The presence of both cell-mediated and humoral immunity is important in protection from and clearance of a number of infectious pathogens. We describe novel vaccine regimens using combinations of plasmid DNA, poxvirus and protein to induce strong Ag-specific T cell and Ab responses simultaneously in a murine model. Intramuscular (i.m.) immunization with plasmid DNA encoding the middle Ag of hepatitis B (DNA) concurrently with a commercial hepatitis B virus (HBV) vaccine (Engerix-B) followed by boosting immunizations with both modified vaccinia virus Ankara (MVA) encoding the middle Ag of HBV and Engerix-B induced high levels of CD4+ and CD8+ T cells and high titer Ab responses to hepatitis B surface Ag (HbsAg). Substitution of Engerix-B with adjuvant-free rHBsAg induced similar T cell responses and greatly enhanced Ab levels. Repeated immunizations with recombinant or nonrecombinant MVA mixed with Ag induced higher titers of Abs compared with immunization with either Ag or Engerix-B further demonstrating this novel adjuvant effect of MVA. The poxviruses NYVAC, fowlpox (FP9) and ALVAC, and to a lesser extent, adenovirus, also displayed similar adjuvant properties when used in combination with rHBsAg. The use of poxviruses as an adjuvant for protein to concurrently induce Ag-specific T cells and Abs could be applied to the development of vaccines for many diseases, including HIV and malaria, where both cell mediated and humoral immunity may be important for protection.
The major aim of vaccination is to generate sufficiently strong immune responses of the correct nature to protect against disease. Most vaccines that are in used in humans, except bacillus Calmette-Guérin, have been developed by optimizing their capacity to induce Ab responses. Novel candidate vaccines that have been developed to date either induce immune responses that are dominated by cellular or humoral immune responses. For example, vaccines against intracellular pathogens, such as HIV or liver stage malaria, have been designed either to induce strong T cell responses (1, 2, 3) or to induce Ab responses against the extracellular form of the virus or parasite (4, 5). However, it is now widely believed that the generation of both cellular and humoral responses may be important in vaccine efficacy for diverse causes of disease including viruses, bacteria, parasites, and cancer (4, 6, 7, 8). The licensed prophylactic vaccine to hepatitis B virus (HBV),3 Engerix-B, is an example of a vaccine that induces Ab responses. It induces protective Abs in >85% of healthy recipients (9) and comprises the small surface Ag, S (HBsAg), produced in yeast cells (Saccharomyces cerevisiae) adsorbed to the adjuvant aluminum phosphate (alum). A vigorous CD8+ CTL response is detectable in individuals with acute self-limited HBV but is absent in individuals with chronic infection (10, 11, 12). This suggests that a cell-mediated immune response may be important in eliminating virus at an early stage of infection. A vaccine that effectively induces both cellular and humoral immunity may improve the efficacy of a HBV vaccination.
Recombinant DNA and viral vectors have been used for a number of years in laboratory animals and in human clinical trials as vectors for the Ag(s) of interest to induce strong effector and memory CD4+ and CD8+ T cell responses in both the prevention and treatment of disease (2, 13, 14, 15, 16). Widely used viral vectors include modified vaccinia virus Ankara (MVA) and NYVAC, both attenuated forms of vaccinia virus, the avipoxviruses FP9 and ALVAC, and also nonreplicating adenovirus (ADV) (2, 15, 16, 17, 18, 19, 20, 21). We and others have previously demonstrated in a pre-erythrocytic malaria vaccine model, where induction of IFN-γ secreting T cells correlates with protection, that heterologous prime-boost immunization with DNA and viral vectors is protective in mice, primate and human challenge models (2, 13, 16, 22, 23, 50). With respect to HBV vaccines, DNA and MVA encoding the middle Ag of HBV, comprising S and Pre-S2 Ags, have been designed in this laboratory as a potential therapeutic vaccine for chronic HBV infection. This prime-boost regimen has demonstrated potent induction of specific T cells in mice and humans and is currently in phase II clinical trials (J. Schneider, personal communication).
