, N-terminal domain of middle protein of the HBV envelope; HBcAg, hepatitis B core antigen; TT, tetanus toxoid; PBMC, peripheral blood mononuclear cell; HLA, human leukocyte antigen; IL, interleukin; ELISpot, enzyme-linked immunosorbent spot; rVV, recombinant vaccinia virus.

Patients and Methods

Nine HBeAg-positive chronic HBV carriers and 1 HBeAg-negative patient infected with a precore HBV mutant were enrolled. All of the patients were male, with a median age of 43 years (range 20–58), biopsy-proven chronic hepatitis, active HBV replication (documented for a period of at least 6 months), and no decompensated liver disease. All were long-term HBV carriers, mostly contaminated during childhood, and had not responded to IFN-α and/or lamivudine therapy (Table 1). Exclusion criteria included coinfection with human immunodeficiency, hepatitis C or delta viruses, alcohol consumption of more than 40 g/d, and active intravenous drug usage.

Table 1. Clinical and Serological Characteristics of Patients
PatientPrevious Treatments (Time Interval)*HLA-AHistologyHBV Serology
Before Vaccination (M0)After Vaccination (M15)
ALTHBeAg/Anti-HBeHBV DNA (pg/mL)ALTHBeAg/Anti-HBeHBV DNA (pg/mL)
  • NOTE. Necroinflammatory activity (A) and fibrosis (F) were semiquantitatively assessed and ranked by calculating Metavir score in the 6 months preceding inclusion. Alanine aminotransferase (ALT) concentration is expressed as a multiple of the upper limit of the normal range.

  • Abbreviation: M, months.

  • *

    Time interval is expressed in months from cessation of virological treatment or development of lamivudine resistance to first DNA injection.

  • HBV DNA was repeatedly detected before inclusion.

  • For patients receiving adefovir dipivoxyl treatment, values after vaccination are given before adefovir treatment.

HB13Lamivudine (9 M)A1-33A1F10.5−/+<2.50.6−/+16
HB14IFN-α–Lamivudine (14 M)A1-31A1F10.8+/−5520.9+/−1,195
HB16IFN-α–Lamivudine (25 M)A2-33A1F30.9+/−1741.9+/+2 (M8)
HB17Lamivudine (20 M)A24A2F40.9+/−8040.4+/−836 (M10)
HB18IFN-α–Lamivudine (4 M)A2-33A1F23.5+/−15,4284.3+/−5,380 (M12)
HB21Lamivudine (6 M)A29-31A1F10.8+/−100.6−/+<2.5
HB22IFN-α–Lamivudine (38 M)A3A1F21.9+/−1,3485.0+/−4,493 (M8)
HB23Lamivudine (2 M)A32A1F12.0+/−9021.3+/−1,384
HB24IFN-α–Lamivudine (13 M)A1-25A1F22.0+/−3,9682.8+/−1,175
HB31IFN-α–Lamivudine (19 M)A2-3A1F11.4+/−6572.5+/−348

We injected 0.5 mg of DNA vaccine simultaneously into each deltoid muscle at months 0, 2, and 4. Six patients received an additional injection at month 10. By decision of investigators and after its availability, adefovir dipivoxyl was given to 4 patients (see Table 1). This study was approved by the ethics committee of Necker Hospital, and all study participants gave informed, written consent for participation in line with French ethical guidelines.

DNA Vaccine.

The pCMV-S2.S DNA vaccine26 was produced under GMP conditions (Qiagen GmbH, Hilden, Germany). This plasmid encodes the small (S) and middle (preS2 + S) proteins of the HBV envelope (ayw subtype)—hereafter referred to as HBsAg—under control of the human cytomegalovirus promoter and has been shown to induce anti-HBs and anti-preS2 antibodies in mice and chimpanzees.27, 28 HBV-specific IFN-γ–secreting T cells have been found in mice, including HLA-A∗02 transgenic mice,29 after pCMV-S2S injection. DNA was formulated in endotoxin-free 0.9 % NaCl to give a dose of 1 mg/mL.

HBV Antigens and Synthetic Peptides.

