key: cord-0958778-b05q2v24
authors: Gagne, Matthew; Moliva, Juan I.; Foulds, Kathryn E.; Andrew, Shayne F.; Flynn, Barbara J.; Werner, Anne P.; Wagner, Danielle A.; Teng, I-Ting; Lin, Bob C.; Moore, Christopher; Jean-Baptiste, Nazaire; Carroll, Robin; Foster, Stephanie L.; Patel, Mit; Ellis, Madison; Edara, Venkata-Viswanadh; Maldonado, Nahara Vargas; Minai, Mahnaz; McCormick, Lauren; Honeycutt, Christopher Cole; Nagata, Bianca M.; Bock, Kevin W.; Dulan, Caitlyn N.M.; Cordon, Jamilet; Flebbe, Dillon R.; Todd, John-Paul M.; McCarthy, Elizabeth; Pessaint, Laurent; Van Ry, Alex; Narvaez, Brandon; Valentin, Daniel; Cook, Anthony; Dodson, Alan; Steingrebe, Katelyn; Nurmukhambetova, Saule T.; Godbole, Sucheta; Henry, Amy R.; Laboune, Farida; Roberts-Torres, Jesmine; Lorang, Cynthia G.; Amin, Shivani; Trost, Jessica; Naisan, Mursal; Basappa, Manjula; Willis, Jacquelyn; Wang, Lingshu; Shi, Wei; Doria-Rose, Nicole A.; Zhang, Yi; Yang, Eun Sung; Leung, Kwanyee; O’Dell, Sijy; Schmidt, Stephen D.; Olia, Adam S.; Liu, Cuiping; Harris, Darcy R.; Chuang, Gwo-Yu; Stewart-Jones, Guillaume; Renzi, Isabella; Lai, Yen-Ting; Malinowski, Agata; Wu, Kai; Mascola, John R.; Carfi, Andrea; Kwong, Peter D.; Edwards, Darin K.; Lewis, Mark G.; Andersen, Hanne; Corbett, Kizzmekia S.; Nason, Martha C.; McDermott, Adrian B.; Suthar, Mehul S.; Moore, Ian N.; Roederer, Mario; Sullivan, Nancy J.; Douek, Daniel C.; Seder, Robert A.
title: mRNA-1273 or mRNA-Omicron boost in vaccinated macaques elicits similar B cell expansion, neutralizing antibodies and protection against Omicron
date: 2022-03-25
journal: Cell
DOI: 10.1016/j.cell.2022.03.038
sha: 69951c20b5bbc35f945a509168a6625552453623
doc_id: 958778
cord_uid: b05q2v24

SARS-CoV-2 Omicron is highly transmissible and has substantial resistance to neutralization following immunization with ancestral spike-matched vaccines. It is unclear whether boosting with Omicron-matched vaccines would enhance protection. Here, nonhuman primates that received mRNA-1273 at weeks 0 and 4 were boosted at week 41 with mRNA-1273 or mRNA-Omicron. Neutralizing titers against D614G were 4760 and 270 reciprocal ID50 at week 6 (peak) and week 41 (pre-boost), respectively, and 320 and 110 for Omicron. Two weeks after boost, titers against D614G and Omicron increased to 5360 and 2980 for mRNA-1273 boost and 2670 and 1930 for mRNA-Omicron. Similar increases against BA.2 were observed. Following either boost, 70-80% of spike-specific B cells were cross-reactive against WA1 and Omicron. Equivalent control of virus replication in lower airways was observed following Omicron challenge one month after either boost. These data show that mRNA-1273 and mRNA-Omicron elicit comparable immunity and protection shortly after the boost.

The COVID-19 mRNA vaccines BNT162b2 and mRNA-1273 provide highly effective 19 protection against symptomatic and severe infection with ancestral SARS-CoV-2 (Baden et al., 20 2021b; Dagan et al., 2021; Pilishvili et al., 2021; Polack et al., 2020) . More recently, protective 21 efficacy has declined due to both waning vaccine-elicited immunity (Baden et al., 2021a; 22 Bergwerk et al., 2021; Goldberg et al., 2021) and antigenic shifts in variants of concern (VOC) 23 including B.1.351 (Beta) and B.1.617.2 (Delta) (Planas et al., 2021; Wang et al., 2021a; Wang et 24 al., 2021b) . Importantly, the introduction of a boost after the initial vaccine regimen enhances 25 immunity and vaccine efficacy against symptomatic disease, hospitalization and death across a 26 broad range of age groups (Andrews et al., 2022; Bar-On et al., 2021; Barda et al., 2021; Garcia-27 Beltran et al., 2022; Pajon et al., 2022) . However, the timing and selection of a boost is a major 28 scientific and clinical challenge during this evolving pandemic in which emerging VOC have 29 distinctive patterns of transmission and virulence and against which vaccine-elicited antibody 30 neutralization is reduced. 31

The most recent VOC, B.1.1.529, henceforth referred to by its WHO designation of Omicron, 33 was first identified in South Africa in November 2021 and was associated with a dramatic 34 increase in COVID-19 cases (Cele et al., 2022; Maslo et al., 2022) . Omicron is highly 35 contagious, with a significant transmission advantage compared to Delta, which until recently 36 was the dominant VOC worldwide (Viana et al., 2022) . It remains unclear, however, if this 37 advantage is due to differences in cell entry, enrichment in respiratory aerosols, or the ability to few months after immunization has been estimated as 44% in California, USA and 37% in 49

Denmark (Hansen et al., 2021; Tseng et al., 2022) and a complete loss of protection within six 50 months (Accorsi et al., 2022) . Multiple reports have suggested that Omicron has reduced 51 virulence compared to prior VOC in humans, mice and hamsters (Davies et al., 2022; Halfmann 52 et al., 2022; Suryawanshi et al., 2022) . It is possible that reduced virulence of Omicron may 53 result from preferential replication in the upper airway compared to the lungs, perhaps due to 54 altered cellular tropism not reliant on expression of transmembrane serine protease 2 55 (TMPRSS2) (Meng et al., 2022; Willett et al., 2022) . However, the effect of any reduction in 56 intrinsic viral pathogenicity may be somewhat offset in the context of reduced vaccine efficacy 57 and enhanced virus transmission in human populations worldwide. Together these data reinforce 58 the value of boosting to limit the extent of infection from Omicron. 59 60 Variant-matched boosts have been suggested as a strategy to enhance neutralizing and binding 61 antibody titers to the corresponding VOC beyond the levels conferred by existing FDA-approved 62 boosts, which are homologous to the original ancestral WA1-matched primary vaccine regimen. 63 J o u r n a l P r e -p r o o f following either a homologous or heterologous boost (Fig. 1E) , and titers to Omicron were still 110 lower than all other variants. 111 112 Neutralizing antibody titers were then assessed using a live virus assay ( Fig. 1C and Table S2 ). 113

