Intracellular Retention of Hepatitis B Virus Surface Proteins Reduces Interleukin-2 Augmentation after Genetic Immunizations

Abstract

We have previously shown that hepatitis B virus (HBV) surface antigens (HBsAgs) are highly immunogenic after genetic immunization. Compared to the secreted middle HBV surface proteins (MHBs) or small HBV surface proteins (SHBs), the nonsecreted large HBV surface protein (LHBs), however, induced significantly weaker humoral and cellular immune responses that could not be augmented by genetic coimmunizations with cytokine expression plasmids. In order to understand the mechanisms underlying this phenomenon, we examined the effect of coimmunizations with an interleukin-2 (IL-2) DNA expression plasmid on the immunogenicity at the B- and T-cell level of nonsecreted wild-type LHBs, a secreted mutant LHBs, wild-type SHBs, and a nonsecreted mutant SHBs. Coimmunizations of mice with plasmids encoding wild-type SHBs or the secreted mutant LHBs and IL-2 increased anti-HBs responses, helper T-cell proliferative activity and cytotoxic T-lymphocyte killing. By contrast, coimmunizations of plasmids encoding wild-type LHBs or nonsecreted mutant SHBs and IL-2 had no significant effects on immune responses. Interestingly, mice immunized with cytokine expression plasmids 14 days after the injection of the wild-type LHBs plasmid showed augmented immune responses compared to animals simultaneously injected with both expression constructs. Anti-HBs responses in mice injected with plasmids encoding secreted forms of HBsAgs were detectable about 10 days earlier than those in mice immunized with plasmids encoding nonsecreted forms of HBsAgs. Based on these observations, we conclude that cytokines produced by DNA plasmids at the initial site of antigen presentation cannot augment LHBs specific immune responses because LHBs is not produced at high enough levels or is not accessible for uptake by antigen-presenting cells.

Hepatitis B virus (HBV) is a noncytopathic, hepatotropic virus. Worldwide, more than 350 million individuals are infected (21). HBV is a leading cause of chronic hepatitis, cirrhosis, and hepatocellular carcinoma (3, 19). The cellular immune response to HBV is thought to be responsible for viral clearance and pathogenesis of liver disease, including hepatocellular carcinoma. The observation of spontaneous HBV clearance in some chronically infected individuals implies that the suboptimal cellular immune response may be reversible. Therefore, strategies designed to boost the HBV-specific T-cell immune response, to alter the balance between the cytopathic and the regulatory components of the response, or to mimic the regulatory functions of the T-cell response in the liver may terminate persistent infection.

For these reasons, we chose genetic immunization as an immunotherapeutic approach to chronic HBV infection because this approach offers the potential advantage of inducing cellular and humoral immune responses against conserved viral epitopes because vaccination is based on DNA expression plasmids rather than proteins. This strategy involves the transfer of a viral gene into muscle cells and antigen-presenting cells by a plasmid vector with subsequent endogenous production and intracellular processing of the viral structural proteins into smaller antigenic peptides. Such peptides are subsequently expressed on the cell surface in the context of major histocompatibility complex molecules (23, 25) and therefore have been shown to induce CD8⁺ cytotoxic T-lymphocyte (CTL) and helper T-cell type 1 (TH1) responses against various viral antigens (24). Using this approach, several groups have demonstrated that HBV surface and nucleocapsid antigens are highly immunogenic at both the T-cell and B-cell levels in mice (6, 9, 12, 15, 22). Immunogenicity of the secreted middle HBV surface protein (MHBs) was significantly better than that of the nonsecreted large HBV surface protein (LHBs) (12).

