key: cord-0867354-zpo0vthx
authors: Quandt, Jasmin; Muik, Alexander; Salisch, Nadine; Lui, Bonny Gaby; Lutz, Sebastian; Krüger, Kimberly; Wallisch, Ann-Kathrin; Adams-Quack, Petra; Bacher, Maren; Finlayson, Andrew; Ozhelvaci, Orkun; Vogler, Isabel; Grikscheit, Katharina; Hoehl, Sebastian; Goetsch, Udo; Ciesek, Sandra; Türeci, Özlem; Sahin, Ugur
title: Omicron breakthrough infection drives cross-variant neutralization and memory B cell formation
date: 2022-04-01
journal: bioRxiv
DOI: 10.1101/2022.04.01.486695
sha: 04847607377cf43a9d90e08434136eff538721b7
doc_id: 867354
cord_uid: zpo0vthx

Omicron is the evolutionarily most distinct SARS-CoV-2 variant (VOC) to date and displays multiple amino acid alterations located in neutralizing antibody sites of the spike (S) protein. We report here that Omicron breakthrough infection in BNT162b2 vaccinated individuals results in strong neutralizing activity not only against Omicron, but also broadly against previous SARS-CoV-2 VOCs and against SARS-CoV-1. We found that Omicron breakthrough infection mediates a robust B cell recall response, and primarily expands preformed memory B cells that recognize epitopes shared broadly by different variants, rather than inducing new B cells against strictly Omicron-specific epitopes. Our data suggest that, despite imprinting of the immune response by previous vaccination, the preformed B cell memory pool has sufficient plasticity for being refocused and quantitatively remodeled by exposure to heterologous S protein, thus allowing effective neutralization of variants that evade a previously established neutralizing antibody response. One Sentence Summary Breakthrough infection in individuals double- and triple-vaccinated with BNT162b2 drives cross-variant neutralization and memory B cell formation.

Containment of the current COVID-19 pandemic requires the generation of durable and sufficiently broad immunity that provides protection against circulating and future variants of SARS-CoV-2. The titer of neutralizing antibodies to SARS-CoV-2, and the binding of antibodies to the spike (S) glycoprotein and its receptor-binding domain (RBD) are considered correlates of 5 protection against infection (1, 2) . Currently available vaccines are based on the ancestral Wuhan-Hu-1 strain and induce antibodies with a neutralizing capacity that exceeds the breadth elicited by infection with the Wuhan strain, or with variants of concern (VOCs) (3) . However, protective titers wane over time (4-7) and routine booster vaccinations are thought to be needed to trigger recall immunity and maintain efficacy against new VOCs (8) (9) (10) (11) . 10 Long-lived memory B (BMEM) cells are the basis for the recall response upon antigen reencounter either by infection or booster vaccination. They play an important role in the maintenance and evolution of the antiviral antibody response against variants, since low-affinity selection mechanisms during the germinal center reaction and continued hypermutation of BMEM cells expand the breadth of viral variant recognition over time (12, 13) . 15 How vaccine-mediated protective immunity will evolve over time and will be modified by iterations of exposure to COVID-19 vaccines and infections with increasingly divergent viral variants, remains poorly understood and is of particular relevance with the emergence of antigenically distinct VOCs. Omicron is the evolutionary most distant reported VOC with a hitherto unprecedented number of amino acid alterations in its S glycoprotein, including at least 20 15 amino acid changes in the RBD and extensive changes in the N-terminal domain (NTD) (14).

These alterations are predicted to affect most neutralizing antibody epitopes (15) (16) (17) (18) . In addition, 4 Omicron is highly transmissible, and its sublineages BA.1 and BA.2 have spread rapidly across the globe, outcompeting Delta within weeks to become the dominant circulating VOC (19, 20) .

