FcRn-Targeted Mucosal Vaccination against Influenza Virus Infection

The respiratory tract is a common site for pathogen exposure. The respiratory tract can resist infection and facilitate the clearance of invading pathogens through a variety of mechanisms, including the airway barrier of polarized epithelial cells and various innate or adaptive immune responses (1, 2). Adaptive immunity, including effector and memory T or B lymphocytes and local and circulating Abs, can prevent infections or decrease the severity of subsequent respiratory infections (3, 4). For example, tissue-resident memory (TRM) T cells that reside in the lung are a recently appreciated subset of memory T cells and are required for optimal protection against previously encountered pathogens (5–7). Presently, most vaccines against respiratory infections are designed for delivery via parenteral routes, including the muscle or skin, for intended protection of the lung. However, nonmucosal delivery routes elicit relatively poor mucosal immune responses in the respiratory tract, even though they often induce robust systemic immunity (8–11). A partial reason for the failure of systemic vaccination is the lack of strong mucosal Ab and cell-mediated immunity, including TRM T cells that reside in the lung tissue and their availability in the event of pathogen exposure. To prevent respiratory infections, an ideal way is to develop a mucosal vaccine that mimics natural respiratory infections and engenders beneficial lung immunity. This goal is best achieved by direct administration of vaccines via the respiratory route (12, 13). However, our ability to deliver vaccine Ags safely and effectively across the respiratory mucosal barrier is limited. Mucosal vaccines must avoid inducing excessively robust inflammatory responses that may lead to lung damage and exacerbate chronic diseases such as asthma or chronic obstructive pulmonary disease. Because respiratory infections are more prevalent in young and elderly individuals, certain types of mucosal vaccines, such as live-attenuated vaccines, are not preferred for these vulnerable populations. Given the high impact of respiratory infections on public health, developing safe and effective mucosal vaccines is an urgent, unmet need (12).

Epithelial monolayers lining the respiratory, intestinal, and genital tracts, as well as the placenta, polarize into the apical and basolateral plasma membrane domains, which are separated by intercellular tight junctions. The neonatal Fc receptor (FcRn) is expressed in these epithelial monolayers and mediates the transfer of IgG Ab across the epithelium (14–16). By IgG transcytosis, FcRn provides a line of humoral defense at mucosal surfaces (16–19), in addition to transferring maternal immunity to neonates. A hallmark of FcRn is its interaction with IgG Ab in a pH-dependent manner, binding IgG at acidic pH (pH 6–6.5) and releasing IgG at neutral or higher pH (20). FcRn primarily resides within low pH endosomes and binds IgG through the Fc region. Normally, IgG enters epithelial cells via pinocytotic vesicles that fuse with acidic endosomes. IgG bound to FcRn then enters a nondegradative vesicular transport pathway within epithelial cells. Bound IgG is transported to the apical or basolateral surface and released into the lumen or submucosa upon physiological pH (21). Evidence of IgG transport across the respiratory epithelia by FcRn suggests that FcRn might also transport a vaccine Ag, if fused with the Fc portion of IgG, across the respiratory mucosal barrier. Our previous studies showed that FcRn can induce protective genital immunity by efficiently transporting HSV gD or HIV-1 gag Ags across epithelial barrier (22, 23).

To test the possibility that FcRn can elicit a protective immunity to respiratory infections, we used the influenza A virus as a model pathogen. The hemagglutinin (HA) primarily mediates interactions of influenza virions with cell surface sialic acid residue receptors. After binding, virions are internalized through endocytic pathways to infect epithelial cells. The HA protein consists of the membrane-distal immunodominant globular head domain and the membrane-proximal HA stalk domain. The head domain shows high structural plasticity, which is strongly affected by antigenic drift; in contrast, the stalk domain exhibits a high degree of conservation. The HA plays a critical role in the early steps of viral infection and elicits both humoral and cellular immunity (24). In this study, we determined the ability of FcRn to deliver the viral HA protein fused to an Fc region of IgG across the respiratory epithelial barrier. We defined protective immune responses and mechanisms relevant to this route for mucosal vaccination in the lung in a mouse model. Our data suggest that FcRn-mediated intranasal (i.n.) delivery of influenza virus HA Ag induces high levels of long-lasting Ab and T cell responses, including TRM T cells in the lung, to provide potent protection against lethal influenza virus challenge. Our data demonstrate that FcRn-targeted delivery of an influenza virus vaccine Ag in the respiratory tract comprises an effective vaccine strategy and may be developed as a universal influenza vaccine against seasonal infection or for protection against pandemic influenza viruses or other common respiratory infections.

Chinese hamster ovary (CHO) cells were purchased from the American Tissue Culture Collection. Madin–Darby canine kidney (MDCK) cells were maintained in Opti-MEM complete medium (Invitrogen Life Technologies), and CHO cells were maintained in complete DMEM (Invitrogen Life Technologies), both supplemented with 10% FBS, 2 mM l-glutamine, nonessential amino acids, and penicillin (100 U/ml)/streptomycin (100 μg/ml). Recombinant CHO cells were grown in a complete medium with G418 (500 μg/ml). All cells were grown at 37°C in 5% CO₂. Influenza A/Puerto Rico/8/34/H1N1 (PR8) virus was provided by Dr. P. Palese (Icahn School of Medicine at Mount Sinai) and was amplified in 10–11-d-old embryonated chicken eggs and titrated by 50% end point dilution assay. The HRP-conjugated streptavidin and anti-mouse IgG, IgG1, IgG2b, and IgG2c were obtained from SouthernBiotech (Birmingham, AL). HA Abs were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). HA stalk–specific mAb KB2, 6F12, FI6v3, CR6261, and CR8020 were provided by Dr. F. Krammer (Icahn School of Medicine at Mount Sinai) and Dr. J. Boyington (National Institutes of Health). Anti-HA Ab (RA5-22, mouse IgG1) was purchased from Santa Cruz Biotechnology (SC-52025). HRP-conjugated goat anti-mouse IgG2a (1080-05) and HRP-conjugated rat anti-mouse IgG1 (1144-05) were from SouthernBiotech. Recombinant HA was purchased from Sino Biologicals (Shanghai, China) or from Biodefense and Emerging Infections Research Resources Repository (NR-19240; BEI Resources, Manassas, VA). HIV gp120–specific IgG mAb B12 was also from BEI Resources.

