Novel Vaccine Regimen Elicits Strong Airway Immune Responses and Control of Respiratory Syncytial Virus in Nonhuman Primates

ABSTRACT

Induction of long-lasting immunity against viral respiratory tract infections remains an elusive goal. Using a nonhuman primate model of human respiratory syncytial virus (hRSV) infection, we compared mucosal and systemic immune responses induced by different DNA delivery approaches to a novel parenteral DNA prime-tonsillar adenoviral vector booster immunization regimen. Intramuscular (i.m.) electroporation (EP) of a DNA vaccine encoding the fusion protein of hRSV induced stronger systemic immune responses than intradermal EP, tattoo immunization, and conventional i.m. DNA injection. A single EP i.m., followed by two atraumatic tonsillar immunizations with the adenoviral vector, elicited strong systemic immune responses, an unique persistent CD4⁺ and CD8⁺ T cell response in the lower respiratory tract and protection from intranasal hRSV challenge. Thus, parenteral DNA priming followed by booster immunization targeted to a mucosal inductive site constitutes an effective vaccine regimen for eliciting protective immune responses at mucosal effector sites.

IMPORTANCE The human respiratory syncytial virus (hRSV) is the most common cause of severe respiratory tract disease in infancy and leads to substantial morbidity and morality in the elderly. In this study, we compared the immunogenicity and efficacy of several gene-based immunization protocols in rhesus macaques. Thereby, we found that the combination of an initially parenterally delivered DNA vaccine with a subsequent atraumatic tonsillar adenoviral vector immunization results in a strong systemic immune response accompanied by an exceptional high T-cell response in the mucosa. Strikingly, these animals were protected against a RSV challenge infection controlling the viral replication indicated by a 1,000-fold-lower viral load in the lower respiratory tract. Since mucosal cellular responses of this strength had not been described in earlier RSV vaccine studies, this heterologous DNA prime-tonsillar boost vaccine strategy is very promising and should be pursued for further preclinical and clinical testing.

INTRODUCTION

The human respiratory syncytial virus (hRSV) is the most common cause of severe respiratory tract disease in infancy. Almost all children experience their first hRSV infection before they reach the age of 2 years, but in most cases these are accompanied by relatively mild upper respiratory tract disease. However, in 1 to 2% of the cases, especially in infants less than 6 months of age, hRSV infection induces severe disease that may lead to hospitalization. In addition, severe hRSV infection during early life is associated with an increased risk of development of asthma and other atopic diseases (reviewed in reference 1). In the elderly, reinfection with hRSV has been reported to be as severe with respect to the rate of hospitalization, the need for intensive care, and mortality as seasonal influenza (2). In total, hRSV is responsible for an estimated 66,000 to 199,000 deaths every year worldwide (3). Although great efforts have been conducted to develop a safe and efficacious vaccine against RSV, no vaccine has yet been licensed (reviewed in reference 4). Only a neutralizing monoclonal antibody, Palivizumab, has been approved for prophylaxis in infants (5), but high costs and the need for repeated administration puts limitations on its widespread use (6).

Gene-based vaccines, including DNA vaccines, provide a platform for the endogenous expression to target antigens within the recipient's own cells. In contrast to vaccines based on whole inactivated viruses or viral subunits, endogenous expression facilitates exposure to antigens in their natural conformation and with appropriate posttranslational modifications, both of which appear to be favorable for the induction of neutralizing antibodies and balanced cellular immune response. Since intramuscular (i.m.) DNA immunization has also been shown to induce Th1 immune responses (reviewed in reference 7), DNA immunization appears to be a promising hRSV vaccine approach with the potential to avoid the immune response profile correlated with the occurrence of vaccine mediated disease potentiation (reviewed in reference 8). Furthermore, DNA vaccines have the potential of being used in combination against different respiratory tract infections in humans without increasing the number of required vaccinations (9). Initial difficulties to extrapolate promising results from mice (10) to larger species, including nonhuman primates (NHPs), were partially overcome by more sophisticated DNA delivery approaches such as jet injection and DNA electroporation (11, 12). DNA vaccine candidates are currently being assessed for safety and immunogenicity in early clinical trials for infectious disease indications, including HIV, hepatitis C virus (HCV), and the influenza virus. Initial interest in DNA and other gene-based vaccines against hRSV using the fusion protein of hRSV (hRSV-F) as vaccine antigen was tempered by poor immunogenicity and efficacy in NHPs (13–15). It subsequently turned out that expression of hRSV-F cDNA is impaired by premature polyadenylation and inefficient nuclear export (16, 17). However, high levels of the hRSV-F protein, as well as its secreted form, could be expressed from a codon-optimized hRSV-F sequence. For the immunization in the presented study, the secreted form of hRSV-F was used as a vaccine antigen since expression of a fusion active full-length hRSV-F was shown to be cytotoxic (18). Consistent with other studies (19, 20), we had previously observed that immunization with DNA or adenoviral vectors (AdV) vaccines expressing the soluble hRSV-F protein, resulted in a superior immune response and protective efficacy in mice, which opened the possibility to further explore these genetic vaccines against hRSV in NHPs (21, 22). Structural data further support the concept of induction of neutralizing antibodies by soluble forms of the hRSV-F (23).

