Intranasal Delivery of MVA Vector Vaccine Induces Effective Pulmonary Immunity Against SARS-CoV-2 in Rodents

Antigen-specific tissue-resident memory T cells (Trms) and neutralizing IgA antibodies provide the most effective protection of the lungs from viral infections. To induce those essential components of lung immunity against SARS-CoV-2, we tested various immunization protocols involving intranasal delivery of a novel Modified Vaccinia virus Ankara (MVA)-SARS-2-spike vaccine candidate. We show that a single intranasal MVA-SARS-CoV-2-S application in mice strongly induced pulmonary spike-specific CD8+ T cells, albeit restricted production of neutralizing antibodies. In prime-boost protocols, intranasal booster vaccine delivery proved to be crucial for a massive expansion of systemic and lung tissue-resident spike-specific CD8+ T cells and the development of Th1 - but not Th2 - CD4+ T cells. Likewise, very high titers of IgG and IgA anti-spike antibodies were present in serum and broncho-alveolar lavages that possessed high virus neutralization capacities to all current SARS-CoV-2 variants of concern. Importantly, the MVA-SARS-2-spike vaccine applied in intramuscular priming and intranasal boosting treatment regimen completely protected hamsters from developing SARS-CoV-2 lung infection and pathology. Together, these results identify intramuscular priming followed by respiratory tract boosting with MVA-SARS-2-S as a promising approach for the induction of local, respiratory as well as systemic immune responses suited to protect from SARS-CoV-2 infections.

Antigen-specific tissue-resident memory T cells (Trms) and neutralizing IgA antibodies provide the most effective protection of the lungs from viral infections. To induce those essential components of lung immunity against SARS-CoV-2, we tested various immunization protocols involving intranasal delivery of a novel Modified Vaccinia virus Ankara (MVA)-SARS-2-spike vaccine candidate. We show that a single intranasal MVA-SARS-CoV-2-S application in mice strongly induced pulmonary spike-specific CD8 + T cells, albeit restricted production of neutralizing antibodies. In prime-boost protocols, intranasal booster vaccine delivery proved to be crucial for a massive expansion of systemic and lung tissue-resident spike-specific CD8 + T cells and the development of Th1 -but not Th2 -CD4 + T cells. Likewise, very high titers of IgG and IgA anti-spike antibodies were present in serum and broncho-alveolar lavages that possessed high virus neutralization capacities to all current SARS-CoV-2 variants of concern. Importantly, the MVA-SARS-2-spike vaccine applied in intramuscular priming and intranasal boosting treatment regimen completely protected hamsters from developing SARS-CoV-2 lung
In contrast to intramuscular vaccine application, inhalative vaccination induces high levels of antigen-specific IgA antibodies and tissue-resident memory T cells (Trms) in the respiratory tract (17)(18)(19). Secretory IgA antibodies cover lung mucosa and protect the lungs by immune exclusion, complexing, and neutralization of invading microorganisms (20). Trms originate from effector T cells that enter the lungs during initial infection and, after pathogen clearance, remain resident within the lung parenchyma and in the airways (21). Upon re-infection, Trms get rapidly activated, secrete cytokines, proliferate, and recruit other leukocytes, thus enabling accelerated clearance of the pathogens (22). Along these lines, respiratory delivery of coronavirus 2019 (COVID-19) vaccine candidates induced robust lung immunity and protected against SARS-CoV-2 infection (23)(24)(25)(26). Moreover, inhaled COVID-19 vaccine candidate based on parainfluenza virus type 5 vector efficiently blocked animal-to-animal transmission of the SARS-CoV-2 (27). Currently, there are several COVID-19 vaccine candidates in development for inhaled administration (28) and the latest clinical results proved safety and immunogenicity of aerosolized adenovirus type-5 vector-based COVID-19 vaccine (29).
We already hypothesized that respiratory delivery of a Modified Vaccinia virus Ankara (MVA)-based vaccine would be a favorable approach to combat COVID-19 (30). MVA is a highly attenuated strain of vaccinia virus that is growth adapted to avian cells and replication-deficient in mammalian cells (31,32). Nevertheless, MVA retained the ability to infect mammalian cells and to induce efficient humoral and cellular immune responses (33) and became a third generation smallpox vaccine, now licensed in Canada and the European Union (34). Moreover, unaffected synthesis of viral proteins in MVA-infected cells enables high levels of recombinant protein production from recombinant MVA (35,36), making it an advanced viral vector platform for numerous vaccines, including those against coronaviruses (37)(38)(39)(40).
Beside systemic immune responses, respiratory delivery of MVA induces strong antigen-specific CD8 immunity and IgA production within lungs (41)(42)(43). Moreover, as we have described in detail intranasal application of MVA also leads to development of bronchus-associated lymphoid tissue (BALT) in the lungs of mice (44). BALT are tertiary lymphoid organs composed of B cell follicles surrounded by a para-follicular area rich in T cells and antigen-presenting cells (44)(45)(46) and serve as a general priming site for the induction of adaptive immune responses (44).
In this study, we used respiratory delivery of a MVA vector vaccine expressing the spike (S) protein of SARS-CoV-2 (MVA-SARS-2-S) to induce protective lung and systemic immunity in rodents. In line with previous preclinical and clinical data (37,41,(47)(48)(49), respiratory delivery of MVA-SARS-2-S has been shown to be both safe and immunogenic. More importantly, our profound analysis in mice revealed that exclusively immunization protocols involving intranasal boosting induce S-specific Trms and neutralizing IgA antibodies in lung. Particularly, neutralizing antibodies developed by MVA-SARS-2-S were also effective against SARS-CoV-2 VoC and this immunization protocol also efficiently protected hamsters from SARS-CoV-2 infection.

