Inactivated Influenza Vaccine That Provides Rapid, Innate-Immune-System-Mediated Protection and Subsequent Long-Term Adaptive Immunity

Inoculation with split IAV vaccine formulated with R₄Pam₂Cys confers rapid antiviral protection.

Our previous work (35) demonstrated that treatment of animals with Pam₂Cys in the absence of antigen at up to 7 days prior to influenza virus challenge results in a reduction in viral loads postchallenge. To assess whether these protective effects extended to the use of an influenza vaccine formulation containing R₄Pam₂Cys, we used a split influenza virus vaccine derived from PR8 influenza virus. This vaccine preparation contained proteins corresponding to the molecular weights of the major viral surface glycoproteins, as well as the internal nucleoprotein (NP) and the M1 matrix protein (see Fig. S1 in the supplemental material). By determining lung viral titers in mice challenged within 7 days of vaccination, i.e., before any adaptive immune response would have a significant effect, we were able to determine any protective effects mediated by the innate immune system (Fig. 1A).

The results (Fig. 1B) show that mice inoculated with a single dose of split virus vaccine formulated with R₄Pam₂Cys and challenged after 24 h with PR8 or X31 virus had significantly reduced lung viral titers 5 days postinfection compared to naive mice or those inoculated with split virus in the absence of R₄Pam₂Cys. A greater reduction in viral titers in mice inoculated with split virus plus R₄Pam₂Cys was detected when vaccinations were carried out 3 days prior to challenge. These protective effects were still evident 7 days following vaccination.

To evaluate protection against a lethal viral infection, vaccinated animals were challenged with a higher dose (500 PFU) of PR8 (Fig. 2A). While lung viral titers were not reduced in mice vaccinated in the presence of R₄Pam₂Cys and challenged a day after vaccination (Fig. 2B), significant reductions in viral titers in these mice compared to the control animals were detected when vaccinations were carried out 3 to 7 days prior to challenge.

Mice vaccinated with split virus plus R₄Pam₂Cys 3 days before lethal viral challenge were also protected from the substantial weight loss (Fig. 2C) and development of severe disease symptoms associated with infection compared to mice that received split virus alone, which by day 7 all had to be culled, having reached the defined humane endpoint (Fig. 2D). No protective effects were observed in similarly vaccinated TLR2^−/− mice, indicating that protection mediated by R₄Pam₂Cys was dependent on recognition and signaling operating through TLR2. These observations demonstrate that an influenza vaccine delivered with Pam₂Cys can induce rapid protection to reduce the impact of infection irrespective of the viral subtype. These protective effects are also achieved with a lesser dose of Pam₂Cys than was previously used to confer antiviral protection (35).

Antibody responses induced by vaccination of minimal doses of split IAV vaccine formulated with R₄Pam₂Cys.

The induction of subsequent adaptive (antigen-specific) immunity was studied by measuring CD8⁺ T cell and antibody responses at several time points after vaccination. Animals were also challenged 35 days after vaccination, and recall CD8⁺ T cell responses and viral loads were analyzed 5 days postinfection (Fig. 3A). To evaluate antibody responses, mice were inoculated via the intranasal, subcutaneous, or intramuscular route. A low dose of split vaccine alone was unable to induce titers of antibody as high as those seen with adjuvanted split vaccine irrespective of the inoculation route used (Fig. 3B). In contrast, administration of the vaccine with R₄Pam₂Cys resulted in significant titers of antibody, with an ~600-fold improvement in titers following inoculation via the intranasal route and an ~20-fold to 60-fold improvement following inoculation by the subcutaneous or intramuscular route.

Serum antibody isotype profiles induced by intranasal inoculation of R₄Pam₂Cys-formulated vaccine were dominated by IgM, IgG1, and IgG2b (Fig. 3C). Similar isotype profiles were observed in the lung with the additional presence of IgA, indicating a significant mucosal antibody response. Taken together, these results indicate that formulation of an otherwise ineffective low dose of split virus vaccine combined with R₄Pam₂Cys elicits strong antibody responses characterized by the presence of a number of antibody isotypes, including significant levels of IgA, especially when administered intranasally.

