The adenoviral vectors vaccines and adjuvants are replication-deficient (ΔE1ΔE3) and based on the human serotype 5 (Ad5). Ad-HA and Ad-NP vaccines were generated from a codon-optimized gene sequence derived from A/H1N1/Puerto Rico/8/1934 (PR8) as described previously (33). The vaccine Ad-F was produced from a codon-optimized gene sequence from the RSV-A2 F protein as described before (37). High-titer viral stocks of the vaccines and the control vector lacking the transgene expression (Ad-Empty) were obtained in collaboration with Sirion Biotech (Martinsried, Germany). The established AdEasy system was used to obtain the mucosal adjuvants, Ad-TGFβ and Ad-CCL17. The respective ORFs were cloned into the vector pShuttle-CMV, which was then used for homologous recombination with the adenoviral vector pAdEasy-1 (38). The recombinant viruses were propagated in HEK 293 cells and viral particles (vp) were purified and concentrated with the Vivapure Adenopack20 (Sartorius). The concentration of total Ad was measured by optical density at 260 nm (OD₂₆₀) and the infectious particles by Reed-and-Muench TCID₅₀ (39). Ratios of total to infectious particles were typically bellow 200:1.

Six to eight weeks old female BALB/cJRj mice were purchased from Janvier (Le Genest-Saint-Isle, France) and kept in individually ventilated cages in accordance with German law and institutional guidelines under pathogen-free (SPF) condition, with constant temperature (20–24°C) and humidity (45–65%) on a 12 h/12 h-light/dark cycle. The research staff was trained in animal care and handling in accordance to the FELASA and GV-SOLAS guidelines. The study was approved by the Government of Lower Franconia, which nominated an external ethics committee that authorized the experiments. Studies were performed under the project license AZ. 55.2.2–2532-2–1085. Mice were intranasally immunized with a dose of 2x10⁸ vp (calculated from optical density measurement) of each antigen-encoding vector (Ad-HA, Ad-NP and Ad-F) in combination with 1x10⁹ vp encoding the adjuvant (Ad-TGFβ, Ad-CCL17 or Ad-empty). The vaccine was slowly pipetted in a total volume of 50 µl into one nostril under general anesthesia (100 mg/kg ketamine and 15 mg/kg xylazine). In all experiments, unvaccinated animals served as naïve negative control to define background levels. Blood samples were collected from the retro-orbital sinus under light anesthesia with inhaled isoflurane. Animals were euthanized with isoflurane and bronchoalveolar lavage fluids (BAL) were collected by washing the lungs with 2x1 ml PBS 1x via the cannulated tracheae. In addition, lungs and spleens were removed for further analysis.

96-well ELISA plates (Lumitrac, high binding, Greiner Bio-One) were coated with heat-inactivated (30 min, 56°C) 5x10⁵ PFU/well PR8 or RSV-A2 diluted in carbonate buffer overnight at 4°C. Afterwards, free binding sites were blocked with 5% skimmed milk in PBS-T_0.05 (PBS containing 0.05% Tween-20, Sigma-Aldrich). After a washing step with PBS-T_0.05, diluted sera or BAL were added and incubated for one hour at room temperature. Subsequently, plates were washed and the detection antibodies, HRP-coupled polyclonal anti-mouse IgG (1:3000, PA1–84631, Invitrogen) or anti-mouse IgA (1:5000, A90–103P, Bethyl Laboratories), were added for one hour. After washing with PBS-T_0.05 and the addition of an ECL substrate, the signals were acquired on a microplate luminometer (VICTOR X5, Perkin Elmer) with the PerkinElmer 2030 Manager software.

Neutralizing antibody titers against influenza virus were determined by a microneutralization assay as previously described (40). A two-fold serial dilution of complement-inactivated (56°C, 30 min) serum samples were incubated with 2000 PFU of PR8. The samples and the virus were diluted in Dulbecco’s modified Eagle’s medium containing 0.6% ml BSA, 100 units/ml penicillin/streptomycin, and 1.2 µl/ml Trypsin. Afterwards, the serum-virus mix was added on top of confluent MDCK II cells in a 96-well plate. Four days later, plaques were identified by crystal violet staining. The neutralization titer corresponded to the highest dilution in which complete inhibition of infectivity was observed.

