|Home | About | Journals | Submit | Contact Us | Français|
The use of adenoviruses (Ad) as vaccine vectors against a variety of pathogens has demonstrated their capacity to elicit strong antibody and cell-mediated immune responses. Adenovirus serotype C vectors, such as Ad serotype 5 (Ad5), expressing Ebolavirus (EBOV) glycoprotein (GP), protect completely after a single inoculation at a dose of 1010 viral particles. However, the clinical application of a vaccine based on Ad5 vectors may be hampered, since impairment of Ad5 vaccine efficacy has been demonstrated for humans and nonhuman primates with high levels of preexisting immunity to the vector. Ad26 and Ad35 segregate genetically from Ad5 and exhibit lower seroprevalence in humans, making them attractive vaccine vector alternatives. In the series of studies presented, we show that Ad26 and Ad35 vectors generate robust antigen-specific cell-mediated and humoral immune responses against EBOV GP and that Ad5 immune status does not affect the generation of GP-specific immune responses by these vaccines. We demonstrate partial protection against EBOV by a single-shot Ad26 vaccine and complete protection when this vaccine is boosted by Ad35 1 month later. Increases in efficacy are paralleled by substantial increases in T- and B-cell responses to EBOV GP. These results suggest that Ad26 and Ad35 vectors warrant further development as candidate vaccines for EBOV.
Replication-defective adenovirus (rAd) vectors are powerful inducers of cellular immune responses and have therefore come to serve as useful vectors for gene-based vaccines, particularly for lentiviruses and filoviruses, as well as other nonviral pathogens (14, 34, 39, 40, 43, 44, 46). Adenovirus-based vaccines have several advantages as human vaccines—they can be produced to high titers under good manufacturing practice (GMP) conditions and have proven to be safe and immunogenic in humans (2, 6, 12, 16, 18). While most of the initial vaccine work was conducted using rAd serotype 5 (rAd5) due to its significant potency in eliciting broad antibody and CD8+ T-cell responses, preexisting immunity to rAd5 in humans may limit efficacy (5–7, 29). This property might restrict the use of rAd5 vectors in clinical applications for many vaccines that are currently in development, including those for Ebolavirus (EBOV) and Marburg virus (MARV).
To circumvent the issue of preexisting immunity to rAd5, several alternative vectors are currently under investigation. These include adenoviral vectors derived from rare human serotypes and vectors derived from other animals, such as chimpanzees (1, 39, 49). Research on the use of animal-derived adenoviral vectors is relatively nascent, while human adenoviruses possess the advantages of having well-characterized biology and tropism on human cells, as well as documented manufacturability (48). Immunogenicity of these vectors and their potential as vaccines has been demonstrated with animal models, primarily as prime-boost combinations with heterologous vectors (1, 41).
Adenovirus seroprevalence frequencies are cohort dependent (28), but among the large group of 51 human adenoviruses tested, Ad35 and Ad11 were the most rarely neutralized by sera from six geographic locations (49). rAd35 vector vaccines have been shown to be immunogenic in mice, nonhuman primates (NHPs), and humans and are able to circumvent Ad5 immunity (4, 30, 31, 36, 47). rAd35 vectors grow to high titers in cell lines suitable for production of clinical-grade vaccines (13) and have been formulated for injection as well as stable inhalable powder (15). These vectors show efficient transduction of human dendritic cells (8, 26) and thus have the capability to mediate high-level antigen delivery and presentation.
Prime-boost regimens based on vectors derived from closely related adenovirus serotypes, such as Ad11 and Ad35, both from subgroup B, are less immunogenic than combinations of more genetically and immunologically distinct adenoviral vectors, most probably as a result of low levels of cross-reactive neutralizing antibodies (NAbs) elicited by Ad35 and Ad11(22, 47). Therefore, Ad26, from subgroup D, was the second vector selected for its ability to circumvent Ad5 preexisting immunity. Although Ad26 seroprevalence can be significant in certain adult populations, Ad26 neutralizing antibody titers remain markedly lower than Ad5 neutralizing antibody titers (1, 28). Studies have shown that rAd26 vectors can be grown to high titers in Ad5 early region 1 (E1)-complementing cell lines suitable for manufacturing these vectors on a large scale and at clinical grade (1), and this vector has been shown to induce humoral and cell-mediated immune responses in prime-boost vaccine strategies (1, 25).
In this paper, we report the immunogenicity of rAd35 and rAd26 vectors upon single inoculation as well as heterologous prime-boost combination. There are distinct advantages associated with either single-shot or prime-boost immunization depending on the need for immediate versus long-term immunity, and these must be taken into account when optimizing immunization regimens. EBOV and other filovirus outbreaks tend to occur suddenly and spread quickly among populations in which medical facilities are scarce. Thus, under these circumstances, short vaccine regimens may be desirable. For this reason, single-shot vaccinations with rAd5 vectors containing EBOV glycoprotein (GP) and nucleoprotein (NP) genes have been developed for nonhuman primates (43). Such vaccines have been shown to elicit strong immune responses within 1 month (44), probably due to high expression levels of the inserts and the tropism of Ad5 for dendritic cells. On the other hand, long-term protective immunity will likely require a prime-boost vaccine regimen comprising two or more administrations that can induce durable T-cell memory. Therefore, we designed a series of experiments to test immunogenicity and potency for both single inoculation and a prime-boost combination using rAd35 and rAd26 vectors, and the results of these studies are presented herein.
