key: cord-1016958-a4k85jai
authors: Corbett, Kizzmekia S.; Gagne, Matthew; Wagner, Danielle A.; Connell, Sarah O’; Narpala, Sandeep R.; Flebbe, Dillon R.; Andrew, Shayne F.; Davis, Rachel L.; Flynn, Barbara; Johnston, Timothy S.; Stringham, Christopher; Lai, Lilin; Valentin, Daniel; Van Ry, Alex; Flinchbaugh, Zackery; Werner, Anne P.; Moliva, Juan I.; Sriparna, Manjari; O’Dell, Sijy; Schmidt, Stephen D.; Tucker, Courtney; Choi, Angela; Koch, Matthew; Bock, Kevin W.; Minai, Mahnaz; Nagata, Bianca M.; Alvarado, Gabriela S.; Henry, Amy R.; Laboune, Farida; Schramm, Chaim A.; Zhang, Yi; Wang, Lingshu; Choe, Misook; Boyoglu-Barnum, Seyhan; Shi, Wei; Lamb, Evan; Nurmukhambetova, Saule T.; Provost, Samantha J.; Donaldson, Mitzi M.; Marquez, Josue; Todd, John-Paul M.; Cook, Anthony; Dodson, Alan; Pekosz, Andrew; Boritz, Eli; Ploquin, Aurélie; Doria-Rose, Nicole; Pessaint, Laurent; Andersen, Hanne; Foulds, Kathryn E.; Misasi, John; Wu, Kai; Carfi, Andrea; Nason, Martha C.; Mascola, John; Moore, Ian N.; Edwards, Darin K.; Lewis, Mark G.; Suthar, Mehul S.; Roederer, Mario; McDermott, Adrian; Douek, Daniel C.; Sullivan, Nancy J.; Graham, Barney S.; Seder, Robert A.
title: Protection against SARS-CoV-2 Beta Variant in mRNA-1273 Boosted Nonhuman Primates
date: 2021-08-12
journal: bioRxiv
DOI: 10.1101/2021.08.11.456015
sha: 6489fb3a5d309cddf78c2b451169872d28a12411
doc_id: 1016958
cord_uid: a4k85jai

Neutralizing antibody responses gradually wane after vaccination with mRNA-1273 against several variants of concern (VOC), and additional boost vaccinations may be required to sustain immunity and protection. Here, we evaluated the immune responses in nonhuman primates that received 100 µg of mRNA-1273 vaccine at 0 and 4 weeks and were boosted at week 29 with mRNA-1273 (homologous) or mRNA-1273.β (heterologous), which encompasses the spike sequence of the B.1.351 (beta or β) variant. Reciprocal ID50 pseudovirus neutralizing antibody geometric mean titers (GMT) against live SARS-CoV-2 D614G and the β variant, were 4700 and 765, respectively, at week 6, the peak of primary response, and 644 and 553, respectively, at a 5-month post-vaccination memory time point. Two weeks following homologous or heterologous boost β-specific reciprocal ID50 GMT were 5000 and 3000, respectively. At week 38, animals were challenged in the upper and lower airway with the β variant. Two days post-challenge, viral replication was low to undetectable in both BAL and nasal swabs in most of the boosted animals. These data show that boosting with the homologous mRNA-1273 vaccine six months after primary immunization provides up to a 20-fold increase in neutralizing antibody responses across all VOC, which may be required to sustain high-level protection against severe disease, especially for at-risk populations. One-sentence summary mRNA-1273 boosted nonhuman primates have increased immune responses and are protected against SARS-CoV-2 beta infection.