One scope of the present study was to explore the immunogenicities of different DNA delivery methods in NHPs, including i.m. DNA immunization delivered either by conventional injection or by electroporation (EP), intradermal (i.d.) EP, and tattoo immunization (24). Direct mucosal immunization is considered the best way to target immune responses to mucosal sites, since primed immune cells are preferentially homing to the effector sites corresponding to the initial inductive sites (reviewed in reference 25). Moreover, a recent study introduced an alternative strategy of parenteral priming with the vaccine antigen, followed by mucosal application of chemokines to pull the primed T cells to the effector site (26). Instead of antigen-independent recruitment of T cells to the effector site by chemotactic stimuli, we explored whether systemic immune responses could be redirected to the airways by tonsillar booster immunizations. In addition, the DNA priming has been successfully combined with AdV boosting in several studies (27, 28). Therefore, we used DNA expressing soluble hRSV-F in combination with an AdV expressing the same antigen, which was delivered by the tonsillar route. Since the parenteral DNA prime-tonsillar boost immunization induced striking systemic and mucosal cellular and humoral immune responses, we determined the persistence of the mucosal immune responses and challenged these animals with hRSV, revealing significant protective efficacy.

MATERIALS AND METHODS

Ethics statement.

Both the DNA immunization study and the hRSV challenge study were reviewed and approved by an external ethics committee authorized by the Lower Saxony State Office for Consumer Protection and Food Safety and were performed under project license numbers 33.9.42502-04-017/07 and 33.9.42502-04-10/025733.9.42502-04-017/07 issued by the same State Office. All animals were housed at the German Primate Center under conditions in accordance with the German Animal Welfare Act, complying with the European Union guidelines on the use of nonhuman primates for biomedical research. This includes measures of animal welfare and amelioration of suffering in all work such as a 12:12 light-dark schedule, provision of monkey biscuits supplemented with fresh fruit twice a day, and constant access to water. In addition, the monkeys were kept under permanent medical supervision. In cases of suffering predefined by a scoring system on termination criteria, monkeys were killed humanely. Healthy volunteers donating serum samples for the testing of hRSV specific antibodies in the present study provided a written informed consent.

Vaccines.

The DNA vaccine, pVax-Fsol, was constructed by cloning the codon-optimized open reading frame of the soluble form of the hRSV-F from the A2 Long strain by XhoI and HindIII from the previously described pcD-Fsol (16) into pVAX1 (Invitrogen, Germany). The optimization process, construction, and in vitro and in vivo analysis of the codon-optimized open reading frame (GenBank accession no. EF566942) were previously described (16). Plasmid DNA was purified according to the instructions of the manufacturer by use of an EndoFree plasmid Giga kit (Qiagen, Hilden, Germany). The concentration of endotoxin was measured as described previously (21). Recombinant replication deficient adenoviral vector (rAdV) carrying the soluble form of the codon optimized hRSV-F gene (rAdV-Fsol) was described previously (22). The rAdV-Fsol vaccine preparation used had a particle concentration of 10¹¹ gene transferring units per ml.

Animals and sampling.

Thirty-four purpose-bred young adult rhesus macaques (Macaca mulatta) of Indian origin and of both sexes were assigned to the present study. They were housed under BSL2 conditions at the German Primate Center in accordance with the German Animal Welfare Act and in compliance with the European Union guidelines on the use of nonhuman primates for biomedical research. In order to collect blood samples, animals were sedated via intramuscular injection (i.m.) with 10 mg of ketaminhydrochloride per kg of body weight. For all other measures and treatments, animals received a deeper level of anesthesia through an i.m. injection with a mixture of 5 mg of ketaminhydrochloride, 1 mg of xylazinehydrochloride, and 0.01 mg of atropine sulfate per kg of body weight. Bronchoalveolar lavage (BAL) samples were obtained as reported elsewhere (29). Nasal and throat swabs were collected by using flocked nylon swabs (minitip flocked swab, MicroRheologics, Brescia, Italy), and the samples were released in 1 ml of phosphate-buffered saline and stored at −80°C.

Immunization.

Each vaccinee received a total of 1 mg of pVax-Fsol DNA per immunization, except for animals immunized by tattoo delivery method; these animals received three consecutive applications with 0.33 mg each at 3-day-intervals. The skin of the vaccination area was shaved before application, except for the conventional i.m. immunization. The DNA vaccine was diluted in injectable isotonic 0.9% saline solution (Braun, Germany) and filled in syringes in volumes of 250 μl for the i.m.-immunized animals, 200 μl for the EP i.d.-immunized animals, and 100 μl for the tattooed animals and stored overnight at 4°C. Animals that were immunized via the conventional i.m. application received 250 μl of DNA solution each (2 mg/ml in isotonic NaCl) into the right and left thighs. The control animals remained untreated. For the immunization via i.m. electroporation (EP) the animals were, as previously described, vaccinated with a “needle-style” EP device (TriGrid delivery system; Ichor Medical Systems, San Diego, CA) (30–32). For the i.d. EP, 100 μl of DNA solution (5 mg/ml) was injected into the dermis of both arms of the animals using the Mantoux technique, resulting in a visible bleb. Thereafter, the plate applicator of a Cliniporator (IGEA, Italy) was placed on the skin positioning the bleb between the two plate electrodes filled with sonographic gel. The electric parameters were two electric pulses of 450 V for 50 μs, followed by eight pulses of 110 V for 20 ms. For tattoo immunization, a tattoo applicator vessel was filled with 100 μl of DNA solution (3.3 mg ml⁻¹), and the DNA was subsequently delivered by a tattoo machine as previously described (33). The tonsillar spray immunization with rAdV was carried out as described previously (34). From weeks 32 to 98 all of the group E animals (animals 13916, 13928, 13931, 13932, and 13935) were also immunized with unrelated simian immunodeficiency virus (SIV) antigens through DNA and viruslike particle vaccines as reported previously (31, 35).