Low seroprevalent E1/E3-deleted rAd26 and rAd35 vaccine vectors expressing EBOV GPs were constructed, amplified, and purified as described previously (1). An Ad5 vector, not expressing EBOV GP, was constructed, amplified, and purified by the same method and used to induce immunity to Ad5 in selected animals. EBOV GP inserts spanning the open reading frames of Zaire (ZEBOV) and Sudan/Gulu species [GP(Z) and GP(S/G), respectively] were cloned under transcriptional control of the human cytomegalovirus (CMV) promoter and the simian virus 40 (SV40) polyadenylation sequence into a plasmid containing the left portion of the Ad genome, including left inverted terminal repeat (ITR) and packaging signal. Cotransfection of this plasmid with a cosmid containing the remaining (E3-deleted) Ad sequence to PER.C6 cells yielded an E1/E3-deleted replication-deficient recombinant Ad26 or Ad35 vaccine vector. To facilitate replication of rAd26 and rAd35 vectors on PER.C6 cells, the native E4orf6 regions were replaced by the Ad5 E4orf6 sequence (13). The rAd viruses were plaque purified, and one plaque of each was expanded up to a production scale of approximately 2.4 liters. A two-step cesium chloride gradient ultracentrifugation procedure was used to purify rAd vectors. The purified rAd vector EBOV vaccines were stored as single-use aliquots at below −65°C. Viral particle titers were determined by measuring the optical density at 260 nm (27). Infectivity was assessed by measuring the 50% tissue culture infective dose (TCID50) using 911 cells (9). Adenovirus-mediated EBOV GP expression was assessed by infection of A549 cells followed by analysis of culture lysates by Western blotting. The identity of the purified vectors was confirmed by PCR, and the complete transgene regions, including flanking sequences, were checked using DNA sequencing.
The phylogenetic tree was constructed using full-length adenovirus hexon amino acid sequences. Amino acid sequences were aligned using the Clustal X program (19), and the tree was built using the Clustal X neighbor-joining method and bootstrapped 1,000 times. The tree was visualized and plotted using the Drawtree program from the Phylip Phylogeny Inference package, version 3.68 (http://evolution.genetics.washington.edu/phylip.html).
Animal experiments were conducted in full compliance with all relevant federal guidelines and NIH policies. Cynomolgus macaques (Macaca fascicularis) aged 3 to 5 years and weighing between 2 and 3 kg were obtained from Covance for all studies. Monkeys were housed individually and given enrichment regularly as recommended by the Guide for the Care and Use of Laboratory Animals (30a). Animals were anesthetized with ketamine prior to blood sampling or vaccination. Each vaccine group in this study contained three or four cynomolgus macaques, and each control group contained a single cynomolgus macaque. Four weeks after EBOV vaccination, animals were transferred to the Maximum Containment Laboratory (biosafety level 4 [BSL-4]) for infection with a target dose of 1,000 PFU of ZEBOV delivered by the intramuscular (i.m.) route into the caudal thigh. The ZEBOV challenge stock was prepared from a human fatality in the 1995 outbreak in the former Zaire. In order to minimize the use of animals in infectious challenge experiments, we rely on historical unvaccinated controls, and the number of controls has been reduced to one unvaccinated NHP per experiment. All historical unvaccinated cynomolgus macaques that received the same exposure dose, strain, and route (i.m.) of ZEBOV as in these studies have succumbed to lethal infection (n > 23 macaques). A group size of 4 vaccinated subjects allows >80% power at the 95% confidence level to detect a difference in survival rates, assuming 4 of 4 vaccinated animals survive. Animals remained in BSL-4 containment through completion of the study. While in the BSL-4 facility, the monkeys were fed and checked a minimum of twice daily.
Animal studies performed under BSL-4 biocontainment conditions at the U.S. Army Medical Research Institute for Infectious Diseases (USAMRIID) were approved by the USAMRIID Institutional Animal Care and Use Committee. Animal research was conducted in compliance with the Animal Welfare Act (47a) and other federal statutes and regulations relating to animals and experiments involving animals and adheres to the principles stated in the Guide for the Care and Use of Laboratory Animals (30a). The facilities used are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International.
Subjects received intramuscular vaccinations in the bilateral deltoids by needle and syringe with the doses and vectors indicated in figure legends and the text. Selected animals, as indicated for each experiment, were preimmunized with 1011 particle units (PU), equivalent to virus particles (vp), of an empty Ad5 vector in a single bilateral deltoid inoculation to induce Ad5 immunity. Anti-Ad5 neutralizing antibody titers were measured as described below to confirm Ad5-immune status in macaques prior to EBOV vaccination.
Ad5-specific neutralizing antibody titers were assessed by luciferase-based virus neutralization assays as described previously (42). Briefly, serum is heat inactivated and serially diluted 2-fold (starting dilution is 1:16). Ad5.Luc solution (1 × 108 vp/ml) is added to each well at 500 virus particles per cell. A549 cells are added at 1 × 104 cells/well, and plates are incubated at 37°C and 10% CO2 for 24 to 26 h. After incubation, the medium is discarded, phosphate-buffered saline (PBS) is added, and plates are stored frozen overnight. Plates are allowed to thaw at room temperature (RT), Luciferase Steady-Lite substrate is added, and the lysate is transferred into black-and-white isoplates. Luminescence counts are recorded using a 1450 MicroBeta Trilux counter. Titers were determined by validated ANAM (adenovirus neutralization assay macro) software.