geometric mean titers (GMT) against live SARS-CoV-2 D614G and the b variant, were 4700 48 and 765, respectively, at week 6, the peak of primary response, and 644 and 553, respectively, at 49 a 5-month post-vaccination memory time point. Two weeks following homologous or 50 heterologous boost b-specific reciprocal ID50 GMT were 5000 and 3000, respectively. At week 51 38, animals were challenged in the upper and lower airway with the b variant. Two days post-52 challenge, viral replication was low to undetectable in both BAL and nasal swabs in most of the 53 boosted animals. These data show that boosting with the homologous mRNA-1273 vaccine six 54 months after primary immunization provides up to a 20-fold increase in neutralizing antibody 55 responses across all VOC, which may be required to sustain high-level protection against severe 56 disease, especially for at-risk populations. 57 58 One-sentence summary: 59 mRNA-1273 boosted nonhuman primates have increased immune responses and are protected 60 against SARS-CoV-2 beta infection.

individuals; this group additionally served as a homologous vaccine control for the b challenge 173 (Fig. S1 ). S-specific IgG GMT for WA-1 and b were 5100 and 8900 two weeks after the first 174 immunization of mRNA-1273.b and increased 4-5-fold after a second immunization (Fig. S2A) . 175

Two doses of mRNA-1273.b resulted in D614G and b live virus ID50 GMT of 198 and 788, 176 respectively (Fig. S2B) . Potent S-specific antibody responses against WA-1 or b were also 177 detected in BAL (Fig. S2C ) and nasal washes (Fig. S2D) . Mucosal antibody responses were 178 further exhibited by ACE2 binding inhibition, where mRNA-1273.b induced BAL antibodies 179 that reduced ACE2 binding to WA-1 and b S by a median of 21% and 48%, respectively (Fig.  180   2E) . In nasal washes, median reduction of ACE2 binding to WA-1 and b S binding was 30% and 181 56%, respectively (Fig. 2F) . Together, these data show mRNA-1273.b given as a primary 182 regimen elicits higher b-specific responses as compared to WA-1. 183 184 Serum antibody repertoire elicited by homologous or heterologous prime boost 185

To evaluate the impact of homologous (WA-1) or heterologous (β) boosting antigen on serum 186 antibody epitope specificity, the absolute value (RUs) and the relative proportion (percent 187 competition) of serum antibodies against 16 distinct antigenic sites (Table S2 ) on WA-1 SARS-188

CoV-2 S was measured using surface plasmon resonance (SPR). We evaluated serum antibody 189 specificity at week 35 (6 weeks post-boost), to allow sufficient time for B cell expansion 190 following exposure to naïve antigen in heterologous-boosted animals. The epitope specificity 191 spanned the S2, N-terminal domain (NTD), S1, and RBD subdomains of S protein and the 192 breadth was similar whether the boost was homologous or heterologous (Fig 3A) . The specificity 193 of serum antibodies was qualitatively similar across all animals within each group for the NTD 194 subdomain (average standard deviation 32.1-36.9), but not for RBD subdomain (average 195 standard deviation 50.2-56.1) (Fig 3A) . Between vaccine groups, there was more NTD site A 196 specificity in homologous than in heterologous boosted animals (Fig. 3A) . Both absolute and 197 relative serum reactivity to NTD site A (mAb 4-8, Table S2 ) were higher in homologous than 198 heterologous boosted animals at all time points post-boost (Fig. 3B) , indicating mutations found 199 in the β S protein (Δ242-244, R246I) may render the heterologous boost unable to recall a 200 primary mRNA-1273 immunization memory B-cell response to this antigenic site. In contrast, 201 RBD site H (mAb A23-97.1, Table S2 ) showed differences in relative but not absolute serum 202 reactivity, with heterologous inducing higher reactivity than homologous boosting at all post 203 memory boost time points (Fig 3C) . Because A23-97.1 binds equivalently to both WA-1 and β S 204 ( Fig. S3) , the higher relative (but not absolute) reactivity against this site for β is likely due to the 205 decreased contribution of NTD-A reactivity, as shown by the ratio of mRNA-1273.β to mRNA-206 1273 relative reactivity (Fig. 3D) . Overall, mRNA-1273 to b ratios decrease over time, 207 indicating a contraction of the naïve β-directed response and return to a memory, long-lived 208 epitope profile (Fig. 3D) , and suggesting heterologous boost does not alter absolute epitope 209 reactivity over time. In contrast, following primary immunization with mRNA-1273.β we 210 observed significantly increased absolute serum reactivity to RBD sites B, C, D, and F ( Fig. S4 ) 211 as compared to homologous boost, with sites C and D (A19-46.1, A19-61.1, Table S2 ) being 212 associated with strong neutralization potency against β (27). These data highlight differences 213 between primary immunization and heterologous boost with mRNA-1273.β and indicate primary 214 immunization plays a key role in shaping the overall serum antibody epitope repertoire. 215