In addition, recent studies demonstrated that coimmunization of interleukin-2 (IL-2) and granulocyte-macrophage colony-stimulating factor (GM-CSF) DNA expression plasmids enhanced humoral and cellular immune responses to rabies glycoprotein (26), HBV small surface protein (SHBs) and MHBs (5) and the hepatitis C virus core protein (8, 10). Different from the findings with MHBs DNA, coimmunizations of LHBs encoding DNA with either IL-2 or GM-CSF expression plasmids did not augment cellular and humoral immune responses to HBV envelope proteins (11). This finding was not due to an inhibition of the secretion of IL-2 and GM-CSF by LHBs. The effects of LHBs on the immune response augmenting properties of IL-2 and GM-CSF in vivo, therefore, were not related to inhibition of their secretion from the cell by LHBs. Conversely, IL-2, gamma interferon (INF-γ) and tumor necrosis factor alpha (TNF-α) did not down-regulate HBV surface gene expression in several mouse cell lines with different genetic backgrounds (11).

We and others have recently demonstrated that the anti-HBs response to an LHBs DNA expression construct is detectable about 10 to 14 days later than the responses to MHBs (6, 12). This may be due to the intracellular retention of LHBs in transfected muscle cells with a subsequent delay in accessibility of the antigen for the initiation of the immune response. To test this hypothesis, in the present study we designed plasmids producing a secreted and a nonsecreted form of LHBs and SHBs, respectively, without changing antigenicity and determined the immunogenicity of these proteins in vivo using the genetic immunization approach.

MATERIALS AND METHODS

DNA expression vectors.

pSVL encodes wild-type LHBs and carries mutations of both the MHBs and SHBs start codons to the threonine codon ACG. pSVblaL carries a bacterial β-lactamase secretion signal sequence upstream from the LHBs coding sequence and encodes a secreted LHBs. MHBs and SHBs start codons are mutated in the same manner as in pSVL. pSVs25L corresponds to pSVblaL but contains a truncated nonfunctional signal sequence as well as wild-type MHBs and SHBs start codons. These plasmids have been described in detail (1). The pSVBX24H vector encodes SHBs (14). The pSVBX24HSer65 construct corresponds to plasmid pSVBX24H, except for codon 65 of the S-gene, which has been mutated from cysteine (TGT) to serine (TCT) by site-directed mutagenesis (Kunkel method). This mutated SHBs protein is not secreted by the cell (16, 25a). The DNA expression construct coding for murine IL-2 (pcD/3-IL2) was cloned and purified as previously described (8).

Cell lines.

The mastocytoma cell line P815 (H-2^d) and the mouse myoblastoma cell line G8 were obtained from the American Type Culture Collection. P815L cells stably express LHBs (22).

Mice.

Female BALB/c (H-2^d) mice were kept in the animal facility of the University Hospital of Freiburg. Mice were obtained from Charles River Labs (Wilmington, Mass.) and used between the ages of 10 to 25 weeks.

In vitro studies.

Intracellular expression of mutant and wild-type LHBs and MHBs proteins in G8 cells after transfection using Lipofectamine (Gibco, Gaithersburg, Md.) with the corresponding DNA plasmids was determined by immunoblot analysis using the pre-S1-specific monoclonal antibody (MAb) MA18/7 (a gift from W. H. Gerlich) and a polyclonal goat anti-HBVs antiserum (Dako, Carpinteria, Calif.) as previously described (12).

The pcD/3-IL2 vector was transfected into G8 cells. Cell culture medium was collected 2 days later. IL-2 in cell culture supernatant was measured using a commercial enzyme-linked immunosorbent assay (ELISA) kit (R&D, Minneapolis, Minn.). To control for transfection efficiency, cells were cotransfected with a pSV-β-galactosidase expression vector (Promega, Madison, Wis.).

Genetic immunization.