Recombinant HBsAg particles were produced in Chinese hamster ovary cells (Aventis Pasteur, Val de Reuil, France; Michel et al.30). A recombinant preparation of full-length ayw subtype hepatitis B core antigen (HBcAg) purified from Escherichia coli was provided by Darrell L. Peterson, Department of Biochemistry, Commonwealth University, Richmond, VA. Purified tetanus toxoid (TT) (Aventis Pasteur, Marcy l'étoile, France) was used as a positive control in proliferation assays. Three peptides (HBs109–128, HBs124–148, HBs139–163) covering the entire preS2 and 26 overlapping 15-mer peptides covering the S sequence (amino acids 164–389) were used as pools to stimulate human peripheral blood mononuclear cells (PBMCs). HBs-derived peptides29 and one irrelevant control peptide known to bind human leukocyte antigen HLA-A∗02 major histocompatibility complex class I molecules were synthesized (Neosystem, Strasbourg, France). Peptides were numbered from the first methionine of the HBV ayw subtype preS1 domain. A promiscuous Th peptide from the core protein recognized by most of the chronic carriers (HBc50–69, TPPATRPPNAPIL) was used as an internal control.31

Virological Assessment.

HBsAg, HBeAg, anti-HBs, and anti-HBe antibody levels were determined with commercial enzyme immunoassay kits (Biorad, Marnes la Coquette, France). HBV DNA was quantified using branched DNA Versant HBV 1.0 (Bayer Diagnostics, Puteaux, France). The detection limit was estimated to be 2.5 pg/mL or approximately 700,000 viral genome equivalents. Anti-preS2 antibodies were determined via specific enzyme-linked immunosorbent assay as described previously.17

Isolation of PBMCs and Proliferation Assay.

For immunological studies, blood was taken at months −2, 0, 1, 3, 5, 10, 11, and 15. PBMCs were recovered from heparin-treated blood via Ficoll-Hypaque density gradient centrifugation. Freshly isolated PBMCs were resuspended in complete medium (RPMI 1640 medium, Life Technologies, Gaithersburg, MD) supplemented with 2 mmol/L L-glutamine, 1 mmol/L sodium pyruvate, nonessential amino acids, 100 U/mL penicillin, 100 μg/mL streptomycin, and 10% human AB serum (Institut Jacques Boy, Reims, France). PBMCs were cultured in triplicate (1.5 × 105 cells/well) in 96-well round-bottom microplates for 4 days in the presence of HBsAg (3 μg/mL), HBcAg (1 μg/mL), TT (10 μg/mL), or medium alone. On day 4, cultures were labeled by incubation for 8 hours with 1 μCi 3H-thymidine/well (specific activity 25 μCi/mmol/L; Amersham, Bucks, United Kingdom). The proliferative response was evaluated by determining 3H-thymidine uptake with a beta counter (LKB/Pharmacia, Uppsala, Sweden). The stimulation index was calculated as the ratio of the number of counts per minute in the presence and absence of antigen.

In Vitro Expansion of PBMCs.

Freshly prepared or frozen PBMCs were incubated with complete medium plus 20 ng/mL interleukin (IL)-7 (Roche Diagnostics, Meylan, France) in 24-well plates (6 × 106 cells/well). PBMCs from 5 unvaccinated patients with chronic active hepatitis B (already described by Pol et al.18) were used as controls. Cells were stimulated via incubation for 3 days with 10 μg/mL HLA-A*02–restricted peptides or with 3 pools of 15-mer peptides derived from the S region (1 μg/mL of each peptide), a pool of 3 preS2 peptides, and a promiscuous core-derived Th epitope (each at 10 μg/mL). Half the medium was replaced every 3–4 days with complete medium supplemented with recombinant IL-2 (50 IU/mL) (Roche Diagnostics). After 2 weeks of culture, IFN-γ–producing cells were specifically quantified via enzyme-linked immunosorbent spot assay (ELISpot).

ELISpot Assay.

Sterile nitrocellulose HA 96-well plates (Millipore, Bedford, MA) were coated with 15 μg/mL anti–IFN-γ monoclonal antibody (1-DIK; Mabtech, Stockholm, Sweden) or 7.5 μg/mL anti–IL-5 monoclonal antibody (TRFK5; BD Biosciences, Le Pont de Claix, France) in 50 μL 0.1 mol/L bicarbonate buffer (pH 9.6) overnight at 4°C. Wells were blocked with 200 μL of 5% human AB serum in phosphate-buffered saline at room temperature for 2 hours. The coated wells were filled in triplicate with in vitro–stimulated cells (2 × 105 cells/well) in complete medium and the appropriate peptides (1 μg/mL). For the determination of ex vivo CD8+ T cell frequencies, 5 × 105 freshly isolated PBMCs were added to each well with various recombinant vaccinia viruses (rVV) encoding the large HBV envelope protein (rVVS1.S2.S) or the core protein (rVVHBc) at a multiplicity of infection of 20. Cells in culture medium or infected with wild-type vaccinia virus were used as negative controls to evaluate background.

