Efficacy and breadth of adjuvanted SARS-CoV-2 receptor-binding domain nanoparticle vaccine in macaques

Emergence of novel variants of the severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) underscores the need for next-generation vaccines able to elicit broad and durable immunity. Here we report the evaluation of a ferritin nanoparticle vaccine displaying the receptor-binding domain of the SARS-CoV-2 spike protein (RFN) adjuvanted with Army Liposomal Formulation QS-21 (ALFQ). RFN vaccination of macaques using a two-dose regimen resulted in robust, predominantly Th1 CD4+ T cell responses and reciprocal peak mean neutralizing antibody titers of 14,000-21,000. Rapid control of viral replication was achieved in the upper and lower airways of animals after high-dose SARS-CoV-2 respiratory challenge, with undetectable replication within four days in 7 of 8 animals receiving 50 µg RFN. Cross-neutralization activity against SARS-CoV-2 variant B.1.351 decreased only ∼2-fold relative to USA-WA1. In addition, neutralizing, effector antibody and cellular responses targeted the heterotypic SARS-CoV-1, highlighting the broad immunogenicity of RFN-ALFQ for SARS-like betacoronavirus vaccine development. Significance Statement The emergence of SARS-CoV-2 variants of concern (VOC) that reduce the efficacy of current COVID-19 vaccines is a major threat to pandemic control. We evaluate a SARS-CoV-2 Spike receptor-binding domain ferritin nanoparticle protein vaccine (RFN) in a nonhuman primate challenge model that addresses the need for a next-generation, efficacious vaccine with increased pan-SARS breadth of coverage. RFN, adjuvanted with a liposomal-QS21 formulation (ALFQ), elicits humoral and cellular immune responses exceeding those of current vaccines in terms of breadth and potency and protects against high-dose respiratory tract challenge. Neutralization activity against the B.1.351 VOC within two-fold of wild-type virus and against SARS-CoV-1 indicate exceptional breadth. Our results support consideration of RFN for SARS-like betacoronavirus vaccine development.

The coronavirus infectious disease 2019  pandemic, precipitated by severe acute 74 respiratory syndrome coronavirus-2 (SARS-CoV-2), continues to threaten global public health 75 and economies. Threats of future outbreaks also loom, as evidenced by three emergent SARS-76 like diseases caused by zoonotic betacoronaviruses in the last two decades. While several 77 emergency use authorized (EUA) vaccines currently in use are expected to curb both disease and 78 transmission of SARS-CoV-2 (1-6), the emergence of circulating variants of concern (VOC) less 79 sensitive to vaccine-elicited immunity has raised concerns for sustained vaccine efficacy (7). 80 Logistic challenges of vaccine production, distribution, storage and access for these vaccines will 81 need to be resolved equitably to achieve resolution to the pandemic (8,9). The rapid and 82 unparalleled spread of SARS-CoV-2 has driven an urgent need to deploy scalable vaccine 83 platforms to combat the ongoing pandemic and mitigate future outbreaks. 84 85 Current vaccines primarily focus the immune response to the spike glycoprotein (S) on the virion 86 surface as it mediates host cell viral fusion and entry. The receptor-binding domain (RBD) of S 87 engages the primary host cell receptor, Angiotensin-converting enzyme 2 (ACE2), for both 88 SARS-CoV-2 and SARS-CoV-1, making RBD a promising domain for vaccine elicited immune 89 focus (10-12). Moreover, many of the potently neutralizing monoclonal antibodies isolated 90 against SARS-CoV-2 target the RBD (13,14). Vaccination of nonhuman primates with RBD-91 encoding RNA or DNA protects against respiratory tract challenge, indicating that immune 92 responses to the RBD can prevent viral replication (15,16). RBD vaccination also elicits cross-93 reactive responses to circulating SARS-CoV-2 VOC in both animals and humans (17,18), with 94 decrements against the more difficult to neutralize B.1.351 variant similar to that seen with S 95 immunogens (19). The breadth of RBD immunogenicity is further supported by the ability of 96 RBD-specific monoclonal antibodies isolated from SARS-CoV-1 convalescent individuals to 97 cross-neutralize SARS- 21). These findings suggest potential for RBD-based 98 vaccines being efficacious against SARS-CoV-2 variants and other coronavirus species. 99 100 Approaches to improve immunogenicity of S or RBD protein vaccines include optimizing 101 antigen presentation and co-formulating with adjuvants to enhance the protective immunity. One 102 common approach to enhance the elicitation of adaptive immune responses is the multimeric 103 presentation of antigen, for example, on the surface of nanoparticles or virus-like particles (22). 104 Presenting RBD in ordered, multivalent arrays on the surface of self-assembling protein 105 nanoparticles is immunogenic and efficacious in animals (23-28), with improved 106 immunogenicity relative to monomeric soluble RBD and cross-reactive responses to variants (17, 107 24, 26). However, it is unknown whether RBD nanoparticle vaccines are able to protect against 108 infection in primates, which have become a standard model for benchmarking performance of 109 vaccine candidates. Liposomal adjuvants incorporating QS-21, such as that used in the 110 efficacious varicella zoster vaccine, SHINGRIX ® , may augment protective immunity to  CoV-2 vaccines. Such adjuvants have previously demonstrated superior humoral and cellular 112 immunogenicity relative to conventional adjuvants (29,30). 113

