Comparative Antigenicity and Immunogenicity of Hep
http://www.100md.com
病菌学杂志 2005年第21期
Vaccine Research Institute of San Diego, San Diego, California 92109
Department of Biochemistry and Molecular Biophysics, Commonwealth University, Richmond, Virginia
Merck Research Laboratories, West Point, Pennsylvania
Division of Clinical Virology, Karolinska Institute, Karolinska University Hospital, Huddinge, Sweden
ABSTRACT
The hepatitis B virus core protein (HBcAg) is a uniquely immunogenic particulate antigen and as such has been used as a vaccine carrier platform. The use of other hepadnavirus core proteins as vaccine carriers has not been explored. To determine whether the rodent hepadnavirus core proteins derived from the woodchuck (WHcAg), ground squirrel (GScAg), and arctic squirrel (AScAg) viruses possess immunogen characteristics similar to those of HBcAg, comparative antigenicity and immunogenicity studies were performed. The results indicate that (i) the rodent core proteins are equal in immunogenicity to or more immunogenic than HBcAg at the B-cell and T-cell levels; (ii) major histocompatibility complex (MHC) genes influence the immune response to the rodent core proteins (however, nonresponder haplotypes were not identified); (iii) WHcAg can behave as a T-cell-independent antigen in athymic mice; (iv) the rodent core proteins are not significantly cross-reactive with the HBcAg at the antibody level (however, the nonparticulate "eAgs" do appear to be cross-reactive); (v) the rodent core proteins are only partially cross-reactive with HBcAg at the CD4+ T-cell level, depending on MHC haplotype; and (vi) the rodent core proteins are competent to function as vaccine carrier platforms for heterologous, B-cell epitopes. These results have implications for the selection of an optimal hepadnavirus core protein for vaccine design, especially in view of the "preexisting" immunity problem that is inherent in the use of HBcAg for human vaccine development.
INTRODUCTION
The virus family Hepadnaviridae includes hepatotropic, partially double-stranded DNA viruses with a replication strategy unique for animal DNA viruses consisting of reverse transcription of an RNA intermediate (35). This family is divided into two groups, the Orthohepadnavirus genus and the Avihepadnavirus genus. In addition to being identified in humans (hepatitis B virus [HBV]), orthohepadnaviruses have been identified in rodents such as woodchucks (woodchuck hepatitis virus [WHV]) (39) and ground (17) and arctic (40) squirrels (ground squirrel hepatitis virus and arctic squirrel hepatitis virus) and more recently in Old World as well as New World primates such as woolly monkeys (14), orangutans (43, 44), gorillas (11), chimpanzees (26, 42), and gibbons (15). The first avian hepadnavirus (duck hepatitis B virus) was identified in Pekin ducks (18, 48). Avian hepadnaviruses have also been isolated from other avian species such as the gray heron (37), Ross' goose and snow goose (6), and white stork (30) and most recently from cranes (29). The nonhuman primate viruses are most closely related to HBV, and their structural proteins share antigenic cross-reactivity. The rodent hepadnaviruses are more distantly related to HBV (55 to 70% nucleotide identity), and the avian hepadnaviruses are highly divergent from HBV (approximately 40% nucleotide identity) (13).
Interestingly, the nucleocapsid of HBV, the hepatitis B core antigen (HBcAg), is an extremely powerful immunogen and is significantly more immunogenic than HBV envelope (HBsAg) proteins during natural infection (12) and after immunization of recombinant proteins in mice (24), although both HBcAg and HBsAg are particulate antigens. For example, in contrast to HBsAg, HBcAg elicits immunoglobulin G (IgG) and IgM anti-HBc antibody production in athymic (i.e., T-cell-independent) mice (21); HBcAg preferentially activates Th1-type T cells (23); HBcAg up-regulates B7.1 and B7.2 costimulatory molecules on resting B cells (19); and HBcAg is an efficient vaccine platform to carry heterologous epitopes (41). Because these immunologic characteristics are unique to the particulate HBcAg and do not pertain to a nonparticulate secreted form of this protein, designated HBeAg, structural characteristics of the HBcAg may explain its enhanced immunogenicity. Recent cryoelectron microscopy (5, 7) and crystallographic (46) studies have elucidated the structure of HBcAg. A clustering of dimer subunits produces spikes on the surface of the core shell, which consist of radial bundles of four long -helices (5, 7). The orientation of the array of protein spikes distributed over the surface of the HBcAg particle may be optimal for cross-linking B-cell membrane immunoglobulin antigen receptors (19), especially because dominant B-cell epitopes appear to be positioned on or near the tip of the spikes (2). Similar structural analyses of the other orthohepadnavirus core particles have not been performed.
Therefore, to determine whether the immunologic characteristics are unique to HBcAg we have performed immunogenicity-antigenicity studies with mice by comparing HBcAg with the core proteins derived from the rodent orthohepadnaviruses, namely, WHV (WHcAg), ground squirrel hepatitis virus (GScAg), and arctic squirrel hepatitis virus (AScAg). We have compared (i) relative degrees of immunogenicity at the B- and T-cell levels; (ii) major histocompatibility complex (MHC) influence on responsiveness; (iii) T-cell independence; (iv) antigenic cross-reactivity at the B- and T-cell levels; and (v) the relative ability to function as vaccine carrier platforms for heterologous epitopes. For this purpose we have produced a panel of recombinant native and modified () core particles derived from these four hepadnaviruses. There is no current consensus in the literature regarding the serologic relatedness of orthohepadnavirus core proteins. Although most earlier studies suggested low to no cross-reactivity between antibodies specific for HBcAg and the rodent hepadnavirus core proteins (8, 9, 25, 27, 28, 34), several studies have indicated a greater degree of antibody cross-reactivity (36, 38, 45). Early serologic comparisons between hepadnaviral core proteins often employed infectious sera and/or partially purified core proteins and various experimental assays. The availability of recombinant core proteins derived from each viral species and monoclonal anti-HBc/e reagents and murine polyclonal species- specific anti-core antisera has made these comparisons less problematic and the quantitation of relative cross-reactivity feasible. The CD4+ T-cell responses to the various orthohepadnavirus core proteins have not been compared in the same species (i.e., mouse) previously.
MATERIALS AND METHODS
Animals. The C57BL/10 (B10), B10.S, and (B10 x B10.s)F1 mice were obtained from the breeding colony of the Vaccine Research Institute of San Diego. The B10, H-2 congenic strain (B10.BR, B10.D1, B10.D2, B10.M, B10.PL, and B10.RIII) and BALB/c euthymic and athymic mice were obtained from the Jackson Laboratory (Bar Harbor, ME). The acutely infected woodchuck sera were kindly provided by Michael Roggendorf (Institute of Virology, University of Essen, Essen, Germany). All animal care was performed according to National Institutes of Health standards as set forth in the Guide for the Care and Use of Laboratory Animals (1996).
Recombinant core proteins and synthetic peptides. Full-length recombinant HBcAg of the ayw3 subtype and a recombinant HBeAg corresponding in sequence to serum-derived HBeAg encompassing the 10 precore amino acids remaining after cleavage of the precursor and residues 1 to 149 of HBcAg were produced as described below. The presence of the 10 precore amino acids prevents particle assembly, and HBeAg is recognized efficiently by HBeAg-specific MAbs but displays little HBc antigenicity.
The recombinant rodent core proteins WHcAg, WHeAg, GScAg, and AScAg and the "modified" core proteins WHcAg, GScAg, and WHeAg were expressed or derived (GScAg and AScAg) from the pUC-WHcAg vector expressing the full-length woodchuck core. Briefly, the full-length sequence of WHcAg (NCBI accession no. NC_004107) served as a template to create synthetically the two other rodent full-length core antigens correlating to their accession numbers (for GScAg, accession no. NP_040993; for AScAg, accession no. NC_001719). The core proteins were either full length or truncated at amino acid (aa) 149 and had a cysteine added at residue 150 (i.e., 150C). Addition of a cysteine at residue 150 has been previously shown to stabilize truncated HBcAg (20). WHeAg corresponds to a truncated and less stable (i.e., more prone to degradation in vitro) version of WHcAg (149 aa) and was obtained after excision of the BseAI-BamHI fragment of the full-length WHcAg and replaced after ligation with the following pair of oligonucleotides digested by BseAI-BamHI: 5' TCCGGAACATACAGTCATTTGAGGATCC 3' and 5' GGATCCTCAAATGACTGTATGTTCCGGA 3'.
WHcAg, GScAg, and WHeAg indicate, unless otherwise specified, the insertion of heterologous sequences of 20 to 30 aa in the loop of the corresponding core proteins. All insertions were accomplished using the EcoRI-XhoI sites specifically engineered to not disrupt the reading frame and to encode a linker Gly-Ile-Leu on the N terminus and Leu-Glu on the C terminus of the inserted heterologous epitopes. Hybrid core proteins containing a specific Plasmodium falciparum circumsporozoite (CS) repeat sequence (NANPNVDPNANP3) inserted in the external loop were also produced (i.e., HBc-M78, WHc-M78, and GS-M83). For greater detail regarding constructs, refer to the companion paper (4). All the core constructs have been sequenced in both directions (Retrogen, Inc., San Diego, Calif.). Production and purification of the plasmids (MoBio Laboratories, Inc., San Diego, Calif.) and ligation experiments (La Roche) were performed according to the manufacturer's procedures. Cloning and subcloning have been performed following protocols described by Sambrook et al. (32). The Top10 Escherichia coli strain was purchased from Invitrogen. Transformation of chemically competent Top10 by heat shock was carried out according to the protocol of the manufacturer (Invitrogen).
Purification of core antigens. The core proteins were precipitated from the bacterial lysate by the addition of solid ammonium sulfate to 45% saturation (277 g/liter). The precipitates were collected by centrifugation, redissolved in a minimum of buffer (10 mM sodium phosphate buffer, pH 6.8), and dialyzed extensively against the same buffer. The protein solutions were then applied to a Bio-Rad BioGel HTP hydroxyapatite column (5 cm by 5 to 10 cm, depending on amount of protein) and eluted with 50 mM sodium phosphate buffer, pH 6.8. The core antigens pass through unretained. The proteins were then applied to a Sepharose CL4B column (5 by 100 cm). Characterization of the purified core proteins is illustrated in Fig. 1.
Endotoxin was removed from the core preparations by a modification of a phase separation with Triton X-114 (1, 16). A solution of the protein at a concentration of 5 mg/ml was made with 1% Triton X-114 and incubated at 4 C for 30 min with constant stirring. The solution was then incubated at 37 C for 10 min and then centrifuged at 20,000 x g for 10 min. The protein solution was recovered from above the detergent. This procedure was repeated four times. Finally, the protein was precipitated by lowering the pH to 5. Residual detergent remained in solution. The protein was recovered by centrifugation and dissolved in endotoxin-free buffer. Prior to Triton X-114 treatment, the core preparations contained approximately 10 to 25 ng of endotoxin/μg HBcAg; after phase separation with Triton X-114, the endotoxin content was from 0.1 ng/μg HBcAg to an undetectable amount as determined by the QCL-1000 chromogenic Limulus amoebocyte lysate endpoint assay (Cambrex, East Rutherford, NJ).
Synthetic peptides derived from the HBcAg or WHcAg sequences were synthesized by the simultaneous peptide synthesis method as previously described (31).
Immunizations and serology. Groups of three to five mice were immunized intraperitoneally (i.p.) with the core proteins (usually 10 μg) emulsified in complete Freund's adjuvant or incomplete Freund's adjuvant (IFA) for both antibody production and T-cell experiments. For antibody experiments, mice were bled retro-orbitally and sera were pooled from each group. Individual mouse serums were tested periodically to confirm the fidelity of the pooled serum results. Anti-core or anti-insert IgG antibodies were measured in murine sera by an indirect solid-phase enzyme-linked immunosorbent assay (ELISA) by using the homologous or heterologous core proteins (50 ng/well) or synthetic peptides (0.5 μg/well), representing the inserted sequence, as solid-phase ligands as described previously (33). Serial dilutions of both test sera and preimmunization sera were made, and the data are expressed as antibody titers representing the reciprocals of the highest dilutions of sera required to yield an optical density at 492 nm (OD492) three times that seen with an equal dilution of preimmunization sera. IgG isotype-specific ELISAs were performed by using IgG1-, IgG2a-, IgG2b-, and IgG3-specific peroxidase-labeled secondary antibodies (Southern Biotechnology, Birmingham, AL). Because secondary antibody specific for woodchuck IgG is not available, protein A-peroxidase (Prot A-HROP; Sigma Chemical Co., St. Louis, Mo.) (1 μg/ml) was used to develop the ELISA for the WHV-infected woodchuck sera.