To make an IgG Fc fusion protein, a pCDNA3 plasmid encoding the hinge, CH2, and CH3 domains of mouse IgG2a Fc (22) served as a template for the Fc fragment. The rationale for using IgG2a is that it has the highest affinity for activating FcɣRI but the lowest affinity for FcɣRIIB. In IgG2a Fc, the E318, K320, and K322 residues were replaced with alanine residues to remove the complement C1q binding site (25). In addition, to produce a mutant (mut) form of IgG Fc protein that cannot bind to FcRn, the H310/Q311 and H433/N434 residues were changed to A310/D311 and A433/Q434 residues, respectively, to eliminate FcRn binding sites (26).

To generate a trimeric HA that is fused to the Fc, we first converted the Cys224, Cys227, and Cys229 residues to serine residues within the Fc using a DNA mutagenesis kit (Clontech Laboratories), resulting in a monomeric Fc fragment. To make a trimeric HA-Fc fusion gene, the extracellular portion of PR8 HA, excluding the signal peptide sequence, was amplified by PCR from a plasmid containing full-length PR8 HA using the primer pair 5′-GGATCAGGCGGGGGTGGGTCCGGAGGAGGTGGCTCGGGATCTGACA CAATATGTATAGGCTACCATGC-3′ and 5′-CCTCTGGGCACCAGGCTTCTTGATCCTGAGCCT GATCCCTGATAGATCCCCATTGATTCC-3′. The IgG Fc antisense primer and the HA sense primer contain complementary glycine and serine codons to produce a 14GS linker to bridge the IgG Fc and HA fragments. A protein trimerization domain was amplified from a plasmid containing the T4 fibritin foldon sequence (27). Similarly, the HA antisense primer and the foldon sense primer contain complementary glycine and serine codons to introduce a 6GS linker between the HA and foldon fragments. The Fc, HA, and foldon fragments were fused by overlapping PCR and ligated into the pCDNA3 vector. All the resultant plasmids were confirmed by dsDNA sequencing to verify the fidelity of PCR amplification and DNA cloning.

Protein concentration and quality were assessed by 8–12% SDS-PAGE gel under reducing and nonreducing conditions. Protein in gels was either stained with Coomassie blue dye or used for transferring onto nitrocellulose membranes (Schleicher & Schuell). The membranes were blocked with 5% milk in PBS and 0.05% Tween 20 (PBST) and incubated overnight with anti–IgG2a–HRP (1:10,000) or anti-HA Abs (1:2000). For HA probing, membranes were further incubated with the anti-mouse IgG1–HRP Ab (1:5000) for 2 h. SuperSignal West Pico PLUS ECL substrate (Thermo Fisher Scientific) was used to visualize protein in membranes, and images were developed and captured by the ChemiDoc XRS system (Bio-Rad Laboratories).

The different HA-Fc plasmids were transfected into CHO cells using PolyJet reagent (SignaGen Laboratories). Stable cell lines were selected and maintained under G418 (0.5–1 mg/ml). Expression and secretion of HA-Fc fusion proteins were determined by immunofluorescence assay, SDS-PAGE, and Western blotting analysis. The soluble HA-Fc proteins were produced by culturing CHO cells in a complete medium containing 5% FBS with ultralow IgG. The proteins were purified by protein A column (Thermo Fisher Scientific) for the HA-Fc/wt protein and anti-mouse IgG (Rockland) conjugated agarose beads for the HA-Fc/mut protein. Protein concentrations were determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific).

The trimerization of HA-Fc was determined by the bis(sulfosuccinimidyl)suberate (BS³; Thermo Fisher Scientific) cross-linker method. Briefly, HA-Fc proteins (0.1 mg) were incubated with BS³ in 50-fold molar excess for 2 h on ice. The reaction was then quenched by adding 1 M Tris-HCl (pH 7.5) to a final concentration of 50 mM Tris-HCl and further incubated for 15 min at room temperature. The protein samples were subjected to electrophoresis and subsequently analyzed by Western blotting analysis with anti-HA and anti-IgG2a Abs in Western blotting.

An FcRn binding assay was performed. CHO cells were either transfected with plasmids expressing mouse FcRn and β₂-microglobulin (β₂m) or mock transfected. Twenty-four hours later, the transfected cells were seeded in a six-well plate for 6 h. Cells were subsequently equilibrated with medium under either pH 6 or pH 7.4 condition at 4°C for 30 min; then, 3 μg of trimeric HA-Fc/wt, HA-Fc/mut, or HA was added into each well or left untreated for 1 h. The cells were washed with a corresponding pH buffer to remove the unbound proteins. The cells were finally lysed in cold PBS (pH 6 or 7.4) with 0.5% CHAPS (Sigma-Aldrich) and protease inhibitor mixture (Calbiochem) mixture on ice for 1 h. The soluble proteins (10 μg) were subjected to Western blot analysis and blotted with biotin-labeled anti-HA primary Ab and streptavidin–HRP-conjugated secondary Ab.

Surface plasmon resonance analysis of affinity between mouse FcRn and trimeric HA-Fc/wt was performed by ACROBiosystems (Newark, DE). In brief, the purified recombinant FcRn was diluted to 1 μg/ml with 10 mM sodium acetate (pH 4.5) and was immobilized onto a CM5 biosensor chip (Biacore, Uppsala, Sweden) using an amine coupling kit (Biacore). The activator is prepared by mixing 400 mM 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride and 100 mM N-hydroxysuccinimide (GE Healthcare) immediately prior to injection. The CM5 sensor chip is activated for 420 s with the mixture at a flow rate of 10 μl/min, which typically results in immobilization levels of 100 resonance units. The chip is deactivated by 1 M ethanolamine hydrochloride–NaOH (GE Healthcare) at a flow rate of 10 μl/min for 420 s. The reference surface channel was prepared in the same way as the active surface but without injecting mouse FcRn. The HA-Fc/wt proteins were diluted with the running buffer B (PBST [pH 6.0]) to 62.5, 31.25, 15.625, 7.813, 3.906, 1.953, and 0 nM. The HA-Fc/wt proteins were injected and allowed to flow at a rate of 30 μl/min for an association phase of 90 s, followed by 210 s for dissociation in the running buffer B. The affinity was analyzed by Biacore T200 Evaluation Software 3.0 in Biacore T200.

