Immunogenicity and Protective Efficacy of an Intranasal Live-attenuated Vaccine Against SARS-CoV-2 in Preclinical Animal Models

The global deployment of an effective and safe vaccine is currently a public health priority to curtail the coronavirus disease 2019 (COVID-19) pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Here, we evaluated a Newcastle disease virus (NDV)-based intranasal vectored-vaccine in mice and hamsters for its immunogenicity, safety and protective efficacy in challenge studies with SARS-CoV-2. The recombinant (r)NDV-S vaccine expressing spike (S) protein of SARS-CoV-2 administrated via intranasal route in mice induced high levels of SARS-CoV-2-specific neutralizing immunoglobulin A (IgA) and IgG2a antibodies and T cell-mediated immunity. Hamsters vaccinated with two doses of vaccine showed complete protection from clinical disease including lung infection, inflammation, and pathological lesions after SARS-CoV-2 challenge. Importantly, a single or double dose of intranasal rNDV-S vaccine completely blocked SARS-CoV-2 shedding in nasal turbinate and lungs within 4 days of vaccine administration in hamsters. Taken together, intranasal administration of rNDV-S has the potential to control infection at the site of inoculation, which should prevent both the clinical disease and transmission to halt the spread of the COVID-19 pandemic.

INTRODUCTION lungs of hamsters. Altogether, intranasal immunization of rNDV-S has the potential to control SARS-CoV-2 infection at the site of inoculation, which should prevent both virusinduced disease and transmission, representing an excellent valid option for the prevention of SARS-CoV-2 infection and associated COVID-19 disease.

Intranasal vaccination with rNDV-S successfully establishes a viral load with no accompanying adverse pathology
In order to explore the potential of intranasal live-attenuated vector vaccine against SARS-CoV-2, we engineered rNDV-S (i.e., avian orthoavulavirus 1, AOaV-1) encoding a human codon-optimized SARS-CoV-2 full-length S glycoprotein gene, including the ectodomain, transmembrane domain, and cytoplasmic domain. The S gene was cloned in a pre-optimized gene junction between phosphoprotein and matrix gene of NDV. As we demonstrated previously 33 , the rNDV-S replicated comparably to rNDV-WT in both cell culture and avian eggs and spreads within cells independently of exogenous trypsin, proposing a competitive vaccine candidate.
For safety and immunogenicity assessment of rNDV-S vaccine in mice, groups of 12-week-old BALB/c mice were immunized by intranasal inoculation with 10 6 PFU of test vaccine rNDV-S or wild type NDV (rNDV-WT) or were mock-vaccinated with phosphate buffer saline (PBS) (Fig. 1a). The rNDV-WT and one rNDV-S group of mice received a booster dose of 10 6 PFU of rNDV-WT or rNDV-S, respectively, a week later, while other groups were mock-boosted. All animals were euthanized on day 19 post-vaccination for safety assessment, as well as pathological and immunological responses. Mice were monitored daily for weight loss, health status, and feed intake. Initial starting weights did not significantly differ in any experimental group prior to treatment ( Fig. 1b and Supplementary Fig. 1), and there was no significant alteration in the percentage of weight in groups immunised with rNDV-S and the control mock-vaccinated group across the time course of the experiment (Fig.1b). The first dose of rNDV-S induced a significant reduction in percentage of basal weight at 2-and 3-days post-vaccination (DPV), but weights returned to comparable levels to mock immunised mice on the following day and for the rest of the time course (Fig.1b). A reduction in percentage weight was also seen at 14 and 15 DPV in the rNDV-S (prime + boost) group, but weights also quickly returned to mock levels for the rest of the study. Analysis of daily chow intake revealed no significant differences in daily chow consumption among groups (Fig. 1c). Mice were scored daily assessing movement, hair coat, temperature, eyes, and signs of hunching on a 14-point scale. No adverse clinical disease signs were observed with indicated daily score "0" in the duration of experiments. Following the end of the experiment at day 19 post-vaccination, the gross and histopathological assessment of right lung lobe showed no significant lesions in any of the treated groups (Fig. 1d, e). Further to confirm that intranasally delivered rNDV-S replicated in immunised mice, we examined the expression of matrix (M) gene of NDV in various tissues at 19 DPV using qRT-PCR. Results indicate a high number of NDV copies, and a comparable replication of rNDV-S and rNDV-WT was observed in the respiratory tract (nasal turbinate, trachea, and lungs) (Fig. 1f-h).
However, comparatively reduced replication was observed in the gut of mice vaccinated with rNDV-S or rNDV-WT (Fig. 1i). Collectively, these data indicate that rNDV-S intranasal vaccination showed no adverse pathology in the tissues examined and the rNDV-S vaccine replicated significantly in the target tissues comparable to the rNDV-WT.