Experimental Animals
C57BL/6N mice were purchased from Charles River and bred and maintained in the Central Animal Facility of Hannover Medical School (Hannover, Germany) under specific pathogenfree conditions and used for experiments at the age of 7 -13 weeks. Male syrian hamster (10 week-old, Mesocricetus auratus; breed HsdHan ® :AURA) were purchased from Envigo RMS Inc. (Indiampolis, United States). Hamsters were maintained under specified pathogen-free conditions, had free access to food and water, and were allowed to adapt to the facilities for at least one week before vaccination experiments were performed. All animal experiments were handled in compliance with the European and national regulations for animal experimentation (European Directive 2010/63/EU; Animal Welfare Acts in Germany) and Animal Welfare Act, approved by the Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES) Lower Saxony, Germany).

Immunization With MVA-Spike
The recombinant virus MVA-SARS-2-S was constructed and characterized in detail as described elsewhere (50). In the present study, a quality controlled vaccine lot was applied serving for preclinical characterization of MVA-SARS-2-S as in preparation of clinical Phase-1 evaluation. For i.n. administration, animals were lightly anaesthetized and indicated plaque forming units (PFU) of MVA-SARS-2-S diluted in 40 ml saline, 15 mM Tris pH7.7, 3% sucrose, 0.005% Tween 80 were applied to the nostrils. For intra muscular immunization, 25 µl of indicated PFU of MVA-SARS-2-S dissolved in 30 mM Tris pH7.7, 6% sucrose, 0.01% Tween 80 were injected into the quadriceps at one of the hind legs.
In mice, we tested MVA-SARS-2 in two different setups: (i) prime only (PO), (ii) primeboost (PB) vaccination regiment. In the PO setting, mice received one single i.n. priming dose of MVA-SARS-2-S doses indicated at day 0 and were analyzed at day 11, 24, 35 or 40 post immunizations. For PB immunization protocol, mice primed on day 0 received a second MVA-SARS-2 immunization (boost) on day 24 and were analyzed on day 40 post prime. In all PB experiments, three immunization protocols were compared (1): intranasal (na) priming and boosting (na-na) (2), intramuscular (mu) priming intra nasal boosting (mu-na) and (3) intranasal priming intramuscular boosting (na-mu). For priming and/or boosting of mice we used 10 6 , 10 7 or 10 8 PFU of the vaccine or vehicle as control. In some experiments untreated mice were used as controls. In some experiments consecutive blood samples were collected at 16 hours, 3 days and 10 days post boost. Hamsters were treated with MVA-SARS-CoV-2-S according to mu-na PB protocol. Immunizations were performed using intramuscular applications with vaccine suspension with 10 8 PFU recombinant MVA-SARS2-S or non-recombinant MVA (mock) into the quadriceps muscle of the left hind leg under isoflurane anesthesia. Boost vaccinations were performed using intranasal applications with vaccine suspension with 10 7 PFU recombinant MVA-SARS2-S or non-recombinant MVA (mock) under isoflurane anesthesia.

SARS-CoV-2 Infection
SARS-CoV-2 (BetaCoV/Germany/BavPat1/2020p.1, European Virus Archive Global #026V-03883) received from European Virus Archive was propagated in Vero E6 cells (ATCC #CRL-1586) in Dulbecco's Modified Eagle's Medium (DMEM) (Sigma-Aldrich GmbH) supplemented with 2% fetal bovine serum, 1% penicillin-streptomycin and 1% L-glutamine at 37°C. All infection experiments with SARS-CoV-2 were performed in the biosafety level 3 (BSL-3) laboratories at the Research Center for Emerging Infections and Zoonoses (RIZ), University of Veterinary Medicine Hannover, Germany. Hamsters were infected via the intranasal route with 1x10 4 tissue culture infectious dose 50 (TCID50) SARS-CoV-2 under isoflurane anesthesia. After SARS-CoV-2 challenge infection hamsters were monitored daily at least twice for well-being, health constitution and clinical signs, such as body temperature, anorexia, diarrhea/loose stool, vomiting, lethargy/depression and respiratory symptoms using a clinical score sheet. Weights of all hamsters were checked daily.