Antibody assays used Zaire GP as the target antigen in order to match the challenge virus species. Polyvinyl chloride enzyme-linked immunosorbent assay (ELISA) plates (Dynatech, Vienna, VA, or Nunc, Rochester, NY) were coated with 100 μl of antigen per well and incubated at 4°C overnight. Subsequent incubations were performed at room temperature. Transmembrane-deleted EBOV GP (EBOV GPΔTM) generated by calcium phosphate-mediated transient transfection of 293T cells served as the antigen. Plates were washed six times with PBS containing Tween 20 after antigen coating. Test sera were serially diluted to seven concentrations ranging from 1:50 to 1:50,000 and added to the antigen-coated wells for 60 min. The plates were washed six times followed by incubation with the detection antibody, goat anti-human IgG (H+L; Chemicon/Millipore, Billerica, MA) conjugated to horseradish peroxide. SigmaFast o-phenylenediamine dihydrochloride (Sigma, St. Louis, MO) substrate was added to the wells, and the optical density was determined (450 nm). A prevaccination serum sample for each animal was run every time the assay was performed. A positive-control serum sample from a single animal with a known ZEBOV GP IgG response was run every time the assay was performed. Background-subtracted ELISA titers are expressed as reciprocal optical density values, representing the dilution at which there is a 90% decrease in antigen binding (EC90).
T-cell assays used Zaire GP as the target antigen in order to match the challenge virus species. Whole-blood samples from cynomolgus macaques were subjected to density gradient centrifugation over Ficoll to isolate peripheral blood mononuclear cells (PBMC). Approximately 1 × 106 cells were stimulated in 100 μl RPMI medium containing 10% heat inactivated fetal calf serum for 6 h at 37°C with anti-CD28 (clone CD28.2) and -CD49d (clone L25) antibodies (BD Biosciences), Brefeldin A (Sigma-Aldrich, St. Louis, MO), and either dimethyl sulfoxide (DMSO) or a pool of peptides spanning the entire ZEBOV GP open reading frame. The peptides were 15-mers overlapping by 11 amino acids reconstituted in fresh sterile DMSO at a final concentration of 2.5 μg/ml for each peptide. For each sample, an equivalent aliquot was stimulated with staphylococcal enterotoxin B (SEB) as a positive control. After the 6-h stimulation, PBMC were stained with a mixture of antibodies against lineage markers CD3-Cy7-allophycocyanin (APC), clone SP34-2 (BD Biosciences), CD4-QD605, clone M-T477 (BD Biosciences), CD8-Texas Red phycoerythrin (TRPE), clone RPA-T8, CD95 Cy5-phycoerythrin (PE), clone DX2 (BD Biosciences), and CD45RA QD655, clone 5H3, at room temperature for 20 min. The CD45RA QD655 and CD8-TRPE antibodies were conjugated according to standardized protocols as previously described (18). After two washes, the cells were fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences) followed by staining with antibodies against cytokines tumor necrosis factor alpha (TNF-α)-APC, clone monoclonal antibody 11 (MAb11; BD Biosciences), and interleukin-2 (IL-2) PE, clone MQ17H12 (BD Biosciences). These cytokines identify Ebolavirus GP-specific CD8+ and CD4+ (respectively) T-cell responses in vaccinated macaques (44, 45). The viability dye ViViD (Invitrogen) was included to allow discrimination between live and dead cells (33). Samples were acquired with an LSR II cytometer (BD Biosciences), collecting up to 1,000,000 events, and analyzed using FlowJo 9.1 software. Cytokine-positive cells were defined as a percentage within CD4+ and CD8+ T-cell memory subsets. Memory subsets were defined as CD45RA±/CD95+ or CD28±/CD95+.
For challenge studies, blood was collected from the NHPs on days 0, 3, 6, 10, 14, and 28 after ZEBOV infection. Serum samples were tested for concentrations of aspartate aminotransferase (AST) by using a Piccolo point-of-care blood analyzer (Abaxis, Sunnyvale, CA).
Comparison of anti-GP ELISA IgG titers and intracellular cytokine production by T-cell memory subsets was done using a two-tailed t test. Kaplan-Meier survival proportion comparisons were conducted using the Mantel-Cox log-rank test (GraphPad Prism).
