Virus-Like Particle Vaccine Containing the F Protein of Respiratory Syncytial Virus Confers Protection without Pulmonary Disease by Modulating Specific Subsets of Dendritic Cells and Effector T Cells

ABSTRACT

There is no licensed vaccine against respiratory syncytial virus (RSV) since the failure of formalin-inactivated RSV (FI-RSV) due to its vaccine-enhanced disease. We investigated immune correlates conferring protection without causing disease after intranasal immunization with virus-like particle vaccine containing the RSV fusion protein (F VLP) in comparison to FI-RSV and live RSV. Upon RSV challenge, FI-RSV immune mice showed severe weight loss, eosinophilia, and histopathology, and RSV reinfection also caused substantial RSV disease despite their viral clearance. In contrast, F VLP immune mice showed least weight loss and no sign of histopathology and eosinophilia. High levels of interleukin-4-positive (IL-4⁺) and tumor necrosis factor alpha-positive (TNF-α⁺) CD4⁺ T cells were found in FI-RSV immune mice, whereas gamma interferon-positive (IFN-γ⁺) and TNF-α⁺ CD4⁺ T cells were predominantly detected in live RSV-infected mice. More importantly, in contrast to FI-RSV and live RSV that induced higher levels of CD11b⁺ dendritic cells, F VLP immunization induced CD8α⁺ and CD103⁺ dendritic cells, as well as F-specific IFN-γ⁺ and TNF-α⁺ CD8⁺ T cells. These results suggest that F VLP can induce protection without causing pulmonary RSV disease by inducing RSV neutralizing antibodies, as well as modulating specific subsets of dendritic cells and CD8 T cell immunity.

IMPORTANCE It has been a difficult challenge to develop an effective and safe vaccine against respiratory syncytial virus (RSV), a leading cause of respiratory disease. Immune correlates conferring protection but preventing vaccine-enhanced disease remain poorly understood. RSV F virus-like particle (VLP) would be an efficient vaccine platform conferring protection. Here, we investigated the protective immune correlates without causing disease after intranasal immunization with RSV F VLP in comparison to FI-RSV and live RSV. In addition to inducing RSV neutralizing antibodies responsible for clearing lung viral loads, we show that modulation of specific subsets of dendritic cells and CD8 T cells producing T helper type 1 cytokines are important immune correlates conferring protection but not causing vaccine-enhanced disease.

INTRODUCTION

Respiratory syncytial virus (RSV) is a major human pathogen that causes bronchiolitis in infants and young children, as well as serious respiratory illness in elderly and immunocompromised adults. It is estimated that approximately 3.4 million children are annually hospitalized due to RSV-related illnesses and 160,000 people die from RSV infection worldwide (1). Despite extensive attempts to develop RSV vaccines, there have been significant obstacles and challenges. This is partially due to the disastrous outcome of formalin-inactivated, alum-adjuvanted RSV (FI-RSV) vaccine in the 1960s. In this trial, children who were vaccinated with FI-RSV developed vaccine-enhanced respiratory disease (ERD) resulting in hospitalizations and two deaths during the next epidemic season (2). Atypical T helper type 2 (Th2)-biased T cell responses were reported to be associated with enhanced histopathology following experimental immunization with FI-RSV in small animals (3–5). In addition, a high rate of RSV reinfection is observed during childhood and throughout life, although RSV is effectively cleared after primary infection and both RSV-specific antibody and T-cell responses are induced (6). Illness associated with RSV reinfection includes sinus complications with upper respiratory tract infections and increased airway resistance as lower airway disease (7, 8). Thus, it is suggested that a protective immune response to an ideal vaccine should differ quantitatively or qualitatively from that induced by natural infection.

Virus-like particles (VLPs) have morphologies similar to live viruses in size and external structure but do not have viral genomes. It was demonstrated that intramuscular immunization of mice with Newcastle disease virus-based VLPs containing the chimeric RSV attachment (G) or both the chimeric G and the fusion (F) proteins induced protection against RSV, although the roles of T cells in protection were not investigated (9, 10). Influenza M1-based VLPs containing the RSV F protein (F VLP) was produced using the recombinant baculovirus expression system and shown to induce protection (11, 12). A cocktail vaccination of RSV F and G VLPs and F DNA was recently demonstrated to induce protection without an obvious sign of ERD (13). However, cellular phenotypes of immune cells contributing to the protection or ERD after RSV mucosal immunization and infection are poorly understood partially because there is no licensed RSV vaccine.

The licensed RSV monoclonal antibody drug (Synagis [palivizumab]) is known to recognize an epitope in the RSV F protein (14–16). Thus, RSV F is considered a promising RSV vaccine antigen. An important determinant for protection against RSV may be the ability of the vaccine to induce mucosal and systemic immunity. Here, we investigated humoral and cellular immune correlates for protection in mice that were intranasally immunized with RSV F VLPs. We also analyzed innate and adaptive immune cells possibly contributing to RSV protection and/or disease by comparing F VLPs with FI-RSV and live RSV. The results in this study suggest that, in addition to inducing RSV-neutralizing antibodies, the modulation of specific subsets of CD8α⁺ and CD103⁺ dendritic cells (DCs), the induction of a Th1 type cytokine-inducing pulmonary microenvironment, and CD8 T cells producing IFN-γ by F VLP vaccination are important immune correlates for conferring protection against RSV without causing ERD.

MATERIALS AND METHODS

Cells, virus, and reagents.

Spodoptera frugiperda 9 (Sf9) insect cells (CRL-1711; American Type Culture Collection [ATCC], Manassas, VA) were maintained in suspension in serum-free SF900-II medium (Gibco-BRL, Grand Island, NY) and used for production of recombinant baculoviruses (rBVs) and VLPs. Human RSV A2 was kindly provided by Martin Moore (Emory University, Atlanta, GA). HEp-2 cells were purchased from the ATCC (Rockville, MD). Monoclonal mouse anti-RSV fusion protein (131-2A) was obtained from Millipore (Billerica, MA). RSV F-specific palivizumab antibody (MedImmune, Gaithersburg, MD) was kindly provided by Frances Eun-Hyung Lee. Horseradish peroxidase (HRP)-conjugated anti-mouse antibody IgG, IgG1, IgG2a, and IgA were obtained from Southern Biotech (Birmingham, AL).

Preparation of RSV and formalin-inactivated RSV virus.

HEp-2 cells were cultured in Dulbecco modified Eagle medium (DMEM; Gibco-BRL, Grand Island, NY) with 10% fetal bovine serum (FBS), 2 mM glutamine, penicillin (1% vol/vol), and streptomycin (1% vol/vol) at 37°C with 5% CO₂. RSV A2 virus in serum-free DMEM was added to HEp-2 cells and adsorbed for 1 h. DMEM containing 5% FBS was added to the flask, followed by incubation at 37°C for 72 h. RSV-infected cells were harvested and centrifuged at 2,000 rpm for 15 min at 4°C to remove cell pellets and then filtered.

