SARS-CoV-2 infection and vaccination elicit distinct pharyngeal mucosal B cell responses in children

SARS-CoV-2 infection and vaccination elicit distinct pharyngeal mucosal B cell responses in children | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article SARS-CoV-2 infection and vaccination elicit distinct pharyngeal mucosal B cell responses in children Kalpana Manthiram, Qin Xu, Lihong Shi, Liya Wang, Foo Cheung, and 29 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7428491/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Mucosal immunity is an important correlate of protection against respiratory infections such as SARS-CoV-2. Comparing B cell responses in the upper respiratory tract following vaccination and infection may offer unique insights into mucosal immunity. Here, we characterized antigen-specific B cells in the tonsils, adenoids, and peripheral blood of children who had been infected with SARS-CoV-2 or vaccinated with SARS-CoV-2 mRNA vaccines. SARS-CoV-2-specific switched memory B cells (B SM ) and germinal center B cells were found in the blood and pharyngeal lymphoid tissues after vaccination or infection. However, infection generated a higher proportion of IgA + B SM and CXCR3 + CD21 + B SM , which showed distinct spatial localization, greater clonal expansion and increased propensity for plasma cell differentiation compared to their CXCR3 - counterparts, accompanied by persistent activation of innate and T follicular helper cells in the tissues. Our data provide evidence for tissue-specific B cell memory after either SARS-CoV-2 vaccination or infection, but with distinct characteristics that can influence the quality, durability, and localization of immunity. Biological sciences/Immunology/Mucosal immunology Biological sciences/Immunology/Translational immunology Biological sciences/Immunology/Lymphoid tissues/Tonsils Biological sciences/Immunology/Adaptive immunity/Humoral immunity/Immunological memory Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 INTRODUCTION The high mortality and far-reaching effects of the COVID-19 pandemic triggered the rapid development of several vaccine platforms including two mRNA-based vaccines, BNT162b2 (Pfizer) and mRNA-1273 (Moderna), which were shown to have high efficacy in preventing severe COVID-19 1, 2 . These vaccines generate high-titers of serum antibodies with neutralizing activity against early and more later circulating SARS-CoV-2 strains, as well as virus-specific memory B and T cells in the peripheral blood 3, 4, 5 . However, the immunity afforded by these vaccines wanes with time and population-based studies suggest that the durability of immune protection following vaccination may be lower than that from prior infection with SARS-CoV-2 6, 7, 8, 9 . SARS-CoV-2 initially enters and infects upper respiratory tract tissues 10, 11 , and immunity in the upper respiratory tract, particularly mucosal IgA levels, has been shown to correlate with protection against COVID-19 12, 13 . Therefore, characterizing upper respiratory tract immunity provided by these intramuscularly delivered vaccines is important to understand their efficacy and to provide insights for future vaccine development against a variety of pathogens. A major component of protective immune memory to viruses, including SARS-CoV-2, is the development and maintenance of long-lasting, high-affinity memory B cells, which are poised to rapidly differentiate into antibody-secreting cells upon secondary antigen encounter 14, 15 . A subset of memory B cells in the tissues, known as tissue-resident memory B cells, do not recirculate and have been shown to provide effective localized immunity 16 . Secondary lymphoid organs enable the development of germinal centers (GC), which are essential for creating long-lived and high-affinity memory B cells. In these specialized structures, B cells undergo somatic hypermutation (SHM) and affinity maturation with help from T follicular helper (Tfh) cells. COVID-19 mRNA vaccines have been shown to generate SARS-CoV-2-specific Tfh and GC B cells (GCBs) in the draining axillary lymph nodes, which can persist at least 6 months 17, 18, 19, 20 ; however, the ability of intramuscular vaccines to elicit immunity in upper respiratory tract lymphoid tissues remains an important question. Furthermore, whether immune memory in the tissue differs after vaccination versus infection is largely unknown. The tonsils and adenoids are secondary lymphoid structures at the mucosal surface of the upper respiratory tract that contain lymphoid cells not found in the peripheral blood; viral-specific GCBs and tissue-resident memory T and B cells have been described in these mucosal tissues following respiratory infections 21, 22, 23 . Assessing SARS-CoV-2-specific-B cells in these tissues following intramuscular vaccination may, therefore, offer insights into the characteristics, magnitude, and durability of tissue immunity at the sites where the host first encounters the virus. Here, we characterized SARS-CoV-2-specific B cells in the tonsils, adenoids, and peripheral blood of children immunized with a SARS-CoV-2 mRNA vaccine compared with those of children previously infected with SARS-CoV-2, taking advantage of the unique period during the COVID-19 pandemic to study immunity to an unencountered airborne pathogen and new vaccines. We found SARS-CoV-2-specific B cells, which were primarily switched memory B cells (B SM ) but also some GCBs, in the tonsils and adenoids of children both post-infection and post-vaccination, indicating that immune memory can be found and maintained in the upper respiratory tract after intramuscular vaccination as well as after infection. Nonetheless, vaccination and infection generated B SM with different characteristics, including increased induction of a population defined by CXCR3 and CD21 expression post-infection. This CXCR3 + CD21 + B SM population showed a greater propensity for plasma cell differentiation and mucosal homing, had unique spatial distribution, and correlated with persistent adaptive and innate immune cell activation in mucosal lymphoid tissues. Our findings provide a framework for understanding responses to vaccination