Mucosal prime-boost immunization with live murine pneumonia virus-vectored SARS-CoV-2 vaccine is protective in macaques

Immunization via the respiratory route is predicted to increase the effectiveness of a SARS-CoV-2 vaccine. We evaluated the immunogenicity and protective efficacy of one or two doses of a live-attenuated murine pneumonia virus vector expressing SARS-CoV-2 prefusion-stabilized spike protein (MPV/S-2P), delivered intranasally/intratracheally to rhesus macaques. A single dose of MPV/S-2P was highly immunogenic, and a second dose increased the magnitude and breadth of the mucosal and systemic anti-S antibody responses and increased levels of dimeric anti-S IgA in the airways. MPV/S-2P also induced S-specific CD4+ and CD8+ T-cells in the airways that differentiated into large populations of tissue-resident memory cells within a month after the boost. One dose induced substantial protection against SARS-CoV-2 challenge, and two doses of MPV/S-2P were fully protective against SARS-CoV-2 challenge virus replication in the airways. A prime/boost immunization with a mucosally-administered live-attenuated MPV vector could thus be highly effective in preventing SARS-CoV-2 infection and replication.


INTRODUCTION
Since emerging in late 2019, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has caused over 700 million cumulative cases and more than 6 million deaths worldwide 1 .Rapid deployment of SARS-CoV-2 vaccines in late 2020 helped reduce the burden of disease associated with the pandemic.
Currently, three SARS-CoV-2 vaccines are available in the United States: two mRNA-based vaccines and a protein subunit vaccine.A fourth vaccine based on a replication-incompetent adenovirus vector was available in the US until May 7, 2023.All of these are injectable vaccines, based on the spike (S) surface glycoprotein.These injectable vaccines stimulate systemic immunity and their use signi cantly reduced severe cases of coronavirus infectious disease (COVID).However, they do not directly induce mucosal immunity in the respiratory tract, limiting their ability to prevent SARS-CoV-2 infection and replication at the primary site of infection 2 .SARS-CoV-2 continues to circulate in communities worldwide, allowing for continuous evolution of new SARS-CoV-2 variants 3 .Mucosal immunization with a live, replicating viral vector would be expected to elicit robust local mucosal immune memory with the ability to effectively restrict SARS-CoV-2 replication at airway surfaces, reducing transmission and the emergence of new variants.
We previously evaluated murine pneumonia virus (MPV) as a live-attenuated vaccine vector for mucosal immunization 4 .MPV is a pneumovirus that is the murine homolog of human respiratory syncytial virus (RSV).Like RSV, MPV is an enveloped virus with a single-stranded negative-sense RNA genome of approximately 15,000 nucleotides that replicates in the cytoplasm.MPV is a promising live vector for intranasal immunization against respiratory viruses because (i) it is highly attenuated in nonhuman primates due to host range restriction, (ii) it has a tropism for the respiratory tract, and (iii) pre-existing immunity against the MPV vector in humans is low 5 .In a previous study, recombinant MPV expressing the RSV fusion (F) protein, the major RSV neutralization antigen, was highly attenuated for replication in rhesus macaques, but induced high RSV-neutralizing serum antibody titers.Thus, even though its replication in nonhuman primates is low, the MPV vector was strongly immunogenic 4,5 .
In the present report, we performed a prime-boost immunization with an MPV vector expressing the prefusion-stabilized version of the SARS-CoV-2 S protein (MPV/S-2P) in rhesus macaques, with a detailed analysis of the local and systemic immune responses and evaluation of protective e cacy.