FI-RSV was prepared by a modification of a previously described method (17, 18). Briefly, virus clarified by centrifugation and filtering was inactivated with formalin (1:4,000 [vol/vol]) for 72 h at 37°C. The formalin-treated virus was pelleted by ultracentrifugation at 4°C (30,000 rpm, 1 h) and then resuspended in serum-free medium. Inactivation of FI-RSV was confirmed by an immune-plaque assay and then adsorbed to aluminum hydroxide (4 mg/ml) for FI-RSV vaccine.

Preparation of RSV F VLPs.

F VLP expressing RSV A2 fusion (F) protein and influenza virus matrix 1 (M1) core protein was produced by the Sf9 insect cell expression system and characterized as previously described (11). Briefly, Sf9 insect cells were cotransfected with rBVs expressing influenza virus M1 core protein and RSV F protein in serum-free SF900-II medium for 3 days. The culture supernatants were collected by centrifugation (6,000 rpm, 20 min) to remove insect cells. The cleared supernatants containing F VLPs were concentrated by the QuixStand hollow fiber-based ultrafiltration system (GE Healthcare, Piscataway, NJ) and purified by sucrose gradient ultracentrifugation with layers of 30 to 60% (wt/vol) as previously described (11). The incorporation of F proteins into VLPs was confirmed by Western blotting, and the particle sizes ranged from 60 to 100 nm in diameter, as examined by electron microscopic (11).

Mouse immunization, RSV infection, and sample collection.

Female BALB/c mice (n = 5; Charles River Laboratories, Inc., Wilmington, MA) aged 6 to 8 weeks were immunized intranasally (i.n.) twice with 25 μg of F VLPs, 5 μg of FI-RSV in alum (20 μg) adjuvant, or phosphate-buffered saline (PBS; naive mock control), or they were i.n. inoculated twice with live RSV (10⁶ PFU) at 4-week intervals. Blood samples (five mice per group) were collected at 3 weeks after the prime and boost administrations. Naive, vaccine-immunized, or live RSV-reinfected mice were i.n. infected with 2 × 10⁶ PFU of RSV A2 under isoflurane anesthesia at 3 or 12 weeks after boost administration to determine the efficacy of protection. The body weight changes were monitored for 9 days and individual organs, including the lungs, mediastinal lymph nodes (MLN), spleens, bone marrow (BM), and bronchoalveolar lavage fluids (BALF) were collected at 3, 5, or 9 days postchallenge (dpc). All animal studies were performed according to the guidelines of Georgia State University Institutional Animal Care and Use Committee.

Antibody responses, cytokines, and chemokine production by ELISA.

F protein-specific antibody responses were determined in immune sera, BALF, and lung homogenates by enzyme-linked immunosorbent assay (ELISA). Briefly, 96-well plates were coated with purified F protein (200 ng/ml; BEI Resources, Manassas, VA) per well and then blocked with 0.05% Tween 20 in PBS containing 1% bovine serum albumin for 1.5 h at 37°C. After a washing step, the diluted sera were added, followed by incubation for 1.5 h at 37°C. Antibody responses were detected using HRP-conjugated goat anti-mouse IgG, IgG1, IgG2a, and IgA secondary antibodies (Southern Biotechnology), developed with TMB (3,3′,5,5′-tetramethylbenzidine; Sigma-Aldrich), and stopped with 1 M H₃PO₄. The optical density was determined at 450 nm using a Bio-Tek ELISA plate reader. The total antibody amount was quantified using the standard curve for each IgG isotype antibody. The levels of cytokines and chemokines, such as gamma interferon (IFN-γ), interleukin-4 (IL-4), IL-13, tumor necrosis factor alpha (TNF-α), and IL-12p70 (eBioscience, San Diego, CA) and eotaxin (R&D Systems, Minneapolis, MN), in BALF or lung homogenates were measured using ELISA kits according to the manufacturer's instructions.

Lung viral titration and RSV neutralization assay.

RSV titers for clarified live RSV stocks and lung samples collected at 3, 5, and 9 dpc, and RSV neutralizing activity of immune sera were examined by an immunoplaque assay as previously described (11). Briefly, serially diluted live RSV and lung homogenates were added to the HEp-2 cell monolayer plates, adsorbed for 2 h at 37°C, and then incubated at 37°C for 3 days. The plates were fixed with 5% formaldehyde in PBS and developed by mouse anti-RSV F monoclonal antibody and HRP-conjugated anti-mouse IgG antibody using 3,3′-diaminobenzidine tetrahydrochloride (DAP) substrate (Invitrogen). Virus titers were represented as plaques per gram of lung tissue. The viral load detection limit is ∼50 PFU from the lung samples of mice in this assay.

For the antibody neutralizing assay, immune sera were inactivated at 56°C 30 min and then diluted in serum-free DMEM. An equal volume of virus (200 PFU) was mixed with serum samples, and then the mixture or virus alone (as a positive control) was added to the confluent monolayers of HEp-2 cells for RSV plaque formation. Neutralizing antibody titers were defined as the reverse of serum dilution factors resulting in 60% plaque reduction.

Determination of antibody- and cytokine-secreting cells by ELISpot assay.

Splenocytes, bone marrow, and lung cells from individual mouse were prepared at 5 or 9 dpc for enzyme-linked immunospot (ELISpot) analysis. To assess F protein-specific antibody-secreting cells, multiscreen 96-well plates (Millipore) were coated with purified F protein (400 ng/ml) overnight at 4°C. After blocking, 10⁶ splenocytes and bone marrow cells were added in triplicate to F protein-coated plates, followed by incubation for 1 or 5 days. The plates were developed with biotinylated anti-mouse IgG and streptavidin-alkaline phosphatase (Southern), and the spots were visualized with diaminobenzidene substrates.

To detect IL-4 spot-forming cells, splenocytes (5 × 10⁵ cells/well) and lung cells (2 × 10⁵ cells/well) collected at 5 or 9 dpc were added to wells coated with anti-mouse IL-4 monoclonal antibodies (BD Biosciences, San Diego, CA) in the presence of 4 μg/ml RSV peptides F_92–106 (ELQLLMQSTPATNNR) (19), F_85–93 (KYKNAVTEL), and M2_82–90 (SYIGSINNI) (20). The cytokine spot-forming cells were developed with biotinylated mouse IFN-γ, IL-4 antibodies, and alkaline phosphatase-labeled streptavidin (BD Pharmingen). The spots were visualized with DAP substrates and counted by using an ELISpot reader (BioSys, Miami, FL).