The durability of vaccine-induced antibody responses is an integral component of the pandemic 218 response as efforts continue to mitigate the ongoing spread of SARS-CoV-2. Durable humoral 219 immunity is driven by the ability to generate and sustain memory B cells and long-lived 220 plasmablasts (28), and recent data in humans shows the potential for such responses after 221 vaccination with mRNA or primary infection (29, 30). To define how B cell specificity matures 222 overtime in mRNA-1273 immunized NHP and to assess how mRNA-1273 priming imprints 223 homologous or heterologous boost, we performed temporal analysis of WA-1 and b S-specific 224 memory B cells (Fig. S5) . Following the primary immunization series with mRNA-1273, the 225 frequency of memory B cells expressing antibody receptors dual reactive for both WA-1 and b S 226 at week 6 was 2-3%, with a much lower proportion of single WA-1 or b-specific B cells (Fig.  227 4A-B). Six months later, at week 29, there was ~10-fold reduction in the frequency of double-228 positive WA-1 and b S-specific memory B cells (Fig. 4A-B ), but these were restored (~10-fold 229 increase) following a memory boost with mRNA-1273 or mRNA-1273.b (Fig. 4) . In contrast, 230 boosting did not cause an increase in the frequency of single-positive WA-1 or b S-specific B 231 cells (Fig. 4) . This finding demonstrates a rapid recall response of primary vaccination B cells 232 and coincides with an increase in neutralizing antibody responses (Fig. 1E-J) . The mRNA-233 1273.b immunized NHP had a memory B cell response that also consisted of WA-1 and b S-234 specific double-positive specificity, but a higher proportion of single positive b S-specific cells 235 (Fig. 4) , consistent with b-specific skewing of antibody responses following mRNA-1273.b 236 primary immunization (Fig. S2) . These data confirm the observation from serum antibody 237 epitope mapping that both homologous and heterologous boosting can efficiently expand 238 memory B-cell responses that are maintained after primary immunization. SARS-CoV-2 from BAL or NS that had sgRNA_E levels <1.0 x10 4 RNA copies/mL (Fig. 6E ) 279 and <4.7 x 10 3 RNA copies/swab (Fig. 6F) , respectively. Nucleocapsid (N)-specific sgRNA 280 measurements followed the same trend, albeit to with higher detection sensitivity (Fig. S7) . 281

We also evaluated lung samples for pathology and detection of viral antigen (VAg) 7-9 days post 282 SARS-CoV-2 b challenge. Inflammation was minimal to mild and was similar across lung 283 samples from vaccinated NHP, with rare cases showing a moderate to severe response (Fig. S8  284 and Table S3 ). The inflammatory changes in the lung were characterized by a mixture of macrophages and polymorphonuclear cells present within some alveolar spaces and mild to 286 moderate expansion of alveolar capillaries with mild Type II pneumocyte hyperplasia at day 7 287 post-challenge to changes more consistent with lymphocytes, histiocytes and fewer 288 polymorphonuclear cells associated with more prominent and expanded alveolar capillaries, 289 occasional areas of perivascular and peribronchiolar inflammation, and Type II pneumocyte 290 hyperplasia at days 8-9 post-challenge. SARS-CoV-2 antigen (Ag) was detected in 6/6 control 291 NHP evaluated at days 7, 8, or 9 after infectious challenge, with 3/6 showing Ag present in 292 multiple lobes. In vaccinated NHP, Ag was detected in only 1/24 animals, and in a single lobe 293 (mRNA-1273x3, Table S3 ). These results, together with sgRNA data, confirm that boosting 294 mRNA-1273 immunized NHP limits b viral replication in the lower and upper airway. 295