and infection, including novel parameters for assessing vaccine-induced mucosal immunity. RESULTS SARS-CoV-2-specific B cells are found in the pharyngeal tissues and blood post-vaccination To assess SARS-CoV-2-specific immunity in vaccinated individuals, we collected serum, peripheral blood mononuclear cells (PBMC), and tonsil and adenoid tissues from 21 children undergoing tonsillectomy and/or adenoidectomy from December 2021 to September 2022 at Children’s National Hospital in Washington, DC, USA, who had received at least one dose of the monovalent mRNA vaccines, BNT162b2 or mRNA-1273 (Figure 1a, Supplemental Tables 1 and 2). These vaccines encode the ancestral Wuhan strain-like spike protein. Of these vaccinated subjects, 10 had no evidence of prior SARS-CoV-2 infection by history of positive PCR or antigen test, nor by positive serum titers against nucleocapsid (NC) and/or open reading frame 8 (ORF8), which have been used for serodiagnosis of SARS-CoV-2 infection 24, 25 (Supplemental Table 3). These 10 subjects comprised our vaccinated only cohort (VAC). We compared these vaccinated children to subjects we previously recruited from late 2020 to early 2021 who had prior SARS-CoV-2 infection (INF), before the availability of vaccination for children (Figure 1a, Supplemental Tables 1 and 2) 23 . All had mild or asymptomatic infection. Based on the timing of sample collection and known infection dates, most of these subjects were likely infected with the D614G or alpha strains, which bear sequence similarity to the spike mRNA used in the BNT162b2 or mRNA-1273 vaccines administered to the VAC cohort. The interval from infection to surgery among INF participants was comparable to the interval from the last vaccination to surgery in the VAC group (Figure 1b, Extended Data Figure 1a). We also included a group of unvaccinated pediatric controls (CON) with no serologic or cellular evidence of prior COVID-19, who were recruited during our initial study (Figure 1a, Supplemental Table 2 and 3). To compare VAC and INF groups, we used both direct statistical comparisons (Mann-Whitney U) and linear models correcting for age. All VAC and INF participants had serum neutralizing antibodies to the WA-1 strain (Figure 1c, Supplemental Table 3) with no significant differences noted between the two groups. Similar to our INF cohort, VAC subjects showed a trend towards lower neutralizing titers to WA-1 with greater time from infection (Figure 1d), consistent with previous reports 26 . We also evaluated neutralizing titers to omicron, a variant that emerged in late 2021 and rapidly became the dominant strain, causing numerous breakthrough infections in those who were previously infected or vaccinated 27, 28 . Although neutralizing titers were lower to omicron than to WA-1 in both groups, they were on average higher in VAC than INF, with a higher proportion of VAC subjects having positive titers (Figure 1c, Supplemental Table 3). We then used fluorescently-labeled probes for the receptor-binding domain (RBD) and S1 portion of the spike protein from the original Wuhan strain to identify SARS-CoV-2-recognizing B cells (S1 + RBD + ) by flow cytometry (Figure 1e). As previously reported, S1 + RBD + CD19 + B cells were found in both tissues and blood of most INF participants (Supplemental Table 3). Notably, almost all VAC subjects also had S1 + RBD + CD19 + B cells in both their pharyngeal tissues and blood (Figure 1f, Supplemental Table 3). A higher frequency of S1 + RBD + cells was noted among B cells in peripheral blood of VAC compared to INF individuals but these trended towards a higher frequency in the adenoids of INF subjects (Figure 1f). These findings were confirmed comparing S1 + RBD + B cells in the blood versus tissues within individual subjects by group (Figure 1g). Among INF subjects, the percentage of S1 + RBD + B cells in the PBMCs, adenoids, and tonsils were all significantly correlated, as previously reported (Figure 1h) 23 . In contrast, among VAC subjects, the percentage of S1 + RBD + B cells in the PBMCs correlated significantly with serum neutralizing titers, but not significantly with the percentages of S1 + RBD + B cells in the adenoid and tonsil (Figure 1h). We also identified omicron-recognizing B cells (RBD-Omi + ) using two fluorescently labelled RBD probes from the omicron strain (Extended Data Figure 1b-c). VAC participants had a higher proportion of detectable S1 + RBD + B cells that also recognized omicron in the blood, adenoid and tonsil compared to INF (Extended Data Figure 1d), suggesting vaccination provides broader coverage of variants than infection in both the blood and tissues 4, 19, 29 . Thus, although vaccinated children had greater B cell responses in the peripheral blood compared to the mucosal tissues, we could detect SARS-CoV-2-specific B cells in the secondary lymphoid tissue of the upper respiratory tract, distal from the site of intramuscular immunization in nearly all vaccinated children. Infection induces a higher proportion of IgA + SARS-CoV-2-specific B SM than vaccination To provide insight into SARS-CoV-2-specific B cells, we used a high dimensional flow cytometry panel consisting of 29 markers including fluorescently labelled SARS-CoV-2 probes as well as surface markers to describe B cell subsets, isotype, activation, tissue residence, and homing. In both INF and VAC, the majority of S1 + RBD + B cells in PBMCs, tonsils, and adenoids were B SM (Figure 1i). Unlike neutralizing titers which declined with greater time from vaccination/infection, the percentage of S1 + RBD + B SM in the tissues were stable (VAC) or increased (INF) with time (Figure 1j). Mucosal IgA is protective against SARS-CoV-2 infection, but mRNA vaccination has been shown to engender less mucosal IgA in the respiratory tract compared to infection 9, 12, 30, 31 . SARS-CoV-2-specific B SM were predominantly IgG + post-infection and post-vaccination (Figure 1k). However, INF individuals had a greater proportion of IgA + cells (and less IgG + cells) among S1 + RBD + B SM in PBMC and tissues compared to VAC subjects (Figure 1k, Extended Data Figure 1e). Thus, infection with early circulating strains was associated with a