In vitro T-cell proliferation and cytokine assays. Spleen cells from groups of three each of the B10 H-2 congenic mice were harvested and pooled 4 to 6 weeks after immunization with the various core proteins. Spleen cells (5 x 105) were cultured with various concentrations of homologous or heterologous core proteins or synthetic peptides. For T-cell proliferation assays spleen cells were cultured for 96 h and during the final 16 h with 1μCi of [3H]thymidine (New England Nuclear) (6 to 7 Ci/μmol). The cells were harvested onto filter strips for determination of [3H]thymidine incorporation. The data are expressed as counts per minute corrected for background proliferation in the absence of Ag ( cpm). For cytokine assays, culture supernatants were harvested at 96 h and gamma interferon (IFN-) production was measured by a two-site ELISA using MAb 170 and a polyclonal goat anti-mouse IFN- (Genzyme Corp., Boston, MA).
RESULTS
Eight H-2 congenic murine strains which differ only in MHC haplotype were immunized i.p. with equal 10-μg doses of recombinant truncated (i.e., 150C) WHcAg, HBcAg, GScAg, and AScAg emulsified in IFA. As shown in Fig. 2, all four core proteins elicited strong antibody responses to the homologous protein as early as 4 weeks after a single injection. The rodent core proteins elicited responses equal to or higher than those seen with HBcAg in all eight strains. In fact, the rodent core proteins appeared to be somewhat more immunogenic in mice than HBcAg. Therefore, enhanced immunogenicity is not unique to HBcAg. This analysis also indicated that there was an absence of genetic nonresponders to WHcAg, GScAg, or AScAg, as previously reported for HBcAg (21). However, a hierarchy of responsiveness did exist among the different H-2 haplotypes for each core protein and the response hierarchy differed between the core proteins. The rank order of anticore production among H-2 congenic mice for each core protein is shown in Table 1. Because antibody production is largely regulated by T-helper (Th) cell function, which is MHC class II (MHC-II) restricted, the differing rank order of anti-core production among H-2 haplotypes between the four core proteins suggests major differences in the Th cell recognition of the various core proteins (see Fig. 9 and 10). The IgG isotype distribution of anti-core antibodies was remarkably similar for all four core proteins (i.e., IgG2b > IgG2a >> IgG1 > IgG3) (data not shown).
Cross-reactivity of anti-core antibodies. The primary polyclonal anti-core antibodies depicted in Fig. 2 were analyzed for cross-reactivity between the various core proteins (Fig. 3). The anti-HBc antibodies produced against HBcAg in all eight H-2 congenic strains demonstrated zero to relatively low cross-reactivity for the rodent core proteins. The absence of antibody cross-reactivity in either direction between HBcAg and WHcAg was especially clear. Low but still detectable antibody cross-reactivity between HBcAg and GScAg and between HBcAg and AScAg was present in at least some antisera from the various H-2 congenic strains. In the B10.M strain, for example, polyclonal anti-HBc antibodies were nonreactive with WHcAg but reacted with GScAg and AScAg with approximately 0.8% efficiency compared to HBcAg (i.e., endpoint titers of 1:5,000 versus 1:625,000). Reciprocally, B10.M polyclonal anti-GSc and anti-ASc antibodies cross-reacted with HBcAg with very low efficiency (i.e., 0.037%) (Fig. 3). In fact, the greatest degree of cross-reactivity of polyclonal anti-HBc antibodies for any rodent core protein was 0.8%, as demonstrated with the B10.RIII, B10.S, and B10 strains (Fig. 3). Given the 65 to 67% amino acid homology between the HBcAg and the rodent core proteins, the zero to low level of antibody cross-reactivity may appear somewhat surprising. However, when one considers that the dominant B-cell epitopes on HBcAg reside in the external loop region at the tip of the structural spikes (2) comprising amino acid 74 to 84, the sequence homology between HBcAg and the rodent core proteins is only 18.2% in this region (See Fig. 4 for core protein sequence alignments). An additional B-cell epitope (nonloop epitope) on HBcAg has been mapped to include residues 20 to 22 and 25 to 29 (2), for which the sequence homology between HBcAg and the rodent core proteins is only 37.5%. Considering the location of the B-cell epitopes mapped on HBcAg, the low level of antibody cross-reactivity between HBcAg and the rodent core proteins is consistent with the results of cryoelectron microscopy studies of HBcAg labeled with anti-HBc monoclonal Fab fragments (2). Confirming the lack of serologic cross-reactivity between HBcAg and WHcAg and GScAg observed with polyclonal anti-HBc antisera, a panel of anti-HBc-specific monoclonal antibodies also failed to recognize WHcAg and GScAg. Similarly, a commercial (Abbott Labs) human anti-HBc reagent recognized solid-phase HBcAg but not WHcAg in an ELISA format (Table 2).
In contrast to the lack of antibody cross-reactivity between HBcAg and the rodent core proteins, the rodent core proteins are significantly cross-reactive among themselves at the level of antibody recognition. In all eight H-2 congenic strains, antibodies elicited by immunization with WHcAg, GScAg, and AScAg are cross-reactive to some degree on the other two rodent proteins. Interestingly, anti-WHc polyclonal antibodies cross-reacted more efficiently on AScAg than on GScAg, as demonstrated with sera from six of the eight H-2 congenic strains (i.e., B10.BR, B10.D1, B10.D2, B10.M, B10.PL, and B10) (Fig. 3). Reciprocally, anti-ASc antibodies cross-reacted on WHcAg more efficiently than on GScAg in sera from five of the eight strains. While the amino acid sequences of the rodent core proteins are virtually identical in the nonloop, B-cell epitope regions (i.e., aa 20 to 22, 25 to 29, and 126 to 132), there are significant differences within the external loop region. Significantly, WHcAg and AScAg are more (i.e., 72.7%) homologous in a broadly defined loop region between residues 70 and 91 than are AScAg and GScAg (i.e., 63.6%), which most likely explains the serologic data.
To identify the contribution of the nonloop regions to the immunogenicity and cross-reactivity of the rodent core proteins, we examined the effect of abolishing the dominant loop epitope by inserting different foreign sequences into the loop region of WHcAg (WHcAg) and GScAg (GScAg) (i.e., an influenza A virus sequence was inserted into WHcAg and a malaria CS repeat sequence was inserted into GScAg). As shown in Fig. 5, polyclonal anti-HBc recognized the native GScAg with only low efficiency and did not recognize native WHcAg or modified WHcAg. Similarly, antisera raised against native or modified rodent core proteins did not recognize HBcAg efficiently. Anti-native WHc antibody recognized modified WHcAg 24-fold less efficiently than native WHcAg, demonstrating the dominance of the WHcAg loop epitope. However, note that recognition of GScAg was equivalent to that of the modified WHcAg and recognition of the modified GScAg was almost equivalent to that of the modified WHcAg by the anti-native WHc antisera (Fig. 5, second column). This suggests that WHcAg and GScAg are totally cross-reactive outside the loop region. As shown in column 3 of Fig. 5, antisera raised to modified WHcAg (i.e., nonloop anti- WHc) recognized native WHcAg, modified WHcAg, native GScAg, and modified GScAg equivalently, again indicating that the two rodent core proteins are fully cross-reactive outside of the loop region. Similarly, anti-native GSc antibody recognized GScAg 120-fold less efficiently than native GScAg but recognition of WHcAg and WHcAg was equivalent to that seen with GScAg by anti-native GSc antisera (Fig. 5, column 4). Finally, antisera raised to modified GScAg recognized modified and native WHcAg core proteins similarly and only fourfold less efficiently than GScAg and GScAg. Cumulatively, these results indicate that WHcAg and GScAg are fully cross-reactive outside of the loop region, which is consistent with the 100% sequence homology between these two proteins in the nonloop, B-cell epitope regions mapped on the HBcAg. Conversely, the lack of cross-reactivity between HBcAg and the rodent core proteins outside the loop region is also consistent with the relatively low (i.e., 37.5%) sequence homology in the aa regions 20 to 22 and 25 to 29. The relatively high (71.4%) sequence homology between HBcAg and the rodent core proteins in the aa 126 to 132 non-loop region suggests that this region may not contribute in a major way to a nonloop, B-cell epitope(s).
Cross-reactivity of anti-E antibodies. Although HBcAg and the rodent core proteins were from minimally cross-reactive to non-cross-reactive at the antibody level, it was of interest to determine whether the nonparticulate forms (i.e., HBeAg and WHeAg) were cross-reactive. A panel of eight HBeAg-specific MAbs were tested for binding to solid-phase HBeAg and WHeAg in an ELISA format (Table 3). All eight HBeAg-specific MAbs were able to bind solid-phase WHeAg with efficiency that was equal to or reduced compared to that seen with HBeAg. In a solution-phase inhibition ELISA format, soluble WHeAg inhibited both the binding of an HBeAg-specific MAb (CE1-E) and the ability of a polyclonal anti-HBe antibody to bind to solid-phase HBeAg; however, the right-shifted WHeAg concentration curves suggested low-affinity binding to the WHeAg compared to the HBeAg results (Fig. 6). To determine whether the apparent low-affinity binding of WHeAg to MAb CE1-E had a structural basis as opposed to merely reflecting reduced amino acid homology within the HBeAg/WHeAg epitope, a modified WHeAg was used as the inhibitor. The modified WHeAg contained a heterologous sequence inserted behind residue 78 in addition to lacking a C-terminal cysteine and was less stable than truncated WHeAg without an insert, as determined on the basis of immunogenicity data (i.e., it was less immunogenic). The modified WHeAg was significantly more inhibitory than WHeAg, which suggests that the "eAg" epitope recognized by MAb CE1-E is shared between HBeAg and WHeAg and can be more or less exposed depending on structure. Because similar results were obtained with the HBeAg-specific MAbs 904 and 905 (data not shown), which recognize distinct HBeAg epitopes, it appears that HBeAg and WHeAg share more than one "eAg" B-cell epitope.
Serologic cross-reactivity of infected woodchuck sera. An independent method of determining degrees of cross-reactivity between the capsid antigens of the HBV and the WHV is measurement of antibody specificities in sera from WHV-infected woodchucks. Therefore, sequential serum samples from two woodchucks infected at day 0 with 106 50% infective doses of a titrated WHV stock serum were tested for anti-WHc, anti-HBc, and anti-HBe antibody specificities (Fig. 7). Anti-WHc antibody was detectable at day 40 and peaked between days 80 and 100 after infection. Anti-WHc antibody cross-reacted on HBcAg very minimally, as the anti-WHc titers were 4 orders of magnitude higher than the anti-HBc titers in both animals. In contrast, anti-HBe antibodies of significantly higher titer than anti-HBc antibodies were observed in both woodchucks and likely represent anti-WHe antibody production cross-reacting on HBeAg. The use of WHeAg to detect anti-WHe antibody is problematic because of the very high anti-WHc antibody titers that may obscure anti-WHe detection. However, because anti-WHc antibody is non-cross-reactive on HBc/HBeAg and because HBeAg and WHeAg are cross-reactive, HBeAg can be used to detect anti-WHe antibodies in infected woodchuck sera.
WHcAg can act as a T-cell-independent antigen. We have previously shown that the HBcAg can act as a T-cell-independent antigen in athymic mice (21). To determine whether WHcAg possessed this characteristic, euthymic and athymic BALB/c mice were immunized with WHcAg or WHeAg and in vivo antibody production levels were determined (Table 4). Whereas both the WHcAg and the WHeAg were immunogenic in euthymic BALB/c mice, only the stable particulate WHcAg elicited anti-WHc antibodies in athymic BALB/ c (nu/nu) mice, although these were of a significantly lower titer (i.e., 1:10,240) compared to euthymic mouse results, indicating the important role played by Th cells in the humoral response to WHcAg.
Comparative Th cell responses to hepadnavirus core proteins. The eight H-2 congenic murine strains were immunized i.p. with 10 μg of HBcAg, WHcAg, GScAg, or AScAg emulsified in IFA, and 4 to 6 weeks later spleen cells were harvested and cultured with various concentrations of the core particles used for immunization and IFN- production was measured in the culture supernatants. Figure 8 depicts the results obtained with B10.D2 mice. A single injection of each of the four core proteins elicited significant splenic IFN- production that was recalled by the homologous core protein in vitro and was core antigen dose dependent. In similarity to the findings for anti-core antibody production, all the eight H-2 congenic strains were capable of producing core-specific IFN- in recall cultures (see Fig. 9).
It was also of interest to determine the degree of cross-reactivity between the HBcAg and the rodent core proteins at the level of Th cell recognition in a common neutral (i.e., nonhuman, nonwoodchuck, nonsquirrel) host, in this case, the mouse. Splenic Th cells from the eight H-2 congenic strains primed with HBcAg, WHcAg, GScAg, or AScAg were cultured with the heterologous core proteins as well as with the same core protein used for immunization, and comparative IFN- production recalled by the various in vitro antigens (1.0 μg/ml) was determined (Fig. 9). Regarding the Th cell cross-reactivity between HBcAg and the rodent core proteins, two basic patterns were observed. The first pattern was illustrated by the first four strains shown in Fig. 9, which demonstrate little to no cross-reactivity between HBcAg and the rodent core proteins regardless of whether HBcAg was used as the immunogen and the rodent core proteins were the recall antigens in vitro or vice versa. The second pattern, which is characterized by limited but significant Th cell cross-reactivity between HBcAg and the rodent core proteins, especially WHcAg, is illustrated most prominently by the B10.S and B10 strains and somewhat by the B10.PL and B10.RIII strains (Fig. 9, bottom panels). The Th cell cross-reactivity between HBcAg and WHcAg occurs in both directions in the B10.S and B10 strains and is more unidirectional (i.e., HBcAg-primed Th cells are recalled in vitro with WHcAg more efficiently than WHcAg-primed Th cells are recalled in vitro by HBcAg) in the case of the B10.PL and B10.RIII strains.