rAd5 genetic vaccines for EBOV provide potent protective immunity in macaques and have been proven safe and immunogenic in human clinical trials (2, 12, 16, 21). Studies of macaques and humans have shown that preexisting vector-directed immunity can limit the potency of viral vector-based vaccines (5, 29). However, the effect of preexisting Ad5 immunity on rAd5-based vaccine efficacy may not be universal, since the protection threshold for vaccine-induced immune responses is pathogen dependent. Therefore, we tested whether an rAd5-GP vaccine, demonstrated previously to provide uniform protection against lethal EBOV challenge in macaques, was protective in Ad5-immune subjects. Ad5 immunity was generated in a group of three cynomolgus macaques (P1, P2, P3) by vaccination with an rAd5 vector lacking any antigen insert as described in Materials and Methods. Anti-Ad5-neutralizing antibody titers measured before vaccination with rAd5-GP (Fig. 1A) fell within the median range of titers observed for humans (17). Subjects were vaccinated with 1010 particles of rAd5-GP, and 4 weeks later, the vaccinated animals and one unvaccinated control subject were exposed to 1,000 PFU of Zaire Ebolavirus (ZEBOV) by intramuscular injection. Hepatic enzymes were measured regularly after infectious challenge, since elevations in these markers are characteristic of productive EBOV infection in macaques (Fig. 1B). In all vaccinated subjects, circulating levels of aspartate transaminase (AST) demonstrated elevations that were equal to or greater than levels in the control, unvaccinated animal, and all subjects in the study succumbed to the lethal effects of infection by day 10 postchallenge (Fig. 1C). These data demonstrated that rAd5-GP vaccine efficacy is reduced in Ad5-immune subjects.
Since seroprevalence data suggest that a large proportion of humans worldwide have experienced natural infection with Ad5, and the rAd5-GP vaccine was not protective in Ad5-immune subjects, we evaluated other adenovirus serotypes for use as vaccine vectors (Fig. 2A). Ad35, a group B adenovirus, and Ad26, a group D adenovirus, segregate genetically from rAd5, in group C, so we hypothesized that vaccine vectors derived from these serotypes would be less sensitive to Ad5 immunity in primates. Although Ad35 and Ad26 vectors use receptors distinct from Ad5 usage, they nonetheless demonstrate efficient transduction of monocyte-derived dendritic cells, and circumvent Ad5 immunity in mice (26, 49). Therefore, GP inserts from the Zaire or Sudan/Gulu species of EBOV were cloned under transcriptional control of the human CMV promoter into the E1 region of rAd35 and rAd26 vectors (Fig. 2B). Both vector genomes have the E1 genes deleted in order to render them replication deficient and reduce the potential for recombination in vaccinated subjects.
Initial studies were conducted with a single EBOV species vaccine encoding the GP from ZEBOV, GP(Z), to test the ability of rAd35 vectors to induce immune responses in Ad5-naïve macaques and also to evaluate vector potency within the context of preexisting immunity to Ad5. Six cynomolgus macaques, three of which were Ad5 naïve and three Ad5 immune, were each vaccinated intramuscularly with 1010 particles of rAd35-GP(Z) by needle injection. At 3 weeks postvaccination, antigen-specific antibody and T-cell responses were evaluated using peripheral blood samples obtained from individual subjects. Antibodies against EBOV-GP(Z) assessed by ELISA were induced in all subjects, demonstrating that rAd35-GP(Z) vectors mediated successful in vivo transduction of target cells and efficient antigen presentation (Fig. 3A). Ad5-naïve and Ad5-immune subjects generated serum antibody titers ranging from approximately 1:700 to 1:3,000. These antibody levels are in the range that has been observed for Ad5-based vaccines containing GP(Z) inserts and exceeded the minimum value (1:500) that has been associated with immune protection against EBOV infection in this macaque Ad vaccine model (45). Although significant antibody titers were induced in all vaccinees, none of the subjects exceeded the reported protection threshold titer for Ad5-GP vaccine vectors in macaques. It is noteworthy, however, that a comparison of antibody titers for Ad5-naïve versus Ad5-immune subjects showed that there was no significant difference in average titers elicited among these groups (1:1,600 versus 1:1,800, respectively), suggesting that rAd35 vectors are effective vaccines in subjects who have been exposed previously to Ad5.
Cellular immune responses were evaluated by intracellular cytokine staining (ICS) for either TNF-α (CD8+) or IL-2 (CD4+) after stimulation of the subject PBMC with overlapping peptides spanning the EBOV GP(Z) open reading frame. Surface staining of lymphocytes using CD45RA and CD95 was performed to assess antigen-specific immune responses in the memory subpopulations of CD4+ and CD8+ T cells (Fig. 3B and C). As observed for antibody responses, macaques vaccinated with the rAd35 vector generated cellular immunity against EBOV-GP, and the frequency of antigen-specific T cells was not affected by Ad5 immune status. The rank order magnitude of cellular responses in both CD4+ and CD8+ lymphocytes across subjects was similar to that for antibody responses, although the frequency of antigen-specific T cells for one subject, V3, was below detectable levels. Previous studies of macaques have shown that rAd5-based vaccine vectors induce CD8+ T-cell frequencies that are dominant over CD4+ responses. In the present study, rAd35-vaccinated subjects generated GP-specific CD4+ and CD8+ lymphocytes at similar frequencies. However, given the relatively small number of subjects tested, it is possible that differences, if present, could not be revealed. Altogether, these results demonstrate that rAd35-GP is immunogenic in cynomolgus macaques and that vector potency for the induction of antigen-specific humoral and cell-mediated immune responses is not reduced in subjects with preexisting immunity to Ad5.