greater proportion of IgA + B SM than vaccination, both in tissues and blood. In humans, B SM can be subdivided based on expression of CD21 and CD27 32, 33 . CD27 + CD21 + conventional or “resting” memory B cells (cMBCs) are quiescent, highly affinity-matured cells that emerge from GC reactions. Following infection or vaccination, responding B cells become activated and display heterogenous phenotypes including downregulation of CD21 15, 34, 35, 36 37, 38 . Those that express CD27 (CD27 + CD21 - ) are referred to as activated memory cells (acMBC), while those that express low CD21 and CD27 are called atypical MBCs (atMBC); both are expanded in chronic infections, autoimmune diseases, and shortly after acute infection and vaccination including COVID-19 vaccines 35, 38, 39, 40, 41, 42, 43, 44 . The atMBCs (and some acMBCs) express CD11c, FCRL3, FCRL4, FCRL5, and CD85J, as well as Tbet 45, 46, 47 . Most of the SARS-CoV-2-specific B SM in both VAC and INF tissues and blood were cMBC (Extended Data Figure 1f-g, Supplemental Table 4). However, a larger proportion of acMBCs were noted in VAC PBMC and tonsils compared to INF, with a similar trend in the adenoids (Extended Data Figure 1f-g). Neutralizing titers correlated with the proportion of acMBCs among S1 + RBD + B SM in both blood and adenoids (Extended Data Figure 1h). Furthermore, like neutralizing titers, the frequency of CD21 lo/- B SM populations (acMBC and atMBC) in the blood declined with time from vaccination (Extended Data Figure 1i). SARS-CoV-2-specific B SM in the blood are phenotypically different post-infection and post-vaccination To further compare the phenotypes of S1 + RBD + B cells between the two groups beyond traditional B SM markers, we concatenated the S1 + RBD + B cells from all VAC and INF participants and performed unsupervised clustering analyses based on expression of 19 B cell surface markers. Cells from the peripheral blood and tissues were assessed separately due to differences in cell populations. Unsupervised clustering of S1 + RBD + B cells from PBMCs generated 12 phenotypically distinct clusters, 11 of which were B SM subsets (Figure 2a-c, Supplemental Table 5). VAC and INF samples largely segregated on principal component analysis (PCA) on PC2; a few VAC subjects were distinct on PC1 (Figure 2d). We compared the proportion of each cluster in the two groups by Mann-Whitney U or a linear model correcting for age (Figure 2c, e; Extended Data Figure 2a), and found that INF subjects had higher proportions of S1 + RBD + B cells in cluster 10 with a trend towards more in clusters 2 and 6 compared to VAC individuals; these three clusters represented CXCR3 + B SM (Figure 2b, e-h). Cluster 10, the most significant cluster, represented IgA + CXCR3 + B SM . CXCR3 is a chemokine receptor induced by IFN-g that directs B SM to areas of inflammation 48 . Although CXCR3 can be expressed on CD21 - atMBCs along with CD11c and other markers regulated by Tbet 49, 50 , these three clusters did not express atMBC markers and were largely CD21 + (Figure 2a-b, h) 5, 42 . Conversely, VAC subjects had a higher portion of S1 + RBD + B cells in several CD21 lo clusters, including clusters 1, 3, 4 and 7 (CD21 lo IgG + B SM ), and cluster 11 (CD21 lo IgA + B SM cluster found in 2 subjects) (Figure 2a-c, e-h). Clusters 4, 7 and 11 expressed atMBC markers including CD11c, CD85J, FCRL3/5 and CD95 and had a mixture of CD27 + and CD27 - cells but were not CXCR3 + (Figure 2b). Cluster 3 was an IgG + CD21 - CD27 + acMBC cluster that expressed variable CXCR3. Thus, unbiased analyses revealed phenotypic differences in SARS-CoV-2-specific B cells in the peripheral blood post-infection and post-vaccination, including enrichment of CXCR3 + CD21 + B SM post-infection. Distinct SARS-CoV-2-specifc GCBs and B SM are in the pharyngeal tissues of INF and VAC subjects We next examined the characteristics of SARS-CoV-2-specific B cells in the pharyngeal tissues, generating 14 clusters representing naïve and unswitched memory B cells (USM), GC and pre-GC B cells (CD38 + CD10 + CD95 + CD71 + ), and B SM (Figure 3a-c, Extended Data Figure 2b-c, Supplemental Table 6). Like PBMCs, PCA analyses showed separation between VAC and INF subjects (Figure 3d). We previously found that SARS-CoV-2 infection induces local GC reactions in the pharyngeal lymphoid tissues 23 . As expected, S1 + cells were observed in GCB clusters in most (15/24, 62.5%) INF individuals; however, two out of 10 (20%) VAC subjects also had at least 10 S1 + RBD + B cells within GCB clusters in either the tonsil or adenoid, a threshold based on a recent study 51 (Figure 3e; Supplemental Table 7). These two VAC subjects had their most recent (second) vaccine doses 48 days and 232 days prior to surgery, suggesting this was not temporally associated with recent vaccination (Supplemental Table 2). INF tonsils had a higher proportion of S1 + RBD + B cells in the IgG + GCB cluster (cluster 6) and IgA + GCB cluster (cluster 11, by Mann-Whitney U) than VAC (Figure 3a-c, f-h; Extended Data Figure 2d), supporting the greater effectiveness of natural infection in generating and maintaining GC responses in pharyngeal mucosal tissues over vaccination. Similar to PBMCs, VAC tissues had a higher frequency of S1 + RBD + B cells in certain clusters containing CD27 + IgG + B SM with lower or intermediate CD21 expression and lacking CXCR3 expression. These included a greater proportion in cluster 2 in VAC tonsils, and in cluster 7 (CD62L + ) in VAC adenoids and tonsils compared to INF (Figure 3a-c, f-h; Extended Data Figure 2c-d). In contrast, tonsils and adenoids of INF subjects contained a higher fraction of SARS-COV-2-specific B cells in CXCR3 + CD21 + containing clusters compared to VAC individuals including IgA + B SM populations (clusters 4 and 12) and IgG + B SM populations (clusters 5 and 10 in the adenoids and cluster 3 in tonsils) (Figure 3a-c, f-h, Extended Data Figure 2c-d). Increased percentages of CXCR3 + IgA and IgG cells in INF subjects were confirmed by manual gating (Figure 3i). CXCR3 + has been shown to be important for establishment of tissue-residenceamong memory B cells in murine lungs and production of mucosal IgA in mice, in