After 20 hours of incubation at 37°C, plates were washed and incubated with 50 μL of 1 μg/mL biotinylated anti–IFN-γ monoclonal antibody (7B6-1; Mabtech, Stockholm) or biotinylated anti–IL-5 monoclonal antibody (JES1-5A10; BD Biosciences) for 2 hours at room temperature. Plates were then washed and antibody binding detected as described previously.29 A Zeiss ELISpot automated counter was used to score the number of spots. The response was considered positive if the median number of spot-forming cells in triplicate wells was at least twice that in control wells containing irrelevant peptide and at least 5 spots were detected per 2 × 105 PBMCs after background subtraction.

Flow Cytometry Detection of T Cells After Tetramer or Intracellular Staining.

The HLA-A*02 tetrameric complexes were obtained from Proimmune (London, United Kingdom). The major histocompatibility complex–peptide complexes were formed with two HBV immunodominant epitopes from envelope protein (HBs183–191 and HBs348–357). CD8 tetramer binding cell counts were performed on fresh whole blood samples, and standard lysis procedure of red blood cells was performed with four-color flow cytometry analysis using the following associations: CD3-ECD, CD8-PCy5, CD16-FITC, and tetramer-PE (Beckman-Coulter-Immunotech, Paris, France). Event accumulations were followed up until they reached 50,000 living cells on the lymphocyte gate. The cut-off value was 0.07% obtained with HLA-A*02 tetramer loaded with irrelevant peptide.

For intracellular cytokine staining experiments, in vitro–expanded PBMCs were cultured overnight in the presence or absence of peptide (1 μg/mL) and Brefeldin A (2 μg/mL). After washing, cells were stained with anti–CD8-PerCP– and anti–CD4-PE–conjugated antibodies and washed again. Cells were fixed, permeabilized in the presence of anti–IFN-γ–FITC (BD Biosciences) to allow flow cytometry analysis of peptide-specific T cells.

Statistical Analysis.

Numbers of IFN-γ–producing T cells derived from vaccinated or unvaccinated patients were compared using the Mann-Whitney U test. Forty-nine samples were considered for vaccinated patients, while 20 samples were considered for unvaccinated patients.


Th Cell Proliferation.

We assessed the ability of PBMCs from chronically infected individuals to proliferate in response to the HBV envelope proteins (HBsAg) encoded by the DNA vaccine, HBcAg, or recall antigen TT (Fig. 1). None of the patients displayed a proliferative response to HBsAg before the first injection. At the end of the vaccination protocol, 2 of the 10 patients responded to this antigen (see Fig. 1A, HB18, HB21). In patient HB21, this response increased dramatically between months 9 and 11.

Figure 1.

Longitudinal analysis of proliferative responses in the peripheral blood of chronic hepatitis B patients during the vaccination protocol. PBMCs were incubated in the presence of (A) HBsAg, (B) HBcAg, or (C) TT. Proliferative responses were analyzed before vaccination and 1 month after each pCMV.S2.S injection (open arrows). The dark arrows in panel C indicate revaccination of patients HB13, HB14, and HB16 against tetanus. The results are expressed as stimulation index. Stimulation index of less than 2.1 for HBsAg and TT and of less than 6 for HbcAg was considered insignificant (shaded area). Several curves of interest are indicated. SI, stimulation index; HBsAg, hepatitis B surface antigen; HBcAg, hepatitis B core antigen; TT, tetanus toxoid.

We also investigated whether vaccination with the vector encoding HBV surface proteins affected the proliferative response to the nucleocapsid protein, which is produced during HBV replication. Before immunization, a T-cell proliferative response to HBcAg was detected in only 1 patient (see Fig. 1B). During the vaccination protocol, seven other patients showed transient proliferative responses to HBcAg. The patient with the strongest HBs-specific proliferative response also displayed a strong response to core antigen. This HBc-specific T-cell response was detectable transiently 2 months before the HBs-specific response. All patients displayed significant T-cell responses to TT throughout the study. The intensity of the response was independent of pCMV-S2.S DNA vaccination schedule. During the trial, 3 patients (HB13, HB14, HB16) were revaccinated against tetanus at months 2, 4, and 5, respectively, and TT-specific T-cell response increased following these injections (see Fig. 1C). Thus, specific T-cell recall responses to other antigens may occur in chronic HBV carriers.