To enhance cellular uptake of plasmid DNA, the quadriceps muscles of BALB/c mice were injected at multiple sites with a total of 100 μl of 0.25% bupivacaine per mouse. Five days later, the plasmid constructs were injected into the same region at five different sites in a final volume of 100 μl of 0.9% NaCl. After 24 days, the mice were sacrificed, and sera and spleen cells were collected for subsequent immunological assays. There were 10 groups of mice with five animals each. The mice received the following immunizations: group 1, 50 μg of pSVBX24H plus 50 μg of mock DNA; group 2, 50 μg of pSVBX24H plus 50 μg of pcD/3-IL2; group 3, 50 μg of pSVBX24HSer65 plus 50 μg of mock DNA; group 4, 50 μg of pSVBX24HSer65 plus 50 μg of pcD/3-IL2; group 5, 50 μg of pSVL plus 50 μg of mock DNA; group 6, 50 μg of pSVL plus 50 μg of pcD/3-IL2; group 7, 50 μg of pSVblaL plus 50 μg of mock DNA; and group 8, 50 μg of pSVblaL plus 50 μg of pcD/3-IL2. Group 9 was immunized with 50 μg of pSVL at day 1 and 50 μg of pcD/3-IL2 at day 14 into the same site. Group 10 was immunized with 100 μg of mock DNA plasmid vector (pSV65). Four additional groups of BALB/c mice each containing five animals were immunized with pSVBX24H, pSVBX24HSer65, pSVL, and pSVblaL for determination of anti-HBVs kinetics.

HBV serology.

Surface antigens in culture supernatant or lysates of transfected cells were quantitated by an ELISA recognizing conformational as well as linear epitopes on SHBs (Murex HBsAg; Murex Diagnostica GmbH, Burgwedel, Germany). The cell culture supernatants were collected 48 h after transfection, and cells were lysed using three cycles of freeze-thawing in phosphate-buffered saline (PBS) or RIPA-lysis buffer (50 mM Tris [pH 6.8], 100 mM dithiothreitol, 2% sodium dodecyl sulfate, 10% glycerol) for viral protein studies. Anti-HBs antibodies were measured with a commercial ELISA kit (AUSAB; Abbott Laboratories, Chicago, Ill.). In all experiments, unpooled individual sera of mice were assayed for anti-HBVs responses. For determination of anti-HBs isotypes, microtiter plates were coated with 1 μg of recombinant HBsAg subtype ad (Biodesign, Kennebunk, Maine), blocked with 10% fetal calf serum (FCS)–PBS and subsequently incubated with 50 μl of serial dilutions of serum samples at 4°C overnight. After washing, bound proteins were detected by horseradish peroxidase-conjugated anti-mouse immunoglobulin G1 (IgG1) or IgG2a antibodies (PharMingen, San Diego, Calif.), and o-phenylenediamine (Abbott, Chicago, Ill.) was subsequently used as a substrate for color development.

T-cell proliferation assay.

Mice were anesthetized with isoflurane (Aerrane, Anaquest, N.J.). Blood was removed by retrobulbar puncture, and spleen cells were harvested. For all T-cell proliferation and cytotoxicity assays, unpooled individual spleens of mice were used. Erythrocytes were lysed by incubation in 8.3% NH₄C1–0.17 M Tris (pH 7.4) for 10 min at 37°C. Spleen cells were cultured in triplicate in 96-well flat-bottomed plates at 5 × 10⁵ cells per well in 100 μl of complete Dulbecco’s modified Eagle medium (DMEM) (Gibco, Gaithersburg, Md.) containing 10% FCS. Spleen cells were stimulated with recombinant HBsAg subtype ad (Biodesign) at different concentrations (1 and 10 μg/ml). Finally, 2-mercaptoethanol was added to a final concentration of 50 μM. As a control for antigen specificity, effector cells were stimulated with 10 μg of recombinant hCGβ (the β subunit of human chorionic gonadotropin)/ml, which is secreted from the cell and has recently been shown to be a strong T-cell immunogen (13). Spleen cells were stimulated for 3 days. After the addition of bromodeoxyuridine (BrdU), cells were incubated for 10 h. BrdU incorporation into DNA was measured by ELISA using a commercial cell proliferation kit (Boehringer, Mannheim, Germany).

Cytotoxicity assay.