addition to attracting cells to the mucosa 52, 53, 54 . Although we also observed CXCR3 + B SM in the blood, we saw a trend towards higher proportions of CXCR3 + cells, particularly CXCR3 + IgA + cells, among S1 + RBD + B SM in the adenoid and tonsil compared to blood, supporting the idea that CXCR3 + B SM may preferentially home to mucosal tissues (Extended Data Figure 2e). Thus, infection and vaccination were associated with phenotypically distinct memory B cell populations with a higher proportion of CXCR3 + CD21 + B SM including CXCR3 + IgA + cells post-infection, and B SM populations with lower CD21 expression and lacking CXCR3 post-vaccination in the tissues, analogous to the peripheral blood. Furthermore, a small portion of VAC individuals showed evidence of SARS-CoV-2-specific GCBs in the pharyngeal lymphoid tissue. Evaluation of B SM based on CXCR3 and CD21 expression Given the differential expression of CXCR3 and CD21 on S1 + RBD + B SM from INF and VAC individuals, we used these two markers to define 4 populations among B SM : P1 (CD21 + CXCR3 - ); P2 (CD21 + CXCR3 + ); P3 (CD21 - CXCR3 + ); and P4 (CD21 - CXCR3 - ) (Figure 4a). The proportion of P2 (CD21 + CXCR3 + ) cells among S1 + RBD + B SM was significantly higher in INF PBMCs and tissues compared to VAC (Figure 4b). In contrast, VAC individuals had a higher proportion of CXCR3 - populations, P4 in blood and tissues and P1 in tissues, highlighting differences in CXCR3 expression between antigen-specific B cells in INF and VAC subjects. We then evaluated the expression of additional markers including transcription factors in each population in tonsil and blood. We found that atMBC markers including T-bet, CD95, CD11c, FCRL4, and FCRL3/5, were enriched primarily in P3 and P4, suggesting that these CD21 - populations contained the bulk of atMBCs (Figure 4c-d). Back-gating revealed that P3 and P4 contained mixtures of atMBC and acMBC (Extended Data Figure 3a). Furthermore, the frequency of P3 among S1 + RBD + B SM correlated with neutralizing titers to WA.1 (Figure 4e), consistent with correlations we saw between acMBC and neutralization titers (Extended Data Figure 1h). In contrast, most cells in P2 did not express atMBC markers but expressed an intermediate-low level of T-bet, suggestive of previous exposure to IFN-g that may have driven and maintained expression of CXCR3 54, 55 . Moreover, P1 and P2, the CD21 + populations, expressed high levels of CXCR5, supporting a follicular origin, and were comprised primarily of cMBCs (Figure 4c-d and Extended Data Fig 3a). CD69 is a marker of tissue resident memory B cells (B RM ) that are poised to provide rapid local protection in response to infection or immune challenges 56 . In tonsil and adenoid B SM , we found that P2 had the highest frequency of CD69 + cells, suggesting that this population is enriched for B RM (Figure 4d, Extended Data Figure 3b). Although CD69 + S1 + RBD + B SM were noted post-infection and post-vaccination, the percentage of CD69 + cells among S1 + RBD + B SM was significantly higher in INF tonsils compared to VAC (Figure 4f). Thus, CXCR3 + CD21 + SARS-CoV-2-specific B SM with B RM features were enriched in upper respiratory tract lymphoid tissue after infection compared to vaccination. Tissue CXCR3 + CD21 + S1 + B SM have distinct transcriptomic features To further characterize these populations, we sorted S1 + and S1 - B cells from tonsils and/or adenoids of a subset of 8 INF, 4 VAC, and 2 CON subjects, including the two VAC subjects in whom S1 + GCBs were identified by flow cytometry, and performed Cellular Indexing of Transcriptomes and Epitopes by Sequencing (CITE-seq) using a panel of 20 antibodies including anti-CXCR3 and CD21. In total, 9370 S1 + and 172919 S1 - B cells were captured and assessed for surface protein expression, gene expression and BCR sequencing at the single cell level. Unsupervised clustering based on surface protein expression revealed 6 clusters representing naïve/USM, B SM , GCB, and plasma cells/plasmablasts (PC/PB) (Figure 5a-b). S1 + B cells were predominantly in a B SM cluster (cluster 1) in both INF and VAC tissues (Figure 5c-d). However, a portion of S1 + B cells in both INF and VAC subjects were in cluster 2, which had a gene expression signature consistent with GCBs (Figure 5e-f), providing further support that spike-specific GCBs can be found in upper respiratory lymphoid tissues post-vaccination. Comparison of surface protein to mRNA levels revealed that most CXCR3 + B cells had little or no detectable CXCR3 transcripts (Extended Data Figure 4a). We, therefore, defined B SM populations using CXCR3 and CD21 surface markers (P1-P4) (Figure 5g) to compare their gene expression. As expected, we saw a higher proportion of P2 among S1 + B cells post-infection and higher proportions of P1 and P4 (both CXCR3 - ) post-vaccination (Figure 5h, Extended Data Figure 4b). Trajectory analyses using Slingshot suggested distinct patterns of S1 + B SM development in INF and VAC conditions (Figure 5i). In INF tissues, two branches originating from naive B cells both landed at the P2 state, while in VAC tissues, two diverging trajectories from naïve B cells emerged, including one leading to the P4 population (Figure 5i). By PCA, the CD21 + B SM populations (P1 and P2) segregated distinctly from the CD21 - populations (P3 and P4) on PC1 (Extended Data Fig 4c). Both CD21 - populations (P3 and P4) had high expression of atMBC signature genes including TBX21, FCRL4, NKG7 , ITGAX, and IL2RB 47, 57 as well as HCK and FGR, two SRC family kinases linked to integrin signaling (Extended Data Figure 4d) 58 . P3 and P4 also had high expression of gene sets related to atMBCs, as well as antigen presentation and processing and type I and type II interferon responses, which have been previously reported to be characteristic of atMBC (Figure 5j) 39, 59, 60 . In contrast, both CD21 + populations (P1 and P2) in tissues exhibited features of cMBCs including expression of CR2 and FCER2 and expressed high levels of the tissue-residence marker CD69 (Figure 5j, Extended Data Figure 4d) 59 . We then evaluated differences between the CXCR3 + and CXCR3 - cMBC (P2 and P1, respectively), which