Vaccination with an Adenoviral Vector That Encodes and Displays a Retroviral Antigen Induces Improved Neutralizing Antibody and CD4⁺ T-Cell Responses and Confers Enhanced Protection

Abstract

We present a new type of adenoviral vector that both encodes and displays a vaccine antigen on the capsid, thus combining in itself gene-based and protein vaccination; this vector resulted in an improved vaccination outcome in the Friend virus (FV) model. For presentation of the envelope protein gp70 of Friend murine leukemia virus on the adenoviral capsid, gp70 was fused to the adenovirus capsid protein IX. When compared to vaccination with conventional FV Env- and Gag-encoding adenoviral vectors, vaccination with the adenoviral vector that encodes and displays pIX-gp70 combined with an FV Gag-encoding vector resulted in significantly improved protection against systemic FV challenge infection, with highly controlled viral loads in plasma and spleen. This improved protection correlated with improved neutralizing antibody titers and stronger CD4⁺ T-cell responses. Using a vector that displays gp70 without encoding it, we found that while the antigen display on the capsid alone was sufficient to induce high levels of binding antibodies, in vivo expression was necessary for the induction of neutralizing antibodies. This new type of adenovirus-based vaccine could be a valuable tool for vaccination.

Adenoviruses have been a focus of interest as vaccine vectors for more than a decade and have been tested in various preclinical and clinical studies for vaccination against viral and bacterial infections (reviewed in reference 38). This interest is based on the ability of adenoviral vectors to induce high antibody titers and robust cytotoxic T-lymphocyte (CTL) responses and on the high immunogenicity of the vector, which might have an adjuvant effect on vaccination (17). Adenoviral vectors have also been extensively evaluated for immunization against HIV (reviewed in reference 1), where they were used either alone or in combination with plasmid DNA or protein in prime-boost immunizations. However, vaccination with adenoviral vectors against HIV showed no effectiveness in a large phase IIb study (4), but it is conceivable that the observed lack of effectiveness was due to the choice of vaccine antigen rather than the vector itself, as the vaccine relied exclusively on the induction of CTL responses, and the outcome was unexpected given previous results from studies in nonhuman primates (33, 42). The findings of the phase IIb study brought about a shift of focus from the CTL response to a more balanced immune response, including neutralizing antibodies, that is now expected to be necessary for protection from HIV infection.

Apart from adenoviral vectors that encode vaccine antigens, there have also been approaches to modify adenoviral capsid proteins to include antigenic epitopes. These were mostly inserted into external loops of the hexon protein (5, 22, 25, 26, 43), which is the main component of the adenovirus capsid, but also other components of the capsid, such as fiber, protein IX, and penton base, have been evaluated (22). These studies showed that incorporation of single epitopes into capsid proteins of adenovirus leads to induction of antibody and CD4⁺ T-cell responses, suggesting that incorporation of epitopes into the adenovirus capsid is a useful tool for epitope-based vaccination.

Fusion of a polylysine sequence or an arginine-glycine-aspartic acid motif to adenovirus pIX has been shown to be a tool for redirection of adenovirus tropism to heparan sulfate and α_vβ integrins, respectively (9, 41). By fusing green fluorescent protein and luciferase to the C terminus of pIX, it was shown that relatively large proteins can be displayed on the adenovirus capsid while maintaining the protein's conformation and function as well as virion integrity (24, 28).

Here we describe a novel vaccination approach that combines genetic and protein vaccination by using adenoviral vectors not only as gene expression vectors but also as nanoparticle carriers for a vaccine antigen to improve the vaccination efficiency through enhanced induction of antibodies. Display of the vaccine antigen on the adenovirus capsid was achieved by fusion of the antigen to the C terminus of the adenovirus capsid protein pIX. It was shown before that the presentation of antigens in ordered arrays leads to improved antibody responses by cross-linking of B-cell receptors (13). As the adenoviral capsid is highly structured, we hypothesized that fusion to pIX would result in an ordered display of the antigen, presumably facilitating antibody induction.

We evaluated this vaccine approach using the Friend virus (FV) infection model. FV is an immunosuppressive retroviral complex that consists of Friend murine leukemia virus (F-MuLV) and the replication-deficient, F-MuLV-dependent spleen focus-forming virus. FV infection of susceptible mice induces rapid polyclonal erythroblast proliferation, which leads to splenic enlargement and erythroleukemia and takes a lethal course also in adult mice (14). Protection from FV infection has been shown to require complex immune responses involving antibodies as well as CD4⁺ and CD8⁺ T cells (7). FV is regarded as a useful retrovirus infection model because basic requirements for vaccine protection seem to be similar for FV and HIV infection (8). We demonstrated previously that the FV model is suitable to evaluate and improve adenoviral vectors for antiretroviral vaccination (2), as we showed that a heterologous prime-boost vaccination with adenovirus type 5 (Ad5) and fiber chimeric Ad5F35 vectors led to better protection from FV infection than homologous vaccination, which correlated with improved induction of neutralizing antibodies.

For vaccination with expression/display vectors against FV we constructed a fusion protein of the adenoviral capsid protein pIX and the F-MuLV envelope protein gp70 and produced adenoviral vectors expressing the pIX-gp70 fusion protein, which was incorporated into the viral capsid. We vaccinated FV-susceptible CB6F1 hybrid mice with antigen expression/display vectors or with conventional antigen-expressing adenoviral vectors and analyzed the protection conferred by these two vaccines. Having demonstrated that the expression/display vector leads to better protection of mice from FV challenge, we constructed a panel of expression/display vectors displaying different fusion proteins containing F-MuLV Env or Gag in order to elucidate the underlying immunological mechanisms of the improved protection conferred by the adenoviral expression/display vectors.

