Induction of Broad CD4⁺ and CD8⁺ T-Cell Responses and Cross- Neutralizing Antibodies against Hepatitis C Virus by Vaccination with Th1-Adjuvanted Polypeptides Followed by Defective Alphaviral Particles Expressing Envelope Glycoproteins gpE1 and gpE2 and Nonstructural Proteins 3, 4, and 5

Abstract

Broad, multispecific CD4⁺ and CD8⁺ T-cell responses to the hepatitis C virus (HCV), as well as virus-cross-neutralizing antibodies, are associated with recovery from acute infection and may also be associated in chronic HCV patients with a favorable response to antiviral treatment. In order to recapitulate all of these responses in an ideal vaccine regimen, we have explored the use of recombinant HCV polypeptides combined with various Th1-type adjuvants and replication-defective alphaviral particles encoding HCV proteins in various prime/boost modalities in BALB/c mice. Defective chimeric alphaviral particles derived from the Sindbis and Venezuelan equine encephalitis viruses encoding either the HCV envelope glycoprotein gpE1/gpE2 heterodimer (E1E2) or nonstructural proteins 3, 4, and 5 (NS345) elicited strong CD8⁺ T-cell responses but low CD4⁺ T helper responses to these HCV gene products. In contrast, recombinant E1E2 glycoproteins adjuvanted with MF59 containing a CpG oligonucleotide elicited strong CD4⁺ T helper responses but no CD8⁺ T-cell responses. A recombinant NS345 polyprotein also stimulated strong CD4⁺ T helper responses but no CD8⁺ T-cell responses when adjuvanted with Iscomatrix containing CpG. Optimal elicitation of broad CD4⁺ and CD8⁺ T-cell responses to E1E2 and NS345 was obtained by first priming with Th1-adjuvanted proteins and then boosting with chimeric, defective alphaviruses expressing these HCV genes. In addition, this prime/boost regimen resulted in the induction of anti-E1E2 antibodies capable of cross-neutralizing heterologous HCV isolates in vitro. This vaccine formulation and regimen may therefore be optimal in humans for protection against this highly heterogeneous global pathogen.

The hepatitis C virus (HCV) is responsible for essentially all parentally transmitted non-A, non-B hepatitis cases. An estimated 170 million humans, or 3% of the world's population, are infected with HCV, with an even higher prevalence in the developing parts of the world (27). There is no vaccine available, and the standard combination treatment with pegylated interferon (IFN) and ribavirin is curative in less than one-half of all HCV patients (16). There is therefore an urgent need for alternative therapies and effective prophylactic vaccines.

A key feature of most vaccines is the induction of neutralizing antibodies. In many cases, infusion of neutralizing antibodies is also used for passive postexposure prophylaxis. Preclinical studies with chimpanzees have indicated the ability of polyclonal antibodies derived from plasma of HCV-infected patients to prevent or delay HCV infection. The antibodies were shown to prevent or delay the onset of acute hepatitis C when given before or soon after inoculation of chimpanzees with the virus (13, 14, 22, 63). In addition, vaccination of chimpanzees with recombinant HCV envelope glycoproteins gpE1 and gpE2 induced strong antibody responses that prevented infection from a homologous viral (HCV-1) challenge (8). The HCV 1a strain predominates in the United States. Subsequent studies in which animals were vaccinated with adjuvanted, clade 1a-derived gpE1/gpE2 and then challenged with a heterologous 1a viral strain demonstrated a substantial and statistically significant reduction in the carrier rate of the vaccinees versus a control, unimmunized group of chimpanzees (9, 20). Recently, it was also demonstrated that a sustained anti-E2 antibody response correlates with reduced peak viremia after HCV infection in the chimpanzee (62). Recent studies have also correlated the early induction of HCV cross-neutralizing antibody with recovery from acute infection in humans (28, 39).

Other studies have emphasized the role of the cellular immune response in protection against HCV by showing that broad, multispecific CD4⁺ and CD8⁺ T-cell responses to the virus are associated with naturally resolving infection (10, 11, 12, 15, 17, 29). Furthermore, a series of rechallenge studies with chimpanzees that recovered spontaneously, in which the CD4⁺ or CD8⁺ T-cell compartments were first depleted, have demonstrated the crucial role of both of these cell types in protective immunity against HCV infection (17, 52). This also has been successfully adopted in a vaccine approach using a prime/boost immunization regimen utilizing adenovirus and plasmid DNA expressing HCV nonstructural genes 3, 4, and 5. Most of the naïve chimpanzees vaccinated in this way were protected against the onset of chronic hepatitis and viremia following an experimental challenge with a highly heterologous HCV strain (5). Thus, HCV immunogens able to elicit strong and broad cell-mediated immunity, as well as cross-neutralizing antibodies, may represent the optimal approach to HCV vaccination (20).

Replication-defective alphaviral vectors have been shown to induce robust cellular, humoral, and mucosal immune responses specific to the replicon-expressed antigen in several animal models (6, 18, 21, 38, 42). A number of features make alphavirus replicon vectors attractive for gene-based vaccines, including high-level expression of the heterologous gene, vector amplification through double-stranded RNA intermediates (which stimulates aspects of innate immunity such as activation of the IFN cascade), induction of apoptosis in some cell types (which may enhance immunogenicity via antigen cross-priming), and the overall lack of preexisting immunity in the human population (42). In addition, alphavirus replicon particles can be used for the delivery of antigen to antigen-presenting cells such as dendritic cells (42). It is also possible to repeat immunize with defective alphavirus particles (42).