constituted the majority of the S1 + cells. Differentially expressed genes (DEG) in P1 versus P2 S1 + B SM revealed that cells in P2 expressed higher levels of TCF7, a transcription factor associated with stem-like central memory fate and found in cMBCs 36, 39 (Figure 5k, Supplemental Table 9). Upon secondary challenge, B SM in the lymphoid tissue can differentiate to PB/PC or GCBs based on their transcriptomic and epigenetic profile 61 . DEG in P1 versus P2 S1⁺ B SM suggested distinct cell fates upon rechallenge. P2 S1 + B SM expressed higher FOS , an AP-1 transcription factor family member expressed highly in pre-plasmablasts 62 , and expressed lower levels of BACH2 , a transcription factor that represses BLIMP1 expression and plasma cell differentiation and is found in memory B cells that preferentially differentiate to GCBs 61 (Figure 5k). S1 + P2 B SM also expressed lower IL21R , which is critical for GCB differentiation, maintenance, and proliferation 63, 64, 65 and the GCB marker MEF2B (Figure 5k) 66 . Moreover, pathway analyses revealed enrichment of genes involved PC differentiation, translation, interferon response, antigen presentation and oxidative metabolism pathways, with lower cell cycle associated genes in P2 compared to P1 S1 + B SM (Figure 5l). Together, these differences suggest that P2 B SM are better poised to differentiate to plasma cells rather than GCBs compared to P1. We further compared the epigenetic profiles of sorted bulk P1, P2, P3 and P4 B SM from three tonsils using ATAC-seq. Again, CD21 + (P1 and P2) and CD21 - (P3 and P4) populations clustered distinctly on PCA (Extended Data Figure 4e), with P3 and P4 enriched in motifs related to AP-1 transcription factors (FOS, BATF, JUN) that are associated with atMBCs and PC/PB differentiation 66 (Extended Data Figure 4f). Direct comparison of the cMBC groups, P1 and P2, revealed that P1 had greater overall open chromatin regions compared to P2 with a few notable exceptions including the TCF7 locus, which was more accessible in P2, consistent with the higher TCF7 expression noted in the transcriptome (Extended Data Figure 4g, Supplemental Table 10). The enriched signaling pathways associated with differentially accessible regions in P1 versus P2 mirrored those identified in the transcriptomic analysis, including those downstream of IFN-g, BCR, and antigen receptors (Figure 5l, Extended Data Figure 4h). IRF4 and PRDM1 encode transcription factors essential for PC/PB differentiation 67 ; chromatin accessibility at these loci in P1 and P2 B SM were similar and distinct from plasma cells and P3/P4 (Extended Data Figure 4i). However, differentially accessible regions in P2 compared with P1 had a greater proportion of AP-1-IRF composite (AICE) and interferon sequence response element (ISRE) binding motifs, which are IRF4-binding complex motifs reported to favor plasma cell differentiation. In contrast, the proportion of the Ets-IRF composite (EICE) motif, which is also a IRF4 binding complex motif but not involved in plasma cell differentiation, was comparable between P2 and P1 enriched peaks 67, 68 . This enrichment pattern suggests the P2 B SM population is primed for IRF4-mediated transcriptional changes driving plasma cell differentiation. Our findings suggest that among the cMBCs, P2 B SM are transcriptionally and epigenetically predisposed to antibody-secreting cell differentiation upon antigenic re-exposure compared to P1 B SM . Infection induces greater clonal expansion among S1 + B cells in pharyngeal tissues To further understand differences between VAC and INF responses, we assessed isotype, somatic hypermutation, clonal expansion, and clonal overlap by BCR sequencing of the sorted S1 + and S1 - B cells characterized by CITE-seq. Single-cell BCR sequencing confirmed that the majority of S1 + B cells were IgG1-class switched cells in both VAC and INF tissues, with a greater proportion of IgA1 expressing cells across all B SM subsets (P1-P4) post-infection (Extended Data Figure 5a). A greater proportion of S1 + B cells were part of expanded clones in INF compared to VAC tissues (Figure 6a). S1 + B cells from INF tissues displayed lower clonal diversity compared to S1 - B cells and also to S1 + B cells in VAC tissues, suggesting greater antigen-driven expansion in tissues post-infection compared to post-vaccination (Figure 6b). Furthermore, in INF tissues, P2 S1 + B SM had a trend towards a higher level of expansion than P1 S1 + B SM ; this expansion was not seen in P2 S1 + B SM post-vaccination (Figure 6c). In tonsil, atMBCs have been noted to have slightly lower levels of SHM 59 ; we also found that SHM frequency was lower among bulk S1 - P3 B SM in the tissues. However, SHM frequency was comparable across all four P1-P4 subsets of S1 + B SM in both INF and VAC (Figure 6d), revealing distinctions between SARS-CoV-2-specific cells compared to the bulk. Indeed, SHM frequency and heavy chain CDR3 amino acid length and distribution among S1 + IgG + B SM and GCBs were comparable between INF versus VAC tissues (Extended Data Figure 5b-d). We then assessed BCR clonal overlap to evaluate B SM development. Using STARTRAC (single T cell analysis by RNA sequencing and TCR tracking) pairwise transition index (pTrans) 69, 70 , which assesses developmental links in lymphocyte populations using clonal sharing, we found that antigen-specific S1 + P2 B SM exhibited high pTrans scores with S1 + P1 B SM and S1 + GCB (Figure 6e). Although S1 + P3 B SM were most strongly connected to GCBs, their connectivity was less strong than S1 + P2 B SM (Figure 6f). Bulk S1 - P2 B SM also had a strong association to S1 - GCBs, whereas S1 - P3 B SM were most strongly associated with S1 - P4 B SM and PC/PB, further supporting distinct connections and development among these cell populations (Extended Data Figure 5e-f). We then looked further at BCR sequences from the tissues of one INF subject from whom we sorted a large number of S1 + B cells (CNMC 70). By Jaccard index scores, we observed clonal overlap among S1 + P1, P2, P3, and P4 B SM populations, with particularly strong overlap between P1 and P2. S1 + P1 and P2 B SM also had strong clonal sharing with GCBs, whereas