Flow cytometry.

Bronchoalveolar lavage (BAL) fluids were obtained by infusing 1 ml of PBS into the lungs via the trachea using a 25-gauge catheter (Exelint International Co., Los Angeles, CA) at day 5 postchallenge. Lung single cells were isolated by homogenization of tissues and Percoll gradient (44 and 67%) centrifugation. Lung or spleen cells were stimulated with F or M2 peptide (4 μg/ml), respectively, for 5 h prior to staining of intracellular cytokines, and then the cells were fixed and permeabilized using a Cytofix/Cytoperm kit according to the manufacturer's instructions (BD Biosciences). Intracellular cytokines and surface markers for T cells or eosinophils were stained with antibodies for IFN-γ, IL-4 (eBioscience), TNF-α (BioLegend), CD45, CD3, CD4, CD8, or Siglec F (BD Biosciences). Stained cells were acquired on a FACSCanto flow cytometer (BD) and analyzed using FlowJo software (Tree Star, Inc., Ashland, OR).

To analyze the GC B cells, plasma cells, dendritic cell subsets, MLN, and BM were harvested. The cells were prepared to assess the expression of PNA⁺ germinal center (GC) B cells and CD138⁺ plasma cells from pregating mature B cells, including CD19⁺ IgD⁻ B220^int and CD19⁺ IgD⁻ B220⁺ cells, respectively, from CD45⁺ CD4⁻ CD8⁻ cells. Furthermore, the populations of CD11b⁺, CD103⁺, and CD8α⁺ DCs in the BAL, lungs, and MLN were analyzed from pregating DCs, including CD11c⁺ major histocompatibility complex class II (MHC-II)^hi and CD11c⁺ CD11b⁻ CD103⁻ cells, respectively, from CD45⁺ F4/80⁻ cells using the antibodies of surface markers such as CD19, IgD, F4/80, CD11b, CD11c, PNA (eBioscience), CD45R/B220, CD138, CD8α (BD), and CD103 (BioLegend).

Lung histopathology.

Lung tissues were fixed by 10% formalin in PBS and embedded in paraffin. Sections of 5 μm were stained with hematoxylin and eosin (H&E), periodic acid-Schiff stain (PAS), and hematoxylin/Congo red (H&CR) to assess histologic changes, mucin expression, and eosinophils, respectively. The sections around the airways and the peribronchial and interstitial spaces were scored for the degree of inflammation according to the following scale: 0 (none), 1 (mild), 2 (moderate), and 3 (severe). PAS-positive areas within the airway epithelium were detected in 25 randomly selected airways by Adobe Photoshop CS5.1 software. Infiltrated eosinophils were quantified by counting the H&CR-stained cells in the highly infiltrated regions in the airway, interstitial spaces of the lung sections.

Statistical analysis.

The data obtained using the mean ± the standard errors of the mean (SEM). The statistical analyzes were performed by a one-way analysis of variance (ANOVA) with Tukey or Dunnett's multiple-comparison test and two-way ANOVA in GraphPad Prism version 5 (GraphPad Software, Inc., San Diego, CA). A P value of <0.05 was regarded as statistically significant.

RESULTS

Intranasal immunization with RSV F VLPs induces a higher ratio of IgG2a antibodies than that of FI-RSV and live RSV.

The F protein of RSV is considered a promising target for developing vaccines against RSV. RSV F VLP was characterized in a previous study (11). We determined the incorporation and conformation of RSV F on VLP by using monoclonal antibody 131-2A (which is specific for a postfusion form of F) and palivizumab (which is reactive to both postfusion and prefusion forms of F). The reactivity of 131-2A (Fig. 1A) and palivizumab (Fig. 1B) antibodies suggests that the RSV F protein on VLPs is most likely presented in a postfusion conformation. Here, we focused on determining potential immune correlates contributing to protection without disease after mucosal (i.n.) immunization with F VLP in comparison to live RSV or FI-RSV.

FIG 1.

F VLP antigenicity and RSV F protein-specific antibody responses. (A) Reactivity of F-VLP to monoclonal antibody 131-2A (specific for postfusion conformation of F). (B) Reactivity of F-VLP to palivizumab (reactive to both the postfusion and the prefusion forms of F). The data represent ELISA optical density values at different concentrations of F VLP and M1 VLP (no RSV F) on the ELISA plates. (C to F) BALB/c mice (n = 5 per group) were i.n. immunized with 25 μg of F VLP, 5 μg of FI-RSV, and PBS (naive control) or i.n. inoculated with 10⁶ PFU of live RSV A2 (RSV). Serum samples were collected at 3 weeks after prime or boost administration. (C) The production of antibody specific for F and M1 protein was evaluated in boost immune sera using the RSV F protein and a pool of influenza M1 peptides as an ELISA coating antigen. (D and E) IgG2a and IgG1 isotype antibodies specific for RSV F protein. (F) Ratios of IgG2a to IgG1 isotype antibodies. The results are representative of two independent experiments and are presented as means ± the SEM. Statistical significance was determined using two-way ANOVA in GraphPad Prism. ***, P < 0.001; ND, not detected.

RSV-specific antibody production was determined in the sera of prime and boost immune mice using the purified RSV F protein as an ELISA coating antigen (Fig. 1C, D, E, and F). IgG antibody specific for influenza M1 protein used for VLP core formation was not detected at a meaningful level (Fig. 1C). F VLP prime immune sera showed significantly higher levels of IgG2a antibody specific for F protein compared to those of live RSV or FI-RSV immune mice (Fig. 1D). IgG2a antibody was further increased after boost immunization with F VLPs, which is also significantly higher than those with FI-RSV (Fig. 1D). In contrast, F VLP immune mice did not significantly induce IgG1 antibodies specific for RSV F, whereas FI-RSV immune mice induced higher levels of IgG1 than IgG2a (Fig. 1D and E). The live RSV group showed mixed IgG1 and IgG2a antibodies at high levels (Fig. 1D and E). As a result, the F VLP group showed highest ratios of IgG2a/IgG1 of approximately 8 to 10 (Fig. 1F). Although both live RSV infection and FI-RSV immune mice showed low IgG2a/IgG1 ratios, the live RSV group exhibited higher IgG2a/IgG1 ratios than the FI-RSV group (Fig. 1F). Thus, these results indicate that i.n. immunization with F VLP induces higher ratios of IgG2a T helper type 1 (Th1) isotype antibody responses compared to those with live RSV and FI-RSV.

Comparative studies of protective efficacy among F VLP, FI-RSV, and live RSV immune mice.