In this study, we examined methods of inducing strong cellular and Ab responses in a mouse model of candidate and licensed HBV vaccines. Although this model cannot be used to assess protection against HBV challenge, immunogenicity of Engerix-B can be tested in mice, and this clinically licensed vaccine functions as a good standard to compare our candidate vaccines to. We report here that prime-boost immunization with DNA and MVA (D/M) induces potent T cell responses but poor levels of Ab against the encoded HBsAg. Conversely, repeat immunization with Engerix-B induced a weak T cell response but greatly increased Abs to rHBsAg compared with D/M. The concurrent administration of these vaccines induced both T cells and Abs to HBsAg. We sought to analyze and further improve this coinduction of cell-mediated immunity (CMI) and humoral immunity by combining the viral vectors MVA, FP, ALVAC NYVAC, and ADV with the main component of Engerix-B, rHBsAg. Here we show that poxviruses can enhance T cell and Ab responses to a coadministered protein while retaining strong T cell immunogenicity for their encoded recombinant Ag thereby demonstrating a method of strongly inducing both types of immunity against a target Ag. This combination vaccine approach may be of benefit to the development of efficacious vaccines where this type of immunity is required to protect against pathogen challenge.
Materials and Methods
Animals and immunizations
Female BALB/c (H-2d) mice (BMSU, John Radcliffe Hospital, Oxford, U.K.), 4–6 wk old, were used in all experiments. Before immunization mice were anesthetized i.p. with 1:1:2 solution of Hypnorm (Janssen-Cilag, Saunderton), Hypnovel (Midazolam; Roche) and endotoxin-free water (Sigma-Aldrich).
Plasmid DNA encoding the middle Ag of HBV was generated by insertion of a gene fragment containing the pre-S2 and S sequences of HBV strain ayw into the polylinker cloning region of the plasmid vector pSG2. Expression of the middle HBsAg was driven by the CMV promoter. DNA (50 μg) in endotoxin-free PBS was administered i.m. bilaterally into each musculus tibialis or mixed with 5 μg of Ag and injected intradermally (i.d.).
MVA.HBs (MVA) contains the gene fragment containing the pre-S2 and S sequences of HBV strain ayw surface Ag inserted into the thymidine kinase locus of MVA. Expression of the middle HBsAg gene is driven by the vaccinia early/late P7.5 promoter. MVA.HBs also contains the vaccinia late promoter P11 driving expression of the lacZ marker gene. The viruses MVA, FP and ADV that were nonrecombinant for HBsAg expressed only the lacZ marker gene, termed Mnr, FPnr and ADnr, respectively. Their generation has been previously described (3, 15, 24). Nonrecombinant ALVAC (ALnr) and NYVAC (NYnr) were a gift from Virogenetics (Albany, NY). All viruses were administered at 5 × 106 pfu in endotoxin-free PBS or mixed with 5 μg of rHBsAg and injected i.d. bilaterally.
Recombinant small surface Ag of HBV (rHBsAg), ayw subtype, was manufactured by BiosPacific in S. cerevisiae. When administered alone, rHBsAg was prepared in endotoxin-free PBS and 5 μg was administered s.c. into the scruff of the neck or i.d. bilaterally into the ears.
Engerix-B is a commercially available vaccine containing recombinant small surface Ag (20 μg/ml HBsAg) produced in S. cerevisiae adsorbed to alum (GSK); 5 μg was administered (s.c.) into the scruff of the neck.
Ex vivo IFN-γ ELISPOT
Cells from spleens were assayed for IFN-γ production as described previously (3). Axial lymph nodes (LN), which drain the s.c injection site, or facial LN, which drain the face and ears, were used to assess draining LN (DLN) responses. Axial or facial DLN were pooled within groups. Cells were stimulated for 18–20 h with 1 μg/ml Ld-restricted peptide IPQSLDSWWTSL (Invitrogen Life Technologies) or 5 μg/ml rHBsAg. Spots were counted using an ELISPOT counter (AID) and are represented as spot-forming cells (SFC) per million ±SEM.