Sample Collection From Mice
To label leukocytes present in the blood vessels of lungs, we administered 5 µg of FITC-or BUV717-labeled anti-CD45.2 mAb (clone  in mice terminally anesthetized with an overdose of ketamine/xylazine. Within 3-5 minutes after antibody administration, blood was collected from saphenous veins and serum was separated by centrifugation and stored at -20°C until further analyses. After collection of blood, the spleen was resected before broncho-alveolar lavage (BAL) was performed through a plastic catheter clamped into the trachea using three separate lavage steps with 400µl, 300µl and 300µl of PBS each. Collected BAL samples were centrifuged for 10 min at 300 g at 4°C and the supernatant was stored at -20°C for subsequent analysis. BAL cells from pellet were re-suspended in 50 µl of remaining BAL fluid, counted to determine total cell number and immune phenotyped by spectral flow cytometry as described below. Finally, lungs and bronchial lymph nodes were resected. Right lung lobes were dissected and stored in PBS on ice until single cell suspensions were made. Left lung lobes were filled with mixture of PBS : OCT (1:1) and frozen in OCT (Tissue-Tek) on dry ice for histological analysis.
For flow cytometry, single cell suspensions of spleen and bronchial lymph nodes were obtained by meshing the organs through 70 µm cell strainers. Right lung lobes were digested in RPMI supplemented with 10% FCS, 0.025 mg/ml DNAse I (#11284932001, Roche), 0.5 mg/ml Collagenase D (#11088866001 Roche) for 45 minutes at 37°C. Single cells were separated by smashing samples through 70 µm cell strainers. Single cell suspensions of spleen and lung were subjected to hypertonic red blood cell lysis.
In hamster experiments, animals were sacrificed 6 days post SARS-CoV-2 infection and serum as well as lung tissue samples were taken for analysis of virus loads.

Immunophenotyping Using Spectral Flow Cytometry
For deep phenotyping of immune cells isolated from different organs non-specific antibody binding was blocked by incubating samples 10% rat serum for 10 minutes at 4°C. Next, cells were incubated with MVA-or SARS-CoV-2-S-peptide loaded tetramers for 15 minutes at 37°C. The tetramers were prepared by loading Vaccinia virus WR epitope B8R 20-27 (TSYKFESV) or an 8-amino acid (aa) long immune dominant spike peptide V8L 539-546 (VNFNFNGL) on empty H-2kb tetramers according to the manufacturer's protocol (Tetramer-Shop). Without washing, antibody mixes were added and cells were incubated for additional 15 minutes at 37°C. The full list of antibodies used is listed in Supplementary Table S1. After washing, samples were acquired on Cytek Aurora spectral flow cytometer (Cytek) equipped with five lasers operating on 355nm, 405nm, 488nm, 561nm and 640nm. All flow cytometry data were analyzed using FCS Express V7 (Denovo) or FlowJo V10 (BD).

Measurement of Viral Burden
Tissue samples of sacrificed hamsters were excised from the right lung lobes and homogenized in 600 µl DMEM containing antibiotics (penicillin and streptomycin, Gibco). Tissue was homogenized using the TissueLyser-II (Qiagen), and aliquots were stored at -80°C. Virus titers were determined on Vero cells as median tissue culture infectious dose (TCID50 units). Briefly, Vero cells were seeded in 96-well plates and serial 10-fold dilutions of homogenized lung samples in DMEM containing 5% FBS. After incubation for 96 hours at 37°C, cytopathic effect in Vero cells was evaluated and calculated as TCID50 unit per gram of hamster lungs by using Reed-Muench method. The detection limit of the assay was 316 TCID50 unit. Thus, for samples where no cytopathic effect was detected, data points were set to 157 TCID50 unit for statical analysis purpose.

SARS-CoV-2-S Protein Peptides and Epitope Prediction
253 overlapping peptides, each 15 aa long with 10 aa overlap, spanning the whole length of SARS-CoV-2-S or predicted immune dominant peptides were synthesized at >95% purity (GeneScript). For immune dominant peptide prediction, a fulllength SARS-CoV-2 sequence was used (YP_009724390.1 from National Center for Biotechnology Information) to predict potential CD8 T cell-immuno dominant 8-11 aa long peptides applying MHC I Binding Tool (with filters summary IC50 < 1000 and Percentile Rank < 10) and MHC I Processing Tool from IEDB (https://www.iedb.org/). For prediction of potential CD4 T cell epitopes, we used IEDB MHC II binding tool (with summary IC50 < 1000) and from that list only 15-aa peptides passing MHC II binding threshold on RankPep (http://imed. med.ucm.es/Tools/rankpep.html) were used. A full list of 13 MHC I and 18 MHC II peptides are shown in Supplementary  Table S2. All lyophilized peptides were reconstituted at a stock concentration of 50 mg/ml in DMSO (Sigma-Aldrich) except for 9 SARS-CoV-2-S overlapping peptides (number 24,190,191,225,226,234,244,245 and 246) that were dissolved at 25 mg/ml due to solubility issues. All peptides were stored at -80°C before use.