Intranasal immunization with rNDV-S elicits potent S-protein binding Abs and NAbs in mice
To evaluate the immunogenicity of rNDV-S, sera samples collected at 19 days postvaccination were used to evaluate the total IgG antibody titers using ELISAs and levels of neutralizing antibodies using microneutralization and pseudo-particle entry inhibition assays, respectively. To detect anti-spike serum Abs, full-length spike (S) protein or recombinant Spike protein receptor binding domain (RBD) expressed using the Drosophila S2 cell system were used to coat ELISA plates. After prime immunization only, the sera from mice immunized with rNDV-S contained S protein-specific Abs, which were significantly increased with the boosting dose. In comparison, no S protein-specific Abs were detected in sera from the mice vaccinated with rNDV-WT or mock-vaccinated (Fig.   2a). Next, we assessed the levels of total IgG responses against purified RBD. The rNDV-S only induced significantly high levels of anti-RBD-specific IgG in the prime + boost vaccinated group when compared to mock-immunised or rNDV-WT-immunised mice (Fig.   2b). We next functionally characterized serum Ab responses by the lentiviral entry inhibition and focal-reduction neutralization tests 34,35 . As expected, serum from mock-or rNDV-WT vaccinated mice did not inhibit the entry of pseudoviral particles (Fig. 2c). In contrast, serum from rNDV-S prime or rNDV-S prime + boost groups significantly inhibited the entry of the pseudoviral particles. Correspondingly, compared to mice mockimmunised or immunised with rNDV-WT, a significant level of neutralization of SARS-CoV-2 infection was observed by sera collected from rNDV-S immunised mice, both prime and prime + boost, the latter more efficient in generating NAbs (Fig. 2d). Overall, rNDV expressing SARS-CoV-2 S protein elicited high titters of protein binding and virus NAbs in mice, especially when a prime + boost regime was utilized.

Intranasal vaccination with rNDV-S does not induce overt myeloid inflammatory responses in the lungs, pleural cavity or systemically
We next analysed systemic and local immunological responses to determine whether innate and adaptive responses were beyond the basal response induced by rNDV-WT.
We did not observe significant alterations in cellularity of the spleen, used as a surrogate marker of systemic inflammation ( Supplementary Fig. 2a). We also did not observe significant changes in the rNDV-S prime + boost group percentages of myeloid subsets examined, including neutrophils, monocytes, inflammatory monocytes, or eosinophils ( Supplementary Fig. 2b). However, we observed a significant increase in the percentage of dendritic cells (DCs) in the spleen of rNDV-WT immunised animals as compared to the mock vehicle control, and although no significant increase was seen in the rNDV-S prime immunised group, a similar but non-significant trend in DC levels was seen in murine spleen after prime + boost rNDV-S immunisation ( Supplementary Fig. 2b). Despite the observed slight morphological changes in the cellular architecture in the lung (Fig. 1d, e), we did observe a significant increase in the cellularity of the lung in the rNDV-S (prime + boost) immunized group, as compared to all other experimental groups studied ( Supplementary Fig. 2c). However, when we further examined the myeloid immune subsets, we did not observe significant differences in neutrophils, Ly6C-monocyte/macrophages, alveolar macrophages, and eosinophils as compared to the mock vehicle control group. A significant decrease in the Ly6c+ monocyte/macrophage subset was observed in the rNDV-S (prime + boost) group ( Supplementary Fig. 2d).
Interestingly and similar to the results in the spleen, we only observed a significant increase in the total DC population in the lung of mice receiving the rNDV-WT vaccine as compared to mock-vaccinated group ( Supplementary Fig. 2d).
To investigate further whether a local inflammatory response occurred following intranasal vaccination, we assessed the immune response within the pleural cavity. Our results indicate that there were no statistically significant increases in total CD45 + haematopoietic cells, neutrophils, eosinophils, monocytes, and F4/80 lo MHCII hi myeloid cells within either pleural lavage fluid ( Supplementary Fig. 3a-f) or pericardial adipose tissue ( Supplementary Fig. 3g-k). However, although not significant, a trend for higher amounts of F4/80 hi MHCII lo pleural cells was observed in rNDV-S prime + boost group ( Supplementary Fig. 3f). These data suggest that there is no lasting abhorrent myeloid inflammatory response generated at the site of vaccination or systemically following intranasal delivery of rNDV-S as compared to the mock-vaccinated mice.