RSV-neutralizing antibody titers were determined using a luciferase-based microneutralization assay. Two-fold serial dilution of complement-inactivated serum samples were incubated with 2.55x10⁴ TCID₅₀ of a recombinant RSV-A2 virus expressing firefly luciferase (41). The serum-virus mix was then added onto confluent Hep2 cells, previously seeded in a 96-well plate with serum-free DMEM (2 mM L-Glutamine, 100 units/ml penicillin/streptomycin). Four hours after incubation at 37°C, 2% DMEM was added to the cells (2% FCS, 2 mM L-Glutamine, 100 units/ml penicillin/streptomycin). After 24 h at 37°C, cell lysis was performed with Glo Lysis Buffer 1x (Promega) for 15 min at 37°C. Luciferase luminescence was detected after 3 min incubation with Bright-Glo™ Luciferase Assay System (Promega). The plates were acquired on a microplate luminometer (VICTOR X5, PerkinElmer) using PerkinElmer 2030 Manager Software. The 50% plaque reduction neutralization titers (PRNT50) were defined as the highest dilution that inhibited more than 50% of plaques observed in eight infected control wells without serum treatment.

HEK 293A cell lines expressing viral antigens under the control of a doxycycline-inducible promoter were generated after stable transduction with lentiviral particles encoding either the F protein of RSV-A2 (34), the HA or the NP protein of PR8. Initially, target cells were treated for 24–48 h with doxycycline (100–400 ng/ml) to induce antigen expression. Afterwards, 1x10⁵ cells were plated in a 96-well plate and HA- and F-producing HEK 293A cells were incubated with sera and BAL samples diluted in FACS buffer (PBS with 0.5% BSA and 1 mM sodium-azide) for 30 min at 4°C. Antibodies specific for intracellular NP were analyzed after incubation of permeabilized cells (0.5% saponin in FACS buffer) with samples diluted in permeabilization buffer. Bound HA-, NP- and F-specific antibodies were detected with either a polyclonal anti-mouse IgG-FITC (1:300, 406001, BioLegend) or a polyclonal anti-mouse IgA-FITC (1:300, A90–103F, Bethyl Laboratories) diluted in either FACS or permeabilization buffer. After incubation with the secondary antibodies for 20 min at 4°C, HA- and F-producing cells were fixated (2% PFA in PBS, 1004965000, Sigma-Aldrich). Samples were acquired on an Attune NxT (ThermoFisher) and analyzed with FlowJo™ software (Treestar Inc.).

Fifty-two days after immunization, mice were injected intravenously with 2 µg of anti-CD45-BV510 (clone 30-F11, BioLegend) to distinguish circulating from resident T-cells and sacrificed 3 min later with isoflurane. Lungs were collected and lymphocytes obtained after incubation of cut lung tissue pieces with 500 units Collagenase D (Sigma-Aldrich) and 160 units DNase I (Applichem) in 2 ml R10 medium (RPMI 1640 supplemented with 10% FCS, 2 mM L-Glutamine, 10 mM HEPES, 50 μM β-Mercaptoethanol and 1% penicillin/streptomycin) for 45 min at 37°C. Once digested, lung pieces were homogenized through a 70 µm cell strainer and incubated with an ammonium-chloride-potassium buffer (Gibco) to lyse erythrocytes. In the end, one eighth of the total lung cell suspension was plated in a 96-well round-bottom plate and incubated for 6 h with 100 µl of R10 plus monensin (2 µM), anti-CD28 (1 µg/ml, eBioscience), anti-CD107a-FITC (1:100, clone 1D4B, BD Biosciences) and 5 µg/ml of the respective MHC-I peptides (Genescript); NP₁₄₇-₁₅₅ TYQRTRALV or a F-pool (F_80-94 KQELDKYKNAVTEKQ; F_84-98 DKYKNAVTELQLLMQ; F_243-257 DKYKNAVTELQLLMQ; F_247-261 VSTYMLTNSELLSLI) to analyze T-cell cytokine expression. After restimulation, cells were stained with anti-CD8a-Pacific blue (1:300, clone 53–6.7, BioLegend), anti-CD4-PerCP eFluor710 (1:2000, clone RM4–5, eBioscience) and Fixable Viability Dye eFluor^® 780 (1:100, eBioscience). Subsequently, cells were washed with FACS buffer, fixed in 2% PFA and permeabilized with 0.5% saponin in FACS buffer. Cells were then stained intracellularly with anti-IL-2-APC (1:300, clone JES6-5H4, BioLegend), anti-TNFα-PECy7 (1:300, clone MPG-XT22, BioLegend) and anti-IFNy-PE (1:300, clone XMG1.2, BioLegend) diluted in permeabilization buffer. Data was acquired on Attune NxT (ThermoFisher) and analyzed with FlowJo™ software (Treestar Inc.).