The fine specificity of Th cell recognition of HBcAg and WHcAg in the various H-2 congenic strains most likely explains the two patterns of cross-reactivity. For example, the predominant Th cell recognition site on HBcAg for B10.S (H-2s) mice is residues 120 to 131 and this sequence is highly conserved between HBcAg and WHcAg (Fig. 10). In fact, immunization of B10.S mice with WHcAg primes Th cells with specificity for the woodchuck p120-to-p131 sequence and these Th cells are also efficiently recalled by HBcAg in vitro (Fig. 10a). In contrast, in the B10.D2 (H-2d) strain the predominant Th cell recognition site on WHcAg is Wp60 to Wp80 and this region is not well conserved (33.3%) on HBcAg; therefore, WHcAg-primed Th cells are poorly cross-reactive on HBcAg (Fig. 10b). Reciprocally, a predominant Th cell recognition site on HBcAg in B10.D2 mice is 85 to 100 (22), which is only 50% conserved on WHcAg (See Fig. 4 for sequence alignments). In the B10.M (H-2f) strain WHcAg-specific Th cell recognition is also dominantly focused on the Wp60-to-Wp80 region; therefore, WHcAg-primed Th cells are not recalled efficiently by the HBcAg in vitro (Fig. 10c). The predominant Th cell recognition site on HBcAg in the B10.M strain is p100 to p120 (22). Therefore, Th cell cross-reactivity between HBcAg and WHcAg is only partial and variable between the murine strains and is dependent upon the T-cell site recognized, which is dictated by the MHC-II genotype.
Regarding the Th cell cross-reactivity among the rodent core proteins, as anticipated by their higher degree of sequence homology, a higher level of Th cell cross-reactivity is apparent between the rodent core proteins than between HBcAg and the rodent core proteins (Fig. 9). However, given the approximately 90 to 92% sequence homology between the rodent core proteins, the Th cell cross-reactivity is actually less pronounced than would be expected. The explanation for the unexpected heterogeneity in Th cell fine specificity among the rodent core proteins most likely is the fact that in all eight H-2 congenic strains at least one Th cell recognition site maps within residues 60 to 80 (manuscript in preparation). The 60-to-80 sequence is the least conserved region among the rodent core proteins; note that it incorporates a large part of the highly variable external loop domain that is also the focus of B-cell recognition, which is also variable between rodent core proteins (Fig. 3 and 5).
Ability of rodent core particles to function as carrier platforms for heterologous epitopes. Because of its ability to tolerate the insertion of heterologous sequences and retain self-assembly into chimeric virus-like particles (VLPs), HBcAg has often been used as a carrier platform for vaccine design (41). Although the rodent core proteins are minimally cross-reactive with HBcAg at the B-cell level and only partially cross-reactive at the Th cell level, the rodent core particles do possess a number of immunogenic characteristics (described herein) that suggest that they may also have the ability to serve as effective vaccine carrier platforms for heterologous epitopes. To begin to address this question, we inserted a well-defined P. falciparum malaria circumsporozoite (CS)-neutralizing repeat sequence (20, 47) into the external loop regions of HBcAg at residue 78, WHcAg at residue 78, and GScAg at residue 83 and compared the humoral response to these chimeric VLPs. First, all three core proteins tolerated the internal insertion of the malaria CS repeat sequence and retained the ability to self-assemble into chimeric VLPs (data not shown). The hybrid VLPs were used to immunize groups of mice. The serum IgG antibody responses to the core carriers and the malaria insert were measured by ELISA. As shown in Fig. 11, a single 20-μg dose of the HBc-M78 VLP emulsified in IFA elicited anti-HBc antibodies, which cross-reacted minimally on WHcAg, as well as anti-insert antibodies [measured on an (NANP)5 synthetic peptide]. The anti-insert antibody level increased temporally and peaked at 8 weeks postimmunization at an end point serum titer of 1:625,000. Similarly, the WHc-M78 VLP functioned as an efficient carrier platform and actually elicited anti-carrier and anti-insert antibodies with higher titers than the HBc-M78 particle (Fig. 11). The anti-WHc antibodies were minimally cross-reactive on HBcAg. The GSc-M83 chimeric particle was also capable of eliciting anti-insert antibody, and the peak anti-insert response was reduced (i.e., 1:125,000) compared to that seen with the other two platforms. The anti-GSc antibodies demonstrated significant cross-reactivity with WHcAg (fivefold lower titer) and relatively low cross-reactivity with HBcAg (125-fold lower titer compared to GScAg).
DISCUSSION
Previous studies have documented that WHcAg is a strong immunogen during natural WHV infections (28) and after immunization in woodchucks (34). However, comparison of the immunogenicity of WHcAg with the other rodent core proteins and with HBcAg in the same neutral (i.e., mouse) species has not been reported. It is clear from the comparative studies that the enhanced immunogenicity observed for HBcAg is not unique to the human core protein, and this characteristic is shared by WHcAg, GScAg, and AScAg. Similarly, no antibody or Th cell nonresponder MHC haplotypes, among the eight H-2 congenic strains tested, were identified for any of the rodent core proteins, as previously shown for the HBcAg. Although core protein nonresponder phenotypes were not identified, there was a hierarchy of responder phenotypes among the H-2 congenic strains and high-, intermediate-, and low-responder status could be arbitrarily defined. However, even "low-responder" strains produced anti-core IgG serum end point titers from 1:25,000 to 1:250,000 as early as 4 weeks after a single injection of 10 μg of protein (Table 1) (Fig. 2). Another characteristic shared between HBcAg and WHcAg is the ability to function as a T-cell-independent antigen in the absence of Th cells (i.e., athymic mice). T-cell independence of WHcAg required a stable particulate structure because the truncated version (i.e., WHeAg) did not display T-cell independence.
Although the rodent core proteins may possess a number of characteristics in common with HBcAg, they also differ significantly as well. For example, the rodent core proteins do not share antibody cross-reactivity with HBcAg to any significant extent. This is especially true for WHcAg, whereas very low (<1.0%) cross-reactivity between HBcAg and the two squirrel core proteins was observed in several murine strains (Fig. 3). Serum cross-reactivity between core proteins was examined on relatively early (4-week) primary serum samples, because longer-term serum from hyperimmunized mice tends to contain antibodies to peptidic degradation products, which may overestimate the true serologic cross-reactivity. For example, immunization with denatured HBcAg results predominantly in antibody production to peptide epitopes present between residues 120 and 145 (3). The 120-to-145 region is relatively well conserved between HBcAg and the various rodent core proteins; indeed, anti-peptidic MAbs specific for this region in HBcAg cross-react on WHcAg and GScAg (3).
In contrast to the lack of antibody cross-reactivity between the HBcAg and the rodent core proteins, serologic cross-reactivity is relatively high among the rodent core proteins (Fig. 3 and 5). Cryoelectron microscopy studies have identified several B-cell epitopes within the loop region (aa 74 to 84) and outside the loop region (aa 20 to 22, aa 25 to 29, and aa 126 to 132) of HBcAg (2). As depicted in the sequence alignments (Fig. 4), the loop B-cell epitope region is poorly conserved between HBcAg and the rodent core proteins and is also not well conserved between the rodent core proteins. However, the nonloop, B-cell epitope regions are highly conserved between rodent core proteins but not between HBcAg (especially aa 20 to 22 and aa 25 to 29) and the rodent core proteins. Therefore, the serologic cross-reactivity between the rodent core proteins is likely due to nonloop, B-cell epitopes. To confirm this hypothesis, we disrupted the dominant loop region epitopes by inserting different heterologous sequences into the loop of WHcAg and GScAg to produce modified WHcAg and GScAg. These modified core particles were virtually 100% cross-reactive regardless of whether the antisera were produced against unmodified or modified rodent core particles (See Fig. 5). Because of the low level of sequence homology between HBcAg and the rodent core proteins in the identified B-cell epitope regions, little to no cross-reactivity exists within the loop B-cell epitopes or the nonloop, B-cell epitopes of HBcAg and the rodent core proteins. In summary, the comparative serologic data obtained using species variants of the hepatitis core proteins are consistent with the cryoelectron microscopy studies which mapped B-cell epitopes on the HBcAg (2). Of course, the serologic studies do not preclude the existence of additional nonloop region, B-cell epitopes on the rodent core proteins. The lack of antibody cross-reactivity between the HBcAg and the rodent core proteins is relevant to the consideration of using these core proteins as potential vaccine platforms as discussed below. In light of the absence of significant antibody cross-reactivity between the HBcAg and the rodent core proteins, it is interesting that the nonparticulate forms of the nucleoprotein (i.e., eAgs) do appear to share antibody cross-reactivity (Table 3) (Fig. 6 and 7). This suggests that "HBeAgicity" may not map within the loop region and is consistent with an earlier observation that deletion of a major portion of the HBcAg loop region did not affect "HBeAgicity" (33).
Analysis of hepatitis core protein cross-reactivity at the level of Th (CD4+) cell recognition is a bit more complicated than serologic cross-reactivity because Th cell recognition is MHC restricted. Therefore, the site recognized by Th cells is determined by MHC-II genotype and cross-reactivity between core proteins will depend on the conservation of amino acid sequence within the T-cell site. Thus, amino acid sequence differences between species variants of the hepatitis core proteins may affect the site selected for T-cell recognition by a given MHC genotype as well as the degree of Th cell cross-reactivity between core proteins. The influence of amino acid sequence variation among the HBcAg and rodent core proteins on Th cell site selection is clear from the observation that murine T-cell recognition of the WHcAg is highly focused on residues 60 to 80, whereas murine Th cell recognition of the HBcAg is focused on residues 85 to 140 (22) (See Fig. 4). The cross-reactivity between HBcAg- and WHcAg-specific Th cells can best be described as partial, depending on the conservation of the particular Th cell site recognized on the heterologous proteins. For example, Th cells of B10.S mice, which predominantly recognize aa 120 to 131 on HBcAg, are highly cross-reactive on WHcAg and GScAg because the 120-to-131 region is highly conserved between these core proteins. In contrast, T cells of B10.D2 and B10.M mice primed with WHcAg predominantly recognize aa 60 to 80 on WHcAg; since the 60-to-80 region is not well conserved between WHcAg and HBcAg it is not surprising that WHcAg and HBcAg are not very cross-reactive at the Th cell level in these strains. Of the eight H-2 congenic murine strains examined, significant Th cell cross-reactivity between the HBcAg and the WHcAg was observed only in the B10.S and the B10 strains, which predominantly recognize the 120- to 140-aa region on HBcAg which is conserved between HBcAg and the WHcAg. Significant differences in Th cell recognition sites between the HBcAg and the rodent core proteins may have important implications in the selection of a hepatitis core protein for use as a vaccine platform, as discussed later.
The HBcAg has been used as a carrier moiety for HBV and non-HBV B-cell epitopes in the design of candidate vaccines (41). HBcAg has been proposed as a vaccine platform largely because of its ability to self-assemble even after insertions of heterologous sequence at the NH2 terminus and COOH terminus and internally in the monomer subunit and because of the enhanced immunogenicity of HBcAg. Because the rodent hepatitis core proteins are as immunogenic as and possibly more immunogenic than HBcAg, it was of interest to determine whether they could also function as efficient carrier moieties. Herein we demonstrated that both WHcAg and GScAg can self-assemble after the insertion of a malaria CS repeat sequence into the external loop region; furthermore, the modified core proteins elicited significant anti-insert antibody production after a single injection. In fact, the WHcAg- M78-modified particle elicited higher-titer anti-insert IgG production than HBcAg-M78 (Fig. 11). Because HBcAg is derived from a human pathogen there are a number of practical limitations to its use as a vaccine carrier in humans. For example, preexisting anti-HBc antibody in individuals previously infected with HBV may affect immune clearance of the modified HBcAg through the formation of immune complexes. Secondly, anti-HBc antibodies induced during vaccination with modified HBcAg may compromise the utility of the anti-HBc assay used for diagnostic purposes. Most importantly, the approximately 400 million global chronic carriers of the HBV demonstrate very poor T-cell responses to all HBV structural proteins, including HBcAg, due to immune tolerance and, therefore, an HBcAg-based vaccine is not likely to be very effective in these individuals. Immune tolerance to the HBcAg is especially relevant in countries where HBV infection is endemic, where chronic carrier rates can be as high as 20% of the population. Because the rodent core proteins are not derived from a human pathogen, their use as vaccine platforms may solve a number of the "preexisting immunity" problems associated with the use of HBcAg. For example, because the rodent core proteins are not significantly cross-reactive with HBcAg at the antibody level preexisting anti-HBc antibodies will not immune complex with modified rodent core particles and the anti-core antibodies induced by vaccination with modified rodent core particles should not compromise the anti-HBc diagnostic assay. Furthermore, because the HBcAg and rodent core proteins are minimally cross-reactive at the Th cell level the use of the rodent core proteins may offer a means of circumventing the Th cell tolerance to HBcAg present in chronic carriers of HBV. Although we have compared Th cell recognition between HBcAg and the rodent core proteins in mice, the predominant Th cell recognition sites on HBcAg recognized by humans have been mapped to aa 1 to 20, aa 50 to 69, and aa 117 to 130 (10). The first two of these HBcAg Th cell sites are not highly conserved on the rodent core proteins and predict low levels of Th cell cross-reactivity between HBcAg and the rodent core proteins in humans, as demonstrated herein for mice. Although these issues need to be rigorously investigated, recent additional studies suggest that the use of the rodent core proteins as carrier moieties offers a number of advantages relative to the use of the HBcAg for vaccine design (manuscript in preparation). Therefore, we have chosen to use the WHcAg to develop a novel vaccine carrier platform to address a second major problem that has plagued the HBcAg platform technology, namely, the "assembly problem" (see companion paper) (4).