We next tested whether vaccination with rAd35-GP provided protection against infectious challenge with a high dose of ZEBOV. One week after the assessment of immune responses, the six rAd35-GP-vaccinated macaques and one naïve cynomolgus macaque were exposed to 1,000 PFU of the 1995 Kikwit strain of ZEBOV by intramuscular injection. Circulating levels of AST were evaluated every 3 to 4 days during the acute infection period, through day 10 to day 14 (Fig. 3D), and then on the last day of the 28-day follow-up period (not shown). Plasma AST remained at baseline levels through day 3 after infection in all subjects, indicating normal liver function immediately following infectious EBOV challenge. By day 6 after EBOV exposure, the unvaccinated control subject displayed a 10-fold increase in enzyme levels, indicating active infection in this subject. Blood samples from two subjects in the Ad5-naïve/rAd35-vaccinated group (V1, V3) also exhibited dramatic increases in AST, while the third subject in this group (V2) showed only a marginal increase at a single time point prior to resolution to baseline levels. Similarly, two of three subjects in the Ad5-immune/rAd35-vaccinated group (V4, V5) displayed elevations in AST, though these levels were much lower than that for the unvaccinated control, while results for one subject (V6) remained at normal levels for this parameter of infection. Overall, AST levels were higher in Ad5-naïve than in the Ad5-immune vaccinated subjects. It is noteworthy that the subjects remaining at normal levels for this clinical observation displayed the highest prechallenge, antigen-specific CD8+ and antibody responses within their respective vaccine groups. Plasma viremia levels confirmed EBOV infection in all animals which displayed elevated AST levels (not shown).
The results of this experiment showed that rAd35 is immunogenic when administered at a dose of 1010 particles per subject. The vaccine generated GP-specific immune responses, but this dose and regimen were suboptimal for uniform protection of all subjects (Table 1). Within vaccine groups, protection was observed for those subjects displaying the highest magnitude of antigen-specific antibody and CD8+ T-cell responses.
rAd-based vectors are commonly administered to macaques in doses ranging from 1010 to 1012 particles. Previous results with rAd5-GP have demonstrated 1010 virus particles as the minimal dose to achieve 100% protection of cynomolgus macaques against EBOV infection (43). Since the studies described above were performed at the lower end of this dose range and did not result in uniform protection, it is possible that even marginally lower in vivo antigen expression achieved with the rAd35 vector compared to that of rAd5 vectors could result in suboptimal immune responses. Therefore, we asked whether administration of a higher vaccine dose could elicit a greater degree of immune protection. To minimize the use of nonhuman primates, we proceeded in a stepwise fashion by increasing the vaccine dose by 1 log. The vaccine in this experiment also included rAd35 expressing GP from the Sudan EBOV species [GP(S/G)] in order to compare efficacy to historical data for rAd5 vaccines that comprised GP from both Sudan and Zaire species (43, 45) and because the addition of a second vaccine insert could theoretically enhance or dilute protective efficacy against Zaire Ebolavirus challenge. In the present study, cynomolgus macaques (n = 3 per group) were vaccinated with 1010 or 1011 virus particles (each) of rAd35-GP(Z) and rAd35-GP(S/G), and immune responses were measured 3 weeks after vaccination as in the previous experiment.
GP-specific antibodies against the challenge virus species, Zaire, were generated in all subjects (Fig. 4A), an expected result since the vaccine dose was equal to or higher than that in the previous experiment. However, the titers in two subjects vaccinated with 1010 particles in this experiment (V8, V9) were below the cutoff protection titer in Ad5-GP-vaccinated subjects (45). The maximum antibody titer observed was 1:2,900 (subject V10), which is the same as the maximum titers observed for the rAd35 subjects shown in Fig. 3. Average antibody titers were higher in the 1011 dose group (1:1,500 versus 1:700) and statistically significant at a P value of 0.02. As observed in the lower-dose group, there was one subject whose titer was near the rAd5-GP vaccine cutoff for immune protection (subject V11).
CD4+ and CD8+ T-cell responses were present in both dose groups at 3 weeks postvaccination (Fig. 4B and C). There was no clear dose response in either CD4+ or CD8+ T-cell responses. Since the kinetics of cellular immune responses varies between subjects, especially in outbred animals, and the response measurement is not cumulative over time as it is with antibody levels, group trends are sometimes difficult to capture at a single time point. Within each animal, antigen-specific T-cell frequencies were higher for CD4+ than for CD8+ cells, but when combined with the results from the first experiment, there was no trend toward either CD4+ or CD8+ cell dominance induced by rAd35-GP vectors in these experiments.
One week following the assessment of immune responses, all six vaccinated macaques and one unvaccinated subject were exposed to 1,000 PFU ZEBOV by intramuscular injection and observed for signs of productive infection. Hemorrhagic manifestations of EBOV infection routinely result in the appearance of a maculopapular rash on the face and extremities of infected macaques; subjects also typically reduce food intake and become dehydrated. The earliest appearance of symptoms for the two unvaccinated subjects in both rAd35 studies occurred on day 6 after EBOV exposure; each displayed a full constellation of symptoms by day 7 (data not shown). Consistent with the manifestation of clinical symptoms, the unvaccinated subject demonstrated dramatic elevations in the liver enzyme AST (Fig. 4D and E). Likewise, most vaccinated subjects also displayed increases in hepatic enzymes. Table 1 shows the infectious challenge outcomes and the day of death or euthanasia in extremis of nonsurvivors for all rAd35-vaccinated subjects described thus far. The unvaccinated subjects succumbed to the lethal effects of infection on days 9 and 8 (experiments 1 and 2, respectively), and unprotected vaccinated subjects died between days 6 and 10.