The SARS-CoV-2 mRNA-1273 vaccine shows between 90-100% protection against WA-1 (2), 298 a, or b variants (32) when administered as two doses four weeks apart and assessed within a 2 299 month window. Kinetic analyses of antibody responses following vaccination with mRNA-1273 300 (7) or BNT162b2 (30) show reduction of neutralizing activity from the peak humoral response 301 after the second immunization through day 209 (7, 17). Consistently, the β variant is the most 302 neutralization resistant of all VOC to date. However, it has not yet been established how this 303 resistance affects the durability of protective efficacy in humans. Recent clinical reports from 304

Israel show that cohorts immunized with the BioNTech/Pfizer vaccine over 6 months ago may 305 present with symptomatic disease when infected with VOC (33). These data suggest that it may 306 be important to boost antibody responses, especially against VOC, to sustain protection against 307 severe disease, especially in at-risk cohorts and to reduce the potential for transmission. A related issue is whether the additional boost should be matched to a specific newly arising variant 309 (heterologous) or if homologous boosting with the original vaccine based on ancestral strains 310 will be sufficient to generate broadly neutralizing antibody responses against VOC. Here, we 311

show that boosting NHP ~ 6 months after their primary vaccination series with either mRNA-312 1273 or mRNA-1273.β significantly increased serum neutralizing activity against all variants and 313 mediated high-level protection in the upper and lower airways against challenge with the β 314 variant. 315

Longitudinal neutralizing antibody responses against the benchmark D614G strain compared to β 316 and several additional VOC confirmed data from prior vaccine studies in humans or NHP (7, 30, 317 34), that there was a 3-10 fold reduction in neutralizing activity against several VOC, with β and 318 g variants showing the greatest reduction at the peak time point after the primary immunization 319 series. Here, all neutralizing responses were significantly reduced by week 24. Of note, there was 320 a more rapid decay of binding antibodies to VOCs compared to WA-1, but a relatively slower 321 decay in serum neutralizing activity against VOC. These data suggest that, although binding 322 antibody is an established correlate for short term protection, deeper analysis is needed to define 323 correlates of protection with contracted memory responses. These data also raise a possibility 324 that while there is a quantitative decrease in the magnitude of antibody responses over time 325

following vaccination, the loss of functional activity may be partially mitigated by affinity 326 maturation and improved quality of the response. It was recently reported that there may be 327 discordance in antibody evolution between convalescent individuals and those who receive an 328 mRNA vaccine. Specifically, less affinity maturation and a limited increase in breadth to the 329 RBD subdomain was observed in isolated monoclonal antibodies between 2-and 5-months post 330 mRNA vaccination (30). Our assessment of polyclonal serum antibody affinities (avidity) using the intact WA-1 S protein showed that vaccination with mRNA-1273 led to an increase in 332 avidity over time, confirming a qualitative improvement in the antibody response in NHP (29, 333

Following a homologous mRNA-1273, or heterologous mRNA-1273.β boost at 6 months after the 335 primary immunization series, there was a ~7-21-fold increase in serum antibody neutralizing 336 activity for D614G, β, and the other VOCs tested, and the increase was not significantly different 337 between homologous or heterologous boosting. These data are consistent with recent studies in 338

humans showing that boosting with mRNA-1273 or mRNA-1273.β approximately 6 months after 339 the primary immunization with mRNA-1273 elicited significantly higher neutralizing titers against 340 the D614G, β, and the other VOC tested compared to pre-boost titers (37). Importantly, there was 341 significant boosting of neutralizing responses against the β variant even in individuals that had 342 undetectable responses prior to the boost. Last, neutralizing responses after the boost in our NHP 343 study and in humans were significantly higher than the peak titers observed after the primary 344 vaccination series (17). 345

The increase in antibody responses post-boost suggests that there is significant B-cell memory 346 induced by primary mRNA vaccination that can be rapidly recalled following the boost (7, 29). 347