Serum was collected from tail vein blood samples. Blood was allowed to clot overnight at 4°C then centrifuged at 12,000 rpm for 3 min and the serum collected and stored at −20°C. Individual mouse serum was analyzed for anti-rHBsAg Abs by an indirect ELISA. Briefly, 96-well plates (Maxisorp; Nalge Nunc International) were coated with 500 ng/ml rHBsAg diluted in carbonate-bicarbonate buffer and incubated overnight at 4°C. Plates were washed in PBS containing 0.05% Tween 20 (PBS/T) then blocked with 10% skim milk powder in PBS/T for 1 h at 37°C. Sera diluted to 1/200 in PBS/T was added in duplicate wells and serially diluted. Following 1 h of incubation at 37°C, bound Abs were detected using alkaline-phosphatase-conjugated goat anti-mouse IgG Ab (Sigma-Aldrich) or biotin-conjugated anti-mouse IgG1 or IgG2a Ab (BD Pharmingen) followed by incubation with ExtrAvidin (Sigma-Aldrich) for isotype analysis. Plates were developed by adding p-nitrophenyl phosphate substrate (Sigma-Aldrich) to each well. Optical densities at 405 nm were measured for each well. End point titers were determined at the x-axis intercept of the dilution curve, at three times the absorbance (A405) given for naive mouse serum diluted from 1/200 accordingly.
Statistical analysis was performed using SPSS for Windows version 10. Mann-Whitney U tests were used to determine differences between groups.
Immunogenicity of DNA/MVA.HBs and protein immunization
T cell and Ab responses induced by the prime-boost regime of DNA/MVA were compared with homologous protein immunization with or without alum (Fig. 1).
Immunization with DNA/MVA induced high levels of IFN-γ producing cells against peptide encoding a dominant CD8+ epitope and whole rHBsAg in the spleen (Fig. 1,A) and facial DLN (Fig. 1,B). Immunization with HBsAg or Engerix-B induced minimal peptide and rHBsAg responses in spleen. However, the peptide-specific response induced in axial DLN when immunized s.c. with Engerix-B or Ag was equivalent to the response induced by DNA/MVA in facial DLN. Conversely, DNA/MVA or a single immunization of Engerix-B induced very low levels of Abs against rHBsAg compared with repeat immunization with rHBsAg or Engerix-B (Fig. 1 C). These results show that prime-boost DNA/MVA induces a potent IFN-γ response whereas two Engerix-B immunizations induce high levels of Ab to rHBsAg.
Concurrent induction of CMI and humoral immunity
We assessed the concurrent induction of T cell and Ab responses by coadministration of our optimal CTL-inducing vaccine, DNA/MVA with our strongest Ab-inducing vaccine, Engerix-B (labeled D,E/M,E). We also combined DNA/MVA with rHBsAg to examine responses in the absence of alum (labeled D,Ag/M+Ag).
Coadministration of these vaccines induced high levels of T cells and Abs to rHBsAg concurrently (Fig. 2). Priming with DNA and Engerix-B or rHBsAg then boosting with MVA and either Engerix-B or rHBsAg both induced statistically similar levels of peptide and rHBsAg-specific T cells compared with DNA/MVA in the spleens (Fig. 2,A). Peptide responses in facial DLN were increased in groups receiving DNA/MVA with Engerix-B or rHBsAg (Fig. 2 B).
Interestingly, Ab levels were reduced in the DNA/MVA with Engerix-B (D,E/M,E) group compared with repeat immunization with Engerix-B (E/E). However, DNA/MVA with rHBsAg (D,Ag/M+Ag) significantly increased Abs compared with repeat Engerix-B immunization (Fig. 2 C).
We demonstrate here an adjuvant effect of recombinant MVA on a coadministered protein and describe a method of vaccination whereby potent cellular and humoral immunity can be induced concurrently.