Immunofluorescence was performed as previously described (22). Briefly, cells were grown on coverslips for 48 h. The cells were rinsed with HBSS and fixed with 4% paraformaldehyde (Sigma-Aldrich) in HBSS for 20 min and quenched with 100 mM glycine in PBS for 10 min. Cells were permeabilized with 0.2% Triton-X in HBSS for 5 min and incubated with blocking solution (3% normal goat serum in PBS) for 30 min. Cells were incubated with anti-HA Ab diluted in blocking solution for 2 h. After washing, Alexa Fluor 555–conjugated anti-mouse IgG1 or IgG2a secondary Ab was added for 1 h. Cells were washed with PBS and mounted to slides with ProLong antifade solution (Thermo Fisher Scientific). Images were obtained using a Zeiss LSM 510 confocal microscope and analyzed by LSM Image Examiner software (ZEISS).

All the animal experiments were performed with the approval of the Institutional Animal Care and Use Committee. FcRn knockout (KO) mice are a kind gift from Dr. D. Roopenian (The Jackson Laboratory). Six- to eight-week-old female C57BL/6 mice (Charles River Laboratories) and FcRn KO mice were i.n. immunized with 20 μl of 5 μg HA-Fc/wt, HA-Fc/mut, equivalent molar of recombinant HA, or PBS. All vaccine proteins or PBS were mixed with 10 μg of CpG ODN 1826 (InvivoGen). For i.m. immunizations, mice were injected in the right hind leg with a 50-µl sample containing 5 μg of HA-Fc/wt Ag admixed with 10 μg of CpG. Two weeks later, the mice were boosted with the same vaccine formulations. The mice were i.n. infected with lethal doses (1 × 10⁴ median tissue culture infectious dose [TCID₅₀]_, equal to 5 median lethal dose [MLD₅₀]) of the PR8 virus 2 wk after the boost. For immunizations and challenge, all mice were anesthetized with an i.p. injection of 200 μl of fresh Avertin (20 mg/ml; Thermo Fisher Scientific) and laid down in a dorsal recumbent position to allow for recovery. Postinfection, mice were monitored daily for weight loss and other clinical signs of illness for 14 d. Animals that lost above 25% of their body weight on the day of infection or had become grossly moribund were euthanized.

Bronchoalveolar lavage (BAL) and nasal wash fluids were collected 14 d after boost. Briefly, a small incision was made in the trachea. A syringe with a thin tube inserted at the tip was filled with PBS. The syringe was inserted first into the trachea toward the lungs, and 1 ml of PBS was carefully injected into the lungs, and by keeping the syringe in position, the PBS was retrieved back for the collection of BAL. For sampling the nasal wash, the syringe was similarly inserted into the trachea but toward the nasal cavity. PBS was carefully injected into the nasopharynx and collected when it flowed from the external nares. BAL and nasal wash fluids were then subjected to low-speed centrifugation, and the supernatants were retained.

The single-cell suspensions from the mediastinal lymph nodes (MedLNs) or spleen were made by mechanical abrasion of the organs. For isolation of cells from bone marrow, tibias and femurs were removed, and the ends were clipped. The bone marrow was flushed out with RPMI 1640. Isolation of single cells from the lung was performed as previously described (22). Briefly, after perfusion with 3 ml of PBS, lungs were minced and treated to enzymatic digestion in RPMI 1640 with pronase (1.5 mg/ml), dispase (0.2%), and DNase (0.5 mg/ml) for 40 min at 37°C with rotation. All cells from the MedLNs, spleen, bone marrow, and lung were filtered through a 40-μm nylon cell strainer and treated with RBC lysis buffer (0.14 M NH₄Cl and 0.017 M Tris-HCl [pH 7.2]). All cells were washed and suspended in 2% FBS (Invitrogen Life Technologies) in PBS or RPMI 1640 complete medium with 1–2% FBS. For each experiment, cells were pooled from three to five mice in each animal group.

For i.v. in vivo labeling of circulating T cells, mice were i.v. injected with 3 μg of PerCP-Cy5.5–conjugated anti-mouse CD3e Ab (145-2C11; BD Biosciences). After 10 min, lungs were perfused with 3 ml of PBS, and the single-cell suspensions were made as described above. Fc block (anti-mouse CD16/CD32, 1 μg/sample; BD Biosciences) was added to the lung and spleen cell samples and incubated for 30 min at 4°C. After washing with FACS buffer, cells were incubated with fluorescently conjugated Abs to stain for T cell markers, CD3 PE (145-2C11; BD Biosciences), CD4 APC-Cy7 (RM4-5; BD Biosciences), CD8 APC-Cy7 (53-6.7; BD Biosciences), CD69 APC (H1.2F3; BD Biosciences), CD11a FITC (2D7; BD Biosciences), and CD103 FITC (M290; BD Biosciences) for 1 h at 4°C. Isotype control Abs were included in each experiment. After washing, cells were suspended in 2% paraformaldehyde and analyzed using an FACSAria cytometer (BD Biosciences) and FlowJo software (Tree Star).