At week 6, neutralizing titers were highest to D614G followed by Delta, then Beta and Omicron. 114 Titers to all variants markedly declined by week 41, including a drop in reciprocal 50% 115 inhibitory dilution (ID50) titers for D614G from 5560 at week 6 to 330 at week 41 and for 116

Omicron from 110 at week 6 to 33 at week 41. However, following either boost, neutralizing 117 titers to D614G and Delta were increased similar to week 6 and titers to Beta and Omicron were 118 greater than they had been at week 6 (Beta: P=0.05 and 0.035; Omicron: P=0.041 and 0.01 for 119 mRNA-1273 and mRNA-Omicron, respectively) (Fig. 1E ). We substantiated these findings 120 using a lentiviral pseudovirus neutralization assay similar to the one used to assess immune The increase in neutralizing titers to all VOC tested after the third dose could suggest continued 130 other variants. Consistent with the findings in the BAL, either boost increased nasal antibody 155 titers ~6-7 logs, with GMT of ~10 12 for WA1, Delta and Beta and ~10 10 for Omicron. 156 157 In a number of prior NHP studies, we have not been able to detect antibody neutralizing titers 158 using pseudo-or live-virus assays from NW or BAL. However, based on its high sensitivity, we 159 have used the angiotensin converting enzyme 2 receptor (ACE2) inhibition assay to measure 160 antibody function as a surrogate for neutralization capacity (Corbett et al., 2021a; Gagne et al., 161 2022) . While the antigen for determination of binding titers was wildtype (WT) S, our ACE2 162 inhibition assay used stabilized S-2P (Table S3 ). In the BAL, 25-50% median binding inhibition 163 was observed for all variants at week 8, except for Omicron S-2P in which binding inhibition 164 was low to undetectable (Fig. 2C ). ACE2 binding inhibition declined to a median of <15% for all 165 variants by week 39. Following a boost with either mRNA-1273 or mRNA-Omicron, we 166 observed greater ACE2 inhibition, although this increase did not reach significance due to the 167 small number of animals in each group. Of note, although ACE2 inhibition of Omicron S-2P 168 increased following the boost, it remained lower than all other variants. In the upper airway, 169 ACE2 inhibition was low to undetectable at week 39 following the initial vaccine regimen for all 170 variants. However, after either boost, there was an increase across all variants including Omicron 171 to values higher than the initial peak at week 8 (Fig. 2D) . Thus, boosting with either vaccine was 172 important for enhancing mucosal antibody binding and neutralization responses. 173 174 Similar expansion of cross-reactive S-2P-specific memory B cells following boosting 175

The observation of rapid and significant increases in binding and neutralizing antibody titers to 176

Omicron in both blood and mucosal sites after homologous or heterologous mRNA boost 177 J o u r n a l P r e -p r o o f suggests an anamnestic response involving the mobilization of cross-reactive memory B cells. 178

Thus, we measured B cell binding to pairs of fluorochrome-labeled S-2P probes representing 179 different VOC including Omicron at weeks 6, 41 and 43 (2 weeks post-boost) (Fig. S2) . 180

Of the total S-2P specific memory B cell responses at week 6, 63% were dual-specific and 182 capable of binding both WA1 and Omicron probes, with 33% binding WA1 alone and only 4% 183 which bound Omicron alone (Fig. 3A, 4A ). By week 41, the total S-specific memory B cell 184 compartment in the blood had declined ~90% as a fraction of all class-switched memory B cells 185 Two weeks after boosting, there was an expansion of the total S-specific memory B cell 189 compartment similar to that observed at week 6. Following an mRNA-1273 boost, 24% of all S-190 2P-specific memory B cells were specific for WA1 alone and 71% were dual-specific for WA1 191 and Omicron. After the mRNA-Omicron boost, 81% were dual-specific for WA1 and Omicron., 192 with 12% specific for WA1 only (Fig. 4A ). Of note, we did not observe a population of 193

Omicron-only memory B cells before or after the boost that was clearly distinct from background 194 staining (Fig. 3A) . These data suggest a marked expansion of cross-reactive dual-specific and Omicron-positive B cells for either boost, with mRNA-1273 also expanding WA1-only B 196 cell responses. The increase in cross-reactive B cells for WA1 and Omicron is consistent with the 197 comparable and high-level of neutralizing titers against D614G and Omicron by either boost 198 ( Fig. 1C-D) . To extend these data, serologic mapping of antigenic sites on Omicron and WA1 199 RBD was performed. This analysis revealed that boosting with either mRNA-1273 or mRNA-200 J o u r n a l P r e -p r o o f Omicron elicited serum antibody reactivity with similar RBD specificities (Fig. S4) . Further, 201 both boosts resulted in comparable antibody binding to the antigenic site defined by S309 (parent 202 antibody of sotrovimab) which retains neutralizing capacity against Omicron (VanBlargan et al., 203 2022) . 204

To further explore the effect of boosting on anamnestic B cell responses, we phenotyped the 206 activation status of S-binding memory B cells (Fig. 4E ). WA1 S-2P-and/or Omicron S-2P-207 binding memory B cells predominantly had an activated memory phenotype immediately after 208 both the second and third doses. 209 210 Next, we determined the extent of cross-reactivity of B cells for two other VOC: Delta, which 211 has recently co-circulated, and Beta, which shows significant neutralization resistance. Six weeks 212 after vaccination, 68% of all Delta S-2P and/or Omicron S-2P memory B cells were dual-specific 213 and the remainder of S-binding memory B cells largely bound Delta alone (Fig. 3B, 4B ). 214

Following a third dose, the frequency of dual-specific cells increased to 76% for mRNA -1273 215 and 85% for mRNA-Omicron, consistent with our findings on cross-reactive B cells using WA1 216 and Omicron S-2P probes. 217

