Breaking tolerance in hepatitis B surface antigen (HBsAg) transgenic mice by vaccination with cross-reactive, natural HBsAg variants



Processing exogenous hepatitis B surface antigen (HBsAg) of the hepatitis B virus (HBV) generates the Kb-binding S208–215 epitope 1; processing endogenous HBsAg generates theKb-binding S190–197 epitope 2. Cross-reactive CD8+ T cell responses were primed to epitope 1 but not epitope 2 when mice were immunized with natural HBsAgayw, orHBsAgadw2 variants differing within both epitopes by one or two residues. Expression of HBsAgayw from a transgene in the liver renders (HBs-tg) mice tolerant to epitope 1 of HBsAgayw. CD8+ T cells specific for epitope 1 could be primed in HBs-tg mice by HBsAgadw2; these specific CD8+ T cells cross-reacted with epitope 1 processed from the transgene-encoded HBsAgayw. The liver of vaccinated HBsAgayw transgenic mice showed severe histopathology and contained functional (IFNγ-producing), cross-reactive CD8+ T cells, and vaccinated HBs-tg mice showed reduced antigenemia. Hence, vaccination with natural HBsAg variants from different HBV sero/genotypes can prime cross-reactive, specific CD8+ T cellimmunity that breaks tolerance to HBsAg.


C57BL/6 mice


Knock out




Non-parenchymal cells


Magnetic activated cell sorting


Intrahepatic lymphocyte


Hepatitis B virus


Hepatitis Bsurface antigen

1 Introduction

Chronic hepatitis B virus (HBV) infection affects about 200–400 million people worldwide and represents one of the leading causes of liver cirrhosis and hepatocellular carcinoma 1. Control of hepatitis B virus (HBV) infection correlates with efficient induction of multispecific CD8+ T cell responses against the (pre)core and surface antigen of this virus 2. In contrast, failure to control HBV infection (leading to chronic infection) correlates with inefficient, oligoclonal CD8+ T cell responses against HBV antigens. Specific CD8+ T cells thus deliver critical antiviral effector functions that can control HBV infection. Protocols for the specific immunotherapy of chronic HBV infection thus focus on the induction or expansion of HBV-specific CD8+ T cell reactivity.

About 100–150 subunits of the small HBsAg or S protein (p24/gp27) self-assemble into porous 20–30 nm lipoprotein particles (reviewed in 3). The variability of the S protein in HBV isolates of different serotypes and genotypes is limited 4. The four stable HBsAg serotypes adw, ayw, adr and ayr map to single, allelic residue exchanges at position 122 (d/y) and 160 (w/r) situated close to the immunodominant ‘a’ determinant (residue 124–147). No biological or pathogenetic differences have been traced to infection by HBV of different serotypes 5. In contrast, increasing evidence indicates that the clinical course, the response to interferon treatment, and the long-term prognosis of infections with HBV of different genotypes differ 4, 68. It has been suggested that the course of HBV infection is critically influenced by the different immunogenicity of different HBV genotypes for CD8+ T cells, but not by the different immunogenicity of different HBV serotypes for B cells.

CD8+ T cells from H-2b mice recognize two epitopes of the S protein: the Kb-binding S208–215 epitope 1 9 and the S190–197 epitope 2 10 (Fig. 1). These epitopes are generated in different processing pathways 11, 12. Cells processing exogenous HBsAg particles present the S208–215 epitope 1 but not the S190–197 epitope 2 to T cells 9, 10. In contrast, the S190–197 epitope 2 but not the S208–215 epitope 1 is presented to T cells by cells processing endogenous HBsAg. Many details of the alternative processing pathways that generate these epitopes are unresolved 11. We have shown that some Kb-restricted CD8+ T cell responses to variant epitopes of HBsAgayw and HBsAgadw2 show cross-reactivity (Fig. 1B) 10. We extended the study to transgenic mice that constitutively express HBsAgayw in the liver 13. These mice produce large amounts of HBsAg (evident as persistent antigenemia) and are tolerant to HBsAg. They are considered preclinical models for evaluating specific immunotherapy protocols for chronic HBV infection. We show in this report that transgenic mice vaccinated with HBsAgadw2 develop cross-reactive CD8+ T cell reactivity with antiviral therapeutic potential.

Figure 1.

 HBsAg variants. (A) Amino acid sequence of the small hepatitis B surface antigen (HBsAg) ayw (1) and adw2 (2) serotype are shown. (B) HBsAg ayw- and adw2-derived, Kb-restricted epitope sequences. The epitope 1 (S208–215) was presented only by cells processing exogenous HBsAg, while epitope 2 (S190–197) was presented only by cells processing endogenous HBsAg.