RESULTS
A prime/boost regimen of MPV/S-2P increased the magnitude and breadth of the mucosal anti-S antibody response in the airways, and induced high levels of dimeric anti-S IgA We previously generated an MPV vaccine vector based on a recombinant version of MPV strain 15 with the L ORF encoding the MPV polymerase partially codon-pair optimized for e cient expression in humans 4,6 .Here, we evaluated a version of this MPV vector expressing the prefusion-stabilized SARS-CoV-2 S protein (S-2P; derived from strain Wuhan-1) in rhesus macaques.To study the safety, immunogenicity, and protective e cacy of one or two doses of this vaccine, delivered by the respiratory mucosal route, we included three groups of macaques (Fig. 1A).Animals received a single intranasal/intratracheal (IN/IT) dose of MPV (empty-vector control, group 1) or MPV/S-2P (groups 2 and 3) on day 0; group 3 received a boost on day 28.After the rst dose, shedding of the MPV empty-vector control from the upper airways (UA) was low (< 2 log 10 PFU/ml, 3/4 animals) and sporadic.Shedding of MPV/S-2P (groups 2 and 3) was detected in the UA of 2/8 animals only, with titers at the limit of detection (Fig. 1B, left panel).In the lower airways (LA), MPV control was detectable in all four animals over 8 days, with peak titers on day 4 post-immunization (pi) (median 4.3 log 10 PFU/ml; group 1, n = 4).
MPV/S-2P was detected in the LA of 3/4 and 2/4 macaques (groups 2 and 3), albeit only through day 6, and titers were signi cantly lower compared to those of the MPV empty-vector control (Fig. 1B, right panel).After the second dose of MPV/S-2P vaccine (group 3; boost; n = 4), shedding was detected only in a single animal in the UA on day 3 at the level of detection of the assay (Fig. 1B).No changes in the macaques' vital signs were observed following IN/IT immunization with MPV or MPV/S-2P (Fig. S1).
A single IN/IT dose of MPV/S-2P e ciently induced mucosal anti-S and anti-receptor binding domain (RBD) IgG and IgA in the UA, detectable in nasal wash (NW) by ELISA (Fig. 2A and B; MPV/S-2P prime; groups 2 and 3 combined, n = 8).Surprisingly, despite vaccine replication being undetectable in the UA, the second dose of MPV/S-2P further boosted the mucosal antibody response with a 2.1-to 7.9-fold increase of the anti-S and RBD mucosal IgG and IgA geometric mean titers (GMTs) in the UA from day 25/28 to day 42 (group 3, n = 4, two weeks post-boost).We also evaluated dimeric IgA because it is highly functional for virus neutralization in the respiratory mucosal environment 7 .While one dose induced dimeric anti-S IgA in two of 8 MPV/S-2P primed animals, the second dose induced dimeric IgA in 3 of 4 animals two weeks post-boost (Fig. 2C).In the LA, we detected strong anti-S and anti-RBD IgG and IgA responses after one dose of MPV/S-2P (n = 8) that peaked two weeks after immunization, with an impressive 50-to 119-fold increases in GMTs (n = 4) from day 25/28 to day 37, i.e., 9 days after the second dose (n = 4) (Fig. 2D and E).Even though titers gradually decreased, by four weeks after the boost, anti-S IgA and IgG GMTs in the UA and the LA were still above the peak GMTs induced by one dose.In the LA, MPV/S-2P also induced dimeric anti-S IgA, detectable in 6 of 8 animals on day 14 after priming, and in 4 of 4 animals on day 9 post-boost (Fig. 2F).
A prime/boost regimen of MPV/S-2P increased the magnitude, avidity, and breadth of the anti-S serum antibody response After a single dose of MPV/S-2P, we detected high titers of serum anti-S and anti-RBD IgA and IgG that remained high through week four after priming (Fig. 3A and B, n = 8).The MPV/S-2P boost further increased the peak serum anti-S IgG and IgA GMTs (5.6-and 8.6-fold increase, respectively; n = 4), with no signi cant decrease through week four post-boost.We measured the strength of antibody binding in an ELISA-based avidity assay.The geometric mean of the serum anti-S avidity index increased from 0.87 and 0.94 at four weeks after prime to 0.98 and 0.99 at two weeks after the boost for IgA and IgG, respectively (Fig. 3A and B), suggesting that further antibody maturation occurred after the boost.Furthermore, the serum anti-S antibodies also conferred antibody-dependent cellular phagocytosis (ADCP) activity after priming, which remained high after the boost (Fig. 3C).
The breadth of the serum-neutralizing antibody response was evaluated in live SARS-CoV-2 neutralization assays against the vaccine-matched strain WA1/2020 or representatives of Alpha or Beta lineages.Neutralizing antibody titers to WA1/2020 were variable and relatively low after the prime [detectable in 6/8 macaques, median 50% SARS-CoV-2 neutralizing antibody titers (ND 50 ) of 1.1 log 10 ], with a strong increase (about 9-fold) in 4/4 macaques after the boost (median ND 50 of 2.0 log 10 ) (Fig. 3D), suggesting that animals were effectively primed for anamnestic B-cell responses despite low or undetectable neutralizing activity after the rst dose.The MPV/S-2P boost induced a 13-and 7-fold increase in peak serum neutralizing titers against Alpha and Beta isolates, respectively (Fig. 3D).As expected, no serum anti-S or anti-RBD IgA or IgG or S-speci c ADCP activity was induced by the empty vector control (Fig. 3A-3C), and all macaques developed high serum neutralizing antibody titers against the MPV vector that remained high in the MPV/S-2P-boosted macaques (Fig. 3D).