Lymphocytes were obtained as described in the previous section and stained with either APC-labelled H-2K^D NP_147-155 (1:40, TYQRTRALV, ProImmune) or H-2K^D F₈₅-₉₃ (1:40, KYKNAVTEL, ProImmune) pentamers, followed by a second step staining with anti-CD69-BV421 (1:100, clone H1.2F3, BioLegend), anti-CD4-BV605 (1:1000, clone RM4–5, BioLegend), anti-CD19-BV650 (1:300, clone 6D5, BioLegend), anti-CD8-BV711 (1:500, clone 53–6.7, BioLegend), anti-CD49a-BV785 (1:300, clone Ha31/8, BD Biosciences), anti-CD45-FITC (1:300, clone 104, BioLegend), anti-CD11a-PerCP eFluor 710 (1:300, clone M17/4, Invitrogen), anti-CD103-PE (1:100, clone 2E7, Invitrogen), anti-CD127-PE/Dazzle (1:300, clone A7R34, BioLegend), anti-CD44-PECy5 (1:2000, clone IM7, BioLegend), anti-KLRG1-PECy7 (1:300, clone 2F1, Invitrogen) and Fixable Viability Dye eFluor^® 780 (1:100, eBioscience). Data was acquired on Northern Lights (Cytek) and analyzed with FlowJo™ software (Treestar Inc.).

Fifty- six days after vaccination, mice under general anesthesia (100 mg/kg ketamine and 15 mg/kg xylazine) were inoculated intranasally with 2000 PFU (2–4 x LD50) of mouse-adapted A/pH1N1/Hamburg/4/2009 (pH1N1) (42) or mouse-adapted A/H3N2/Hong Kong/1/1968 (H3N2, kindly provided by Prof. Georg Kochs, University Hospital Freiburg, Germany) in a final volume of 50 µl PBS. Weight loss and clinical score were monitored daily throughout infection and mice were euthanized when reaching the humane endpoint criteria. Lungs were collected and then homogenized in M-tubes with a GentleMACS dissociator (Miltenyi Biotec). Viral RNA was isolated from cell-free lung and BAL supernatants using the NucleoSpin^® RNA Virus kit (Machery-Nagel) according to the manufacturer’s instructions. Viral replication was quantified by SYBR-Green quantitative reverse transcriptase real-time PCR (qRT-PCR) using the GoTaq^® RT-qPCR 1-Step Kit (Promega) or TaqMan™ qRT-PCR using the AgPath-ID One-Step RT-PCR Kit (Thermofisher) with a 5’ FAM/3’ BHQ-1 probe (tcaggccccctcaaagccga). The universal primers used amplify the influenza matrixprotein 2 gene (forward: agatgagtcttctaaccgaggtcg, reverse1: tgcaaaaacatcttcaagtctctg, reverse2: tgcaaagacatcttccagtctctg). The data were log10-transformed and the lower limit of quantification was at 666 copies/ml. Tissue damage was determined by detecting total protein in cell-free BAL using bicinchoninic acid assay (Pierce) according to the manufacturer’s instructions.

RSV challenge was conducted by the intranasal inoculation of mice under general anesthesia (100 mg/kg ketamine and 15 mg/kg xylazine) with 5x10⁵ PFU of RSV-A2 (ATCC VR-1540) in a final volume of 50 µl PBS. Body weight and clinical score were monitored daily throughout infection. Mice were euthanized five days post infection. Lungs were collected and then homogenized in M-tubes with a GentleMACS dissociator (Miltenyi Biotec). Viral RNA was isolated from cell-free lung supernatants using the NucleoSpin^® RNA Virus kit (Machery-Nagel) according to the manufacturer’s instructions. Viral replication was quantified by SYBR-Green qRT-PCR using the GoTaq^® RT-qPCR 1-Step Kit (Promega) with primers for a sequence in the nucleoprotein gene (forward: agatcaacttctgtcatccagcaa, reverse: gcacatcataattaggagtatcaat). The data were log10-transformed and the lower limit of quantification was at 666 copies/ml. Tissue damage was determined by detecting total protein in cell-free BAL using bicinchoninic acid assay (Pierce) according to the manufacturer’s instructions.

All data sets were tested for normal distribution by using Shapiro-Wilk Test. Passing the normality test, one-way ANOVA or Two-way ANOVA with Tukey’s posttest were performed, otherwise, a non-parametric Kuskal Wallis Test with Dunn´s multiple comparisons was used. A p value of <0.05 was considered to be statistically significant. All analyses were done with Prism 9.0 (GraphPad Software, Inc.).