Alphaviruses such as Semliki Forest virus, Venezuelan equine encephalitis virus (VEE), and Sindbis virus (SIN) have all been developed as potential vaccine and gene therapy vectors. Each alphavirus has unique properties in terms of function, potency, and safety when applied to recombinant vector systems. VEE replicon particles have been shown to induce potent and protective immune responses in primates. However, production of VEE replicon particles must be conducted at biosafety level III (38). In contrast, since SIN is not associated with serious human disease, the use of SIN-derived replicon particles largely obviates safety concerns. In addition, particular engineered SIN variants target lymphoid cells, and stable replicon-packaging cell lines have been developed (41), which may simplify large-scale production for human testing of vaccine candidates. A novel VEE/SIN defective-particle chimera that comprises SIN structural outer membrane proteins encapsidating VEE replicon RNA (38) was used in this study to express the HCV E1E2 envelope or nonstructural NS345 protein.

MF59, CpG, and Iscomatrix have been used as adjuvants for various recombinant proteins. The oil-in-water adjuvant MF59 (34) has been shown to be a potent stimulator of cellular and humoral responses to subunit antigens in both animal models and clinical studies (34) and is licensed for human use in the European Union as part of an influenza vaccine. Recombinant HCV gpE1/gpE2 glycoproteins combined with MF59-like adjuvant were shown to be protective against both homologous and heterologous viral challenges (8, 9, 20). CpG 7909 is an immunomodulating synthetic oligonucleotide designed to specifically target Toll-like receptor 9 in dendritic antigen-presenting cells and augments the adaptive immune response to vaccine antigens (23). It also activates memory B cells (24) and thus is a promising adjuvant for vaccines against cancer and infectious diseases (2). Iscomatrix (36, 37, 40) comprises phospholipid-cholesterol particles containing purified saponins. When it was combined with a recombinant HCV nucleocapsid protein (C) and used to immunize rhesus macaques, very strong C-specific Th1-type CD4⁺ T helper cells were elicited, as well as long-lived CD8⁺ cytotoxic T cells (40).

In the present study, we have explored the use of adjuvanted recombinant HCV proteins and defective alphaviral vectors to elicit cross-neutralizing antibodies and broad cellular immune responses to HCV in BALB/c mice. We show that a prime/boost regimen can elicit all of the humoral and cellular immune responses known to be associated with protection against this common and global pathogen.

MATERIALS AND METHODS

Replicon vectors and defective helper constructs.

E1E2 (amino acids [aa] 192 to 746) was generated by PCR amplification of pCMVtpaE1E2p7, which contains the upstream tissue plasminogen activator (TPA) signal sequence at the 5′ end and E1E2p7 (aa 192 to 809) from HCV strain 1a (32, 44, 55), and the cDNA was inserted into the VCR-Chim2.1 replicon vectors (38), resulting in the construct VCRChim-E1E2 (aa 192 to 746). NS345 generated by PCR amplification of HCV strain 1a cDNA (7) and the cDNA (aa 1027 to 3012) were inserted into the VCR-Chim2.1 replicon vectors, resulting in construct VCRChim-NS345. Sequences encoding either capsid or envelope glycoproteins from SIN were inserted into the VEE-based defective helper backbone (VCR-DH) and named VCR-DH-Scap and VCR-DH-Sgly (38).

Production of alphavirus replicon particles.

Replicon particles were generated by coelectroporation of in vitro-transcribed RNAs corresponding to a replicon vector (VCRChim-E1E2or VCRChim-NS345) and two defective helpers (VCR-DH-Scap and VCR-DH-Sgly) (38). Replicon particles (VEE/SIN-E1E2 and VEE/SIN-NS345) expressing HCV E1E2 and NS345 were harvested as culture supernatants at 24 h postelectroporation, clarified by filtration, and purified by cation-exchange chromatography. Replicon particle titers, reported in infectious units (IU) per milliliter, were determined by intracellular staining (ICS) of expressed HCV proteins (E1 and E2 for VEE/SIN-E1E2; NS3, NS4, NS5A, and NS5B for VEE/SIN-NS345) following overnight infection of BHK-21 cells (38). Infected cells were permeabilized and fixed by using a Cytofix/Cytoperm kit (Pharmingen) and then stained with fluorescein isothiocyanate-conjugated antibodies to HCV antigen. By flow cytometry analysis, the percentage of HCV protein-positive cells was determined and used to calculate titers. The absence of contaminating replication-competent virus was established by five consecutive infections of naive BHK-21 cells and determination of titers. Finally, endotoxin levels were measured for all replicon particle samples and shown to be <0.03 U/ml.

HCV recombinant proteins.

HCV strain 1a E1E2 contains E1E2p7 (aa 192 to 809) with the TPA at the 5′ end. E1E2 was expressed from recombinant CHO cells and extracted with Triton X-100 detergent. The expression and purification methods used have been described previously (32). HCV polyprotein NS345core (HCV strain 1a NS345core, aa 1018 to 3012 and 1 to 191) (59) contains HCV strain 1a NS3 to NS5 and has the core protein at the C terminus (59). The methods used for plasmid construction, protein expression, and purification from yeast cells have been described in detail previously (59).