ACKNOWLEDGMENTS
We thank Michael Roggendorf (Germany) for the infectious woodchuck sera, Helen Pederson for editorial assistance, and Rudy Garcia for technical assistance.
This work was supported by National Institutes of Health grants 5 R01 AI 49730 and 5 R01 AI 20720 and grants from the Swedish Cancer Foundation and Swedish Science Council.
REFERENCES
Aida, Y., and M. J. Pabst. 1990. Removal of endotoxin from protein solutions by phase separation using Triton X-114. J. Immunol. Methods 132:191-195.
Belnap, D. M., N. R. Watts, J. F. Conway, N. Cheng, S. J. Stahl, P. T. Wingfield, and A. C. Steven. 2003. Diversity of core antigen epitopes of hepatitis B virus. Proc. Natl. Acad. Sci. USA 100:10884-10889.
Bichko, V., F. Schodel, M. Nassal, E. Gren, I. Berzinsh, G. Borisova, S. Miska, D. L. Peterson, E. Gren, P. Pushko, et al. 1993. Epitopes recognized by antibodies to denatured core protein of hepatitis B virus. Mol. Immunol. 30:221-231.
Billaud, J.-N., D. Peterson, M. Barr, A. Chen, M. Sallberg, F. Garduno, P. Goldstein, W. McDowell, J. Hughes, J. Jones, and D. Milich. 2005. Combinatorial approach to hepadnavirus-like particle vaccine design. J. Virol. 79:13656-13666.
Bottcher, B., S. A. Wynne, and R. A. Crowther. 1997. Determination of the fold of the core protein of hepatitis B virus by electron cryomicroscopy. Nature 386:88-91.
Chang, S. F., H. J. Netter, M. Bruns, R. Schneider, K. Frolich, and H. Will. 1999. A new avian hepadnavirus infecting snow geese (Anser caerulescens) produces a significant fraction of virions containing single-stranded DNA. Virology 262:39-54.
Conway, J. F., N. Cheng, A. Zlotnick, P. T. Wingfield, S. J. Stahl, and A. C. Steven. 1997. Visualization of a 4-helix bundle in the hepatitis B virus capsid by cryo-electron microscopy. Nature 386:91-94.
Feitelson, M. A., M. M. Clayton, and B. Phimister. 1990. Monoclonal antibodies raised to purified woodchuck hepatitis virus core antigen particles demonstrate X antigen reactivity. Virology 177:357-366.
Feitelson, M. A., P. L. Marion, and W. S. Robinson. 1982. Core particles of hepatitis B virus and ground squirrel hepatitis virus. I. Relationship between hepatitis B core antigen- and ground squirrel hepatitis core antigen-associated polypeptides by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and tryptic peptide mapping. J. Virol. 43:687-696.
Ferrari, C., A. Bertoletti, A. Penna, A. Cavalli, A. Valli, G. Missale, M. Pilli, P. Fowler, T. Giuberti, F. V. Chisari, et al. 1991. Identification of immunodominant T cell epitopes of the hepatitis B virus nucleocapsid antigen. J. Clin. Investig. 88:214-222.
Grethe, S., J. O. Heckel, W. Rietschel, and F. T. Hufert. 2000. Molecular epidemiology of hepatitis B virus variants in nonhuman primates. J. Virol. 74:5377-5381.
Hoofnagle, J. H., R. J. Gerety, and L. F. Barker. 1973. Antibody to hepatitis-B-virus core in man. Lancet 2:869-873.
Kidd-Ljunggren, K., Y. Miyakawa, and A. H. Kidd. 2002. Genetic variability in hepatitis B viruses. J. Gen. Virol. 83:1267-1280.
Lanford, R. E., D. Chavez, K. M. Brasky, R. B. Burns III, and R. Rico-Hesse. 1998. Isolation of a hepadnavirus from the woolly monkey, a New World primate. Proc. Natl. Acad. Sci. USA 95:5757-5761.
Lanford, R. E., D. Chavez, R. Rico-Hesse, and A. Mootnick. 2000. Hepadnavirus infection in captive gibbons. J. Virol. 74:2955-2959.
Liu, S., R. Tobias, S. McClure, G. Styba, Q. Shi, and G. Jackowski. 1997. Removal of endotoxin from recombinant protein preparations. Clin. Biochem. 30:455-463.
Marion, P. L., L. S. Oshiro, D. C. Regnery, G. H. Scullard, and W. S. Robinson. 1980. A virus in Beechey ground squirrels that is related to hepatitis B virus of humans. Proc. Natl. Acad. Sci. USA 77:2941-2945.
Mason, W. S., G. Seal, and J. Summers. 1980. Virus of Pekin ducks with structural and biological relatedness to human hepatitis B virus. J. Virol. 36:829-836.
Milich, D. R., M. Chen, F. Schodel, D. L. Peterson, J. E. Jones, and J. L. Hughes. 1997. Role of B cells in antigen presentation of the hepatitis B core. Proc. Natl. Acad. Sci. USA 94:14648-14653.
Milich, D. R., J. Hughes, J. Jones, M. Sallberg, and T. R. Phillips. 2001. Conversion of poorly immunogenic malaria repeat sequences into a highly immunogenic vaccine candidate. Vaccine 20:771-788.
Milich, D. R., and A. McLachlan. 1986. The nucleocapsid of hepatitis B virus is both a T-cell-independent and a T-cell-dependent antigen. Science 234:1398-1401.
Milich, D. R., A. McLachlan, A. Moriarty, and G. B. Thornton. 1987. Immune response to hepatitis B virus core antigen (HBcAg): localization of T cell recognition sites within HBcAg/HBeAg. J. Immunol. 139:1223-1231.
Milich, D. R., A. McLachlan, S. Stahl, P. Wingfield, G. B. Thornton, J. L. Hughes, and J. E. Jones. 1988. Comparative immunogenicity of hepatitis B virus core and E antigens. J. Immunol. 141:3617-3624.
Milich, D. R., A. McLachlan, G. B. Thornton, and J. L. Hughes. 1987. Antibody production to the nucleocapsid and envelope of the hepatitis B virus primed by a single synthetic T cell site. Nature 329:547-549.
Millman, I., T. Halbherr, and H. Simmons. 1982. Immunological cross-reactivities of woodchuck and hepatitis B viral antigens. Infect. Immun. 35:752-757.
Norder, H., J. W. Ebert, H. A. Fields, I. K. Mushahwar, and L. O. Magnius. 1996. Complete sequencing of a gibbon hepatitis B virus genome reveals a unique genotype distantly related to the chimpanzee hepatitis B virus. Virology 218:214-223.
Ponzetto, A., P. J. Cote, E. C. Ford, R. Engle, J. Cicmanec, M. Shapiro, R. H. Purcell, and J. L. Gerin. 1985. Radioimmunoassay and characterization of woodchuck hepatitis virus core antigen and antibody. Virus Res. 2:301-315.
Ponzetto, A., P. J. Cote, E. C. Ford, R. H. Purcell, and J. L. Gerin. 1984. Core antigen and antibody in woodchucks after infection with woodchuck hepatitis virus. J. Virol. 52:70-76.
Prassolov, A., H. Hohenberg, T. Kalinina, C. Schneider, L. Cova, O. Krone, K. Frolich, H. Will, and H. Sirma. 2003. New hepatitis B virus of cranes that has an unexpected broad host range. J. Virol. 77:1964-1976.
Pult, I., H. J. Netter, M. Bruns, A. Prassolov, H. Sirma, H. Hohenberg, S. F. Chang, K. Frolich, O. Krone, E. F. Kaleta, and H. Will. 2001. Identification and analysis of a new hepadnavirus in white storks. Virology 289:114-128.
Sallberg, M., U. Ruden, L. O. Magnius, E. Norrby, and B. Wahren. 1991. Rapid "tea-bag" peptide synthesis using 9-fluorenylmethoxycarbonyl (Fmoc) protected amino acids applied for antigenic mapping of viral proteins. Immunol. Lett. 30:59-68.
Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning, a laboratory manual, 2nd ed., vol. 3. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Schodel, F., A. M. Moriarty, D. L. Peterson, J. A. Zheng, J. L. Hughes, H. Will, D. J. Leturcq, J. S. McGee, and D. R. Milich. 1992. The position of heterologous epitopes inserted in hepatitis B virus core particles determines their immunogenicity. J. Virol. 66:106-114.
Schodel, F., G. Neckermann, D. Peterson, K. Fuchs, S. Fuller, H. Will, and M. Roggendorf. 1993. Immunization with recombinant woodchuck hepatitis virus nucleocapsid antigen or hepatitis B virus nucleocapsid antigen protects woodchucks from woodchuck hepatitis virus infection. Vaccine 11:624-628.
Seeger, C. 1991. Hepadnavirus replication, p. 213-226. Molecular biology of the Hepatitis B virus. CRC Press, Boca Raton, Fla.
Shanmuganathan, S., J. A. Waters, P. Karayiannis, M. Thursz, and H. C. Thomas. 1997. Mapping of the cellular immune responses to woodchuck hepatitis core antigen epitopes in chronically infected woodchucks. J. Med. Virol. 52:128-135.
Sprengel, R., E. F. Kaleta, and H. Will. 1988. Isolation and characterization of a hepatitis B virus endemic in herons. J. Virol. 62:3832-3839.
Stannard, L. M., O. Hantz, and C. Trepo. 1983. Antigenic cross-reactions between woodchuck hepatitis virus and human hepatitis B virus shown by immune electron microscopy. J. Gen. Virol. 64(Pt. 4):975-980.
Summers, J., J. M. Smolec, and R. Snyder. 1978. A virus similar to human hepatitis B virus associated with hepatitis and hepatoma in woodchucks. Proc. Natl. Acad. Sci. USA 75:4533-4537.
Testut, P., C. A. Renard, O. Terradillos, L. Vitvitski-Trepo, F. Tekaia, C. Degott, J. Blake, B. Boyer, and M. A. Buendia. 1996. A new hepadnavirus endemic in arctic ground squirrels in Alaska. J. Virol. 70:4210-4219.
Ulrich, R., M. Nassal, H. Meisel, and D. H. Kruger. 1998. Core particles of hepatitis B virus as carrier for foreign epitopes. Adv. Virus Res. 50:141-182.
Vaudin, M., A. J. Wolstenholme, K. N. Tsiquaye, A. J. Zuckerman, and T. J. Harrison. 1988. The complete nucleotide sequence of the genome of a hepatitis B virus isolated from a naturally infected chimpanzee. J. Gen. Virol. 69(Pt. 6):1383-1389.
Verschoor, E. J., K. S. Warren, S. Langenhuijzen, Heriyanto, R. A. Swan, and J. L. Heeney. 2001. Analysis of two genomic variants of orangutan hepadnavirus and their relationship to other primate hepatitis B-like viruses. J. Gen. Virol. 82:893-897.
Warren, K. S., J. L. Heeney, R. A. Swan, Heriyanto, and E. J. Verschoor. 1999. A new group of hepadnaviruses naturally infecting orangutans (Pongo pygmaeus). J. Virol. 73:7860-7865.
Werner, B. G., J. M. Smolec, R. Snyder, and J. Summers. 1979. Serological relationship of woodchuck hepatitis virus to human hepatitis B virus. J. Virol. 32:314-322.
Wynne, S. A., R. A. Crowther, and A. G. Leslie. 1999. The crystal structure of the human hepatitis B virus capsid. Mol. Cell 3:771-780.
Zavala, F., J. P. Tam, P. J. Barr, P. J. Romero, V. Ley, R. S. Nussenzweig, and V. Nussenzweig. 1987. Synthetic peptide vaccine confers protection against murine malaria. J. Exp. Med. 166:1591-1596.