114
Here, we evaluate the use of a ferritin nanoparticle vaccine presenting the SARS-CoV-2 RBD 115 (RFN) adjuvanted with the Army Liposomal Formulation QS-21 (ALFQ) (31). Both ferritin 116 nanoparticles and ALFQ have been evaluated for vaccination against multiple pathogens in 117 humans in phase 1 clinical trials (32)(33)(34). We demonstrate in a nonhuman primate model that 118 immunization with RFN induces robust and broad antibody and T cell responses, as well as 119 protection against viral replication and lung pathology following high-dose respiratory tract 120 challenge. 121 122 RESULTS 123

Humoral responses to vaccination 138
Multiple vaccine-matched humoral immune responses were measured longitudinally in serum 139 following vaccination. First, binding antibody responses to the SARS-CoV-2 prefusion stabilized 140 S protein (S-2P) (35) were assessed by MSD. Immunization with either 5 or 50 μg of RFN 141 elicited S-specific IgG two weeks following the prime (21,896 and 79,109 AU/mL, respectively) 142 (Fig. 1B). These responses increased two weeks following the second immunization (420,249 143 and 293,509 AU/mL). Boosting was greater with the 5 µg dose, achieving a 19-fold increase 144 relative to post-prime versus ~3.7-fold with 50 µg. Responses continued to marginally increase 145 four weeks following the second immunization as well as two weeks post-challenge. 146 Unvaccinated control animals mounted responses ~1,000-fold over baseline within two weeks 147 post challenge and these responses were ~65-fold lower than those in vaccinated animals after 148 challenge. 149

150
Given the importance of the RBD in mediating viral entry and the majority of neutralizing 151 antibody responses targeting this domain, RBD-specific humoral responses were also measured. 152 RFN induced binding antibodies four weeks following the second immunization, with no 153 significant difference between vaccine dose groups (Fig. 1C). Responses in vaccinated animals 154 were ~60,000-fold over the background in unvaccinated controls and comparable in magnitude 155 to those against the S protein, consistent with an RBD-focused response. To confirm these 156 findings, the on-rate association between serum antibodies and RBD antigen was measured by 157 biolayer interferometry and longitudinal responses tracked with S-binding and pseudovirus 158 neutralizing responses (Fig. S1). Again, vaccine dose groups did not differ. Functional activity of 159 serum antibodies to inhibit ACE2 binding to the RBD antigen was also measured and present, 160 with high magnitude responses elicited by RFN at both the 5 and 50 μg doses (Fig. 1D). 161

162
Neutralizing antibody responses against SARS-CoV-2 using a pseudovirus neutralization assay 163 followed a similar pattern as the S-specific binding responses (Fig. 1E). Peak ID50 titers of 164 14,540 and 21,298 were observed two weeks following the boost for the 5 and 50 μg RFN doses,165 respectively. Neutralizing responses increased markedly between the prime and boost, rising 48-166 and 32-fold between study weeks two and six. Among the 50 μg RFN vaccinated animals 167 followed two weeks post-challenge, neutralizing responses declined six weeks post-boost by 168 approximately one log relative to peak values, indicating neutralizing responses may decay more 169 quickly than binding antibodies. 170