Adjuvants.

MF59 was used at a 1:1 dilution with the immunogens immediately before immunization. CpG oligodeoxynucleotides (CpG) 7909 (10 μg per dose; Coley Pharmaceutical Group, Wellesley, MA) (35) were mixed with the protein plus MF59 or Iscomatrix (5 μg per mouse) immediately before immunization. Iscomatrix was diluted with isotonic buffer (sodium chloride at 145 mM, disodium hydrogen phosphate anhydrate at 2.5 mM and sodium dihydrogen phosphate dihydrate at 7.7 mM; pH 6.2) to 100 μg/ml and used at a 1:1 dilution with isotonic-buffer-diluted polyprotein immediately before immunization.

Immunizations.

Ten female BALB/c mice per group were injected intramuscularly in the tibialis anterior muscle with a total volume of 100 μl (i.e., 50 μl per thigh) of the indicated vaccine formulations at weeks 0, 3, and 6, and serum samples were collected at week 8. For the prime-boost studies, the mice were primed at weeks 0 and 3 and boosted at week 6. We used 4 × 10⁶ IU of VEE/SIN-E1E2 and 2 μg of HCV strain 1a E1E2 protein (E1E2P7 derived from E1E2-809) (55) mixed with MF59 in the presence or absence of CpG for injection. The mice were sacrificed at week 8, and the spleens were harvested for further assays. For nonstructural protein, 5 × 10⁶ replication particles of VEE/SIN-NS345 and 50 μg of polyprotein were mixed with 5 μg of Iscomatrix (from CSL) (36, 37, 40) for injection. The mice were injected and sacrificed at week 8 to detect T-cell responses in the spleen and antibody responses in serum.

ICS.

Spleen cells (1 × 10⁶) were stimulated with the indicated peptides or proteins at 10 μg/ml (Table 1) for 6 h at 37°C in the presence of anti-CD28 antibody (1 μg/ml) (BD Biosciences, San Jose, CA) and brefeldin A (BD Biosciences, San Jose, CA) and then stained with antibodies against CD4 (anti-CD4 PerCP-Cy5.5 conjugate, clone RM4-5; BD Pharmingen) and CD8 (anti-CD8α fluorescein isothiocyanate conjugate, clone 53-6.7; BD Pharmingen), permeabilized with Cytofix/Cytoperm (Pharmingen), and IFN-γ (clone XMG1.2, allophycocyanin conjugate; BD Pharmingen). Stained cells were analyzed with a FACScalibur flow cytometer (BD Bioscience). The mean frequencies of cytokine-positive cells were calculated for each pair of duplicates. The antigen-specific frequency was determined by comparing the unstimulated mean frequency (no peptide) with the stimulated mean frequency (with HCV peptides, Table 1), and P < 0.05 is considered statistically significant by t test (Table 1).

TABLE 1.

Peptides and proteins used in this study

Name	Peptide or protein information	Reference(s)
E1 pool	20-mer overlapping peptides covering E1 region	64
E2 pool	20-mer overlapping peptides covering E2 region	64
CD4 E2 pep	QTHTTGGQAGHQAHSLTGLFSPGAKQN	43, 64
CD8 E2 pep	DATYSRCGSGPWITPRCLVD	64
NS3 pool	20-mer overlapping peptides covering NS3	53, 61
NS4 pool	20-mer overlapping peptides covering NS4	53, 61
NS5A pool	20-mer overlapping peptides covering NS5A	53, 61
NS5B pool	20-mer overlapping peptides covering NS5B	53, 61
NS3-1 pep	LVALGINAVAYYRGL	54
NS3-2 pep	TTVRLRAYMNTPGLP	54
NS3-3 pep	SSPPVVPQSF	3, 4
NS5B pep	MSYSWTGALVTPCAAE	51, 58
SOD-C100 (NS34)	Recombinant NS34 protein purified from yeast (HCV strain 1a aa 1569-1931)	25, 30
SOD-NS5	Recombinant NS5A/B protein purified from yeast (HCV strain 1a aa 2054-2995)	48

CD81 inhibition assay.

E1E2 protein at 5 μg/ml was mixed with 5E5/H7*Eu3⁺ (anti-E2 specific monoclonal antibody labeled with europium) at 0.33 μg/ml and then incubated with serial diluted serum samples. The mixture was transferred to CD81 recombinant receptor-coated 96-well plates. After washing, the plates were treated with Enhancement Solution (Wallac) and then placed in a Wallac 1420 Multilabel Counter for reading by the europium protocol. Fifty percent inhibition multiplied by the dilution factor is used to estimate the CD81 titer. The preimmunized blood of a given subject at the assay dilution of 1:10 is used as the negative control (1, 19, 47).

Luminex multiplex cytokine assay.

Spleen cells (1 × 10⁶) were stimulated with HCV peptides or proteins at 10 μg/ml (Table 1) for 24 h at 37°C in the presence of anti-CD28 antibody (1 μg/ml) (BD Biosciences, San Jose, CA), and the supernatant was harvested and stored at −20°C for cytokine assay. The Beadlyte Mouse Multi-Cytokine Flex Kit Beadlyte (Upstate, Charlottesville, VA) was used to detect interleukin-5 (IL-5), IL-10, tumor necrosis factor alpha (TNF-α), and IFN-γ levels in cell culture supernatants according to the manufacturer's protocol. Cytokine concentrations from the mean fluorescence values obtained were calculated from standard curves of each cytokine tested with Miraibio Master Plex QT software (Miraibio, Alameda, CA).