We have previously reported that dual-specific WA1 S-2P and Delta S-2P memory B cells 219 accounted for greater than 85% of all memory B cells which bound either spike after two 220 immunizations with mRNA-1273 (Gagne et al., 2022) . Here we confirmed and extended these 221 findings and show that after either boost, ~95% of all WA1-and/or Delta-binding memory B 222 cells are dual-specific (Fig. 3C, 4C ). Similar findings were obtained with WA1 and Beta S-2P 223 probes, in which the dual-specific population was 85% at week 6 and 90% following either boost 224 ( Fig. 3D, 4D ). Of note following the mRNA-Omicron boost, very few B cells were detected that 225 only bound WA1 epitopes when co-staining for Delta or Beta. Overall, the data show that cross-226 reactive cells were expanded following a boost with either mRNA-1273 or mRNA-Omicron 227 while only mRNA-1273 was capable of boosting memory B cells specific for WA1 alone (Fig.  228   S3) . 229 230 Primary responses following variant-matched immunization 231

In addition to understanding how mRNA-1273 and mRNA-Omicron influenced immunity as a 232

boost, it was also important to assess responses elicited by variant-matched vaccines used in a 233 primary regimen. Thus, we immunized naïve NHP with two doses of 100µg mRNA-Omicron at 234 weeks 0 and 4. While a single dose of mRNA-Omicron in mRNA-1273-immunized NHP had 235 failed to elicit a new population of memory B cells with unique specificities for Omicron S-2P 236 ( Fig. 3A, 4A ), primary immunization with mRNA-Omicron, in contrast, induced both 237 WA1/Omicron cross-reactive memory B cells as well as B cells that bound only Omicron but not 238 WA1 S-2P. Such responses were observed at two weeks after the prime and two weeks after the 239 second dose of mRNA-Omicron although the kinetics of the response varied among animals 240 (Fig. S5A ). Two weeks after either the first or second dose of mRNA-Omicron in naïve animals, 241 the magnitude of this response as measured by the frequency of total WA1-Omicron+ B cells 242 was greater at a similar time after immunization than in the vaccinated NHP who had received 243 mRNA-1273 prime and boost followed by the heterologous mRNA-Omicron boost (Fig. S5B) . 244

These data highlight the differences between naïve and vaccine-primed animals in generating a 245 population of memory B cells specific only for Omicron. 246

Further, recent studies in naïve mice show that while vaccination with mRNA encoding WT or 248 Delta S elicits neutralization that is cross-reactive to all the variants, vaccination with Omicron 249 mRNA induces neutralization to Omicron only (Lee et al., 2022) . Thus, we compared two-dose 250 prime/boost regimens of mRNA-1273, mRNA-Omicron and mRNA-Beta in naïve NHP (Fig.  251   S6 ). While vaccination with mRNA-1273 or mRNA-Beta elicited neutralizing responses to all 252 variants tested, mRNA-Omicron elicited responses predominantly biased towards Omicron, with 253 lower titers to the other variants. These data reveal a striking difference between using mRNA-254

Omicron as a boost to broadly enhance prior cross-reactive immunity compared to its use in a 255 primary immunizing regimen. 256 257 S-2P-specific T cell responses in blood and BAL following vaccination 258

We have previously shown that mRNA-1273 immunization elicits TH1, TFH and a low frequency 259 of CD8 responses to S peptides in NHP and humans (Corbett et al., 2020; Corbett et al., 2021a; 260 Corbett et al., 2021c; Gagne et al., 2022; Jackson et al., 2020) . Consistent with the prior studies, 261

we show that mRNA-1273 elicits TH1, TFH and low-level CD8 T cell responses to WA1 S 262 peptides at the peak of the response (week 6) that decline over time ( Fig. S7 and S8 ). Boosting 263 with either mRNA-1273 or mRNA-Omicron increased TFH responses which could be important 264 for expanding the S-specific memory B cell population following the boost (Johnston et al., 265 2009; Nurieva et al., 2009; Tangye et al., 2002) . T cell epitopes within Omicron S have been 266

shown to be largely conserved (Choi et al., 2022) ; indeed, responses to Omicron-specific 267 peptides after boosting were similar to those of WA1-specific peptides (Fig. S9) . Finally, we also 268 detected TH1 and CD8 T cells in BAL at week 8 that decreased to undetectable levels at week 39. 269

Such responses were increased with either mRNA-1273 or mRNA-Omicron (Fig. S8) . replication. As sgRNA encoding for the N gene (sgRNA_N) are the most abundant transcripts 286 produced due to the viral discontinuous transcription process (Kim et al., 2020) , the sgRNA_N 287 qRT-PCR assay was chosen for its enhanced sensitivity. On day 2 post-infection in the BAL, 288 unvaccinated NHP had geometric mean copy numbers of 1x10 6 sgRNA_N per mL whereas the 289 vaccinated NHP had 3x10 2 and 2x10 2 for the mRNA-1273 and mRNA-Omicron cohorts, 290 respectively (Fig. 5A ). By day 4, all vaccinated NHP had undetectable levels of sgRNA_N while 291 copy numbers in the unvaccinated group had only declined to 3x10 5 per mL (either boost vs 292 control on days 2 and 4: P<0.0001). 293

In the nose, sgRNA_N copy numbers at days 1 to 4 were low for most animals and were not 295 different between the control and vaccinated cohorts, so protection following vaccination could 296 not be determined (Fig. 5B ). At day 4, 5/8 controls had detectable virus in the nose as compared 297

to 3/8 vaccinated NHP, with no clear difference between the boost cohorts. However, by day 8, 298 4/8 controls still had detectable sgRNA_N including 2 animals with increased copy numbers 299 while none of the vaccinated NHP had detectable sgRNA. 300

In assessing sgRNA_N in the throat, it is noteworthy that 2 days after challenge, only 1/8 302 vaccinated NHP (in either boost group) had detectable virus in the throat compared to 6/8 control 303 NHP (Fig. 5C) . 304