Intranasal vaccination with rNDV-S drives T-cell IFNγ responses both in the lung and systemically
We next examined T-cell subsets, asking what antigen specific cytokine responses were produced against the purified full length S protein both systemically and locally at the vaccination site. Systemically, within the spleen, there was no alteration in the percentage of splenic CD4+ or CD8+ T-cell subsets in any of the experimental groups studied ( Supplementary Fig. 4a). We also did not see a significant increase in CD8+ T-cell TNF + or CD8+ T-cell IFNγ + cell populations in response to the full-length S protein Ag in any of the groups studied ( Supplementary Fig. 4b). Conversely, splenic CD4+ T-cell IFNγ + (Fig.   3a) and NK T-cell TNF + (Fig. 3b) were significantly increased in the rNDV-S boosted group indicating a systemic SARS-CoV-2 S specific response following vaccination with rNDV-S.
Locally, within the lung homogenate, a significant decrease in CD4+ T cells and an increase in CD8+ T cell percentage was observed following vaccination with all NDV containing groups ( Supplementary Fig. 4c). Looking at cytokine responses we saw a significant reduction in CD8+ T-cell TNFα production in both the NVD-WT and NVD-S groups and a significant increase in IFNγ production in the rNDV-S boosted group as compared to the mock vehicle group (Fig. 3c, e). However, despite an overall decrease of CD4 T cell percentage and no alteration in CD4+ T cell subpopulations producing IL-17 or TNF, a significant increase in CD4+ T cell IFNγ + (6.4-fold) was observed also in the rNDV-S boosted group as compared to the mock-vaccinated group (Fig. 3d, e).
Collectively, these data indicate both a systemic and local IFNγ antigen specific T-cell response following intranasal vaccination with rNDV-S.

SARS-COV-2 S protein
Fat associated lymphoid clusters (FALCs) are sites of local Ab production within the pericardium and mediastinum, known for their early response to intranasal challange 36 .
Using whole-mount immunofluorescence staining, we were able to detect an expansion of FALCs within the mediastinum at 19 DPV in the rNDV-S boosted group (Fig. 4a, b).
Within FALCs of mice receiving rNDV-S boost, and supporting an earlier response, there was a significant increase of the level of Ki67 + proliferation, representative of cellular division, (Fig. 4c, d). Although not reaching significance, a trend for increased percentage area of CD4 was seen, suggesting a local accumulation of CD4 T cells within the mediastinum (Fig. 4e). No increase in the % area of IgM (B cell marker) was seen within the pericardium at day 19 DPV (Fig. 4e).
IgM is the first polyclonal Ab induced during an immune response, and following induction of adaptive immunity B cells, IgM class switches to other Ab subclasses defined by the cytokine milieu. As FALCs are the site from which Abs detected in pleural fluid are produced we next confirmed the local activation of B cells by assessing the presence of Abs against SARS-CoV2 S RBD within the pleural fluid. Within the pleural fluid no antigen specific IgM was detected (Fig. 4g), however a significant increase in both RBD-specific IgA (Fig. 4h) and IgG2a ( Fig. 4i) but not IgG1 (Fig. 4j) levels were present within the pleural fluid of mice in the rNDV-S boosted group compared to control mice and those receiving rNDV-WT. These data suggest that B cells producing the Abs had class switched and is consistent with the antigen specific T cell derived IFNγ + detected upon re-stimulation, driving a switch to IgA and/or IgG2a (Fig. 4h, i). We next determined the number of B cells present within the pleural lavage fluid and digested pericardium and did not observe significant differences in the numbers of B1a, B1b, or B2 cells present within the pleural lavage fluid or pericardium when comparing rNDV-S boosted to mock control mice ( Supplementary Fig. 5). Furthermore, we also did not detect a significant increase in pericardial B cell proliferation ( Supplementary Fig. 6a, b), suggesting that Abs detected within the pleural fluid may have been secreted earlier or resulted from an ongoing systemic B cell response within primary lymphoid organs. To further support this, systemic anti-full-length spike and anti-RBD Ab responses were observed which neutralized SARS-CoV-2 (Fig. 2).