Induction of IFN-γ–Secreting T Cells.

We characterized HBV-specific CD8+ effector T cells in vaccinated patients by means of an ex vivo IFN-γ ELISpot assay using freshly isolated PBMCs mixed with rVVs expressing various HBV antigens.29, 32 Significant responses to HBs- and HBc-expressing rVVs were detected only in patient HB18, 1 month after the first injection and 1 month after the fourth injection (Fig. 2). The frequency of IFN-γ–secreting T cells was found to be similar in PBMCs from a HBV-infected patient with clinical features suggestive of a spontaneous disease flare-up (0.012% for HBs-specific T cells and 0.017% for HBc-specific T cells). In contrast, no IL-5–secreting T cells were detected in any patient during the protocol.

Figure 2.

Quantification of IFN-γ–secreting T cells in patient HB18 upon stimulation with HBV antigens. Ex vivo ELISpot assays were carried out on freshly isolated PBMCs infected with wild-type vaccinia virus and with recombinant vaccinia viruses encoding large HBV envelope protein (rVVS1.S2.S) or HBV core protein (rVVHBc). Results are expressed as the number of IFN-γ–secreting T cells/106 PBMCs. IFN-γ, interferon γ; PBMCs, peripheral blood mononuclear cells; VVWT, wild-type vaccinia virus; rVV, recombinant vaccinia virus; M, month.

ELISpot assays were also used to evaluate the memory T-cell responses elicited by the vaccine. Freshly isolated PBMCs were stimulated for 2 weeks with peptides derived from the preS2 or S region of HBV. Despite possible effects of growth in vitro on the initial cytokine profile, we chose this strategy because the small number of cells and low frequency of antigen-specific T cells in chronic HBV carriers precluded the detection of IFN-γ–secreting T cells in all but 1 patient (see above) in the absence of in vitro stimulation. Before vaccine injection, no IFN-γ–secreting T cells were detected in the PBMCs of 7 of the patients (Fig. 3). HBs-specific T cells were found after the first (HB16, HB23), second (HB14, HB21, HB24), or third injection (HB13, HB17, HB31) (see Fig. 3A). However, HBs-specific responses were detected in 2 patients (HB18, HB22) before vaccination. We therefore compared the numbers of HBs-specific IFN-γ–producing T cells found in vaccinees (n = 49 samples) with those found in 5 unvaccinated control patients (n = 20 samples); a significant difference was observed between the 2 groups (P = .028). Sustained preS2-specific T cells were detected in 3 patients after vaccine injection (see Fig. 3B). However, these responses were not significantly different from those of control patients (P = .31). Four of the 7 patients responding to the vaccine (HB13, HB16, HB17, HB21) responded to 2 envelope proteins (S and preS2), and for 6 patients, responses were detected at more than one time point.

Figure 3.

Follow-up of IFN-γ–secreting T-cell responses in vaccinated patients. Number of IFN-γ–secreting T cells was determined by ELISpot after 2 weeks of in vitro stimulation. PBMCs were stimulated either with pools of (A) S peptides or (B) preS2 peptides. Data are expressed as the number of IFN-γ–secreting T cells/106 PBMCs and represent corrected values after background subtraction (PBMCs cultured with an irrelevant peptide). Responses were analyzed before vaccination and 1 month after each pCMV.S2.S injection. Patients are represented alternately by white and striped bars for reasons of clarity. IFN-γ, interferon γ; PBMCs, peripheral blood mononuclear cells; S, small protein of the HBV envelope; PreS2, N-terminal domain of middle protein of the HBV envelope.

Overall T-Cell Response.