MATERIALS AND METHODS

Cells and cell culture.

The human embryonic kidney cell lines 293 (Microbix Biosystems, Toronto, ON, Canada) and 293T (CRL-11268; American Type Culture Collection, Manassas, VA) were propagated in Dulbecco's modified Eagle medium with high glucose. The cell line 293.pIX-gp70 was generated by transfection of 293 cells with a pcDNA vector encoding pIX-gp70 followed by neomycin selection. A murine fibroblast cell line from Mus dunni (23) was maintained in RPMI medium (Invitrogen/Gibco, Karlsruhe, Germany). Cell culture media were supplemented with 10% heat-inactivated fetal bovine serum (Invitrogen/Gibco) and 50 μg/ml gentamicin. Cell lines were maintained in a humidified 5% CO₂ atmosphere at 37°C.

Vaccines.

The adenoviral vectors Ad5.env, Ad5F35.env, Ad5.gag, and Ad5F35.gag (2) encode full-length F-MuLV Env or Gag proteins amplified by PCR from F-MuLV clone FB29; vectors were obtained using the AdEasy system and vectors pAdTrackCMV, pAdEasy-1, and pAdEasy-1/F35.

For the construction of pIX-fusion protein-encoding vectors we employed a modified pShuttle vector from which pIX was deleted by digestion with MfeI and BglII and religation after exonuclease treatment. A cassette originating from pAdTrack-CMV containing a cytomegalovirus (CMV) promoter, multiple cloning site, and polyadenylation signal was inserted into exonuclease-treated KpnI-, SalI-restricted pShuttleΔpIX, resulting in pShuttleΔpIXCMVpA. For construction of the fusion proteins, the entire pIX sequence was PCR amplified. The sequences encoding F-MuLV clone FB29 Env gp70 or the complete Gag were amplified by PCR and added to the C terminus, resulting in fusion proteins pIX-gp70 and pIX-gag. For construction of pIX-p15E, the extracellular part of the transmembrane envelope protein p15E was PCR amplified and added to the C terminus of pIX. In the construct pIX-gp70∼p15E, gp70 and the extracellular domain of p15E were linked via a glycine-serine linker by overlap extension PCR and added to the C terminus of pIX to present a larger part of Env on the capsid. The coding sequences for the fusion proteins were inserted into pShuttleΔpIXCMVpA, and Ad5 and Ad5F35 vectors were generated by homologous recombination with pAdEasy-1 and pAdEasy-1/F35, respectively, and transfection into 293 cells as described before (2).

To obtain adenoviral vectors that carry the fusion protein on their surface without encoding it, a full-length Gag-encoding, pIX-deleted vector was constructed using pShuttleΔpIX. To obtain pIX-gp70-displaying adenoviral particles, 293.pIX-gp70 cells were used for vector production.

For vaccination against ovalbumin (Ova), full-length Ova was fused to pIX, resulting in the adenoviral vector Ad5.pIX-ova. For a conventional Ova-expressing adenoviral vector, we used the previously described vector Ad-ΔGM-OVA (40), which shall be referred to as Ad5.ova in this publication. Ova-coated exosomes (Exo-ova) were produced as described previously (39). The Ova amounts on Ad5.pIX-ova and Exo-ova were determined by enzyme-linked immunosorbent assay (ELISA).

All adenoviral vectors were purified with the Vivapure AdenoPACK 100 kit (Vivascience, Hannover, Germany). The adenovirus particle concentrations were determined by spectrophotometry as described previously (29) and expressed as viral particles (VP)/ml. The particle-to-PFU ratio of all vector preparations was ∼30:1. Expression levels of the viruses and incorporation of the fusion proteins into the particles were verified by Western blotting using F-MuLV gp70-specific monoclonal antibody (MAb) 720 (31).

For plasmid vaccination, gp70 (comprising the gp70 coding sequence without leader peptide) and pIX-gp70 were expressed from plasmid pCG. Plasmid preparations were purified using a NucleoBond PC2000 endotoxin-free Mega-prep kit (Macherey-Nagel, Düren, Germany).

Mice.

We purchased female C57BL/6 mice (H-2^b/b Fv1^b/b Fv2^r/r Rfv3^r/r) from Elevage Janvier (Le Genest-St.-Isle, France) and female CB6F1 hybrid mice (BALB/c × C57BL/6 F₁; H-2^b/d Fv1^b/b Fv2^r/s Rfv3^r/s) from Charles River Laboratories (Sulzfeld, Germany). OT-I mice that express a transgenic T-cell receptor specific for the OT-I peptide SIINFEKL were bred in the animal facility of the Ruhr University Bochum under pathogen-free conditions. All mice entered the study when they were between 9 and 10 weeks of age and were treated in accordance with the regulations and guidelines of the institutional animal care and use committee of the Ruhr University Bochum.

FV immunization.

CB6F1 mice underwent a homologous prime-boost vaccination schedule with a 21-day interval. For adenovirus-based immunization, 5 × 10⁹ VP each of Env and Gag, or the respective Env or Gag fusion protein-encoding vectors were injected in 100 μl of phosphate-buffered saline (PBS) into both hind footpads (50 μl each) as described before (2). Ad5 vectors were used for the prime immunizations, and fiber-chimeric Ad5F35 vectors were used for boost immunizations.