Spleen cells from immunized mice were suspended in complete DMEM containing 10% FCS and 50 μM 2-mercaptoethanol and analyzed for cytotoxic activity 5 days after in vitro stimulation. Recombinant murine IL-2 was added once at day 2 at a concentration of 10 U/ml, and responder cells (4 × 10⁷) were cocultured with 1 × 10⁷ irradiated syngeneic cells (8,000 rad) stably expressing LHBs (P815L). Cytotoxic effector lymphocyte populations were harvested after 6 days of incubation. A 5-h ⁵¹Cr-release assay was performed in a 96-well round-bottomed plate using as target cell lines ⁵¹Cr-labeled P815L or parental P815 cells. CTL assays were performed at lymphocyte effector/target (E:T) ratios of 20:1, 5:1, and 2:1. The HBV envelope specificity of CTL activity was confirmed when effector cells were also stimulated with irradiated parental P815 cells prior to the ⁵¹Cr-release assay. Results were expressed according to the following formula: percentage specific lysis = (experimental release − spontaneous release)/(maximum release − spontaneous release). Experimental release represents the mean counts per minute released by target cells in the presence of effector cells. Total release represents the radioactivity released after total lysis of target cells with 5% Triton X-100. Spontaneous release represents the radioactivity present in medium derived from target cells only.

Statistical analysis.

For comparison of results between the different groups, we used a nonparametric Mann-Whitney U test. P values of <0.05 were considered statistically significant. Numbers for P values according to CD4⁺ and CD8⁺ T-cell responses are derived from all rHBsAg concentrations used for stimulation and from E:T ratios, respectively.

RESULTS

Structure and characteristics of plasmids encoding wild-type and mutant HBsAgs.

To investigate the impact of secreted forms of HBsAgs on immune responses after cytokine genetic coimmunizations, we took a genetic approach and constructed and characterized the following proteins (Fig. 1): (i) a wild-type LHBs derivative which is not secreted by the cell and encoded by pSVL and (ii) a mutant LHBs carrying only external pre-S domains which is secreted and encoded by pSVblaL. For this purpose, the first 32 amino acids of the bacterial β-lactamase, which contain a secretion signal, were fused to amino acid 7 of the pre-S sequence. The N-terminal signal (blaL) causes the cotranslational entry of the pre-S domain into the endoplasmic reticulum (ER) lumen and is cleaved between amino acids 23 and 24 by a signal peptidase. (iii) A wild-type SHBs which is secreted by the cell and encoded by pSVBX24H. (iv) An altered SHBs which contains a mutation of cysteine at position 65 to serine, resulting in retention of the mutant SHBs in the ER membrane (25a). This plasmid was designated pSVBX24HSer65. It is important to note that the plasmids pSVL and pSVblaL described above carry mutations of both the pre-S2 and S start codons to the threonine codon ACG. Therefore, no internal translation initiation with the subsequent production of secreted MHBs and SHBs can occur.

FIG. 1.

Map of the transmembrane topology of HBV envelope proteins. DNA expression constructs. (A) DNA expression constructs with the corresponding HBV surface protein-encoding sequences (e.g., pre-S1, pre-S2, and S) and the corresponding intact or deleted ATG translation initiation start codons. (B) Expected transmembrane topology at the ER of the different HBsAgs. Effective (G) and potential (∗) glycosylation sites are marked.

As a control for internal translation initiation and potential alteration of LHBs structure due to the blaL sequence an additional mutant (pSVs25L) was used which contains a truncated nonfunctional signal sequence. Therefore, s25L and cleaved blaL have an identical primary amino acid sequence except for the first two amino acids, but their pre-S domains have an opposite orientation directly after synthesis at the ER membrane. In addition, pre-S2 and S start codons are intact in plasmid pSVs25L.