Vaccine immunogenicity was determined after intranasal immunization of BALB/c mice with an Ad5-based vaccine encoding for the HA and NP proteins of PR8 and RSV-F protein (Ad-HA/NP/F) in combination with either Ad-TGFβ or Ad-CCL17 as mucosal adjuvants ( Figure 1 ). As a control group, we used mice vaccinated with Ad-HA/NP/F plus an empty Ad5-vector construct to standardize the amount of particles each mouse received (Ad-Empty). A second control group consisted of naïve, non-vaccinated animals to validate background levels in the immunological assays. None of the mice showed any signs of adverse effects and histological analyses of the lung did not reveal signs of tissue damage or extensive inflammation two weeks after immunization (not shown).

Using whole virus particles as coating agent, we detected IAV- and RSV-specific IgG antibodies in serum ( Figures 2A, D ) 42 days and IgA antibodies in BAL ( Figures 2B, E ) 52 days after a single shot of the combinatory vaccine. Furthermore, the neutralizing capacity of sera was confirmed in IAV-and RSV-specific neutralization assays ( Figures 2C, F ). In an independent flow cytometry based assay, IgG and IgA antibodies against all three antigens could also be detected, however none of the adjuvants significantly impacted vaccine-induced antibody responses ( Supplementary Figure S1 ). Nevertheless, IgA responses to HA and RSV-F were slightly higher after co-administration of Ad-TGFβ ( Supplementary Figure S1 ).

Next, we performed a phenotypic analysis of the lung-derived lymphocytes and investigated the effect of the immunization strategy on the development of mucosal CD4⁺ T- and B-cells. To distinguish circulating from resident lymphocytes, a fluorescent anti-CD45 antibody was administered intravenously before sacrifice. Since resident T- and B-cells do not circulate, these cells are protected from the intravenous staining and therefore are considered CD45 iv^- (gating strategy shown in Supplementary Figure S2 ).

In the B-cell compartment of the vaccinated animals, higher proportions of CD45 iv^- cells (30–50%) were detected compared to the naïve controls (less than 20% of all B-cells; Figure 3A ). Since we can not specify the antigen-reactivity of those B-cells, the high proportion of iv^- cells in the naïve group might indicate some limitations of the intravascular staining for the analyses of B_RM. Nevertheless, to characterize the B_RM compartment in more detail, we performed an analysis using the markers CD11a, CD69 and CD103 within the CD19⁺ iv^- population (gating strategy shown in Supplementary Figure S2 ). In infection settings, B_RM have been described to not express CD103 (27). Here we investigated whether similar phenotypes were found in the vaccine context. Indeed, CD103 expression was also not found in B_RM induced after mucosal immunization with Ad-HA/NP/F. Most of the CD45 iv^- B-cells found in the lungs of mice express the integrin CD11a, but lack expression of CD69 and CD103. This B-cell population was increased upon vaccination, while the mucosal adjuvants seem to further promote its formation ( Figure 3B ). Furthermore, a second CD11a⁺ B_RM population expressing additionally the activation marker CD69 was found at the highest frequencies in the adjuvanted mice ( Figure 3B ), but the individual variability within the groups was too high to reach statistical significant differences.

The presence of vaccine–induced, antigen-experienced CD4⁺ T_RM was also assessed in the lungs of mice via the surface staining with anti-CD44 in combination with the intravenous staining (gating strategy shown in Supplementary Figure S2 ). Although the overall proportion of CD44⁺ cells among lung-derived CD4⁺ T-cells was already increased in all vaccine groups, statistical significance was reached when considering specifically the CD45 iv^- compartment ( Figure 3C ).

Similarly to the B_RM analyses, the presence of CD11a, CD69 and CD103 were used for a detailed characterization of the CD4⁺ T_RM population (gating strategy shown in Supplementary Figure S2 ).

Here, a heterogeneous T_RM profile was induced by vaccination and a small percentage of cells expressing CD103 was also detected ( Figure 3D ). However, CD11a⁺CD69⁺CD103^- was the most predominant phenotype found among CD4⁺ T_RM, followed by the single expression of CD11a. The percentage of these T_RM phenotypes was significantly increased upon vaccination, and the mucosal adjuvant Ad-TGFβ seemed to have a superior effect on their formation compared to Ad-CCL17-adjuvanted or the non-adjuvanted vaccine. Since TGFβ is not only important for T_RM formation, but also involved in the development of regulatory CD4⁺ T cells (T_reg), we also quantified Foxp3⁺CD25⁺ CD4⁺ T-cells in the lungs of immunized mice (gating strategy shown in Supplementary Figure S3B ). Even though, compared to the naïve animals, frequencies were increased and reached statistical significance for the two adjuvanted groups, there was no significant difference in the frequency of T_reg among the vaccine groups ( Supplementary Figure S3C ). Therefore, the mucosal delivery of TGFβ and CCL17 did not seem to further increase the frequency of vaccine induced T_reg in the lungs of mice.