We also measured the amount of culturable virus using a tissue culture infectious dose assay 306 (TCID50). No virus was detected in the BAL of any vaccinated NHP, while 8/8 and 7/8 control 307 NHP had detectable virus 2 and 4 days after challenge, respectively (either boost vs control on 308 day 2: P<0.0001; day 4: P=0.0005) (Fig. 5D ). In the NS, 1/8 boosted animals had culturable 309 virus at any timepoint. In the unvaccinated control animals, 2/8 and 3/8 NHP had culturable virus 310 in the nose 2 and 4 days after challenge, respectively (Fig. 5E) . To assess lung pathology in NHP, 2 of the animals in each group were euthanized on day 8 314 following Omicron challenge, and the amount of virus antigen (SARS-CoV-2 N) and 315 inflammation in the lungs were assessed (Fig. 6) . N antigen was detected in variable amounts in 316 the lungs of both control animals. When present, virus antigen was often associated with the 317 alveolar capillaries and, occasionally, nearby immune cells. There was no evidence of virus 318 antigen in the lungs of the vaccinated NHP. 319 320 Animals from both boost groups displayed histopathologic alterations that were classified as 321 minimal to mild or moderate. Inflammation was largely characterized by mild and patchy 322 expansion of alveolar capillaries, generalized alveolar capillary hypercellularity, mild and 323 regional type II pneumocyte hyperplasia and, less frequently, scattered collections of immune 324 cells within some alveolar spaces. In contrast, unvaccinated animals were characterized as 325 having a moderate to severe pathology. Lung sections from controls included features 326 characterized by moderate and often diffuse alveolar capillary expansion, diffuse 327 hypercellularity, moderate type II pneumocyte hyperplasia and multiple areas of perivascular 328 cellular infiltration. Together, these data indicate that protection against Omicron was robust in 329 the lungs regardless of boost selection. 330 331

Omicron has become the dominant global variant of SARS-CoV-2 due to its transmission 333 advantage relative to Delta and its ability to evade prior immunity (Grabowski et al., 2022; Viana 334 et al., 2022) . Vaccine efficacy against infection with Omicron has declined and boosting with a 335 third dose of an mRNA COVID-19 vaccine matched to the prototype strain has been shown to 336 restore immunity and protection (Accorsi et al., 2022; Garcia-Beltran et al., 2022; Hansen et al., 337 2021; Pajon et al., 2022; Tseng et al., 2022) . Here, we immunized NHP with 2 doses of mRNA-338 1273 (100g) and boosted them ~9 months later with 50g of either mRNA-1273 or mRNA-339

Omicron prior to challenge with Omicron virus. The principal findings were: (1) 9 months after 340 the two-dose regimen, neutralizing and binding antibody titers to Omicron had declined 341 substantially in blood and mucosal airways; (2) after the boost, neutralizing antibody titers to 342 ancestral strains were restored and those to Omicron were increased compared to the peak 343 response after the initial prime and boost; (3) both boosts expanded cross-reactive memory B 344 cells but only the homologous boost was capable of expanding B cells specific for epitopes 345 unique to the ancestral strain; and (4) both boosts provided complete protection in the lungs and 346 limited protection in the upper airway after Omicron challenge. 347

Following either mRNA-1273 or mRNA-Omicron boost, there was essentially complete 349 protection in the lower airway with no culturable virus by day 2 and no detectable sgRNA_N by 350 day 4. These data are comparable to our previous findings of equivalent upper and lower airway 351 protection following Beta challenge 2 months after boosting mRNA-1273-immunized NHP with 352 either mRNA-1273 or mRNA-Beta (Corbett et al., 2021a) . In contrast to the lower airway, there 353 were no clear and consistent differences in sgRNA_N copy number at days 2 or 4 in the upper 354 airway of vaccinated or control NHP. Of note, more of the control animals had detectable 355 sgRNA at day 4 and increased sgRNA at day 8 as compared to the boosted animals. We would 356 also note that the amount of Omicron replication as assessed by sgRNA or culturable virus in the 357 control animals is demonstrably different than in our prior studies in which NHP were 358 challenged with WA1, Delta or Beta (Corbett et al., 2020; Corbett et al., 2021c; Gagne et al., 359 2022) . These findings are consistent with evidence for reduced overall severity of Omicron 360 infection in animal models of COVID-19 compared to other variants (Bentley et al., 2021; 361 Halfmann et al., 2022; Suryawanshi et al., 2022) . Overall, the findings here of high-level 362 protection in the lungs recapitulate observations in vaccinated humans of reduced disease 363 severity following infection with Omicron (Abdullah et al., 2021; Sigal, 2022; Wolter et al., 364 2022) . Our data also complement recent findings on protection from Omicron infection in 365 hamsters immunized with replicating RNA matched to the ancestral strain or Omicron S 366 (Hawman et al., 2022) . 367

Neutralizing antibodies to Omicron in the blood or ACE2 binding inhibitory antibodies in the 369 airway mucosa were low after the first 2 doses of mRNA-1273 at weeks 6-8 and low to 370 undetectable ~9 months later. Importantly, either mRNA-1273 or mRNA-Omicron boosts were 371 able to significantly increase neutralizing antibody titers against Omicron and Beta beyond their 372 initial peak consistent with a rapid recall B cell response. This also implies that neutralizing 373 antibody titers at extended times after vaccination may not be a reliable surrogate either for 374 vaccine efficacy in the lower airway or for predicting responses following a boost or infection as 375 they may not reflect the recall capacity of the underlying memory B cell population. 376

The observation that boosting with either mRNA-1273 or mRNA-Omicron resulted in the 378 expansion of a similarly high frequency of cross-reactive B cells likely stems from the recall of 379 prior immune memory after a related antigenic encounter. This principle has been termed 380 original antigenic sin, imprinting, and back boosting (Fonville et al., 2014; Francis, 1960; Kohler 381 et al., 1994) . Recall of prior immunity may be deleterious or beneficial as exemplified by the 382 impact of the circulating influenza A subtypes at the time of an individual's first exposure after 383 birth on patterns of disease susceptibility to subsequent pandemic influenza A outbreaks (Gostic 384 et al., 2016; Worobey et al., 2014) . The current worldwide distribution and evolution of SARS-385