For determining T cell–derived cytokine levels, intracellular cytokine staining was performed as described (28). Briefly, single-cell suspensions from the lungs were stimulated with 2 μg of HA for 5 h at 37°C. Cells were then incubated with GolgiStop (BD Biosciences) for an additional 5 h. After wash, cells were incubated with Fc block and then stained with fluorescently conjugated Abs for T cell surface markers, CD3 PE (145-2C11; BD Biosciences), CD4 APC-Cy7 (RM4-5; BD Biosciences), and CD8 APC-Cy7 (53-6.7; BD Biosciences). Cells were fixed and permeabilized by incubating with BD CytoFix/Perm. After FACS buffer wash, cells were stained for cytokines IFN-γ APC (XMG1.2; BD Biosciences) and TNF-α Alexa Fluor 488 (MP6-XT22; BD Biosciences). All block, incubation, and permeabilization steps were performed for 20 min at 4°C. After wash, cells were suspended in 2% paraformaldehyde and analyzed by flow cytometry as described above.

Viral titers were determined by 50% end point dilution assay and hemagglutination assay as described (29, 30). Briefly, three extra mice were randomly selected from each immunized group and euthanized at day 4 postinfection. Mouse lungs were collected, and individual lungs were homogenized in the TissueLyser LT (QIAGEN). After centrifuging the homogenates, the supernatants were serially diluted and incubated on MDCK cells for 1 h. The supernatants were removed from cells and replaced with serum-free Opti-MEM with 1 μg/ml tosyl phenylalanyl chloromethyl ketone (TPCK)–treated trypsin. After incubation at 37°C for 3 d, the supernatant (50 μl) was mixed with chicken RBC (50 μl) and incubated for 35 min. Samples were scored for agglutination, and virus titers were calculated by the Reed–Muench method.

To examine the lung pathology, lungs were removed from at least three mice in each group and photographed to observe gross pathology. Lungs were then fixed in 10% buffered formalin solution. The lungs were paraffin embedded, sectioned in 5-μm thickness by American HistoLabs (Gaithersburg, MD), and stained with H&E. To determine the level of pulmonary inflammation, the lung inflammations were scored by an investigator who was blinded to the experimental design. A semiquantitative scoring system, ranging from 0 to 5, was used to evaluate the following parameters: alveolitis, parenchymal pneumonia, inflammatory cell infiltration, peribronchiolitis, perivasculitis, and lung edema (31, 32). The inflammatory scores are defined as follows: 0, normal; 1, very mild; 2, mild; 3, moderate; 4, marked; and 5, severe. An increment of 0.5 was assigned if the inflammatory score falls between two integers.

For the detection of HA-specific Abs in serum, BAL fluid, and nasal washes, ELISA plates (MaxiSorp, Nunc) were coated with 3 μg/ml of the trimeric HA protein (NR-19240; BEI Resources) in coating buffer and incubated overnight at 4°C. For determination of the interaction between the HA-Fc and HA stalk–specific Abs, plates were coated with serially diluted mAbs, starting from 3 μg/ml. Plates were then washed three times with PBST and blocked with 2% BSA in PBST for 1 h at room temperature. Samples were serially diluted in 2% BSA–PBST, or HA-Fc (0.5 μg/well) was added for 2-h incubation. After washing three times, HRP-conjugated rabbit anti-mouse IgG Ab (1:20,000; Pierce) or anti-mouse subclass-specific Ab (1:5000; SouthernBiotech) was added. For use of biotin-labeled goat anti-mouse IgG–specific Fab (1:2000), the streptavidin–HRP (1:8000) was added. The reaction was visualized in a colorimetric assay using substrate tetramethyl benzidine and analyzed using VICTOR³ microplate reader (PerkinElmer). Titers represent the highest dilution of samples showing a 2-fold increase over average OD₄₅₀ nm values of negative controls.

For measuring HA-specific Ab-producing plasma cells, 96-well ELISpot plates (MilliporeSigma) were prewetted with 35% ethanol and washed with PBS. The plates were then coated with 5 μg/ml HA protein overnight at 4°C and blocked with RPMI 1640 complete medium with 10% FBS for 2 h at 37°C under 5% CO₂. Serial dilutions of single-cell suspensions from bone marrow were prepared in RPMI 1640 and added to the coated wells for 24 h at 37°C in 5% CO₂. After the incubation, the cells were removed, and the plates were washed five times with PBST and then incubated with biotin-labeled goat anti-mouse IgG–specific Fab Ab (1:2000) for 2 h. After washing with PBST, the streptavidin-conjugated HRP (1:3000) was added and incubated for 1 h. The samples were developed with 3-amino-9-ethylcarbazole substrate (BD Biosciences). After washing, the plates were stored upside down in the dark to dry overnight at room temperature. Spots were counted with an ELISpot reader and analyzed by ZellNet Consulting (Fort Lee, NJ).

Neutralizing Abs were measured by a standard microneutralization assay on MDCK cells as previously described (33). Briefly, receptor-destroying enzyme–treated serum samples were serially diluted in PBS with 1× antibiotics/antimycotics. Then, 100 TCID₅₀ of the PR8 virus was added to each well and incubated at 37°C for 1 h. MDCK cells were incubated with the serum/virus mixture for 1 h at 37°C. After removing the mixture, serum-free Opti-MEM containing 1 μg/ml TPCK-treated trypsin was added to each well and incubated for 3 d at 37°C. Cytopathic effects were observed daily, and the presence of virus was determined by HA assay as described elsewhere. Neutralizing Ab titers were determined as the reciprocal of the highest serum dilution, preventing the appearance of cytopathic effects. Each assay was done in triplicate. The average neutralizing Ab titer was determined for each immunization and control group.

To compare the Kaplan–Meier survival curves, we used multiple Mantel–Cox tests. Differences in Ab titers, cytokine percentages, virus titers, inflammation scores, and IgG-secreting cell numbers were assessed by using one-way ANOVA with Dunnett multiple comparison tests. GraphPad Prism 5.01 software was used for the statistical analyses.