MPV/S-2P induced S-and RBD-speci c peripheral blood Bcell responses
To extend our analysis of the systemic S-speci c immune response to MPV/S-2P, we characterized the kinetics and phenotypes of S-speci c peripheral blood B cells post prime and boost (Fig. 4).Peripheral blood mononuclear cells (PBMCs) were stained with uorochrome-labeled RBD and S-2P protein probes to identify S-speci c B cells that recognize epitopes of the RBD (RBD + /S-2P + ), or outside the RBD (RBD − /S-2P + ).The cells were further stained using a cocktail of uorochrome-labeled monoclonal antibodies and analyzed by ow cytometry (Fig. 4A).Analysis was done on live, single, non-naïve (IgD − ) CD95 +/− B cells from PBMC collected before immunization and on six time points after prime (groups 2 and 3 combined; n = 8) or boost (group 3, n = 4; Fig. 1A, see Fig. S3A for the gating strategy).
After MPV/S-2P prime, the frequency of S-speci c (RBD + and RBD − combined) B cells in the blood peaked at day 14 (median 1.1%), and slightly declined by day 28 (median 0.8%).By day 9 after boost, a rapid recall response was detected, with median frequencies of 1.3%, followed by a steady decrease through day 28 post-boost (n = 4, median 0.5% on day 28 post-boost, Fig. 4B).At each time point, RBD + -B cells represented approximately one third of S-speci c B cells, with the remaining two thirds (RBD − /S-2P + ) exhibiting speci city to undetermined epitopes of S-2P (Fig. 4C).
The majority of S-speci c RBD + and RBD − B cells had an IgG isotype (80% median frequency) regardless of the time post-immunization (Fig. 4D and E).The frequency of S-speci c IgM B cells was highest at day 14 post-prime (median 12.5% of all S-speci c B cells; n = 8 macaques) and decreased thereafter, suggestive of isotype switching.S-speci c IgA B cells also were detected at each time point.Interestingly, after MPV/S-2P boost, S-speci c RBD + IgA B cells (Fig. 4E) were about 4-fold more abundant than RBD − IgA B cells, representing about 16% and 4% of the S-speci c B cells, respectively (n = 4 animals).About 60% of the S-speci c IgG B cells exhibited an activated memory (AM) phenotype (CD21 − /CD27 + ), with similar proportions for RBD + and RBD − cells (Fig. 4F and G, left).The phenotypes of the S-speci c IgA and IgM B cells were more variable, with a lower proportion of activated memory B cells, a greater proportions of cells with tissue-resident like memory phenotype (CD21 − /CD27 − ), and, especially after the boost, resting memory phenotype (CD21 + /CD27 + or CD21 + /CD27 − ) (Fig. 4G, center and right).Thus, mucosal immunization with MPV/S-2P induced S-speci c B cells in the blood that target the RBD and epitopes on other regions of S, and a recall response after boost of mostly activated memory IgG B cells occurred rapidly.

Priming with MPV/S-2P by the IN/IT route induces S-speci c CD4 + and CD8 + T-cells in blood and airways that were reactivated by boosting
We also characterized S-speci c CD4 + and CD8 + T-cells in the blood and airways following IN/IT immunization with MPV/S-2P (Fig. 5) by stimulation with pools of overlapping peptides covering the entire SARS-CoV-2 S (PBMC and BAL) or N proteins (control; BAL only).Stimulated cells and unstimulated controls were analyzed by ow cytometry, with gating on live, single, non-naïve (CD95 + ), non-regulatory (FoxP3 − ), CD4 + or CD8 + T-cells (Fig. S3B).Results from representative macaques are shown in Fig. 5A, D, G, and I, while data from all macaques are shown in Fig. 5B, C, E, F, H and J (post-MPV prime: n = 4, post-MPV/S-2P prime: n = 8; post-MPV/S-2P boost: n = 4).S-speci c IFNγ + /TNFα + CD4 + T-cells in the blood peaked on day 9 after prime (median 0.2%) and declined by day 25/28 (median 0.1%).CD4 + T-cells were restimulated 9 days after boost (median 0.2%, Fig. 5A-B) and maintained (median 0.1%) until challenge.The frequency of S-speci c IFNγ + /TNFα + CD8 + T-cells in the blood was variable over the course of the experiment, but all macaques had measurable S-speci c IFNγ + /TNFα + CD8 + T-cells at each day pi (median 0.12-0.21%,Fig. 5A and C).
In the airways, MPV/S-2P induced a greater frequency of S-speci c IFNγ + /TNFα + CD4 + and CD8 + T-cells compared to blood (Fig. 5D-5F).Airway T-cell responses peaked on day 14 post-prime (medians 8.6% and 7.5% for CD4 + and CD8 + T-cells, respectively) and decreased by day 25/28 (medians 6.0% and 5.2% for CD4 + and CD8 + T-cells, respectively).The MPV/S-2P boost induced a recall response of the IFNγ + /TNFα + S-speci c T-cells that did not signi cantly decline between days 9-28 post-boost (medians of 10.5% and 6.0% for CD4 + and CD8 + T-cells on day 14 post-boost, respectively).As expected, CD4 + and CD8 + T-cells from BAL did not respond to the pool of N peptides above background (Fig. 5D), and the empty MPV vector did not induce any S-speci c T-cells in the blood and airways (Fig. 5A-5F).S-speci c IFNγ + /TNFα + CD4 + and CD8 + T-cells in both the blood and BAL of MPV/S-2P-immunized macaques expressed high levels of the proliferation marker Ki-67 on day 9 post-prime that steadily decreased until day 28 post prime (Fig. 5G-5J).As expected, Ki-67 expression in S-speci c T-cells substantially increased after the boost.Ki-67 expression returned to baseline before challenge (Fig. 5H  and J).