CoV-2, however, is quite different from that of influenza A. Whereas multiple subtypes of 386 influenza A circulate with different levels of co-dominance, SARS-CoV-2 distribution has 387 generally become rapidly dominated by a single variantcurrently Omicronbefore 388 replacement by another that, for various reasons, is more transmissible. The question therefore is 389 whether there is added value from boosting with a heterologous vaccine matched to the dominant 390 circulating variant, or whether cross-reactive B cell recall immunity elicited by boosting with the 391 original vaccine is sufficient to reduce infection and disease severity. As we have now shown in 392 two different NHP studies, boosting animals with either mRNA-Beta (Corbett et al., 2021a) or 393 mRNA-Omicron has not yet been shown to provide any significant advantage over mRNA -1273 394 in recalling high titer neutralizing antibodies across all variants tested in the short-term and 395

protecting from virus replication after challenge. These considerations may apply to the large 396 numbers of individuals with prior immunity from vaccination or infection with current and 397 previous variants. Importantly, the conclusions related to post-boost immunity are limited to the 398 short duration of immune assessment in our studies. It is conceivable that mobilization of pre-399 existing memory B cells may dominate the initial immune response to a booster dose and that 400 further longitudinal analysis could reveal B cell populations with new specificities to the 401 matched variant boost (Sokal et al., 2021) . 402

Looking to the future, however, if Omicron, or a closely antigenically related variant, remains 404 the dominant circulating variant for some years to come, then it is possible that a change in the 405 initial vaccine regimen would be warranted, particularly in immunologically naïve populations 406 such as children as they reach the age of eligibility for approved Importantly, it would need to be established that a switch in COVID-19 vaccine design to match 408 the current dominant variant would not limit responses against variants which may be 409 antigenically distant from Omicron but close to the prototype. In fact, we show that in naïve 410 animals, mRNA-Omicron as an initial prime and boost regimen skewed neutralizing responses 411 predominantly towards Omicron with more limited neutralization against past variants, 412 consistent with recent data from primary Omicron infection in humans (Richardson et al., 2022; 413 Rössler et al., 2022) . Thus, a combination or bivalent vaccine to generate B cells specific for the 414 current variant as well as cross-reactive to other variants might ensure greater breadth of 415 neutralization in naïve hosts (Lee et al., 2022) . 416

In summary, our findings highlight two important factors that will impact management of this 418 pandemic. The first is the design of the vaccine and whether it should be changed based on the 419 currently circulating variant. At present, boosting previously vaccinated NHP and humans with 420 mRNA-1273 provides significant increases in neutralizing antibodies and is sufficient to prevent 421 severe disease after exposure from all known variants (Accorsi et al., 2022; Andrews et al., 2022; 422 Choi et al., 2021; Corbett et al., 2021a; Garcia-Beltran et al., 2022; Pajon et al., 2022; greater neutralization towards Omicron. In addition, we did not assess a bivalent boost of 448 mRNA-1273 and mRNA-Omicron which might elicit higher neutralizing responses than either 449 alone. Finally, since we sought to compare two different mRNA boosts, we did not have an 450 unboosted group to determine whether the boost enhanced protection. As all the boosted NHP 451 were completely protected in the lungs, we were unable to determine an immune threshold for 452 protection. 453

We would like to thank G. Alvarado for experimental organization and administrative support. Table S1 for mRNA-Omicron sequence and Table S2 for were gated as singlets and live cells on forward and side scatter and a live/dead aqua blue stain. 653 CD3 + events were gated as CD4 + or CD8 + T cells. Total memory CD8 + T cells were selected 654 based on expression of CCR7 and CD45RA. Finally, SARS-CoV-2 S-specific memory CD8 + T 655 cells were gated according to co-expression of CD69 and IL-2, TNF or IFNγ. The CD4 + events 656 were defined as naïve, total memory or central memory according to expression of CCR7 and 657 CD45RA. CD4 + cells with a TH1 phenotype were defined as memory cells that co-expressed 658 CD69 and IL-2, TNF or IFNγ. CD4 + cells with a TH2 phenotype were defined as memory cells 659 that co-expressed CD69 and IL-4 or IL-13. TFH cells were defined as central memory CD4 + T 660 cells that expressed CXCR5, ICOS and PD-1. TFH cells were further characterized as IL-21 + , 661 CD69 + or CD40L + , CD69 + . 662 J o u r n a l P r e -p r o o f 

This study did not generate new unique reagents. 705 706 Data and code availability 707

• All data reported in this paper will be shared by the lead contact upon request. 708

• This paper does not report original code. 709

• Any additional information required to reanalyze the data reported in this paper is 710 available from the lead contact upon request. 711 712 Experimental model and subject details 713

A sequence-optimized mRNA encoding prefusion-stabilized SARS-CoV-2 S protein containing 715 2 proline stabilization mutations (S-2P) (Pallesen et al., 2017; Wrapp et al., 2020) for WA1, 716

Omicron and Beta were synthesized in vitro and formulated (Hassett et al., 2019) . Control 717 mRNA "UNFIX-01 (Untranslated Factor 9)" was synthesized and similarly formulated into lipid 718 nanoparticles as previously described (Corbett et al., 2021a) . nanoparticles diluted in phosphate-buffered saline (PBS) into the right quadricep as previously 728 described (Corbett et al., 2020; Corbett et al., 2021c; Gagne et al., 2022) . At week 41 (~9 months 729 after the second immunization), the eight macaques were split into two groups of 4 and boosted 730 with 50μg mRNA-1273 or 50μg mRNA-Omicron. Animals in the control group were immunized 731 with 50μg control mRNA at the time of the boost. Isolation and sequencing of EHC-083E (D614G SARS-CoV-2), Delta, Beta and Omicron for 743 live virus neutralization assays were previously described (Edara et al., 2022; Edara et al., 2021a; 744 Edara et al., 2021b; Vanderheiden et al., 2020) . Viruses were propagated in Vero-TMPRSS2 745 cells to generate viral stocks. Viral titers were determined by focus-forming assay on VeroE6-746 TMPRSS2 cells. Viral stocks were stored at -80C until use. 747 748

NEBNext Ultra II RNA Prep reagents and multiplex oligos (New England Biolabs) were used to 750

prepare Illumina-ready libraries, which were sequenced on a MiSeq (Illumina) as described 751

previously (Corbett et al., 2021c; Gagne et al., 2022) . Demultiplexed sequence reads were 752 analyzed in the CLC Genomics Workbench v.21.0.3 by (1) trimming for quality, length, and 753 adaptor sequence, (2) mapping to the Wuhan-Hu-1 SARS-CoV-2 reference (GenBank no. 