2 Results

2.1 Adoptive transfer of Kb-restricted CD8+ T cell lines specific for either epitope 1, or epitope 2 induce hepatic injury in HBs-tg C57BL/6 mice

We generated short-term CD8+ T cell lines specific for epitope 1 or epitope 2 (Fig. 1B) of HBsAg from the spleen of C57BL/6 (B6) mice immunized with pCI/Sayw plasmid DNA. Within these lines >95% of the cells were CD8+, and specific IFN-γ expression could be induced in >80% of these CD8+ T cells. The adoptive transfer of 5×106 cells of these lines into congenic B6 hosts that express HBsAgayw in the liver from a transgene induced acute liver injury evident by a short but high elevation of serum transaminases (Fig. 2). Serum transaminase levels normalized 5–6 days post-transfer, and no transferred CD8+ T cells were detectable in the host at these late time points. Transfer of an equal number of polyclonal (mitogen-activated) CD8+ T blasts did not induce liver injury. Hence, (i) specific CD8+ T cells efficiently induce liver injury in HBs-tg mice (as shown previously 14); (ii) HBsAg epitopes generated by processing endogenous, or exogenous HBsAg are presented in the transgene-expressing liver; and (iii) adoptively transferred CD8+ T cells are rapidly cleared from the trans-genic host. Transferred CD8+ T cells with different specificities of HBsAg can thus access the liver and are activated in situ but cannot be stably engrafted.

Figure 2.

 Transfer of epitope 1 or epitope 2 specific CTLL into HBs-tg hosts establish transient liver damage. Spleen cells were removed from pCI/Sayw DNA immune B6 mice and restimulated in vitro with syngeneic, RBL5 cells pulsed with the Kb/S208–215 binding peptide 1 (ILSPFLPL) or the Kb/S190–197 binding peptide 2 (VWLSVIWM), or stimulated with Con A. 5x106/mouse CD8+ CTLL were injected intravenously into HBs-tg mice and mean serum ALT levels were determined.

2.2 Kb-restricted CTL recognizing HBsAg epitopes 1 and 2 are found in spleen and liver

We tested if vaccine-primed, HBsAg-specific CD8+ T cells can access the liver of normal or transgenic, HBsAg-expressing (HBs-tg) B6 mice (Fig. 3). Spleen cells and non-parenchymal liver cells (NPC) were isolated from B6 mice immunized 12–15 days previously with the pCI/Sayw DNA vaccine. CD8+ T cells specific for epitope 1 or epitope 2 were found in splenic and hepatic CD8+ T cell populations from normal B6 mice (Fig. 3A). In most experiments, mice were vaccinated with the pCI/S DNA vaccines that encode secreted small HBsAg. Similar data were obtained when mice were vaccinated with ODN-adjuvanted HBsAg particles 10. Vaccination with exogenous HBsAg (HBsAg particles) or endogenous HBsAg (encoded by a DNA vaccine) thus efficiently primes CD8+ T cell responses to epitope 1 and epitope 2. Although the frequency of HBsAg-specific CD8+ T cells was high within hepatic CD8+ T cell populations, their absolute numbers were lower than in the spleen (data not shown). In contrast, no CD8+ T cell reactivity was detectable in HBsAgayw-tg B6 mice vaccinated with the DNA vaccine encoding HBsAgayw (Fig. 3B). Neither three boost injections (at 3-weekly intervals) with this DNA vaccine, nor repeated vaccinations with oligonucleotide-adjuvanted HBsAg particles elicited HBsAg-specific CD8+ T cell immunity in HBs-tg mice (data not shown). Hence, vaccination protocols using the same HBsAg variant against which the mouse is tolerant can not prime this antiviral CD8+ T cell immunity.

Figure 3.

Ex vivo detection of HBsAg-specific CD8+ T cells in liver and spleen of vaccinated mice. B6 mice were immunized intramuscularly by a single injection of 100 μg pCI/Sayw DNA. Specific CD8+ T cells were detected 12 days post-immunization. Isolated liver MNC and spleen cells were restimulated in vitro for 4 h (in the presence of BFA) with the Kb/S208–215 binding peptide 1 (ILSPFLPL) or the Kb/S190–197 binding peptide 2 (VWLSVIWM). The mean frequency of CD8+ IFN-γ+ T cells/105 CD8+ T cells ± SD of four to six mice cells (from two independent experiments) per group are shown.