MPV/S-2P immunization induced S-speci c T-cells that transitioned to a tissue-resident memory phenotype in airways
We characterized in greater detail the phenotype of the S-speci c T-cells in the blood (Fig. S4) and BAL airway samples (Fig. 6).In addition to expressing IFNγ and TNFα, S-speci c CD4 + T-cells in the airways expressed IL-2 (Fig. 6A and C), consistent with a Th1-biased phenotype.Subsets of these Th1-biased Sspeci c CD4 + T-cells and the large majority of the S-speci c IFNγ + /TNFα + CD8 + T-cells expressed the degranulation markers CD107ab and granzyme B, suggesting that they exhibited cytotoxic activity and were highly functional (Fig. 6A-6D).The proportion of polyfunctional S-speci c CD4 + and CD8 + T-cells remained stable after boost (Fig. 6C-6D).The functionality of S-speci c T-cells in the airways and in the blood was overall comparable (Fig. S4A-D).Interestingly, MPV/S-2P induced a small population of Sspeci c IL-17 + CD4 + T-cells in the LA that peaked on day 14 post-prime (median = 0.3%), declined on day 28, and responded to boosting (Fig. 6E and F; median = 0.4% on day 42, i.e. day 14 post-boost).
To con rm that immunization with MPV and MPV/S-2P induced a Th1-biased environment in the airways, we characterized the expression of 36 cytokines in BAL after the priming using a multiplex beadbased immuno-assay (Fig. S5).A transient and moderate increase of the Th1-related cytokines IFNγ, TNFα, and granzyme B was detected on day 9 post-prime with MPV and MPV/S-2P (Fig. S5A); Th2related cytokines such as IL-4, IL-5, and IL-13 (Fig. S5B) or the Th17-related cytokine IL-17 (Fig. S5C) were not increased, further con rming a Th1-biased response in the airways of macaques.Furthermore, a transient and moderate increase of interferons and chemokines also was measured on day 9 post-prime, but notably, the in ammatory response in the airways of MPV/S-2P-immunized macaques was signi cantly milder than in MPV-immunized macaques (Fig. S5C).

MPV/S-2P immunized macaques were protected from SARS-CoV-2 challenge replication in both upper and lower airways
To evaluate the protective e cacy of prime and prime/boost regimens with MPV/S-2P, macaques were challenged by the IN/IT route with 6.3 log 10 TCID 50 of the vaccine-matched WA1/2020 strain of SARS-CoV-2 on days 31 or 32 days after a single dose of MPV or MPV/S-2P (n = 4 per group), or on day 58 (30 days after the boost) for the animals that received two doses of MPV/S-2P (n = 4) (Fig. 1A).Nasal swabs and BAL were obtained on indicated days post-challenge (pc) to evaluate SARS-CoV-2 replication in the UA and LA.On day 6 pc, animals were euthanized, and lung tissues were obtained.
Challenge virus was quanti ed by RT-qPCR using SARS-CoV-2 speci c TaqMan assays for subgenomic E (sgE) or genomic N (gN) RNA (Fig. 7); sgE mRNA is a marker for active SARS-CoV-2 transcription/replication, and gN RNA is a marker for input and newly-replicated challenge virus 8 .NS and BAL of 3/4 and 4/4 macaques immunized with the MPV empty-vector control were sgE positive, with maximal copy numbers on day 2 pc (medians: 3.0 and 4.7 log 10 sgE copies/ml in the UA and LA, Fig. 7A-7B, left panels).sgE RNA remained detectable in the UA of 2 of 4 MPV-immunized macaques on day 6 pc.In MPV/S-2P-primed macaques, low levels of sgE RNA were detectable only in one of four animals on day 2 pc in the UA.No sgE RNA was detected in the UA of any of the macaques that had received a prime/boost immunization with MPV/S-2P, nor in the LA of macaques from the MPV/S-2P prime or prime/boost groups.SARS-CoV-2 gN RNA was detected in the UA of all macaques on day 2 pi, con rming that all macaques had received challenge virus (Fig. 7A, right panel).In the LA, high copy numbers of gN were detected in MPV-immunized macaques (median of 6.6 log 10 copies/ml on day 2 pi, Fig. 7B, right panel), presumably re ecting the presence of mature progeny virions.
After necropsy, high levels of sgE were detected in the lungs of all four MPV empty-vector controlimmunized macaques (up to 6.5 log 10 sgE copies/g), indicating a high level of active genome replication, while all lung tissue specimens from animals that had received one or two doses of MPV/S-2P were negative for sgE mRNA, indicative of the absence of replicating challenge virus.gN RNA was detected in a single sample from the MPV/S-2P one-dose group, while high copy numbers were measured in lung samples from MPV-control immunized macaques (median of 6.4 log 10 copies/g in left middle area of the lungs, Fig. 7C, right panel).Thus, SARS-CoV-2 replication was undetectable in the UA and LA of macaques that received two doses of MPV/S-2P, indicative of a high level of protection.
Finally, we evaluated the CD4 + and CD8 + T-cell responses in the blood and airways on day 6 post challenge, focusing on S-speci c IFNγ + /TNFα + cells (Fig. 5B-5F and S6).In the blood, we detected a postchallenge increase in S-speci c IFNγ + /TNFα + CD4 + T-cells in 3/4 and 1/4 MPV/S-2P-primed andprimed/boosted macaques, respectively, with cells from all animals expressing the proliferation marker Ki-67 (Fig. S6C, top).In these animals, there was no detectable increase in S-speci c blood CD8 + T-cells nor an increase in their Ki-67 + expression (Fig. 5B, C, and S6A, C, bottom).In the airways, we did not detect an increase in S-speci c IFNγ + /TNFα + CD4 + T-cells (Fig. 5E, F and S6B) after challenge, although we detected an increase in their Ki-67 expression in a subset of animals, suggesting CD4 + T-cell reactivation upon challenge (Fig. S6D, top).There also was no increase in airway S-speci c IFNγ + /TNFα + CD8 + T-cells, and no increase in their Ki-67 expression (Fig. 6D, bottom).