Protein expression by pcD/3-IL2 expressing murine IL-2 has been described previously in detail (8). G8 cells were transfected with pSVL, pSVblaL, pSVBX24H, and pSVBX24HSer69. To control for transfection efficiency, cells were cotransfected with a pSV-β-galactosidase-containing vector. As predicted, pSVBX24H- and pSVblaL-transfected cells secreted high amounts of HBsAg, as measured by an HBsAg-specific ELISA format which recognizes conformational as well as linear epitopes on SHBs (Fig. 2). The amount of secreted HBsAg from cells transfected with pSVblaL, however, was significantly lower compared to that from cells transfected with pSVBX24H. By contrast, cell culture supernatant from pSVL- and pSVBX24HSer69-transfected cells did not contain detectable amounts of HBsAg. In pSVs25L-transfected G8 cells HBsAg could be measured in culture supernatant despite a nonfunctional secretion signal. This may be due to cosecretion of wild-type MHBs and SHBs which were expressed from their autologous PS2 promoter within the pre-S1 sequence of pSVs25L. The corresponding translation initiation sites for MHBs and SHBs expression had been mutagenized in pSVL and pSVblaL. Mutated blaL and S proteins were conformationally intact, since HBsAg could easily be detected in cell culture supernatants as well as in cell lysates without denaturing agents (e.g., freeze-thaw lysis) (Fig. 2).

FIG. 2.

HBsAg levels in transfected G8 cells. HBsAg in culture supernatant and lysates from transfected cells were measured by an ELISA which recognizes conformational as well as linear epitopes on SHBs. The cell culture supernatants were collected 48 h after transfection. Cells were lysed using three cycles of freeze-thawing in PBS for viral protein studies. All results were corrected for transfection efficiency by using a β-galactosidase reporter assay and represent three different experiments with 1:20 dilutions in PBS of all individual samples. OD 450 nm, optical density at 450 nm.

To allow for the comparisons of the immune responses after immunizations of mice with LHBs and SHBs wild-type and mutant plasmids, we analyzed the level of intracellular expression of HBsAg in pSVL-, pSVblaL-, pSVBX24H-, and pSVBX24HSer69-transfected cells. Figure 2 demonstrates that the intracellular amount of HBsAg produced in transfected G8 cells was comparable. No HBsAg could be detected in mock-transfected cells.

These results were confirmed by immunoblot analysis using a pre-S1-specific MAb. In cells transfected with pSVL, LHBs was detected as double bands with molecular masses of 39 and 42 kDa. The blaL mutant formed four bands between 39 and 48 kDa, confirming the previously described presence of a triple N-glycosylation in association with translocation to the ER lumen by the N-terminal secretion signal (Fig. 3A). For comparison, the s25L protein was glycosylated like wild-type LHBs, indicating the expected cytosolic location of its pre-S domains.

FIG. 3.

Western blot analysis of HBsAgs. Intracellular expression of mutant and wild-type LHBs and SHBs in G8 cells after transfection with the corresponding DNA plasmids was determined by immunoblot analysis using the pre-S1-specific MAb MA18/7 (A) and a polyclonal goat anti-HBVs antiserum (B) as previously described (12). S corresponds to P24 and GP27 SHBs antigens. Wild-type L corresponds to P39 and GP42 LHBs. The blaL mutant formed four bands between 39 and 48 kDa, confirming the previously described presence of a triple N-glycosylation in association with translocation to the ER lumen by the N-terminal secretion signal.

Using a polyclonal SHBs-specific antibody, we were able to demonstrate that no internal translation initiation at the MHBs and SHBs start codons occurred in pSVL and pSVblaL. In contrast, P24 and GP27 SHBs but not GP33 and GP36 MHBs antigens could be detected in cells after transfection with pSVs25L, which carries functional pre-S2 and S start codons (Fig. 3B).

Although the cysteine mutant of pSVBX24HSer69 was not secreted, it was still glycosylated to the same extent as wild-type SHBs encoded by pSVBX24H (data not shown [16]). This indicates that translocation across the ER membrane was not noticeably affected. Since the mutant was also detectable in the cell lysates by ELISA, the structure of the major antigenic determinant was presumably not affected by the mutation.

Anti-HBs response and isotypes.