To date, over 1 billion people worldwide have been vaccinated with the mRNA-based COVID-19 vaccine BNT162b2 and have received the primary 2-dose series or further boosters (21) . This vaccine is contributing substantially to the pattern of population immunity in many regions on 5 which further immune editing and effects of currently spreading variants will build upon.

To characterize the effect of Omicron breakthrough infection on the magnitude and breadth of serum neutralizing activity and BMEM cells, we studied blood samples from individuals that were double-or triple-vaccinated with BNT162b2.

As an understanding of the antigen-specific B cell memory pool is a critical determinant of an 10 individual's ability to respond to newly emerging variants, our data will help to guide future vaccine development. 5 

Blood samples have been sourced from the biosample collection of BNT162b2 vaccine trials, and a biobank of prospectively collected samples from vaccinated individuals with subsequent SARS-CoV-2 Omicron breakthrough infection. Samples were selected to investigate biomarkers 5 in four independent groups, namely individuals who were (i) double-or (ii) triple-vaccinated with BNT162b2 without a prior or breakthrough infection at the time of sample collection (BNT162b2 2 , BNT162b2 3 ) and individuals who were (iii) double-or (iv) triple-vaccinated with BNT162b2 and who experienced breakthrough infection with the SARS-CoV-2 Omicron variant after a median of approximately 5 months or 4 weeks, respectively (BNT162b2 2 + Omi, 10 BNT162b2 3 + Omi) (see materials and methods). Immune sera were used to characterize Omicron infection-associated changes to the magnitude and the breadth of serum neutralizing activity. PBMCs were used to characterize the VOC-specificity of peripheral BMEM cells recognizing the respective full-length SARS-CoV-2 S protein or its RBD (Fig. 1 , Tables S1 to S3). 15 

To evaluate the neutralizing activity of immune sera, we used two orthogonal test systems: a well-characterized pseudovirus neutralization test (pVNT) (22, 23) to investigate the breadth of inhibition of virus entry in a propagation-deficient set-up, as well as a live SARS-CoV-2 20 neutralization test (VNT) designed to evaluate neutralization during multicycle replication of authentic virus with the antibodies maintained throughout the entire test period. For the former, we applied pseudoviruses bearing the S proteins of Omicron sublineages BA.1 or BA.2, other 6 SARS-CoV-2 VOCs (Wuhan, Alpha, Beta, Delta) to assess breadth and of SARS-CoV (herein referred as SARS-Cov-1) to detect potential pan-Sarbecovirus neutralizing activity (24).

As reported previously (22, 25, 26) , in Omicron-naïve double-vaccinated individuals 50% pseudovirus neutralization (pVN50) geometric mean titers (GMTs) of Beta and Delta VOCs were reduced, and neutralization of both Omicron sublineages was virtually undetectable (Fig. 2a, fig.   5 S1a, Table S4 ). In Omicron-naïve triple-vaccinated individuals, pVN50 GMTs against all tested VOCs were substantially higher with robust neutralization of Alpha, Beta and Delta. While GMTs against Omicron BA.1 were significantly lower compared to Wuhan (GMT 160 vs 398), titers against Omicron BA.2 were also considerably reduced at 211 (Fig. 2a, fig. S1a , Table S5 ).

Thus, triple vaccination induced a similar level of neutralization against the two Omicron 10 sublineages.

Omicron breakthrough infection had a marked effect on magnitude and breadth of the neutralizing antibody response of both double-and triple-vaccinated individuals, with slightly higher pVN50 GMTs observed in the triple-vaccinated individuals (Fig. 2a, fig. S1b , Table S6 ).

The pVN50 GMT of double-vaccinated individuals with breakthrough infection against Omicron 15 BA.1 and BA.2 was more than 100-fold and 35-fold above the GMTs of Omicron-naïve double- GMTs comparable to or above those against the Wuhan reference in Omicron-naïve double- 10 vaccinated individuals (GMT≥120 , Table S4 and S6).