171
Neutralizing responses were also evaluated using an authentic SARS-CoV-2 virus neutralization 172 assay (USA-WA1 isolate). Robust neutralizing titers were detected in all RFN vaccinated 173 animals ( Fig. 1F). Median EC50 -1 titers were ~3,800 for both dose groups, though slightly more 174 variable with 5 µg dosing. This result paralleled responses assessed by a pseudovirus assay (Fig.  175 1E,G). Since serum from convalescent COVID-19 human cases is frequently used as a 176 benchmarking reference for vaccine immunogenicity in clinical and pre-clinical studies, we 177 compared RFN-vaccinated macaque pseudovirus neutralizing titers to those of 41 convalescent 178 individuals 4-8 weeks post-COVID infection. Responses in the 50 μg group were on average 13-179 fold higher than the convalescent individuals, indicating that RFN elicited higher antibody titers 180 than observed in the first months following human infection. Summarizing, RFN vaccination 181 generated strong RBD-specific binding antibodies with potent neutralization activity which block 182 the interaction between the RBD region of SARS-CoV-2 S and the host ACE2 receptor. 183 184 Non-neutralizing antibody effector functions associated with vaccine-mediated protection against 185 other viruses may also be important for SARS-CoV-2 (36, 37). Strong IgG-mediated cellular 186 opsonization responses were observed following the second immunization, while IgM and IgA 187 were more modest ( Fig. S2A-C). Serum antibody-dependent phagocytosis mediated by either 188 monocytes (ADCP) or neutrophils (ADNP) as well as complement deposition (ADCD) 189 responses were also robust in both vaccinated groups and consistently peaked at week six (Fig. 190 S2D-F). A similar pattern was seen for antibody-dependent trogocytosis (38) (Fig. S2G). of the animals vaccinated with 5 µg RFN (Fig. 2C). The average frequency in responders was 213 0.15%. The CD4+ T cell activation marker, CD40L, which promotes B cell antibody isotype 214 switching, was highly expressed by S-specific cells (Fig. 2D). Responses ranged from ~1-7% 215 after 50 µg RFN and were observed in all eight animals, while response rates and magnitude 216 were slightly reduced with the 5 µg dose (~0.7-2% in six of seven animals). Overall, these data 217 show that adjuvanted RFN induced robust Th1-polarized polyfunctional CD4+ T cells favorable 218 for viral clearance and with critical B-cell help capability. 219 220

SARS-CoV-2 challenge efficacy 221
To assess the protective efficacy of RFN vaccination, animals were challenged with high-dose 222 (10 6 TCID50) SARS-CoV-2 USA-WA1 administered via the simultaneous IN/IT routes four 223 weeks following the second immunization. The presence of viral RNA was assessed in both the 224 upper (NP swabs and saliva) and lower (BAL) respiratory tract. Measurements were made of 225 both total RNA and subgenomic E mRNA (sgmRNA), considered a more specific indicator of 226 viral replication (42,43). Unvaccinated control animals all showed evidence of a robust 227 infection, with mean levels of sgmRNA in the BAL of ~10 6 copies/mL, and in the NP swabs of 228 ~10 7 copies/mL at day 2 post-challenge (Fig. 3). Moreover, viral replication was sustained at 229 >10 4 copies/mL sgmRNA for 7 days in the upper airways. In RFN vaccinated animals, the 230 magnitude and duration of viral replication was markedly reduced. In the 50 μg group, day 1 231 sgmRNA was reduced by 1 and 2 logs in the BAL and NP swabs, respectively. Rapid clearance 232 was observed by day 2 in five of eight animals in the upper airways and four of eight in the lower 233 airways. Both airways were void of replicating virus in all but one animal by day 4. Viral 234 control was also apparent after 5 µg RFN vaccination, though with slightly more breakthrough 235 replication early after challenge. The majority of animals had no detectable sgmRNA by day 4 in 236 both the upper and lower airways. starting on day 2. The kinetics of SARS-CoV-2 total RNA, which is more likely to reflect 243 material from the viral inoculum, paralleled results described above for sgmRNA in BAL, NP 244 swabs and saliva (Fig. S4). 245 246