2.3 Kb-restricted T cell responses to epitopes of HBsAgayw and HBsAgadw2 variants

The 226-residue small HBsAgayw and HBsAgadw2 proteins from the HBV isolates used differ in 16 residues (thus sharing 93% of their residues). The sequence of the HBsAgayw protein we used is identical to the sequence of the transgene-encoded HBsAgayw expressed by HBs-tg B6 mice. The sequence of the Kb-binding epitopes 1 and 2 of HBsAgayw and HBsAgadw2 selected for study differ by one or two residue exchanges within the epitope, but have identical flanking sequences (Fig. 1A, B). The S208–215 epitope 1 of HBsAgayw and HBsAgadw2 differ at two positions: in adw2, a valine (V) is substituted for leucine (L) at position 2, and an isoleucine (I) is substituted for leucine (L) at position 6 (Fig. 1B). Binding affinity of epitope 1 of HBsAgayw to Kb is low; the HBsAgadw2 variant of epitope 1 shows higher binding affinity to Kb than the HBsAgayw variant of this epitope (Table 1). In contrast, binding affinity of epitope 2 to Kb is high (Table 1).

B6 mice immunized with the pCI/Sayw or pCI/Sadw2 DNA vaccine generated a CD8+ T cell response to the Kb-binding epitope 1 detected after a 5-h ex vivo restimulation of primed splenic CD8+ T cells with cells pulsed with either HBsAgayw or HBsAgadw2 particles, or antigenic peptide S208–215 of HBsAgayw or HBsAgadw2 (Fig. 4A, groups 2, 3). The ayw and the adw2 variant of epitope 1 were cross-reactive because (i) epitope 1-specific CTL were primed by pCI/Sayw, or pCI/Sadw2; and (ii) cells pulsed with either HBsAgayw, or HBsAgadw2 particles, or pulsed with either peptide ILSPFLPL (ayw), or peptide IVSPFIPL (adw2) presented epitope 1 to primed CD8+ T cells. Hence, two substitutions within the 8mer epitope 1 did not impair productive processing, Kb binding or presentation of this epitope 10.

CD8+ T cells primed by the pCI/Sayw DNA vaccine recognized epitope 2 (S190–197) of HBsAgayw or HBsAgadw2 (Fig. 5A; group 2). This was detected ex vivo after a 5-h restimulation using peptide-pulsed cells, or transfectants expressing HBsAgayw. Primed CD8+ T cells did not recognize transfectants expressing endogenous HBsAgadw2. Immunization with the pCI/Sadw2 DNA vaccine did not prime epitope 2-specific T cells (Fig. 5A, group 3). CD8+ T cells primed by the pCI/Sadw2 (but not the pCI/Sayw) DNA vaccine recognized an adw2-specific epitope of unknown epitope/restriction specificity presented by transfectants that was not further investigated (Fig. 5, group 3). The substitution of the amino acid at position 5 [exchange of a hydrophobic valine (V) against a hydrophobic alanine (A)] thus impairs generation of epitope 2, but not its presentation by Kb molecule 10.

Figure 4.

 HBsAg-specific CTL responses to epitope 1 in HBs-tg mice. HBs-tg mice that express HBsAgayw in the liver were intramuscularly immunized three times (at 4-weekly intervals) with DNA vaccines encoding HBsAg subtype ayw (pCI/Sayw) or adw2 (pCI/Sadw2), or negative control pCI (vector without insert). Spleen cells removed from immune mice 12 days after the last injection were restimulated in vitro for 4 h (in the presence of brefeldin A) with RBL5 cells pulsed with HBsAg particles of the ayw (RBL5/SPayw) or adw2 (RBL5/SPadw2) subtype, or with the Kb/S208–215 binding peptide 1 from either HBsAgayw (ILSPFLPL) or HBsAgadw2 (IVSPFIPL). Mean numbers of splenic IFN-γ+ CD8+ T cells/105 CD8+ T cells ± SD of four to six mice (from two independent experiments) per group are shown.

Figure 5.

 HBsAg-specific CTL responses to epitope 2 in HBs-tg mice. Spleen cells were removed from mice vaccinated as described in the legend of Fig. 4, and restimulated in vitro with either syngeneic RBL5/Sayw or RBL5/Sadw2 transfectants, or the Kb/ S190–197 epitope 2 from HBsAgayw (VWLSVIWM) or HBsAgadw2 (VWLSAIWM). Mean numbers of splenic IFN-γ+ CD8+ T cells / 105 CD8+ T cells ± SD of four mice per group are shown.