Immunohistology
Frozen tissue blocks of mice lungs were cut in 8µm thick cryo sections and fixed 10 minutes with acetone on ice. Fixed cryo sections containing main stem bronchi were rehydrated for 5 minutes in TBS (PBS, 0.05% Tween) and washed twice with TBS. After washing, cryo sections were blocked (5% rat serum,  Table S3) and washed twice with TBS. The sections were then stained with DAPI (1 µg/ml) for 3 minutes and washed with TBS again before there were embedded in Moviol. Images were acquired using an Axioscan Z1 (Zeiss). The amount of BALT was quantified as described before (51). In brief, panoramic images of whole sections from different central planes (close to main bronchi and vessels) were collected and analyzed. Individual BALT structures were enumerated, their surface measured, and the cumulative BALT size was calculated as the sum of surface areas of all individual BALT structures present on one central lung section (ZEN 2.3 blue software, Zeiss).

Plaque Reduction Neutralizing Test for SARS-CoV-2
Sera of infected hamsters were used to analyze neutralization capacity against SARS-CoV-2 as previously described (53). We 2-fold serially diluted heat-inactivated serum samples in Dulbecco modified Eagle medium starting at a dilution of 1:10 in 50 mL. We then added 50 ml of virus suspension (600 TCID50) to each well and incubated at 37°C for 1 h before placing the mixtures on VeroE6 cells. After incubation for 1 h, we washed, cells supplemented with medium, and incubated for 8 h. After incubation, we fixed the cells with 4% formaldehyde/PBS and stained the cells with a polyclonal rabbit antibody against SARS-CoV-2 nucleoprotein (clone 40588-T62, Sino Biological) and a secondary peroxidase-labeled goat anti-rabbit IgG (Dako, Agilent). We developed the signal using a precipitate forming TMB substrate (True Blue, KPL SeraCare) and counted the number of infected cells per well by using the ImmunoSpot ® reader (CTL Europe GmbH). The reciprocal of the highest serum dilution allow reduction of >90% plaque formation was calculated as the serum neutralization titer (PRNT90).

Surrogate Virus Neutralization Test (sVNT)
sVNT test for detection of neutralizing anti-SARS-CoV-2-S antibodies was done as described before (54)

Enzyme-Linked Immunosorbent Assay (ELISA)
To determine anti-spike IgA and IgG antibodies in sera or BAL samples, ELISA plates were coated overnight at 4°C with 200 ng SARS-CoV-2 trimer in its perfusion conformation (55) fused to mNEONgreen and expressed in Drosophila S2 cells (56) (detailed production will be described elsewhere) in 100µl PBS per well. On the next day, plates were washed three times with washing buffer (PBS, 0.05% (w/v) Tween 20), blocked with 2% (w/v) BSA for 1h at 37°C and incubated for 2h at RT with serial dilutions of sera or BAL. For the detection of anti-spike specific isotype antibodies goat anti-mouse IgG-Fc antibody HRPconjugated (cat.# 1013-05, SouthernBiotech) and goat antimouse IgA-Fc antibody HRP-conjugated (cat.# 1040-05, SouthernBiotech) were used at a final dilution of 1:4000. All assays were conducted in duplicates and quantified using a SpectraMax iD3 microplate reader (Molecular Devices).

Statistics
Statistical analysis was done using Prism 7 or 8 (GraphPad). Statistical analysis was done on log-transformed values using ordinary or Welch's ANOVA followed by Dunnett's T3 multiple comparisons test. P values < 0.05 were considered statistically significant.