Respiratory tract pathology and antigen expression 247
Vaccine efficacy was also assessed by histopathologic analysis of lung tissue from 3-5 macaques 248 from each group necropsied at day 7 post-challenge. By this point, all unvaccinated animals had 249 developed histopathologic evidence of multifocal, mild to moderate interstitial pneumonia (

Cross-reactive immunity to emergent SARS-CoV-2 variants and SARS-CoV-1 268
Given concerns about circulating SARS-CoV-2 viral variants' increased resistance to currently 269 available vaccines, we assessed RFN vaccinated macaque serum for neutralizing antibody 270 responses against two variants of concern, B.1.1.7 and B.1.351. In an authentic virus 271 neutralization assay, reciprocal neutralization ID50 GMT titers against B.1.1.7 were 73,983 two 272 weeks following the second 50 µg dose (Fig. 5A). This translated to ~3.8-fold greater titers than 273 those against the wild-type, vaccine-matched USA-WA1 strain. Titers against the two strains 274 were similar when measured by the pseudovirus neutralization assay (Fig. 5B). These trends 275 were observed regardless of vaccine dose, though responses were slightly lower with 5 µg RFN. In addition to SARS-CoV-2 variants of concern, another open question in the field is the ability 286 of existing SARS-CoV-2 vaccine platforms to protect against future SARS-like coronavirus 287 outbreaks. Cross-protective vaccine-elicited immunity against SARS-CoV-1 may be a useful 288 metric to address this question. We measured IgG antibody responses able to bind SARS-CoV-1 289 RBD by biolayer interferometry in macaque serum at week 2 following the second vaccination. 290 All RFN vaccinated animals developed cross-reactive binding antibodies to SARS-CoV-1 at 291 levels approximately half those to SARS-CoV-2 ( Fig. 5D, Fig. S1). Binding responses were also 292 measured to a series of SARS-CoV-1 and SARS-CoV-2 antigens using a Luminex assay (Fig. 293 S5), with strong binding responses observed to the SARS-1 S1 subunit and RBD, but not against 294 the S2 subunit or N-terminal domain. SARS-CoV-1 RBD-specific binding antibody responses 295 were ~70% that of the SARS-CoV-2 response. The functional capacity of these cross-binding 296 antibodies to mediate effector activity was assessed in an ADCP assay using SARS-CoV-1 297 trimeric S antigen. SARS-CoV-1 ADCP responses were observed in plasma of all vaccinated 298 animals and were comparable between the dose groups ( Fig. 5E). 299 300 Neutralizing titers against SARS-CoV-1 were measured using both authentic virus and 301 pseudovirus neutralization assays, with cross-neutralizing responses observed in most RFN 302 vaccinated animals (Fig. 5F,G). Significant authentic virus neutralization titers were elicited by 303 50 μg RFN two weeks following the second immunization. SARS-CoV-1 pseudovirus 304 neutralization activity was also observed in both the 50 and 5 μg groups, though background in a 305 subset of control animals limited interpretation of both assays. 306 307 To assess T cell immunity cross-reactivity to SARS-CoV-1, we evaluated whether the RFN 308 vaccine-elicited T cells could recognize SARS-CoV-1 S. PBMC stimulated with SARS-CoV-1 S 309 peptide pools were stained for intracellular cytokine expression to quantitate cross-reactive T 310 cells. Significant CD4+ T cell Th1 responses were observed following the 50 µg RFN 311 vaccination series, though they were ~5-fold lower in magnitude than those to SARS-CoV-2 S 312 ( Fig. 5H). SARS-CoV-1 S-specific CD40L responses were comparable to the Th1 responses for 313 both dose groups (Fig. S7A). IL-21 and Th2 CD4+ T cell responses were minimal or negligible 314 ( Fig. S7B,C). Significant cross-reactive CD8+ T cells were elicited and similar in magnitude to 315 SARS-CoV-2-specific responses (~0.1-0.3%) (Fig. 5I), suggesting that the CD8+ T cell RBD 316 epitope specificities elicited by RFN vaccination may be relatively conserved. Again, responses 317 trended greater with the higher vaccine dose. These data indicate that S-specific CD4+ and 318 CD8+ T cells generated by ALFQ-adjuvanted RFN were able to cross-react with sequence 319 divergent SARS-CoV-1. 320