Here, we show that after primary immunization with mRNA-1273, the peak response yields a total 348 frequency of WA-1 and b-specific memory B cells of ~ 3%, and the majority (~ 80%) of these 349 responses recognized both S proteins, with a small proportion specific for WA-1 only. There was 350 a very low frequency of B cells specific for b only. After 6 months, all responses contracted to 351 exhibit a 10-fold decrease in the frequency of WA-1 and b-specific memory B cells, but these were 352 rapidly restored to a frequency of 3% after boosting. The relative frequency of B cells specific for 353 WA-1, b, or both did not change post-boost suggesting that priming with mRNA-1273 imprinted 354 the B-cell repertoire. 355

The boosting of B cells specific for WA-1 and β highlights the overlapping epitopes in S between 356 the two strains and likely accounts for neutralization breadth against VOCs. 3b (Fig. S1 ) are similar to prior studies (18, 19, 26, 34) . 

Total S-specific IgG in BAL and nasal washes was determined by meso-scale ELISA (MSD) as 443 previously described (46). 444

Avidity was assessed using a sodium thiocyanate (NaSCN)-based avidity ELISA against SARS-446

CoV-2 S-2P as previously described (18, 26) . The avidity index (AI) was calculated using the 447 ratio of IgG binding to S-2P in the absence or presence of NaSCN and reported as the average of 448 two independent experiments, each containing duplicate samples. 449

Serum epitope mapping competition assays were performed using a Biacore 8K+ (Cytiva) was captured on active sensor surface at a set concentration for 10 minutes. Monoclonal 479 antibodies (mAbs) were injected over both active and reference surfaces and allowed to bind to 480 saturation. Graphs represent reference-subtracted relative "analyte binding late" report points 481 (RUs), determined by aligning sensorgrams to Y (Response Units) = 0, beginning at the mAb 482 association stage using Biacore 8K Insights Evaluation Software (Cytiva). All assays were 483 performed using a Biacore 8K+ (Cytiva) surface plasmon resonance (SPR) spectrometer. 484

ACE2-binding inhibition was completed, as previously described (47), using 1:40 diluted BAL 486 and nasal wash samples. 487

Pseudotyped lentiviral reporter viruses were produced by the co-transfection of plasmids 489 encoding SARS-CoV-2 proteins from multiple variants, a luciferase reporter, lentivirus 490 backbone, and human transmembrane protease serine 2 (TMPRSS2) genes as previously 491 described (17, 37); reagent details are in Table S5 . Sera were tested, in duplicate, for 492 neutralizing activity against the pseudoviruses and percent neutralization was calculated in Prism 493 v9.0.2 (GraphPad). The lower limit of quantification was 1:40 ID50. 494

SARS-CoV-2 pseudotyped recombinant VSV-ΔG-firefly luciferase viruses were made by co-496 transfection of plasmid expressing full-length S, and subsequent infection with VSV∆G-firefly-497 luciferase and neutralization assays were completed on sera samples as previously described (15, 498 42); reagent details are in Table S5 . The lower limit of quantification was 1:40 ID50. 499

FRNT assays were performed on sera samples, in duplicate, as previously described (34); and 501 reagent details are in Table S5 . For samples that do not neutralize 50% of virus at the limit of 502 detection, 5 was plotted and used for geometric mean calculations. residual red blood cells were lysed using BD FACS Lysing Solution (BD Biosciences) for 10 520 minutes at room temperature. Following two additional washes, cells were fixed in 0.5% 521 formaldehyde (Tousimis Research Corp.) All antibodies were previously titrated to determine the 522 optimal concentration. Samples were acquired on an BD FACSymphony flow cytometer and 523 analyzed using FlowJo version 10.7.2 (BD, Ashland, OR). 524

Cryopreserved PBMC were thawed and rested overnight in a 37C/5% CO2 incubator. The next 526 morning, cells were stimulated with SARS-CoV-2 Spike protein peptide pools (S1 and S2) that All antibodies were previously titrated to determine the optimal concentration. Samples were 543 acquired on an BD FACSymphony flow cytometer and analyzed using FlowJo version 9.9.6 544 (Treestar, Inc., Ashland, OR). Gating schema is represented in Fig. S6 . 545