For DNA immunization, C57BL/6 mice were injected with 50 μg of plasmid pCGgp70 or pCGpIX-gp70. DNA was applied intramuscularly followed by electroporation (two low-voltage pulses of 150 V for 25 ms) using a Cliniporator (IGEA, Italy).

FV and challenge infection.

Uncloned, lactate dehydrogenase-elevating virus (LDV)-free FV stock was obtained from BALB/c mouse spleen cell homogenate (10%, wt/vol) 14 days postinfection with a B-cell-tropic, polycythemia-inducing FV complex (3). C57BL/6 and CB6F1 mice were challenged by the intravenous injection of 10⁴ and 250 spleen focus-forming units (FFU), respectively. In CB6F1 mice, the course of disease was monitored twice a week by palpation of the spleen of each animal under general anesthesia. The spleen size was rated on a scale ranging from 1 (normal spleen size) to 4 (severe splenomegaly), as described previously (16).

Viremia assay.

Ten days postchallenge (p.c.), plasma samples from CB6F1 mice were obtained, and viremia was determined in a focal infectivity assay (34). Serial dilutions of plasma were incubated with M. dunni cells for 3 days under standard tissue culture conditions. When cells reached ∼100% confluence, they were fixed with ethanol and labeled first with F-MuLV Env-specific MAb 720 (31) and then with a horseradish peroxidase (HRP)-conjugated rabbit anti-mouse Ig antibody (Dako, Hamburg, Germany), followed by aminoethylcarbazole (Sigma-Aldrich, Deisenhofen, Germany) as substrate to detect foci. Foci were counted, and the FFU/ml plasma was calculated.

Viral load determinations in spleens by infectious center assay or quantitative reverse transcription-PCR (RT-PCR).

FV-infected animals were sacrificed by cervical dislocation, the spleens were removed and weighed, and single-cell suspensions were prepared. Serial dilutions of isolated spleen cells were seeded onto M. dunni cells and incubated under standard tissue culture conditions for 3 days, fixed with ethanol, and stained as described for the viremia assay. Resulting foci were counted, and infectious centers (IC)/spleen values were calculated.

RNA was isolated from spleen cells using TRIzol reagent (Invitrogen, Karlsruhe, Germany), and quantitative RT-PCR was performed on a Rotor-Gene 3000 (Corbett Research, Mortlake, Australia) using the QuantiTect probe RT-PCR kit (Qiagen, Hilden, Germany) and primers and the fluorogenic PCR probe that were described before (37).

F-MuLV-binding antibody ELISA.

MaxiSorp ELISA plates (Nunc, Roskilde, Denmark) were coated with whole F-MuLV antigen, blocked with fetal calf serum, and incubated with serum dilutions. Binding antibodies were detected using a polyclonal rabbit-anti-mouse HRP-coupled anti-IgG antibody (Dako Deutschland GmbH, Hamburg, Germany) and the substrate tetramethylbenzidine (Dako Deutschland GmbH, Hamburg, Germany). Sera were considered positive if the optical density at 450 nm was 3-fold higher than that obtained with sera from naïve mice.

Assay for complement-requiring F-MuLV-neutralizing antibodies.

To detect neutralizing antibodies, serial dilutions of plasma in PBS were mixed with purified F-MuLV and guinea pig complement (Institut Virion/Serion GmbH, Würzburg, Germany), incubated at 37°C for 60 min, and then added to M. dunni cells that had been plated at a density of 7.5 × 10³ cells per well in 24-well plates the day before. Seventy-two hours later cells were stained as described for the viremia assay. Dilutions that resulted in a reduction of foci number by 50% or more were considered neutralizing.

Tetramer staining of F-MuLV Env-specific T cells.

Spleens of CB6F1 mice were removed 3 days p.c., and single-cell suspensions were prepared. Spleen cells were stained with a phycoerythrin (PE)-coupled major histocompatibility complex (MHC) class II tetramer (containing the I-Ab-restricted F-MuLV Env epitope EPLTSLTPRCNTAWNRLKL [20]; kindly provided by the MHC Tetramer Core Facility of the National Institutes of Health, National Institute of Allergy and Infectious Disease, Atlanta, GA), peridinin chlorophyll protein (PerCP)-anti-CD4, and fluorescein isothiocyanate (FITC)-anti-CD11b (Becton Dickinson, Heidelberg, Germany). Data were acquired on a flow cytometer (FACSCalibur; Becton Dickinson, Mountain View, CA) and analyzed using CellQuest Pro software (version 4.0.1; Becton Dickinson).

Intracellular cytokine staining.

Cytotoxic T cells were characterized by intracellular cytokine staining after vaccination with Ova-encoding vectors. C57BL/6 mice were immunized with10⁸ infectious particles of Ad5.ova or Ad5.pIX-ova or with exosomes containing 200 ng Ova protein as determined by ELISA. Two weeks after immunization with Ova vectors, spleen cells were stimulated in vitro with OT-I peptide (SIINFEKL; Genaxxon, Biberbach, Germany) for 6 h. Cells were stained with PerCP-anti-CD8, PE-anti-gamma interferon (IFN-γ), allophycocyanin (APC)-anti-intereukin-2 (IL-2), and FITC-anti-tumor necrosis factor alpha (TNF-α); all from Becton Dickinson, Heidelberg, Germany, and analyzed by flow cytometry.