DISCUSSION 321
New SARS-CoV-2 vaccines may be needed to address concerns regarding emerging virus 322 variants less sensitive to immunity elicited by current vaccines (1,(44)(45)(46)(47). In this study, we 323 antibodies were also reactive to SARS-CoV-1, which is 36% amino acid sequence divergent 334 from SARS-CoV-2 in the RBD (50). Overall, these data indicate broad, potent and efficacious 335 immunity elicited by  337 This study provides strong evidence that RBD-directed vaccination in primates is able to protect 338 against SARS-CoV-2 infection and elicit neutralization breadth against variant B.1.351, which 339 has shown the greatest resistance to neutralization by vaccinee sera (19,51,52). While many 340 RBD-based immunogens have been shown to be immunogenic in small and large animal models 341 vaccines, which ranged from 408-1,862 (48, 49, 55). While neutralizing activity is unlikely to be 359 the sole determinant of vaccine-mediated protection, it has been predictive of efficacy in human 360 trials (1). Therefore, ~10-50-fold greater neutralizing titers by RFN relative to those elicited in 361 NHP studies by efficacious vaccines currently in clinical use strongly suggests that RFN would 362 be protective in humans. In addition, breadth against the B.1.351 VOC appears greater than that 363 of EUA vaccines in humans, as the modest ~2-fold reduction in B.1.351 neutralization activity 364 relative to wild-type virus reported here is less than the ~6-12-fold reduction in mRNA vaccinee 365 sera (19). The most advanced platform closest in design and composition to RFN is NVX-366 CoV2373, a prefusion spike nanoparticle vaccine delivered with a saponin-based Matrix-M 367 adjuvant. NVX-CoV2373 elicited neutralizing antibody titers of 6,400-17,000 in macaques (56, 368 57), greater than those achieved by current EUA vaccines and less than or similar to those 369 elicited by RFN. T cell immunity was also more pronounced with RFN. S-specific Th1 CD4+ T 370 cells ranged from 0.5-5% following 50 µg RFN, compared to peak values of 0.1-0.2% in NHP 371 vaccinated with the EUA vaccines (48,49,55). The overall magnitude of these differences was small and suggests that both the RBD and S 391 proteins are similarly immunogenic and protective when complexed to ferritin nanoparticles and 392 administered with ALFQ adjuvant at these vaccine doses. However, S-based immunogens may 393 offer the advantage of broadening the specificity of the immune response to other domains and 394 subdomains of the spike protein, limiting potential for viral escape. These findings support 395 further clinical development of both products. 396

397
There exists a strong potential for future pandemics arising from zoonotic SARS-like 398 betacoronaviruses entering into humans. We report SARS-CoV-2 RFN vaccine-elicited 399 responses that cross-react with the S glycoprotein SARS-CoV-1, including binding antibody 400 titers within an order of magnitude of those to SARS-CoV-2. The observed cross-neutralizing 401 and binding reactivity to SARS-CoV-1 suggests that adjuvanted RFN may be a viable candidate 402 for vaccination against future betacoronavirus outbreaks. Work is ongoing to elucidate the 403 potential mechanisms of cross-protective responses in this study, including epitope mapping of 404 the antibody binding responses. Taken together, these findings support continued development of 405 RFN vaccines for managing COVID-19 and SARS-like betacoronavirus outbreaks. 406

MATERIALS AND METHODS 407
Vaccine and adjuvant production 408

Immunogen formulation 453
RFN was diluted in dPBS to 0.1 mg/mL or 0.01 mg/mL and mixed 1:1 with 2X ALFQ on a tilted 454 slow-speed roller at room temperature for 10 min, followed by incubation at 4 o C for 50 min. 455 Reagents were equilibrated to room temperature before use and immunizations were performed 456 within 4 h of vaccine formulation. Each vaccine comprised a 1.0 mL solution of RFN formulated 457 with ALFQ. The 3D-PHAD ® and QS-21 doses were 200 and 100 µg, respectively. 458