Intranasal MVA-SARS-2-S Immunization Induces BALT Formation
In agreement with our previous results (44) Figure 1C). Similarly, at d11 p.i. we observed massive recruitment of T cells, B cells, macrophages, monocytes, NK cells, and dendritic cells in lung tissue, as revealed by the absence of staining of an anti-CD45 mAb that had been intravenously (i.v.) injected 10min before the mice were sacrificed; Figures 1F-H and Supplementary Figure 1D). The number of T cells and interstitial macrophages within the tissue remained elevated even 40 days post i.n. MVA-SARS-2-S delivery ( Figure 1H and Supplementary Figure 1D), reflecting the presence of BALT observed by histology ( Figure 1B). To specifically measure S-specific CD8 + T cell responses, we established an ex vivo stimulation assay with 10-amino acid overlapping 15-mer peptides covering the entire S-protein.
Quantification of intracellular interferon g (IFN-g) expression by flow cytometry revealed that immunization with MVA-SARS-2-S induced CD8 + T cells specific to peptides 1-129 (amino acids 1-655), covering almost the entire S1 domain of the S protein (Supplementary Figures 3A, B). In addition, the Immune Epitope Database (IEDB) (57) allowed us to identify an immunodominant SARS-CoV-2-S H2-Kb epitope at S539-546 (VNFNFNGL; S-V8L) of the S1 C-terminal domain (Supplementary Figures 3C-F and Supplementary Table S2). Integrating S-V8L tetramers in spectral FACS analysis, we found that S-specific CD8 + T cells peaked in lungs, BALs, bLNs and spleens at d11 p.i. and slowly decreased to day 40 (Figures 2A, B and Supplementary Figure 3G). Importantly, even at this time-point their number remained significantly elevated in lungs, BAL, and spleen compared to non-immunized controls, indicating development of immunological memory. The majority of those cells produced tumor necrosis alpha (TNF-a) in addition to IFN-g upon re-stimulation with S protein-derived peptides (Supplementary Figure 3H). Of note, the kinetics of S-specific CD8 + cells matched the kinetics of CD8 + T cell specific for a known immuno-dominant epitope of MVA in C57BL/6 mice (TSYKFESV; MVA-B8R) (58) (Supplementary Figures 3I, K).
Analysis of ex vivo cultured cells stimulated with S proteinderived peptides further revealed that S-specific IFN-g producing CD4 + T cells were also present in lungs, bLNs, and spleens of immunized animals up to day 24 post i.n. vaccination (Figures 2C, D and Supplementary Figure 4A, B). Importantly, we could not detect IL-5 producing CD4 + T cells upon ex vivo re-stimulation (Supplementary Figure 4C). Together these data indicate that i.n. MVA-SARS-2-S vaccination induces, in addition to CD8 immunity, also specific T helper type 1 (Th1), but not Th2 responses.
Profiling of B cell responses after a single i.n. dose of the vaccine revealed increased numbers of germinal center (GC) B cells (CD19 + CD138 -IgD -IgM -GL7 + CD73 + ) in bLN and lungs but not in spleen (Supplementary Figures 4D, E). Despite the increase in GC B cell counts, anti-S antibodies were only present at relatively low levels in all mice ( Figure 2E). Mean EC 50 values were in the range of 1:100 (10 2 ) at day 11 and increased approximately 10 fold at later time points analyzed ( Figure 2F). Similarly, BAL from immunized mice also possessed low anti-S IgG antibody titers that stayed constant over time, but no anti-S IgA antibodies could be detected ( Figure 2G). Together, these data indicate that a single i.n. immunization with MVA-SARS-2-S induces strong local and systemic cellular but only moderate humoral immune responses directed against the spike protein. ( Figure 3A). According to our central hypothesis that delivery of the vaccine to the respiratory tract is required for the induction of protective immunity in the lung, we compared single intranasal application using 10 7 PFU (PO 7na-d24 ) with three different prime-boost immunization protocols (1): priming and boosting by i.n. vaccination (PB 7na-7na ) (2), i.n. priming followed by an intramuscular boost (PB 7na-7mu ), and (3) i.m. priming followed by an i.n. boost (PB 7mu-7na ). Of note, when compared to PO 7na-d24 group, increased amounts of BALT were only induced by i.n. but not intramuscular boost, independent of the priming route applied ( Figure 3B). This finding was confirmed by flow cytometry quantifying the major immune cell subsets in BAL and lung (Supplementary Figures 5A, B). The composition of T cells residing in the lung parenchyma was comparable between the different prime-boost immunization protocols but groups that received the i.n. boost had considerably more cells than primed-only mice or mice in PB 7na-7mu group ( Figures 3C, D). Likewise, CD8 + T cells specific for S-V8L as well as MVA-B8R were strongly increased in lungs, bLNs and BALs of the PB 7na-7na and PB 7mu-7na groups, while such differences were not found in spleens ( Figure 3E and Supplementary Figure 5C). Further, in vitro-re-stimulation with S-specific peptides identified increased frequencies of IFNg-producing CD4 + T cells in the lungs ( Figure 3F).  In addition to stronger cellular immune responses, all primeboost protocols induced also significantly higher S-specific IgG titers in serum compared to the mice receiving only a single i.n. vaccination ( Figure 3G). Similar results were found for anti-S IgG in the BAL, although mice receiving intramuscular boost developed clearly lower titers compared to intranasal boosted mice ( Figure 3H). Most importantly, protocols applying i.n. boost (sPB 7na-na and PB 7mu-na ) induced anti-S IgA antibodies in the BAL (Figure 3I), the antibody isotype that protects respiratory mucosal surfaces from infection with air-borne pathogens.