Induction of Abs and T cells by homologous immunization with MVA and rHBsAg
Repeated immunization with the same vaccine may improve efficacy in the field. We therefore aimed to induce T cells and Abs simultaneously using one formulation via a clinically relevant route (Fig. 3). Coadministration of MVA and rHBsAg twice (M+Ag/M+Ag) induced a high frequency of CD8+ T cells in spleen (Fig. 3,A) and DLN (Fig. 3,B). Immunizing with nonrecombinant MVA (MVAnr) and rHBsAg induced weak T cell responses in the spleen (Fig. 3,A) although comparatively stronger responses in DLN (Fig. 3,B). Priming with a mixture of DNA and rHBsAg followed by boosting with MVA and rHBsAg (D+Ag/M+Ag) induced higher levels of peptide and rHBsAg specific T cells in spleen and DLN compared with repeat MVA and rHBsAg (M+Ag/M+Ag) immunization (Fig. 3 A).
Immunization with MVA and rHBsAg induced a higher titer of Abs to rHBsAg compared with Engerix-B (Fig. 3,C). Interestingly, priming with a mixture of DNA and rHBsAg followed by boosting with MVA and rHBsAg (D+Ag/M+Ag) or repeat MVAnr and rHBsAg immunization also increased Ab levels compared with Engerix-B (Fig. 3 C).
This demonstrates that two immunizations with rHBsAg mixed with MVA is a potent stimulator of T cells and Abs. However, intracellular expression of the protein, provided by recombinant MVA or DNA, is required for induction of specific T cells in the spleen.
Heterologous immunization with FP and ADV
Lower T cell levels induced by homologous MVA immunization may be due to anti-MVA immune response (25). We have previously demonstrated that heterologous priming with FP or ADV and boosting with MVA induces strong cellular immunity and, in some cases, Ab responses compared with homologous immunization (15, 16). To determine whether heterologous immunization could further increase Ab and T cell responses in our system, mice were primed with DNA i.m. and a mixture of rHBsAg with either nonrecombinant FP (D, FPnr+Ag) or ADV (D, ADVnr+Ag) i.d. All groups were boosted with recombinant MVA and rHBsAg i.d. (M+Ag) (Fig. 4). Intracellular Ag was found to be essential for high cellular responses (Fig. 3 A). DNA was included in the prime to provide an intracellular Ag delivery system for HBsAg in the FPnr and ADVnr groups.
Adding FPnr to the DNA and rHBsAg regimen did not further improve the induction of T cell responses (Fig. 4,A). However, these strong class I responses were reduced when DNA was removed from the prime (Fig. 4,A). In contrast, priming with Adnr+Ag i.d. in the presence or absence of DNA i.m. induced similar T cell responses suggesting that ADnr adjuvants a weak CD8+ T cell response that is independent of intracellular delivery of Ag by DNA (Fig. 4,A). However, these responses were greatly reduced compared with priming with DNA i.m and Ag i.d. (Fig. 4,A). All regimens induced strong peptide responses in DLN, although regimens that included a DNA prime were consistently stronger (Fig. 4 B).
Including FPnr in the DNA and Ag prime (D, FPnr+Ag) enhanced Ab levels compared with DNA and rHBsAg alone (Fig. 4,C). Interestingly, priming with FPnr+Ag i.d. without DNA also induced an Ab response that was not significantly different to immunizing with D, FPnr+Ag/M+Ag, demonstrating that the enhanced Ab response is independent of intracellular Ag expression at the time of priming with this vaccine. Nonrecombinant ADV, with or without DNA, failed to increase Ab responses when combined with rHBsAg (Fig. 4,C). Heterologous immunization with rHBsAg, FPnr, and recombinant MVA (Fig. 4,C) provides a potent method of inducing Abs at similar levels to those induced by homologous immunization with recombinant MVA and rHBsAg (Mnr+Ag/Mnr+Ag) (Fig. 3) (D, FPnr+Ag/M+Ag; p = 0.478, FPnr+Ag/M+Ag; p = 0.794, compared with M+Ag/M+Ag). This demonstrates that FPnr exhibits similar adjuvant properties to MVA in its ability to induce high levels of Abs to rHBsAg whereas, as expected, heterologous FPnr/MVA immunization with DNA and rHBsAg (i.e., D, FPnr+Ag/MVA+Ag) (Fig. 4,A) induces stronger CD8+ responses compared with homologous MVA+rHBsAg immunization (Fig. 3 A).