Hemagglutination inhibition (HI) and neuraminidase inhibition (NI) titers in sera obtained from mice inoculated with split virus plus R₄Pam₂Cys were found to be comparable to or higher than the antibody titers induced by PR8 infection (Table 1). In contrast, titers were not detected in sera of animals vaccinated with split virus alone. Not surprisingly, no detectable HI or NI reactivity was observed using sera from mice inoculated with vaccine alone or with vaccine formulated with R₄Pam₂Cys when tested against the serologically distinct H3N2 X31 virus. These results demonstrate that antibodies induced by inoculation of split virus can prevent viral HA and NA activity of homologous but not heterologous strains.

To investigate the induction of T cell responses, lymphocytes from the mediastinal lymph nodes (MLNs) and lungs were analyzed for the presence of NP_147–155-specific CD8⁺ T cells by tetramer straining 7 days after vaccination. Compared to vaccination of split virus alone, vaccination of split virus in the presence of R₄Pam₂Cys resulted in the detection of higher numbers of NP_147–155-specific CD8⁺ T cells at both these sites, especially in the lungs (Fig. 3D). Furthermore, higher numbers of gamma interferon (IFN-γ)-, tumor necrosis factor alpha (TNF-α)-, and interleukin-2 (IL-2)-producing NP_147–155-specific CD8⁺ T cells were also detected in lymphocyte populations derived from animals vaccinated with split virus plus R₄Pam₂Cys following in vitro restimulation with NP_147–155 peptide for a further 7 days (Fig. 3E).

Protection against homologous and heterologous viral challenge.

To evaluate the in vivo protective efficacy of split virus formulations, mice were challenged 35 days following inoculation with split virus vaccine alone or formulated with R₄Pam₂Cys and lung viral titers were determined 5 days later. Mice challenged with the homologous PR8 virus (Fig. 4A) were not protected if vaccinated with split virus or with R₄Pam₂Cys alone as these groups of mice had lung viral titers not significantly different from those of phosphate-buffered saline (PBS)-vaccinated control mice. In contrast, animals that received low-dose split virus formulated with R₄Pam₂Cys showed complete elimination of virus from the lungs.

A similar approach was used to determine if vaccination could clear infection with X31 (Fig. 4B). As with PR8 challenge, inoculation of mice with split virus alone or R₄Pam₂Cys alone provided no significant reduction of viral loads in lungs. The titers of virus in the lungs of mice that had been inoculated with split vaccine formulated with R₄Pam₂Cys, however, were significantly reduced. These results demonstrate that the presence of R₄Pam₂Cys in a split virus formulation not only enhances virus-clearing adaptive immune responses to the homologous virus strain but also enables the vaccine to provide cross-reactive virus-clearing immunity to heterologous virus.

Ability of antibodies induced by vaccination to mediate heterosubtypic protection.

To investigate the possible presence of X31-neutralizing antibodies that were not detected in standard HI and NI assays, sera from vaccinated mice were tested for their ability to inhibit infection of MDCK cell monolayers in an in vitro virus neutralization assay. The results from these experiments (Table 2) demonstrated that sera obtained from naive mice or from mice inoculated with split virus alone were ineffective at specifically preventing PR8 and X31 infection of cells. In contrast, sera obtained from mice inoculated with split virus administered with R₄Pam₂Cys or from mice previously primed with PR8 virus effectively neutralized PR8 infection. They did not, however, inhibit infection by X31, indicating the absence of cross-neutralizing antibodies. Neutralization of X31 was achieved only using sera from mice previously infected with the homologous (X31) virus strain. These results are consistent with the presence of HI and NI antibodies in these sera (Table 1). We conclude that, although coformulation of the split vaccine with R₄Pam₂Cys results in a strong homologous neutralizing antibody response, detectable levels of neutralizing antibodies against a heterologous strain could not be demonstrated in vitro.