Zhou, Y. J. 1980. A virus possibly associated with hepatitis and hepatoma in ducks. Shangai Med. J. 3:641-644.(Jean-Noel Billaud, Darrel)
Department of Biochemistry and Molecular Biophysics, Commonwealth University, Richmond, Virginia
Merck Research Laboratories, West Point, Pennsylvania
Division of Clinical Virology, Karolinska Institute, Karolinska University Hospital, Huddinge, Sweden
ABSTRACT
The hepatitis B virus core protein (HBcAg) is a uniquely immunogenic particulate antigen and as such has been used as a vaccine carrier platform. The use of other hepadnavirus core proteins as vaccine carriers has not been explored. To determine whether the rodent hepadnavirus core proteins derived from the woodchuck (WHcAg), ground squirrel (GScAg), and arctic squirrel (AScAg) viruses possess immunogen characteristics similar to those of HBcAg, comparative antigenicity and immunogenicity studies were performed. The results indicate that (i) the rodent core proteins are equal in immunogenicity to or more immunogenic than HBcAg at the B-cell and T-cell levels; (ii) major histocompatibility complex (MHC) genes influence the immune response to the rodent core proteins (however, nonresponder haplotypes were not identified); (iii) WHcAg can behave as a T-cell-independent antigen in athymic mice; (iv) the rodent core proteins are not significantly cross-reactive with the HBcAg at the antibody level (however, the nonparticulate "eAgs" do appear to be cross-reactive); (v) the rodent core proteins are only partially cross-reactive with HBcAg at the CD4+ T-cell level, depending on MHC haplotype; and (vi) the rodent core proteins are competent to function as vaccine carrier platforms for heterologous, B-cell epitopes. These results have implications for the selection of an optimal hepadnavirus core protein for vaccine design, especially in view of the "preexisting" immunity problem that is inherent in the use of HBcAg for human vaccine development.
INTRODUCTION
The virus family Hepadnaviridae includes hepatotropic, partially double-stranded DNA viruses with a replication strategy unique for animal DNA viruses consisting of reverse transcription of an RNA intermediate (35). This family is divided into two groups, the Orthohepadnavirus genus and the Avihepadnavirus genus. In addition to being identified in humans (hepatitis B virus [HBV]), orthohepadnaviruses have been identified in rodents such as woodchucks (woodchuck hepatitis virus [WHV]) (39) and ground (17) and arctic (40) squirrels (ground squirrel hepatitis virus and arctic squirrel hepatitis virus) and more recently in Old World as well as New World primates such as woolly monkeys (14), orangutans (43, 44), gorillas (11), chimpanzees (26, 42), and gibbons (15). The first avian hepadnavirus (duck hepatitis B virus) was identified in Pekin ducks (18, 48). Avian hepadnaviruses have also been isolated from other avian species such as the gray heron (37), Ross' goose and snow goose (6), and white stork (30) and most recently from cranes (29). The nonhuman primate viruses are most closely related to HBV, and their structural proteins share antigenic cross-reactivity. The rodent hepadnaviruses are more distantly related to HBV (55 to 70% nucleotide identity), and the avian hepadnaviruses are highly divergent from HBV (approximately 40% nucleotide identity) (13).
Interestingly, the nucleocapsid of HBV, the hepatitis B core antigen (HBcAg), is an extremely powerful immunogen and is significantly more immunogenic than HBV envelope (HBsAg) proteins during natural infection (12) and after immunization of recombinant proteins in mice (24), although both HBcAg and HBsAg are particulate antigens. For example, in contrast to HBsAg, HBcAg elicits immunoglobulin G (IgG) and IgM anti-HBc antibody production in athymic (i.e., T-cell-independent) mice (21); HBcAg preferentially activates Th1-type T cells (23); HBcAg up-regulates B7.1 and B7.2 costimulatory molecules on resting B cells (19); and HBcAg is an efficient vaccine platform to carry heterologous epitopes (41). Because these immunologic characteristics are unique to the particulate HBcAg and do not pertain to a nonparticulate secreted form of this protein, designated HBeAg, structural characteristics of the HBcAg may explain its enhanced immunogenicity. Recent cryoelectron microscopy (5, 7) and crystallographic (46) studies have elucidated the structure of HBcAg. A clustering of dimer subunits produces spikes on the surface of the core shell, which consist of radial bundles of four long -helices (5, 7). The orientation of the array of protein spikes distributed over the surface of the HBcAg particle may be optimal for cross-linking B-cell membrane immunoglobulin antigen receptors (19), especially because dominant B-cell epitopes appear to be positioned on or near the tip of the spikes (2). Similar structural analyses of the other orthohepadnavirus core particles have not been performed.
Therefore, to determine whether the immunologic characteristics are unique to HBcAg we have performed immunogenicity-antigenicity studies with mice by comparing HBcAg with the core proteins derived from the rodent orthohepadnaviruses, namely, WHV (WHcAg), ground squirrel hepatitis virus (GScAg), and arctic squirrel hepatitis virus (AScAg). We have compared (i) relative degrees of immunogenicity at the B- and T-cell levels; (ii) major histocompatibility complex (MHC) influence on responsiveness; (iii) T-cell independence; (iv) antigenic cross-reactivity at the B- and T-cell levels; and (v) the relative ability to function as vaccine carrier platforms for heterologous epitopes. For this purpose we have produced a panel of recombinant native and modified () core particles derived from these four hepadnaviruses. There is no current consensus in the literature regarding the serologic relatedness of orthohepadnavirus core proteins. Although most earlier studies suggested low to no cross-reactivity between antibodies specific for HBcAg and the rodent hepadnavirus core proteins (8, 9, 25, 27, 28, 34), several studies have indicated a greater degree of antibody cross-reactivity (36, 38, 45). Early serologic comparisons between hepadnaviral core proteins often employed infectious sera and/or partially purified core proteins and various experimental assays. The availability of recombinant core proteins derived from each viral species and monoclonal anti-HBc/e reagents and murine polyclonal species- specific anti-core antisera has made these comparisons less problematic and the quantitation of relative cross-reactivity feasible. The CD4+ T-cell responses to the various orthohepadnavirus core proteins have not been compared in the same species (i.e., mouse) previously.
MATERIALS AND METHODS
Animals. The C57BL/10 (B10), B10.S, and (B10 x B10.s)F1 mice were obtained from the breeding colony of the Vaccine Research Institute of San Diego. The B10, H-2 congenic strain (B10.BR, B10.D1, B10.D2, B10.M, B10.PL, and B10.RIII) and BALB/c euthymic and athymic mice were obtained from the Jackson Laboratory (Bar Harbor, ME). The acutely infected woodchuck sera were kindly provided by Michael Roggendorf (Institute of Virology, University of Essen, Essen, Germany). All animal care was performed according to National Institutes of Health standards as set forth in the Guide for the Care and Use of Laboratory Animals (1996).
Recombinant core proteins and synthetic peptides. Full-length recombinant HBcAg of the ayw3 subtype and a recombinant HBeAg corresponding in sequence to serum-derived HBeAg encompassing the 10 precore amino acids remaining after cleavage of the precursor and residues 1 to 149 of HBcAg were produced as described below. The presence of the 10 precore amino acids prevents particle assembly, and HBeAg is recognized efficiently by HBeAg-specific MAbs but displays little HBc antigenicity.
The recombinant rodent core proteins WHcAg, WHeAg, GScAg, and AScAg and the "modified" core proteins WHcAg, GScAg, and WHeAg were expressed or derived (GScAg and AScAg) from the pUC-WHcAg vector expressing the full-length woodchuck core. Briefly, the full-length sequence of WHcAg (NCBI accession no. NC_004107) served as a template to create synthetically the two other rodent full-length core antigens correlating to their accession numbers (for GScAg, accession no. NP_040993; for AScAg, accession no. NC_001719). The core proteins were either full length or truncated at amino acid (aa) 149 and had a cysteine added at residue 150 (i.e., 150C). Addition of a cysteine at residue 150 has been previously shown to stabilize truncated HBcAg (20). WHeAg corresponds to a truncated and less stable (i.e., more prone to degradation in vitro) version of WHcAg (149 aa) and was obtained after excision of the BseAI-BamHI fragment of the full-length WHcAg and replaced after ligation with the following pair of oligonucleotides digested by BseAI-BamHI: 5' TCCGGAACATACAGTCATTTGAGGATCC 3' and 5' GGATCCTCAAATGACTGTATGTTCCGGA 3'.
WHcAg, GScAg, and WHeAg indicate, unless otherwise specified, the insertion of heterologous sequences of 20 to 30 aa in the loop of the corresponding core proteins. All insertions were accomplished using the EcoRI-XhoI sites specifically engineered to not disrupt the reading frame and to encode a linker Gly-Ile-Leu on the N terminus and Leu-Glu on the C terminus of the inserted heterologous epitopes. Hybrid core proteins containing a specific Plasmodium falciparum circumsporozoite (CS) repeat sequence (NANPNVDPNANP3) inserted in the external loop were also produced (i.e., HBc-M78, WHc-M78, and GS-M83). For greater detail regarding constructs, refer to the companion paper (4). All the core constructs have been sequenced in both directions (Retrogen, Inc., San Diego, Calif.). Production and purification of the plasmids (MoBio Laboratories, Inc., San Diego, Calif.) and ligation experiments (La Roche) were performed according to the manufacturer's procedures. Cloning and subcloning have been performed following protocols described by Sambrook et al. (32). The Top10 Escherichia coli strain was purchased from Invitrogen. Transformation of chemically competent Top10 by heat shock was carried out according to the protocol of the manufacturer (Invitrogen).
Purification of core antigens. The core proteins were precipitated from the bacterial lysate by the addition of solid ammonium sulfate to 45% saturation (277 g/liter). The precipitates were collected by centrifugation, redissolved in a minimum of buffer (10 mM sodium phosphate buffer, pH 6.8), and dialyzed extensively against the same buffer. The protein solutions were then applied to a Bio-Rad BioGel HTP hydroxyapatite column (5 cm by 5 to 10 cm, depending on amount of protein) and eluted with 50 mM sodium phosphate buffer, pH 6.8. The core antigens pass through unretained. The proteins were then applied to a Sepharose CL4B column (5 by 100 cm). Characterization of the purified core proteins is illustrated in Fig. 1.
Endotoxin was removed from the core preparations by a modification of a phase separation with Triton X-114 (1, 16). A solution of the protein at a concentration of 5 mg/ml was made with 1% Triton X-114 and incubated at 4 C for 30 min with constant stirring. The solution was then incubated at 37 C for 10 min and then centrifuged at 20,000 x g for 10 min. The protein solution was recovered from above the detergent. This procedure was repeated four times. Finally, the protein was precipitated by lowering the pH to 5. Residual detergent remained in solution. The protein was recovered by centrifugation and dissolved in endotoxin-free buffer. Prior to Triton X-114 treatment, the core preparations contained approximately 10 to 25 ng of endotoxin/μg HBcAg; after phase separation with Triton X-114, the endotoxin content was from 0.1 ng/μg HBcAg to an undetectable amount as determined by the QCL-1000 chromogenic Limulus amoebocyte lysate endpoint assay (Cambrex, East Rutherford, NJ).
Synthetic peptides derived from the HBcAg or WHcAg sequences were synthesized by the simultaneous peptide synthesis method as previously described (31).
Immunizations and serology. Groups of three to five mice were immunized intraperitoneally (i.p.) with the core proteins (usually 10 μg) emulsified in complete Freund's adjuvant or incomplete Freund's adjuvant (IFA) for both antibody production and T-cell experiments. For antibody experiments, mice were bled retro-orbitally and sera were pooled from each group. Individual mouse serums were tested periodically to confirm the fidelity of the pooled serum results. Anti-core or anti-insert IgG antibodies were measured in murine sera by an indirect solid-phase enzyme-linked immunosorbent assay (ELISA) by using the homologous or heterologous core proteins (50 ng/well) or synthetic peptides (0.5 μg/well), representing the inserted sequence, as solid-phase ligands as described previously (33). Serial dilutions of both test sera and preimmunization sera were made, and the data are expressed as antibody titers representing the reciprocals of the highest dilutions of sera required to yield an optical density at 492 nm (OD492) three times that seen with an equal dilution of preimmunization sera. IgG isotype-specific ELISAs were performed by using IgG1-, IgG2a-, IgG2b-, and IgG3-specific peroxidase-labeled secondary antibodies (Southern Biotechnology, Birmingham, AL). Because secondary antibody specific for woodchuck IgG is not available, protein A-peroxidase (Prot A-HROP; Sigma Chemical Co., St. Louis, Mo.) (1 μg/ml) was used to develop the ELISA for the WHV-infected woodchuck sera.
In vitro T-cell proliferation and cytokine assays. Spleen cells from groups of three each of the B10 H-2 congenic mice were harvested and pooled 4 to 6 weeks after immunization with the various core proteins. Spleen cells (5 x 105) were cultured with various concentrations of homologous or heterologous core proteins or synthetic peptides. For T-cell proliferation assays spleen cells were cultured for 96 h and during the final 16 h with 1μCi of [3H]thymidine (New England Nuclear) (6 to 7 Ci/μmol). The cells were harvested onto filter strips for determination of [3H]thymidine incorporation. The data are expressed as counts per minute corrected for background proliferation in the absence of Ag ( cpm). For cytokine assays, culture supernatants were harvested at 96 h and gamma interferon (IFN-) production was measured by a two-site ELISA using MAb 170 and a polyclonal goat anti-mouse IFN- (Genzyme Corp., Boston, MA).