To activate virus infectivity, the HA precursor molecule HA0 is cleaved into HA1 and HA2 (34). To produce the noncleavable HA0 protein, mutagenesis at the cleavage sites (SIQS→QRET) of PR8 HA ensured that the expressed HA would remain in the HA0 precleavage state. The HA exists as a trimer on the virions or virally infected cells. It is likely that a trimeric HA Ag fused to an Fc would more closely mimic a native HA structure. Because IgG Fc forms a disulfide-bond dimer, we created a monomeric Fc by eliminating the disulfide bonds formed by three cysteines at positions 224, 227, and 229 by substituting with serine residues. We also generated an Fc mut that was unable to bind FcRn owing to histidine residue substitutions at positions 310 and 433 (23) (Fig. 1A). In both wild-type (wt) and mut Fc for FcRn binding, the complement C1q-binding motif was eliminated (24) (Fig. 1A). To facilitate trimerization, we engineered a foldon domain from T4 bacteriophage fibritin (25) to the C terminus of HA0. We fused the monomeric IgG Fc/wt or Fc/mut in frame with the HA0-Foldon, respectively (Fig. 1A), generating plasmids that expressed trimeric HA-Fc/wt or HA-Fc/mut proteins.

We observed that both HA-Fc/wt and HA-Fc/mut proteins secreted from stable CHO cells were monomers under reducing or nonreducing conditions (Fig. 1B, 1D), suggesting the removal of the disulfide bonds in the Fc. The fusion proteins were expressed in the HA fused at the C or N terminus of the Fc. FcRn binds IgG at acidic pH, but not neutral pH, conditions (21). To determine whether HA-Fc/wt or HA-Fc/mut protein binds to FcRn, we incubated CHO cells expressing mouse FcRn and β₂m with 3 μg of HA-Fc/wt, HA-Fc/mut, or HA protein under pH 6 or pH 7.4 condition for 1 h at 4°C. In this way, FcRn would bind the HA-Fc proteins only at pH 6. As shown in (Fig. 1C, the HA-Fc/wt and FcRn proteins were detected with anti-HA– or anti-FcRn–specific Ab (lane 1). However, the HA-Fc/mut (lane 2) or HA (lane 3) proteins were not found. Surface plasmon resonance assay determined that the K_D of the trimeric HA-Fc binding to immobilized mouse FcRn was 2.38 nM (Supplemental Fig. 1). Therefore, the trimeric HA-Fc/wt protein maintains the structural integrity required to interact with FcRn.

We further determined whether the HA portion of the HA-Fc indeed maintains its trimeric conformation. First, the BS³, a hydrophilic, 11-Å cross-linker that covalently links proteins, can stabilize trimeric influenza HAs (35, 36). Thus, we cross-linked the HA-Fc/wt or HA-Fc/mut proteins with BS³, and the treated proteins were subjected to SDS-PAGE analysis under a reducing and denaturing condition. As shown in (Fig. 1E, the treated HA-Fc/wt and HA-Fc/mut proteins migrated to a position at ∼330 kDa in comparison with the untreated HA-Fc protein that migrated at 110-kDa position, suggesting the HA-Fc protein exists as a trimer. The HA proteins migrated at 70 kDa without BS³ treatment but at ∼210 kDa with BS³ treatment. Second, we used broadly neutralizing HA Abs to probe the epitopes on HA-Fc/wt. The HA-Fc/wt interacted with mAbs 6F12 and KB2 in an immunofluorescence staining (Fig. 1F) in CHO stable cell lines or with mAbs CR6261, FI6v3, and 6F12 in soluble form in a concentration-dependent manner in ELISA (Fig. 1G). CR8020 mAb showed the binding with low affinity because it preferably binds to the HA stalk of group 2 influenza virus. All of these HA-specific mAbs are conformation dependent (37–40). Together, we showed that the HA portion of the HA-Fc proteins forms a trimer and maintains the correct conformational structure, whereas its monomeric Fc portion retains its ability to interact with FcRn.

We tested whether FcRn-dependent transport augments the immunogenicity to HA protein. Mice were immunized i.n. with 5 μg of HA-Fc, HA protein (equal molar amount), or PBS, all in combination with 10 μg of CpG, and boosted after 2 wk. The specific engagement of FcRn in enhancing immunity was demonstrated in wt mice that were immunized with trimeric HA-Fc/mut proteins or FcRn KO mice that are immunized with trimeric HA-Fc/wt proteins. The HA unlinked to an Fc fragment allowed us to evaluate FcRn-independent effects in vivo and determine the magnitude of any observed enhancement in immune responses conferred by targeting the HA-Fc to FcRn. Therefore, these control groups allow us to evaluate the extent that interactions between FcRn and Fc contribute to the immune responses. We coadministrated CpG as a mucosal adjuvant (41). Significantly higher titers of total IgG (Fig. 2A), together with IgG1, IgG2b, and IgG2c (Supplemental Fig. 2), were seen in the HA-Fc/wt–immunized mice when compared with the HA-, HA-Fc/mut–, and HA-Fc/wt/KO–immunized and PBS-treated groups of mice. We found that CpG was necessary to enhance the Ab immune responses when the HA-Fc was targeted to FcRn. Moreover, sera from the HA-Fc/wt–immunized mice exhibited strong neutralizing activity relative to those from other control groups (Fig. 2B). Likewise, HA-Fc/wt proteins induced strong IFN-ɣ– or TNF-α–producing CD8⁺ and CD4⁺ T cell responses, as evidenced by significantly higher percentages of IFN-ɣ– or TNF-α–producing CD4⁺ (Fig. 2C, 2D, Supplemental Fig. 3A) and CD8⁺ (Fig. 2E, 2F, Supplemental Fig. 3B) T cells in response to HA stimulation in the lungs of wt mice immunized with HA-Fc/wt, compared with the other groups. This Th1 response was also supported by a major presence of the IgG2c subclass in the sera of the immunized mice (Supplemental Fig. 2). It remains uncertain whether this polarized Th1 cell response is caused by mucosal immunization as a result of FcRn targeting or, more likely, by the CpG used as adjuvant. Overall, our data demonstrate that engagement of FcRn greatly increased the efficiency by which HA Ag–specific Ab and cellular immune responses were induced.