High Prime Dose Induces Tissue Specific S-Specific T-Cells
Our results so far revealed that prime-boost protocols give stronger immune responses than prime-only vaccine delivery and that intranasal boosting induced the best local adaptive immune response. Next, we investigated immune responses after priming with optimal i.m. priming dose (10 8 PFU) for MVAbased vaccines (50,62), followed by lower booster doses of i.n. vaccine delivery (10 7 or 10 6 PFU) ( Figure 4A). To monitor cytokines induced by i.n. vaccine delivery, we assessed serum concentrations of 11 cytokines at different times point after boosting (Supplementary Table S4). The most prominent changes were observed for IFN-g that was elevated dosedependently but transiently, increasing approximately 150-fold and 60-fold with 10 7 PFU and 10 6 PFU, respectively at 16 hr post boost and returning to the levels of control mice by day 10 post boost ( Figure 4B and Supplementary Table S4). Similar patterns, albeit to a much lower extent were also observed for the T cell proliferation cytokine IL-2, the Th1-cytokine, IL-12p70, as well as IL-6 (approximately a 3-fold increase with the higher vaccine dose). Interestingly, i.n. vaccine boost also induced transient and dose-dependent increase in serum levels of the Th2 cytokine IL-5 (13-fold and 8-fold with higher and lower vaccine dose, respectively; Figure 4B), a cytokine important for maturation of IgA-secreting B cells (63). Importantly, other Th2 cytokines such as IL-4 and IL-13 were not affected (Supplementary Table S4). In line with no effects on body weight and body temperature, the concentrations of the classical pro-inflammatory cytokines TNF-a and IL-1b, the regulatory cytokine IL-10 or Th17 cytokine IL-17A were not affected by the boost (Supplementary Figures 6A, B and Supplementary Table S4) underlining the excellent tolerability of MVA-SARS-2-S after i.n. boost application.
Interestingly, the higher booster dose (10 7 PFU) induced slightly less BALT than the lower dose (10 6 PFU; Figure 4C), accompanied by strong accumulation of all major immune cell subsets in BAL and lung in both boosted groups ( Supplementary  Figures 6C-E). In contrast, the higher booster dose induced slightly more S-V8L CD8 + T cells in lung and bLN, while both protocols were equally efficient in the spleen ( Figure 4D). Likewise, both vaccine doses induced an identical TNF-a/ IFN-g production profile in CD8 + T cells following ex vivo peptide re-stimulation (Supplementary Figure 6F). Detailed flow cytometric profiling of S-V8L CD8 + T cells indicated that they cluster according to the organ of origin ( Figure 4E). Importantly, expression of Trm markers CD103 and CD69 was largely restricted to lung and bLN CD8 + T cells ( Figure 4E and Supplementary Figure 6G).
Similarly to CD8 + T cell responses, the low and the high booster doses were equally efficient in inducing S-specific Th1 cells in lung, bLN and spleen ( Figure 4F). Of note, we found no evidence of IL-5-secreting Th2 or IL-17A-secreting Th17 cells in any organ upon re-stimulation with spike-derived peptides (Supplementary Figures 6H, I). In summary, both booster protocols, PB 8mu-7na and PB 8mu-6na , induced very similar and effective cellular immunity.