DISCUSSION
A SARS-CoV-2 vaccine that can induce immunity directly in the respiratory tract -in addition to inducing systemic immunity -should have increased effectiveness because this is the major site of SARS-CoV-2 entry, replication, shedding, and illness 2,9 .Systemic immunity alone has reduced effectiveness in the respiratory tract.For example, studies in humans indicated an approximate 350-fold gradient between serum and mucosal IgG 10 .In addition, induction of local IgA and Trm B and T-cells in the respiratory tract is highly dependent on local immunization 11,12,13,14,15 .
In the present study, we used the non-human virus MPV as a live-attenuated viral vector for mucosal delivery to the respiratory tract, designed to express a prefusion-stabilized version of SARS-CoV-2 S. The replication of pneumoviruses is restricted to the respiratory tract, an important safety feature for a vaccine vector.In the present and previous studies 4,5 , MPV was strongly attenuated in nonhuman primates, presumably due to a strong host-range restriction.Host-range restriction is typically based on multiple viral genes resulting in high refraction to de-attenuation 16 .Based on their phenotype in primates, MPV vectors are expected to be attenuated but immunogenic in humans.The absence of pre-existing immunity to MPV in humans eliminates concerns of immune restriction of the live-attenuated immunizing vector.
In a previous study, we found that a single IN/IT immunization of macaques with MPV/S-2P induced high levels of serum anti-S and anti-RBD IgG that were comparable to levels measured in the plasma of COVID-19 convalescent individuals 6 .However, in a SARS-CoV-2 neutralization assay, serum neutralizing antibodies titers were variable among MPV/S-2P-immunized macaques.This prompted us to evaluate if a prime/boost regimen of MPV/S-2P would improve the S-speci c immune response in SARS-CoV-2 naïve animals.Indeed, our results show that despite the absence of detectable replication in the UA and LA in 3 of 4 animals, presumably due to inhibition by anti-vector immunity induced by the rst dose, the MPV/S-2P boost strongly increased the magnitude of the mucosal anti-S IgG and IgA antibody response and induced high levels of S-speci c dimeric IgA in the UA and LA.Dimeric IgA is highly neutralizing and non-in ammatory 7 .Nasal dimeric IgA but not IgG has been shown to correlate with nasal neutralization of SARS-CoV-2 17 , or, in controlled human infection models, with protection against RSV 18 .The MPV/S-2P prime/boost regimen also strongly increased the magnitude, breadth, and avidity of serum anti-S antibodies, suggesting further a nity maturation following the boost.Serum anti-S antibodies also displayed Fc receptor-mediated ADCP activity that increased following boost, enhancing the functionality of MPV/S-2P-induced serum antibodies and their breadth against SARS-CoV-2 VoCs 19 .
As expected, the second dose resulted in a quick recall response of mostly activated peripheral blood memory IgG B cells.Approximately one-third of the S-speci c B cells recognized epitopes of the RBD which represents approximately 20% of the S protein.Thus, the B-cell response was modestly biased toward the RBD, consistent with previous studies showing that RBD was the primary antibody target for neutralizing antibodies 20,21 .RBD-speci c IgA antibodies in particular have been associated with SARS-CoV-2 protective immunity 22 .Interestingly, following the MPV/S-2P boost, RBD + IgA B cells were more abundant than RBD − IgA B cells.
The systemic and mucosal IgA recall response after the second dose of MPV/S-2P in macaques was substantial.In a recent study following nasal and plasma antibody responses one year after COVID-19 hospitalization, individuals who later received an injectable SARS-CoV-2 vaccine mounted systemic IgG and IgA recall responses and nasal IgG responses, but the nasal IgA responses were negligible 11 .This suggested that the nasal IgA responses are compartmentalized from the systemic responses 11,23 .Based on the strong mucosal recall responses that were induced by the second dose of MPV/S-2P in our study, we would predict that an intranasal vectored vaccine will induce a mucosal IgA recall response in the respiratory tract in individuals with prior SARS-CoV-2 infection, improving protection against re-infection and replication at the mucosal entry sites of SARS-CoV-2.
MPV/S-2P induced S-speci c peripheral blood CD4 + and CD8 + T-cells with frequencies comparable to other vaccine platforms 24,25 , but their frequencies in the airways largely exceeded what has been described for mRNA vaccines 26 and were comparable to those induced by mucosal immunization with the parain uenza virus vector B/HPIV3/S-6P 27 .A large proportion of these highly functional S-speci c CD4 + T-cells expressed Th1-related cytokines.Interestingly, a small population of S-speci c CD4 + T-cells expressed IL-17 which has also been identi ed following immunization with the B/HPIV3/S-6P vector 27 and thus seems to be induced by live-attenuated viral vectors.Such Th-17 cells were previously shown to facilitate the recruitment of T and B cells in the lung, resulting in a faster IgA recall response 28 .The Sspeci c CD4 + and CD8 + T-cells in the blood and airways appeared to also be highly functional, expressing high levels of granzyme B. These S-speci c CD4 + and CD8 + T-cells were restimulated after boost while they transitioned to a Trm phenotype, providing a critical line of defense for the control of respiratory infections 12,13,14 .In convalescent COVID patients, Trm CD8 + T-cells were detected in the lungs up to 10 months after initial infection, likely contributing to long-term protection against SARS-CoV-2 15 .
Previous studies evaluated intranasal immunization with non-replicating adenovirus-based COVID-19 vaccine candidates in Phase 1 studies 29 or in a heterologous prime/boost study using the antigenicallydistinct candidates Sad23L and Ad49L in nonhuman primates 30 .These candidates were generally welltolerated, but following delivery by the intranasal route, their immunogenicity was relatively low 29,30 .
Another Phase 1 study evaluated the immunogenicity of a Newcastle disease virus vector expressing the SARS-CoV-2 S protein, delivered intramuscularly or intranasally.This vaccine induced S-speci c antibody and T-cell responses, but the immunogenicity following intranasal delivery also was relatively weak 23 .Intranasal delivery of subunit and mRNA vaccines is also being evaluated, but inactivated or nonreplicating vaccines may require adjuvants for an e cient induction of a mucosal T-cell response, adding layers of complexity to the vaccine development path 29 .
At the time of challenge in this study (4 weeks post-prime or 4 weeks post-boost), titers of S-speci c serum antibodies were high, while mucosal antibodies had declined from their peak titers at 2 weeks post-immunization, and the majority of antigen-speci c CD4 + and CD8 + T-cells had differentiated into Trm cells.This combination of systemic and mucosal immunity was strongly protective against SARS-CoV-2 challenge.Speci cally, only one animal had traces of challenge virus replication detectable (measured as sgE RNA) after a single dose of MPV/S-2P despite high levels of challenge virus (measured as gN RNA).In addition, protection increased after the second dose, as no animals had measurable challenge virus replication.This protection from SARS-CoV-2 challenge virus replication predicts that MPV/S-2P immunization may restrict SARS-CoV-2 transmission in vaccinated individuals.
In summary, a single dose of MPV/S-2P elicited strong systemic and mucosal immunity (antibodies, B cells, and T cells) in macaques.The second intranasal dose increased the magnitude, avidity, ADCP, and breadth of the mucosal and systemic anti-S IgG and IgA antibody responses, and induced restimulation and proliferation of S-speci c T-cells.One or two doses were protective against SARS-CoV-2 challenge.Future studies will evaluate the durability of immunity, and we will evaluate MPV/S-2P in heterologous prime/boost studies (e.g. with mRNA vaccines).A Phase 1 study to evaluate the safety and immunogenicity of intranasal immunization with MPV/S-2P is in preparation.