Since neutralizing antibodies play an essential role in conferring protection, we determined neutralizing antibody titers of immune sera by using a plaque reduction assay. Live RSV reinfection sera showed significantly higher titers of neutralizing activity ∼1,024 (≥10 log₂) (Fig. 2A). F VLP and FI-RSV immune sera displayed neutralization antibody titers of ∼8 log₂, which is significantly higher than those observed with unimmunized naive immune sera (Fig. 2A). These results suggest that F VLP i.n. immunization induces significant levels of RSV neutralizing activity.

FIG 2.

RSV neutralizing activity, body weight changes, and lung viral clearance. (A) RSV neutralizing activity. Sera (n = 5 per group) were collected from each group of immunized and PBS control mice at week 3 after boost immunization. The neutralization titers were determined in a plaque reduction assay. The titer was defined as the serum dilution that inhibited virus plaque formation by 60%. (B and C) Body weight changes. Immunized and naive (PBS) mice were i.n. infected with RSV A2 (2 × 10⁶ PFU/mouse) 12 weeks after boost immunization. (D) Lung virus titers were determined from individual mice at 3, 5, and 9 dpc. F VLP, F VLP immune mice; FI-RSV, FI-RSV immune mice; RSV, prior live RSV reinfected mice; PBS, unimmunized control mice. The results are representative of two independent experiments and are presented as means ± the SEM. Statistical significance was determined using two-way ANOVA in GraphPad Prism. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Body weight changes and lung viral clearance are important parameters in assessing protective efficacy after RSV vaccination and challenge in a mouse model. Mice treated with F VLP, live RSV, FI-RSV, and PBS were challenged with RSV (2 × 10⁶ PFU) at 12 weeks after a boost immunization, and the weight changes of the mice were observed for 9 days (Fig. 2B and C). Severe weight loss was observed in FI-RSV immune mice (ca. 18%) compared to F VLP immune mice (<4%) (Fig. 2B). Even live RSV reinfection caused substantial weight loss (ca. 9% loss) upon RSV challenge (Fig. 2C). Naive mice showed a pattern of weight loss similar to that observed in live RSV reinfection mice and then slowly recovered (Fig. 2B).

Lung tissues were collected from individual mice at 3, 5, and 9 dpc, and viral loads were determined by an immunoplaque assay (Fig. 2D). RSV titers of all group were showed higher levels of viral loads at day 5 postchallenge than those an early time point of 3 days dpc. Lung viral loads were cleared by 9 dpc (Fig. 2D). Naive mice exhibited highest lung viral loads at days 3 and 5 postchallenge, whereas other immune mice (F VLP, FI-RSV) showed significantly low virus titers, with live RSV exhibiting the most effective viral clearance, indicating a correlation with the induction of serum neutralizing antibodies.

F VLP i.n. immunization is effective in inducing RSV-specific antibody-secreting cells, as well as germinal center B and plasma cell responses.

We determined mucosal antibody responses in lung extracts (Fig. 3A and B) and BALF (Fig. 3C and D) at 5 dpc. High levels of RSV F-specific IgG and IgA antibodies were produced in BALF and lung homogenates by immunization with F VLP, FI-RSV, or live RSV reinfections compared to those in naive mice with RSV infection (Fig. 3A to D).

FIG 3.

Mucosal RSV F-specific antibodies, antibody-secreting cell responses, and germinal center (GC) B cells. Immune mice (n = 5 per group) were i.n. infected with live RSV A2 (2 × 10⁶ PFU) 12 weeks after boost immunization. BALF, lung homogenates, bone marrow, and splenocytes were harvested from each mouse at 5 dpc. The levels of IgG and IgA antibodies specific for RSV F protein were determined from lung homogenates (A and B) and BALF (C and D), respectively, by ELISA. (E) F-specific antibody-secreting cells in bone marrow. (F) F-specific antibody-secreting cells in spleens. Bone marrow cells and splenocytes were added to plates coated with RSV F protein (400 ng/ml), followed by incubation for 1 day (bone marrow) or 5 days (Spleens). Antibody-secreting cell spots were detected by ELISpot analysis. (G) GC phenotypic B cells (PNA⁺ B220⁺ CD138⁻) in bone marrow were determined flow cytometry. The results are presented as means ± the SEM, and statistical significance was determined using two-way ANOVA in GraphPad Prism. ***, P < 0.001; **, P < 0.01. The results are representative of two independent experiments.

The induction of long-lived plasma cells and memory B cells is a major goal for vaccination. Bone marrow and spleen cells collected at 5 dpc were analyzed for RSV F-specific antibody-secreting cell responses using ELISpot. Significant levels of F-specific antibody-secreting cells were observed in BM cells from immunized (F VLP and FI-RSV) or reinfected (RSV) mice after 1 day of culture (Fig. 3E). In addition, substantial levels of F protein-specific antibody-secreting cells were also detected in spleens on day 5 of culture (Fig. 3F) compared to those in unvaccinated naive mice at day 5 postinfection. We further analyzed B cell phenotypes in bone marrow (Fig. 3G). GC-derived B cells (PNA⁺ B220⁺ CD138⁻ CD19⁺ IgD⁻ phenotypes) were found to be higher in bone marrow (P < 0.001) from F VLP mice than in bone marrow from primary RSV-infected mice (Fig. 3G). These results indicate the generation of plasma, GC-derived B cells, and memory B cells by F VLP immunization.

F VLP i.n. immunization does not induces pulmonary inflammation upon RSV infection.

Developing a safe RSV vaccine that should not cause RSV disease has been a challenge. We examined pulmonary histopathologic changes upon RSV infection for assessing the safety of RSV vaccines (Fig. 4). The lungs of FI-RSV immune mice displayed most severe symptoms of alveolitis and infiltrates in the airways and interstitial spaces (Fig. 4A), as well as PAS-positive mucus production (Fig. 4B). The FI-RSV group showed significantly higher inflammation scores around the peribronchial airways and interstitial spaces. PAS-positive mucus production was significantly higher at 3 dpc than that at 5 dpc and was prolonged until 9 dpc in FI-RSV immune mice (Fig. 4C to E). Mice with prior RSV reinfection exhibited higher scores of infiltrates around the airways and interstitial spaces than did F VLP immune mice (Fig. 4C to E). Unimmunized naive (PBS control) mice after RSV infection showed a certain degree of inflammation in the airways and interstitial spaces, but mucus production was not observed in the airways (Fig. 4C to E). In contrast, no overt sign of cellular infiltration and mucus production around the airway and interstitial spaces was observed in F VLP immune mice (Fig. 4C to E).

FIG 4.