Viral challenge.

For the hRSV challenge, a macaque-adapted hRSV stock was used (3, 36). After thawing of aliquots stored at −135°C, the titer was determined on HEp2 cells resulting in 1.1 × 10⁵ PFU ml⁻¹. Animals were challenged by slowly pipetting of hRSV dropwise into each nostril.

hRSV-F protein purification.

An hRSV plasmid expressing the codon-optimized hRSV-F ectodomain fused to a His tag was transfected in 293T cells using polyethylenimine as described previously (4, 37). Supernatants were harvested and placed on a column packed with 5 ml of cobalt resin (Thermo Scientific, Germany) as recommended by the manufacturer. His-tagged hRSV-F protein was eluted with 25 ml of 300 mM imidazole in Dulbecco modified phosphate-buffered saline (PBS) and concentrated by Vivaspin 20 centrifugation with a 100,000-Da molecular mass cutoff (Sartorius, Germany). Silver staining of polyacrylamide gels revealed an ca. 80% purity of the hRSV-F protein.

Immune monitoring of rhesus monkey hRSV-specific IgG and IgA.

As described previously, serum, plasma, and BAL fluid samples were analyzed for IgA and IgG hRSV and hRSV-F specific antibody levels using a conventional enzyme-linked immunosorbent assay (ELISA) methodology (5, 22). For coating of ELISA plates, heat-inactivated whole virus based on RSV A2 Long strain was used (21, 22). For hRSV-F specific antibodies, 100 ng of recombinant hRSV-F protein per well was coated overnight at 4°C in carbonate coating buffer. For IgG determination, a goat anti-monkey γ-chain-specific antibody conjugated with horseradish peroxidase (Rockland; 1:3,000 diluted in PBS) was used, and for the IgA determination a goat anti-monkey IgA-α-chain-specific antibody (Rockland; 1:300 diluted in PBS) conjugated with peroxidase was used. Coupled antibodies were detected with enhanced chemiluminescence (ECL) detection. ECL detection solution was prepared by mixing solution A containing 0.1 M Tris-HCl (pH 8.6) and 0.025% 3-aminophthalhydrazide (Sigma, Germany) and solution B containing 0.01 M 4-hydrocoumarin (Sigma, Germany) in dimethyl sulfoxide (DMSO) at a ratio of 100 to 1 and starting the reaction by adding 30% H₂O₂ at a ratio of 3,125 to 1. According to its endpoint titer in the RSV IgG ELISA, a human serum sample positive for RSV antibodies was defined to contain 156,250 arbitrary units (AU) of RSV IgG, RSV-IgA, and RSV-F IgG antibodies per ml. A serial dilution of this serum sample in duplicates was used as a standard curve in each ELISA. The AUs of RSV-specific antibody levels in rhesus monkey sera were then calculated from the standard row. For the RSV specific IgG analysis serum samples of one animal of group C was excluded due to high background levels of preimmune sera and one serum sample of group E was missing for RSV-F specific antibody responses at 9 days after hRSV challenge.

hRSV-specific neutralizing antibody assay.

As reported earlier (8, 22, 38), neutralizing antibody titers were determined in a 96-well neutralization format assay by using a recombinant hRSV encoding green fluorescent protein hRSV (rgRSV).

Quantitative hRSV-specific reverse transcription-PCR.

Viral RNA was isolated from 140 μl of BAL fluid or nasal and throat swabs by using a QIAamp viral RNA minikit (Qiagen) as described previously (9, 22).

hRSV-F ELISPOT assay.