While S antigens were used for binding ELISAs, S-2P antigens were used for ACE2 inhibition 777 assays and B cell probe binding. S-2P constructs were made as follows. Biotinylated S probes 778 were expressed transiently for WA1, D614G, Delta, Beta and Omicron strains and purified and 779 biotinylated in a single in-process step (Teng et al., 2021; Zhou et al., 2020) . S-2P for WA1 and 780

Omicron were made as previously described (Olia et al., 2021) . FRNT assays were performed as previously described (Edara et al., 2021a; Edara et al., 2021b; 797 Vanderheiden et al., 2020) . Briefly, samples were diluted at 3-fold in 8 serial dilutions using 798 DMEM (VWR, #45000-304) in duplicates with an initial dilution of 1:10 in a total volume of 799 60l. Serially diluted samples were incubated with an equal volume of WA1, Delta, Beta or 800

Omicron (100-200 foci per well based on the target cell) at 37°C for 45 minutes in a round-801 bottomed 96-well culture plate. The antibody-virus mixture was then added to VeroE6-802 TMPRSS2 cells and incubated at 37°C for 1 hour. Post-incubation, the antibody-virus mixture 803 was removed and 100µl of pre-warmed 0.85% methylcellulose overlay was added to each well. 804

Plates were incubated at 37°C for 18 hours and the methylcellulose overlay was removed and 805 washed six times with PBS. Cells were fixed with 2% paraformaldehyde in PBS for 30 minutes. 806

Following fixation, plates were washed twice with PBS and permeabilization buffer (0.1% BSA, 807 0.1% Saponin in PBS) was added to cells for at least 20 minutes. Cells were incubated with an 808 anti-SARS-CoV S primary antibody directly conjugated to Alexaflour-647 (CR3022-AF647) 809 overnight at 4°C. Foci were visualized and imaged on an ELISPOT reader (CTL). Antibody 810 neutralization was quantified by counting the number of foci for each sample using the Viridot 811 program (Katzelnick et al., 2018) . The neutralization titers were calculated as follows: 1 -(ratio 812 of the mean number of foci in the presence of sera and foci at the highest dilution of respective 813 sera sample). Each specimen was tested in duplicate. The FRNT-50 titers were interpolated using 814 a 4-parameter nonlinear regression in GraphPad Prism 9.2.0. Samples that do not neutralize at 815 the limit of detection (LOD) at 50% are plotted at 20 and was used for geometric mean and fold-816 change calculations. The assay LOD was 20. Delta, Beta and Omicron, the plasmid was altered via site-directed mutagenesis to match the S 831 sequence to the corresponding variant sequence as previously described (Corbett et al., 2021a) . A Avidity was measured as described previously (Francica et al., 2021) in an adapted ELISA assay. 856

Briefly, ELISA against S-2P was performed in the absence or presence of sodium thiocyanate 857 (NaSCN) and developed with HRP-conjugated goat anti-monkey IgG (H+L) secondary antibody 858 (Invitrogen) and SureBlue 3,3′,5,5′-tetramethylbenzidine (TMB) microwell peroxidase substrate 859

(1-Component; SeraCare) and quenched with 1N H2SO4. The avidity index (AI) was calculated 860 as the ratio of IgG binding to S-2P in the absence or presence of NaSCN. 861 862

Serum epitope mapping competition assays were performed, as previously described (Corbett et 864 al., 2021a) , using the Biacore 8K+ surface plasmon resonance system (Cytiva). Briefly, through 865 primary amine coupling using a His capture kit (Cytiva), anti-histidine antibody was 866 Intracellular cytokine staining was performed as previously described (Donaldson et al., 2019; 903 Gagne et al., 2022) . Briefly, cryopreserved PBMC and BAL cells were thawed and rested 904 overnight in a 37°C/5% CO2 incubator. The following morning, cells were stimulated with 905 SARS-CoV-2 S protein peptide pools (S1 and S2, matched to vaccine insert or Omicron variant; 906

JPT Peptides) at a final concentration of 2 μg/ml in the presence of 3mM monensin for 6 hours. sgRNA was isolated and quantified by researchers blinded to vaccine status as previously 926 described (Corbett et al., 2021c) , except for the use of a new probe noted below. Briefly, total 927 RNA was extracted from BAL fluid and nasal swabs using RNAzol BD column kit (Molecular 928

Research Center). PCR reactions were conducted with TaqMan TCID50 was quantified as previously described (Corbett et al., 2021c) . Briefly, Vero-TMPRSS2 943 cells were plated and incubated overnight. The following day, BAL or NS samples were serially 944 diluted, and the plates were incubated at 37 °C/5.0% CO2 for four days. Positive (virus stock of 945 known infectious titer in the assay) and negative (medium only) control wells were included in 946 each assay setup. The cell monolayers were visually inspected for cytopathic effect. TCID50 947 values were calculated using the Reed-Muench formula. 948

Routine histopathology and detection of SARS-CoV-2 virus antigen via immunohistochemistry 951 (IHC) were performed as previously described (Corbett et al., 2020; Gagne et al., 2022) . Briefly, 952 8 days following Omicron challenge, animals were euthanized and lung tissue was processed and 953 stained with hematoxylin and eosin for pathological analysis or with a rabbit polyclonal anti-954 SARS-CoV-2 anti-nucleocapsid antibody (GeneTex, GTX135357) at a dilution of 1:2000 for 955 IHC. Tissue sections were analyzed by a blinded board-certified veterinary pathologist using an 956

Olympus BX43 light microscope. Photomicrographs were taken on an Olympus DP27 camera. Week:

Week:

Week:

Week: Week:

Week:

Week:

Week A   NTD  RBD RBM  FP  HR1   HR2   A67V  69/70  G142D  G339D  S371L   S373P   N440K   G446S   S375F   K417N   S477N  E484A  G496S  N501Y  T547K  D614G  D796Y  D764K  N856K  Q954H  N969K  L981F  H655Y N679K J o u r n a l P r e -p r o o f

Decreased severity of disease during the 973 first global omicron variant covid-19 outbreak in a large hospital in tshwane, south africa

Association Between 3 Doses of mRNA COVID-19 Vaccine 977 and Symptomatic Infection Caused by the SARS-CoV-2 Omicron and Delta Variants