RESULTS
Eight H-2 congenic murine strains which differ only in MHC haplotype were immunized i.p. with equal 10-μg doses of recombinant truncated (i.e., 150C) WHcAg, HBcAg, GScAg, and AScAg emulsified in IFA. As shown in Fig. 2, all four core proteins elicited strong antibody responses to the homologous protein as early as 4 weeks after a single injection. The rodent core proteins elicited responses equal to or higher than those seen with HBcAg in all eight strains. In fact, the rodent core proteins appeared to be somewhat more immunogenic in mice than HBcAg. Therefore, enhanced immunogenicity is not unique to HBcAg. This analysis also indicated that there was an absence of genetic nonresponders to WHcAg, GScAg, or AScAg, as previously reported for HBcAg (21). However, a hierarchy of responsiveness did exist among the different H-2 haplotypes for each core protein and the response hierarchy differed between the core proteins. The rank order of anticore production among H-2 congenic mice for each core protein is shown in Table 1. Because antibody production is largely regulated by T-helper (Th) cell function, which is MHC class II (MHC-II) restricted, the differing rank order of anti-core production among H-2 haplotypes between the four core proteins suggests major differences in the Th cell recognition of the various core proteins (see Fig. 9 and 10). The IgG isotype distribution of anti-core antibodies was remarkably similar for all four core proteins (i.e., IgG2b > IgG2a >> IgG1 > IgG3) (data not shown).
Cross-reactivity of anti-core antibodies. The primary polyclonal anti-core antibodies depicted in Fig. 2 were analyzed for cross-reactivity between the various core proteins (Fig. 3). The anti-HBc antibodies produced against HBcAg in all eight H-2 congenic strains demonstrated zero to relatively low cross-reactivity for the rodent core proteins. The absence of antibody cross-reactivity in either direction between HBcAg and WHcAg was especially clear. Low but still detectable antibody cross-reactivity between HBcAg and GScAg and between HBcAg and AScAg was present in at least some antisera from the various H-2 congenic strains. In the B10.M strain, for example, polyclonal anti-HBc antibodies were nonreactive with WHcAg but reacted with GScAg and AScAg with approximately 0.8% efficiency compared to HBcAg (i.e., endpoint titers of 1:5,000 versus 1:625,000). Reciprocally, B10.M polyclonal anti-GSc and anti-ASc antibodies cross-reacted with HBcAg with very low efficiency (i.e., 0.037%) (Fig. 3). In fact, the greatest degree of cross-reactivity of polyclonal anti-HBc antibodies for any rodent core protein was 0.8%, as demonstrated with the B10.RIII, B10.S, and B10 strains (Fig. 3). Given the 65 to 67% amino acid homology between the HBcAg and the rodent core proteins, the zero to low level of antibody cross-reactivity may appear somewhat surprising. However, when one considers that the dominant B-cell epitopes on HBcAg reside in the external loop region at the tip of the structural spikes (2) comprising amino acid 74 to 84, the sequence homology between HBcAg and the rodent core proteins is only 18.2% in this region (See Fig. 4 for core protein sequence alignments). An additional B-cell epitope (nonloop epitope) on HBcAg has been mapped to include residues 20 to 22 and 25 to 29 (2), for which the sequence homology between HBcAg and the rodent core proteins is only 37.5%. Considering the location of the B-cell epitopes mapped on HBcAg, the low level of antibody cross-reactivity between HBcAg and the rodent core proteins is consistent with the results of cryoelectron microscopy studies of HBcAg labeled with anti-HBc monoclonal Fab fragments (2). Confirming the lack of serologic cross-reactivity between HBcAg and WHcAg and GScAg observed with polyclonal anti-HBc antisera, a panel of anti-HBc-specific monoclonal antibodies also failed to recognize WHcAg and GScAg. Similarly, a commercial (Abbott Labs) human anti-HBc reagent recognized solid-phase HBcAg but not WHcAg in an ELISA format (Table 2).
In contrast to the lack of antibody cross-reactivity between HBcAg and the rodent core proteins, the rodent core proteins are significantly cross-reactive among themselves at the level of antibody recognition. In all eight H-2 congenic strains, antibodies elicited by immunization with WHcAg, GScAg, and AScAg are cross-reactive to some degree on the other two rodent proteins. Interestingly, anti-WHc polyclonal antibodies cross-reacted more efficiently on AScAg than on GScAg, as demonstrated with sera from six of the eight H-2 congenic strains (i.e., B10.BR, B10.D1, B10.D2, B10.M, B10.PL, and B10) (Fig. 3). Reciprocally, anti-ASc antibodies cross-reacted on WHcAg more efficiently than on GScAg in sera from five of the eight strains. While the amino acid sequences of the rodent core proteins are virtually identical in the nonloop, B-cell epitope regions (i.e., aa 20 to 22, 25 to 29, and 126 to 132), there are significant differences within the external loop region. Significantly, WHcAg and AScAg are more (i.e., 72.7%) homologous in a broadly defined loop region between residues 70 and 91 than are AScAg and GScAg (i.e., 63.6%), which most likely explains the serologic data.
To identify the contribution of the nonloop regions to the immunogenicity and cross-reactivity of the rodent core proteins, we examined the effect of abolishing the dominant loop epitope by inserting different foreign sequences into the loop region of WHcAg (WHcAg) and GScAg (GScAg) (i.e., an influenza A virus sequence was inserted into WHcAg and a malaria CS repeat sequence was inserted into GScAg). As shown in Fig. 5, polyclonal anti-HBc recognized the native GScAg with only low efficiency and did not recognize native WHcAg or modified WHcAg. Similarly, antisera raised against native or modified rodent core proteins did not recognize HBcAg efficiently. Anti-native WHc antibody recognized modified WHcAg 24-fold less efficiently than native WHcAg, demonstrating the dominance of the WHcAg loop epitope. However, note that recognition of GScAg was equivalent to that of the modified WHcAg and recognition of the modified GScAg was almost equivalent to that of the modified WHcAg by the anti-native WHc antisera (Fig. 5, second column). This suggests that WHcAg and GScAg are totally cross-reactive outside the loop region. As shown in column 3 of Fig. 5, antisera raised to modified WHcAg (i.e., nonloop anti- WHc) recognized native WHcAg, modified WHcAg, native GScAg, and modified GScAg equivalently, again indicating that the two rodent core proteins are fully cross-reactive outside of the loop region. Similarly, anti-native GSc antibody recognized GScAg 120-fold less efficiently than native GScAg but recognition of WHcAg and WHcAg was equivalent to that seen with GScAg by anti-native GSc antisera (Fig. 5, column 4). Finally, antisera raised to modified GScAg recognized modified and native WHcAg core proteins similarly and only fourfold less efficiently than GScAg and GScAg. Cumulatively, these results indicate that WHcAg and GScAg are fully cross-reactive outside of the loop region, which is consistent with the 100% sequence homology between these two proteins in the nonloop, B-cell epitope regions mapped on the HBcAg. Conversely, the lack of cross-reactivity between HBcAg and the rodent core proteins outside the loop region is also consistent with the relatively low (i.e., 37.5%) sequence homology in the aa regions 20 to 22 and 25 to 29. The relatively high (71.4%) sequence homology between HBcAg and the rodent core proteins in the aa 126 to 132 non-loop region suggests that this region may not contribute in a major way to a nonloop, B-cell epitope(s).
Cross-reactivity of anti-E antibodies. Although HBcAg and the rodent core proteins were from minimally cross-reactive to non-cross-reactive at the antibody level, it was of interest to determine whether the nonparticulate forms (i.e., HBeAg and WHeAg) were cross-reactive. A panel of eight HBeAg-specific MAbs were tested for binding to solid-phase HBeAg and WHeAg in an ELISA format (Table 3). All eight HBeAg-specific MAbs were able to bind solid-phase WHeAg with efficiency that was equal to or reduced compared to that seen with HBeAg. In a solution-phase inhibition ELISA format, soluble WHeAg inhibited both the binding of an HBeAg-specific MAb (CE1-E) and the ability of a polyclonal anti-HBe antibody to bind to solid-phase HBeAg; however, the right-shifted WHeAg concentration curves suggested low-affinity binding to the WHeAg compared to the HBeAg results (Fig. 6). To determine whether the apparent low-affinity binding of WHeAg to MAb CE1-E had a structural basis as opposed to merely reflecting reduced amino acid homology within the HBeAg/WHeAg epitope, a modified WHeAg was used as the inhibitor. The modified WHeAg contained a heterologous sequence inserted behind residue 78 in addition to lacking a C-terminal cysteine and was less stable than truncated WHeAg without an insert, as determined on the basis of immunogenicity data (i.e., it was less immunogenic). The modified WHeAg was significantly more inhibitory than WHeAg, which suggests that the "eAg" epitope recognized by MAb CE1-E is shared between HBeAg and WHeAg and can be more or less exposed depending on structure. Because similar results were obtained with the HBeAg-specific MAbs 904 and 905 (data not shown), which recognize distinct HBeAg epitopes, it appears that HBeAg and WHeAg share more than one "eAg" B-cell epitope.
Serologic cross-reactivity of infected woodchuck sera. An independent method of determining degrees of cross-reactivity between the capsid antigens of the HBV and the WHV is measurement of antibody specificities in sera from WHV-infected woodchucks. Therefore, sequential serum samples from two woodchucks infected at day 0 with 106 50% infective doses of a titrated WHV stock serum were tested for anti-WHc, anti-HBc, and anti-HBe antibody specificities (Fig. 7). Anti-WHc antibody was detectable at day 40 and peaked between days 80 and 100 after infection. Anti-WHc antibody cross-reacted on HBcAg very minimally, as the anti-WHc titers were 4 orders of magnitude higher than the anti-HBc titers in both animals. In contrast, anti-HBe antibodies of significantly higher titer than anti-HBc antibodies were observed in both woodchucks and likely represent anti-WHe antibody production cross-reacting on HBeAg. The use of WHeAg to detect anti-WHe antibody is problematic because of the very high anti-WHc antibody titers that may obscure anti-WHe detection. However, because anti-WHc antibody is non-cross-reactive on HBc/HBeAg and because HBeAg and WHeAg are cross-reactive, HBeAg can be used to detect anti-WHe antibodies in infected woodchuck sera.
WHcAg can act as a T-cell-independent antigen. We have previously shown that the HBcAg can act as a T-cell-independent antigen in athymic mice (21). To determine whether WHcAg possessed this characteristic, euthymic and athymic BALB/c mice were immunized with WHcAg or WHeAg and in vivo antibody production levels were determined (Table 4). Whereas both the WHcAg and the WHeAg were immunogenic in euthymic BALB/c mice, only the stable particulate WHcAg elicited anti-WHc antibodies in athymic BALB/ c (nu/nu) mice, although these were of a significantly lower titer (i.e., 1:10,240) compared to euthymic mouse results, indicating the important role played by Th cells in the humoral response to WHcAg.
Comparative Th cell responses to hepadnavirus core proteins. The eight H-2 congenic murine strains were immunized i.p. with 10 μg of HBcAg, WHcAg, GScAg, or AScAg emulsified in IFA, and 4 to 6 weeks later spleen cells were harvested and cultured with various concentrations of the core particles used for immunization and IFN- production was measured in the culture supernatants. Figure 8 depicts the results obtained with B10.D2 mice. A single injection of each of the four core proteins elicited significant splenic IFN- production that was recalled by the homologous core protein in vitro and was core antigen dose dependent. In similarity to the findings for anti-core antibody production, all the eight H-2 congenic strains were capable of producing core-specific IFN- in recall cultures (see Fig. 9).
It was also of interest to determine the degree of cross-reactivity between the HBcAg and the rodent core proteins at the level of Th cell recognition in a common neutral (i.e., nonhuman, nonwoodchuck, nonsquirrel) host, in this case, the mouse. Splenic Th cells from the eight H-2 congenic strains primed with HBcAg, WHcAg, GScAg, or AScAg were cultured with the heterologous core proteins as well as with the same core protein used for immunization, and comparative IFN- production recalled by the various in vitro antigens (1.0 μg/ml) was determined (Fig. 9). Regarding the Th cell cross-reactivity between HBcAg and the rodent core proteins, two basic patterns were observed. The first pattern was illustrated by the first four strains shown in Fig. 9, which demonstrate little to no cross-reactivity between HBcAg and the rodent core proteins regardless of whether HBcAg was used as the immunogen and the rodent core proteins were the recall antigens in vitro or vice versa. The second pattern, which is characterized by limited but significant Th cell cross-reactivity between HBcAg and the rodent core proteins, especially WHcAg, is illustrated most prominently by the B10.S and B10 strains and somewhat by the B10.PL and B10.RIII strains (Fig. 9, bottom panels). The Th cell cross-reactivity between HBcAg and WHcAg occurs in both directions in the B10.S and B10 strains and is more unidirectional (i.e., HBcAg-primed Th cells are recalled in vitro with WHcAg more efficiently than WHcAg-primed Th cells are recalled in vitro by HBcAg) in the case of the B10.PL and B10.RIII strains.