Because the influenza virus initiates its infection in the airway (42), an important objective for FcRn-targeted mucosal delivery of influenza virus vaccines is to elicit stronger mucosal immune responses, including the presence of antiviral IgA Ab in nasal washes and IgG in the lung. Several lines of evidence demonstrate the outcome. To determine the ability of the FcRn-targeted mucosal immunization to induce local humoral immune responses, we examined HA-specific Abs in mucosal secretions. The nasal wash and BAL fluids were collected 14 d following the boost and analyzed for HA-specific IgG and IgA by ELISA. Significantly increased levels of HA-specific IgA and IgG were present in the nasal washes (Fig. 3A) and BAL (Fig. 3B) of the HA-Fc/wt protein–immunized mice. The wt, but not FcRn KO, mice that received the HA-Fc/wt protein had high levels of HA-specific IgA and IgG in the nasal washes and BAL (p < 0.01; (Fig. 3), suggesting that FcRn-mediated respiratory delivery of HA-Fc/wt induces mucosal IgA and IgG Abs. The formation and maintenance of germinal centers (GCs) generally lead to the differentiation of memory B cells and long-lived plasma cells. Second, we monitored the activated GC reaction in the MedLNs and spleens 10 d after the boost. As shown in (Fig. 3C, the trimeric HA-Fc/wt immunization induced substantially higher levels of FAS⁺PNA⁺B220⁺ B cells in the MedLNs or spleen of wt mice in comparison with those of the control groups. We conclude that HA Ag targeting FcRn, combined with CpG, produced strong Ab and T cell immune responses in the respiratory mucosa.

To test whether the humoral and cellular immune responses elicited by FcRn-targeted i.n. vaccination provides protection, we i.n. challenged all immunized mice with a lethal dose (5 MLD₅₀) of influenza PR8 virus 2 wk following the boost. Mice were monitored and weighed daily for a 14-d period and were euthanized at 25% body weight loss as a study end point. Most of the mice in the control groups had severe weight loss (up to 25%) within 8 d after the challenge (Fig. 4A) and either succumbed to infection or were euthanized. In contrast, only three of the 19 HA-Fc/wt–immunized mice experienced a 25% body weight loss. Hence, the trimeric HA-Fc/wt protein–immunized mice led to the protection in 84% of mice, which was significantly higher than the survival rates of other control groups (Fig. 4B). In addition, we assessed each group for viral replication in the lungs 4 d after lethal challenge (Fig. 4C). We observed markedly lower levels of virus in the lungs of the trimeric HA-Fc/wt–immunized mice. After the lethal challenge, there was a 1.5–3 log significant reduction of virus titer in the HA-Fc/wt–immunized group when compared with the PBS group (Fig. 4C). Other control groups of mice also essentially failed to contain viral replication. Some stochastic protections observed in the HA-Fc/mut and HA-Fc/wt/KO control groups (Fig. 4A) remain unknown; it might be associated with the innate immunity induced by the HA-Fc interacting with FcɣRs in alveolar macrophages of individual mice.

To further demonstrate protection, we characterized the lung pathology of all groups of mice following the challenge. The lung inflammations were scored in a blinded manner. Based on gross pathology, the lungs of mice in all control groups exhibited severe pulmonary lesions, as evidenced by hemorrhage with redness and edema (Fig. 5A). However, the lungs of HA-Fc/wt–immunized mice displayed significantly reduced hemorrhage with an overall pink-like color (Fig. 5A). The lungs of uninfected mice were used as normal control (Fig. 5A). To verify the gross pathology, we then used histopathology to determine the extent of lung inflammation. In agreement with the gross pathology, the histopathology of mouse lungs of all challenged control groups showed remarkable infiltrations of monocytes and lymphocytes, resulting in high levels of inflammation (Fig. 5B). In contrast, the mice immunized with HA-Fc/wt had a significantly lower inflammation score of the lungs compared with those of mice in the control groups (Fig. 5B). Collectively, these findings demonstrate that FcRn-mediated delivery of the trimeric HA-Fc/wt confers significant protection against lethal PR8 challenge, resulting in decreased mortality, viral replication, and pulmonary inflammation. In addition, we did not find a significant difference in the sensitivity of PR8 infection between unimmunized wt and FcRn KO mice.

In addition to providing immediate protection against infection after the boost, a successful influenza virus mucosal vaccine is expected to induce long-lasting immune memory. We wanted to determine whether the FcRn-mediated respiratory vaccination with HA-Fc/wt promotes an effective memory immune response up to 8 wk after the boost. As shown in (Fig. 6A, higher titers of HA-specific serum IgG were detected in the mice immunized with the HA-Fc/wt. To further show that this group of mice also maintains local immune responses, we again measured the IgA Abs in nasal washings and IgG in the BAL. We detected significantly high levels of HA-specific IgA and IgG in the nasal washes (Fig. 6B) and BAL (Fig. 6C) in the HA-Fc/wt–immunized mice, but not in the mice of control groups. By ELISpot, a significantly higher number of HA-specific IgG–secreting plasma cells were detected in the bone marrow of mice immunized with HA-Fc/wt (Fig. 6D). The existence of long-lived plasma cells in the bone marrow niche accounts for the maintenance of high levels of viral Ag-specific IgG in circulation (43). We detected some IgG-secreting plasma cells in the immunized FcRn KO mice; we speculate that the IgG-secreting cells may be contributed by an individual mouse with positive immune responses because pooled samples were used. Also, there was no significant difference in the number of IgG-secreting plasma cells between the mice immunized by HA-Fc/mut and FcRn KO mice immunized by HA-Fc/wt proteins (p > 0.05). It remains to be determined whether HA-specific IgA–secreting plasma cells also develop. These data indicate that HA-specific B cells maintained significant memory immunity potential at least 2 mo after the boost.