2.4 Cross-reactive, Kb-restricted CD8+ T cell responses to HBsAg epitope 1 are primed in HBs-tg B6 mice

HBs-tg B6 mice express HBsAgayw from a transgene in the liver. HBs-tg mice were vaccinated to HBsAgayw (pCI/ Sayw) or HBsAgadw2 (pCI/Sadw2) (Fig. 4, 5B). No CD8+ T cell response was elicited by repeatedly vaccinating HBs-tg B6 mice with the pCI/Sayw DNA vaccine (Fig. 4, 5B, group 2). In contrast, vaccination of HBs-tg B6 mice with the pCI/Sadw2 DNA vaccine elicited a CD8+ T cell response to HBsAg (Fig. 4B, group 3). This shows that a natural variant epitope of HBsAg with a higher affinity for Kb (see Table 1) can ‘break tolerance’ by priming cross-reactive T cell immunity. This cross-reactive CD8+ T cell response recognized cells pulsed with either HBsAgayw, or HBsAgadw2 particles, or with peptides representing the ayw or adw2 variant of epitope 1 (Fig. 4B, group 3). These CD8+ T cells did not recognize RBL5/Sayw transfectants or the Kb-binding S190–197 epitope 2 (Fig. 5B, group 3). The CD8+ T cells showed a subtype-specific reactivity against an unmapped determinant(s) presented by RBL5/Sadw2 but not RBL5/Sayw transfectants (Fig. 5B, group 3).

We tested if specific CD8+ T cell populations can be detected in the antigen-producing liver in transgenic mice vaccinated with pCI/Sadw2. In spleen and liver NPC from HBs-tg B6 mice immunized with pCI/Sadw2, we detected this specific CD8+ T cell reactivity for months (Fig. 6). In contrast to adoptively transferred CD8+ T cells (Fig. 2), vaccine-primed anti-HBV-specific CD8+ T cells can thus access and show stable engraftment in the antigen-bearing target organ for >3 months.

Figure 6.

 S208–215-specific CD8+ T cells are found in the liver of vaccinated HBs-tg mice. Transgenic HBs-tg mice were immunized three times (at 4-weekly intervals) with a DNA vaccine encoding HBsAgadw2 (pCI/Sadw2). Liver and spleen cells were removed from immune mice 12 days after the last injection and restimulated in vitro with the Kb/S208–215-binding peptide ILSPFLPL. Mean numbers of splenic IFN-γ+ CD8+ T cells/105 CD8+ T cells ± SD of four mice per group are shown.

Table 1. Binding affinity of immunogenic HBsAg epitopes to Kb
EpitopeHBsAg variantPeptide sequenceKb-binding (nM)
1adw2IVSPFIPL 773

2.5 Histopathology of the liver of immunized HBs-tg mice developing specific CD8+ T cell reactivity to the HBsAg epitope 1

HBsAg-specific CD8+ T cells induced an inflammatory response in the HBsAg-producing liver. Untreated B6 mice showed a normal liver histology (Fig. 7A, B). Hepatocytes from HBs-tg B6 mice were enlarged and had a finely granular, pale-eosinophilic cytoplasm characteristic for ‘ground-glass hepatocytes’ that appears also in human HBV infection (Fig. 7C, D). No inflammatory infiltrations were detected.

HBs-tg mice immunized with the pCI/Sadw2 (but not the pCI/Sayw) DNA vaccine showed severe liver histopathology (Fig. 7E). Inflammatory infiltrates found in the parenchyma (Fig. 7F) and periportal fields (Fig. 7G) were pre-dominantly mononuclear cells (Fig. 7F). Small lymphoid cells were abundantly scattered throughout the parenchyma and periportal fields. Circumscribed foci of inflammatory cells surrounded apoptotic hepatocytes (Fig. 7H). Enlargement and hydropic swelling of hepatocytes was more pronounced in immune than in untreated HBs-tg mice. Some medium to small-sized nuclei displayed a condensed chromatin and a perinuclear halo (Fig. 7F, arrows), indicating an early stage of apoptosis. In addition, Councilman bodies representing apoptotic liver cells were abundant (Fig. 7H, arrows). Some hepatocytes displayed nuclear vacuolization (Fig. 7I, arrows). No significant cholestasis was detectable.

Figure 7.