Homologous immunization of HBsAg mixed with nonrecombinant MVA, NYVAC, ALVAC, FP or ADV
We next examined whether the adjuvanting capacity of MVA in a homologous prime-boost setting could be extended to other poxviruses or to other viruses. We tested candidate nonrecombinant poxviruses, namely, FP, NYVAC (NYnr) and ALVAC (ALnr) and nonrecombinant ADV (ADVnr) in combination with soluble rHBsAg (Fig. 5) Addition of each of these viruses to rHBsAg increased cellular responses compared with repeat immunization with rHBsAg (Fig. 5,A). The most significant increase in T cell responses compared with rHBsAg immunization was induced by recombinant MVA. However, nonrecombinant ALVAC, MVA, FP, NYVAC and ADV also significantly increased T cell responses compared with rHBsAg immunization. All viruses induced potent responses to peptide in facial DLN although interestingly the most potent responses were induced by ALVACnr (Fig. 5 B).
All poxviruses enhanced the induction of potent Abs against rHBsAg whereas ADVnr failed to increase Ab responses above the level of repeat rHBsAg immunization (Fig. 5 C). The attenuated vaccinia virus, NYVACnr, induced similar Ab responses to nonrecombinant MVA whereas the avipox virus, ALVACnr and recombinant MVA elicited the highest levels of anti-HBsAg Abs.
Homologous immunization with recombinant and nonrecombinant poxvirus mixed with Ag is an efficient method for inducing potent Ab levels against the coadministered protein and, surprisingly, the nonrecombinant poxviruses and ADV vector also increases the T cell response induced by the coadministered protein.
Poxviruses modulate IgG subclasses
The ratio of isotype subclasses IgG1 and IgG2a gives an indication of Th2 or Th1 bias of humoral responses, respectively. The effect of DNA, poxviruses, and ADV on IgG subclass division was determined (Fig. 6).
Engerix-B predominantly induced IgG1 Abs which is indicative of a Th2 biased immune response. This is consistent with the poor IFN-γ production by peptide and rHBsAg stimulated splenocytes from Engerix-B immunized animals (Fig. 1 A). Alum inhibited the induction of IgG2a when DNA/MVA was combined with recombinant HBsAg, as no IgG2a was induced when DNA/MVA was used in combination with Engerix-B (D,E/M,E) compared with soluble Ag (D,Ag/M+Ag). In these groups, the level of IgG1 was not changed by the inclusion of DNA/MVA. Adding FPnr to this combination of DNA and Ag and boosting with MVA and Ag (D,FPnr+Ag/M+Ag) significantly increased the IgG1 and IgG2a endpoint titers demonstrating that FP adjuvanted the prime by DNA and Ag. Removing the source of intracellular Ag, by omitting the DNA in this mix (Fnr+Ag/M+Ag), resulted in significantly decreased IgG1 and IgG2a (D,FPnr+Ag/M+Ag compared with Fnr+Ag/M+Ag; p < 0.05 for both IgG1 and IgG2a). This effect of intracellular Ag was also observed when nonrecombinant MVA was used instead of HBsAg expressing MVA in a homologous prime-boost regimen (Mnr+Ag/Mnr+Ag). However, priming with a mixture of rHBsAg and recombinant MVA or nonrecombinant FP, MVA, ALVAC, or NYVAC with or without DNA consistently increased levels of IgG2a compared with two immunizations with Engerix-B or Ag alone. Combination of rHBsAg with nonrecombinant MVA, NYVAC, or ALVAC, induced a 1:1 ratio of IgG1:IgG2a. In contrast, two immunizations with ADVnr and rHBsAg induced weak IgG1 and undetectable IgG2a levels compared with immunization with rHBsAg alone. Heterologous immunization, using FPnr in the prime and MVA in the boost (FPnr+Ag/M+Ag) induced higher IgG1 and IgG2a Abs compared with two immunizations with FPnr+Ag. The increased levels of IgG2a by including poxviruses with Ag indicates a Th1 biased humoral response, whereas priming with combinations including Engerix-B lead to a higher ratio of IgG1 and therefore Th2 bias. Nonrecombinant MVA, ALVAC and NYVAC increase the IgG2a response. These results indicated that equal Th1 and Th2 or biased responses can be primed to the same Ag depending on the co-delivered viral vector.