Memory CD4⁺ and CD8⁺ T cells are essential to provide protection against the influenza virus (6, 44). A recently appreciated subset is TRM T cells, a subset of T cells that are noncirculating and remain in the lung to provide a rapid response against influenza infections (45, 46). Hence, we determined whether FcRn-mediated immunization could induce TRM T cells in the lung. To differentiate circulating T cells from lung TRM T cells, we used the method of an i.v. in vivo infusion of fluorescently labeled anti-CD3 Ab that targets T cells in circulation, but not CD4⁺ TRM (CD69⁺ CD11a⁺) or CD8⁺ TRM (CD69⁺CD103⁺) T cells within the lung (45, 46). We detected substantial numbers of CD4⁺CD69⁺CD11a⁺ TRM cells (Fig. 6E) and CD8⁺CD69⁺CD103⁺ TRM cells (Fig. 6F) in the lungs, but not in the spleen of HA-Fc/wt–immunized mice (Supplemental Fig. 4), in comparison with that of mice in the control groups. We also failed to detect an appreciable increase in CD4⁺ or CD8⁺ TRM T cells in the lungs of all experimental animals when mice were immunized by the i.m. route (Fig. 6E, 6F). Together, these data suggest that FcRn-targeted respiratory, but not parenteral, immunization can induce lung-resident memory CD4⁺ and CD8⁺ T cells.

Last, to test whether these memory immune responses could provide protection, we again challenged the immunized mice with i.n. PR8 strain 2.5 mo after boost. Mice were weighed daily for a 14-d period and were euthanized at 25% body weight loss as a study end point. Most of the mice in the control groups had severe weight loss within 6–7 d after the challenge (Fig. 7A) and either succumbed to infection or were euthanized. Upon lethal challenge, mice immunized with the HA-Fc/wt exhibited significantly reduced disease severity with a survival rate of 80% (Fig. 7B), whereas mice in control groups succumbed to rapid weight loss and death. Overall, FcRn-targeted mucosal delivery of influenza HA vaccine engendered an effective memory immune response and provided protection against challenge.

Respiratory tract infections are important causes of serious illnesses and death. Conventional vaccination with nonreplicative vaccines is primarily administered by the parenteral routes. However, successful vaccination against respiratory infections may require high levels of potent and durable humoral and cellular responses in the local respiratory tract that are best achieved by direct mucosal immunization. To achieve this goal, a strategy to deliver vaccine Ags via the respiratory route is needed to improve the protective efficacy against respiratory infections. We are developing a novel strategy for vaccine delivery based on exploiting the FcRn-mediated Ab transfer pathway to deliver an influenza virus HA-Fc fusion protein vaccine across the respiratory epithelial barrier.

By using an i.n. delivery route for the nasal spray influenza vaccine, the current study demonstrates that FcRn-targeted respiratory vaccination induced substantial local and systemic immunity against lethal influenza virus infection. Site-specific (nasal and lungs) targeted delivery provided a unique pathway to produce local and systemic immunity against influenza virus infection. Several lines of evidence support this conclusion. First, the HA-Fc/wt–immunized mice have produced significantly high levels of IgG in the blood. Second, the HA-Fc/wt–immunized mice exhibited strong neutralizing Ab activity relative to control groups. Third, the majority of HA-Fc/wt–immunized mice resisted lethal influenza virus infection with reduced virus replication and inflammation in the lung. In contrast, most mice immunized by HA-Fc/mut or HA alone exhibited poor immune responses, increased levels of pulmonary inflammation, and decreased protection against virus challenge. The HA-Fc/mut protein was a control used to show that FcRn was directly mediating delivery of HA-Fc/wt Ag, eliminating the concern of a mucosal leakage. Our data clearly illustrate that the FcRn-mediated delivery pathway is essential to the protection against influenza virus challenge and demonstrate the value of our trimeric fusion protein strategy for directing viral Ags to this pathway. Several mechanisms may account for the protection against respiratory infection by FcRn-targeted mucosal vaccination. Efficient delivery of HA-Fc proteins across the respiratory barrier may increase the half-life of HA-Fc (21) to allow for enhanced Fc receptor–mediated uptake of HA-Fc by APCs such as dendritic cells (47–49). Previous studies showed that HA alone by i.n. route elicited some protective immunity following i.n. immunization (50); in our hands, HA alone was very poorly immunogenic and produced minimal protection against virus challenge. Previous work showed CpG does not increase the permeability of airway respiratory barrier; in contrast, it enhances the tight junction integrity of the bronchial epithelial cell barrier (51). We avoid other agents including volatile chemical anesthetics that are known to increase epithelial barrier permeability. Hence, our results clearly point to the benefits of FcRn-mediated delivery for enhancing the efficacy of respiratory tract–administered influenza virus HA Ags.

Considering influenza virus infects the epithelial cells lining the respiratory tract, an ideal vaccine should induce immunity in the mucosa that can effectively block virus penetration and spread. The local humoral immune response can be characterized by secretion of IgA in the upper respiratory tract or IgG in the BAL and the presence of activated GCs in the draining lymph node and the cytokine secretion by lung-specific T cells (1, 4). First, the trimeric HA-Fc/wt–immunized mice have produced high levels of IgG and IgA Abs in the BAL and nasal secretions. IgA is a major protective Ab in nasal secretions after immunization with influenza vaccine (42, 52). Local secretory Abs represent a primary barrier of immune defense against viral infections of the respiratory tract. Second, the HA-Fc/wt induced a high frequency of IFN-ɣ– or TNF-α–producing CD4⁺ and CD8⁺ T cells in the lung tissues of the immunized mice. IFN-ɣ and TNF-α are clearly indispensable for resistance to influenza infections (53). Third, we detected the presence of activated GCs in the MedLNs draining the lung. The nasopharynx-associated lymphoid tissue and the MedLNs are usually the sites where respiratory immune responses are initiated against Ags administered i.n. after reaching the lung. The presence of activated GCs in the nasopharynx-associated lymphoid tissue merits further investigation. Hence, FcRn-mediated respiratory delivery of influenza virus vaccine Ags promotes potent antiviral humoral and cell-mediated immune responses at the primary site of influenza infection, which is critical for clearance of the virus.