Study Design
The objectives of this study were to characterize in rhesus macaques the safety, immunogenicity, and protective e cacy of one or two mucosal doses of the live MPV-vectored SARS-CoV-2 vaccine MPV/S-2P.We evaluated vaccine shedding, mucosal and systemic antibody and T-cell responses and B cell responses following mucosal administration of one or two doses of MPV/S-2P or empty-vector control.
We also evaluated immune responses and protection following SARS-CoV-2 challenge.We used four animals per group, consistent with prior studies of mucosal vaccines in nonhuman primates.Results from animals primed with MPV/S-2P were combined for some analyses.The sample size for each group is indicated in the gure legends.Animals were randomly assigned to experimental groups.The study was performed unblinded.

Cell lines
African green monkey kidney Vero (ATCC CCL-81) and Vero E6 (ATCC CRL-1586) cells were grown in OptiMEM (Thermo Fisher) supplemented with 5% FBS.Vero cells were used to expand MPV and MPV/S-2P virus stocks.Vero E6 cells were used for SARS-CoV-2 neutralization assays and titrations.SARS-CoV-2 virus stocks were expanded on Vero E6 cells or Vero E6 cells stably expressing human TMPRSS2 31 .

SARS-CoV-2 virus stocks
The SARS-CoV-2 USA-WA1/2020 challenge virus (lineage A; GenBank MN985325 and GISAID accession ID: EPI_ISL_404895; obtained from Natalie Thornburg, Sue Gerber, and Sue Tong, Centers for Disease Control and Prevention [CDC], Atlanta, GA) was passaged twice on Vero E6 cells.The USA/CA_CDC_5574/2020 isolate (lineage B.1.1.7,GISAID: EPI_ISL_751801; CDC) and the USA/MD-HP01542/2021 isolate (lineage B.1.351,GISAID: EPI_ISL_890360; sequence deposited by Christopher Paul Morris, Chun Huai Luo, Adannaya Amadi, Matthew Schwartz, Nicholas Gallagher, and Heba H. Mostafa, The Johns Hopkins University; isolate obtained from Andrew Pekosz, The Johns Hopkins University, Baltimore, MD) were passaged on Vero E6 cells stably expressing TMPRSS2.Titration of SARS-CoV-2 stocks was performed by determination of the TCID 50 in Vero E6 cells.Whole-genome Illumina deep sequencing analysis con rmed that the complete genome sequences of the SARS-CoV-2 used for experiments were identical to that of consensus sequences, except for minor backgrounds of reads (< 10%).All experiments with SARS-CoV-2 were conducted in Biosafety Level-3 containment laboratories approved by the US Department of Agriculture and CDC.

