SARS-CoV-2 mRNA vaccines induce a robust germinal centre reaction in humans

25 Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) messenger RNA 26 (mRNA)-based vaccines are ~95% effective in preventing coronavirus disease 2019 1 – 5 . However, 27 the dynamics of antibody secreting plasmablasts (PBs) and germinal centre (GC) B cells induced 28 by these vaccines in SARS-CoV-2 naïve and antigen-experienced humans remains unclear. Here 29 we examined peripheral blood and/or lymph node (LN) antigen-specific B cell responses in 32 30 individuals who received two doses of BNT162b2, an mRNA-based vaccine encoding the full- 31 length SARS-CoV-2 spike (S) gene 1 . Circulating IgG- and IgA-secreting PBs targeting the S 32 protein peaked one week after the second immunization then declined and were undetectable 33 three weeks later. PB responses coincided with maximal levels of serum anti-S binding and 34 neutralizing antibodies to a historical strain as well as emerging variants, especially in individuals 35 previously infected with SARS-CoV-2, who produced the most robust serological responses. Fine 36 needle aspirates of draining axillary LNs identified GC B cells that bind S protein in all participants 37 sampled after primary immunization. GC responses increased after boosting and were detectable 38 in two distinct LNs in several participants. Remarkably, high frequencies of S-binding GC B cells 39 and PBs were maintained in draining LNs for up to seven weeks after first immunization, with a 40 substantial fraction of the PB pool class-switched to IgA. GC B cell-derived monoclonal antibodies 41 predominantly targeted the RBD, with fewer clones binding to the N-terminal domain or shared 42 epitopes within the S proteins of human betacoronaviruses OC43 and HKU1. Our studies 43 demonstrate that SARS-CoV-2 mRNA-based vaccination of humans induces a robust and 44 persistent GC B cell response that engages pre-existing as well as new B cell clones, which 45 enables generation of high-affinity, broad, and durable humoral immunity. three the S 16,17 (a) a Washington strain (2019n-CoV/USA) a D614G (b) a B.1.1.7 isolate signature in the spike 18 , the 69 70 and 144 145 deletions N501Y, A570D, D614G and P681H and (c) a chimeric SARS-CoV-2 with a B.1.351 spike gene in the Washington strain (Wash SA-B.1.351) that contained the following changes: D80A, deletion, E484K, N501Y, D614G and A701V. Serum neutralizing titers increased boosting,


47
The concept of using mRNAs as vaccines was first introduced over 30 years ago 6,7 . Key 48 refinements that improved the biological stability and translation capacity of exogenous mRNA 49 enabled development of these molecules as vaccines 8,9 . The emergence of SARS-CoV-2 in 50 December, 2019 and the ensuing pandemic has unveiled the potential of this platform 9-11 . In the 51 United States, more than 80 million people have received one of the two SARS-CoV-2 mRNA-52 based vaccines that were granted emergency use authorization by the FDA in December, 2020.

63
Antibody-secreting plasmablasts (PBs) that bound SARS-CoV-2 S protein were measured 64 in blood using an enzyme-linked immune absorbent spot (ELISpot) assay. SARS-CoV-2-S-65 specfic IgG-and IgA-secreting PBs were detected in blood three weeks after primary 66 immunization in 19 of 25 participants with no history of SARS-CoV-2 infection but 0 of 7 67 participants previously infected with SARS-CoV-2. PBs peaked in blood during the first week after 68 boosting (week 4) in all participants, with frequencies varying widely from 3 to 4,100 S-binding 69 PBs per 10 6 PBMC (Fig. 1b, c). Plasma IgG antibody titers against S measured by ELISA 70 increased in all participants over time, reaching peak geometric mean half-maximal binding 71 dilutions of 5,239 and 14,924 among participants without and with history of SARS-CoV-2 72 infection, respectively, and then declining slightly by seven weeks after immunization. IgG titers 73 against the receptor binding domain (RBD) of S showed similar kinetics, reaching peak geometric 74 mean half-maximal binding dilutions of 4,511 and 8,034 among participants without and with 75 history of SARS-CoV-2 infection, respectively before declining (Fig. 1d). The functional quality of 76 serum antibody titers was measured using high-throughput focus reduction neutralization tests

97
The BNT162b2 vaccine is injected into the deltoid muscle, which drains primarily to the lateral 98 axillary lymph nodes. Ultrasound was used to identify and guide FNA of accessible axillary nodes 99 on the side of immunization approximately 3 weeks after primary immunization. In 5 of the 12 100 participants, a second draining LN was identified and sampled following secondary immunization 101 (Fig. 2a). GC B cells, defined as CD19 + CD3 -IgD lo Bcl6 + CD38 int lymphocytes, were detected in 102 all LNs (Fig. 2b, c, Extended Data Fig. 1a). FNA samples were co-stained with two fluorescently 103 labeled S probes to detect S-binding GC B cells. A control tonsillectomy sample with a high 104 frequency of GC B cells collected prior to the SARS-CoV-2 pandemic from an unrelated donor 105 was stained as a negative control. S-binding GC B cells were detected in FNAs from all 12 106 participants following primary immunization. Kinetics of the GC response varied among 107 participants, but S-binding GC B cell frequencies increased at least transiently in all participants 108 after boosting and persisted at high frequency in most individuals for at least 7 weeks (Fig. 2d,

110
To evaluate the domains targeted by the S protein-specific GC response after vaccination, we 111 generated recombinant monoclonal antibodies (mAbs) from single-cell sorted S-binding GC B 112 cells (defined by the surface marker phenotype CD19 + CD3 -IgD lo CD20 hi CD38 int CD71 + CXCR5 + 113 lymphocytes) from three of the participants one week after boosting (Extended Data Fig. 1a).

121
In addition to GC B cells, robust PB responses were detected in the draining LNs of all 12 122 participants sampled. S-binding PBs, defined as CD19 + CD3 -IgD lo CD20 lo CD38 + CD71 + Blimp1 + 123 lymphocytes, were detected in all LNs sampled and increased in frequency after boosting ( Fig.   124   3a, b). The detected lymph node PBs were unlikely a contaminant of blood because CD14 + 125 monocyte/granulocyte frequencies were below 1% in all FNA samples, well below the 10% 126 threshold we previously established 19 . Moreover, S-binding PBs were detected in FNA samples 127 5 and 7 weeks after primary immunization, when they had become undetectable in blood from all 128 participants in the cohort. The vast majority of S-binding LN PBs were isotype-switched 5 weeks 129 after primary immunization, and IgA-switched cells accounted for 25% or more of the PBs in 6 of 130 12 participants (Fig. 3c, d).

131
This study evaluated whether SARS-CoV-2 mRNA-based vaccines induce antigen-specific 132 GC cell responses in humans. The vaccine induced a strong IgG-dominated PB response in blood 133 that peaked one week after the booster immunization. In the draining lymph nodes, we detected 134 robust SARS-CoV-2 S-binding GC and PB responses in aspirates from 12 of 12 participants.

135
These responses were detectable after the first immunization and increased after the second.