Twenty-four days after the immunization, mice injected with pSVBX24H showed strong anti-HBs responses up to 350 mIU/ml. IL-2 and pSVBX24H coimmunizations induced increased anti-HBs responses, which were about 200 mIU/ml higher than those of pSVBX24H-immunized mice (Fig. 4). Immunizations with the SHBs secretion mutant (pSVBX24HSer65) induced weak humoral immune responses, and IL-2 coimmunizations did not significantly increase anti-HBs responses. Similar results were obtained by using the plasmid encoding for wild-type LHBs (pSVL). By contrast, mice immunized with pSVblaL had significantly stronger anti-HBs responses compared to mice immunized with pSVL. It is important to note that IL-2 and pSVblaL coimmunizations increased anti-HBs titers, although anti-HBs titers were lower than in SHBs and IL-2 coimmunized mice. Mice immunized at day 1 with pSVL and injected with pcD/3-IL2 at day 14 showed anti-HBs responses which were about 150 mIU/ml higher than in mice immunized with pSVL and comparable to animals immunized with pSVblaL. Mock-DNA-immunized mice showed no anti-HBs responses. Anti-HBs responses in mice injected with plasmids encoding secreted forms of HBsAgs started to become detectable between days 8 and 10 after immunization (Fig. 5A). In contrast, anti-HBVs responses in mice immunized with pSVL and pSVBX24Ser65 were not detectable before day 18 after immunization.

FIG. 4.

Anti-HBs response. Anti-HBs titers are expressed in mIU/ml and were derived from single mice in each group. Each group comprised five mice. For group designations, see Materials and Methods.

FIG. 5.

Anti-HBs kinetics and antibody subtype. (A) Anti-HBs titers are expressed in mIU/ml and were derived from bleedings of single mice at the indicated time points. Each group comprised five mice. (B) Antibody IgG1 and IgG2a subtypes were determined by ELISA and derived from individual mouse sera diluted 1:50 on day 24 after genetic immunization. The numbers along the top of panel B are anti-HBs IgG1 to IgG2a ratios.

In all mice, a predominant IgG2a response was observed (Fig. 5B). In mice immunized with constructs encoding secreted forms of HBsAgs, however, IgG1-to-IgG2a ratios were significantly higher than those in mice immunized with plasmids encoding nonsecreted antigens (P = 0.001). Coimmunizations with IL-2 did not significantly alter this ratio except in animals immunized with mutant SHBs. In this group a strong decline in the IgG1-to-IgG2a ratios was observed in animals coimmunized with IL-2 and mutant SHBs (P = 0.005). In the other groups, IL-2-coimmunized animals showed insignificantly higher IgG2a and lower IgG1 responses compared to animals immunized without cytokines.

T-cell proliferative response.

Mice immunized with plasmids encoding secreted SHBs and LHBs showed strong T-cell proliferative responses using rHBsAg for in vitro stimulation. IL-2 coimmunizations significantly augmented T-cell proliferative responses (P = 0.004) (Fig. 6). A significant augmentation of T-cell proliferation was observed also in animals immunized sequentially with pSVL and pcD/3-IL2 compared to pSVL-immunized mice (P = 0.005). These responses were even stronger than those in animals injected with pSVblaL. Weak T-cell proliferation was observed in animals immunized with plasmids encoding nonsecreted HBV surface proteins, e.g., pSVL and pSVBX24HSer65. There was no increase of T-cell proliferative responses after coimmunizations with IL-2 and wild-type LHBs- or mutated SHBs-encoding plasmids. All the differences of T-cell proliferative activity described above were particularly significant because results were derived from individual rather than pooled splenocytes. All proliferative responses were HBsAg specific with no specific proliferation after stimulation with hCGβ.

FIG. 6.

T-cell proliferative activity. Spleen cells of individual mice were stimulated with recombinant HBsAg at the indicated concentrations. BrdU incorporation was measured after 3 days by ELISA. In addition, spleen cells were stimulated with recombinant hCGβ as a control for antigen specificity. For group designations, see Materials and Methods. In groups 2 (BX24H + IL-2), 8 (blaL + IL-2), and 9 (L [day 1] + IL-2 [day 14]), optical densities (ODs) greater than 2 are measured values. In our ELISA reading system, the linear range ends at an OD of 2.5 when stimulated with the highest antigen dose. We nevertheless present these data because they are in agreement with the data obtained with lower HBsAg doses that were in the linear range of the assay and therefore representative. OD 450 nm, optical density at 450 nm.