 Liver histopathology of HBs-tg mice vaccinated with the pCI/Sadw2 DNA vaccine. Normal liver histology is observed in B6 mice (A, B). HBs-tg mice (C, D) show moderate cellular enlargement and a marked ground-glass appearance of the cytoplasm (D). The nuclei liver cells are moderately polymorphic. Periportal infiltrations are scarce. Repeated immunization with pCI/Sadw2 DNA induces severe histomorphological alterations of the liver (E–I), consistent with acute viral hepatitis. Inflammatory infiltrations comprising Kupffer cells, lymphocytes and few polymorphonuclear granulocytes are located in the lobular parenchyma (F) and in the periportal fields (G). Hepatocytes are hydropic, and often possess pycnotic nuclei, indicating an early stage of apoptosis (F, arrows). Acidophilic bodies (H, arrows), i.e. apoptotic liver cells, are abundant and often surrounded by focal inflammatory infiltrations. Many liver cell nuclei display conspicuous vacuolization (I, arrows). H&E stainings of formalin-fixed, paraffin-embedded tissue. Original magnifications: (A, C, E) ×10; (B, D, F) ×40; (G–I) ×63.

2.6 Priming HBsAg-specific CD8+ T cells in HBs-tg mice correlates with reduction of antigenemia

Non-treated HBs-tg mice show HBsAg serum levels of 30–50 ng/ml (Fig. 8A). Mice that developed cross-reactive CD8+ T cell responses to epitope 1 after HBsAgadw2 vaccination showed reduced antigenemia (with levels in the 5–15 ng/ml range), while animals vaccinated with HBsAgayw (that did not develop HBsAg-specific CD8+ T cell immunity) showed no change in the level of their antigenemia (Fig. 8A). The partial control of antigenemia thus correlates with the appearance of specific CD8+ T cells in immunized transgenic mice.

Figure 8.

Induction of HBsAg-specific serum antibody responses in HBs-tg mice. B6 and transgenic HBs-tg mice were immunized intramuscularly with DNA vaccines encoding HBsAgadw2 (pCI/Sadw2) or HBsAgayw (pCI/Sayw), and boosted after 3 weeks with the same vaccines. At 4 weeks after the last injection serum samples were analyzed for HBsAg antigen (A) or for HBsAg-specific antibodies (B). Mean antibody titers (mIU/ml) and serum HBsAg levels (ng/ ml) ± SEM of four to six mice/group are shown.

2.7 Anti-HBsAg serum antibodies appear in HBsayw-tg mice vaccinated with HBsAgadw2

In addition to T cell immunity, humoral anti-HBsAg immunity can play a role in the control of antigenemia. We followed the appearance of anti-HBsAg serum antibodies in vaccinated normal and transgenic mice. Normal (non-transgenic) B6 mice and congenic HBs-tg B6 mice were immunized twice with the pCI/Sayw or pCI/Sadw2 DNA vaccine. Their serum antibody titer specific for HBsAg was determined 2 weeks after the last vaccination by the IMxAUSAB test (Abbott) that detects HBsAg of different serotypes. While non-transgenic mice vaccinated with pCI/Sayw or pCI/Sadw2 plasmid DNA developed high serum antibody levels to HBsAg, HBs-tg mice showed an anti-HBsAg serum antibody response only after vaccinations with pCI/Sadw2 (but not with pCI/Sayw) plasmid DNA (Fig. 8B). Similar antibody responses were observed in mice vaccinated with HBsAgayw or HBsAgadw2 particles (data not shown). A serotype-specific ELISA (using HBsAgayw or HBsAgadw2 particle-coated plates) indicated that in normal mice >95% of the antibody response induced by all vaccines is directed against the ‘a’ determinant of HBsAg; in HBs-tg mice >90% of the response is directed against adw2-specific determinants (data not shown). The fine specificity of these antibodies remains to be defined.

3 Discussion

3.1 Priming cross-reactive T cell immunity by natural virus variants

The Kb-restricted HBsAg epitope 1 generated by processing exogenous (but not endogenous) HBsAg contains two conservative residue exchanges in the two natural variants selected for this study. Both variants prime cross-reactive CD8+ T cell responses in normal H-2b mice. In congenic, HBs-tg mice that express transgene-encoded HBsAgayw, only HBsAgadw2 primes CD8+ T cells specific for epitope 1 that cross-reacts to both variants tested. Emergence of epitope 1-specific CD8+ T cells in spleen and liver of HBs-tg mice immunized with HBsAgadw2 was associated with a typical liver histopathology, a reduction in HBsAg antigenemia, and the induction of anti-HBsAg serum antibodies. Natural variants of HBsAg are thus interesting potential candidates for the specific immunotherapy of chronic HBV infection.