Safe intranasal administration of rNDV-S in hamsters
After establishing the basis of immunological and safety profiles in mice, we next attempted to evaluate the safety of aerosol rNDV-S vaccination in golden Syrian hamsters, which have the advantage of being susceptible to SARS-CoV-2 infection. A total of n=8 hamsters in each group were mock (PBS)-vaccinated or vaccinated with 1 x 10 6 PFU of rNDV-WT or rNDV-S once (prime, Fig. 5a) or twice in a 2-week interval (boosted, Fig. 5b).
Body weight was measured daily for 14 (prime) or 28 (boosted) DPV. Prime vaccination with rNDV-S resulted in no apparent clinical disease except slight (<5%) body weight loss at 5 DPV (Fig. 5c), and no mortality (Fig. 5d). Boosted vaccination also induced less than 5% of body weight loss by 5 DPV (Fig. 5e) with all the animals surviving rNDV-S vaccination ( Fig. 5f). Vaccination with rNDV-WT resulted in body weight loss only one day after vaccination in either the prime or boosted vaccinated hamsters (Fig. 5c, e). These results demonstrate that vaccination of golden Syrian hamsters with 1 x 10 6 PFU of rNDV-S is safe without lasting significant changes in body weight, with all the animals surviving aerosol rNDV-S vaccination.

Ability of intranasally administered rNDV-S to induce Ab and NAb responses in hamsters
Serum samples collected at 0, 14, and 28 DPV from experimental hamsters were analysed for Ab and NAb responses against SARS-CoV-2. Total levels of Abs against SARS-CoV-2 were evaluated by ELISA, while NAbs were evaluated by PRNT assay using SARS-CoV-2. Sera from prime vaccinated hamsters (14 DPV) were able to react with extracts from SARS-CoV-2 infected cell homogenates, but not from mock-infected cell homogenates (Fig. 6a). Notably, levels of total Abs against SARS-CoV-2 were higher in sera from boosted rNDV-S-vaccinated hamsters (28 DPV) (Fig. 6a, Supplementary Fig.   7, 8). In terms of NAb responses, only sera from hamsters receiving boosted vaccination (28 DPV) presented NAbs in the PRNT assay in both pre-treatment (Fig. 6b, and Table   1) and post-treatment ( Fig. 6c and Table 1) conditions. These results indicate that boosting is required to induce robust NAb responses against SARS-CoV-2 upon vaccination with rNDV-S.