A peptide library of the hRSV-F sequence was synthesized at the Rockefeller University (New York, NY). The protein sequence of hRSV-F for the peptide library was identical to the encoded hRSV-F protein ectodomain of the genetic vaccines used and corresponded to the viral A2 strain Long. The peptide library was synthesized by the Proteomics Resource Center of the Rockefeller University. A 4-mg portion of each of the 141 peptides containing 15 amino acids starting with the first amino acids of hRSV-F with an overlap of 10 amino acids was provided as a lyophilized pellet and stored at −20°C. Peptides were solubilized in bidistilled water, in 0.1 M acetic acid, in 0.1 M ammonia, in 10% acetonitrile, or in 50 to 100% DMSO, mostly at a concentration of 0.5 or 1 mg ml⁻¹ as recommended by the Proteomics Resource Center. Each of 20 consecutive peptides were pooled, resulting in seven peptide pools with the last pool containing 11 peptides. For peripheral blood mononuclear cell (PBMC) stimulation, the peptide pools were used at a concentration of 10 mg ml⁻¹. An enzyme-linked immunospot (ELISPOT) assay was performed as described previously (10, 39), with the modification that cells were stimulated by each of the seven hRSV-F peptide pools at 2 μg ml⁻¹ for each peptide. The positive control consisted of cells stimulated by Staphylococcus enterotoxin B (SEB) at a final concentration 1 μg ml⁻¹; negative controls comprised cells kept in medium alone or stimulated by an irrelevant peptide pool derived from human hepatitis C virus sequences (11, 12, 34). A background reaction was defined as one of the two negative controls where the higher spot numbers were given, and reactions against hRSV peptide pools were scored after subtraction of the respective background value.

hRSV-F specific T-cell response of BAL cells.

BAL fluid was collected, and cells therein were purified as previously described (13–15, 29). The numbers of cells recovered from BAL were generally within 1 × 10⁶ to 6 × 10⁶ per ml. A total of 10⁶ cells were stimulated with a mixture of two stimulation peptide pools from the hRSV-F peptide library (pools 1 to 4 and pools 5 to 7, respectively) with a final concentration of 2 μg/ml/peptide or with SEB, respectively, or with a peptide pool from HCV (16, 17, 34) or with anti-CD28 as negative controls. Intracellular staining was performed as described previously (21–23, 29) and analyzed by flow cytometry (BD LSRII flow cytometer). All data were analyzed using FlowJo 7.6 (TreeStar).

Statistical analysis.

The statistical analysis was made by using GraphPad Prism 6.0 software. The data were logarithmically transformed as indicated and analyzed with analysis of variance, followed by pairwise multiple comparisons using Tukey's test or Sidak's multiple-comparison test.

RESULTS

Outline of the immunogenicity study.

Although different DNA delivery methods have been explored in mice, it remains to be determined which of them is most immunogenic in primates. Due to differences in the immunogens and in the methods for measuring antigen-specific immune responses, it is difficult to retrospectively compare the magnitude of the immune responses in NHPs induced by different vaccines. We therefore designed a side-by-side comparison of four different delivery approaches with the same DNA vaccine. A codon-optimized plasmid encoding a secreted form of the hRSV-F protein as the major target of neutralizing antibodies was chosen as a DNA vaccine based on our previous results from mice (21, 33). Conventional intramuscular DNA immunization served as standard delivery technology (group A, DNA i.m.) and was compared to intramuscular electroporation of DNA (group B, DNA EP i.m.), intradermal electroporation of DNA (group C, DNA EP i.d.), and application of DNA by tattooing (group D, DNA tattoo). The hRSV-F DNA was administered to groups of five monkeys each with a dose of 1 mg (25, 30, 40, 41) at weeks 0, 9, and 28. In addition, one group of five animals (group E, DNA-AdV) received a heterologous prime-boost regimen starting with DNA by EP i.m. at week 0, followed by booster immunizations at weeks 9 and 28 by spraying an adenoviral vector encoding same F protein antigen of hRSV (22, 26) directly onto the tonsils (27, 28, 34). There were no significant adverse events observed through any of the vaccination methods. Group F served as unvaccinated control group during the immunization phase. An overview of the different immunization regimens is provided in Table 1.

TABLE 1.

Vaccine groups^a

Group	Vaccine, delivery method, route				Challenge (wk 136)
	Prime (wk 0)	Boost 1 (wk 9)	Boost 2 (wk 28)	Boost 3 (wk 126)
A	DNA, i.m.	DNA, i.m.	DNA, i.m.
B	DNA, EP, i.m.	DNA, EP, i.m.	DNA, EP, i.m.
C	DNA, EP, i.d.	DNA, EP, i.d.	DNA, EP, i.d.
D	DNA, tattoo	DNA, tattoo	DNA, tattoo
E	DNA, EP, i.m.	AdV, tonsillar	AdV, tonsillar	DNA, EP, i.m.	hRSV
Fb	None	None	None
Gc				None	hRSV

^aAbbreviations: DNA, recombinant plasmid DNA; i.m., intramuscular; EP, electroporation; i.d., intradermal; AdV, adenoviral vector; hRSV, human respiratory syncytial virus. n = 5 animals per group.

^bGroup F served as an unvaccinated control group for analyses until week 32.

^cGroup G was included as an unvaccinated control group for viral challenge.

Systemic humoral immune responses.