The presence of both CMI and humoral immunity is important in protection from and clearance of a number of pathogens (4, 8). We sought to improve the co-induction of T cell and Abs to HBsAg by using the licensed prophylactic hepatitis B vaccine, Engerix-B, or rHBsAg combined with DNA and viral vectors. The results here describe simultaneous induction of strong CMI and humoral immunity using recombinant and nonrecombinant poxviruses to adjuvant a coadministered protein, thereby demonstrating a novel method for the induction of potent T cell and Abs against a target Ag. We have further demonstrated that responses can also be induced by homologous immunization, providing a cheaper and easier method of manufacture and administration compared with heterologous immunization. Prime-boost immunization with DNA/MVA elicited a high frequency of T cells but a minimal Ab response, whereas two immunizations with Engerix-B induced specific Abs to rHBsAg but poor CMI. Interestingly, combining these vaccines in a concurrent prime-boost regimen induced high levels of specific T cells and Abs simultaneously. The levels of peptide and HbsAg-specific T cells were maintained in the spleen whereas responses to peptide were increased in DLN. In all cases, T cell responses in the spleen were higher than those observed in the DLN. The enhancement of immune responses in DLN by all regimens containing poxviruses or ADVnr may provide important information on the strength, breadth, and longevity of the vaccine (26). T cells stimulated with the chosen peptide represent a dominant CD8+ population (27, 28) whereas rHBsAg stimulation characterized total IFN-γ production by both CD4+ and CD8+ T cells although the measured levels under-represent the total T cell response probably due to limited processing of the protein into peptides. We verified the induction of CD4+ and CD8+ cells by intracellular cytokine staining (data not shown). Our initial investigations demonstrated that a strong cellular and Ab response can be achieved simultaneously. The potent cellular responses induced by the combination regimen were abrogated if MVA and Engerix-B were coadministered subcutaneously. This reduction in T cell responses correlates with the finding that priming with DNA i.m. followed by boosting with MVA s.c. also failed to induce high levels of CTL (data not shown) and s.c. administration was shown previously to be a poor route for T cell induction by recombinant MVA (3). The effect of alum on MVA is unknown. Previous studies have shown that the presence of alum can inhibit CTL induction to particulate Ags (29) and may also impair the capacity of MVA to induce a T cell response. We therefore substituted Engerix-B with rHBsAg to avoid interference by alum. Interestingly, combining DNA/MVA with rHBsAg induced similar T cell responses compared with DNA/MVA with Engerix-B. Mixing and coadministration of DNA/MVA and rHBsAg (D, Ag/M+Ag) clearly increased Ab responses, in contrast to DNA/MVA and Engerix-B immunization suggesting that MVA is adjuvanting rHBsAg but can only do so effectively in the absence of alum.