Protection efficacy of rNDV-S against SARS-CoV-2 infection in hamsters
To access protection efficacy of rNDV-S against SARS-CoV-2 infection, hamsters were vaccinated (prime or boosted) with rNDVs (WT or S) and then challenged with 2 x 10 4 PFU of SARS-CoV-2. For viral titration and pathology, hamsters were sacrificed at 2 and 4 DPI (n=4/group). Mock-vaccinated hamsters either challenged with SARS-CoV-2 or mock challenged were used as internal controls. Lungs from mock-vaccinated hamsters showed mild to moderate, multifocal pneumonic lesions and congestions at 2 DPI, and higher inflammation scores characterized by moderate to severe locally extensive to diffuse bronchopneumonia and foamy exudate in the trachea at 4 DPI (Fig. 7a). Similar results were observed in hamsters vaccinated with rNDV-WT (primed or prime + boost).
Conversely, hamsters vaccinated with rNDV-S (prime + boost) showed significantly lower inflammation scores compared to those of mock or rNDV-WT vaccinated hamsters at both 2 and 4 DPI (Fig. 7a, b). Viral titers from nasal turbinate and lung followed similar trends and further support the lung inflammation scoring. We were not able to detect the presence of SARS-CoV-2 in the nasal turbinate and lungs of hamsters vaccinated with rNDV-S (boosted). Hamsters The average IHC scores showed progressive and widespread distribution of SARS-CoV-2 in mock-and rNDV-WT-vaccinated groups and marked absence of viral N protein staining at 2 and 4 DPI in the rNDV-S-vaccinated groups (Fig. 9b). These results demonstrate that vaccination with rNDV-S can protect from SARS-CoV-2 infection, resulting in lower SARS-CoV-2 titers and protection from lung damage.

DISCUSSION
Due to high transmissibility of SARS-CoV-2 and lack of considerable pre-existing immunity, the COVID-19 pandemic is posing significant health challenges particularly in the elderly and people with existing co-morbidities [37][38][39] . While both treatments and vaccines are being developed [40][41][42][43][44] , there are limited attempts to offer vaccines that can be deployed in frail and resource-limited health care system particularly in low-and middle-income countries. In a thirst to curb the pandemic, upper and middle-income countries are scooping most of the pharmaceutical capacity of vaccine production, which can contribute to reduce the transmission and impact of the pandemic. In this regard, a low-cost vaccine production system for mass immunization of people in low and middleincome countries is required. Here, we describe a promising and scalable live-attenuated vector vaccine based on the use of a rNDV expressing SARS-CoV-2 S, which can be produced robustly and economically in avian embryonated eggs, as well as in US FDAapproved cell lines, that can be administered intranasally to induce protection at the site of SARS-CoV-2 infection.
Our studies establish that intranasal vaccination with rNDV-S induces Abs, including NAbs, against SARS-CoV-2 full S protein and RBD, as well as antigen specific T-cell responses. Although a single dose confers protection against SARS-CoV-2 challenge, a prime and boost regime provided superior protection, reduced pathology in lung and completely blocked shedding of SARS-CoV-2 in hamsters, which is considered an appropriate animal model for vaccine evaluation 45,46 . The intranasal route for immunization is advantageous in comparison to traditional routes (i.e., intramuscular) mainly due to its ability to elicit immunity at the local mucosal level [47][48][49] . Enhancing mucosal protective abilities is desirable as it is the first line of defence in the human body against the majority of pathogens targeting the respiratory tract. Beside the possibility to self-administer and lack of needles, the intranasal delivery of antigen is associated with mucosal immunity including production of IgA and priming of T and B cells in the nasopharynx-associated lymphoid tissues 50 , and may also provide cross-protection by eliciting mucosal immunity at different mucosal sites such as the intestines and genital tract, which have been shown to be potential replication sites for SARS-CoV-2 51 62 . Briefly, the virus stock obtained from BEI Resources considered as passage four (P4) was amplified two more times to generate a P6 working stocks for animal infections. To that end, Vero E6 cells were infected at low multiplicity of infection (MOI, 0.01) for 72 h and tissue culture supernatants (TCS) were collected, clarified, aliquoted and stored at -80°C until use. Virus stocks were titrated in Vero E6 cells by plaque assay and immunostaining as previously described 35,63 . The rNDV-WT and rNDV-S viruses were propagated in chicken embryonated eggs quantified in Vero cells using procedures we described before 33 .

293T cells and Vero cells were maintained in Dulbecco's Modified
Eagle Medium (DMEM) with 10% FBS as described previously 33 .