F VLP immune mice do not show pulmonary inflammation and mucus production upon RSV infection. Lung tissues (n = 5 per group) were collected from individual mice and tissue section were stained with hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS) to assess pulmonary histopathologic changes at 5 dpc. (A) H&E staining was performed on day 5 after RSV infection. Scale bars represent 100 μm (×100 magnification). (B) PAS staining at 3, 5, and 9 dpc. Scale bars represent 100 μm (×200 magnification). (C and D) H&E-stained lung sections were scored for inflammation responses on a scale of 0 to 3 by diagnostic criteria in the airways and interstitial spaces. (E) Stained lung sections were scored for bronchiolar mucus expression as the percentages of PAS positives. The results are given as means ± the SEM, and statistical analysis was performed using one-way ANOVA with Tukey's multiple-comparison test in GraphPad Prism. ***, P < 0.001; **, P < 0.01; *, P < 0.05. UI, uninfected. The results are representative of two independent experiments.

We examined eosinophils infiltration from lung tissues using H&CR staining (Fig. 5A). FI-RSV immune mice showed massive influx of eosinophils compared to F VLP immune mice (P < 0.001, Fig. 5A to D). The H&CR staining results were consistent with high levels of Siglec F⁺ eosinophils from gated CD45⁺ F4/80⁺ CD11b⁺ CD11c⁻ cells, as determined by flow cytometry analysis of BAL fluid and lung cells at 3 dpc (Fig. 5B to D), as well as the high eotaxin levels in lungs from the FI-RSV group but not the F VLP group (Fig. 5E and F). No significant levels of Siglec F⁺ eosinophils and eotaxin were detected in the F VLP group (Fig. 5B to F). The live RSV group displayed a moderate level of Siglec F⁺ eosinophils and eotaxin in BAL samples, as well as H&CR-positive staining. Together, these results suggest that F VLP i.n. immunization is effective in conferring protection against RSV disease involving pulmonary immunopathology, mucus production, and eosinophilia upon RSV infection in a mouse model.

FIG 5.

F VLP i.n. immunization does not induce eosinophilia in the lung. Lung tissues were collected from individual mice from an additional set of experiments (n = 5 per groups). (A) Lung tissue sections were stained with hematoxylin and Congo red (H&CR) to assess pulmonary eosinophilia at 5 dpc. High-magnification images indicate the appearance of eosinophils. (B and C) Populations and numbers of Siglec F⁺ eosinophils gated from granulocytes (CD11b⁺ F4/80⁺ CD45⁺) at 3 dpc as determined by flow cytometry in BAL fluid. (D) Numbers of Siglec F⁺ eosinophils in the lungs at 3 dpc. (E and F) Eotaxin production in BALF (5 dpc) and lung homogenates (5 dpc), as determined by ELISA. The results are presented as means ± the SEM, and statistical analysis was performed using one-way ANOVA with Tukey's multiple-comparison test or two-way ANOVA in GraphPad Prism. ***, P < 0.001; **, P < 0.01. UI, uninfected. The results are representative of two independent experiments.

Th2 cytokines are not induced at local and systemic sites by F VLP immunization.

To evaluate Th1 and Th2 immune responses, Th1 and Th2 cytokines in BALF were determined in immunized mice at 5 dpc with RSV A2. Unimmunized naive (PBS control) mice with RSV infection showed high levels of both Th1 (IFN-γ, TNF-α, and IL-12p70) and Th2 (IL-4) cytokines (Fig. 6A and B). BALF samples of F VLP group detected high levels of IFN-γ, TNF-α, and IL-12p70 (Th1) cytokines but not IL-4 and IL-13 (Th2) cytokines. The live RSV group showed substantial levels of mixed Th1 (TNF-α and IL-12p70) and Th2 IL-4 cytokines, FI-RSV immunization of mice induced the highest levels of Th2 cytokines IL-13 and IL-4 (Fig. 6A and B).

FIG 6.

F VLP i.n. immunization elicits Th1 cytokines but not Th2 cytokines upon RSV infection. BAL fluids, lung, and spleen cells were collected from immunized or naive (PBS) mice (n = 5 per group) at 5 dpc (A to C) or at 9 dpc (D to F), and the cytokine levels were determined by ELISA. (A) Levels of IFN-γ, IL-4, and IL-13 in BALF. (B) Levels of TNF-α and IL-12p70 in BALF. Spot-forming cells (SFCs) secreting IL-4 and IFN-γ in lungs and spleens were detected by ELISpot analysis by stimulating RSV F peptides on plates coated with capture antibodies for IL-4 and IFN-γ at 5 dpc (C) and at 9 dpc (D to F). (F) Comparison of the ratio of IFN-γ to IL-4 by the stimulation of F peptides as CD8 and CD4 T cell epitope in the lung at 9 dpc. The results are presented as means ± the SEM, and statistical analysis was performed using one-way ANOVA with Tukey's multiple-comparison test or two-way ANOVA in GraphPad Prism. ***, P < 0.001; *, P < 0.05. UI, uninfected.

We further assessed the cytokine-secreting cells locally in lungs and systemically in spleens. The FI-RSV group showed highest levels of IL-4-secreting cell spots in both lungs and spleens following in vitro stimulation with F peptides (P < 0.001, Fig. 6C and D). The live RSV group exhibited moderate spot numbers secreting IL-4 in spleens systemically (Fig. 6C) but not in lungs (Fig. 6D). Mice with F VLP immunization or PBS showed the lowest levels of IL-4-secreting cell spots in the lungs and spleens at both 5 and 9 dpc (Fig. 6C and D). In comparison experiments of cells secreting IFN-γ in vitro, the F VLP group induced significantly higher levels of IFN-γ-secreting cells after stimulation with F_85–93, a CD8 epitope, but not with F_51–66, a CD4 epitope (Fig. 6E). The live RSV group showed a reverse pattern of IFN-γ-secreting cells in response to CD4 and CD8 epitopes of F compared to that of the F VLP group (Fig. 6E and F). FI-RSV immune mice elicited IFN-γ-secreting cells at high levels after stimulation with both F_51–66 and F_85–93 peptides (Fig. 6E). The ratios of IFN-γ-producing to IL-4-producing CD4 and CD8 T cells in the F VLP group showed similar patterns to those observed in naive mouse primary RSV infections (Fig. 6F). These results suggest that high levels of Th2 cytokines locally and systemically in FI-RSV mice, but not in F VLP immune mice, likely contribute to lung inflammatory disease. Also, inflammatory Th1 and Th2 cytokines (IFN-γ, TNF-α, IL-4, and IL-12p70) in BALF might be responsible for a moderate weight loss in live RSV reinfection and naive mice after RSV infection.

Lung CD4⁺ T cells producing IL-4 and TNF-α show a correlation with RSV disease.