HCVcc neutralization assay.

pRlucF2Aubi-JFH1 is a full-length JFH1 genome (60) carrying Renilla luciferase (Rluc) as a reporter in a monocistronic construct (T. W. Kwon et al., unpublished data). The RlucF2Aubi-JFH1 virus (RJ1) was generated by electroporation of IFN-cured Huh-7 cells with in vitro-transcribed Rluc-JFH1 RNA. The conditioned medium that was harvested at 10 to 14 days posttransfection was used as the viral stock. For the neutralization assay, Huh-7 cells were seeded at 5 × 10⁴ per well in a 24-well plate the day before infection. Mouse serum was mixed with viral medium (900 tissue culture infective doses/ml) at 1:100 and incubated at 37°C for 1 h. A 200-μl volume of a virus-serum mixture per well was applied to Huh-7 cells. After a 4-h infection at 37°C, the viral medium was removed and replaced with fresh complete cell medium (Dulbecco modified Eagle medium-10% fetal bovine serum). As an indication of infection, Renilla luciferase activity was measured 3 days postinfection by using the Renilla luciferase assay system (Promega).

Statistical analysis.

For antibody responses, each individual animal immune response was counted as an individual value for statistical analysis. The significance of the responses was calculated by Student's t test (two tailed, type 2) with Prism (version 3.0 for Windows; GraphPad, San Diego, CA). A P value of ≤0.05 was considered statistically significant.

RESULTS

Immunization with VEE/SIN particles is optimal for eliciting E1E2-specific CD8⁺ T cells.

Defective alphavirus particles expressing HCV E1E2 (aa 192 to 746) and NS345 (aa 1027 to 3012) were generated and designated VEE/SIN-E1E2 and VEE/SIN-NS345. The expression of glycoproteins E1 and E2, as well as NS3, NS4, NS5A, and NS5B, from the replicons was confirmed by Western blotting (see Fig. S1 in the supplemental material). The particular antigen-adjuvant combinations used in this study were based on individual formulations currently being tested in clinical trials. Anticipated prophylactic formulations comprising E1E2 with MF59 have been used to vaccinate humans in the presence or absence of CpG with a view to eliciting broadly cross-neutralizing antibodies, whereas potential therapeutic vaccine formulations have focused on the use of the Iscomatrix adjuvant to boost optimal cellular immune responses in HCV-infected patients (unpublished data).

To compare the immune responses elicited by defective VEE/SIN vectors encoding E1E2 with adjuvanted E1E2 protein, mice were immunized three times either with 4 × 10⁶ IU VEE/SIN-E1E2 particles or with recombinant E1E2 (aa 192 to 809) protein adjuvanted with MF59 (with or without CpG) (35). Spleen cells were assessed 14 days after the final immunization with each vaccine. The cells were stimulated either with HCV strain 1a-specific peptide pools encompassing E1 or E2 or with single peptides known to represent CD4⁺ or CD8⁺ T-cell epitopes for E2 in BALB/c mice (Table 1). Negative controls consisted of cells treated with medium alone or cells from phosphate-buffered saline (PBS)-injected mice stimulated in vitro with HCV peptides. In contrast to MF59-adjuvanted recombinant E1E2 immunizations, VEE/SIN-E1E2 immunizations stimulated readily detectable HCV-specific CD8⁺ T-cell responses to the E1 peptide pool (64) (Fig. 1A and B; see Fig. S3 in the supplemental material). MF59-adjuvanted E1E2, in either the presence or the absence of CpG, did not stimulate significant HCV-specific CD8⁺ T-cell responses (Fig. 1A and B; see Fig. S3 in the supplemental material). Two immunizations with E1E2/MF59/CpG followed by a boost with VEE/SIN-E1E2 produced strong CD8⁺ responses to an E2 epitope (Fig. 1C; see Fig. S2B and S3 in the supplemental material). This could also be accomplished by administering two immunizations with E1E2/MF59 followed by a third immunization with VEE/SIN-E1E2 (Fig. 1B and C). This response was greater than that obtained with VEE/SIN-E1E2 alone, although surprisingly, CD8⁺ responses to the E1 pool were reduced by the vaccine combination regimen.

FIG. 1.

Priming with E1E2/MF59/CpG followed by boosting with VEE/SIN-E1E2 stimulates strong, good CD4⁺ and CD8⁺ T-cell responses. BALB/c mice (n = 10) received three intramuscular injections (1, 2, 3) at 3-week intervals with the immunogens indicated. The spleen cells were harvested at 2 weeks after the last immunization and stimulated with 10 μg/ml HCV-specific peptides for ICS and fluorescence-activated cell sorter analysis as described in Materials and Methods. Twenty-mer overlapping peptide pools for the HCV strain 1a E1 region (E1 pool) and E2 region (E2 pool) and single E2 peptides specific for CD4⁺ T cells (CD4 E2 pep) and CD8⁺ T cells (CD8 E2 pep) were used for stimulation. The data are presented as the mean total percentage of CD8⁺ IFN-γ⁺ (A, B, and C) and CD4⁺ IFN-γ⁺ (D) cells in two pools (five mice per pool). Data from one representative experiment of two performed are shown. *, P < 0.05 compared with the medium control and the PBS-immunized group.