OT-I cell proliferation assay.

Presentation of antigen to CD8⁺ T cells was analyzed after vaccination with Ova vectors. C57BL/6 mice were immunized once with 10⁷ infectious particles of Ad5.ova or Ad5.pIX-ova or with exosomes containing 200 ng of Ova protein. Injections were given in 100 μl of PBS into both hind footpads. Four days later, 2 × 10⁶ carboxyfluorescein succinimidyl ester (CFSE)-stained OT-I cells were adoptively transferred by tail vein injection. Three days later the brachial, axillary, inguinal, and popliteal lymph nodes were removed, and single-cell suspensions were prepared and stained with APC-anti-CD8 (Becton-Dickinson, Heidelberg, Germany). CFSE staining intensity of CD8⁺ cells was determined by flow cytometry.

Statistical analyses.

Statistical analyses were performed using the software SigmaStat 3.1 (Systat Software GmbH, Erkrath, Germany), testing with a one-way analysis of variance (intracellular staining) or with the Kruskal-Wallis one-way analysis of variance on ranks (all other analyses) and Student-Newman-Keuls multiple comparison procedure.

RESULTS

Vaccination with adenoviral expression/display vectors induces protection from FV-induced splenomegaly.

By fusion of F-MuLV proteins to the adenoviral capsid protein IX, we generated vaccine antigen-displaying adenoviral vectors that were compared to conventional, antigen-expressing adenoviral vectors. We prepared pIX-deleted Ad5- and fiber-chimeric Ad5F35-based vectors encoding a pIX-gag fusion (Ad5.pIX-gag/Ad5F35.pIX-gag) or a fusion of pIX and the surface envelope protein gp70 (Ad5.pIX-gp70/Ad5F35.pIX-gp70). To increase the part of the Env protein presented on the vectors, the extracellular domain of the transmembrane Env protein p15E was fused to pIX-gp70 via a glycine-serine linker (Ad5.pIX-gp70∼p15E/Ad5F35.pIX-gp70∼p15E), and a fusion of pIX and the extracellular part of p15E alone was constructed as a control (Ad5.pIX-p15E/Ad5F35.pIX-p15E). To determine the importance of in vivo antigen expression, a Gag-encoding, pIX-deleted vector was propagated in constitutively pIX-gp70-expressing cells, resulting in a gp70 display-only vector that displayed pIX-gp70 on the capsid without encoding it (Ad5.gag^pIXgp70/Ad5F35.gag^pIX-gp70; vector design is illustrated in Fig. 1A).

FIG. 1.

Vector design and gp70 expression levels and incorporation of pIX-gp70 fusion protein into adenoviral particles. (A) Schematic presentation of Ad5-based conventional (upper panel) and expression/display (lower panel) vectors. The cell lines used for production of the viruses are denoted on the right. The corresponding fiber-chimeric Ad5F35 vectors differ only in the encoded fiber protein. ITR, inverted terminal repeat; Ψ, packaging signal; CMV-IE, CMV immediate-early promoter; pA, polyadenylation signal; ΔpIX, construct with deletion of the pIX coding sequence. (B) Expression levels of the indicated proteins were analyzed in a Western blot assay of lysates of 293 cells transfected with the respective shuttle plasmids. Supernatant of the hybridoma cell line H720 and an HRP-coupled secondary antibody were used for detection. Samples loaded onto the gel were normalized for protein content. (C) For Western blot analysis of purified adenoviral particles, equal amounts of viral particles were submitted for analysis.

gp70 expression levels of the expression/display and conventional adenovirus vectors were compared by Western blot analysis of cell lysates of transfected cells and found to be comparable, with a slightly lower expression of the fusion proteins (Fig. 1B). Incorporation of the gp70 fusion proteins into the particles was verified by Western blot analysis of purified adenovirus particles. Using a defined number of viral particles for Western blot analysis we ascertained that similar amounts of antigen were incorporated into the various vector constructs, including the display-only vector (Fig. 1C).

Highly susceptible CB6F1 mice were immunized twice over a 3-week interval with Ad5- and fiber-chimeric Ad5F35-based vectors in a heterologous prime-boost vaccination. The fiber-chimeric Ad5F35 is an Ad5-based vector with the fiber knob and shaft domains of Ad35. We reported before that this heterologous prime-boost immunization leads to improved immune protection over homologous vaccination (2). Mice were immunized using combinations of Gag- and Env-encoding vectors as detailed in Table 1. Combinations of Gag and Env vectors were used for all vaccinations, as that should result in better protection than vaccination with either Gag or Env vectors alone and it allows the direct comparison of all vaccination groups.

TABLE 1.