The systemic IgG antibody response to hRSV was first measured by ELISA (Fig. 1A). While hRSV-specific antibodies were not detectable after a single conventional i.m. DNA immunization (group A), delivery of the same DNA by the device-based methods under evaluation (i.m. EP, i.d. EP, or tattooing; groups B to D, respectively) resulted in low levels of binding antibodies. The second DNA immunization led to an approximately 4- to 8-fold increase in hRSV antibody levels in all vaccine groups. The device-based DNA delivery approaches led to substantially higher antibody responses compared to the conventional i.m. immunization. The antibodies induced by i.m. EP seemed to persist at higher levels until the time point of the third DNA immunization (Fig. 1, week 28). Antibody titers continued to rise considerably after the second booster immunization in all groups immunized by the device-based DNA delivery approaches, whereas conventional i.m. immunization performed much more poorly. The DNA prime-tonsillar booster immunization boosted the systemic humoral immune responses as efficiently as a second DNA immunization, with antibodies persisting at similar levels as observed after i.m. EP. The antibodies induced by DNA immunization also exhibited neutralizing activity (Fig. 1B). The DNA prime-tonsillar booster immunization elicited neutralizing antibody levels comparable to those induced by the device-based DNA delivery approaches. In all groups, the neutralizing antibodies rapidly declined close to background levels after the first booster immunization (Fig. 1B). Nevertheless, the vaccine-induced neutralizing activity at peak response was as high as that of a panel of serum samples of healthy young adults, who had previously been infected by hRSV (Fig. 1C), indicating that biologically relevant levels of neutralizing antibodies were generated. Furthermore, we analyzed the induction of systemic hRSV-specific IgA antibodies at week 32 after the first immunization, which revealed significantly elevated levels with i.m. EP and the DNA prime-tonsillar booster immunization (Fig. 1D) (groups B and E, respectively).

FIG 1.

Systemic humoral immune responses in monkeys after applying different hRSV immunization approaches. hRSV-specific binding IgG antibody levels (A) and neutralizing antibody titers (B) in sera of immunized (groups A to E) and control (group F) macaques at the indicated weeks after first immunization. (A) Arbitrary units (AU) of hRSV specific IgG antibody levels were logarithmically transformed to calculate the arithmetic means and standard errors for each group. (B) The highest serum dilution inhibiting hRSV infection by >50% was taken as the neutralization titer (NT). Log₂-transformed titers were used to calculate the arithmetic mean and standard error for each group. Dotted lines indicate the detection limit. Time points of immunizations are indicated by arrows. (C) Neutralizing antibody titers in the sera of healthy young adults are compared to macaques of group E at week 13. The results from individual sera and the geometric mean are shown. (D) hRSV-specific IgA levels in sera of all groups at week 32 after the first immunization. Arbitrary units (AU) of hRSV specific IgA antibody levels were logarithmically transformed to calculate the arithmetic means and standard errors for each group. Statistically significant differences of group B and E to group F are marked (**, P = 0.0035; *, P = 0.0107).

Systemic cellular immune responses.

Seven overlapping peptide pools spanning the entire ectodomain of the hRSV-F protein were used to stimulate PBMCs obtained from immunized macaques and the control group at different time points before and during immunization (Fig. 2A). The presence of hRSV-F specific memory T cells was monitored by an IFN-γ ELISPOT assay. A single i.m. EP (groups B and E) was sufficient to induce considerable levels of gamma interferon (IFN-γ)-producing cells, which were much lower in the other vaccine groups (groups A, C, and D). In the latter groups, even after the second immunization, T-cell responses were still around 4- to 5-fold lower than in groups B and E. Although the third immunization had boosted the response of binding antibodies, it had almost no effect on the IFN-γ ELISPOT responses. More importantly, the first tonsillar immunization (group E) boosted the systemic cellular immune response as efficiently as the i.m. EP (group B). The breadth of the cellular immune response, analyzed by different hRSV-F peptide pools, is shown at its peak at week 11 after the first immunization for each single animal (Fig. 2B). In animals vaccinated by i.m. EP or by DNA prime-tonsillar boost immunization, IFN-γ-secreting cells were induced by several different overlapping peptide pools, whereas responses with limited specificities were observed for the other vaccine groups (Fig. 2B).

FIG 2.

Systemic cellular immune response in monkeys following different hRSV immunization approaches. Longitudinal systemic cellular immune responses of the different vaccine groups (A to E) in comparison to an unvaccinated control group (F) are shown at the indicated weeks (A). The IFN-γ ELISPOT responses of PBMCs were determined by stimulation with seven overlapping hRSV-F peptide pools. Cumulative IFN-γ-secreting cells are shown as spots per one million PBMCs at the indicated weeks before and after first immunization. Arithmetic mean values plus standard errors of all animals per group are shown. Statistically significant differences between the immunized group versus the control group are marked by an asterisk at the top of the corresponding column, and a hash mark (#) indicates a statistically significant difference to preimmune ELISPOT responses. (B) Breadth of cellular immune response at 11 weeks after the first immunization. The number of spots per one million PBMCs stimulated by each of the seven peptide pools (PP1 to PP7) is shown for each individual animal of all groups. The number at the tops of the columns indicate the numbers of peptide pools inducing more than 100 spots/10⁶ PBMCs. nd, not determined.

Mucosal immune responses.