Details on the cellular phenotypes producing cytokines in the lung were determined by intracellular cytokine staining flow cytometry. A gating strategy of intracellular cytokine staining flow cytometry is shown to determine phenotypes of T cells producing Th1 or Th2 cytokines (see Fig. 11). Approximately 3.6-, 3.7-, and 2-fold higher levels of CD4 T cells were found to be recruited into the lung of FI-RSV immune mice at 9 dpc compared to those in naive mice, F VLP immune mice, and live RSV immune mice, respectively (Fig. 7A). The cellularity of CD4 T cells was lower in the lungs and BAL fluid from F VLP immune mice than in those from FI-RSV or live RSV immune mice (Fig. 7A and C). We analyzed cytokine-secreting lung CD4 T cells at a later time point (9 dpc) to better understand adaptive T cell responses (Fig. 7B and D). The highest levels of IL-4⁺ and TNF-α⁺ CD4⁺ T cells were recruited into the lung of FI-RSV immune mice but not in the lungs of naive or F VLP immune mice. The highest levels of TNF-α⁺ and IL-4⁺ CD4⁺ T cells, as well as substantial IFN-γ⁺ CD4 T cells, were recruited into the lung (Fig. 7B) and BAL fluid (Fig. 7D) from FI-RSV immune mice. Live RSV reinfection and PBS control mice showed a similar pattern of moderate levels of TNF-α⁺, IL-4⁺, and IFN-γ⁺ CD4 T cells, whereas F VLP immune mice exhibited lower levels of cytokine-secreting CD4 T cells (Fig. 7B and D). Therefore, these results suggest that excess infiltration of CD4⁺ T cells producing IFN-γ, TNF-α, and IL-4 into the lungs might have contributed to pulmonary RSV disease in FI-RSV immune mice.

FIG 11.

Flow cytometry gating strategy for effector T cells and subsets of dendritic cells. (A) Gating strategy for T cells secreting cytokines. Effector CD4⁺ and CD8⁺ T cell populations secreting cytokines (IFN-γ, IL-4, and TNF-α) were gated by phenotypic CD4 and CD8 markers, together with intracellular cytokine staining fluorescent antibodies. (B) Gating strategy to identify CD11b⁺, CD103⁺, or CD8α DCs from BAL fluid, lungs, and MLN. The percentages of CD11b⁺ and CD103⁺ DC populations gated from CD11c⁺ MHC-II⁺ CD103⁻ and CD11c⁺ MHC-II⁺ CD11b⁻ cells, respectively, at 3 dpc are shown.

FIG 7.

Intracellular cytokine staining of pulmonary CD4⁺ T cells in immune mice. Lung and BAL cells were isolated from mice (n = 5 per group) at 9 dpc, and then the cells were stimulated with RSV F peptide to investigate the levels of CD4⁺ T cells secreting intracellular cytokines by flow cytometry. (A) Cellularity of lung CD4⁺ T cells at 9 dpc. (B) Lung CD4⁺ T cells secreting IFN-γ⁺, IL-4, or TNF-α. (C) Cellularity of BAL CD4⁺ T cells at 9 dpc. (D) BAL CD4⁺ T cells secreting IFN-γ⁺, IL-4, or TNF-α at 9 dpc. The results are presented as means ± the SEM, and statistical analysis was performed using two-way ANOVA in GraphPad Prism. ***, P < 0.001; **, P < 0.01; *, P < 0.05. UI, uninfected.

F VLP immune mice induce high levels of F-specific IFN-γ⁺ effector CD8⁺ T cells in the lung and BAL fluid.

CD8 T cell immunity to RSV is characterized by highly immunodominant epitopes in M2 and F proteins. Thus, we determined CD8⁺ effector T cells producing TNF-α and IFN-γ cytokines in the lung and BAL fluid after stimulation with RSV F_85–93 or M2_82–90 peptide by intracellular cytokine flow cytometry analysis (Fig. 8). To determine whether F VLP vaccination induces more effector CD8 T cells than live RSV reinfection before challenge, we analyzed RSV-specific effector CD8 T cell responses in the lung at day 7 after boost immunization (Fig. 8). F VLP immune mice induced higher levels of F_85–93-specific CD8⁺ T cells producing IFN-γ⁺ than live RSV reinfection (Fig. 8A and B). In contrast, live RSV reinfection elicited significantly higher levels of M2_82–90-specific effector CD8⁺ T cells in the lungs than those by F VLP immunization (Fig. 8C and D).

FIG 8.

F VLP induces a higher level of F_85–93-specific CD8⁺ T cells secreting IFN-γ than live RSV. Effector CD8 T cells producing IFN-γ were determined in the lungs 7 days after boost immunization of mice (n = 3) before RSV challenge. Lung cells were in vitro stimulated with F_85–93 or M2_82–90 peptide. (A) Flow cytometry gating populations of F_85–93-specific CD8⁺ T cells producing IFN-γ. (B) Percentages of effector CD8⁺ IFN-γ⁺ T cells as determined by F_85–93 stimulation. (C) Flow cytometry gating populations of M2_82–90-specific CD8⁺ T cells producing IFN-γ. (D) Percentages of effector CD8⁺ IFN-γ⁺ T cells by M2_82–90 stimulation. The results are presented as means ± the SEM, and statistical significance was determined using one-way ANOVA with Tukey's multiple-comparison test performed in GraphPad Prism. ***, P < 0.001.

After challenge, F VLP immune mice showed higher levels of IFN-γ⁺ CD8⁺ T cells and TNF-α⁺ CD8⁺ T cells in the lungs (9.45 and 6.87%, respectively; Fig. 9A and D) and BAL fluid (11.4 and 6.07%, respectively; Fig. 9B and E) upon stimulation of F_85–93 peptide at 9 dpc. In contrast, highest levels of IFN-γ⁺ CD8⁺ T cells and TNF-α⁺ CD8⁺ T cells after M2_82–90 peptide stimulation were observed in the lung from live RSV infection mice (35.70 and 30.20%, respectively, Fig. 9C and F), suggesting that RSV challenge can boost M2-specific CD8 T cell responses despite effective clearance of viral antigens after challenge. These results indicate that F VLP vaccination leads to the induction of F-specific CD8 T cells producing IFN-γ, which is different from live RSV reinfection inducing much higher levels of M2-specific CD8 T cells producing IFN-γ and TNF-α⁺. It is significant that nonreplicating F VLP immune mice can generate RSV F-specific IFN-γ-producing CD8 T cell responses at a higher level than live RSV infection.

FIG 9.