Protection against heterosubtypic viral challenge in µMT antibody-deficient mice.

To provide evidence that the immunity to heterologous virus challenge occurred independently of any antibody-mediated effects, µMT B cell-deficient mice were vaccinated and we evaluated their ability to clear virus following challenge with X31. Prior to challenge, vaccine-specific antibody titers in these mice were compared to those obtained in congenic C57BL/6 mice. In contrast to the significant levels of antibody detected in the serum and lung homogenates of C57BL/6 mice inoculated with split virus formulated with R₄Pam₂Cys, antibody levels in µMT mice that had been similarly immunized were below the limits of detection of the assay (Fig. 5A).

When pulmonary viral titers were measured in vaccinated µMT mice following challenge with X31, significantly lower viral levels were present in mice vaccinated with split virus in the presence of R₄Pam₂Cys than in its absence (Fig. 5B). It therefore appears that, despite a lack of antibody, µMT mice can mount a virus-clearing response to a heterologous virus, indicating that the cross-protective immunity generated in the presence of R₄ Pam₂Cys is antibody independent.

Pulmonary CD8⁺ T cell responses in vaccinated animals following viral challenge.

We next investigated whether protection against X31 was associated with CD8⁺ T cell responses. Following vaccination, lymphocytes in the lungs were analyzed for virus-specific IFN-γ-producing CD8⁺ T cells following ex vivo stimulation with syngeneic antigen-presenting cells (APCs) infected with X31 or pulsed with the nucleoprotein-derived H-2K^d-restricted immunodominant epitope NP_147–155. Our analyses found very few virus- or NP-specific IFN-γ-producing CD8⁺ T cells in the lungs of mice 34 days after vaccination with split virus in the absence or presence of R₄Pam₂Cys (Fig. 6A). Following challenge with X31, however, the number of virus-specific IFN-γ-producing CD8⁺ T cells that were detected in mice inoculated with split virus formulated with R₄Pam₂Cys was approximately 3-fold higher than the number detected in mice that received split virus alone or PBS (Fig. 6A and B). A large proportion of this recall response also appeared to be specific for NP_147–155 (Fig. 6A).

To determine the importance of these responses in mediating heterologous protection, vaccinated animals were depleted of CD8⁺ T cells prior to X31 challenge. Non-CD8⁺ T cell-depleted mice vaccinated with split virus plus R₄Pam₂Cys exhibited significantly higher recall responses to NP_147–155 than mice vaccinated with split virus alone (Fig. 6C) at 5 days postinfection. In comparison, responses in similarly vaccinated but CD8⁺ T cell-depleted animals were absent and correlated with an absence of a reduction in viral titers (Fig. 6D). Taken together, these results support the view that the ability of mice inoculated with split virus formulated with R₄Pam₂Cys to clear virus from the lung is at least in part due to the presence of a strong NP-specific CD8⁺ T cell-mediated recall response.

Reduction in pulmonary viral loads of contacts cohoused with R₄Pam₂Cys-vaccinated index mice.

To determine if the cross-protective responses observed in mice inoculated with split virus plus R₄Pam₂Cys (Fig. 4B) also inhibited transmission of virus to an unvaccinated population, we used a mouse contact-dependent transmission model (36). Mice vaccinated with split virus alone or combined with R₄Pam₂Cys were infected with a transmissible strain (X31) 35 days after vaccination (vaccinated index mice) and were cohoused for 54 h with naive animals (unvaccinated contact mice). Irrespective of whether mice were vaccinated with split virus in the presence or absence of R₄Pam₂Cys, we found similar viral titers (>6 log₁₀ PFU/ml) in the lungs of all vaccinated index mice (Fig. 7) after the cohousing period. This was expected because the recall and virus-clearing effects of pulmonary cytotoxic T cells take at least that long to manifest. Nevertheless, unvaccinated contact mice cohoused with index animals vaccinated with split virus plus R₄Pam₂Cys had undetectable or very low levels of virus in the lung. This is in contrast to the 1 to 5.4 log₁₀ PFU/ml of virus found in more than 60% of contacts cohoused with index mice vaccinated with split virus alone or combined with PBS. It should be noted that transmissibility in this model is known to be dependent on the level of virus present in the saliva of index mice. These results therefore suggest that, despite equivalent pulmonary viral loads early after challenge, reduced levels of virus may be present in the saliva of mice vaccinated with split virus plus R₄Pam₂Cys compared to mice vaccinated with split virus alone (36), resulting in less viral shedding.