Next, we investigated the effect of the mucosal vaccination on the development of CD8⁺ T_RM, as these cells are considered the first line of defense against respiratory tract infections. Circulating and resident cells were distinguished by an intravenous labeling with anti-CD45 and then further characterized with APC-labeled MHC-I pentamers loaded with either the NP_147-155 or F_85-93 peptides (gating strategy shown in Supplementary Figure S2 ). The intranasal immunization with Ad-HA/NP/F significantly increased the presence of total CD8⁺ T-cells in the lungs of mice, which was further enhanced in combination with either Ad-TGFβ or Ad-CCL17 ( Supplementary Figure S4A ). Within this T-cell compartment, we could detect both NP- and F-specific CD8⁺ T-cells after mucosal immunization ( Figures 4A, D ), which belonged almost exclusively to the resident compartment ( Supplementary Figures S4B, C ). The administration of either Ad-TGFβ or Ad-CCL17, as mucosal adjuvants, significantly increased the frequencies of NP-specific CD8⁺ T_RM in comparison to the naïve group ( Figure 4A ). However, only the adjuvant Ad-TGFβ had a slight impact on the formation of F-specific CD8⁺ T_RM ( Figure 4D ). Here, a detailed analysis of the T_RM phenotype was also performed via the surface staining with CD69 and CD103 (gating strategy shown in Supplementary Figure S2 ). Although most of the NP- and F-specific CD8⁺ T_RM expressed CD69 and CD103, a fraction of these cells was also found to be CD69⁺CD103^- ( Figures 4B, E ). The frequency of CD69⁺CD103⁺ NP-specific CD8⁺ T_RM was significantly increased after the mucosal immunization with the adjuvanted vaccines, but only Ad-TGFβ increased vaccine-induced CD69⁺CD103⁺ F-specific CD8⁺ T_RM.

Beside the phenotypic analysis, resident CD8⁺ T-cell responses were also functionally analyzed by intracellular cytokine staining. Lymphocytes isolated from the lungs of mice were restimulated with MHC-I restricted NP and F peptides, followed by the detection of IFNγ, TNFα and IL-2 as well as the granulation marker CD107a within this CD8⁺ T-cell population ( Supplementary Figure S5 ). In all immunized groups, IFNγ- and TNFα–producing CD8⁺ T_RM were observed upon in vitro restimulation of lung lymphocytes, with NP- and F-specific peptides, and this was further increased when using the mucosal adjuvants ( Figures 4C, F ).

After investigating vaccine immunogenicity, we also determined its protective effect against RSV. For this purpose, fifty-six days after vaccination, mice were infected with 5x10⁶ PFU of RSV-A2 ( Figure 1 ). Body weight was measured throughout infection, however no substantial weight loss was observed ( Figure 5A ). Five days after challenge, animals were euthanized and viral load was measured in BAL samples. Immunized mice showed significantly lower copy numbers of viral RNA compared to the naïve group ( Figure 5B ). However, no statistical differences were observed between non-adjuvanted and adjuvanted animals.

Next, we assessed vaccine efficacy against two different heterologous IAV strains. Fifty-six days after vaccination, mice were infected with 2000 PFU of either pH1N1 or H3N2 ( Figure 1 ). Body weight was measured during infection and, at day 8 viral load and lung damage were determined. Independently of the adjuvants, the mucosal immunization with Ad-HA/NP/F protected mice against weight loss upon infection with the pH1N1 ( Figure 6A ). In contrast, unimmunized mice had already lost 4.2% ± 1.5% of their initial weight three days post infection (p.i.) and continuously decreased until the experimental endpoint was reached on day 8 (25.2% ± 8.1%; Figure 6A ). Accordingly, all immunized mice had a significant reduction in viral RNA levels ( Figure 6B ). Furthermore, the amount of protein found in the BAL was highly reduced in the vaccinated animals indicating reduced damage of lung tissue ( Figure 6C ).

The challenge with the more distant H3N2 strain led to a symptomatic infection also in the immunized groups, since all animals showed slight weight loss three days p.i. ( Figure 6D ). Nevertheless, all vaccinated animals regained weight starting at day 7 and recovered from disease, whereas naïve mice reached the humane endpoint at day 8 p.i. ( Figure 6D ). In line with this attenuated disease phenotype, significant reductions in viral loads and protein levels were found in the BAL of immunized animals ( Figures 6E, F ).