3.2 Variant HBsAg epitopes that stimulate CD8+ T cells

The two HBsAg variants (of ayw or adw2 serotype) were selected for this study because of the availability of (i) the DNA vaccines encoding them, (ii) purified particles with an identical sequence produced by the same source, (iii) transgenic mice expressing one of the variants from a transgene in the liver, and (iv) stable transfectants expressing comparable amounts of these HBsAg variants. The variants differ by 16 residue exchanges. All residue exchanges involve conservative replacements of similar hydrophobic amino acids (i.e. L/V, L/I, V/A). These HBsAg variants selected showed residue exchanges within the two Kb-binding HBsAg epitopes. We show that some of these residue exchanges can change the immunogenicity of HBsAg by affecting processing and/or epitope presentation. In addition, residue exchanges may modulate the regulation of CD8+ T cell priming by attenuating the immunodominance of epitopes. In the HBsAg system, the immunodominance of Ld-restricted CD8+ T cell responses is well established 15, 16. Conceivably, single residue exchanges can create or destroy such dominant CD8+ T cell responses thereby suppressing or enhancing other responses. This aspect was not addressed in the present study.

We studied CD8+ T cell priming to two Kb-binding HBsAg epitopes that are mapped, and for which information on the processing pathway that generates them is available. We do not assume that this is the complete repertoire of the multispecific CTL response to HBsAg in H-2b mice. The detection of a ‘new’ CTL-stimulating epitope with unknown epitope/restriction specificity in B6 mice vaccinated with pCI/Sadw2 indicates that additional specificities are recognized in this response. Antigenic peptides may be located in the hydrophobic (transmembrane) regions of HBsAg, and may hence be difficult to synthesize and to use. Furthermore, the study was focused on the two Kb-restricted HBsAg epitopes we have mapped. ‘New’ CD4+ T cell-stimulating epitopes in the HBsAgadw2 (but not HBsAgayw) variant may facilitate priming of epitope 1-specific CD8+ T cell responses and anti-HBsAg antibody responses in HBs-tg mice vaccinated with HBsAgadw2. This aspect is under investigation.

3.3 The Kb-restricted epitope 1 and 2 of HBsAg

Comparing the two epitopes, which bind to the same class I molecule, we demonstrated that one or two conservative residue exchanges can either affect, or not affect productive processing, Kb binding and presentation. The immunogenicity of epitope 1 generated from exogenous HBsAg was not affected by the two residue exchanges within the epitope that do not affect the Kb anchor motifs (i.e. F or Y at position 5; L, V or I at position 8). Epitope 1 is the focus of the study because it primes cross-reactive CD8+ T cell responses to HBsAg in HBs-tg B6 mice, a specific reactivity that is exceedingly difficult to generate in these mice.

Epitope 2 lost its immunogenicity in the HBsAgadw2 variant, and is therefore of no interest in this system. The ‘new’ CTL response primed by HBsAgadw2 is not expected to contribute to the therapeutic effect of pCI/Sadw2 DNA vaccination of HBs-tg mice, although this CTL reactivity is primed in these mice. Because this specific CD8+ T cell reactivity is not detected in mice injected with the pCI/Sayw DNA vaccine, it is probably not presented by the transgene-encoded HBsAgayw.

3.4 Antibody response to HBsAg induced by vaccinating HBs-tg mice

The partial suppression of HBsAg antigenemia in vaccinated HBs-tg mice probably results from the anti-HBsAg antibody response also induced by the vaccination protocol used 17. It confirms that vaccination with HBsAg of a heterologous sero/genotype can break tolerance but probably does not contribute to the therapeutic effect of this post-exposure vaccination. The coexistence of HBsAg antigenemia and anti-HBsAg serum antibodies, and the reduction in antigenemia observed (Fig. 8), indicates that vaccinating HBsayw-tg mice with HBsAgadw2 stimulates antibodies that cross-react with HBsAgayw. The specificity of the serum antibodies differed between immunized normal and transgenic mice. While vaccination of normal mice with all HBsAg variants tested induced a strong, serotype-independent antibody response against the ‘a’ determinant of HBsAg, vaccination of HBs-tg mice with HBsAgadw2 stimulated low titers of adw2-specific antibodies. We are well aware that in a situation in which continuously produced antigen adsorbs a major fraction of the antibodies, analyses of serum antibody titers and their specificities is not very informative.