Effectiveness of COVID-19 booster vaccines against covid-19 related symptoms, 981 hospitalisation and death in England

Phase 3 Trial of mRNA-1273 during the Delta-Variant Surge. 984

Efficacy and Safety of the mRNA-1273 SARS-CoV-2 987 Vaccine

Protection against Covid-19 by BNT162b2 Booster across Age 990 Groups

Effectiveness of a third dose of the BNT162b2 mRNA COVID-19 vaccine for 993 preventing severe outcomes in Israel: an observational study

SARS-CoV-2 Omicron-B.1.1.529 Variant leads to 996 less severe disease than Pango B and Delta variants strains in a mouse model of severe COVID-997 19. bioRxiv

Covid-19 Breakthrough Infections in Vaccinated 1000 Health Care Workers

Omicron extensively but incompletely escapes Pfizer 1003 BNT162b2 neutralization

Safety and immunogenicity of SARS-CoV-2 variant mRNA vaccine boosters in 1006 healthy adults: an interim analysis

T cell 1008 epitopes in SARS-CoV-2 proteins are substantially conserved in the Omicron variant

Evaluation of the mRNA-1273 Vaccine against 1012 SARS-CoV-2 in Nonhuman Primates

Protection against SARS-CoV-2 beta variant in 1015 mRNA-1273 vaccine-boosted nonhuman primates

Immune correlates of protection by mRNA-1273 vaccine 1018 against SARS-CoV-2 in nonhuman primates

mRNA-1273 protects against SARS-CoV-2 beta infection in 1021 nonhuman primates

BNT162b2 mRNA Covid-19 Vaccine in a Nationwide Mass 1024 Vaccination Setting

Outcomes of laboratory-confirmed SARS-1027

CoV-2 infection in the Omicron-driven fourth wave compared with previous waves in the 1028 Western Cape Province, South Africa. medRxiv

COVID-19 vaccine mRNA-1273 elicits a protective 1031 immune profile in mice that is not associated with vaccine-enhanced disease upon SARS-CoV-2 1032 challenge

OMIP-052: An 18-Color Panel for 1034

Measuring Th1, Th2, Th17, and Tfh Responses in Rhesus Macaques

mRNA-1273 and BNT162b2 mRNA vaccines have 1037 reduced neutralizing activity against the SARS-CoV-2 omicron variant

Infection-and vaccine-induced 1040 antibody binding and neutralization of the B.1.351 SARS-CoV-2 variant

Infection and Vaccine-Induced 1044

Neutralizing-Antibody Responses to the SARS-CoV-2 B.1.617 Variants

Antibody landscapes after influenza virus infection or 1048 vaccination

Protective antibodies elicited by SARS-CoV-2 spike 1051 protein vaccination are boosted in the lung after challenge in nonhuman primates

On the Doctrine of Original Antigenic Sin

Evolution of antibody immunity to SARS-1057 CoV-2

CoV-2 Delta one year after mRNA-1273 vaccination in rhesus macaques coincides with 1061 anamnestic antibody response in the lung

mRNA-based COVID-19 vaccine 1064 boosters induce neutralizing immunity against SARS-CoV-2 Omicron variant

Immune correlates analysis of the 1068 mRNA-1273 COVID-19 vaccine efficacy clinical trial

Waning Immunity after the BNT162b2 Vaccine 1071 in Israel

Potent protection 1073 against H5N1 and H7N9 influenza via childhood hemagglutinin imprinting

The Spread of SARS-CoV-2 Variant 1076 Omicron with a Doubling Time of 2.0-3.3 Days Can Be Explained by Immune Evasion

SARS-CoV-2 Omicron virus causes attenuated 1080 disease in mice and hamsters

Vaccine effectiveness against 1083 SARS-CoV-2 infection with the Omicron or Delta variants following a two-dose or booster 1084 BNT162b2 or mRNA-1273 vaccination series: A Danish cohort study. medRxiv

Optimization of Lipid Nanoparticles for 1088 Intramuscular Administration of mRNA Vaccines

Replicating RNA platform enables rapid 1091 response to the SARS-CoV-2 Omicron variant and elicits enhanced protection in naïve hamsters 1092 compared to ancestral vaccine. bioRxiv

The Omicron variant is highly 1095 resistant against antibody-mediated neutralization: Implications for control of the COVID-19 1096 pandemic

An mRNA Vaccine 1099 against SARS-CoV-2 -Preliminary Report

Bcl6 and Blimp-1 are reciprocal and antagonistic regulators of T follicular 1102 helper cell differentiation

Viridot: An automated virus 1105 plaque (immunofocus) counter for the measurement of serological neutralizing responses with 1106 application to dengue virus

The Architecture of 1108 SARS-CoV-2 Transcriptome

Deceptive imprinting in the immune response 1110 against HIV-1

Omicron-specific mRNA vaccine induced potent neutralizing 1113 antibody against Omicron but not other SARS-CoV-2 variants. bioRxiv

Characteristics and Outcomes of Hospitalized Patients in South Africa During the COVID-19 1116

Omicron Wave Compared With Previous Waves

Altered TMPRSS2 usage by SARS-CoV-2 1119 Omicron impacts tropism and fusogenicity

Neutralization of SARS-CoV-2 Omicron by BNT162b2 1122 mRNA vaccine-elicited human sera

Efficient transfer, 1124 integration, and sustained long-term expression of the transgene in adult rat brains injected 1125 with a lentiviral vector

Bcl6 mediates the development of T follicular helper cells

SARS-CoV-2 S2P spike ages through distinct states with altered 1131 immunogenicity

SARS-CoV-2 Omicron Variant Neutralization after mRNA-1134

Booster Vaccination

Immunogenicity and structures of a rationally 1137 designed prefusion MERS-CoV spike antigen

Effectiveness of mRNA Covid-19 1140 Vaccine among U.S. Health Care Personnel

Cross-neutralization of SARS-CoV-2 by a human 1143 monoclonal SARS-CoV antibody

Reduced sensitivity of SARS-CoV-2 variant 1146 Delta to antibody neutralization

Safety and Efficacy of the BNT162b2 mRNA 1149 Covid-19 Vaccine

SARS-CoV-2 Omicron 1152 triggers cross-reactive neutralization and Fc effector functions in previously vaccinated, but not 1153 unvaccinated individuals. medRxiv

Neutralization profile of Omicron variant 1155 convalescent individuals. medRxiv

Neutralizing antibody vaccine for pandemic and pre-emergent 1158 coronaviruses

Plasma Neutralization of the SARS-CoV-2 Omicron 1161 Variant

SARS-CoV-2 variant B.1.1.7 is susceptible to neutralizing antibodies 1164 elicited by ancestral spike vaccines

Milder disease with Omicron: is it the virus or the pre-existing immunity?