Vaccine safety studies in hamsters
To evaluate in vivo safety of rNDV-S, hamsters (n=8 per group) were vaccinated intranasally with 1 x 10 6 PFU of rNDV-WT or rNDV-S, or mock-vaccinated with saline (PBS), in a final volume of 100 µl, after sedation in an isoflurane chamber. Hamsters were vaccinated following a prime or a booster regimen. In the case of the booster immunization, animals were boosted 2 weeks after prime vaccination. After vaccination, hamsters were monitored daily for morbidity (body weight changes and clinical signs of infection) and mortality (survival) for 14 (prime) or 28 (booster) DPV. Sera were collected at 0 and 14 (mock, prime, and booster groups), and 28 (booster) DPV. One mockvaccinated hamster in the vaccination group was removed because of accidental death.

Vaccine efficacy in hamster
After 14 (prime) or 28 (mock and booster groups) DPV, hamsters were challenged intranasally with 2 x 10 4 PFU of SARS-CoV-2 in a final volume of 100 µl under isoflurane sedation. To evaluate SARS-CoV-2 titers in nasal turbinates and lung, hamsters were humanely sacrificed at 2 (n=4) or 4 (n=4) DPI. Nasal turbinates and lungs were harvested, and half of the organs were homogenized in 2 mL of PBS using a Precellys tissue homogenizer (Bertin Instruments) for viral titration and the other half was kept in 10% neutral buffered formalin (NBF, ThermoFisher Scientific) for histopathology and immunohistochemistry (IHC). Tissue homogenates were centrifuged at 21,500 x g for 5 min and supernatants were used to calculate viral titers.

Real-time PCR for quantification of rNDV replication
RNA was extracted from mice organs (nasal turbinate, lungs, trachea, and gut) using TRIzol™ reagent as per manufacturer's instructions (Invitrogen, USA). Real-time qRT-PCR was performed using SuperScript™ III Platinum™ One-Step qRT-PCR Kit (Invitrogen, USA) following the manufacturer's instructions to detect NDV M gene 64 and enable the calculation of NDV genome copies.

Production and purification of recombinant SARS-CoV-2 S and RBD proteins
Recombinant S and RBD proteins were produced and purified as previously described 65 .
Briefly, expression cassettes containing SARS-CoV-2 S and RBD nucleotide sequences were codon optimized for Drosophila melanogaster Schneider 2 (S2) cells. The Nterminus signal sequence of both S and RBD proteins were replaced with Drosophila BiP signal sequence which encodes immunoglobulin-binding chaperone protein, and the C-

ELISA for Ab detection in hamsters
Binding affinity of hamster sera to SARS-CoV-2 antigens was determined by ELISA using The half maximal neutralizing concentration (NC50) for sera was determined using 4parameter nonlinear regression curve fit to raw infectivity data measured as relative light units, or as the percentage of infected cells (GraphPad Prism).

SARS-CoV-2 microneutralization assay
The day before infection, Vero-E6 cells were seeded at 1×10 4 cells/well into 96-well plates and incubated at 37°C in a 5% CO2 incubator overnight. Mice sera samples were heat inactivated at 56°C for 30 min. Sera samples were 10-fold serially diluted in DMEM (Invitrogen, USA) supplemented with 1% BSA and 1x penicillin/streptomycin. Next, the diluted sera samples were mixed with a constant amount of SARS-CoV-2 (100 Median Tissue Culture Infectious Dose (TCID50)/50 μl) and incubated for 1 h at 37°C. The Abvirus-mix was then directly transferred to Vero E6 cells and incubated for 3 days at 37°C in a 5% CO2 incubator. The median microneutralization (MN) 50 (MN50) or 100 (MN100) of each serum sample was calculated as the highest serum dilution that completely protect the cells from cytopathic effect (CPE) in half or all wells, respectively.

Plaque reduction microneutralization (PRMN) assays
PRMN assays were performed to identify the levels of SARS-CoV-2 NAbs as described previously 30  protein (1 µg/ml). Viral neutralization was evaluated and quantified using ELISPOT, and a sigmoidal dose-response, non-linear regression curve was generated using GraphPad Prism to calculate median neutralization titer (NT50) in each of the serum samples.