Induction of influenza-specific memory responses is crucial for a vaccine to provide protection after re-exposure to the influenza virus (6, 44, 54). Immunological memory has been a concern in protein-based subunit mucosal vaccine development. To establish long-lasting protection, a multifaceted memory immune response is essential, including virus-specific memory T and B cells and long-lasting plasma cells. A remarkable feature of this study is that FcRn-mediated mucosal vaccination with HA-Fc/wt induced and sustained higher levels of HA-specific Abs, both IgA and IgG, and plasma cells 2 mo after the boost. More importantly, we detected a higher percentage of CD4⁺ or CD8⁺ TRM T cells in the lungs of mice immunized with HA-Fc/wt, but not in control groups. CD4⁺ T cells are essential for promoting memory CD8⁺ T cell responses, including TRM CD8⁺ T cells (6, 55). TRM CD4⁺ or CD8⁺ T cells in the lung have been shown to promote rapid viral clearance at the site of infection and mediate survival against lethal influenza challenge (55, 56). In addition, we showed that TRM T cells are induced only via i.n. immunization and not by i.m. injections. This result is consistent with other findings that TRM T cells appear in the lung after natural influenza infection (46, 56) or they are induced by i.n. vaccination with live-attenuated influenza virus in a mouse model (45). Our results from FcRn-mediated respiratory delivery of influenza virus HA Ags verifies that the lung-resident T cells can only be induced solely via respiratory vaccination. Corresponding to the induction of memory humoral and cellular immune responses, most HA-Fc/wt–immunized mice resisted lethal influenza infection 2 mo after boost. Further studies are needed to confirm how long this subset of TRM T cells persists in the lung and how they specifically contribute toward long-term protection against influenza virus infections.

The Fc-fused trimeric HA proteins are required to induce a high level of protection from influenza virus infection. We initially immunized mice with a dimeric Fc-fused monomeric HA protein. Although the monomeric HA-Fc/wt induced a strong IgG immune response, it only conferred partial protection to subsequent influenza challenge. This low protection conferred by the monomeric HA vaccine may be interpreted by the fact that the native viral HA exhibits a trimeric presentation, which is essential for inducing conformation-dependent neutralizing Abs that mirror those induced by exposure to natural infection. Hence, we designed and produced a trimeric HA-Fc that mimics the native HA structure, as evidenced by the recognition of the trimeric HA-Fc by conformation-dependent anti-HA Abs and its ability to bind to FcRn. As expected, the mice immunized by the trimeric HA-Fc/wt proteins had high levels of survival and decreased morbidity in HA-Fc/wt–vaccinated mice.

We have shown that the effects of FcRn-targeted mucosal immunization differ considerably between wt and FcRn KO mice or the HA-Fc/wt– and the HA-Fc/mut–immunized mice in terms of mucosal and systemic immune responses, cytokine expression profiles, the maintenance of T and B cell memory, long-lived bone marrow plasma cells, and resistance to infection. In this study, we proved that FcRn-targeted mucosal delivery of influenza virus HA vaccine can provide protection against homologous influenza virus. This study leaves an open question of whether this pathway can be used to deliver a universal influenza vaccine that protects against all strains of influenza virus, eliminating the need for seasonal vaccination, with a potential to protect against pandemic strains. To achieve this goal, an optimal universal influenza vaccine is expected to induce broadly neutralizing Abs and cross-reactive T cells against conserved and protective influenza virus Ags, including the stalk domain of HA, nucleoprotein, matrix 2 ectodomain, and/or neuraminidase (57). We reason that the development of a universal influenza virus vaccine using our FcRn-mediated mucosal delivery of highly conserved influenza virus Ags, such as chimeric HA (24, 58, 59) or HA stalk–based vaccine (39, 60), is very likely. First, our trimeric HA-Fc Ag is readily recognized by several conformation-dependent, stalk-specific Abs (CR6121, FI6v3, 6F12, and CR8020) in a concentration-dependent manner (Fig. 1G, 1H) (37–40). Second, FcRn-mediated influenza HA delivery induces memory immune responses, including TRM T cells. TRM T cells are shown to promote viral clearance and mediate heterosubtypic protection and survival against lethal influenza virus challenge (45, 61). Third, FcRn-mediated mucosal delivery of influenza virus vaccines aimed at stimulating protective immunity in the respiratory tract will make a prospective universal influenza virus vaccine that is very likely to be more effective and efficient. This mucosal response may forestall influenza virus infection in its early stages, thereby contributing significantly to the reduction in influenza clinical infection and spread in the community. Hence, the mucosal delivery of protein Ags by FcRn may bring us closer to the implementation of a universal influenza virus vaccine.

Taken together, our study has demonstrated the important role of FcRn in facilitating i.n. delivery of protective influenza virus vaccine Ags across the respiratory mucosa, highlighting a (to our knowledge) novel approach for formulating influenza virus vaccines that stimulate long-lasting, protective local and systemic immunities. We propose a model for FcRn-targeted respiratory immunization. In general, mucosal DCs take up FcRn-transported Ags and subsequently migrate to MedLNs, where they prime CD4⁺ T cells and initiate the cognate B cell response in the GCs. By increasing the persistence of HA-Fc in tissue and circulation, interactions with FcRn may further enhance the development of long-term humoral and cellular immunity by sustaining high levels of serum IgG Abs and TRM T cells specific for HA. It is expected that FcRn can increase pre-existing influenza immunity because FcRn can transport influenza Ag–Ab complexes across the mucosal barrier (14). Our results imply that FcRn-mediated respiratory immunization could be proven to be an effective and safe strategy for maximizing the efficacy of vaccinations directed against influenza virus infections. Our goal is to develop multivalent mucosal vaccines offering protection against a spectrum of respiratory infections.

We thank Dr. Georgy Belov, Dr. Jeffrey DeStefano, Dr. Kenneth Frauwirth, and Dr. Wenxia Song for helpful discussions and critical reading. We are grateful to Dr. Peter Palese for supplying the PR8 virus. We acknowledge the receipt of HA mAbs from Dr. Jeffrey Boyington and MDCK cells expressing rat FcRn from Dr. Pamela Bjorkman. We are most grateful for the technical help from Dr. Donna Farber, Dr. Yunsheng Wang, Rongyu Zeng, Dr. Chunyan Ma, and Dr. Xiaoling Wu.

The authors have no financial conflicts of interest.