F VLP and live RSV have differential effects on effector CD8 T cells after challenge. Lung and BAL cells were collected from mice (n = 5 per group) at 9 dpc and stimulated with RSV F_85–93 and M2_82–90 peptide to assess the CD8⁺ T cells secreting cytokines by intracellular flow cytometry assays. (A) Representative flow cytometry plots of gating lung CD8 T cells (left) and F_85–93-specific CD8⁺ T cells expressing IFN-γ (middle) or TNF-α (right) after RSV F_85–93 peptide stimulation. (B) BAL CD8⁺ T cells secreting F_85–93-specific IFN-γ (left) and TNF-α (right) after RSV F_85–93 peptide stimulation. (C) Lung CD8⁺ T cells secreting IFN-γ (left) and TNF-α (right) specific for M2_82–90 peptide. The populations of CD8⁺ T cell secreting cytokines as percentages of IFN-γ⁺ and TNF-α⁺ gated cells specific for F_85–93 (D and E)and M2_82–90 (F) in the lungs and BAL were determined. The results are presented as means ± the SEM, and statistical analysis was performed using two-way ANOVA in GraphPad Prism. ***, P < 0.001; **, P < 0.01; *, P < 0.05. UI, uninfected.

F VLP immune mice induce high levels of CD103⁺ and CD8α⁺ dendritic cells.

The dendritic cells (DCs), including CD11b⁺, CD103⁺, and CD8α⁺ subtypes, were shown to play a key role in inducing viral immunity during infection (21–24). The lung-derived DCs, such as CD11b⁺ and CD103⁺ DCs, migrate into the draining MLN and contribute to the induction of an adaptive T cell response after viral infection. We investigated different DC subtypes in BAL fluid, lungs, and MLN. As shown in a DC gating strategy (Fig. 10 and 11), CD11b⁺ DCs (CD11c⁺ MHC-II⁺⁻ CD103⁻ F4/80⁻ CD11b⁺ DCs) were observed at significantly higher levels at 3 dpc in BAL fluid and lungs from live RSV reinfection, FI-RSV immune mice, and PBS mice compared to samples from F VLP immune mice (Fig. 10A). Interestingly, at 3 dpc lung resident or migrating CD103⁺ DCs (CD11c⁺ MHC-II⁺ F4/80⁻ CD11b⁻ CD103⁺ DCs) were significantly higher in the BAL fluid and lungs from the F VLP group than those from the live RSV and FI-RSV groups (Fig. 10B). Accordingly, F VLP immune mice showed the highest ratios of CD103⁺ CD11b⁺ DCs in the BAL fluid and lungs (Fig. 10C). Also, MLN from F VLP immune mice displayed higher levels of CD103⁺ DCs (5 dpc, Fig. 10D) and resident CD8α⁺ DCs (CD11c⁺ CD103⁻ CD11b⁻ F4/80⁻ CD8α⁺ DCs) at 3 and 5 dpc (Fig. 10E) compared to those from FI-RSV immune or live RSV reinfection mice. Thus, the results support a possibility that F VLP immune mice induce IFN-γ-producing CD8⁺ T cell responses through upregulation of CD103⁺ and CD8α⁺DCs in the lung, BAL fluid, and MLN after RSV infection.

FIG 10.

F VLP immune mice induce CD103⁺ and CD8α⁺ DCs in lungs and draining lymph nodes. Lungs, BAL fluid, and MLN cells were collected from mice (n = 5 per group). (A and B) Percentage of CD11b⁺ and CD103⁺ DC populations gated from CD11c⁺ MHC-II⁺ CD103⁻ and CD11c⁺ MHC-II⁺ CD11b⁻, respectively, at 3 dpc. (C) Ratio of CD103⁺ to CD11b⁺ cells at 3 dpc. (D) CD103⁺ DCs in MLN. (E) Percentage of CD8α⁺ DCs gated from CD45⁺ F4/80⁻ CD11c⁺ CD11b⁻ CD103⁻ in MLN. The results are presented as means ± the SEM, and statistical analysis was performed using two-way ANOVA in GraphPad Prism. ***, P < 0.001; **, P < 0.01; *, P < 0.05. UI, uninfected.

DISCUSSION

Developing an effective and safe RSV vaccine has been a difficult challenge, partially due to the potential risks of causing vaccine enhanced disease, as observed in FI-RSV vaccine trials (2). Repeated RSV infections are common even in the absence of significant antigenic differences (6, 8). In the present study, we investigated the immunogenicity and possible immune correlates contributing to protection and/or disease after intranasal immunization of mice with F VLPs or FI-RSV in comparison to live RSV reinfection. FI-RSV, F VLP, or live RSV could induce RSV neutralizing activities in sera contributing to lung viral clearance. Mice with RSV reinfection showed highest RSV neutralizing activities and lowest lung viral loads, indicating that efficacy of lung viral clearance is correlated with RSV neutralizing titers. Regardless of lung viral clearance, cellular phenotypes of inducing immune responses are dependent on the platforms of RSV vaccines. In particular, IFN-γ⁺ CD8 T cells and CD103⁺ DCs in lungs were characteristic cellular immune types by F VLP i.n. immunization that did not induce an overt sign of pulmonary inflammation upon RSV infection. In addition, differential properties of RSV vaccines in stimulating DCs and macrophages might have an impact on potential outcomes of inflammatory disease and a distinct type of T cell immunity.

In a previous study, depletion of both CD4 and CD8 T cells resulted in no significant signs of illness during primary RSV infection despite an extensive delay in lung viral clearance (25), suggesting that cellular immune responses in response to RSV infection rather than viral loads are the main contributors for inflammatory pulmonary disease in mice. In support of this observation, previous studies demonstrated that FI-RSV immune mice and cotton rats well controlled RSV lung viral loads after infection despite enhanced RSV disease (5, 26–28). FI-RSV immune mice showed severe weight loss upon RSV infection despite low lung viral loads compared to naive mice exhibiting high lung viral loads. Also, mice with prior RSV reinfections could not prevent weight loss, displaying similar morbidity of weight loss as seen in the primary infection. Consistent with weight loss, FI-RSV immune mice exhibited most severe pulmonary histopathology, including mucus production. However, it is not clear how weight loss in mice is related to RSV disease in humans, implicating a limitation in the studies of RSV pathogenesis in mouse animal models. Pulmonary inflammation may be a more relevant parameter in assessing vaccine enhanced RSV disease. We observed that live RSV reinfections induced substantial inflammatory RSV disease, which is consistent with a previous study reporting the appearance of peribronchiolar lymphocyte aggregates after reinfections (25). Host immune responses preventing RSV disease after RSV vaccination and infection remain poorly understood probably because there is no licensed RSV vaccine. Most importantly, F VLP immunization did not cause such pulmonary inflammation of RSV disease. Therefore, comparative analysis would provide informative insights into plausible immune correlates conferring safe protection against RSV after vaccination.