3.5 Implications for the specific immunotherapy of chronic HBV infection

Hepatocytes from HBs-tg mice are exquisitely sensitive to the cytotoxic effect of IFN-γ and TNF because they retain the large HBsAg protein within the cell 18. HBs-tg mice are exquisitely sensitive to liver injury induced by adoptively transferred CD8+ T cells specific for either epitope 1, or epitope 2. Hence, direct presentation (of epitope 2) and cross-presentation (of epitope 1) is able to specifically activate virus-specific CTL in the liver. HBs-tg mice do not replicate HBV. These two aspects of the model do not reflect the clinical condition of chronic HBV infection. As a preclinical model in an immunologically well-defined system, the analysis nevertheless can contribute valuable insight into the CD8+ T cell reactivity againstHBV.

The in vitro propagated T cells are rapidly eliminated in vivo, presumably in the liver. Adoptive transfers of specific CD8+ T cells thus deliver a brief antiviral effect to the liver but antiviral immunity is not stably engrafted into the host. This makes direct vaccination protocols attractive. Vaccination is difficult as HBs-tg mice are largely tolerant to HBsAgayw15, 19; this is shown by the finding that even repeated immunizations of HBs-tg mice with the pCI/Sayw DNA vaccine or HBsAgayw particles (with ODN) did not prime epitope 1- or epitope 2-specific CD8+ T cells. Hence, alternative approaches for therapeutic vaccination against chronic HBV are needed.

Priming CD8+ T cells to an epitope 1 variant (from an HBVadw2 isolate) that binds Kb with higher affinity could override unresponsiveness to an epitope 1 variant (from an HBVayw isolate) that binds Kb with lower affinity. In HBs-tg mice, this specific antiviral activity could be primed in vivo, and CD8+ T cell reactivity withthis specificity was found in the liver of vaccinated, transgenic mice. Vaccinated HBs-tg mice developed a liver histopathology but no striking serum transaminase elevation, indicating in situ delivery of CD8+ T cell effector functions. The histological finding in transgenic mice repeatedly immunized with HBsAgadw2 was a diffuse hepatocellular injury associated withperiportal and lobular inflammatory infiltrates and a high rate of liver cell apoptosis. These histological changes reproduce many aspects of the findings characteristic for acute viral hepatitis in man 20. Chronic HBV infection develops because of inefficient CTL-mediated control of the virus 21, 22. For therapeutic vaccination, variants expressing cross-reactive CTL epitopes (perhaps preferentially generated during processing of exogenous HBV antigens) may be an attractive choice. Although the tg mice are exquisitely sensitive to CD8+ T cell effector functions, and the model does not allow us to monitor the effect of vaccination on HBV replication, we propose that vaccination of chronically HBV-infected patients with heterologous HBV isolates of different geno/serotype offers advantages over vaccination with homologous HBV isolates. Hence, the reported mouse data, if confirmed for HBsAg or other HBV antigens in its natural host (i.e. in man), may help to define new ways for the specific therapy of chronic HBV infection.

4 Materials and methods

4.1 Mice

C57BL/6JBom (B6) mice (H-2b) were bred and kept under standard pathogen-free conditions in the animal colony of Ulm University (Ulm, Germany). C57BL/6J-TgN(Alb1HBV)44Bri transgenic (HBs-tg) mice that express HBsAgayw (encoded by the HBV sequence with the accession no. V01460 J02203, 23) were obtained from the Jackson Laboratory (Bar Harbor, ME). Male and female mice were used at 8–16 weeks of age.

4.2 Cells, recombinant HBsAg particles and antigenic HBsAg peptides

The H-2b cell line RBL5 was obtained from Dr. H.-U. Weltzien (Freiburg, Germany). Stable RBL5 transfectants that expressed similar amounts of HBsAgayw and HBsAgadw2 were generated (data not shown). Recombinant HBsAg subtype ayw or adw2 particles were obtained from Rhein-Biotech AG (Düsseldorf, Germany). HBsAg particles produced in the Hansenula polymorpha host strain RB10 were purified as described 24. The synthetic Kb-binding S208–215 ILSPFLPL (ayw) or IVSPFIPL (subtype adw2) peptides, and Kb-binding S190–197 VWLSVIWM (ayw) or VWLSAIWM (adw2) peptides were obtained from Jerini BioTools (Berlin, Germany). Peptides were dissolved in aDMSO solution at a concentration of 10 mg/ml and diluted with culture medium before use.

4.3 Plasmids and DNA vaccination

HBsAgayw and HBsAgadw2 were cloned into the pCI (Promega) and BMGneo vectors as described 25, 26. We used as DNA vaccines the pCI/Sayw and pCI/Sadw2 plasmids that express HBsAgayw and HBsAgadw2 equally well. This was shown by immunoprecipitating HBsAg from cells transiently transfected with DNA of these plasmids (data not shown). Differences in the immunogenicity of the HBsAg epitopes thus do not result from different levels of HBsAg expression by the DNA vaccine or the transfectants. For intramuscular nucleic acid immunization, we injected 50 μl PBS containing 1 μg/μl plasmid DNA into each tibialis anterior muscle as described 25.