Generation of MPV and MPV/S-2P
The recombinant MPV vector used in this study was based on a previously-described reverse-genetics system for MPV strain 15 5 .In this version, the downstream 67% of the L ORF were codon-pair optimized for e cient expression in humans.MPV/S-2P 6 contains an additional gene encoding the full-length ORF encoding the 1,273 amino acid S protein derived from the ancestral Wuhan-Hu-1 sequence (GenBank MN908947), codon optimized for human expression, with two prefusion-stabilizing proline substitutions (aa 986 and 987), and four aa mutations (RRAR to GSAS, aa 682-685) to ablate the furin cleavage site between S1 and S2 32 .Virus stocks were ampli ed in Vero cells, and the genomic sequences were con rmed by Sanger sequencing of overlapping RT-PCR amplicons.

Immunization and challenge of rhesus macaques
The study was approved by the NIAID Animal Care and Use Committee.Twelve juvenile and young adult male Indian-origin rhesus macaques (Macaca mulatta), seronegative for SARS-CoV-2, were immunized intranasally (0.5 ml per nostril) and intratracheally (1 ml) with a total dose of 6.3 log 10 PFU of MPV empty vector or MPV/S-2P.On day 28, 4 of the 12 animals received a second dose of MPV/S-2P.Animals were monitored daily from day − 3 through the end of the study.
The schedule for the macaque experiment is shown in Fig. 1A.On indicated days, nasopharyngeal swabs (NS) and tracheal lavages (TL) were performed to evaluate vaccine virus replication.NS were collected using cotton-tipped applicators, placed in 2 ml Leibovitz (L-15) medium with 1x sucrose phosphate (SP) as stabilizer and vortexed for 10 seconds.TL were mixed 1:1 with L-15 medium containing 2x SP.Nasal washes (NW) for analysis of mucosal antibodies in the UA were performed on indicated days (Fig. 1A) using 1 ml of Lactated Ringer's solution per nostril.Aliquots of all samples were snap-frozen in dry ice and stored at -80˚C.Blood was collected on days − 3, 9, 14, 21, 25 or 28 after the prime and on days 37, 42, 49 and 56 from the macaques that received the MPV/S-2P boost dose for analysis of serum antibodies and/or peripheral blood mononuclear cells (PBMCs).Bronchoalveolar lavages (BAL) were performed for analysis of mucosal antibodies and mononuclear cells in the LA.
Four weeks after immunization (or four weeks after boost for animals that received a prime/boost regimen), macaques were transferred to an Animal Biosafety Level 3 (ABSL3) facility.On day 31-32 for primed macaques (two animals per group per day for technical reasons) or day 58 for primed/boosted macaques, animals were challenged intranasally and intratracheally with 6.3 log 10 TCID 50 of SARS-CoV-2 USA-WA1/2020.Post-challenge samples were collected as described for the post-immunization sampling.On day 6 pc, animals were euthanized and six separate samples from individual lung lobes were collected, snap frozen in dry ice, and stored at -80˚C.
Replication of MPV and derivatives in the upper and lower airways of macaques MPV and MPV/S-2P shedding in the upper and lower airways of macaques was evaluated by dualstaining immunoplaque assay of NS and TL samples.Brie y, tenfold serial dilutions were incubated in duplicate wells on Vero cells under methyl-cellulose overlay.After 11-day incubation at 32°C, monolayers were xed using 80% methanol, MPV was detected using a rabbit hyperimmune MPV antiserum, and SARS-CoV-2 S-2P was detected using the human monoclonal antibody CR3022.Anti-rabbit IRDye680 and anti-human IRDye800 infrared uorophore-labeled secondary antibodies were used to visualize MPV and S.

Dual IgG and IgA ELISAs
Levels of serum and mucosal anti-SARS-CoV-2 S or anti-receptor binding domain (RBD) IgG and IgA elicited by MPV/S-2P were determined using a dual IgG/IgA or IgG/dimeric IgA ELISA.Serum, BAL, and NW samples were heated at 56°C for 30 min to inactivate complement and reduce potential risk from any residual viruses.Black ninety-six-well plates (MaxiSorp, Thermo Fisher Scienti c, cat #437111) were coated with 100 µl/well of SARS-CoV-2 S-2P or S-6P (1 µg/ml) or RBD (2 µg/ml) 31 33 in 50 mM carbonate coating buffer, and incubated overnight at 4°C.Plates were washed three times with 250 µl washing buffer [PBS with 0.1% IGEPAL CA-630] and blocked with 250 µl DPBS containing 5% dry milk (W/V).Samples were serially diluted in sample dilution buffer (PBS + 5% dry milk + 0.2% IGEPAL CA-630) and transferred to antigen-coated assay plates in duplicate.After 1-hour incubation, plates were washed as above.100 µl per well of secondary antibodies [goat anti-monkey IgG(H + L)-HRP (ThermoFisher, cat# PA1-84631, 1:10,000) and goat anti-monkey IgA-alpha chain-speci c biotin (Alpha Diagnostic International, cat# 70049, 1:5,000) or mouse anti-rhesus J chain-biotin (NHPRR, cat# PR-3316, 1:5,000)] in dilution buffer was added, and plates were incubated for 1 h.Plates were washed and 100 µl per well of diluted Streptavidin-Europium (PerkinElmer, cat# 1244 − 360), diluted 1:2,000 in PBS + 0.2% IGEPAL CA-630, was added.Plates were incubated for 1 h and washed.Then, 50 µl of Pierce ECL (ThermoFisher, cat# 32106) per well was added and plates were read on the Synergy neo (BioTek) plate reader to collect IgG luminescence data.Plates were washed, and 100 ml per well of Enhancement Solution (PerkinElmer: 4001-0010) was added.Plates were read again using a program for time-resolved uorescence (TRF; excitation 360/40; emission 620/40) to collect IgA data.Data were processed as follows: (i) the average reading was calculated from duplicate wells, (ii), the average reading from blank samples was subtracted, (iv) the cut-off value was set to the blank average plus three standard deviations.The IgG and IgA titers were determined by interpolating the sigmoid standard curve generated on Prism 9.0.