Adenoviral conventional and expression/display vectors used for vaccination against FV

Vaccination group	Vectors	Transgene(s)	F-MuLV antigen displayed on capsid
Env + Gag	Prime: Ad5.env, Ad5.gag Boost: Ad5F35.env, Ad5F35.gag	env, gag
pIX-gp70 + Gag	Prime: Ad5.pIX-gp70, Ad5.gag Boost: Ad5F35.pIX-gp70, Ad5F35.gag	pIX-gp70, gag	gp70
GagpIX-gp70 + Gag	Prime: Ad5.gagpIX-gp70, Ad5.gag Boost: Ad5F35.gagpIX-gp70, Ad5F35.gag	gag	gp70
Env + pIX-gag	Prime: Ad5.env, Ad5.pIX-gag Boost: Ad5F35.env, Ad5F35.pIX-gag	env, pIX-gag	Gag
pIX-p15E + Gag	Prime: Ad5.pIX-p15E, Ad5.gag Boost: Ad5F35.pIX-p15E, Ad5F35.gag	pIX-p15E, gag	p15E
pIX-gp70∼p15E + Gag	Prime: Ad5.pIX-gp70∼p15E, Ad5.gag Boost: Ad5F35.pIX-gp70∼p15E, Ad5F35.gag	pIX-gp70-GS4 linker-p15E, gag	gp70∼p15E

Mice were challenged with FV 3 weeks after the boost immunization. Spleen palpations were performed twice a week after FV challenge, and spleen sizes were categorized as described before (16). Palpation showed a significant effect of vaccination on the development of splenomegaly (Fig. 2A and B). On day 10 p.c., all vaccinated mice had significantly smaller spleens than unvaccinated control mice, and mice immunized with pIX-gp70 or pIX-gp70∼p15E expression/display vectors had significantly smaller spleens than mice vaccinated with conventional Env- and Gag-encoding vectors. On day 14 p.c., spleens of all but the groups of mice vaccinated with pIX-p15E expression/display or pIX-gp70 display only vectors were significantly smaller than those of unvaccinated control mice, and spleens of mice vaccinated with the pIX-gp70 expression/display vector were significantly smaller than those of mice vaccinated with conventional vectors.

FIG. 2.

Protection from FV-induced splenomegaly by vaccination. Mice were prime-boost immunized with Ad5-/Ad5F35-based vectors encoding the indicated antigens and challenged with FV 3 weeks after the boost vaccination. The progression to splenomegaly was monitored twice a week by palpation of the spleens (A and B). The graphs show the categorized spleen sizes of six mice per group on day 10 p.c. (A) and day 14 p.c. (B) (means + standard errors of the means). (C) On day 21, the spleens were removed and weighed. Statistically significant differences in spleen size (P < 0.05) compared to unvaccinated control mice or Env- plus Gag-vaccinated mice are indicated by * and #, respectively. Data are representative of two independent experiments.

On day 21 p.c., spleens were removed and weighed (Fig. 2C). All vaccinated mice had lower spleen weights than unvaccinated control mice. Mice vaccinated with pIX-gp70, pIX-gag, or pIX-gp70∼p15E expression/display vectors had significantly lower spleen weights than mice vaccinated with conventional vectors.

Vaccination with adenoviral expression/display vectors presenting gp70 reduces viral load after FV challenge.

Viremia of CB6F1 mice vaccinated with conventional or expression/display adenoviral vectors was analyzed 10 days p.c. (Fig. 3). All vaccinated mice had significantly lower viral loads in the plasma than unvaccinated control mice. For all mice vaccinated with vectors encoding a gp70-containing fusion protein, no viremia was detectable, which was a significant reduction compared to mice vaccinated with the conventional adenoviral vectors. Vaccination with the vector encoding pIX-p15E resulted in significantly higher viremia than vaccination with the conventional vectors, which demonstrated the importance of antigenic sites located in gp70 for protection from viremia. Vaccination with the display-only vectors also resulted in significantly higher viremia than vaccination with conventional vectors, which suggested that presentation of antigen on the virus surface is not sufficient for the full protective effect of adenoviral expression/display vectors. Vaccination with the vector encoding pIX-gag did not significantly affect viremia compared to mice vaccinated with conventional vectors.

FIG. 3.

FV viral load in plasma 10 days after FV challenge. Mice were vaccinated twice in a heterologous prime-boost immunization regimen with Ad5- and Ad5F35-based vectors encoding the indicated antigens and challenged with FV 3 weeks after boost immunization. Ten days p.c. FV viremia was analyzed in a focal infectivity assay. The graph shows viremia levels as FFU/ml of plasma, and horizontal lines mark the mean values. Statistically significant differences in viremia levels (P < 0.05) compared to unvaccinated control mice or Env- plus Gag-vaccinated mice are indicated by * and #, respectively. Data are representative of two independent experiments.

The viral load in the spleen on day 21 p.c. was significantly reduced by all but the vaccinations with pIX-p15E and the gp70 display-only vectors, with mean viral loads lowered by ∼10³- to 10⁵-fold (Fig. 4). Compared to vaccination with the conventional vectors, vaccination with gp70-containing expression/display vectors led to a reduction of the median viral load by ∼100-fold, but due to higher variations the differences did not reach statistical significance.

FIG. 4.

FV viral load in the spleen 21 days p.c. Mice were vaccinated twice in the heterologous prime-boost immunization regimen with Ad5- and Ad5F35-based adenoviral vectors encoding the indicated antigens, respectively, and challenged with FV 3 weeks after boost immunization. The viral load in the spleen was analyzed 21 days p.c. The graph shows the viral load as IC/spleen, and horizontal lines mark the median values. Statistically significant differences in viral load (P < 0.05) compared to unvaccinated control mice are indicated by *. Data are representative of two independent experiments.

Presentation of pIX-gp70 on the adenoviral capsid is necessary for the beneficial effect on vaccination.