4.4 Determination of specific, splenic and hepatic CD8+ T cell frequencies

Spleen cell suspensions 10 and the preparation of hepatic NPC cells has been described 27, 28. Spleen cells and liver NPC (1×106/ml) were incubated for 1 h in RPMI 1640 medium with either 5 μg/ml HBsAg-derived peptides, or HBsAg-expressing transfectants (106/ml), or HBsAg particle-pulsed cells. Thereafter, 5 μg/ml brefeldin A (BFA) (; Sigma) was added, and the cultures were incubated for further 4 h. Cells were harvested and surface stained with anti-CD8 mAb, fixed, permeabilized and stained for cytoplasmic IFN-γ. We determined the frequencies of CD8+ IFN-γ+ CTL by FCM analyses. The mean number of CD8+ IFN-γ+ T cells/105 CD8+ spleen or liver T cells are shown.

4.5 Transfer of specific CD8+ T cell lines

CD8+ T cell lines were obtained from spleen of B6 mice immunized by the pCI/Sayw DNA vaccine. Spleen cells were restimulated in vitro with syngeneic, RBL5 cells pulsed with the Kb/S208–215 binding peptide 1 (ILSPFLPL) or the Kb/S190–197 binding peptide 2 (VWLSVIWM). In lines expanded for approximately 2 weeks in vitro, >80% of the CD8+ T cells were of the expected epitope specificity, as evident in specific IFN-γ expression analyses. Cells were washed, and 5×106 cells of these lines were injected intravenously. Controls were nonspecific CD8+ T blasts isolated from 3-day Con A-stimulated cultures.

4.6 Determination of transaminases, HBsAg and anti-HBsAg antibody in serum

Serum samples were repeatedly obtained from individual, immunized or control mice by tail bleedings at different time points post-injection. Serum alanine aminotransferase (ALT) activity was determined in blood using the Reflotron® test (; Roche Diagnostics GmbH). The HBsAg concentration in the serum of transgenic mice was determined by the commercial ELISA AUSZYME II (Abbott Laboratories, Wiesbaden, Germany). Antibodies against HBsAg were detected in mouse sera using the commercial IMxAUSAB test (cat. no.7A39-20; Abbott). Antibody levels were quantified usingsix standard sera. The tested sera were diluted so that the measured OD values were between standard serum one and six. Values presented in this report are calculated by multiplying the serum dilution with the measured antibody level (mIU/ml).

4.7 Histology

Thin slides of liver tissue (<3 mm) were fixed in 4% formalin (pH 7.0) for 24 h and embedded in paraffin. Paraffin sections, 2 μm thick, were stained with hematoxylin-eosin (H&E).

4.8 Binding of HBsAg peptides to Kb

Affinity-purified MHC class I molecules Kb were incubated for 48 h at 18°C with increasing concentrations of test peptide and a fixed concentration (about 2 nM) of radiolabelled VSV NP 52–59 indicator peptide in the presence of 3 μM human β2m as previously described 29, 30. Subsequently, the binding of peptides to MHC class I complexeswas examined by Sephadex G50 spun column gel filtration 29. The radiolabelled VSV NP 52–59 peptide is distributed into the excluded void volume (MHC-bound peptide) and the included volume (free peptide). These are measured by gamma spectrometry and the fraction of the test peptide bound to the MHC relative to the total amount of test peptide applied was calculated. The concentration of test peptide needed to obtain 50% inhibition of the binding of the indicator peptide (the IC50) was determined. The lower the IC50 value the better the binding of the test peptide. To avoid ligand depletion, a concentration of MHC sufficient to obtain no more than 15–25% binding was used in all binding experiments. Under these conditions, the IC50 is an approximation of the KD. All binding experiments were conducted as inhibition experiments.


We greatly appreciate the helpful comments and suggestions of Drs. F. V. Chisari (Scripps Institute, La Jolla, CA) and A. Bertoletti (UC London, UK). We greatly appreciate the expert technical assistance of Tanja Güntert, Beate Wotschke and Katrin Ölberger. This work was supported by grants from the Deutsche Forschungsgemeinschaft (Re 549/10–1) and theEuropean Commission (CTLALTVAX; QLRT-2001–00700 ) to R.S. and J.R.


  1. 1


  2. 2


  3. 3


  4. 4


  5. 5


  6. 6


  7. 7


  8. 8