Authentic live SARS-CoV-2 virus neutralization assays conducted with Wuhan, Beta, Delta and Omicron BA.1 confirmed these observations (Fig 2b, fig. S1c , d, Tables S7 to S9). In BNT162b2 double-and triple-vaccinated individuals, Omicron infection was associated with a strongly increased neutralizing activity against Omicron BA.1 with 50% virus neutralization (VN50) 15 GMTs in the same range as against the Wuhan strain ( Table S11 ).

Next, we investigated the phenotype and quantity of SARS-CoV-2 S protein specific B cells in 10 these individuals. To this aim, we employed flow cytometry-based B cell phenotyping assays for differential detection of variant-specific S protein-binding B cells in bulk PBMCs. We found that all S protein-and RBD-specific B cells in the peripheral blood were of a BMEM phenotype (BMEM; CD20 high CD38 int/neg , fig. S3 ), as antigen-specific plasmablasts or naïve B cells were not detected (data not shown). The assays therefore allowed us to differentiate for each of the SARS- 15 CoV-2 variants between BMEM cells recognizing the full S protein or its RBD that is a hotspot for amino acid alterations, and variant-specific antigenic epitopes (15, 16) (Fig. 3a) .

As expected, the overall frequency of antigen-specific BMEM cells varied across the different groups. Consistent with prior reports (27) , the frequency of BMEM cells in Omicron-naïve doublevaccinated individuals was low at an early time point after vaccination and increased over time: 20 At 5 months as compared to 3 weeks after the second BNT162b2 dose, S protein-specific BMEM cells almost quadrupled, RBD-specific ones tripled across all VOCs thereby reaching quantities 9 similar to those observed in Omicron-naïve triple-vaccinated individuals (Fig. 3b, c, fig. S4a , b, c, Table S12 ).

Double or triple BNT162b2-vaccinated individuals with a SARS-CoV-2 Omicron breakthrough infection exhibited a strongly increased frequency of BMEM cells, which was higher than those of Omicron-naïve triple-vaccinated individuals (Fig. 3b, d fig. S4d , e, k, l). 5 In all groups, including Omicron-naïve and Omicron infected individuals, BMEM cells against We then compared the ratios of RBD-to S protein-binding BMEM cells within the different 10 groups and found that they are biased towards S protein recognition for the Omicron BA.1 VOC, particularly in the Omicron-naïve groups ( Figure 3f ). In the Omicron-experienced groups this ratio is higher, indicating that an Omicron breakthrough infection improved Omicron BA.1 RBD recognition. 15 

Our findings imply that Omicron infection in vaccinated individuals boosts not only neutralizing activity and BMEM cells against Omicron BA.1, but broadly augments immunity against various 20 VOCs. To investigate the specificity of antibody responses at a cellular level, we performed multi-parameter analyses of BMEM cells stained with fluorescently labeled variant-specific S or RBD proteins. 10 By applying a combinatorial gating strategy, we sought to distinguish between BMEM cell subsets that could identify only single variant-specific epitopes of Wuhan, Alpha, Delta or Omicron BA.1, versus those that could identify any given combination thereof (Fig 4a, fig. S3 ).

In a first analysis, we evaluated BMEM cell recognition of Wuhan and Omicron BA.1 S and RBD proteins (Fig. 4b, c, d) . 

Omicron-naïve double-vaccinated individuals, and even more predominantly from triple-10 vaccinated individuals were directed against epitopes shared by both Wuhan and Omicron BA.1 SARS-CoV-2 variants. Consistent with the fact that vaccination with BNT162b2 can elicit immune responses against Wuhan epitopes that do not recognize the corresponding altered epitopes in the Omicron BA.1 S protein (Fig. 4b, c) , we found in most individuals a smaller but clearly detectable proportion of BMEM cells that recognized only Wuhan S protein or RBD. 15 Consistent with the lack of exposure, no BMEM cells binding exclusively to Omicron BA.1 S or RBD protein were detected in these Omicron-naïve individuals.