Synthesis of the cationic lipopeptide R₄Pam₂Cys.

The synthesis of the cationic lipopeptide R₄Pam₂Cys was carried out manually using conventional solid-phase methodologies as described previously (34). Synthetic lipopeptide preparations contained no detectable lipopolysaccharide (LPS) (<0.05 endotoxin units [EU]/ml) as determined using the Limulus amebocyte lysate assay (Lonza, Walkersville, MD).

Vaccination and infection of mice.

All animal experimentation was performed with approval from the University of Melbourne’s Animal Ethics Committee. C57BL/6, BALB/c, TLR2^−/−, and B cell-deficient (µMT) mice (52) were bred and maintained under specific-pathogen-free conditions and used when the mice were 6 to 10 weeks of age. µMT mice carry a disrupted gene encoding the μ-chain constant region of IgM which arrests B cell development and the ability to mount antibody responses (52). Mice were anesthetized with isoflurane and inoculated via the intranasal (i.n.) route with split virus vaccine that was provided by bioCSL Limited, Australia. The vaccine contained 1 µg of HA and was derived from A/Puerto Rico/8/34 (PR8; H1N1) virus. The vaccine was administered either in saline solution or formulated with 5 nmol of R₄Pam₂Cys (9.91 µg) in a total volume of 40 µl. For some experiments, mice were inoculated via the subcutaneous route at the base of the tail (50 µl at each side) or via the intramuscular route (40 µl into each quadriceps muscle).

After 1, 3, 7, or 35 days, mice were challenged intranasally with 10^4.5 PFU of A/HK×31 (X31) (H3N2) influenza virus or with 50 or 500 PFU of PR8 influenza virus. X31 is a recombinant virus containing the HA and NA segments from a 1968 Hong Kong influenza virus but sharing the internal viral proteins of the PR8 virus. Lungs were harvested 5 days later, homogenized in 3 ml RPMI 1640 (Invitrogen, Australia), and centrifuged at 300 × g for 30 s, and supernatants were stored at −80°C. Titers of virus in the supernatants of lung homogenates were then determined using a Madin-Darby canine kidney (MDCK) plaque assay as previously described (53).

ELISAs.

Levels of antibody present in serial dilutions of sera or the supernatants of lung homogenates obtained from mice 34 days following vaccination were determined by an enzyme-linked immunosorbent assay (ELISA) as previously described (34). A panel of horseradish peroxidase (HRP)-conjugated rat anti-mouse IgG1-, IgG2a-, IgG2b-, IgG3-, IgA-, and IgM-specific antibodies (Southern Biotech, USA) were used in a separate assay to determine the isotypes of antibodies. Titers of antibody are expressed as the reciprocal of the highest dilution of serum required to achieve an optical density of 0.2.

Hemagglutination inhibition (HI), neuraminidase inhibition (NI), and virus neutralization assays.

The presence of hemagglutination inhibition (HI) antibodies and of neuraminidase inhibition (NI) antibodies was determined as described elsewhere (54, 55). Virus neutralization assays were carried out as described previously (55).

In vivo depletion of CD8⁺ T cells.

Mice were inoculated via the intraperitoneal route with 100 µg (200 µl) of anti-CD8α antibody (clone 2.43; National Cell Culture Center, USA) 29 days after vaccination. Antibody was administered daily for 3 consecutive days and then once every 3 days thereafter for the duration of the experiment.