Cytotoxic T-cell response.

There were no significant differences in CTL killing activity against P815L target cells between animals exclusively immunized with the various wild-type and mutant LHBs and SHBs constructs (Fig. 7). At an E:T ratio of 20:1, about 25% lysis was observed, whereas lysis values at an E:T ratio of 2:1 were only marginal compared to unspecific lysis values against parental P815 target cells. Effector cells derived from mice following a single coimmunization with secreted forms of LHBs or SHBs and IL-2, however, displayed significantly higher CTL activity against P815L target cells at all E:T ratios tested (P = 0.01). Importantly, this was observed also in mice immunized with pSVL at day 1 and pcD/3-IL2 at day 14 (P = 0.002). Lysis values reached 35 to 40% at an E:T ratio of 20:1 and about 15% at an E:T ratio of 2:1. No significant increase in background lysis values against parental P815 cells was observed in these mice. By contrast, coimmunization of mice with nonsecreted forms of LHBs or SHBs and IL-2 did not increase CTL activity compared to animals immunized with pSVL or pSVBX24HSer65 only. Mock-DNA-immunized animals displayed no specific CTL activity against P815L cells. Additional experiments in some mice confirmed the antigen specificity of the CTL activity since in vivo-primed effector cells stimulated in vitro using parental P815 cells did not display CTL activity against P815L target cells (data not shown).

FIG. 7.

CTL responses of 10 groups of BALB/c mice immunized with HBV surface plasmids. Mice were injected once with a total of 100 μg of plasmid DNA encoding the different HBsAgs and IL-2. Single spleen cell suspensions were assayed after in vitro stimulation with syngeneic pre-S1 protein-expressing cells (P815L) for 6 days. The effector cells were then tested against P815L (■) and “wild-type” parental P815 cells (●) in a ⁵¹Cr-release assay at the E:T ratios indicated. Values are means of triplicate determinations and were derived from responder mice only. The numbers in the upper right hand corners represent responder mice/total mice studied.

DISCUSSION

Genetic immunization against HBV provides an excellent model with which to evaluate the strategy of prophylactic or therapeutic immunization and to define the immunogenicity of the different HBV antigens. Several groups have recently shown that HBV surface and nucleocapsid antigens are highly immunogenic at both the T-cell and B-cell levels in mice (6, 9, 12, 15, 22). It was interesting to note that antibody and TH1 responses to the LHBs were significantly weaker compared to responses to the MHBs (12) even though studies in HBV-infected humans and in mice immunized with recombinant protein have shown that additional B-cell and T-cell epitopes exist within the pre-S1 region of the HBV surface protein (17, 18). In addition, LHBs has been shown to be highly immunogenic at the T-cell level in HBV-infected humans (7).

LHBs cannot be secreted by itself in hepatoma and other cell lines, including myoblast cells, even though it is targeted cotranslationally to the rough ER (2, 4, 20). In addition, it inhibits the secretion of MHBs and SHBs by forming heteromultimers when produced in large amounts. LHBs itself can form 20-nm particles which are retained in a post-rough-ER compartment, possibly by interacting with calnexin and other regulatory ER proteins (27). In natural HBV infection, however, LHBs, together with MHBs and SHBs, is part of the Dane particles and filaments and therefore may be regarded as a secreted antigen because it is presented as an exogenous antigen (and, in addition, as an endogenous processed antigen in HBV-infected cells) to the immune system of the host. These biological properties of LHBs during HBV infection may explain the differences in immunogenicity compared to genetic immunizations. Using LHBs expression plasmids, the antigen is primarily expressed intracellularly and released only after some time from damaged muscle fibers.