Initial experiments indicated that combination vaccines could induce potent cellular and humoral responses but were complicated by their varied components and delivery sites which would make them less effective from a practical viewpoint. We therefore sought to simplify our vaccine regimen while still using the adjuvant capacity of recombinant MVA through homologous immunization regimens. Potent levels of T cells and Abs were induced by repeat immunization with a mixture of MVA and Ag. When compared with our initial combination vaccines used in heterologous prime-boost regimes, there was some reduction in T cell responses, although a significant increase in specific Abs was elicited compared with our previous best regimen for inducing Abs, DNA/MVA, and rHBsAg. Nonrecombinant MVA also demonstrated the capacity to enhance Ab levels to rHBsAg; however, optimal Ab production was induced by homologous immunization when MVA encoded the coadministered Ag. The viral vector MVA is therefore adjuvanting B cell responses but intracellular production of Ag is required to enhance induction of CD8+ T cells and to a lesser degree, Abs. MVA may achieve this adjuvant capability through induction of TNF-α and IL-6 (25) both of which have been shown to support plasma cell survival (30). It is unclear exactly how intracellular Ag expression is able to enhance Ab responses, although it is possible that HBsAg may be secreted from the infected cell (31). Alternatively, endogenous production of Ag may allow it to access an alternative route of processing and presentation, such as using nascent as well as recycling MHC class II molecules, which may enhance CD4+ T cell responses. Indeed, T cell responses to whole Ag in vitro were increased in all cases when Ag was delivered by DNA or a vectored vaccine, suggesting that CD4+ T cell responses were induced. These CD4+ Th cells may provide B cell help resulting in increased Ab responses. Immunization with recombinant MVA leads to intracellular production of HBsAg and presentation in MHC class I molecules. Nonrecombinant MVA insufficiently stimulated class I processing of the coadministered exogenous protein, hence a reduced CD8+ T cell response in this regimen. However, we demonstrate the adjuvant capabilities of other nonrecombinant poxviruses, including ALVAC, NYVAC, and FP, for Ab induction. Boudet et al. (19) demonstrated that cellular infiltration triggered by the presence of ALVAC enhanced the humoral immune response to coadministered HIV gp160 but CD8+ and other T cell responses were not measured. MVA, NYVAC, and FP9 are all poxviruses and MVA has also been shown to induce an inflammatory response in vitro (32). Evidence suggests that viral factors and/or interaction of virus with the host are important mediators of cellular immunity to HBsAg (33). In contrast to this study (33), the poxviruses we used are nonreplicative in mammalian cells (34) but are still capable of inducing CMI and Abs to a coadministered protein. This suggests that the interaction between host and virus is sufficient to enhance immunogenicity, most likely through affecting some form of innate immunity. Viruses induce cellular infiltration and induction of cytokines such as type I IFNs, TNF-α, and IL-6 (19, 25) probably through TLR-mediated signaling (35), which may lead to increased uptake and presentation of encoded and delivered Ag. Previous studies suggest that dendritic cells (DC) take up Ag or are infected by virus in the periphery before homing to the DLN where they present to T and B cells (36, 37, 38, 39). MVA has also been shown to affect DCs in vivo by inducing their activation and maturation (40). It has previously been shown that inoculation of mice with influenza and simultaneous immunization with protein induces DC maturation, which was associated with specific immunity to the coadministered protein (41). The viruses we have used may be inducing DC maturation leading to enhanced uptake and presentation of coadministered rHBsAg. These Ag-loaded DC may then migrate to the DLN where they stimulate naive and memory T cells, a possible explanation for why we see such potent responses following immunization with the viral vectors.
It is well established that recombinant poxviruses induce in vivo expansion of pre-primed CD8+ T cells (2, 17, 42) but are also immunologically potent when different viruses encoding the same Ag are used in prime and boost (15, 16, 43, 44). We recently reported that heterologous prime-boost immunization with FP and MVA, both encoding the same malarial Ag, induced greater numbers of specific CD8+ T cells and protection from challenge when compared with homologous immunization with either vector (16). To determine whether heterologous prime-boost could improve immunogenicity in our system, animals were primed with DNA, rHBsAg, and nonrecombinant FP9 then boosted with MVA and rHBsAg. DNA was included in the prime to provide intracellular expression of HBsAg and for enhancing T cell responses. Inclusion of FP9 at prime increased T cell responses whereas similar Ab levels were maintained when compared with homologous MVA and rHBsAg immunization. This result supported heterologous immunization as a more effective strategy than homologous immunization for T cell induction. Kent et al. (45) demonstrated that Ag-specific T cells primed by a single DNA immunization were subsequently boosted with a non-FP9 strain of recombinant FP virus, but this FP boost resulted in reduced Abs. In our study, including nonrecombinant FP9 in the prime with rHBsAg and DNA immunization likewise primed Ab and cellular responses, both of which were further amplified following boosting with MVA and rHBsAg.