Study design and procedures 459
Twenty-three male and female specific-pathogen-free, research-naïve Chinese-origin rhesus 460 macaques (age 3 -7 years) were distributed-on the basis of age, weight and sex-into 3 cohorts 461 of 7-8 animals (Table S1). Animals were vaccinated intramuscularly with either 50 or 5 μg of 462 RFN, formulated with ALFQ, and control group animals received 1 mL of PBS, in the 463 anterior proximal quadriceps muscle, on alternating sides with each dose in the series. washed with buffer before the addition of reference standard and calibrator controls. Serum 496 samples were diluted at 1:1,000 -1:100,000 in diluent buffer, then added to each of four wells. 497 Plates were incubated for 2 h at room temperature while shaking at 700 rpm, then washed. MSD 498 SULFO-TAG TM anti-IgG antibody was added to each well. Plates were incubated for 1 h at RT 499 with shaking at 700 rpm and washed, then MSD GOLD TM Read buffer B was added to each 500 well. Plates were read by the MESO SECTOR SQ 120 Reader. IgG concentration was calculated 501 using DISCOVERY WORKBENCH ® MSD Software and reported as arbitrary units (AU)/mL. 502 For binding antibodies that block S or RBD binding to ACE2, antigen-coated plates were 503 blocked and washed as described above. Assay calibrator and samples were diluted at 1:25 -504 1:1,000 in MSD Diluent buffer, then added to the wells. Plates were incubated for 1 h at room 505 temperature while shaking at 700 rpm. ACE2 protein conjugated with MSD SULFO-TAG TM was 506 added and plates were incubated for 1 h at room temperature while shaking at 700rpm and 507 washed and read as described above. 508 Binding antibody measurements by octet biolayer interferometry were made using 509 HIS1K biosensors hydrated in PBS prior to use, using an Octet FortéBio Red96 instrument 510 (Sartorius, Fremont CA). All assay steps were performed at 30°C with agitation set at 1,000 rpm. 511 Baseline equilibration of the HIS1K biosensors (Sartorius, Fremont, CA)) was carried out with 512 PBS for 15 sec, prior to SARS-CoV2 RBD molecules (30 µg/mL diluted in PBS) loading for 120 513 sec. Biosensors were dipped in assay buffer (15 sec in PBS), dipped in the serum samples (100-514 fold dilution) for 180 sec, and binding response (nm) was recorded for 180 sec. 515

SARS-CoV-1 and SARS-CoV-2 pseudovirus neutralization 517
The S expression plasmid sequence for SARS-CoV-2 was codon optimized and modified to 518 remove an 18 amino acid endoplasmic reticulum retention signal in the cytoplasmic tail to 519 improve S incorporation into pseudovirions (PSV). PSV were produced by co-transfection of 520 Madison, WI). Neutralization dose-response curves were fitted by nonlinear regression using the 533 LabKey Server. Final titers are reported as the reciprocal of the serum dilution necessary to 534 achieve 50% inhibition SARS-CoV-2 (ID50, 50% inhibitory dose) or 90% inhibition for SARS-535 CoV-1 (ID90, 90% inhibitory dose). Assay equivalency was established by participation in the 536

SARS-CoV-2 Neutralizing Assay Concordance Survey run by the Virology Quality Assurance 537
Program and External Quality Assurance Program Oversite Laboratory at the Duke Human 538 Vaccine Institute. 539

Authentic SARS-CoV-2 wild-type neutralization assay 540
Authentic virus neutralization was measured using SARS-CoV-2 (2019-nCoV/USA_WA1/2020) 541 obtained from the Centers for Disease Control and Prevention and passaged once in Vero CCL81 542 cells (ATCC). Rhesus sera were serially diluted and incubated with 100 focus-forming units of 543 SARS-CoV-2 for 1 h at 37°C. Serum-virus mixtures were added to Vero E6 cells in 96-well 544 plates and incubated for 1 h at 37°C. Cells were overlaid with 1% (w/v) methylcellulose in 545 MEM. After 30 h, cells were fixed with 4% PFA in PBS for 20 min at room temperature then 546 washed and stained overnight at 4°C with 1 µg/mL of CR3022 antibody in PBS supplemented 547 with 0.1% saponin and 0.1% bovine serum albumin. Cells were subsequently stained with HRP-548 conjugated goat anti-human IgG for 2 h at room temperature. SARS-CoV-2-infected cell foci 549 were visualized with TrueBlue peroxidase substrate (KPL) and quantified using ImmunoSpot 550 microanalyzer (Cellular Technologies). Neutralization curves were generated with Prism 551 software (GraphPad Prism 8.0). 552