Intracellular cytokine staining (ICS) assay and tetramer staining.

Lungs from mice were finely minced and treated with 2 mg collagenase A (Roche Diagnostics, Germany)–2 ml RPMI 1640 per lung for 30 min at 37°C. Cell suspensions were passed through a wire sieve before centrifugal sedimentation of cells and treatment with 0.15 M NH₄Cl–17 mM Tris-HCl at pH 7.2 for 5 min at 37°C. In some experiments, cells were obtained from bronchoalveolar lavage (BAL) fluids or mediastinal lymph nodes (MLNs). Lymphocyte cell suspensions (1 × 10⁶) were cultured in the presence of the H2K^d-restricted influenza virus nucleoprotein (NP)-derived immunodominant epitope NP_147–155 (TYQRTRALV; 1 mg/ml) at 37°C in a mixture with 200 µl of supplemented RPMI 1640 containing 55 mM 2-mercaptoethanol (2-ME), BD GolgiPlug (1 mg/ml) from a Cytofix/Cytoperm Plus kit (Becton, Dickinson), and recombinant IL-2 (Roche, Mannheim, Germany) (10 U/ml). For some experiments, lymphocyte cell suspensions were cultured in the presence of 2 × 10⁵ virus-infected syngeneic P815 target cells. These cells (4 × 10⁶/ml) had been infected previously with X31 (at a multiplicity of infection of 20) for 1 h at 37°C in Opti-MEM (Life Technologies, Australia) supplemented with gentamicin, glutamine, penicillin, and streptomycin at the concentrations described above and had been washed extensively in fetal calf serum (FCS) before use. Infection of cells was confirmed by the phenotypic expression of surface HA as verified by flow cytometry.

After 6 h, lymphocytes were washed with fluorescence-activated cell sorter (FACS) wash buffer (1% FCS–5 mM EDTA–PBS) and stained with a peridinin chlorophyll protein (PerCP) Cy5.5-conjugated rat anti-mouse CD8 antibody (clone 53-6.7; Becton, Dickinson) for 30 min at 4°C. Fixation and permeabilization were then performed for 20 min at 4°C using Cytofix/Cytoperm solution (Becton, Dickinson) according to the manufacturer’s instructions. Cells were washed once and stained for intracellular IFN-γ, IL-2, or TNF-α with fluorochrome-conjugated antibodies for 30 min at 4°C before flow cytometric analysis (FACSCanto II or LSR II; BD Biosciences). Data analysis was performed using FlowJo software (Treestar, USA).

For in vitro restimulation of lymphocytes over 7 days, cells (2 × 10⁶) were cultured in 200 µl of supplemented RPMI 1640 media containing NP_147–155 peptide (5 µg/ml) and IL-2 (10 U/ml) with medium changes performed every 2 days. ICS assays were then performed as indicated above.

Tetramer staining of lymphocytes from the MLNs and lungs was performed at room temperature using a 1:200 dilution in a mixture with 50 µl of 10% FCS–PBS for 30 min followed by the addition of fluorochrome-conjugated anti-CD8 antibody for 30 min. Cells were extensively washed before analysis.

Contact-dependent virus transmission.

The transmissibility of virus was evaluated in a mouse model developed for the study of contact-dependent viral transmission (36). Vaccinated mice were challenged after 35 days by the i.n. route with 10^4.5 PFU of X31 virus. After 6 h, these index mice were introduced into a clean cage and cohoused with naive contact mice (2 index mice for 3 contact mice) for 2 days. Index mice were euthanized directly after the cohousing period, and contact mice were euthanized 4 days after their initial exposure to the index mice. Viral titers in lung homogenates of all animals were then determined in a standard plaque assay.

Statistical analyses.

Analysis of variance (ANOVA) and P values were obtained using nonparametric one-way ANOVA; Tukey’s post hoc range tests, two-way ANOVA, and Bonferroni posttests were performed using the Prism 5 software package (GraphPad Software, La Jolla, CA, USA).