Table 2 lists the positive T-cell responses detected by IFN-γ ELISpot assay in vaccinated patients. T-cell responses specific for preS2- and S-envelope domains were detected at baseline in 11% and 20% of patients, respectively. Following injections, the response rate increased gradually to 100%, with all patients recognizing at least one peptide derived from the S HBV envelope protein after the third injection. Responses to the preS2 HBV envelope protein increased from 11% to 50% during the same period. Thus, the DNA vaccine tested either induced the de novo production of envelope-specific IFN-γ–secreting T cells or activated HBV-specific memory T cells in chronic HBV carriers. Because only 6 patients received the fourth injection, T-cell responses in months 11–15 correspond both to responses to the fourth injection and to persistent responses to the third injection. This final injection, performed 6 months after the third injection, clearly provided no additional benefit in terms of HBs-specific T-cell activation or persistence of the induced responses. As an internal control, we used an HBc-derived Th1 peptide recognized by all HBV-infected patients. Interestingly, the frequency of patients with core-specific Th cell responses also increased during follow-up—from 30% at baseline to 60%—after the third injection.

Table 2. Induction of T-Cell Responses During the Vaccination Protocol
MonthNumber of Positive Responders (%)
Core 50–69PreS2S
  • *

    Number of patients with T-cell responses from month 11–15 receiving either 3 or 4 injections.

 03/10 (30%)1/9 (11%)2/10 (20%)
 12/9 (22%)3/7 (43%)2/9 (22%)
 32/10 (20%)4/9 (44%)5/10 (50%)
 53/7 (43%)4/8 (50%)5/8 (63%)
 5–106/10 (60%)5/10 (50%)10/10 (100%)
11–15*1/10 (10%)2/10 (20%)4/10 (40%)
 3 injections0/41/42/4
 4 injections1/61/62/6

Phenotype and Specificity of Vaccine-Activated T Cells.

In HLA-A*02+ patients, direct ex vivo frequency of HBV-specific CD8+ T cells was tested with HLA-A*02 tetramers specific for envelope HBs183–191 and 348–357. Patients HB16 and HB18 had HBs348–357-specific CD8+ T cells that were tetramer-positive from month 5 (Fig. 4A). Frequency increased over time from 0% (month 0) to 0.2% and 0.14% (month 11) above the background level. Staining with tetramer HBs183–191 was always found negative during follow-up. In addition, after in vitro expansion using HLA-A*02 peptides in ELISpot assay, patient HB18 responded to peptide HBs338–347 throughout the protocol, whereas patient HB16 developed an HBs370–379 CD8+ T-cell response after vaccination from months 2–13 (data not shown). For non–HLA-A*02 patients, we used 15-mer peptides, which can react with both major histocompatibility complex class I and II molecules. Most HBs-specific T cells were CD4+ T cells, and several different peptides were recognized. Intracellular IFN-γ staining of activated CD4+ T cells is shown for patient HB14, who developed an HBs-specific T-cell response from months 3–15 (Fig. 4B).

Figure 4.

Characterization of HBV-specific CD8+ and CD4+ T cells by flow cytometry. (A) Quantification of HBs-specific CD8+ T cells by using HLA-A∗02/HBs348–357 tetramer. Dot plots of PBMCs from patients HB16 and HB18 stained with HLA-A∗02 tetramer and anti-CD8 antibody are shown at the indicated time points. (B) Phenotyping of IFN-γ–producing T cells expanded from PBMCs taken at month 15 of a representative patient (HB14), cultured in the absence (upper panel) or presence (lower panel) of HBs184–198 peptide and stained with anti–IFN-γ–FITC and anti–CD4-PE monoclonal antibodies. M, month; IFN-γ, interferon γ; FITC, fluorescein isothiocyanate.

Clinical Outcome.

Although this trial was not designed to evaluate the clinical efficacy of the vaccine, we determined serum HBV DNA and aminotransferase levels throughout the study to follow the kinetics of HBV replication and liver disease. Four patients displayed decreases in viral DNA of 48%–83% after 3 injections of pCMV-S2.S (HB17, HB18, HB24, HB31) (two of them are shown in Fig. 5). However, this decrease was not sustained, because HBV DNA levels returned to baseline values before month 10 (see Supplementary Fig. 1). Patient HB16 displayed a sharp decrease (>88% at month 8) in HBV DNA levels but underwent an adefovir therapy from month 9. A complete clearance was observed in patient HB21 after 3 vaccine injections, and the virus remained undetectable up to month 15. For patients HB16 and HB21, HBV DNA clearance was associated with a 16-fold and 11-fold increase in transaminase levels, respectively, and with the detection of anti-HBe antibody at month 6 and month 15, respectively (see Fig. 5). PreS2 antibodies were also found in these 2 patients, from month 11 and month 5 onwards, respectively (data not shown). Sustained tetramer-positive staining was found after the third vaccine injection for 2 HLA-A*02+ patients (HB16, HB18). IFN-γ–secreting T-cell responses peaked after vaccine injections (see Fig. 5) but, because of the low number of patients enrolled in this trial, no correlation between the induced T cell responses and clinical improvement could be made.