S1 + P3 and P4 B SM had GC overlap but to a lower extent than S1 + P1 and P2 B SM (Figure 6g). Evaluation of clonal trees showed S1 + B SM bearing various phenotypes (P1-P4, GCB) and isotypes emerging from a common ancestor in both INF (Figure 6h) and VAC (Figure 6i) tissues, implying broad differentiation potential or plasticity in B cell phenotypes during clonal expansion. Several clonotypes were also shared among adenoid and tonsil S1 + B cells, particularly in P1 and P2 B SM and GCB populations, supporting an immunologic connection between these two oropharyngeal lymphoid tissues, as we previously reported 23 (Figure 6h-i, Extended Data Figure 5g). Thus, although similar levels of SHM were noted post-infection and vaccination, infection induced greater antigen-specific B cell expansion in the tissues, particularly among P2 B SM , which likely have a GC origin. CXCR3 + CD21 + B SM are more responsive to BCR stimulation and poised to differentiate to plasmablasts To determine whether these differences reflect functional differences among P1-P4 B SM populations, we evaluated responsiveness to BCR stimulation, proliferation capacity, and ability to differentiate to antibody-secreting cells. Tonsil cells from a separate group of pediatric subjects who were either post-infection or had hybrid immunity to SARS-CoV-2 (both infected and vaccinated, Supplemental Table 2) were stimulated with soluble anti-human IgA, IgG, and IgM for 2 minutes and phosphorylation of the tyrosine kinase Syk and phospholipase Cg2 (PLC-g2), two downstream signaling proteins rapidly phosphorylated upon BCR stimulation, were measured 71 . As noted previously, the CD21 lo B SM populations (P3 and P4) were poorly responsive to BCR stimulation with soluble immunoglobulin 72, 73 (Figure 7a). In contrast, the P2 B SM population had the strongest induction of p-Syk and/or p-PLCg2 among total B SM and S1 + B SM (Figure 7a-b, Extended Data Figure 6a-b). This phenotype was not limited to B SM , as the CXCR3 + CD21 + (P2) subpopulation also had the strongest response to BCR stimulation among GCBs (Extended Data Figure 6c). In general, as B cells transition to PC/PB, membrane-associated IgG declines due to altered splicing, generating secreted IgG 36, 74 . We analyzed surface IgG and IgA expression (geometric mean fluorescence intensity or gMFI of anti-IgG and anti-IgA antibody staining) among B SM and PC/PB in our flow cytometry data. As expected, CD27 hi CD38 hi PC expressed marked reductions in surface IgG compared to B SM in both the peripheral blood and tissues (Figure 7c). Surface IgA was also reduced on plasma cells in peripheral blood, but to a lesser extent (Figure 7c) and was not reduced on tissue PC compared to GCBs and tissue B SM (Figure 7d). These observations are consistent with findings of functional surface IgA on gut and bone marrow PC in humans 75 and suggest that surface IgA may also play a functional role on PC in the oropharyngeal lymphoid tissue. We then compared surface IgG expression on the two major S1 + RBD + B SM populations, P1 and P2, as a surrogate for PC differentiation 36 . In INF tissues and blood, S1 + RBD + IgG + P2 B SM had lower surface immunoglobulin expression than their P1 counterparts, suggesting the SARS-CoV-2-specific P2 B SM post-infection are more prone to differentiate to PC (Figure 7e). Significant differences were not noted for P1 and P2 in VAC tissues, although a trend was seen in the tonsils (Figure 7e). Moreover, in the tonsil, S1 + RBD + IgG + B SM from INF subjects had lower surface immunoglobulin expression compared to VAC subjects (Extended Data Figure 6d); a similar trend was noted in the adenoid, suggestive of a propensity of tissue B SM induced by infection to differentiate to PC/PB. We then directly assessed the ability of P1 and P2 B SM from tissues to differentiate to plasmablasts and proliferate in vitro . We sorted bulk P1 and P2 B SM cells from tonsils of three subjects with hybrid immunity and labelled them with carboxyfluorescein succinimidyl ester (CFSE). We cultured the labelled P1 and P2 B SM populations with B cell-depleted PBMCs of an unrelated donor, R848 (TLR7/8 agonist), and IL-2 and we assessed proliferation index and percentage of plasmablasts in culture using flow cytometry after 4 days. Due to low cell numbers and poor viability after sorting, we were unable to reproducibly culture and assess P3 and P4 B SM . Compared to P1, a greater proportion of P2 B SM differentiated to CD38 + CD20 lo plasmablasts (Figure 7f). P2 B SM also exhibited a higher proliferation index in culture (Figure 7g). Thus, P2 B SM , which were enriched among INF samples, were better primed for plasma cell differentiation compared to P1 B SM , which may result in greater protective antibody production in the upper respiratory tract tissues upon re-exposure to SARS-CoV-2 in infected individuals compared to vaccinated individuals. Infection, but not vaccination, is associated with persistent immunologic activation in pharyngeal tissues We previously found that SARS-CoV-2 infection early in the pandemic induced persistent expansion of CD4 + and CD8 + T cell populations associated with antiviral, GC and IFN-g responses in the tonsils and adenoids compared to uninfected controls, with the most significant effects in adenoid 23 . However, whether vaccination also induces long-lasting immunologic changes in the upper respiratory tract lymphoid tissues is unclear. Consistent with our previous findings, a higher proportion of CXCR3 + CCR6 - CD57 + PD-1 hi CD4 + T cells and CXCR3 + CCR6 - pre-Tfh cells were noted in INF compared to CON adenoids (Figure 8a; Extended Data Figure 7a). In contrast, VAC adenoids did not exhibit changes in these T cell populations compared to CON (Figure 8a). COVID-19 induces persistent activation and epigenetic remodeling of innate immune cells including peripheral blood monocytes and dendritic cells well into convalescence 76, 77 . The frequency of CD14 + monocytes/macrophages, conventional dendritic cells (cDC), and plasmacytoid dendritic cells (pDC) did not differ among the VAC, INF, and CON tissues (Extended Data Figure 7b). However, these innate immune populations expressed higher