A prime/boost regimen with E1E2/MF59/CpG followed by VEE/SIN-E1E2 elicits strong CD4⁺ responses.

To analyze HCV-specific CD4⁺ T-cell responses after immunization, spleen cells were stained for CD4 and IFN-γ after in vitro stimulation with HCV-specific peptides (Table 1). Spleen cells obtained from PBS-immunized mice and from vaccinated mice without in vitro peptide stimulation (medium) were both used as negative controls. Immunization with E1E2/MF59/CpG clearly elicited the strongest CD4⁺ responses to the E1 pool, with weaker responses to the E2 pool (P < 0.05 compared with E1E2/MF59 or VEE/SIN-E1E2 immunization; Fig. 1D; see Fig. S2A in the supplemental material). These responses were also obtained in mice that were primed by two immunizations with E1E2/MF59/CpG followed by boosting with VEE/SIN-E1E2 (Fig. 1D). Interestingly, lower responses were obtained if VEE/SIN-E1E2 was used to prime CD4⁺ responses prior to boosting with E1E2/MF59/CpG (Fig. 1D). Therefore, an optimal prime/boost regimen for eliciting E1E2-specific CD4⁺ T cells (priming with recombinant E1E2 adjuvanted with MF59-CpG followed by boosting with VEE/SIN-E1E2) was also nearly optimal at eliciting E2-specific CD8⁺ T cells (Fig. 1C; see Fig. S2B and S3 in the supplemental material). It should be noted that VEE/SIN-E1E2 induced a high, peptide-independent background of IFN-γ production in CD4⁺ T cells in general, but no HCV-specific CD4⁺ T-cell responses were detected (Fig. 1D).

CD4⁺ Th1 cells assist in priming antigen-specific CTLs by activating DCs and secreting IL-2. They are also required for the maintenance of CD8⁺ T-cell number and function (33). Clinical data indicate that patients who spontaneously clear viremia have stronger multiclonal Th1-based T-cell responses to HCV antigens than those who do not (45, 49). Therefore, we analyzed the Th cell status of the immunized mice by measuring the cytokines released from the spleen cells after in vitro stimulation with HCV-specific peptides. E1E2/MF59 stimulated high IL-5 production but not IFN-γ or TNF-α, indicating the induction of a Th2 response (Fig. 2). On the other hand, E1E2/MF59/CpG induced high levels of IFN-γ and TNF-α, but not IL-5 or IL-10, indicating a Th1 response. VEE/SIN alphavirus particles (VEE/SIN-E1E2, VEE/SIN-E1E2p7, and VEE/SIN-NS345) induced nonspecific cytokine production from CD4⁺ T cells in both Th1 (IFN-γ and TNF-α) and Th2 (IL-5 and IL-10) cytokines, as no difference in cytokine production was observed after in vitro stimulation with either control medium or HCV peptides (Fig. 2). However, the prime/boost regimen using E1E2/MF59/CpG to prime followed by a boost with VEE/SIN-E1E2 induced the strongest and broadest HCV-specific Th1 cytokine responses (Fig. 2). On the other hand, the prime/boost regimen with E1E2/MF59 and VEE/SIN-E1E2 induced both HCV-specific Th1 and Th2 cytokines (Fig. 2).

FIG. 2.

Prime/boost regimen with E1E2/MF59/CpG and VEE/SIN-E1E2 stimulated strong HCV-specific Th1 cytokine responses. Spleen cells from the immunized mice (immunogens as indicated) were stimulated with HCV-specific peptides for 24 h, and the supernatants were collected for detection of the cytokines IFN-γ, TNF-α, IL-5, and IL-10 with the Beadlyte Luminex mouse multicytokine detection system (Upstate, Charlottesville, VA). The data are presented as the means of pooled spleens from the same group (two pools per vaccine regimen, five spleens per pool). Data from one representative experiment of two performed are shown.

There is strong stimulation of IFN-γ and TNF-α production by the E2 peptide pool using E1E2/MF59/CpG or E1E2/MF59 (priming) followed by VEE/SIN-E1E2 (boosting) in the bead-array assay (Fig. 2), but this regimen did not strongly stimulate IFN-γ from CD8⁺ (Fig. 1C) or CD4⁺ T cells (Fig. 1D) by the E2 peptide pool. The cytokines detected by the bead array assay might therefore be derived from other spleen cell populations in addition to CD4⁺ and CD8⁺ T cells.

Elicitation of optimal cellular immune responses against HCV nonstructural proteins.