The improved efficacy of the gp70-containing expression/display vector could be either due to the display of gp70 on the adenoviral particle or due to an improved immunogenicity of the pIX-gp70 fusion protein itself. To exclude the latter, we vaccinated mice with plasmid DNA encoding either the pIX-gp70 fusion protein or gp70 alone. C57BL/6 mice, which are not susceptible to FV-induced disease, were used for DNA immunization, as a lower vaccination efficiency was expected that would not lead to protection of the highly susceptible CB6F1 mice used for adenovirus-based vaccination. Mice immunized by either DNA vaccine had significantly lower viral loads in the spleen after challenge infection compared to unvaccinated mice (Fig. 5). However, no significant difference was found between the mice vaccinated with plasmids encoding pIX-gp70 or gp70, indicating that no intrinsic properties of the fusion protein, but its incorporation into the adenoviral capsid and presentation, were responsible for improved immune protection.

FIG. 5.

FV-infected spleen cells in DNA-immunized C57BL/6 mice 21 days p.c. C57BL/6 mice were immunized twice with plasmid DNA encoding gp70 or the pIX-gp70 fusion protein by intramuscular injection followed by electroporation. Three weeks after boost immunization mice were challenged with FV, and 21 days p.c. the viral load in the spleens was determined by quantitative RT-PCR. The graph shows viral load as the number of RNA copies/10⁷ spleen cells, and horizontal lines mark the median values. Statistically significant differences in viral load (P < 0.05) compared to unvaccinated control mice are indicated by *.

Presentation of gp70 on the adenoviral capsid improves induction of FV binding and neutralizing antibodies.

To determine the immune mechanisms underlying the improved protection after vaccination with the adenoviral expression/display vectors, we first analyzed the induction of binding and neutralizing antibodies after boost vaccination and after FV challenge (Fig. 6).

FIG. 6.

Humoral immune responses to FV vaccination with adenoviral vectors. Binding and neutralizing antibodies were measured 18 days after boost immunization and 10 days p.c. The upper graphs show binding (A) and neutralizing (B) antibodies after boost immunization, and the lower graphs show binding (C) and neutralizing (D) antibodies after FV challenge. The graphs show the reciprocal titers, and horizontal lines mark the mean values. Statistically significant differences in antibody titers (P < 0.05) compared to unvaccinated control mice or Env- plus Gag-vaccinated mice are indicated by * and #, respectively. Data are representative of two independent experiments.

After boost immunization all vaccinated mice had measurable titers of binding antibodies, and all mice vaccinated with expression/display vectors as well as the display-only vector had significantly higher binding antibody titers than mice vaccinated with conventional vectors (Fig. 6A). At this time point, only a few of the vaccinated animals had detectable neutralizing antibody titers, and there was no significant difference between the vaccination groups (Fig. 6B). Ten days after FV challenge infection, binding antibody titers of all vaccinated mice were significantly higher than those of unvaccinated control mice, but there was no longer a significant difference between the vaccination groups (Fig. 6C). At that time point, all vaccinated mice had measurable neutralizing antibody titers (Fig. 6D). Neutralizing antibody titers of mice vaccinated with conventional or expression/display vectors were significantly higher than those of unvaccinated control mice. Vaccination with the pIX-gp70 expression/display vector resulted in significantly higher neutralizing antibody titers than vaccination with conventional vectors, whereas vaccination with the pIX-p15E expression/display vector or the gp70 display-only vector resulted in significantly lower neutralizing antibody titers, suggesting that domains in gp70 are important for antibody induction and that antigen display on the capsid alone is not sufficient for efficient neutralizing antibody production.

Vaccination with adenoviral gp70 expression/display vectors results in improved FV Env-specific CD4⁺ T-cell responses.

The CD4⁺ T-cell responses to Env were analyzed by staining with MHC class II tetramers specific for an F-MuLV gp70 epitope 3 days p.c. (Fig. 7). At this early time point, unvaccinated mice do not exhibit a significant number of tetramer⁺ CD4⁺ T cells, indicating that tetramer⁺ CD4⁺ T cells found in vaccinated mice are vaccine primed. Vaccination with conventional adenoviral vectors resulted in a detectable but weak CD4⁺ T-cell response. Significantly higher CD4⁺ T-cell responses were observed in mice vaccinated with pIX-gp70 and pIX-gp70∼p15E expression/display vectors. For the pIX-gag expression/display vector, only a slight increase was observed. No significant CD4⁺ T-cell response was found in mice immunized with the gp70 display-only adenoviral vector, suggesting that also the Env protein expressed after infection by the vaccine vector contributes significantly to the induction of CD4⁺ T cells. For mice vaccinated with the pIX-p15E construct, no CD4⁺ T-cell response was detectable with MHC-II tetramers because the presented epitope is a part of gp70, which is not contained in this vaccine.

FIG. 7.

F-MuLV Env-specific CD4⁺ T-cell response after adenovirus-based vaccination. The Env-specific CD4⁺ T-cell response was analyzed in mice 3 days p.c. using MHC-II tetramers presenting the F-MuLV Env gp70-derived epitope EPLTSLTPRCNTAWNRLKL. The graph shows the percentages of MHC-II tetramer⁺ CD4⁺ T cells, and horizontal lines mark the mean values. Statistically significant differences in viremia levels (P < 0.05) compared to unvaccinated control mice or Env- plus Gag-vaccinated mice are indicated by * and #, respectively. Data are representative of two independent experiments.

Adenoviral expression/display vectors lead to an improved proliferative response of CD8⁺ T cells but not to enhanced activation.