F VLP immunization was found to be most effective in inducing IgG2a isotype antibodies predominantly, a characteristic Th1 type immune response. Meanwhile, FI-RSV immunization induced more IgG1 than IgG2a antibody, and live RSV infection raised similar levels of IgG2a and IgG1 antibodies. This dichotomy of immune responses among immune mouse groups was reflected by cytokine profiles in the lung milieu. FI-RSV immune mice showed highest levels of IL-13 in the airways, which correlated with Siglec F⁺ eosinophil infiltration. IL-13 was shown to be sufficient for eosinophilia in mice vaccinated with RSV G glycoprotein after RSV challenge (29), which is an IL-4 cytokine-independent mechanism (30). The most striking difference among groups was found in the IL-4-secreting cell spots. F VLP immunization did not induce IL-4-producing cell spots in lungs and spleens. In contrast, FI-RSV immune mice showed highest levels of IL-4 spot-forming cells in both lungs and spleens. Anti-IL-4 antibody treatment during FI-RSV immunization was shown to reduce illness and enhance cytotoxic T cell activity in mice (31). It is likely that Th2 biased-priming during immunization might be predisposing lung inflammatory illness upon RSV infection later. Noticeably, RSV reinfection of mice induced high levels of IFN-γ spot-forming cells locally and systemically. Previously, it was also shown that high levels of IFN-γ production can contribute to RSV disease and airway hyper-responsiveness (32). F VLP immune mice showed a moderate level of IFN-γ-producing but not IL-4-producing cells in lungs. Unvaccinated naive mice exhibited high levels of both Th1 and Th2 cytokines (IFN-γ, IL-4, and TNF-α) in the airways after RSV infection. Therefore, we provide evidence here that the induction of regulated Th1 immunity may be needed to prevent severe RSV-related disease.

Cellular phenotypes producing Th1 and Th2 cytokines would be important determinants for protection after RSV vaccination. FI-RSV immune mice induced the highest levels of CD4 T cellularity in lungs, as well as of CD4 T cells producing IL-4 or TNF-α. This is consistent with previous studies that CD4 T cells secreting Th2 cytokines are responsible for eosinophilia and vaccine-enhanced RSV-related disease (3, 33). Live RSV infection is often used as a positive protection control for testing experimental RSV vaccines in preclinical studies. In contrast to FI-RSV, live RSV reinfection and challenge induced the highest levels of IFN-γ producing CD4 T cells. IFN-γ-deficient mice that were immunized with a recombinant vaccinia virus expressing RSV F were experienced less weight loss than wild-type controls (32). F VLP immune mice showed a low level of CD4 T cells producing IFN-γ. It is possible that high levels of CD4 T cells producing IFN-γ observed in the live RSV group might have contributed to substantial pulmonary inflammation around airways and interstitial spaces, as well as a degree of weight loss. Therefore, high levels of IFN-γ Th1 CD4 T cells might have both beneficial protective roles and detrimental inflammation effects during RSV infections.

Interestingly, intracellular cytokine staining lung cells in the present study demonstrated that F VLP immune mice induced the highest level of CD8 T cells producing IFN-γ. This suggests that F VLP i.n. immunization might induce higher levels of F-specific CD8 T cells but not CD4 T cells making IFN-γ compared to live RSV reinfection. M2-specific effector CD8 T cells secreting IFN-γ or TNF-α were significantly increased to ca. 30% in live RSV-reinfected mice, which are significantly higher than those in RSV primary-infected mice. Our results suggest that viral antigen-specific T cell types producing Th1 or Th2 cytokines can be an important factor in the outcome of RSV protection and/or disease. Since F VLP immune mice displayed less weight loss and did not show signs of pulmonary inflammation and mucus production compared to RSV reinfection, the level of IFN-γ⁺ CD8 T cells can be more protective in the prevention of ERD than the level of IFN-γ⁺ CD4 T cells. Also, RSV-specific CD8 T cells were reported to inhibit Th2-mediated pulmonary eosinophilia by unknown mechanisms (3).

It is also not clear how F VLP immune mice could induce RSV F-specific IFN-γ CD8 T cells better than other groups. In the line with IFN-γ CD8 T cells, CD8α⁺ DCs were found at the highest level in MLN from F VLP immune mice. CD8α⁺ DCs are known to be highly effective in inducing CD8 cytotoxic T cells (22), Th1 priming, cross-presentation, and IFN-γ secretion (34). In particular, CD8α⁺ DCs were demonstrated to be the principle DC subsets priming CD8 cytotoxic T cell immunity to infection by influenza virus, herpes simplex virus, and vaccinia virus (23). In line with this, high levels of CD8α⁺ DCs that were found in the MLN of F VLP immune mice might be a major contributor to inducing IFN-γ CD8 T cells after RSV challenge. Also, compared to the live RSV and FI-RSV group, F VLPs appeared to be more effective in recruiting CD103⁺ DCs in the BAL fluid and lungs after RSV challenge. Skin-derived CD103⁺ DCs were shown to be effective in the cross-presentation of viral antigens (21). It is possible that CD103⁺ DCs carry viral antigens and present them to MLN resident DCs such as CD8α⁺ DCs and lymphocytes. The underlying mechanisms for how these subsets of DCs impact the outcome of protection and/or disease after RSV challenge are as yet unknown. Different Th1 or Th2 cytokines and chemokines in the lung microenvironments of different immune mice are expected to significantly influence the possible outcome of protection or disease, as presented here.

In an environment of producing IL-6, lung DCs were shown to be biased toward inducing Th2 responses (35, 36). In addition, TNF-α cytokine is suggested to contribute to weight loss in mice (37). The IFN-γ CD4 Th1 cytokine was also reported to contribute to have beneficial and detrimental effects on RSV protection and disease after vaccination and infection (32). FI-RSV immune mice showed high levels of both IL-4 and IFN-γ CD4 T cells in lungs and spleens upon RSV F_51–66 peptide stimulation. Live RSV reinfection, but not F VLP vaccination, induced high levels of F_51–66 peptide-specific IFN-γ CD4 T cells, as well as M2_82–90 peptide-stimulated IFN-γ CD8 T cells. In relating to severe to moderate disease in FI-RSV and live RSV reinfection mice in the present study, IFN-γ-deficient mice were demonstrated to exhibit less weight loss than wild-type mice upon RSV F vaccination and infection (32). These findings in this and other studies are consistent with previous reports indicating that RSV M2_82–90 peptide-specific IFN-γ CD8 T cells play a role in causing CD8 T cell-mediated immunopathology, including weight loss in mice upon RSV challenge (38, 39). RSV F is known to have Toll-like receptor 4 agonist (40). RSV G exhibits the mimicry activity of CX3C chemokine binding to CX3CR1, which is known to be associated with enhanced disease and pulmonary eosinophilia, as well as leukocyte chemotaxis (41, 42). The intrinsic properties of different vaccine components and platforms might have significant impact on the outcome of a pattern of T cell immunity and inflammatory RSV disease, but further studies are needed to support this possible correlation.