SARS-CoV-2 virus titrations
Nasal turbinate and lungs from SARS-CoV-2-infected golden Syrian hamsters were homogenized in 2 ml of PBS for 20 s at 7,000 rpm using a Precellys tissue homogenizer (Bertin Instruments). Confluent monolayers of Vero E6 cells (96-plate format, 4 x 10 4 cells/well, duplicates) were infected with 10-fold serial dilutions of supernatants obtained from the nasal turbinate or lung homogenates from SARS-CoV-2 infected golden Syrian hamsters. Virus from serially diluted samples was adsorbed at 37°C for 1 h followed by incubation in post-infection media containing 2% FBS and 1% Avicel at 37°C for 24 h.
After viral infection, plates were submerged in 10% NBF for 24 h for fixation/inactivation. For immunostaining, cells were washed 3X with PBS and permeabilised with 0.5% Triton X-100 for 10 min at room temperature, followed by blocking with 2.5% bovine serum albumin (BSA) in PBS for 1 h at 37°C. Cells were incubated with the N protein 1C7 monoclonal Ab (1 µg/ml) diluted in 1% BSA for 1 h at 37°C. Then cells were washed 3X with PBS and stained with the Vectastain ABC kit and developed using the DAB Peroxidase Substrate kit (Vector Laboratory, Inc, CA, USA), as previously described 1 .
Virus titers were calculated as PFU/ml.

Cell re-stimulation and flow cytometry staining
Spleens were removed from mice and disaggregated through a 100 µm sieve.
Pericardium was digested using 1mg/ml Collagenase D (Roche) in a shaking heat block at 37°C, digestion was stopped after 35mins by addition of 5mM EDTA (Fisher) followed by passing through a 100µm cell strainer. Pleural exudate cells (PLEC) were isolated via lavage of the pleural cavity with 10ml of ice cold dPBS (Sigma). Left lung lobes were excised and finely diced using scissors prior to digestion in 2 ml lung digest medium (PBS (Sigma), 0.1mg/ml Liberase TM (Roche), 50 µg/ml DNAse I (Roche)) for 30 min in a shaking incubator at 37°C. Digestion was stopped after 30mins by adding 5mM EDTA

Pathologic observation
Mouse lung tissues were inflated and fixed in 10% neutral buffered formalin (NBF) solution and embedded in paraffin prior to H&E. After mounting, positive cells were enumerated in field of view and alveolar space enumerated via NIH Image J software.
In the hamster studies, lungs were collected at 2-and 4-DPI from mock-and SARS-CoV-2-challenged golden Syrian hamsters. Lung samples were photographed to show gross lesions on both the dorsal and ventral views. Images were used for macroscopic pathology scoring analysis by measuring the distributions of pathological lesions, including consolidation, congestion, and pneumonic lesions using NIH ImageJ software.
The area of pathological lesions was converted into percent of the total lung surface area.
Half of the hamster lungs fixed with 10% neutral buffer formalin (NBF) were embedded in paraffin blocks, and sectioned (5 µm

Immunohistochemistry (IHC) assays
Immunostainings were performed as previously described 54,55 . Briefly, 5 um lung tissues sections were mounted on Superfrost Plus Microscope slides, deparaffinized and antigen retrieval was conducted using the HIER method (Heat Induced Epitope Retrieval). The Mann-Whitney test was used for statistical analysis.

Confocal imaging
Mouse mediastinum samples were fixed for one hour on ice in 10% NBF (Sigma) and then permeabilised in PBS 1% Triton-X 100 (Sigma) for 15 min at room temperature prior to staining with primary Abs for one hour at room temperature in PBS 0.5% BSA 0.5% Triton. DAPI was added for the final 10 min of the 1 h incubation. After washing in PBS, tissues were mounted in fluoromount G and confocal images were acquired using a Zeiss LSM880 confocal laser scanning microscope. Samples were analysed using Fiji software.
Cluster area was delineated using IgM, and percentage area of Ki67 & CD4 staining within each cluster was calculated.

Statistical analysis
Results are expressed as mean ± SEM. Where statistics are quoted, two experimental groups were compared via the Student's t test or the Mann-Whitney for non-parametric data. Three or more groups were compared with ANOVA, with Sidaks, Dunnett's or Bonferroni's post-test as indicated. A P value of <0.05 was considered statistically significant. *, P<0.05; **, P<0.01; or ***, P<0.005 for indicated comparisons, error bars represent standard error of means.