To evaluate whether the different vaccine regimens induced cellular immune responses at the surface of the respiratory tract, cells were purified from bronchoalveolar lavage (BAL) samples 4 weeks after the second booster immunization and they were stimulated with the hRSV-F peptide pools. Cytokine expression (IFN-γ, interleukin-2 [IL-2], and tumor necrosis factor alpha [TNF-α]) was determined in CD4⁺ and CD8⁺ cells through intracellular cytokine staining. Antigen-specific CD4⁺ T cells were only detectable in group E (Fig. 3A). A total of 11.2% of all CD4⁺ T cells in BAL fluid from this group responded to the hRSV-F peptide stimulation through the expression of at least one of the cytokines, and 87.1% of the responding cells produced at least two of the cytokines, demonstrating the polyfunctional character of the hRSV-F specific mucosal T helper cells (Fig. 3A, pie chart). The polyfunctionality of antigen-specific T cells is considered to be indicative of persistent and protective cellular immune responses (16, 42, 43).

FIG 3.

Mucosal cellular and humoral immune responses in vaccinees after the second hRSV booster immunization. Cells were enriched from BAL samples of all vaccinated and control monkeys 4 weeks after the second booster immunization and stimulated with two mixed pools of the hRSV-F peptide pools 1 to 4 and 5 to 7, respectively. The percentages of IFN-γ-, IL-2-, and/or TNF-α-secreting CD4⁺ (A) and CD8⁺ T cells (B) among all CD3⁺ CD4⁺ and CD3⁺ CD8⁺ T cells, respectively, were determined by intracellular cytokine staining after stimulation with peptide pools 1 to 4 or 5 to 7. The percentage of cytokine-positive cells was summed up for the two peptide pools for each animal, and the arithmetic mean, and standard errors are shown for each experimental group. Pie charts for groups with detectable mucosal cytokine responses depict the mean percentages of CD4⁺ or CD8⁺ T cells reacting to peptide stimulation with expression of at least one cytokine in the center and the proportion of T cells producing one (+1), two (+2), or three (+3) cytokines as segments. (C) hRSV-F specific IgG levels in BAL samples obtained at the same time stated above. Arbitrary units (AU) of hRSV-F specific IgG antibody levels in 1:5,000-diluted BAL samples were logarithmically transformed to calculate the arithmetic mean and standard error for each group. Statistically significant differences of each group compared to control group F are marked either with hatches (#, P < 0.0001) or asterisks (**, P = 0.0008; *, P = 0.0025).

Mucosal antigen-specific CD8⁺ T cells were induced by i.m. EP and the DNA prime-tonsillar booster immunization regimen. Remarkably, in the latter group the percentage of hRSV-F specific CD8⁺ T cells reached almost 30% of all CD8⁺ T cells in the BAL fluid samples, whereas only 4.6% of CD8⁺ T cells reacted after i.m. EP (Fig. 3B). There was also a striking difference in the quality of the mucosal CD8⁺ T-cell responses, which were induced in these two groups. The vast majority (87.6%) of the hRSV-F specific CD8⁺ T cells induced by i.m. EP produced only IFN-γ, while the DNA prime-tonsillar booster immunization paved the way for a polyfunctional response in >87.1% of the reacting T cells (Fig. 3B, pie charts).

Moreover, hRSV-specific antibody responses in the BAL samples were analyzed 4 weeks after the second booster immunization (week 32 after first immunization). hRSV-F specific IgG antibodies were present in the BAL fluid samples of all vaccinated groups (Fig. 3C). High levels in the same order of magnitude were observed by EP i.d., EP i.m., and by the DNA prime-tonsillar boost regimen, whereas the levels in groups receiving conventional i.m. and tattoo immunization (groups A and D, respectively) were roughly 10-fold lower.

Efficacy of the systemic DNA prime-mucosal adenoviral vector booster immunization regimen.

Given the striking systemic and mucosal cellular and humoral immune responses induced by the DNA prime-tonsillar booster immunization (group E), we decided to extend the study in order to explore the protective efficacy of this immunization regimen. Due to other research commitments, the animals of group E were first included in an immunogenicity study with unrelated SIV antigens between weeks 60 and 84 (35). At 32 weeks after the last SIV immunization, corresponding to 126 weeks after the first RSV-F DNA immunization and 98 weeks after the last tonsillar booster immunization with the RSV-F adenoviral vector, RSV-F specific antibodies were still detectable in the blood and the BAL fluid (Fig. 4A to C). Although systemic cellular immune responses had declined (Fig. 4D), ca. 5% of the CD4⁺ and CD8⁺ T cells recovered from BAL fluid samples were RSV-F specific. Due to concerns that anti-adenoviral vector immunity might limit the booster effect of a third adenoviral vector immunization, we immunized group E with another DNA EP i.m. at week 126. This boosted the systemic and mucosal antibody responses (Fig. 4A to C). Cellular immune responses as determined by IFN-γ ELISPOT assay from PBMCs (Fig. 4D) were also boosted, whereas no changes were observed for hRSV-F specific CD4⁺ and CD8⁺ T cells recovered from BAL samples (Fig. 4E and F).

FIG 4.