HLA-DR in INF adenoids compared to VAC (and trended toward higher expression compared to CON), revealing persistent myeloid cell activation in adenoids post-infection but not post-vaccination (Figure 8b). Unlike the tissues, peripheral blood CD14 + monocytes/macrophages and cDC in INF subjects did not express higher HLA-DR post-infection, highlighting localized differences in the impact of infection in upper respiratory tract tissues compared to blood (Extended Data Figure 7c). However, the antiviral type I interferon-producing and CXCR3-expressing pDC population was slightly more activated in both INF blood and tissues (Figure 8b, Extended Data Figure 7c) 78 . Thus, convalescence was associated with local activation of innate cells in tissues, which may reflect differences in the primary lymphoid site of antigen-presentation and GC response and/or antigen persistence in infection versus vaccination. We then assessed correlations between populations of interest. The proportion of P2 cells among S1 + RBD + B SM was positively associated with the proportion of CXCR3 + CCR6 - pre-Tfh cells, an IFN-g producing population that is involved in GC reactions and important for CXCR3 + B RM development in mice 23 , 54 (Figure 8c). The frequency of the CXCR3 + CCR6 - CD57 + PD-1 hi Tfh cell population also significantly correlated with the percentage of P2 S1 + RBD + B SM (Figure 8c): these CXCR3 + Tfh populations also correlated with CXCR3 + total and/or S1 + RBD + B SM (Extended Data Figure 7d). Moreover, the proportion of P2 B SM and CXCR3 + pre-Tfh cells were also positively associated with the activation levels of monocytes/macrophages, cDC, and pDC, pointing to the potential relevance of innate cell activation in development or accumulation of CXCR3 + lymphocytes (Figure 8c). Thus, infection is associated with prolonged activation of innate or expansion of adaptive immune populations in tonsils/adenoids, which may contribute to the generation of CXCR3 + B SM . CXCR3 + lymphocytes and CXCL9 -expressing myeloid cells are in close proximity in the interfollicular region Using multicolor immunofluorescence imaging, we compared the location of CXCR3 + CD20 + B and CD4 + T cells and to their CXCR3 - counterparts in 4 tissue samples, derived from tonsils and adenoids of one INF and one VAC subject (Figure 8d, Supplemental Table 2). CXCR3 + BCL6 - CD20 + B cells and CXCR3 + BCL6 - CD4 + T cells were primarily located in the extrafollicular region (Figure 8d, Extended Data Figure 7e-f). Compared to their CXCR3 - counterparts, a smaller portion of CXCR3 + BCL6 - B cells were in follicles, which includes the GC and surrounding mantle (Extended Data Figure 7e). A similar pattern of distribution was seen among CXCR3 + CD4 + T cells (Extended Data Figure 7f). CXCR3 + BCL6 - CD4 + T cells and CXCR3 + BCL6 - CD20 + B cells appeared co-localized or aggregated near one another compared to CXCR3 - BCL6 - CD4 + T and CXCR3 - BCL6 - CD20 + B cells which were more diffusely distributed in the extrafollicular area (Figure 8d and e). We then assessed spatial proximity of CXCR3 + and CXCR3 - B and CD4 + T cells using spatial patterning analysis of cellular ensembles (SPACE), which assesses spatial relationships among multiple cell populations 79 . Peak areas of abundance of BCL6 + and BCL6 – populations were distinct, clearly delineating GCs (Figure 8f). Within BCL6 - regions, the CXCR3 + BCL6 - CD20 + B cell and CXCR3 + BCL6 - CD4 + T cell peaks were in close proximity to each other; similarly, in BCL6 + regions, which represent GCs, the CXCR3 + BCL6 + CD20 + B cell and CXCR3 + BCL6 + CD4 + T cell were closely aligned, suggesting that CXCR3 + B and T cells were adjacent both within and outside GCs (Figure 8f). Moreover, compared to their CXCR3 - counterparts, CXCR3 + BCL6 - B and T cells were closer to BCL6 + populations, and in particular, closer to CXCR3 + BCL6 + B and T cells, further implying proximity and possible interactions among CXCR3 + Tfh, GCB, and B SM populations in the GC and T-B border regions (Figure 8f). We performed spatial transcriptomics analysis with the Xenium 5000 human gene panel of the same 4 samples from an adjacent section of tissue. With unsupervised clustering and annotation, we identified 25 cell types, which localized to particular regions (Figure 8g, Extended Data Figure 7g). CXCL9, -10, and -11 are ligands of CXCR3 and enable migration and retention of CXCR3-expressing lymphocytes in tissue 80 . In tonsils and adenoids, we found that macrophages/monocytes and dendritic cells were the primary expressors of CXCL9 , CXCL10 , and CXCL11, with CXCL9 being the most abundantly expressed (Extended Data Figure 7h). Moreover, these innate immune cell populations were located in the interfollicular area, where we detected expression of CXCL9 and where we found CXCR3 + B and T cells by immunofluorescence (Figure 8d, g-k). Thus, spatial analyses suggest that the CXCR3 and CXCL9/10/11 axis facilitates interaction of innate and adaptive immune cells in lymphoid tissue, which may shape both the position and characteristics of memory B cells during viral infection. DISCUSSION Maintenance of virus-specific adaptive immune memory in the respiratory mucosa is important for immune protection against respiratory viral infections including COVID-19.By assessing tonsils and adenoids of children undergoing tonsillectomy/adenoidectomy, we found SARS-CoV-2-specific B cells, including B SM , B RM and a few GCBs, in these upper respiratory tract lymphoid tissues following vaccination. Distribution of B SM across tissues, including mucosal tissues, after intramuscular vaccination has been noted 51, 81, 82 . However, whether and how these cells differ from memory cells found post-infection has not been clear. Using high-dimensional analyses, we show that immune memory is maintained in mucosal lymphoid tissues after intramuscular COVID-19 immunization, but reveal distinct features compared to those generated post-infection. In particular, we identified a subset of cMBC defined by CXCR3 + CD21 + , that was enriched in infected individuals, with increased propensity for plasma cell differentiation and distinct localization within mucosal lymphoid