MVA-SARS-2-S Boosting Induces High Titers of Neutralizing Anti-SARS-CoV-2-S Antibodies
In addition to S-specific T cells, both booster doses were equally efficient in inducing anti-S antibodies in serum and BAL ( Figures 5A-C). We further measured neutralization capacity of those antibodies in samples of the PB 8mu-7na group using our recently established surrogate virus neutralization tests (sVNT) for different SARS-CoV-2 VoC (14,54,64). The sVNT is based on ELISA technology and allows high-throughput quantitative analysis of neutralizing antibody (nAb) levels by measuring the reduction in the binding of the receptor-binding-domain (RBD) of SARS-CoV-2-S protein to ACE2 in vitro. Compared to serum samples from unimmunized mice, all mice from PB 8mu-7na group developed high amounts of nAb in the serum (Figures 5D, E). Furthermore, these antibodies also efficiently blocked alpha and delta variants, while the protection against beta and gamma variants was less effective but still present in 7 out of 10 immunized mice ( Figures 5D, E). Importantly, nAb against Wuhan, alpha, and delta but not against beta and gamma variants were detectable also in BAL (Figures 5F, G (65,66). We immunized two groups of golden Syrian hamsters, one with MVA-SARS-CoV-2-S and the other with MVA vector (MVA-WT), using the PB 8mu-7na protocol, which proved to be the most effective protocol for in mice. Forty days post prime, the hamsters were infected with 1 × 10 4 tissue culture infectious dose 50 (TCID50) SARS-CoV-2 (isolate BavPat1/ 2020 isolate, European Virus Archive Global #026V-03883). The course of infection was tracked for 6 days by animal weight loss as well as changes in spontaneous behavior and general condition summarized in a clinical score ( Figure 6A). In contrast to MVA-WT-immunized control hamsters, vaccination with MVA-SARS-2-S completely blocked development of clinical signs of infection ( Figures 6B, C). Moreover, at day 6 post SARS-CoV-2 infection,  Table S2) for 6hr. Pooled data from 2 experiments with n = 6-10 per group. Individual values (signs) and mean group value (line). Statistical analysis was done on log-transformed values using ordinary or Welch's ANOVA followed by Dunnett's T3 multiple comparisons test. **p < 0.01, ***p < 0.001, ****p < 0.0001. when substantial virus RNA loads were found in the lungs of MVA-WT-immunized control hamsters, minimal SARS-CoV-2 titers were found in only one of 7 lungs from MVA-SARS-2-S-immunized animals ( Figure 6D). Similarly, using immunohistochemistry we detected SARS-CoV-2 nucleoprotein in the lungs of all MVA-WTimmunized control hamsters but in none of the MVA-SARS-2-Simmunized animals ( Figures 6E, F). These spots of viral replication in the lungs of all MVA-WT-immunized control hamsters were usually located at areas of prominent leukocyte infiltration and consolidation that affected more than one quarter of the lungs ( Figures 6G, H and Supplementary Figure 7). In contrast, 6 out of 7 hamsters immunized with MVA-SARS-2-S had only single small foci of mild inflammatory infiltrates, affecting max. 1% of the lungs ( Figures 6G, H and Supplementary Figure 7). These data indicated that immunity induced by respiratory vaccination with MVA-SARS-2-S so rapidly cleared SARS-CoV-2 infection that infection and inflammation could not spread to large lung areas. Interestingly, we did not found any evidences of BALT structures in any of the analyzed groups. While it is possible that the SARS-CoV-2 infection led to the dissolution of the BALT structures, these data could also be a consequence of species-dependent differences in the MVA-induced BALT development that require further investigation. Lastly, we measured levels of neutralizing antibodies in sera of control and MVA-SARS-2-S-immunized hamsters. SARS-CoV-2 S protein-pseudotyped-vesicular stomatitis virus vector-based neutralization assay revealed that at this time point post infection the MVA-WT immunized animals also had detectable nAb against SARS-CoV-2 in serum ( Figures 7A, B), which is in line with previous reports (67,68). Not surprisingly, the nAb levels in MVA-SARS-CoV-2-S immunized animals were significantly higher ( Figures 7A,  B). In line with the data from vaccinated mice, sVNT revealed that these nAbs are very effective against Wuhan, alpha, and delta SARS-CoV-2 variants, while protection against beta and gamma variants was less prominent but still clearly observable in MVA-SARS-CoV-2-S vaccinated hamsters ( Figure 7C).