In Omicron convalescent individuals, frequencies of BMEM cells recognizing S protein epitopes shared between Wuhan and Omicron BA.1 were significantly higher than in the Omicron-naïve ones (Fig 4b, c) . In most of these subjects, we also found a small proportion of exclusively 20 Wuhan S protein-specific BMEM cells, as well as a slightly lower frequency of exclusively Omicron BA.1 variant S protein-specific ones. 11 A similar but slightly different pattern was observed by B cell staining with labeled RBD proteins (Fig 4b, d) . Again, Omicron breakthrough infection of double-/triple-vaccinated individuals was found to primarily boost BMEM cells reactive with conserved epitopes. A moderate boost of Wuhan-specific reactivities was observed; however, we could hardly detect only Omicron-RBD-specific BMEM cells in the tested individuals (Fig. 4d) . 5 Next, we employed the combinatorial gating approach to identify the subsets of S protein or RBD binding BMEM cells that either bind exclusively to Wuhan or Omicron BA.1, or to common epitopes conserved broadly across all four variants, Wuhan, Alpha, Delta and Omicron BA.1 (Fig 4e) . Across all four study groups, we found that the frequency of BMEM cells recognizing S protein epitopes conserved across all tested variants accounted for the largest fraction of the pool 10 of S protein-binding BMEM cells (Fig. 4f, all 4+ve) . The S protein of the Wuhan strain does not have an exclusive amino acid change that distinguishes it from the spike proteins of the Alpha, Delta, or Omicron BA.1 VOCs. Accordingly, we hardly detected BMEM cells exclusively recognizing the Wuhan S protein in any individual (Fig. 4f) . In several individuals with Omicron breakthrough infection, we detected a small proportion of BMEM cells that bound exclusively to 15 Omicron BA.1 S protein (Fig. 4f) , whereas almost none of the individuals displayed a strictly Omicron BA.1 RBD-specific response (Fig. 4g) .

Our findings indicate that SARS-CoV-2 Omicron breakthrough infection in vaccinated individuals primarily expands a broad BMEM cell repertoire against conserved S protein and RBD epitopes rather than inducing large numbers of Omicron-specific BMEM cells. 20 To further dissect the nuances of this response, we characterized the BMEM subsets directed against the RBD. We used the combinatorial Boolean gating approach to discern BMEM cells with distinct binding patterns in the spectrum of strictly variant-specific and common epitopes shared 12 by several variants. Multiple sequence alignment revealed that the Omicron BA.1 RBD diverges from the RBD sequence regions conserved in Wuhan, Alpha and Delta by 13 single amino acid alterations ( fig. S5 ). We found that all Omicron convalescent individuals had robust frequencies of BMEM cells that recognized Wuhan, Alpha as well as the Delta VOC RBDs, but not Omicron BA.1 RBD, while BMEM cells exclusively reactive with Omicron BA.1 RBD were almost absent 5 in most of those individuals (Fig. 4h) . We also did not detect BMEM cells that exclusively recognized the Omicron BA.1 and Alpha RBDs, or the Omicron BA.1 and Delta RBDs.

Furthermore, in all individuals we identified two additional subsets of RBD-specific BMEM cells.

One subset was characterized by binding to Wuhan, Alpha as well as Omicron BA.1, but not the Delta, RBD. The other population exhibited binding to Wuhan and Alpha but not Omicron BA.1 10 or Delta RBD (Fig. 4h) . Sequence alignment identified L452R as the only RBD mutation unique for Delta that is not shared by the other 3 variant RBDs (Fig. 4i top) . Similarly, the only RBD site conserved in Wuhan and Alpha but altered in Delta and Omicron BA.1 was found to be T478K (Fig 4i bottom) . Both L452R and T478K alterations are known to be associated with the evasion of vaccine induced neutralizing antibody responses (29, 30) . Of note, no BMEM cells were 15 detected in all combinatorial subgroups in which multiple sequence alignment failed to identify unique epitopes in the RBD sequence that satisfied the Boolean selection criteria (e.g., Wuhan only or Wuhan and Omicron BA.1, but not Alpha, Delta). These observations indicate that the BMEM cell response against RBD is driven by specificities induced through prior vaccination with 

only neutralizing activity and BMEM cells against Omicron but broadly augments immunity against various VOCs.