Maturation and persistence of the anti-1169 SARS-CoV-2 memory B cell response

Limited cross-variant immunity after 1172 infection with the SARS-CoV-2 Omicron variant without vaccination. medRxiv

Isotype switching by 1175 human B cells is division-associated and regulated by cytokines

Molecular probes of spike ectodomain and its 1178 subdomains for SARS-CoV-2 variants, Alpha through Omicron. bioRxiv

Effectiveness of mRNA-1273 against SARS-CoV-2 Omicron and 1181 Delta variants

An infectious SARS-CoV-2 B.1.1.529 1184 Omicron virus escapes neutralization by therapeutic monoclonal antibodies

Development of a Rapid Focus 1187

Reduction Neutralization Test Assay for Measuring SARS-CoV-2 Neutralizing Antibodies

CCR2 Signaling Restricts SARS-CoV-2 Infection. 1191 mBio

Rapid epidemic expansion of the SARS-CoV-2 Omicron 1194 variant in southern Africa

Ultrapotent antibodies against diverse and highly transmissible 1197 SARS-CoV-2 variants

Analysis of SARS-CoV-2 variant mutations reveals neutralization escape mechanisms 1200 and the ability to use ACE2 receptors from additional species

The hyper-transmissible SARS-CoV-2 Omicron variant 1203 exhibits significant antigenic change, vaccine escape and a switch in cell entry mechanism

Virological assessment of hospitalized patients 1207 with COVID-2019

Early assessment of the clinical severity of the 1210 SARS-CoV-2 omicron variant in South Africa: a data linkage study

Genesis and pathogenesis of the 1918 1212 pandemic H1N1 influenza A virus

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

Structure-Based Design with Tag-Based Purification and

Biotinylation Enable Streamlined Development of SARS-CoV-2 Spike Molecular Probes

2021a) N/A Goat anti-human IgD-FITC (polyclonal) Southern Biotech

RRID:AB_2795624 PerCP-Cy5.5 mouse anti-human IgM (clone G20-127) BD

RRID:AB_10611998 DyLight 405 AffiniPure goat anti-human serum IgA, α chain specific (polyclonal) Jackson ImmunoResearch

RRID:AB_2337789 Brilliant Violet 570 anti-human CD20 antibody (clone 2H7

RRID:AB_2563805 Brilliant Violet 650 anti-human CD27 antibody

RRID:AB_2562096 Brilliant Violet 785 anti-human CD14 antibody (clone M5E2

RRID:AB_2563425 BUV496 mouse anti-human CD16 (clone 3G8) BD

RRID:AB_2870224 BUV737 mouse anti-human CD4 (clone SK3) BD

RRID:AB_2870079 CD19-APC

Alexa Fluor 700 mouse anti-human IgG (clone G18-145) BD

APC-Cy7 mouse anti-human CD3 (clone SP34-2) BD

RRID:AB_396863 Anti-human CD38 PE (clone OKT10) Caprico Biotechnologies Cat#100826 PE-Cy5 mouse anti-human CD21 (clone B-ly4) BD

RRID:AB_394028 Mouse anti-human CD185 (CXCR5) monoclonal antibody, PE-Cyanine7, eBioscience (clone MU5UBEE) ThermoFisher Scientific

RRID:AB_2573540 Mouse anti-human CD4 monoclonal antibody, PE-Cyanine5.5 (clone S3

ThermoFisher Scientific Cat#MHCD0418

RRID:AB_10376013 Brilliant Violet 570 anti-human CD8a antibody (clone RPA-T8

RRID:AB_2563213 PE-Cy5 mouse anti-human CD45RA (clone 5H9) BD

RRID:AB_394517 Brilliant Violet 650 anti-human CD197 (CCR7) antibody (clone G043H7) Biolegend Cat#353234

RRID:AB_2563867 Mouse anti-human CD185 (CXCR5) monoclonal antibody, PE, eBioscience (clone MU5UBEE) ThermoFisher Scientific Cat#12-9185-42

PE/Cyanine7 anti-human/mouse/rat CD278 (ICOS) antibody (clone C398.4A)

RRID:AB_10643411 Mouse anti-human CD69-ECD

Alexa Fluor 700 anti-human IFN-γ antibody (clone B27) Biolegend Cat#506516

RRID:AB_961351 BV750 rat anti-human IL-2 (clone MQ1-17H12) BD

RRID:AB_2739710 High parameter custom BB700 conjugate (rat anti-human IL-4) (clone MP4-25D2)

BD Biosciences Cat#624381 FITC mouse anti-human TNF (clone MAb11) BD Biosciences Cat#554512

RRID:AB_395443 BV421 rat anti-human IL-13 (clone JES10-5A2)

RRID:AB_2738290 Brilliant Violet 605 anti-human IL-17A antibody (clone BL168

Alexa Fluor 647 mouse anti-human IL-21 (clone 3A3-N2.1) BD Biosciences Cat#560493

RRID:AB_1645421 Brilliant Violet 785 anti-human CD154 antibody

RRID:AB_2572187 SARS-CoV-2 (COVID-19) nucleocapsid antibody (polyclonal

RRID:AB_2868464 Bacterial and virus strains SARS-CoV-2 B.1.1.529 (challenge stock)

SARS-CoV-2 B.1.1.529 (neutralization assay) Mehul Suthar

Biological samples Chemicals, peptides, and recombinant proteins SARS-CoV2-WT-S2P-AVI-bio Vaccine Research Center

6789: 7%%;< !

6789: 02 +,-.6?+5@20' A%;< !

6789: 7%%;< &'()* $%&'(

$49 6:;<.0=>-3?@=A "$49

5&67893BC&#7'9D&

Integrated DNA Technologies (Francica et al., 2021; Wolfel et al., 2020) (Naldini et al., 1996) N/A VRC5602: pCMV ∆R8.2 (Naldini et al., 1996)