Recombinant ADV is another commonly used vector for inducing strong responses to encoded Ags (15, 43, 46, 47). Priming with ADV expressing a Plasmodium yoelii Ag and boosting with VV encoding the same Ag induced specific T cells and Abs and provided complete protection from P. yoelii challenge in mice (43). In this study, priming with DNA and nonrecombinant ADV mixed with rHBsAg followed by boosting with MVA and rHBsAg resulted in a weaker peptide-specific T cell response, but similar Ab levels, compared with coimmunization of DNA/MVA and rHBsAg. In contrast to poxviruses, the adjuvant effect of ADV was much weaker on T cell responses and had no effect on Ab levels. Responses in DLNs were similar to more potent regimens suggesting that ADVnr may be adjuvanting the coadministered protein although only locally. Furthermore, CD8+ T cell responses were induced in the absence of intracellular Ag in the heterologous ADVnr+Ag/M+Ag regimen, which were of a similar magnitude to those induced when DNA was present in the prime. This suggests that ADV is adjuvanting T cell responses by introducing the soluble Ag into a MHC class I processing pathway. Repeated administration of ADVnr in a homologous prime-boost regimen increased the T cell response to a similar extent to repeated use of some poxviruses but this strategy did not enhance Ab responses above those levels observed by immunization with Ag alone. It is known that ADVnr induces strong anti-vector responses against itself and may therefore be unsuitable for use in homologous immunization strategies (20, 21). The differential adjuvant effect of ADV on cellular vs humoral immunity may be related to the induction of a qualitatively different anti-vector immune response after the priming immunization with ADV compared with poxviruses, which effects the capacity of ADV to boost humoral immunity compared with cellular immune responses. Alternatively, it may be related to different cytokines induced by each of these viruses, which then exert varied effects on humoral and cellular immunity.
The example of HBV prevention provides a good illustration of how our strategies might be of clinical value. Approximately 5% of HBV-vaccinated recipients do not generate protective Ab levels (9). In our regimens, we demonstrate not only an increase in specific Abs but also a boost in CD4+ T cells, which may in part be responsible for this amplification of Ab levels. Increasing the number of CD4+ T cells may overcome the problem of nonresponders to this vaccine and possibly to other vaccines. Clinical trials of investigational vaccines involve the use of recombinant viruses and protein that are generally used separately at different time points (48, 49). We demonstrate that combination vaccines, whereby the virus and protein-based vaccines are concurrently administered is an alternative strategy that induces enhanced immunogenicity. The mouse model is unsuitable as a HBV challenge model, therefore we cannot directly demonstrate that coadministering viruses and recombinant HBsAg results in improved protection against virus. However, we believe that the significant increase in the magnitude and robustness of the Ab and T cell responses induced by these combination vaccines may result in improved protection against pathogen challenge. We have recently successfully applied this strategy to a malaria challenge model in mice (C. Hutchings, A. Birkett, A. Moore, and A. Hill, manuscript in preparation).
Here we describe a novel vaccine regimen capable of inducing CD4+ and CD8+ T cells and high levels of specific Abs concurrently. We have demonstrated that combination and route of administration of the vaccine components is critical for inducing optimal cellular and humoral responses to the encoded and exogenous Ag. We used a well-described subunit protein and recombinant viral vectors, both acceptable for clinical use, to enhance immune responses. These results may have important implications for the rational design of future vaccination strategies.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Wellcome Trust. A.V.S.H. is a Wellcome Trust Principal Fellow.
Abbreviations used in this paper: HBV, hepatitis B virus; HBsAg, small hepatitis B surface Ag; MVA, modified vaccinia virus Ankara; ADV, adenovirus; ADVnr, nonrecombinant adenovirus; CMI, cell-mediated immunity; DC, dendritic cell.; LN, lymph node; DLN, draining LN; FP, fowlpox; i.d., intradermal; SFC, spot-forming cells.