DISCUSSION
Our study showed that respiratory delivery of a MVA-SARS-2-S vaccine efficiently protected hamsters from SARS-CoV-2 infection. Importantly, not only was the virus completely cleared from the lungs but vaccinated hamsters also had only minimal traces of resolving lung inflammation. The rapid clearance of SARS-CoV-2 combined with the minimal immune response can only be attributed to the existence of two major pillars of adaptive immunity in the respiratory tract: the local production and secretion of IgA antibodies and tissue resident memory T cells directed against the spike protein. Indeed, our extensive profiling in mice indicated that an intra nasal vaccine boost is crucial for the induction of high titers of neutralizing anti-S-RBD antibodies in BAL. Importantly, these antibodies were successful in neutralizing two predominant SARS-CoV-2 VoC, alpha and delta. These neutralizing anti-S-RBD antibodies play a crucial role in preventing (re-)infections with the virus, as they block binding of the S protein to its cellular receptor ACE2, thus preventing virus cell entry (54,64,69,70). Moreover, i.n. boosting led to a massive accumulation of spike-specific CD8 + T cells expressing CD103 and CD69, markers for tissue residency in the lungs (71). In case neutralizing antibody protection is breached, these CD8 + T cells can rapidly be re-activated for instant and efficient virus control (72). Although tissue-resident memory CD8 T cells might have limited longevity in the lung (73,74), they appear to be crucial for limiting the severity of lung pathology induced by immune response to viral infection (75). Vaccination with MVA-SARS-2-S via the respiratory tract, therefore, provided the local armamentarium needed for effective protection against SARS-CoV-2, which is in agreement with previous reports indicating that i.n. administration of MVAbased vaccines can protect against various respiratory pathogens (38,43,(76)(77)(78). Additionally, intranasal delivery of MVA-SARS-2-S in mice also led to the induction of BALT, as reported earlier for MVA, the backbone for the current vaccine (44)(45)(46). Interestingly, BALT induced by i.n. booster delivery of MVA-SARS-2-S affected larger lung areas and persisted longer than BALT induced by a single i.n. dose of the vaccine, which is in line with the hypothesis that the extent and the nature of the inflammatory stimuli control BALT formation (79)(80)(81). On the other hand, hamsters immunized with the same vaccine did not develop BALT, suggesting species-specific mechanisms important for BALT induction after MVA administration. Of note, lack of BALT did not affect vaccine efficacy. BALT does not seem to contribute to the maintenance of IgA-producing plasma cells (30,79) or tissue resident memory CD8 + T cells (82). However, BALT serves an important site for T cell activation as well as B cell selection and maturation fighting pathogens that reach the lower respiratory tract (44,79). MVA-SARS-2-Sinduced BALT might therefore provide the infrastructure that also facilitates immune responses to other pathogens. In line with this idea, the presence of BALT has been shown to significantly accelerate clearance of different respiratory viruses, including influenza, SARS-CoV-1 and pneumovirus (83).
Interestingly, intramuscular boosting of intranasally primed mice failed to induce neutralizing antibodies and to stimulate the expansion of spike-specific resident T cells within lungs, further emphasizing the requirement of an antigen-driven local reactivation of virus-specific lymphocytes for the establishment of a protective environment in the lung. In contrast, the route used for priming seems to be of less importance, as we only observed rather subtle differences in the adaptive immune response between PB na-na and PB mu-na protocols ( Figure 3). These results confirm the findings that local vaccination induces anti-virus immunity at the organism level (84) and suggest that local (intranasal) boosting helps to target the immune system to the particular organ.
Profound immune responses in the lung induced by intranasal boosting were not achieved at the expense of systemic immunity. Induction of systemic immunity is important because infections with SARS-CoV-2 are not restricted to the respiratory system but can affect basically all other organs. Serum titers of S-specific IgG antibodies induced by the various PB protocols of the present study are comparable to titers reported after intramuscular or intraperitoneal prime-boost immunization with other MVAbased SARS-CoV-2 vaccines (50,60,61). Our data clearly indicate that upon successful systemic priming even low vaccine doses are sufficient for efficient boosting. Furthermore, we used the sVNT to quantify anti-S nAbs, which is done in a speciesindependent manner and thus allows comparison of immune responses between different species. It is therefore of interest to note that mice i.n. boosted with 10 7 PFU by the various PB protocols developed nAb titers approximately 10-fold higher than those present in convalescent individuals recovering from severe COVID-19 (54) and in the range of those vaccinees that received heterologous ChAdOx1-nCov-19 (Vaxzevria, ChAd)/BNT162b2 (Comirnaty, BNT) vaccination (14). These data strongly indicate that humoral immune responses induced by MVA-SARS-2-S could efficiently protect against SARS-CoV-2 infection.
The MVA-SARS-2-S vaccine used in this study was very well tolerated, similarly to other MVA vector vaccines delivered via the respiratory route in preclinical (42,43) and clinical (48) studies. Importantly, our data show that once the immune system has been successfully primed even low doses (10 6 PFU) of the MVA vector vaccines are sufficient to allow for full boosting via the respiratory tract. Of note, systemic priming followed by aerosol boosting of a MVA vector vaccine for tuberculosis, MVA85A, led to transient but clinically significant respiratory adverse effects in some volunteers (49). These adverse effects can probably be attributed to the high vaccine dose (5x10 7 PFU) used for aerosol delivery and/or to cellular impurities possibly present in the vaccine. Although MVA-based vaccines appear to be very well tolerated after intramuscular application (48,49,85), the respiratory route of vaccine delivery is not well established in humans. Therefore, it will be important to carefully assess safety of the respiratory delivery of MVA-SARS-2-S in clinics and to monitor for any potential adverse effects, such as Bell's palsy reported after intranasal application of inactivated flu vaccine (86).
In summary, we demonstrate efficacy of MVA-SARS-2-S after intranasal delivery that is mediated by fulminant induction of cytotoxic CD8 + T cells, a strong Th1 response and high titers of neutralizing antibodies against SARS-CoV-2 in serum and BAL. Together with the results of other preclinical studies (23)(24)(25)(26)(27), these data support respiratory vaccine delivery as a promising application route to interfere with the spread of SARS-CoV-2 and to prevent COVID-19.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The animal studies were reviewed and approved by Niedersächsisches Landesamt für Verbraucherschutz und Lebensmittelsicherheit (LAVES), Lower Saxony, Germany.

FUNDING
This work was supported by Deutsche Forschungsgemeinschaft, (DFG, German research Foundation) Excellence Strategy EXC 2155 "RESIST" to RF and TK (Project ID39087428); by funds of the state of Lower Saxony (14-76103-184 CORONA-11/20) to RF and TK; by funds of the BMBF ("NaFoUniMedCovid19" FKZ: 01KX2021; Projects "COVIM"and "B-FAST") to RF and TK and Deutsche Forschungsgemeinschaft, Projektnummer 158989968 -SFB 900 (Projects B1 to RF and B10 to TK); Projektnumber Fo 334/7-1 and the Bundesministerium für Bildung und Forschung (DZIF 01.92100 and 01KX2026 to GSu, ZOOVAC 01KI1718 and RAPID 01KI1723G to AV) and by a BMBF (Federal Ministry of Education and Research) project entitled RAPID (Risk assessment in re-pandemic respiratory infectious diseases), 01KI1723G and by the Ministry of Science and Culture of Lower Saxony in Germany (14 -76103-184 CORONA-15/20) to WB.