Our studies support this hypothesis since after genetic immunization, secreted forms of LHBs had similar immunogenicity at the B- and T-cell levels compared to the secreted and highly immunogenic wild-type SHBs. The weaker anti-HBs responses in pSVblaL versus pSVBX24H immunized mice may be due to lower levels of secreted HBsAg as determined by in vitro experiments. Furthermore, the mutated nonsecreted form of SHBs induced significantly weaker B- and T-cell responses compared to the secreted wild-type SHBs. All mice displayed predominant IgG2a anti-HBs responses. This is in agreement with a recent study in which plasmids encoding nonsecreted wild-type LHBs and secreted wild-type MHBs induced a predominant TH1 proliferative response (12). In the present study, however, there was a significant tendency to elevated IgG1 levels in mice immunized with plasmids encoding secreted forms of HBsAgs. In addition, IL-2 coimmunizations seemed to partially revert this tendency towards elevated IgG2a levels. Further studies are required to examine this observation with respect to differentiation and induction of TH cell responses in type 1 or type 2 subtypes at the clonal level.

In addition, we studied the effects of coimmunizations of HBV DNA with cytokine DNA expression plasmids on cellular and humoral immune responses to HBsAgs. Recent studies demonstrated that coimmunizations of IL-2 and GM-CSF DNA expression plasmids enhanced humoral and cellular immune responses to rabies glycoprotein (26) and the hepatitis C virus core protein (8, 10). While recent studies demonstrated that helper T- and B-cell responses to SHBs and MHBs were enhanced by coimmunization with IL-2 and GM-CSF DNA expression plasmids and induced an immune response in nonresponder mice (5, 11), coimmunizations of LHBs encoding DNA with either IL-2 or GM-CSF plasmids had no augmenting effect on cellular and humoral immune responses to HBsAgs (11). Furthermore, we demonstrated that the effects of LHBs on the immune response augmenting properties of IL-2 and GM-CSF in vivo were not related to inhibition of their secretion from the cell. Finally, IL-2, TNF-α and IFN-γ cytokines had no suppressive effect on HBV surface protein expression in vitro (11).

In this study, we have shown a possible mechanism for the lack of augmentation of immune responses after a single IL-2 and LHBs coimmunization. We and others have recently demonstrated that the anti-HBs response to LHBs is detectable about 14 days later than the responses to MHBs or SHBs (6, 12). We demonstrate here that this is due to the intracellular retention of LHBs in transfected muscle cells, because the LHBs secretion mutant restored early anti-HBs responses at days 8 to 10 after one immunization compared to days 18 to 20 after wild-type LHBs injection. The intracellular retention of LHBs in transfected muscle cells may result in a delayed accessibility of the antigen for initiation of the immune response. Our results favor the hypothesis that cytokines, such as IL-2, when produced early by DNA expression constructs after one immunization event at the initial site of antigen presentation cannot augment antigen-specific immune responses because, at this point in time, LHBs is not produced at high enough levels or is not accessible for uptake by antigen-presenting cells. This results in the failure of IL-2 to augment anti-HBs, T-cell proliferative, and CTL responses after coimmunizations with nonsecreted forms of LHBs or SHBs. It is important to note that coimmunizations with secreted SHBs or LHBs and IL-2 or sequential immunizations with LHBs and IL-2 plasmids restored the augmenting effects of IL-2 on humoral and cellular immune responses against HBsAgs.

Future studies will address the effects of repeated coimmunizations on the differences between secreted and nonsecreted HBsAgs. It may be that the differences between secreted and nonsecreted HBsAgs diminish with repeated immunizations, since muscle-infiltrating lymphocytes and macrophages may lead to muscle damage and, consequently, to the release of intracellular cytoplasmic antigens. Previous studies with nonsecreted hepatitis C virus core protein support this hypothesis, because IL-2 and GM-CSF augmented immune responses after booster genetic coimmunizations (8).

In conclusion, we have demonstrated a possible mechanism by which intracellular retention of SHBs and LHBs may induce a lack of immune response augmentation after a single genetic coimmunization with an IL-2 DNA expression plasmid. These results may be of interest for the optimal design of DNA expression plasmids for therapeutic DNA vaccination of patients with chronic HBV.