ACKNOWLEDGEMENTS
We want to thank Dr. Thomas Moran at the Icahn School of Medicine at Mount Sinai for providing us with the SARS-CoV cross-reactive N protein monoclonal Ab 1C7. We also thank BEI Resources for providing the SARS-CoV-2 USA-WA1/2020 isolate (NR-52281).
We would also like to thank members at our institutes for their efforts in keeping them fully operational during the COVID-19 pandemic, and the BSC and IACUC committees at Percentage weight change of mice following instillation of PBS or indicated rNDV constructs over the experimental time-course. (c) Daily chow intake by mock-vaccinated or rNDV-S vaccinated mice. Mean interstitial space size of right lung following H&E staining and assessed via ImageJ (d) and representative images (e). Viral replication of intranasally administered rNDV in nasal turbinate (f), trachea (g), lung (h) and gut (i). Data (n = 4-5 mice/group); *, P<0.05 and NS, non-significant between naïve and vaccinated groups, error bars represent SE of means via repeated t-test or ANOVA with Dunnett's post-test.

Fig. 2. Intranasal administration of rNDV-S elicits high titers of binding and NAbs in mice.
(a) S-specific serum IgG titers measured by ELISA. Sera from animals at 19 days after-prime and after prime + boost were isolated and used in the ELISA against a recombinant trimeric S protein. Sera from mock-vaccinated mice and mice vaccinated with rNDV-WT were used as control. (b) Corresponding sera from all group of mice were used to detect Abs against RBD of spike protein. (c) Lentiviruses expressing the S protein of SARS-CoV-2 were used to perform pseudoparticle entry inhibition assay. Sera samples collected from mice after prime or after prime + boost inhibited the entry of lentiviral pseudoparticles compared to sera collected from mice vaccinated with rNDV-WT or mock-vaccinated (d) Sera samples from all groups were used to measure neutralization of SARS-CoV-2. Sera from prime from group of mice which were vaccinated with rNDV-S prime only or after prime + boost group of mice showed marked virus neutralization compared to sera from mock-treated or vaccinated with rNDV-WT. Data (n = 4-5 mice/group); *, P<0.05; **, P<0.01; or ***, P<0.005 between naïve and vaccinated groups, error bars represent SE of means via repeated T-test or ANOVA with Dunnett's.      Lung lesions are primarily characterized by suppurative inflammation within and surrounding the bronchi/bronchiole (black asterisks) at 2 DPI and extending to bronchointerstitial inflammation (arrows) by 4 DPI. Scale bars = 500 µm. (c) Average inflammation scoring: The average inflammation scores were determined based on percent of inflamed lung area: grade 0 = no histopathological lesions; grade 1 = minimal (< 10%) histopathological lesions; grade 2 = mild (10 to 25%) histopathological lesions; grade 3 = moderate (25 to 50%) histopathological lesions; grade 4 = marked (50 to 75%) histopathological lesions; and grade 5 = severe (> 75%) histopathological lesions. (d) Neutrophil infiltration: The neutrophil infiltration scores were graded base on lesion severity as follows: grade 0 = no lesions observed; grade 1 = <10 cells; grade 2 = <25 cells; grade 3 = <50 cells; grade 4 = >100 cells; and grade 5 = too many cells to count. The Mann-Whitney test was used for statistical analysis. *, P<0.05; **, P<0.01; or ***, P<0.005 for indicated comparisons. Lines represent the geometric mean. @ represents one mock-vaccinated hamster that was removed because of accidental death. Scores were determined on the presence of SARS-CoV-2 N protein staining: grade 0 = no or rare immunostaining; grade 1 = viral N protein staining only in bronchi/bronchiole; grade 2 = viral N protein staining in bronchi/bronchiole and surrounding alveolar septa; grade 3 = viral N protein staining in alveolar septa distant from the bronchi and bronchioles; grade 4 = viral N protein staining throughout the lung. @ represents one mock-vaccinated hamster that was removed because of accidental death. The Mann-Whitney test was used for statistical analysis. *, P<0.05; **, P<0.01; or ***, P<0.005 for indicated comparisons.