Serum antibody avidity assay
The avidity of serum antibodies to S-6P was determined by modi cation of the dual IgG and IgA ELISA protocol.After incubation with diluted sera and before addition of secondary antibody mixtures, one set of plates was incubated with 100 µl PBS, while the other set of duplicate plates were treated with 100 µl for 1 h at 37˚C.Mixtures were added to quadruplicate wells of Vero E6 cells in 96-well plates and incubated for four days.The 50% neutralizing dose (ND 50 ) was the highest dilution of serum that prevented cytopathic effect in 50% of the wells.Serum neutralizing antibody titers against the MPV vector were also quanti ed on Vero cells using 60% plaque reduction neutralization tests (PRNT 60 ).

ACE2 binding inhibition assays
As a complement to the neutralization assay, we evaluated the ability of heat-inactivated sera and BAL uid to inhibit binding of ACE2 to SARS-CoV-2 spike proteins (Meso Scale Diagnostics, cat# K15586U, K15609U, K15671U).Sera were diluted 1:20 and BAL and NW samples were diluted 1:2.Each sample was evaluated in duplicate, and plates were prepared and analyzed following the manufacturer's instructions as previously described 27 .Brie y, plates in which each well was coated with 10 different spike proteins were blocked for 1h using MSD blocker A buffer, followed by a wash with MSD washing buffer.Diluted test samples were added, and plates were further incubated for 2h on a plate shaker.Sulfo-Tag labelled soluble ACE2 was added and after 1h incubation, plates were washed.The MSD GOLD electrochemiluminescence read buffer B was added, and chemiluminescence of bound ACE2-Sulfo-Tag was acquired on a Meso1300 Quickplex reader.The ACE2 binding inhibition is calculated as percent inhibition relative to no-sample controls.

Identi cation of S-speci c B cells in the blood of immunized macaques
To characterize S-speci c B cells in the blood of immunized macaques, single-cell suspensions of PBMCs were plated in 96 well plates at ≤ 3.5x10 6 cells/ml in 200 µl PBS + 1% FBS.Cells were centrifuged at 544 syndrome coronavirus 2 (SARS-COV-2) spike protein" led by the United States, Department of Health and Human Services.
Materials & correspondence: Materials described in the manuscript are available upon request under a material transfer agreement with the NIAID.Requests should be addressed to UJB.A, D) and IgA (B, E) was measured using ELISA and dissociation-enhanced lanthanide uorescent (DELFIA) assays, respectively (limit of detection: 1.6 log 10 , dotted line).Dimeric anti-S IgA in the UA and LA was also evaluated by DELFIA assay (C, F; limit of detection is 1.0 log 10 , dotted line).(G) BAL samples were analyzed for their ability to block binding of tagged, soluble ACE2 to puri ed S protein from the vaccine-matched SARS-CoV-2 S protein (Wuhan strain) or variants of concern.ACE2 binding inhibition is expressed as % inhibition relative to a no-serum control (see also Fig. S2A).Medians, min, and max values, 25 th

1. 0
M sodium thiocyanate solution (NaSCN; Sigma-Aldrich) for 15 min to strip low-a nity S-6P binding antibodies.Plates were washed prior to addition of secondary antibodies.Anti-S IgG and IgA titers of each serum sample in presence or absence of NaSCN were calculated.The avidity index (AI) was calculated as the ratio of the NaSCN-treated IgG or IgA titers to the PBS-treated serum IgG or IgA titers.A maximum avidity index of 1 corresponds to no loss of antibody binding to S following NaSCN treatment.Serum neutralization assay of SARS-CoV-2 and MPV Neutralization assays to determine serum antibody titers against SARS-CoV-2 WA1/2020, B.1.1.7,or B.1.351 of immunized macaques were done in a BSL3 laboratory.Sera were heat-inactivated at 56°C for 30 min and two-fold serially diluted in Opti-MEM prior to incubation 1:1 with 100 TCID 50 of SARS-CoV-2

Figure 1 Evaluation
Figure 1

Figure 3 Prime
Figure 3

Figure 5 IN
Figure 5