Altogether, the studies using rAd35 as a vaccine vector with GP(Z) alone or in combination with GP(S/G) showed that antigen delivery and presentation was sufficient to generate antigen-specific immune responses, but at levels below what is required for absolute immune protection. In some animals, reduced protective immunity was associated with antibody levels that are consistent with what has been shown using rAd5 vectors to be marginal for protection. Higher doses of the rAd35 vaccine could, theoretically, be protective. However, due to the absence of a clear dose benefit from 1010 to 1011 here, and in pilot mouse immunogenicity studies (not shown), we decided instead to evaluate alternative vectors.
We next evaluated a recombinant Ad26-based vaccine, a group D adenovirus, for its ability to generate protective immunity against EBOV infection. This serotype uses the same cellular receptor (CD46) as Ad35 but has been shown to generate slightly higher immune responses when used as a priming vaccine vector (24, 25). For these studies, dose escalation was conducted over a range of three orders of magnitude in two separate infectious challenge experiments. In the first study, we tested the vaccine at doses of 1010 or 1011 particles for each vector, rAd26-GP(Z) and rAd26-GP(S/G), and in the second study, we used a dose of 1012 particles each. The first study tested the vaccine in Ad5-immune cynomolgus macaques in order to evaluate whether rAd26, like rAd35, could elicit antigen-specific immune responses in the presence of preexisting immunity to Ad5. Four Ad5-immune cynomolgus macaques per group were vaccinated by intramuscular injection, and blood samples were obtained 3 weeks later to assess circulating humoral and cellular immune responses against EBOV GP (Fig. 5A). The average circulating anti-GP antibody titers showed a dose-response trend across dose groups, 1:700 for subjects receiving 1010 particles and 1:4,500 for subjects receiving 1011 particles, though this difference did not reach statistical significance with the available group sizes (P = 0.056). The average titer for three of four subjects in the 1010 dose group was just above the minimum threshold for immune protection in rAd5-vaccinated subjects (1:500), but subject V16 generated only a marginal antibody response, 1:100. In contrast, subject V19 was vaccinated with 1011 particles of the rAd26 vaccine and generated a very high antibody titer, 1:10,500, exceeding nearly 3-fold the absolute protection threshold, while the others in this vaccine group displayed intermediate titers that do not definitively predict survival outcome. In study 2, four subjects received 1012 particles of each rAd26 vector and generated antibody responses very similar to those in the 1011 dose group, with the majority of subjects' titers between 1:1,000 and 1:4,000. The average anti-GP antibody titer for this group was 1:3,000.
T-cell immune responses (Fig. 5B and C) were measured by ICS as in the rAd35 studies and also trended toward a dose response in study 1, but the difference between 1010 and 1011 dose groups was not significant (P values of 0.12 and 0.26 for CD4+ and CD8+, respectively). Average antigen-specific CD4+ T-cell frequencies were 0.14% for vaccinees receiving 1010 particles of rAd26-GP versus 0.24% at the higher vaccine dose, and one vaccinee in the lower-dose group, subject V14, had an undetectable CD4+ response (Fig. 5B). The rAd26 vaccine did not skew cellular immune responses toward either CD4+ or CD8+ dominance; the CD8+ frequencies, 0.13% and 0.25% (for 1010 and 1011 vaccine doses, respectively) essentially mirrored the magnitude of CD4+ responses. In the case of CD8+ T cells, there were two subjects in the low-dose group, V13 and V14, with undetectable antigen-specific responses (Fig. 5C). In rAd26 study 2, average antigen-specific CD8+ frequencies (0.34%) were higher than CD4+ responses (0.08%), but this apparent skewing toward CD8+ responses was driven primarily by a single subject (V24) with very high CD8+ frequencies and low CD4+ responses. Otherwise, overall cellular immune responses were similar to those observed in rAd26 study 1.
Infectious ZEBOV challenges were performed by i.m. injection of 1,000 PFU ZEBOV at 4 weeks after vaccination for each of the rAd26 vaccine studies, and liver enzyme levels were measured to monitor disease (Fig. 5D). Unvaccinated subjects exhibited manifestations of hepatic injury between days 3 and 6 of infection. Animals that were vaccinated with 1011 or 1012 particles of rAd26-GP either exhibited normal levels of AST or lagged behind the unvaccinated control for signs of hepatic injury. All subjects receiving the lowest-dose rAd26 vaccine, 1010 particles, showed clinical symptoms of disease similar to those of the unvaccinated subject, though AST levels increased at a lower rate. As predicted from this clinical indicator, all subjects in this group succumbed to the lethal effects of infection by day 8 after ZEBOV challenge (Fig. 6). The antibody and T-cell responses were low or undetectable in some subjects in this group. Differences in the magnitude of immune responses between the dose groups in Ad26 study 1 (1010 and 1011) were generally reflected in the survival rates, with a higher survival outcome, 2 out of 4 protected, in subjects vaccinated with 1011 particles of rAd26-GP (P = 0.01). Ad26 at 1011 particles was not only superior to the lower dose of Ad26 but also provided greater protection than rAd35-GP when matched for dose (P = 0.01). This difference was reflected in both CD8+ T-cell and ELISA IgG responses, but the ICS differences were within the range of interassay variability observed for these studies. Finally, rAd26 given at a dose of 1012 particles yielded the highest number of survivors (3 out of 4) for any vaccine regimen tested in these studies (Fig. 6); however, this survival rate was obtained using a higher dose for rAd26 than for rAd5, 1012 versus 1010 particles, respectively, suggesting a potential potency difference between these vectors in this animal model.