Figure 5.

Longitudinal analysis of HBV DNA load and serum alanine aminotransferase in relation to HBV-specific immune responses. HBV DNA (bars) level is expressed as a percentage of the maximal recorded HBV DNA concentration for each patient. Alanine aminotransferase levels are expressed as a multiple of the upper limit of normal range (open triangles). HBs- and preS2-specific IFN-γ–secreting T cells are expressed as in Fig. 3 (filled squares). HBeAg/anti-HBe serological markers are indicated at the top of each panel. Positive HBs-specific proliferative response and tetramer staining are displayed below the graphs. DNA injections are indicated by arrows. The duration of adefovir therapy is indicated by the shaded areas. ALT, alanine aminotransferase; XN, fold upper limit of normal; IFN-γ, interferon γ; HBsAg, hepatitis B surface antigen.

Vaccine Safety.

All individuals tolerated each dose of the vaccine well, with only mild adverse events observed. No clinically significant serological abnormalities were observed. However, serum alanine aminotransferase levels transiently increased to more than 10 times baseline values in 2 patients, possibly reflecting a disease flare-up or the destruction of hepatocytes due to the immune response after vaccination. These 2 patients (HB16, HB21) withdrew from the study before the fourth DNA injection, and one of them (HB16) received adefovir dipivoxyl when it became available.


Active immunotherapy to induce strong, sustained HBV-specific T-cell responses is a promising strategy for the treatment of patients with chronic hepatitis B infection. Advances in our understanding of the characteristics of the immune response to HBV in subjects recovering from hepatitis spontaneously or after IFN treatment and in patients with chronic hepatitis B infection or who are treated with antiviral drugs suggest that the restoration of efficient HBV-specific immunity is feasible.33

Vaccines based on HBV envelope antigens were recently shown to reduce HBV replication in humans, chimpanzees, and other animal models of hepatitis B.18, 21, 22, 34–36 Recent results obtained with DNA-based vaccines in human trials and animal models led us to evaluate the use of such vaccines to stimulate the strong Th1 immune cell response correlated with clearance and recovery after HBV infection. However, potential problems related to the generation of an overly strong T-cell response must be addressed in a disease in which both liver injury and viral control are mediated by the immune response. We therefore designed this phase I clinical trial of vaccine therapy in chronic HBV carriers with two major aims: to monitor changes in immune response induced by vaccination and to evaluate the safety of a HBV DNA vaccine in HBV-infected subjects.

The primary aim of this study was to determine whether the vaccination of chronic HBV carriers with a DNA-based vaccine induced a measurable immune response. Proliferative lymphocyte responses to HBsAg absent at the beginning of the study were detected in 2 patients during follow-up. In patient HB21, this response increased strongly 6 months after the last DNA injection, when HBV DNA levels were below the detection threshold. Up to 0.2% of envelope-specific CD8+ T cells able to bind HLA-A*02/peptide tetrameric complexes were found ex vivo in 2 patients after vaccine injections, whereas they were undetectable before the vaccine trial (see Figs. 4 and 5). These frequencies were comparable or even higher than those found in PBMCs from resolved acute HBV-infected patients tested more than 6 months postinfection (0.04%–0.11% of HBV-env tetramer positive CD8+ T cells38, 37). Moreover, these cells were not detected in acute patients tested in early phase of infection after complete normalization of serum alanine aminotransferase.39