Systemic and mucosal humoral and cellular immune responses in the vaccine group and in the control group before and after hRSV challenge. Systemic hRSV-F specific binding antibodies levels (A) and neutralizing antibody titers (B) were determined at the indicated weeks after the first immunization and at the indicated days after hRSV challenge (dpc) in vaccinated (group E) and unvaccinated control (group G) macaques. The time points of booster immunization in group E and the hRSV challenge are marked by black and gray arrows, respectively. Arbitrary units (AU) of hRSV-F specific IgG antibody levels (A) in 1:5,000-diluted serum samples were used. (B) The highest serum dilution inhibiting hRSV infection by >50% was taken as the neutralization titer (NT). (C) hRSV-F specific IgG levels in BAL samples obtained at the indicated weeks after first immunization. Arbitrary units of hRSV-F specific IgG antibody levels in undiluted BAL samples are presented. Black and gray arrows indicate the time points for the third booster immunization and the hRSV challenge, respectively. For all data, logarithmic transformed titers were used to calculate the arithmetic mean and standard error for each group. Significant differences between vaccine groups and controls are marked at the tops of the corresponding values by asterisks (****, P < 0.0001; ***, P = 0.0078; **, P = 0.0053; *, P = 0.0396). (D) Systemic cellular immune responses were determined by IFN-γ ELISPOT of PBMCs obtained at the indicated time points with hRSV-F and stimulated with peptide pools 1 to 4 and peptide pools 5 to 7. The sum of IFN-γ-positive ELISPOT assays after stimulation with the different peptide pools was determined for each animal. Arithmetic means and standard errors are shown for each group. (E to H) Cells were isolated from BAL samples from animals of both groups at the indicated time points and stimulated with the hRSV-F peptide pools 1 to 4 and 5 to 7, respectively. Cumulative percentages and standard errors of CD4⁺ (E) and CD8⁺ (F) T cells secreting at least one of the cytokines analyzed. Cytokine expression profiles of RSV-F specific CD4⁺ (G) and CD8⁺ (H) T cells are shown for day 21 postchallenge (dpc). nd, not determined.

Ten weeks after the last immunization (week 136 after first immunization) vaccinees of group E, together with unvaccinated controls of group G, were challenged intranasally with a macaque-adapted hRSV strain (16, 36). All animals remained clinically healthy after hRSV infection. However, an increase in hRSV-specific systemic and mucosal humoral (Fig. 4A to C) and cellular immune response (Fig. 4D to F) was observed after challenge in the vaccine group. The hRSV-F specific systemic IFN-γ ELISPOT response 2 weeks after the last DNA booster immunization on week 128 was at least as high as the one observed 2 and 3 weeks after hRSV infection (Fig. 4D). In contrast to the strong CD4⁺ and CD8⁺ T cell responses observed after tonsillar immunization (Fig. 3A and B), and in contrast to the robust anamnestic responses observed after hRSV challenge in the vaccinated group, hRSV infection in naive macaques did not induce striking mucosal cellular immune responses against the hRSV-F (Fig. 4E and F). However, the cytokine profile of the hRSV-F reacting CD4⁺ and CD8⁺ T cells was quite similar after hRSV challenge in vaccinated and naive animals, whereas the quantity was nearly 4-fold higher in the previously vaccinated animals (Fig. 4G and H).

Protection from hRSV infection was analyzed by measuring the viral RNA load in nasal and throat swabs and in BAL fluid. Although the viral load in the nasal cavity of control animals peaked on day 6, the vaccine group was able to suppress viral replication significantly (Fig. 5A). In the throat swabs of the vaccinees, the mean viral load was already reduced by ∼100-fold on day 3 (Fig. 5B), indicating rapid control of virus replication at the site of AdV immunization. Compared to nasal and throat swabs, hRSV RNA levels in the lower respiratory tract of the naive control group were quite low on day 3 but increased >100-fold by day 6, whereas hRSV RNA was barely detectable in BAL fluid from the vaccinees, and peak viral loads were reduced >1,000-fold (Fig. 5C).

FIG 5.

Viral load in the upper and lower respiratory tract after intranasal hRSV challenge. Vaccinated macaques (group E) and control animals (group G) were challenged intranasally 136 weeks after a first immunization with 10⁵ infectious particles of a monkey adapted hRSV strain. hRSV RNA copy numbers in eluted nasal swabs (A), eluted throat swabs (B), and BAL samples (C) were determined at the indicated days after challenge (dpc). BAL samples obtained at week 126 prior to challenge were also analyzed. The hRSV RNA copies number per ml of eluted nasal and throat swabs and BAL fluid were logarithmically transformed to calculate arithmetic means and standard deviations. The detection limit of 3.57 × 10³ copies per ml of received sample is shown by a dotted line. Significant differences between the two groups are marked with asterisks (****, P < 0.001; ***, P = 0.0011; **, P = 0.0315;*, P = 0.0493).

A formalin-inactivated RSV vaccine has previously been associated with immunopathology, including a pulmonary eosinophilic infiltration and enhancement of RSV disease in breakthrough infections (reviewed in reference 44). Characterization of the cellular composition of our BAL samples after challenge did not reveal any evidence for an eosinophilic infiltration (data not shown), and no clinical abnormalities were observed.

DISCUSSION

The short incubation period of viral respiratory tract infections suggests that it is important to develop vaccines that induce immune responses at the mucosal portal of virus entry and replication for rapid control of virus spread. The route of immunization and thus the inductive site of antigen-specific immune responses have been shown to modulate homing-receptor expression of antigen-specific T cells (reviewed in references 21 and 45), thereby directing the antigen-specific T cells to different mucosal effector sites.