Graphs show data from individual animals with dotted lines indicating assay limits of detection. 547

Groups were compared for viral load and immune responses using Welch's t-tests; data were 548 analyzed on the log10 scale for viral loads and appropriate immune assays. Changes over time 549

were summarized using the change on the log10 scale, and statistical significance was determined 550 using paired t-tests. When more than two variants are compared, as in Figure Figure S1 . Intracellular staining (ICS) was performed on PBMCs at weeks 6, 29, 31, and 36 to assess T cell responses to SARS-CoV-2 S protein peptide pools, S1 and S2. Responses to S1 and S2 individual peptide pools 

DNA vaccine protection against SARS-CoV-2 in rhesus macaques

606 24. N. van Doremalen et al., ChAdOx1 nCoV-19 vaccine prevents SARS-CoV-2 pneumonia 607 in rhesus macaques

Human neutralizing antibodies against SARS-CoV-2 require intact 609

Fc effector functions for optimal therapeutic protection

Protective antibodies elicited by SARS-CoV-2 spike protein 612 vaccination are boosted in the lung after challenge in nonhuman primates

Ultrapotent antibodies against diverse and highly transmissible SARS-615 CoV-2 variants

Annual Review of Immunology 30

SARS-CoV-2 infection induces long-lived bone marrow plasma cells 619 in humans

Antibody Evolution after SARS-CoV-2 mRNA Vaccination. bioRxiv

Nucleoside-modified mRNA vaccines induce potent T follicular helper 623 and germinal center B cell responses

mRNA-1273 COVID-19 vaccine effectiveness against the B.1.1.7 626 and B.1.351 variants and severe COVID-19 disease in Qatar

BNT162b2 mRNA Vaccine Effectiveness Given Confirmed Exposure; 628 Analysis of Household Members of COVID-19 Patients. medRxiv

Evaluation of mRNA-1273 against SARS-CoV-2 B.1.351 Infection 631 in Nonhuman Primates. bioRxiv

Evolution of antibody immunity to SARS-CoV-2

Naturally enhanced neutralizing breadth against SARS-CoV-2 one year 635 after infection

An mRNA Vaccine against SARS-CoV-2 -Preliminary Report

CoV-2 mRNA vaccines confer broad neutralization as primary 641 or booster series in mice. bioRxiv

Cryo-EM structure of the 2019-nCoV spike in the prefusion 643 conformation

Immunogenicity and structures of a rationally designed prefusion 645 MERS-CoV spike antigen

Optimization of Lipid Nanoparticles for Intramuscular Administration 648 of mRNA Vaccines

Generation of VSV pseudotypes using recombinant ΔG-VSV for studies on 650 virus entry, identification of entry inhibitors, and immune responses to vaccines

OMIP-052: An 18-Color Panel for Measuring 653

Rhesus Macaques

Farzan for the ACE2-overexpressing 293 cells. We thank B. Zhang for production and 662 purification of published mAbs

DiPiazza for contributions to probe production and validation. We thank M

Whitt for kind support on recombinant VSV-based SARS-CoV-665 2 pseudovirus production. Funding: Intramural Research Program of the VRC

Office of the Assistant Secretary for Preparedness 667 and Response, Biomedical Advanced Research and Development Authority

Education

Emory Executive Vice President for Health 670

Infections and Vaccines and Children's Healthcare of

Sciences Center 2020 COVID-19 CURE Award

prepared figures and tables. All authors contributed 681 to discussions about and editing of the manuscript. Competing Interests: K.S.C. and B.S.G. are 682 inventors on U

/081318 entitled "Prefusion Coronavirus Spike Proteins and Their Use

are inventors on US Patent Application No. 62/972,886 entitled "2019-nCoV Vaccine

serves on the scientific board of advisors for Moderna. Data and materials availability: 691 All data are available in the main text or the supplementary materials

We thank J. Stein and M. Young for technology transfer and 658 administrative support, respectively. We thank members of the NIH NIAID VRC Translational