Reagents for the FV model are limited, and the only MHC-I tetramer available is specific for an epitope located in the leader region of the Gag protein. Thus, we constructed an adenoviral expression/display vector presenting Ova and used the Ova model to analyze proliferation of CD8⁺ T cells in response to antigen presentation by dendritic cells and induction of CD8⁺ T cells.

C57BL/6 mice were immunized once with an adenoviral Ova expression/display vector, a conventional Ova-expressing adenoviral vector, or with Exo-ova. The exosomes display Ova that is anchored to the exosome membrane by fusion to the transmembrane domain of the G protein of vesicular stomatitis virus (39). Exo-ova served as an antigen display-only control. The amounts of Ova contained in Ad5.pIX-ova and Exo-ova preparations were analyzed by ELISA. For immunization, exosomes containing 200 ng Ova were administered, and 10⁸ infectious particles of Ad5.pIX-ova used for immunization included 60 pg of Ova.

Four days after vaccination mice were injected with CFSE-stained Ova-specific OT-I T cells, and 3 days later lymph nodes were recovered and proliferation levels of donor OT-I cells were analyzed. Vaccination with all vectors led to significant induction of OT-I cell proliferation (Fig. 8A). The expression/display vector led to significantly stronger proliferation of OT-I cells than the conventional Ova-encoding vector, indicating a higher level of epitope presentation to or stronger stimulation of the OT-I cells by antigen-presenting cells after vaccination with expression/display vectors than after vaccination with a conventional vector or with exosomes.

FIG. 8.

CD8⁺ T-cell induction in the Ova model. (A) To analyze the proliferation of OT-I cells in response to vaccination, mice were vaccinated once with an Ova-encoding or an Ova expression/display adenovirus or with Ova exosomes, and 4 days later CFSE-stained OT-I cells were adoptively transferred. CFSE staining intensity of OT-I cells reisolated from lymph nodes was determined 3 days later. Percentages of OT-I cells with diluted CFSE intensity due to proliferation are given. Horizontal lines represent the mean values. Statistically significant differences (P < 0.05) compared to unvaccinated control mice or Ad5.ova-vaccinated mice are indicated by * and #, respectively. Data are representative of two independent experiments. (B) Induction of polyfunctional CD8⁺ T cells by vaccination was analyzed by intracellular staining. Mice were vaccinated once, and 2 weeks later spleen cells were stimulated in vitro with OT-I peptide before staining for expression of IFN-γ, IL-2, and TNF-α. The graph shows the percentages of IFN-γ⁺ IL-2⁺ TNF-α⁺ CD8⁺ T cells, and horizontal lines represent the mean values. *, statistically significant difference compared to unvaccinated control mice (P < 0.05). Data are representative of two independent experiments.

To determine whether vaccination with expression/display vectors would also result in improved induction of polyfunctional CD8⁺ T cells after vaccination, an intracellular cytokine staining of CD8⁺ T cells was performed 2 weeks after a single vaccination. We found significant cytokine production by CD8⁺ T cells in all vaccinated mice (Fig. 8B), but contrary to the OT-I cell proliferation experiment, here both adenoviral vector groups had significantly higher cytokine expression levels than the exosome group but did not show a significant difference between one another.

DISCUSSION

In the FV model we evaluated a novel vaccination strategy that combines genetic and protein vaccination by using adenoviral vectors not only as a gene transfer tool but also for display of entire vaccine antigens on the adenovirus capsid by fusion to capsid protein IX. Vaccination with vectors that encode the F-MuLV protein gp70 and display it on the adenoviral capsid led to significantly improved protection in comparison to vaccination with conventional, antigen-expressing adenoviral vectors that corresponded with increased induction of neutralizing antibodies and gp70-specific CD4⁺ T-cell responses.

To determine the importance of in vivo expression of the antigen displayed on the capsid for the beneficial effect on vaccination, we evaluated an adenoviral vector that displayed gp70 on the capsid without encoding it. Vaccination with this vector resulted in diminished control of viral load and disease development compared to vaccination with conventional, antigen-expressing vectors, no appreciable induction of gp70-specific CD4⁺ T cells, and lower neutralizing antibody titers. Furthermore, vaccination with plasmid DNA encoding the fusion protein pIX-gp70 did not lead to better protection compared to vaccination with plasmid DNA encoding gp70 alone. This suggests that the fusion protein pIX-gp70 itself was not responsible for the enhanced vaccine protection observed in the adenovirus-based vaccination. These two findings indicate that for the beneficial effect observed after adenoviral expression/display vector vaccination the antigen has to be both incorporated into the adenoviral capsid and expressed in vivo.

We found both improved neutralizing antibodies and higher CD4⁺ T-cell responses after expression/display vector vaccination compared to vaccination with conventional, antigen-encoding vectors. We hypothesized at the outset that the array-like structure of gp70 on the capsid should result in enhanced antibody responses (13). The hepatitis B virus surface protein and the major capsid protein L1 of human papilloma virus that self-assemble into virus-like particles (10, 15) are prominent examples for efficient protein vaccines where antigens are presented in such ordered structures. Data from vaccination with the gp70 display-only vector shed light on the mechanism of antibody induction by the adenoviral expression/display vectors. The display of gp70 on the display-only particles was sufficient to induce very high binding antibody titers that were comparable to the titers induced by the expression/display vectors. Neutralizing antibody titers on the other hand were markedly lower than after vaccination with both expression/display and conventional vectors. Correlating with the neutralizing antibody titers were the Env-specific CD4⁺ T-cell responses, for which we also found no appreciable induction by the display-only vector. These findings suggest that the ordered display on the capsid might lead to cross-linking of B-cell receptors and induction of Env-binding antibodies even in the absence of detectable T helper cell responses. However, for the induction of neutralizing antibodies, the display on the particle alone was not sufficient, suggesting that the CD4⁺ T helper cells induced by in vivo expression of the antigen by the expression/display vector promote the affinity maturation and the development of a neutralizing antibody response.