Therefore, because of the limited number of envelope-specific effector T cells present in the peripheral blood of HBV-immune subjects and chronic HBV carriers (0.02%–0.11%),37 we also performed ELISpot assays after in vitro stimulation of PBMCs taken from patients before and during therapy. This assay detects memory effector cells40 that proliferate and secrete IFN-γ in response to stimulation with peptides derived from HBV antigens. Although specific immune responses to the HBV envelope were detected before vaccination in 11%–20% of patients, numbers of specific T cells were significantly higher in patients receiving DNA injections than in unvaccinated patients. Together these results suggest that DNA vaccination is able to induce circulating effector T cells or to activate envelope-specific memory T cells in persistently infected patients. However, this induction was followed by a progressive decline in T-cell reactivity, despite a fourth DNA injection in some patients. The restoration of HBV-specific T-cell reactivity therefore appears to be transient, and this biphasic behavior of T-cell responses to HBV has been reported in previous studies of vaccination or antiviral treatment.18, 41 Lamivudine therapy has been shown to be associated with an early but transient increase in CD4+ T-cell reactivity to HBeAg and HBcAg. We found that envelope-specific CD4+ and CD8+ T cells were activated during vaccine therapy. During self-limited HBV infection, high frequency of HBV-specific memory effector CD8+ T cells were found to proliferate, to mount efficient CTL activity and to produce IFN-γ.42 On the other hand, a population of HBs-specific CD8+ T cells was found to persist in chronic hepatitis B patients despite the presence of large amounts of viral antigen.37 These cells are not fully anergic and may therefore be activated by DNA-based vaccination. The transient nature of T-cell detection in the periphery may result from their redistribution in the liver and/or exhaustion as a result of the large amounts of HBsAg persisting in the sera of long-term carriers.

As previously reported, approximately 30% of patients had significant numbers of IFN-γ–secreting T cells specific for the core Th peptide before therapy.11 This frequency doubled during vaccination. We cannot exclude the possibility of nonspecific activation caused by the CpG motifs present in the plasmid backbone, but this increase is probably due to a bystander effect of cytokines produced by vaccine-activated HBs-specific T cells. A similar transient increase in IFN-γ–producing T cells has also been observed in chronically infected patients responding to antiviral treatment.11

Intramuscular injections of a DNA vaccine encoding the small and the middle HBV envelope proteins (1 mg per injection, 3 injections at 2-month intervals) were well tolerated in all patients. A transient increase in transaminase activity was observed in 2 patients, associated with the fall in HBV DNA level. The rise in IFN-γ–secreting T cells observed before the alanine aminotransferase flare suggests that both events could be linked. Alternatively, liver disease reactivation may have occured spontaneously during chronic HBV infection. These patients were withdrawn from the study and did not receive the fourth DNA injection. Nevertheless, no induction of severe hepatitis was observed, suggesting that DNA-based vaccination could be used in patients with chronic disease without major side effects, as previously observed in animal models.20–22

Analysis of the clinical efficacy of this DNA vaccine was not the main aim of this phase I trial. Patients had not responded to approved therapies (IFN-α and lamivudine), and this negative bias makes it impossible to draw conclusions concerning efficacy. However, we did monitor HBV DNA levels and seroconversion from HBeAg to anti-HBe antibody. The series of 3 injections markedly decreased viral load in 2 patients, and transient decreases of more than 50% were observed in 4 patients. For the patient with the strongest and the most long-lasting decrease in HBV DNA level, HBeAg elimination and seroconversion to anti-HBe antibody occurred 2 months after the disease flare-up. However, the 2 patients who displayed a decrease in viral load and an increase in transaminase activity had the lowest viral loads at the beginning of the trial. Although no firm conclusions can be drawn from these 2 patients, this result is consistent with HBV-specific T-cell recovery being related to a certain threshold in HBV DNA titer.43

In conclusion, this report demonstrates the feasibility of using DNA as a therapeutic vaccine in chronic HBV carriers. No major adverse effects were observed despite stimulation of the HBV-specific cellular immune response. However, the transient nature of the restoration of HBV T-cell response suggests that optimization of the dose and vaccination strategy is required. This includes formulation of DNA with adjuvants such as aluminium phosphate; coadministration with plasmid DNA expressing cytokines, chemokines, or costimulatory molecules; and the use of prime-boost regimen.44 Nucleoside analogues cannot completely eradicate HBV infection, but a combination of potent antiviral drugs (lamivudine and/or adefovir dipivoxyl) with efficient immunomodulation (e.g., via DNA vaccination) may have a synergic effect leading to recovery from HBV infection.


The authors extend special thanks to Dr. Souad Benali and Dr. Marc Bourlière, Hôpital St. Joseph, Marseille, France; Dr. Marianne Maynard, Prof. Christian Trépo, and Prof. Fabien Zoulim, INSERM and Hôtel Dieu, Lyon, France, for recruiting patients for this study; and Françine De Felice for help in collecting samples. We thank Yves Rivière for critical reading of the manuscript. Some of this work was performed in the laboratory of Prof. Pierre Tiollais, whom we would like to thank for his interest in this study.