The dose-response characteristics for rAd26-GP-mediated immune protection and the high survival outcome using 1012 vector particles suggested that this vector efficiently induces GP antigen expression. Since it has been shown for EBOV and other pathogens, in both nonhuman and human subjects, that heterologous prime-boost vaccination can elicit more potent immunity than single-shot immunization (10, 18, 39, 46) we asked whether rAd26-GP immune responses could be boosted with a heterologous vector to improve protection against EBOV infection. We chose a vaccine dose of 1011 particles with an eye toward potential clinical testing, as studies have shown this dose of related vectors to be tolerated in human subjects (3, 32). Four cynomolgus macaques were inoculated with 1011 particles (each) of rAd26-GP(Z) and -GP(S/G). One month later, all subjects received a boost vaccination with the same dose of rAd35-GP(Z) and rAd35-GP(S/G). Immune responses were assessed immediately prior to and after the boost. Figure 7A shows that antibody responses against EBOV-GP(Z) were efficiently induced by the priming vaccination. Individual subjects generated EC90 antibody titers against GP of from 1:2,700 to 1:7,100, and the average titer for the group was 1:4,000, consistent with the responses observed in the previous study testing 1011 rAd26 as a single-inoculation vaccine (1:4,500). This study included for comparison a single subject inoculated with 1010 particles (each) of rAd5-GP(Z) and rAd5-GP(S/G), whose postvaccination antibody titer was 1:6,800. Subsequent inoculation of the rAd26-GP-primed subjects with 1011 particles of rAd35-GP vectors boosted antigen-specific antibody levels approximately one order of magnitude for most vaccinees, to an average titer of 1:32,000, except for subject V27, whose postprime antibody titers were exceptionally high. Interestingly, the boost vaccination generated more uniform titers across subjects, with a standard deviation across subjects of just under 10%, than the postprime titers, which exhibited a standard deviation of 54%.
Cellular immune responses were also boosted markedly by rAd35-GP administration to rAd26-vaccinated macaques (Fig. 7B and C). CD4+ T cells were boosted in all except one subject, V25; the average increase after rAd35-GP administration was 2-fold across all subjects, and boosting revealed a measurable response in subject V27, whose response was undetectable prior to boosting. Final, postboost CD4+ T-cell frequencies were comparable to those generated by rAd5-GP vaccination. The boost effect was greatest for CD8+ T cells, and all subjects exhibited the boost effect in this cellular compartment, yielding in two subjects (V27 and V28) responses exceeding those generated by rAd5-GP vaccination. Average GP-specific CD8+ T-cell frequencies measured by ICS were 0.09% after primary immunization, and these increased 4.7-fold to 0.41% 3 weeks after secondary vaccination with rAd35-GP.
Altogether, the immunogenicity results above showed that rAd35-GP vectors are potent for boosting rAd26-GP-primed macaques. Postboost GP-directed antibodies were induced to an average level that is nearly a log higher than the level predictive for 100% immune protection in rAd5-GP-vaccinated primates (45). Importantly, the rAd35-GP boost provided a substantial enhancement of CD8+ T-cell frequencies, also shown to associate with immune protection against EBOV infection. Therefore, 1 week after assessment of immune responses (4 weeks postboost), all vaccinated subjects and one unvaccinated control macaque were exposed to 1,000 PFU of ZEBOV by intramuscular injection. The control subject exhibited clinical symptoms characteristic of EBOV infection and succumbed to lethal effects at day 6 after challenge (Fig. 7D and E). In contrast, circulating AST levels in all vaccinated subjects remained normal, and these subjects exhibited no evidence of hemorrhagic disease in gross pathology evaluation at study termination (not shown). All four vaccinated subjects survived infectious challenge and remained symptom-free throughout the 28-day follow-up period until study termination, demonstrating a statistically significantly higher survival rate than that of unvaccinated controls (P = 0.01). These results showed that rAd26/rAd35 vectors administered as a heterologous prime-boost vaccine regimen provide uniform immune protection against ZEBOV infection and demonstrate the potential utility of this approach for achieving additive or synergistic results with combination vaccines.
Adenoviruses perform well as vaccine vectors for the delivery of a variety of viral, bacterial, and parasitic antigens (20). rAd5 vectors in particular generate potent antigen-specific immune responses in mice, nonhuman primates, and humans, as we have observed when EBOV-GP is the target antigen. However, preclinical and human clinical studies have suggested that the potency of rAd5-based vectors may be compromised in individuals who have been exposed previously to Ad5 if they have a high level of immunity against the vector. The aim of the studies herein was to identify rAd vectors that can deliver the EBOV GP antigen in both Ad5-naïve and Ad5-immune subjects.