The fine specificity of Th cell recognition of HBcAg and WHcAg in the various H-2 congenic strains most likely explains the two patterns of cross-reactivity. For example, the predominant Th cell recognition site on HBcAg for B10.S (H-2s) mice is residues 120 to 131 and this sequence is highly conserved between HBcAg and WHcAg (Fig. 10). In fact, immunization of B10.S mice with WHcAg primes Th cells with specificity for the woodchuck p120-to-p131 sequence and these Th cells are also efficiently recalled by HBcAg in vitro (Fig. 10a). In contrast, in the B10.D2 (H-2d) strain the predominant Th cell recognition site on WHcAg is Wp60 to Wp80 and this region is not well conserved (33.3%) on HBcAg; therefore, WHcAg-primed Th cells are poorly cross-reactive on HBcAg (Fig. 10b). Reciprocally, a predominant Th cell recognition site on HBcAg in B10.D2 mice is 85 to 100 (22), which is only 50% conserved on WHcAg (See Fig. 4 for sequence alignments). In the B10.M (H-2f) strain WHcAg-specific Th cell recognition is also dominantly focused on the Wp60-to-Wp80 region; therefore, WHcAg-primed Th cells are not recalled efficiently by the HBcAg in vitro (Fig. 10c). The predominant Th cell recognition site on HBcAg in the B10.M strain is p100 to p120 (22). Therefore, Th cell cross-reactivity between HBcAg and WHcAg is only partial and variable between the murine strains and is dependent upon the T-cell site recognized, which is dictated by the MHC-II genotype.
Regarding the Th cell cross-reactivity among the rodent core proteins, as anticipated by their higher degree of sequence homology, a higher level of Th cell cross-reactivity is apparent between the rodent core proteins than between HBcAg and the rodent core proteins (Fig. 9). However, given the approximately 90 to 92% sequence homology between the rodent core proteins, the Th cell cross-reactivity is actually less pronounced than would be expected. The explanation for the unexpected heterogeneity in Th cell fine specificity among the rodent core proteins most likely is the fact that in all eight H-2 congenic strains at least one Th cell recognition site maps within residues 60 to 80 (manuscript in preparation). The 60-to-80 sequence is the least conserved region among the rodent core proteins; note that it incorporates a large part of the highly variable external loop domain that is also the focus of B-cell recognition, which is also variable between rodent core proteins (Fig. 3 and 5).
Ability of rodent core particles to function as carrier platforms for heterologous epitopes. Because of its ability to tolerate the insertion of heterologous sequences and retain self-assembly into chimeric virus-like particles (VLPs), HBcAg has often been used as a carrier platform for vaccine design (41). Although the rodent core proteins are minimally cross-reactive with HBcAg at the B-cell level and only partially cross-reactive at the Th cell level, the rodent core particles do possess a number of immunogenic characteristics (described herein) that suggest that they may also have the ability to serve as effective vaccine carrier platforms for heterologous epitopes. To begin to address this question, we inserted a well-defined P. falciparum malaria circumsporozoite (CS)-neutralizing repeat sequence (20, 47) into the external loop regions of HBcAg at residue 78, WHcAg at residue 78, and GScAg at residue 83 and compared the humoral response to these chimeric VLPs. First, all three core proteins tolerated the internal insertion of the malaria CS repeat sequence and retained the ability to self-assemble into chimeric VLPs (data not shown). The hybrid VLPs were used to immunize groups of mice. The serum IgG antibody responses to the core carriers and the malaria insert were measured by ELISA. As shown in Fig. 11, a single 20-μg dose of the HBc-M78 VLP emulsified in IFA elicited anti-HBc antibodies, which cross-reacted minimally on WHcAg, as well as anti-insert antibodies [measured on an (NANP)5 synthetic peptide]. The anti-insert antibody level increased temporally and peaked at 8 weeks postimmunization at an end point serum titer of 1:625,000. Similarly, the WHc-M78 VLP functioned as an efficient carrier platform and actually elicited anti-carrier and anti-insert antibodies with higher titers than the HBc-M78 particle (Fig. 11). The anti-WHc antibodies were minimally cross-reactive on HBcAg. The GSc-M83 chimeric particle was also capable of eliciting anti-insert antibody, and the peak anti-insert response was reduced (i.e., 1:125,000) compared to that seen with the other two platforms. The anti-GSc antibodies demonstrated significant cross-reactivity with WHcAg (fivefold lower titer) and relatively low cross-reactivity with HBcAg (125-fold lower titer compared to GScAg).
DISCUSSION
Previous studies have documented that WHcAg is a strong immunogen during natural WHV infections (28) and after immunization in woodchucks (34). However, comparison of the immunogenicity of WHcAg with the other rodent core proteins and with HBcAg in the same neutral (i.e., mouse) species has not been reported. It is clear from the comparative studies that the enhanced immunogenicity observed for HBcAg is not unique to the human core protein, and this characteristic is shared by WHcAg, GScAg, and AScAg. Similarly, no antibody or Th cell nonresponder MHC haplotypes, among the eight H-2 congenic strains tested, were identified for any of the rodent core proteins, as previously shown for the HBcAg. Although core protein nonresponder phenotypes were not identified, there was a hierarchy of responder phenotypes among the H-2 congenic strains and high-, intermediate-, and low-responder status could be arbitrarily defined. However, even "low-responder" strains produced anti-core IgG serum end point titers from 1:25,000 to 1:250,000 as early as 4 weeks after a single injection of 10 μg of protein (Table 1) (Fig. 2). Another characteristic shared between HBcAg and WHcAg is the ability to function as a T-cell-independent antigen in the absence of Th cells (i.e., athymic mice). T-cell independence of WHcAg required a stable particulate structure because the truncated version (i.e., WHeAg) did not display T-cell independence.
Although the rodent core proteins may possess a number of characteristics in common with HBcAg, they also differ significantly as well. For example, the rodent core proteins do not share antibody cross-reactivity with HBcAg to any significant extent. This is especially true for WHcAg, whereas very low (<1.0%) cross-reactivity between HBcAg and the two squirrel core proteins was observed in several murine strains (Fig. 3). Serum cross-reactivity between core proteins was examined on relatively early (4-week) primary serum samples, because longer-term serum from hyperimmunized mice tends to contain antibodies to peptidic degradation products, which may overestimate the true serologic cross-reactivity. For example, immunization with denatured HBcAg results predominantly in antibody production to peptide epitopes present between residues 120 and 145 (3). The 120-to-145 region is relatively well conserved between HBcAg and the various rodent core proteins; indeed, anti-peptidic MAbs specific for this region in HBcAg cross-react on WHcAg and GScAg (3).
In contrast to the lack of antibody cross-reactivity between the HBcAg and the rodent core proteins, serologic cross-reactivity is relatively high among the rodent core proteins (Fig. 3 and 5). Cryoelectron microscopy studies have identified several B-cell epitopes within the loop region (aa 74 to 84) and outside the loop region (aa 20 to 22, aa 25 to 29, and aa 126 to 132) of HBcAg (2). As depicted in the sequence alignments (Fig. 4), the loop B-cell epitope region is poorly conserved between HBcAg and the rodent core proteins and is also not well conserved between the rodent core proteins. However, the nonloop, B-cell epitope regions are highly conserved between rodent core proteins but not between HBcAg (especially aa 20 to 22 and aa 25 to 29) and the rodent core proteins. Therefore, the serologic cross-reactivity between the rodent core proteins is likely due to nonloop, B-cell epitopes. To confirm this hypothesis, we disrupted the dominant loop region epitopes by inserting different heterologous sequences into the loop of WHcAg and GScAg to produce modified WHcAg and GScAg. These modified core particles were virtually 100% cross-reactive regardless of whether the antisera were produced against unmodified or modified rodent core particles (See Fig. 5). Because of the low level of sequence homology between HBcAg and the rodent core proteins in the identified B-cell epitope regions, little to no cross-reactivity exists within the loop B-cell epitopes or the nonloop, B-cell epitopes of HBcAg and the rodent core proteins. In summary, the comparative serologic data obtained using species variants of the hepatitis core proteins are consistent with the cryoelectron microscopy studies which mapped B-cell epitopes on the HBcAg (2). Of course, the serologic studies do not preclude the existence of additional nonloop region, B-cell epitopes on the rodent core proteins. The lack of antibody cross-reactivity between the HBcAg and the rodent core proteins is relevant to the consideration of using these core proteins as potential vaccine platforms as discussed below. In light of the absence of significant antibody cross-reactivity between the HBcAg and the rodent core proteins, it is interesting that the nonparticulate forms of the nucleoprotein (i.e., eAgs) do appear to share antibody cross-reactivity (Table 3) (Fig. 6 and 7). This suggests that "HBeAgicity" may not map within the loop region and is consistent with an earlier observation that deletion of a major portion of the HBcAg loop region did not affect "HBeAgicity" (33).
Analysis of hepatitis core protein cross-reactivity at the level of Th (CD4+) cell recognition is a bit more complicated than serologic cross-reactivity because Th cell recognition is MHC restricted. Therefore, the site recognized by Th cells is determined by MHC-II genotype and cross-reactivity between core proteins will depend on the conservation of amino acid sequence within the T-cell site. Thus, amino acid sequence differences between species variants of the hepatitis core proteins may affect the site selected for T-cell recognition by a given MHC genotype as well as the degree of Th cell cross-reactivity between core proteins. The influence of amino acid sequence variation among the HBcAg and rodent core proteins on Th cell site selection is clear from the observation that murine T-cell recognition of the WHcAg is highly focused on residues 60 to 80, whereas murine Th cell recognition of the HBcAg is focused on residues 85 to 140 (22) (See Fig. 4). The cross-reactivity between HBcAg- and WHcAg-specific Th cells can best be described as partial, depending on the conservation of the particular Th cell site recognized on the heterologous proteins. For example, Th cells of B10.S mice, which predominantly recognize aa 120 to 131 on HBcAg, are highly cross-reactive on WHcAg and GScAg because the 120-to-131 region is highly conserved between these core proteins. In contrast, T cells of B10.D2 and B10.M mice primed with WHcAg predominantly recognize aa 60 to 80 on WHcAg; since the 60-to-80 region is not well conserved between WHcAg and HBcAg it is not surprising that WHcAg and HBcAg are not very cross-reactive at the Th cell level in these strains. Of the eight H-2 congenic murine strains examined, significant Th cell cross-reactivity between the HBcAg and the WHcAg was observed only in the B10.S and the B10 strains, which predominantly recognize the 120- to 140-aa region on HBcAg which is conserved between HBcAg and the WHcAg. Significant differences in Th cell recognition sites between the HBcAg and the rodent core proteins may have important implications in the selection of a hepatitis core protein for use as a vaccine platform, as discussed later.
The HBcAg has been used as a carrier moiety for HBV and non-HBV B-cell epitopes in the design of candidate vaccines (41). HBcAg has been proposed as a vaccine platform largely because of its ability to self-assemble even after insertions of heterologous sequence at the NH2 terminus and COOH terminus and internally in the monomer subunit and because of the enhanced immunogenicity of HBcAg. Because the rodent hepatitis core proteins are as immunogenic as and possibly more immunogenic than HBcAg, it was of interest to determine whether they could also function as efficient carrier moieties. Herein we demonstrated that both WHcAg and GScAg can self-assemble after the insertion of a malaria CS repeat sequence into the external loop region; furthermore, the modified core proteins elicited significant anti-insert antibody production after a single injection. In fact, the WHcAg- M78-modified particle elicited higher-titer anti-insert IgG production than HBcAg-M78 (Fig. 11). Because HBcAg is derived from a human pathogen there are a number of practical limitations to its use as a vaccine carrier in humans. For example, preexisting anti-HBc antibody in individuals previously infected with HBV may affect immune clearance of the modified HBcAg through the formation of immune complexes. Secondly, anti-HBc antibodies induced during vaccination with modified HBcAg may compromise the utility of the anti-HBc assay used for diagnostic purposes. Most importantly, the approximately 400 million global chronic carriers of the HBV demonstrate very poor T-cell responses to all HBV structural proteins, including HBcAg, due to immune tolerance and, therefore, an HBcAg-based vaccine is not likely to be very effective in these individuals. Immune tolerance to the HBcAg is especially relevant in countries where HBV infection is endemic, where chronic carrier rates can be as high as 20% of the population. Because the rodent core proteins are not derived from a human pathogen, their use as vaccine platforms may solve a number of the "preexisting immunity" problems associated with the use of HBcAg. For example, because the rodent core proteins are not significantly cross-reactive with HBcAg at the antibody level preexisting anti-HBc antibodies will not immune complex with modified rodent core particles and the anti-core antibodies induced by vaccination with modified rodent core particles should not compromise the anti-HBc diagnostic assay. Furthermore, because the HBcAg and rodent core proteins are minimally cross-reactive at the Th cell level the use of the rodent core proteins may offer a means of circumventing the Th cell tolerance to HBcAg present in chronic carriers of HBV. Although we have compared Th cell recognition between HBcAg and the rodent core proteins in mice, the predominant Th cell recognition sites on HBcAg recognized by humans have been mapped to aa 1 to 20, aa 50 to 69, and aa 117 to 130 (10). The first two of these HBcAg Th cell sites are not highly conserved on the rodent core proteins and predict low levels of Th cell cross-reactivity between HBcAg and the rodent core proteins in humans, as demonstrated herein for mice. Although these issues need to be rigorously investigated, recent additional studies suggest that the use of the rodent core proteins as carrier moieties offers a number of advantages relative to the use of the HBcAg for vaccine design (manuscript in preparation). Therefore, we have chosen to use the WHcAg to develop a novel vaccine carrier platform to address a second major problem that has plagued the HBcAg platform technology, namely, the "assembly problem" (see companion paper) (4).