BALB/c mice were immunized with an HCV NS345core polyprotein or with VEE/SIN particles expressing NS345. T-cell responses were detected by ICS of IFN-γ on CD4⁺ or CD8⁺ T cells after in vitro stimulation with overlapping HCV strain 1a peptide pools for either NS3, NS4, NS5A, or NS5B, as well as recombinant HCV NS34 protein [SOD-C100 (NS34); aa 1569 to 1931 (25, 30)] and recombinant HCV NS5 protein (SOD-NS5A/B; aa 2054 to 2995 [48]) (Table 1). Iscomatrix was used as an adjuvant to the HCV NS345core polyprotein, with or without CpG, since previous studies have shown Iscomatrix to be a very effective adjuvant for priming Th1-type CD4⁺ and CD8⁺ CTL responses to the HCV nucleocapsid protein in nonhuman primates (40). HCV NS345 polyprotein adjuvanted with Iscomatrix plus CpG elicited strong CD4⁺ T-cell responses against the NS4 and NS5b peptide pools, as well as against HCV NS34protein (SOD-C100) and recombinant HCV NS5 protein (SOD-NS5) (48) (Fig. 3A). In the absence of CpG, Iscomatrix-adjuvanted NS345 polyprotein still primed CD4⁺ T cells but at an approximately fivefold lower level (no. 2 in Fig. 3A). VEE/SIN particles expressing HCV NS345 could also prime HCV-specific CD4⁺ T cells (no. 4 in Fig. 3A) but at a much lower level than the Iscomatrix-CpG-adjuvanted NS345 polyprotein. Again, however, a prime/boost regimen comprising priming with the latter followed by boosting with VEE/SIN-NS345 recapitulated the optimal CD4⁺ T-cell responses (no. 7 in Fig. 3A).

FIG. 3.

A prime/boost regimen using NS345Poly/IMX/CpG and VEE/SIN-NS345 is optimal for stimulating HCV-specific CD4⁺ (A) and CD8⁺ (B) T-cell responses. BALB/c mice in each test group (n = 10) received three injections (1, 2, 3) at 3-week intervals with immunogens as indicated. Intracellular cytokine staining was performed 2 weeks after the last injection, and the spleen cells were stimulated with 10 μg/ml HCV-specific peptides for ICS and fluorescence-activated cell sorter analysis as described in Materials and Methods. Twenty-mer overlapping peptide pools for the HCV strain 1a NS3 (NS3 pool), NS4 (NS4 pool), NS5A (NS5A pool), and NS5B (NS5B pool) regions; single peptides from the NS3 region (NS3-1 pep, NS3-2 pep, and NS3-3 pep) and NS5 (NS5B pep); and recombinant proteins SOD-C100 (NS34) and SOD-NS5 were used for stimulation. The data are presented as the means of pooled spleens from the same group (two pools per vaccine regimen, five spleens per pool). Data from one representative experiment of two performed are shown. *, P < 0.05 compared with the medium control and the PBS-immunized group. IMX, Iscomatrix.

This same prime/boost regimen was also clearly optimal for priming HCV-specific CD8⁺ T cells (no. 7 in Fig. 3B). Adjuvanted HCV NS345core polyprotein did not elicit CD8⁺ T cells, and while VEE/SIN-NS345 was effective, maximal levels were achieved via the prime/boost regimen.

The cytokine profile of spleen cells from immunized mice was also analyzed after stimulation with HCV-specific peptide pools. Poly/Iscomatrix/CpG induced strong Th1 cytokines (IFN-γ and TNF-α) and low or undetectable Th2 cytokines (IL-10 and IL-5) (Fig. 4). Interestingly, a prime/boost regimen using Poly/Iscomatrix/CpG for the first two immunizations followed by a third immunization with VEE/SIN-NS345 stimulated the highest IFN-γ and TNF-α production, indicating a strong Th1 response (Fig. 4).

FIG. 4.

Prime/boost regimen with NS345Poly/IMX/CpG and VEE/SIN-NS345 stimulates strong HCV-specific Th1 cytokine responses. Spleen cells from immunized mice (immunogens as indicated) were stimulated with HCV-specific peptides for 24 h, and the supernatants were collected for the detection of the cytokines IFN-γ, TNF-α, IL-5, and IL-10 with the Beadlyte Luminex mouse multicytokine detection system (Upstate, Charlottesville, VA). The data are presented as the means of pooled spleens from the same group (two pools per vaccine regimen, five spleens per pool). Data from one representative experiment of two performed are shown. IMX, Iscomatrix.

Also of note are the high but nonspecific background levels of IL-10 and IL-5 elicited by three immunizations with VEE/SIN-NS345, which are similar to those found following immunization with VEE/SIN-E1E2 (Fig. 2). The nonspecific induction of Th1 and Th2 cytokines may be due to the ability of SIN and VEE to infect dendritic cells and induce proinflammatory cytokine production (31, 57).

A prime/boost regimen using MF59-CpG-adjuvanted E1E2 followed by VEE/SIN-E1E2 elicits virus-cross-neutralizing antibodies.