Intradermal delivery of viral vector vaccines in mice has been shown to induce mucosal antibody and cellular immune responses, which were also detectable in lymphatic tissues of the respiratory tract (22, 46, 47). Intradermal DNA immunization in a phase I clinical trial revealed moderate systemic immunity (29, 48). One focus of our study was therefore to explore in a NHP model within a comparative manner the potential of device-based i.d. DNA delivery approaches to induce immune responses systemically and in the respiratory tract. According to our procedures of the delivery of DNA by EP i.m. or i.d. or by tattooing they clearly improved the systemic antibody response in comparison to conventional i.m. immunization. Minor differences in the kinetics of the systemic humoral immune responses suggest that the antibody responses induced by i.m. EP persisted at higher levels. Although i.d. EP and i.m. EP elicited higher levels of hRSV-F specific IgG antibodies in the BAL fluid in comparison to a conventional i.m. immunization, only the i.m. EP improved the systemic IgA antibody levels substantially. Systemic cellular immune responses as assessed by IFN-γ-secreting T cells determined by ELISPOT assay were also highest after i.m. EP. An i.d. delivery of DNA did not provide any obvious benefits with respect to the magnitude or the breadth of induced cellular immune responses in comparison to conventional i.m. immunization. The i.m. EP method also induced a cellular immune response at the mucosal site, but this was limited to CD8⁺ T cells secreting IFN-γ only.

Given the strong systemic immune responses induced by various DNA prime-viral vector boost immunization regimens (30–32, 49), we also analyzed the immunogenicity of an i.m. EP, followed by two tonsillar booster immunizations with an AdV encoding the same antigen. The systemic cellular immune response after i.m. EP, either as single modality or combined with the tonsillar booster immunization, peaked at a level of ca. 0.2% hRSV-F specific T cells determined by IFN-γ ELISPOT assay (Fig. 2A). This is in a range similar to that observed for other systemic DNA prime-viral vector boost immunization regimens against different antigens (33, 50) and indicates efficient induction of robust systemic cellular immune responses by both vaccine regimens. However, there was a striking difference between the two vaccine regimens in the cellular immune responses detectable in the BAL fluid. After the tonsillar booster immunization, the percentage of hRSV-F specific CD4⁺ and CD8⁺ T cells in the BAL fluid reached 11 and 22%, respectively (Fig. 3A B). In addition, only the tonsillar booster immunization induced polyfunctional mucosal CD4⁺ and CD8⁺ T cell responses.

This suggests that the tonsillar booster immunization leads to selective recruitment of hRSV-F-specific T cells to and/or expansion of the T cells in the respiratory tract. Thus, booster immunizations at the inductive site of a mucosal lymphatic tissue seem to be sufficient to target previously primed systemic cellular immune response to mucosal effector sites.

One concern that arises when using adenoviral vectors is that repeated immunizations are attenuated by rapidly induced vector-specific immune responses, although we previously observed that the neutralizing antibody response to the adenoviral vector after tonsillar immunization is only detectable after the second immunization and still lower than after intramuscular adenoviral vector immunization (34). In addition, it is still unclear as to whether preexisting immunity due to natural adenovirus infections reduces the immunogenicity of the tonsillar adenoviral vector booster immunizations. However, delivery of adenoviral vectors by small aerosols to the lungs of rhesus macaques resulted in potent immune responses even in animals with preexisting adenoviral vector immunity (31, 35, 51), suggesting that such vector immunity can be overcome by mucosal delivery of adenoviral vectors. In addition, several adenovirus vectors have been developed based on chimpanzee or gorilla adenoviruses, which combine efficacy with evasion of preexisting vector specific neutralizing antibody responses (52, 53).

The challenged virus exposure revealed that the systemic DNA prime-tonsillar booster immunization led to a potent reduction of viral load in the upper respiratory tract and an even more pronounced reduction in the lower respiratory tract. Whether this is due to the unique and potent mucosal immune responses induced by the systemic DNA prime-tonsillar booster immunization remains to be determined. Since the DNA booster immunization at week 126 did not increase mucosal RSV-F specific cellular immune responses, it is conceivable that the mucosal immune responses induced by the initial DNA prime-tonsillar adenoviral vector booster immunization might have been sufficient for protection. The persistence of RSV-F specific antibodies and CD4⁺ and CD8⁺ T cells in the respiratory tract from the last adenoviral vector immunization at week 28 to the DNA booster immunization at week 126 supports this hypothesis. A contribution of the immunizations against SIV antigens between weeks 60 and 84 to the protective efficacy observed seems highly unlikely given the unrelatedness of the vaccine antigens.

In summary, the striking efficacy observed with the systemic DNA prime-tonsillar booster immunization regimen is, to our knowledge, superior to previous studies using DNA of AdV vaccines against hRSV in comparable NHP-animal models (14, 15, 54) and exhibits levels of viral reduction comparable to that achieved with Sendai virus vaccine or with chimeric parainfluenzavirus expressing hRSV antigens (55–57). Our results therefore suggest that the novel DNA prime-tonsillar booster regimen is a promising vaccine approach against hRSV and other respiratory viruses that should be pursued further.