Since preexisting immunity against any viral vector has the potential to limit its effectiveness, we focused our attention on viruses that infect humans relatively rarely, as indicated by the prevalence of seropositive subjects and/or the low levels of neutralizing antibodies. Vesicular stomatitis virus (VSV) has been used successfully as a vector for filovirus vaccines, and antibody frequency against the vector in human subjects is presumed to be low, since it is primarily a veterinary pathogen and rarely causes symptomatic clinical infection in humans. However, the discovery of Chandipura virus as the causative agent of febrile illness and encephalitis in humans (35) as well as documented human VSV exposure coincident with livestock outbreaks in the United States (37) suggest that natural exposure may occur more frequently than has been assumed, but surveys have not been performed to evaluate human seroprevalence. In contrast, extensive sampling has been undertaken to determine the frequency of anti-adenovirus antibodies in human sera (1, 28). The adenovirus serotypes Ad35 and Ad26 were selected for vaccine development in the current work due to their low seroprevalence in humans.
rAd35-GP vaccination of macaques generated antigen-specific antibody and T-cell responses in individual subjects within the range observed previously when rAd5 was used as the delivery vector. The average anti-GP antibody titer for all rAd35 vaccinees (irrespective of dose), 1:1,400, was lower than the average for all historical subjects vaccinated with 1010 rAd5-GP (1:11,000, n = 17), providing an initial indication that vector potency might differ between the two serotypes if antibody titer is an immune correlate of protection for rAd35 as it is for rAd5 EBOV vaccines. CD4+ and CD8+ T-cell responses were detectable in most subjects prior to infectious challenge, though the absolute magnitude cannot be compared to that for rAd5 vaccinees not included in these studies in the absence of PBMC samples for assay bridging controls.
Vaccination with rAd35 vectors effectively induced antigen-specific antibody and T-lymphocyte immune responses in both rAd5-naïve or rAd5-immune subjects, suggesting that rare serotype vector genomes are sufficiently distant from common serotypes to resist heterologous vector-directed immunity. This feature of vector performance will be important to circumvent preexisting immunity stemming not only from natural viral infection but also from the use of heterologous vectors in priming immunizations or vaccination against other pathogens. Indeed, rAd35-GP inoculation provided a potent boost of both cell-mediated and antibody responses in macaques primed with rAd26-GP. This result was intriguing, since it demonstrates a clear difference in vector potency for the induction of secondary versus primary immune responses; the ability of rAd35-GP to boost the immune response was not predicted by the magnitude of responses observed after the priming immunization. These data may indicate that rAd35 and other rAd vectors have a higher transduction efficiency in certain populations or activation states of target dendritic cells, as suggested recently by Lindsay et al. (23), which may be, in this case, more abundant or accessible during secondary immune responses.
The rAd26/rAd35 prime-boost vaccine provided uniform protection, and rAd26 proved to be more potent than rAd35 as a single-shot vaccine against EBOV infection, mediating survival in up to 75% of vaccinated macaques at the highest dose tested. rAd26-GP vaccines demonstrated a clear dose response for the induction of protective immunity, suggesting that marginal improvements in antigen expression could increase the potency of rAd26-based vaccines to generate uniform protection against high doses of EBOV challenge, such as those used herein. Interestingly, the high degree of protection offered by rAd26-GP vectors compared to rAd35-GP at a matched dose (1011 particles) was associated with higher ELISA anti-GP titers, 1:4,500 versus 1:1,400, respectively. These data raise the possibility that prechallenge antibody titers could serve as an immune correlate of protection against ZEBOV infection for the Ad26/Ad35 combination in addition to within-vector groups, as has been observed for rAd5-GP vaccines. The order of potency for induction of antibody responses predicted the rank order for protection across vector groups.
The studies herein tested vaccine vectors by comparing them singly and in combination, demonstrating the utility of alternative-serotype rAd for use as vaccine vectors in primates. rAd26 and rAd35 were less potent vectors than rAd5 or VSV (11) when used as single-modality vaccines, but they performed well when combined, mediating full protection in macaques. This promising result in macaques suggests that these vaccines will be immunogenic in humans for the induction of anti-EBOV responses. Although it is not yet known whether the protective doses used here will be tolerated in humans, an rAd26-based HIV vaccine was tolerated at similar doses in a recent human clinical trial (3). The nonhuman primate studies here suggest that these vaccines may be most useful in a prime-boost combination. By administering rAd26 with Ad35 in a heterologous vaccination regimen, humoral and cellular immune responses were efficiently increased. Because of the high magnitude of antigen-specific responses achieved by heterologous prime-boost, it has been proposed that long-term immunity may be optimally achieved by priming rAd with DNA (38). Since DNA requires multiple primes and does not induce rapid protection like rAd vectors do, heterologous rAd prime-boost may provide an optimal opportunity to generate a balance between the inductions of rapid and long-lasting protective immunity.
We are grateful for generous support from Gary Nabel throughout the course of these studies. We thank Mythreyi Shastri, Denise Jeffers, Ati Tislerics, and Brenda Hartman for help with manuscript preparation, Michael Cichanowski for graphics, Samuel Franks for assistance with flow cytometry, and Jessica Liu for data management, J. P. Todd, Vi Dang, Srinivas Rao, and Sarah Norris for helpful discussions, and Kathryn Kenyon for editorial support.
Animal studies performed under BSL-4 biocontainment conditions at USAMRIID were approved by the USAMRIID Institutional Animal Care and Use Committee. Animal research was conducted in compliance with the Animal Welfare Act and other Federal statutes and regulations relating to animals and experiments involving animals and adheres to the principles stated in the National Research Council's Guide for the Care and Use of Laboratory Animals. The facilities used are fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International.
Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the U.S. Army or the Department of Defense.
Published ahead of print on 16 February 2011.