Using the Ova model we found stronger proliferation of Ova-specific OT-I cells in response to vaccination with the adenoviral expression/display vector than from a conventional vector, but this did not translate into a significantly stronger induction of cytokine-secreting CD8⁺ T cells after vaccination. This is consistent with our efficacy study in the FV model. Since Gag is hardly accessible to antibodies, an improved protection after vaccination with pIX-gag would have had to be mediated by an improved CD8⁺ T-cell response. As we found no significant difference in protection after vaccination with a pIX-gag expression/display vector compared to vaccination with conventional vectors, the findings from the FV vaccination gave no indication of an enhanced CD8⁺ T-cell response. Vaccination with expression/display vectors might still have a secondary effect on CD8⁺ T-cell responses through the induction of stronger CD4⁺ T-cell responses, providing improved help for CD8⁺ T cells.

Ova-containing exosomes induced a higher degree of proliferation of OT-I T cells than a conventional adenoviral vector but to a lesser extent than adenoviral expression/display vectors, although the expression/display vectors carried a far smaller amount of Ova protein. This indicates that antigen displayed on the adenovirus capsid is far more efficiently processed and presented than when applied as a protein vaccine, a finding that might be due to the capacity of adenovirus for endosomal escape, which is an important step in the adenovirus infection process (reviewed in reference 27). Vaccination with exosomes led only to marginal induction of CD8⁺ T cells that produced cytokines after stimulation with OT-I peptide. Thus, for the induction of CD8⁺ T cells, as seen in the FV model for the induction of CD4⁺ T cells and neutralizing antibodies, in vivo expression seems to play an important role.

For the presentation of the vaccine antigen we chose the pIX protein, because in contrast to hexon, fiber, and penton base, which have been used before for the display of single epitopes (5, 22, 25, 26, 43), pIX allows the incorporation of large proteins without loss of function and virion integrity (24, 28). Incorporation of a whole protein is an interesting option, as the presence of multiple relevant epitopes in the vaccine allows the induction of a broader and more diverse immune response. For HIV vaccination there have been attempts at using rhinovirus (35, 36, 44) and poliovirus (6, 11) for the display of HIV gp41- or gp120-derived epitopes, but these epitope-based approaches are more restrictive than our expression/display vector system. Furthermore, B-cell epitopes are more likely to be in the correct conformation when incorporated in the original protein rather than in some scaffold protein, a problem mentioned for incorporation of a poliovirus epitope into hexon protein (5). Another possible advantage of pIX for the display of Env proteins is its arrangement into trimers when incorporated into the capsid (12). Also, the vaccine antigens fused to pIX molecules should be present in close proximity to one another and might be able to form trimers on the adenoviral particle, possibly allowing for the induction of conformation-dependent antibodies.

Previously, it was shown that the display of small protein tags fused to pIX could be improved when a 70-Å spacer was inserted between pIX and the tag (41). Here, incorporation of a 70-Å spacer in the fusion protein did not change protection levels or induction of antibodies and CD4⁺ T-cell responses compared to the pIX-gp70 fusion protein (data not shown).

The observation that the addition of the ectodomain of the transmembrane Env protein p15E to the fusion protein did not further enhance immunization efficiency and that vaccination with p15E as the sole Env component in the vaccine gravely diminished protection corresponds with earlier observations in the FV model where no immunologically significant epitopes were found in p15E (16).

The only vaccine that has been shown to confer complete protection from FV disease in highly FV-susceptible mice so far is attenuated F-MuLV (7, 32), with which survival and absence of infected cells after FV challenge were demonstrated. Vaccination of highly susceptible CB6F1 mice with a peptide containing a CD4⁺ T-cell epitope emulsified in complete Freund's adjuvant resulted in ∼80% survival of mice after FV challenge, but not sterile immunity (19, 21). Protein vaccination with gp70 in complete Freund's adjuvant induced protection of highly susceptible mice from FV-induced disease, but those study authors did not demonstrate an absence of infected spleen cells (18). Vaccination with recombinant vaccinia viruses expressing F-MuLV Env conferred a high level of protection, but mice transiently showed signs of disease (16, 30). After vaccination with gp70 protein, those authors reported low levels of binding antibodies (1/20) 21 days after immunization, and neutralizing antibodies were reported to have been detected 41 days p.c. For the vaccination with recombinant vaccinia virus, no antibodies were shown before FV challenge.

In our current experiments, vaccination with gp70 expression/display vectors, while not resulting in sterile immunity, conferred strong protection from FV-induced disease, where mice were efficiently protected from splenomegaly and had no detectable FV viral loads in plasma. In contrast to other vaccination approaches reported before, we could demonstrate very high binding antibody titers and in some mice also neutralizing antibodies before challenge with FV. As the initial protection from a retroviral infection is of paramount importance to prevent integration of viral DNA into host cells and establishment of latency, these findings suggest that expression/display vectors are very promising tools for antiretroviral vaccination.