Previous chimpanzee vaccination studies using adjuvanted recombinant E1E2 vaccines generated sterilizing immunity against a homologous viral challenge. This sterilization correlated directly with both total serum anti-E1E2 antibody titers (enzyme immunoassay) and titers of antibodies which block the binding of gpE2 to the essential HCV receptor component CD81 (8, 20, 46). To explore if the immunogens we tested could stimulate HCV-neutralizing antibodies, we first analyzed serum samples obtained from mice 2 weeks after the final immunization for antibodies that could block the binding of gpE2 to CD81. Immunizations with E1E2/MF59 were best at eliciting these antibodies, whereas immunization with defective VEE/SIN particles stimulated poor antibody responses (Fig. 5A and B). However, a prime/boost regimen using E1E2/MF59/CpG followed by VEE/SIN-E1E2 was similar to E1E2/MF59/CpG in stimulating antibodies that block the binding of HCVgpE2 to CD81, although their titers are lower than those obtained by E1E2/MF59 immunization (Fig. 5A). To determine if these immunizations could stimulate antibodies capable of neutralizing the infectivity of HCV derived from infected cell cultures (HCVcc), we generated HCVcc derived from HCV JFH-1 cDNA fused to a luciferase reporter (JFH1-Rluc) as described by Wakita et al. (60). This HCV strain is part of the 2a subtype, whereas the vaccine is derived from a 1a subtype. Mice immunized with PBS were used as controls. The data are expressed as the percent inhibition of luciferase activity relative to that obtained from control mice. Mice immunized with E1E2/MF59 elicited substantial cross-neutralizing antibody titers against this highly heterologous HCV strain, indicating the induction of B-cell responses to a conserved neutralizing epitope(s) within E1E2 (Fig. 5B).

FIG. 5.

Induction of cross-neutralizing antibodies in vaccinated mice. (A) A prime/boost regimen with E1E2/MF59/CpG and VEE/SIN-E1E2 stimulates antibodies that block the binding of HCV gpE2 to CD81. BALB/c mice were immunized with immunogens as indicated. Mouse serum was collected 2 weeks after the last injection. The antibody titers are expressed as reciprocal values of serum dilutions which inhibit 50% of E2 binding to CD81. Serum samples from each individual mouse were measured. (B) Induction of antibodies that cross-neutralize the infectivity of JFH-1 2a HCVcc for Huh-7 cell lines. Serum samples collected as described for panel A were mixed at a 1:100 dilution with JFH-1 2a HCVcc and preincubated at 37°C for 1 h. The mixture was then applied to Huh-7 cells in triplicate, and the cell lysate was measured for luciferase activity at day 3 postinfection. The luciferase activity from mouse serum with PBS immunization was used as a control. The data are shown as percent inhibition of the luciferase activity from serum with different immunizations relative to the control. (C) A prime/boost regimen with E1E2/MF59/CpG and VEE/SIN-E1E2 stimulates cross-neutralizing antibodies against JFH-1 2a HCVcc infection of Huh-7 cells. Serum samples from immunized mice (antigens as indicated) were diluted 1:100 and mixed with JFH-1 2a HCVcc. The experiment was performed as described for panel B. The luciferase activity detectable with mouse serum following PBS immunization was used as a control. The data are shown as the percent inhibition of the luciferase activity from serum with different immunizations relative to the control. *, P < 0.05 relative to PBS immunization.

In addition, the immunization regimen optimal for priming HCV-specific CD4⁺ and CD8⁺ T cells (priming with E1E2/MF59/CpG followed by boosting with VEE/SIN-E1E2) elicited somewhat lower (about 20% less relative to the E1E2/MF59 group) but still significant cross-neutralizing antibody titers against this diverse 2a genotype (Fig. 5B and C).

Eliciting humoral and cellular immune responses to HCV structural and nonstructural proteins in the same immunization regimen.

A combination immunization regimen was then tested in order to determine if all of the above immune responses could be elicited successfully. BALB/c mice were immunized twice with adjuvanted recombinant E1E2 and NS345 polyprotein prior to boosting with defective VEE/SIN alphaviral particles expressing these genes. Figure 6 shows that, compared with an individual regimen for either structural or nonstructural proteins, a combination regimen elicited comparable CD4⁺ (Fig. 6A and C) and CD8⁺ (Fig. 6B and D) T-cell responses. While this combination vaccine elicited somewhat lower CD4⁺ T-cell responses to E1 (Fig. 6A), it increased CD8⁺ T-cell responses to NS34 (Fig. 6D). In addition, cross-neutralizing antibody titers were elicited that were equivalent to those obtained with the single-immunization regimen for E1E2 (Fig. 6E).

FIG. 6.

A combination prime/boost immunization regimen elicits broad cellular responses to HCV structural and nonstructural proteins, as well as cross-neutralizing antibodies. Groups of 10 BALB/c mice were immunized as indicated. For the combination immunization regimen, priming with E1E2/MF59/CpG and NS345Poly/IMX vaccines and boosting with VEE/SIN-E1E2 and VEE/SIN-NS345 were done individually to separate muscles for two different antigens at weeks 0, 3, and 6. At week 8, two pools of spleen cells were prepared (five spleens per pool) and stimulated with HCV-specific peptides prior to staining for intracellular IFN-γ and fluorescence-activated cell sorter analysis. The mean values obtained from the two pools are shown. Panels: A, CD4⁺ cells stimulated by E1 and E2 peptides; B, CD8⁺ cells stimulated by E1 and E2 peptides; C, CD4⁺ cells stimulated by NS3, -4, and -5 peptides; D, CD8⁺ cells stimulated by NS3, -4, and -5 peptides; E, cross-neutralizing antibody activity elicited against JFH-1 2a HCVcc. Serum obtained at week 8 was diluted 1:100 and preincubated with virus for 1 h at 37°C prior to infection of Huh-7 cells. Three days later, luciferase activity in cell lysates was determined. Data are expressed as percent inhibition (based on means from triplicate assays) relative to control mice immunized with PBS. *, P < 0.05 compared with the medium control and the PBS-immunized group. IMX, Iscomatrix.