Our study also provides insights into how this broad immunity is achieved. Our data suggest that L452. An increased focus of the immune response on this domain could help restore immune 20 protection by stimulating BMEM cells that produce neutralizing antibodies against RBD epitopes unaltered in Omicron. 15 Third, the induction of broadly neutralizing antibodies. We found that the majority of sera from Omicron-convalescent but not from Omicron-naïve vaccinated individuals robustly neutralized SARS-CoV-1. This may suggest that Omicron infection in vaccinated individuals stimulates BMEM cells that form neutralizing antibodies against spike protein epitopes conserved in the SARS-CoV-1 and SARS-CoV-2 families. Broadly neutralizing antibodies have recently been 5 described in SARS-CoV-1 infected individuals vaccinated with BNT162b2 (24). Such pan-Sarbecovirus immune responses are thought to be triggered by neutralizing antibodies to highly conserved S protein domains (35, 36) . The greater antigenic distance of the Omicron spike protein from the other SARS-Cov-2 strains may promote targeting of conserved subdominant neutralizing epitopes as recently described to be located in the C-terminal portion of the spike 10 protein (37, 38) . Future studies mapping monoclonal antibodies derived from Omicron-specific BMEM cells will provide further insight into the relevance of these findings.

In aggregate, our results suggest that despite possible imprinting of the immune response by previous vaccination, the preformed B-cell memory pool can be refocused and quantitatively remodeled by exposure to heterologous S proteins to allow neutralization of variants that evade a 15 previously established neutralizing antibody response.

Limitations of this study are that the data were generated from blood samples obtained a few weeks after Omicron infection. Later time points might reveal formation of memory responses against novel epitopes that are not yet visible. Also, the analysis presented here has evaluated sample sets from multiple studies with a limited sample size, different sampling time points and 20 baseline or demographic characteristics. The analysis was limited to BMEM cells while long-lived bone marrow-derived plasma cells (BMPCs) which are known to be BNT162b2 vaccination induced (28) , could not be investigated as they cannot be cryopreserved. 16 In conclusion, while the data are based on samples from individuals exposed to the Omicron S protein as a result of infection, our observations suggest that a vaccine adapted to the Omicron S protein could similarly reshape the B-cell memory repertoire and therefore may be more beneficial than an extended series of boosters with the existing Wuhan-Hu-1 spike based vaccines. Schematic in (a) was created with BioRender.com demographics support.

This work was supported by BioNTech. Participant baseline characteristics are provided in Table S1 to Table S3 and Table S10 . The neutralization titers are provided in Tables S4 to S9 and S11. The frequencies of SARS-CoV-2 variant S protein/RBD-specific BMEM cells are provided in Table S12 . Materials are available 15 from the authors under a material transfer agreement with BioNTech.

Figs. S1-S6 20 Tables S1-S12

Neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection. Nature medicine

Immune correlates analysis of the mRNA-1273 COVID-19 vaccine 5 efficacy clinical trial

Immune imprinting, breadth of variant recognition, and germinal center response in human SARS-CoV-2 infection and vaccination

Neutralizing antibody responses elicited by SARS-CoV-2 mRNA vaccination wane over time and are boosted by breakthrough infection

Antibody titers against SARS-CoV-2 decline, but do not disappear for several months

Dynamics of SARS-CoV-2 neutralising antibody responses and duration of immunity

Waning Immunity after the BNT162b2 Vaccine in Israel. The New England journal of medicine