Total and subgenomic messenger (sgm) RNA quantification 663
Real-time quantitative PCR was carried out for subgenomic messenger RNA (sgmRNA) and 664 viral load RNA quantification from NP swab, BAL fluid and saliva samples. Primers targeted the 665 envelope (E) gene of SARS-CoV-2 (Table S2) TaqPath TM 1-Step RT-qPCR (Life Technologies, Thermo Fisher Scientific, Inc.). Amplification 674 cycling conditions were: 2 min at 25°C, 15 min at 50°C, 2 min at 95°C and 45 cycles of 3 sec at 675 94°C and 30 sec at 55°C with fluorescence read at 55°C. An RNA transcript for the SARS-CoV-676 2 E gene was used as a calibration standard. RNA copy values were extrapolated from the 677 standard curve and multiplied by 45 to obtain RNA copies/mL. A negative control (PBS) and 678 two positive controls, contrived using heat-inactivated SARS-CoV-2 (ATCC, VR-1986HK), at 679 10 6 and 10 3 copies/mL, were extracted and used to assess performance of both assays. 680

Histopathology 681
Formalin-fixed sections of lung tissue were evaluated by light microscopy and 682 immunohistochemistry. Lungs were perfused with 10% neutral-buffered formalin. Lung sections 683 were processed routinely into paraffin wax, then sectioned at 5 µm, and resulting slides were 684 stained with hematoxylin and eosin. Immunohistochemistry (IHC) was performed using the 685 Dako Envision system (Dako Agilent Pathology Solutions, Carpinteria, CA, USA). Briefly, after 686 deparaffinization, peroxidase blocking, and antigen retrieval, sections were covered with a 687 mouse monoclonal anti-SARS-CoV nucleocapsid protein (#40143-MM05, Sino Biological, 688 Chesterbrook, PA, USA) at a dilution of 1:4,000 and incubated at room temperature for 45 min. 689 They were rinsed, and the peroxidase-labeled polymer (secondary antibody) was applied for 30 690 min. Slides were rinsed and a brown chromogenic substrate 3,3' Diaminobenzidine (DAB) 691 solution (Dako Agilent Pathology Solutions) was applied for 8 min. The substrate-chromogen 692 solution was rinsed off the slides, and slides were counterstained with hematoxylin and rinsed. 693 The sections were dehydrated, cleared with Xyless, and then cover slipped. Tissue section slides 694 were evaluated by a board-certified veterinary anatomic pathologist who was blinded to study 695 group allocations. Immunohistochemistry (IHC) was performed with Dako Envision. Three 696 tissue sections from each of the right and left lung lobes were used to evaluate the lung 697 pathology. The histopathology of each section was evaluated on a scale of 0-5: 0 -absent, 1 -698 minimal (<10% of tissue section affected); 2 -mild (11-25% of tissue section affected); 3 -699 moderate (26-50% of tissue section affected); 4 -marked (51-75% affected); 5-severe (>75% of 700 tissue section affected). Sections were evaluated for edema, hyaline membranes, cellular 701 infiltrates, alveolar histiocytes, type II pneumocyte hyperplasia, interstitial fibroplasia, BALT 702 hyperplasia, bronchiolar degeneration, megakaryocytes in capillaries, and thrombosis. The 703 scores for all six sections of each pathologic finding were combined to create the final score 704 (TIIPH score) for individual animals. 705

Statistical analysis 706
Primary immunogenicity outputs of binding and neutralizing antibody titers as well as T cell 707 responses were compared across vaccination groups using the Kruskal-Wallis test. Non-708 parametric pairwise comparisons between groups were made using the post-hoc Dunn's test. 709 Statistical significance was preset at an alpha level of 0.05. 710

Data Availability 711
All data are available in the manuscript or the supplementary materials.    Figure 5