ACKNOWLEDGMENTS
We thank Michael Roggendorf (Germany) for the infectious woodchuck sera, Helen Pederson for editorial assistance, and Rudy Garcia for technical assistance.
This work was supported by National Institutes of Health grants 5 R01 AI 49730 and 5 R01 AI 20720 and grants from the Swedish Cancer Foundation and Swedish Science Council.
REFERENCES
Aida, Y., and M. J. Pabst. 1990. Removal of endotoxin from protein solutions by phase separation using Triton X-114. J. Immunol. Methods 132:191-195.
Belnap, D. M., N. R. Watts, J. F. Conway, N. Cheng, S. J. Stahl, P. T. Wingfield, and A. C. Steven. 2003. Diversity of core antigen epitopes of hepatitis B virus. Proc. Natl. Acad. Sci. USA 100:10884-10889.
Bichko, V., F. Schodel, M. Nassal, E. Gren, I. Berzinsh, G. Borisova, S. Miska, D. L. Peterson, E. Gren, P. Pushko, et al. 1993. Epitopes recognized by antibodies to denatured core protein of hepatitis B virus. Mol. Immunol. 30:221-231.
Billaud, J.-N., D. Peterson, M. Barr, A. Chen, M. Sallberg, F. Garduno, P. Goldstein, W. McDowell, J. Hughes, J. Jones, and D. Milich. 2005. Combinatorial approach to hepadnavirus-like particle vaccine design. J. Virol. 79:13656-13666.
Bottcher, B., S. A. Wynne, and R. A. Crowther. 1997. Determination of the fold of the core protein of hepatitis B virus by electron cryomicroscopy. Nature 386:88-91.
Chang, S. F., H. J. Netter, M. Bruns, R. Schneider, K. Frolich, and H. Will. 1999. A new avian hepadnavirus infecting snow geese (Anser caerulescens) produces a significant fraction of virions containing single-stranded DNA. Virology 262:39-54.
Conway, J. F., N. Cheng, A. Zlotnick, P. T. Wingfield, S. J. Stahl, and A. C. Steven. 1997. Visualization of a 4-helix bundle in the hepatitis B virus capsid by cryo-electron microscopy. Nature 386:91-94.
Feitelson, M. A., M. M. Clayton, and B. Phimister. 1990. Monoclonal antibodies raised to purified woodchuck hepatitis virus core antigen particles demonstrate X antigen reactivity. Virology 177:357-366.
Feitelson, M. A., P. L. Marion, and W. S. Robinson. 1982. Core particles of hepatitis B virus and ground squirrel hepatitis virus. I. Relationship between hepatitis B core antigen- and ground squirrel hepatitis core antigen-associated polypeptides by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and tryptic peptide mapping. J. Virol. 43:687-696.
Ferrari, C., A. Bertoletti, A. Penna, A. Cavalli, A. Valli, G. Missale, M. Pilli, P. Fowler, T. Giuberti, F. V. Chisari, et al. 1991. Identification of immunodominant T cell epitopes of the hepatitis B virus nucleocapsid antigen. J. Clin. Investig. 88:214-222.
Grethe, S., J. O. Heckel, W. Rietschel, and F. T. Hufert. 2000. Molecular epidemiology of hepatitis B virus variants in nonhuman primates. J. Virol. 74:5377-5381.
Hoofnagle, J. H., R. J. Gerety, and L. F. Barker. 1973. Antibody to hepatitis-B-virus core in man. Lancet 2:869-873.
Kidd-Ljunggren, K., Y. Miyakawa, and A. H. Kidd. 2002. Genetic variability in hepatitis B viruses. J. Gen. Virol. 83:1267-1280.
Lanford, R. E., D. Chavez, K. M. Brasky, R. B. Burns III, and R. Rico-Hesse. 1998. Isolation of a hepadnavirus from the woolly monkey, a New World primate. Proc. Natl. Acad. Sci. USA 95:5757-5761.
Lanford, R. E., D. Chavez, R. Rico-Hesse, and A. Mootnick. 2000. Hepadnavirus infection in captive gibbons. J. Virol. 74:2955-2959.
Liu, S., R. Tobias, S. McClure, G. Styba, Q. Shi, and G. Jackowski. 1997. Removal of endotoxin from recombinant protein preparations. Clin. Biochem. 30:455-463.
Marion, P. L., L. S. Oshiro, D. C. Regnery, G. H. Scullard, and W. S. Robinson. 1980. A virus in Beechey ground squirrels that is related to hepatitis B virus of humans. Proc. Natl. Acad. Sci. USA 77:2941-2945.
Mason, W. S., G. Seal, and J. Summers. 1980. Virus of Pekin ducks with structural and biological relatedness to human hepatitis B virus. J. Virol. 36:829-836.
Milich, D. R., M. Chen, F. Schodel, D. L. Peterson, J. E. Jones, and J. L. Hughes. 1997. Role of B cells in antigen presentation of the hepatitis B core. Proc. Natl. Acad. Sci. USA 94:14648-14653.
Milich, D. R., J. Hughes, J. Jones, M. Sallberg, and T. R. Phillips. 2001. Conversion of poorly immunogenic malaria repeat sequences into a highly immunogenic vaccine candidate. Vaccine 20:771-788.
Milich, D. R., and A. McLachlan. 1986. The nucleocapsid of hepatitis B virus is both a T-cell-independent and a T-cell-dependent antigen. Science 234:1398-1401.
Milich, D. R., A. McLachlan, A. Moriarty, and G. B. Thornton. 1987. Immune response to hepatitis B virus core antigen (HBcAg): localization of T cell recognition sites within HBcAg/HBeAg. J. Immunol. 139:1223-1231.
Milich, D. R., A. McLachlan, S. Stahl, P. Wingfield, G. B. Thornton, J. L. Hughes, and J. E. Jones. 1988. Comparative immunogenicity of hepatitis B virus core and E antigens. J. Immunol. 141:3617-3624.
Milich, D. R., A. McLachlan, G. B. Thornton, and J. L. Hughes. 1987. Antibody production to the nucleocapsid and envelope of the hepatitis B virus primed by a single synthetic T cell site. Nature 329:547-549.
Millman, I., T. Halbherr, and H. Simmons. 1982. Immunological cross-reactivities of woodchuck and hepatitis B viral antigens. Infect. Immun. 35:752-757.
Norder, H., J. W. Ebert, H. A. Fields, I. K. Mushahwar, and L. O. Magnius. 1996. Complete sequencing of a gibbon hepatitis B virus genome reveals a unique genotype distantly related to the chimpanzee hepatitis B virus. Virology 218:214-223.
Ponzetto, A., P. J. Cote, E. C. Ford, R. Engle, J. Cicmanec, M. Shapiro, R. H. Purcell, and J. L. Gerin. 1985. Radioimmunoassay and characterization of woodchuck hepatitis virus core antigen and antibody. Virus Res. 2:301-315.
Ponzetto, A., P. J. Cote, E. C. Ford, R. H. Purcell, and J. L. Gerin. 1984. Core antigen and antibody in woodchucks after infection with woodchuck hepatitis virus. J. Virol. 52:70-76.
Prassolov, A., H. Hohenberg, T. Kalinina, C. Schneider, L. Cova, O. Krone, K. Frolich, H. Will, and H. Sirma. 2003. New hepatitis B virus of cranes that has an unexpected broad host range. J. Virol. 77:1964-1976.
Pult, I., H. J. Netter, M. Bruns, A. Prassolov, H. Sirma, H. Hohenberg, S. F. Chang, K. Frolich, O. Krone, E. F. Kaleta, and H. Will. 2001. Identification and analysis of a new hepadnavirus in white storks. Virology 289:114-128.
Sallberg, M., U. Ruden, L. O. Magnius, E. Norrby, and B. Wahren. 1991. Rapid "tea-bag" peptide synthesis using 9-fluorenylmethoxycarbonyl (Fmoc) protected amino acids applied for antigenic mapping of viral proteins. Immunol. Lett. 30:59-68.
Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning, a laboratory manual, 2nd ed., vol. 3. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.
Schodel, F., A. M. Moriarty, D. L. Peterson, J. A. Zheng, J. L. Hughes, H. Will, D. J. Leturcq, J. S. McGee, and D. R. Milich. 1992. The position of heterologous epitopes inserted in hepatitis B virus core particles determines their immunogenicity. J. Virol. 66:106-114.
Schodel, F., G. Neckermann, D. Peterson, K. Fuchs, S. Fuller, H. Will, and M. Roggendorf. 1993. Immunization with recombinant woodchuck hepatitis virus nucleocapsid antigen or hepatitis B virus nucleocapsid antigen protects woodchucks from woodchuck hepatitis virus infection. Vaccine 11:624-628.
Seeger, C. 1991. Hepadnavirus replication, p. 213-226. Molecular biology of the Hepatitis B virus. CRC Press, Boca Raton, Fla.
Shanmuganathan, S., J. A. Waters, P. Karayiannis, M. Thursz, and H. C. Thomas. 1997. Mapping of the cellular immune responses to woodchuck hepatitis core antigen epitopes in chronically infected woodchucks. J. Med. Virol. 52:128-135.
Sprengel, R., E. F. Kaleta, and H. Will. 1988. Isolation and characterization of a hepatitis B virus endemic in herons. J. Virol. 62:3832-3839.
Stannard, L. M., O. Hantz, and C. Trepo. 1983. Antigenic cross-reactions between woodchuck hepatitis virus and human hepatitis B virus shown by immune electron microscopy. J. Gen. Virol. 64(Pt. 4):975-980.
Summers, J., J. M. Smolec, and R. Snyder. 1978. A virus similar to human hepatitis B virus associated with hepatitis and hepatoma in woodchucks. Proc. Natl. Acad. Sci. USA 75:4533-4537.
Testut, P., C. A. Renard, O. Terradillos, L. Vitvitski-Trepo, F. Tekaia, C. Degott, J. Blake, B. Boyer, and M. A. Buendia. 1996. A new hepadnavirus endemic in arctic ground squirrels in Alaska. J. Virol. 70:4210-4219.
Ulrich, R., M. Nassal, H. Meisel, and D. H. Kruger. 1998. Core particles of hepatitis B virus as carrier for foreign epitopes. Adv. Virus Res. 50:141-182.
Vaudin, M., A. J. Wolstenholme, K. N. Tsiquaye, A. J. Zuckerman, and T. J. Harrison. 1988. The complete nucleotide sequence of the genome of a hepatitis B virus isolated from a naturally infected chimpanzee. J. Gen. Virol. 69(Pt. 6):1383-1389.
Verschoor, E. J., K. S. Warren, S. Langenhuijzen, Heriyanto, R. A. Swan, and J. L. Heeney. 2001. Analysis of two genomic variants of orangutan hepadnavirus and their relationship to other primate hepatitis B-like viruses. J. Gen. Virol. 82:893-897.
Warren, K. S., J. L. Heeney, R. A. Swan, Heriyanto, and E. J. Verschoor. 1999. A new group of hepadnaviruses naturally infecting orangutans (Pongo pygmaeus). J. Virol. 73:7860-7865.
Werner, B. G., J. M. Smolec, R. Snyder, and J. Summers. 1979. Serological relationship of woodchuck hepatitis virus to human hepatitis B virus. J. Virol. 32:314-322.
Wynne, S. A., R. A. Crowther, and A. G. Leslie. 1999. The crystal structure of the human hepatitis B virus capsid. Mol. Cell 3:771-780.
Zavala, F., J. P. Tam, P. J. Barr, P. J. Romero, V. Ley, R. S. Nussenzweig, and V. Nussenzweig. 1987. Synthetic peptide vaccine confers protection against murine malaria. J. Exp. Med. 166:1591-1596.
Zhou, Y. J. 1980. A virus possibly associated with hepatitis and hepatoma in ducks. Shangai Med. J. 3:641-644.(Jean-Noel Billaud, Darrel)