SARS-CoV-2 Neutralization with BNT162b2 Vaccine Dose 3. The New 20 England journal of medicine

Safety and immunogenicity of SARS-CoV-2 variant mRNA vaccine boosters in healthy adults

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

WHO Technical Advisory Group on COVID-19 Vaccine Composition (TAG-CO-VAC), Interim statement on COVID-19 vaccines in the context of the circulation of the Omicron SARS

CoV-2 variant

Memory B cells, 10 but not long-lived plasma cells, possess antigen specificities for viral escape mutants. The Journal of experimental medicine

Distinct germinal center selection at local sites shapes memory B cell response to viral escape. The Journal of experimental medicine

Headquarters (HQ), WHO Health Emergencies Programme, Enhancing response to

CoV-2 variant: Technical brief and priority actions for Member States

SARS-CoV-2 variants, spike mutations and immune escape

The receptor binding domain of the viral spike protein is an 20 immunodominant and highly specific target of antibodies in SARS-CoV-2 patients

The Omicron variant is highly resistant against antibody-mediated neutralization

SARS-CoV-2 Omicron-B.1.1.529 leads to widespread escape from neutralizing antibody responses

European Centre for Disease Prevention and Control, Assessment of the further spread and potential impact of the SARS-CoV-2 Omicron variant of concern in the EU/EEA

COVID-19

A global database of COVID-19 vaccinations

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

Neutralization of SARS-CoV-2 lineage B.1.1.7 pseudovirus by BNT162b2 vaccine-elicited human sera

Pan-Sarbecovirus Neutralizing Antibodies in BNT162b2-Immunized SARS-CoV-1 Survivors. The New England journal of medicine

BNT162b2 vaccine induces neutralizing antibodies and poly-specific T cells in humans

BNT162b2-elicited neutralization of B.1.617 and other SARS-CoV-2 variants

SARS-CoV-2 Omicron-neutralizing memory B-cells are elicited by two doses of BNT162b2 mRNA vaccine

Germinal centre-driven maturation of B cell response to mRNA vaccination

Structural basis for SARS-CoV-2 Delta variant recognition of ACE2 receptor and broadly neutralizing antibodies

Complete map of SARS-CoV-2 RBD mutations that escape the monoclonal antibody LY-CoV555 and its cocktail with LY-CoV016

Omicron extensively but incompletely escapes Pfizer BNT162b2 neutralization

Cryo-EM structure of the SARS-CoV-2 Omicron spike

Omicron escapes the majority of existing SARS-CoV-2 neutralizing antibodies

SARS-CoV-2 breakthrough infections elicit potent, broad, and durable 15 neutralizing antibody responses

Structural basis for broad coronavirus neutralization

Cross-neutralization of SARS-CoV-2 by a human monoclonal SARS-CoV 20 antibody

Structural basis and mode of action for two broadly neutralizing antibodies against SARS-CoV-2 emerging variants of concern

Broad betacoronavirus neutralization by a stem helix-specific human 25 antibody

We thank the BioNTech German clinical Phase 1/2 trial (NCT04380701, EudraCT: 2020-001038-36), the German Phase 2 rollover booster trial (NCT04949490, EudraCT: 2021-002387-50) the global clinical Phase 2 trial (NCT04380701) participants, and the Omicron convalescent Research Study participants from whom the post-immunization human sera and PBMCs were 5 obtained. We thank the many colleagues at BioNTech and Pfizer who developed and produced the BNT162b2 vaccine candidate. We thank Sabrina Jägle and Nina Beckmann for logistical support. We thank the VisMederi team for performing excellent work on live virus-neutralising antibody assays. We thank Svetlana Shpyro, Sayeed Nadim, Christina Heiser, Ayca Telorman, Claudia Müller, Amy Wanamaker, Nicki Williams and Jennifer VanCamp for sample