DISCUSSION

Our first-generation HCV vaccine was based on the use of recombinant, native strain 1a E1E2 glycoproteins derived from mammalian cell lines that were adjuvanted with oil-water emulsions. Such vaccines have been shown to be highly effective at preventing the development of chronic infection in the large majority (>80%) of vaccinated chimpanzees that were subsequently challenged with either homologous 1a virus (8) or heterologous 1a virus (9, 20). Recently, vaccination of guinea pigs, chimpanzees, and humans with this formulation has been shown to elicit antibodies that cross-neutralize HCV/HIV pseudoparticles displaying HCV E1E2 envelope glycoproteins derived from multiple global HCV clades. Substantial T helper responses to E1E2 were also elicited in vaccinated individuals (56; M. Houghton et al., unpublished data). While this vaccine represents a highly promising candidate, further research is aimed at developing a second-generation vaccine of even higher potency.

The present work aimed to identify a vaccine regimen capable of eliciting all of the immune responses known to be associated with protection from HCV infection, including virus-cross-neutralizing antibodies (8, 28, 39, 46) and broad cellular immune responses against HCV. The latter includes broad Th1-type CD4⁺ T helper responses (12, 17), as well as broad CD8⁺ T-cell responses (10, 52). Since the Hepacivirus genus is very heterogeneous, a global HCV vaccine needs to elicit protective, broadly cross-neutralizing antibodies, as well as widely cross-reactive cellular immune responses.

The present study explored the use of defective alphavirus particles expressing HCV genes and various adjuvanted recombinant proteins in order to achieve these immunogenicity objectives. The data presented allow the following conclusions. Firstly, it is clear that optimal Th1-type CD4⁺ T helper responses were obtained through the use of adjuvanted recombinant E1E2. The inclusion of CpG with MF59 substantially increases Th1-type CD4⁺ T-cell responses to E1E2 (Fig. 1D). While defective chimeric alphaviral particles derived from both SIN and VEE (38) could not elicit such good CD4⁺ T-cell responses (Fig. 1D), they were effective at eliciting E1E2-specific CD8⁺ T-cell responses (Fig. 1A and B; see Fig. S2B and S3 in the supplemental material). Priming with MF59-CpG-adjuvanted E1E2 followed by boosting with VEE/SIN expressing E1E2 was found to be optimal for eliciting both CD4⁺ and CD8⁺ T-cell responses (Fig. 1C and D; see Fig. S2B and S3 in the supplemental material). This regimen also elicited substantial titers of virus-cross-neutralizing antibody (Fig. 5). Although these titers were somewhat (∼20%) lower than those obtained by the use of recombinant E1E2 adjuvanted with MF59 alone (Fig. 5), the prime/boost regimen allowed the induction of all protective components of the immune response against E1E2 (Fig. 1 and 5; see Fig. S2 and S3 in the supplemental material). The addition of at least 125 μg CpG to MF59-adjuvanted E1E2 has been associated with some systemic side effects in a recent phase 1 clinical trial (S. E. Frey et al., unpublished data), and so it will be important in future trials to explore a lower dose range in order to establish a well-tolerated dose for these HCV antigen formulations.

It should be noted that some variation in the T-cell responses (as measured by ICS assays) was observed between experiments, but a positive result was confirmed in at least three independent experiments (Fig. 1; see Fig. S3 in the supplemental material).

In order to recapitulate cellular immune responses to nonstructural proteins, which have also been associated with protection against HCV infection (12), we also explored the use of an adjuvanted recombinant NS345core polyprotein derived from yeast, as well as replication-defective chimeric VEE/SIN particles expressing NS345 genes. Again, we found that CD8⁺ T-cell responses were induced efficiently by the alphaviral particles and poorly by adjuvanted NS345core polyprotein (Fig. 3B). However, NS345core polyprotein adjuvanted with Iscomatrix plus CpG was optimal in mice for the elicitation of Th1-type CD4⁺ T-cell responses (Fig. 3A). As with E1E2-specific immune responses, optimal levels of both CD4⁺ and CD8⁺ T-cell responses to the NS3, -4, and -5 proteins could be induced via a regimen comprising priming with adjuvanted NS345core polyprotein followed by boosting with VEE/SIN particles expressing NS345 (Fig. 3A and B). It is known that many T-cell epitopes are conserved between different HCV clades (50) and also that eliciting broad cellular immune responses to the HCV NS3, -4, and -5 gene products is associated with the protection of chimpanzees against the development of chronic, persistent infection by highly diverse HCV clades (15, 26). It is to be expected, therefore, that the broad cellular immune responses elicited here by this prime/boost immunization regimen utilizing multiple HCV gene products will be broadly cross-protective against multiple HCV clades.

Importantly, a single immunization regimen was shown to be capable of eliciting these broad T-cell responses, as well as cross-neutralizing antibodies (Fig. 6). We conclude, therefore, that the prime/boost regimen identified here involving priming with adjuvanted envelope glycoproteins and nonstructural proteins of HCV, followed by boosting with defective VEE/SIN particles expressing these genes, is optimal for eliciting the array of known protective immune responses against HCV infection. It should be noted that this vaccine may have potential therapeutic use, as well as prophylactic efficacy, especially when combined with antiviral drugs.