tissues. This population may be responsible for some of the qualitative and clinical differences in humoral mucosal immunity observed with infection and vaccination. Prior studies have found SARS-CoV-2-specfic GC and Tfh cells in the draining axillary lymph nodes following mRNA vaccination 17, 18, 19, 20, 83 , but no SARS-CoV-2-specific GC B cells were noted in a small number of contralateral lymph nodes 17 . However, spike-reactive GCBs have been noted in lung-associated lymph nodes of a few (2/18) vaccinated (and uninfected) organ donors 51 and tonsils of a few vaccinated adults, although whether these latter subjects were truly SARS-CoV-2 infection-naïve is unclear since some had B cells reactive to NC 60 . To address this concern, we measured both anti-NC antibodies and anti-ORF8 antibodies, the latter of which are more sensitive than antibodies to NC 24 . We too noted the presence of spike-specific GCBs in some vaccinated participants in distal mucosal lymphoid sites, although fewer than among infected subjects. Whether these result directly in response to immunization, from cross-reactivity to common cold coronaviruses, or exposure to SARS-CoV-2 that did not result in a productive infection, perhaps due to the presence of high affinity antibodies generated by vaccination 84, 85 , remains an interesting question. Although COVID-19 mRNA vaccines have high vaccine efficacy and prevent hospitalization and death from COVID-19, epidemiologic studies suggest that infection provides longer-lasting immunity than vaccination 6, 7, 8 . This may result from differences in the route of viral antigen exposure to the host’s immune system (intramuscular vs. inhaled), breadth of antigenic exposure (spike only vs. all viral antigens), nature of antigens presented (lipid nanoparticles vs. virions) or duration of antigen exposure. These factors, in turn, affect host immune responses, and, indeed, we found that infection and vaccination drive the generation of phenotypically different SARS-CoV-2-specific B cells, suggesting unique B SM developmental pathways depending on the type of exposure. The clonal expansion and persistence of CXCR3 + B SM post-infection suggest that IFN-g, induced by viral infection and likely produced by activated and expanded innate and Tfh cells in the tissue, may be a driving force for these developmental differences in B cell memory. The CXCR3 + P2 B SM population expressed low levels of Tbet, which may reflect prior IFN-g exposure during infection, and were distinct from atMBCs which expressed high levels of Tbet. Our results suggest that CXCR3 expression affords two key characteristics incMBCs that may contribute to enhanced mucosal protection post-infection. First, CXCR3 promotes homing and retention of B SM cells, particularly IgA + cells, to mucosal lymphoid tissues, where CXCR3 + cells bear B RM markers, as supported by murine studies 52, 54 . Second, we found that among cMBCs, CXCR3 + CD21 + B SM have a predilection for plasma cell differentiation upon secondary challenge compared to CXCR3 - CD21 + B SM, enabling a more robust antibody response upon secondary challenge. Although CXCR3 + B SM were enriched in the tissues, the difference in CXCR3 expression between INF and VAC was also clear in PBMCs suggesting CXCR3 may be a useful marker for evaluating potential for trafficking even in blood; this is supported by our recent findings in blood of vaccinated adults with hybrid immunity 86 , indicating the applicability of our findings across age groups. Moreover, although our study focused on mRNA COVID-19 vaccines, another recent study found the Ad26.COV2.S, elicited more CXCR3 + B SM without atMBC features compared to mRNA vaccines 5 , suggesting similar cells may be generated by viral vaccine platforms. Nonetheless, the mRNA component of mRNA-lipid nanoparticle vaccines can stimulate IFN-g production by Tfh cells 87 , potentially explaining the presence of some CXCR3 + B SM in the blood and tissues of vaccinated subjects in our study. Of note, the frequency of the P2 population among B SM did not correlate with serum neutralizing titers, suggesting this population reflects other aspects of immunity induced by infection and that neutralizing titers cannot be used as a proxy for this population. Conversely, we found that mRNA vaccination induced a greater proportion of virus-specific CD21 lo and particularly CXCR3 - B SM populations than infection in both blood and tissues 29, 38, 60 . Although some studies suggest that CD21 lo atMBCs have lower degrees of somatic hypermutation than cMBC and emerged from extrafollicular reactions particularly in the context of chronic viral or autoimmune conditions 39, 45 , we found that SARS-CoV-2-specific CD21 lo populations in tonsils/adenoids had comparable levels of SHM as CD21 + B SM , highlighting the importance of assessing antigen-specific versus bulk cell populations and evaluating cells in patients without conditions characterized by strong extrafollicular reactions. The frequency of CD21 -/lo B SM in the blood declines with time after influenza or COVID-19 vaccination, suggesting these cells may be shorter lived 34, 35 . It is possible that the enrichment of SARS-CoV-2-specific CD21 - cellspost-vaccination accounts for the steeper decline in immunity months after mRNA vaccination compared to infection. Nonetheless, FCRL5 + Tbet + memory B cells with atMBC features have been found 1 year after influenza vaccination and contributed significantly to serum antibody levels upon secondary challenge; similarly FCRL4 lo atMBCs were present at least 6 months post-influenza vaccination, suggesting the picture is more complex 36, 37 . We also note that vaccination engendered broader variant coverage, as determined by cross-reactivity to omicron, highlighting an important feature induced by mRNA vaccination. Thus, the plasticity and function of different memory populations following vaccination versus infection remain important questions. As humans are confronted with new pathogens, the development of vaccinations for respiratory pathogens that elicit strong mucosal immunity, which may reduce mucosal shedding and transmission, is a priority. Although intramuscular SARS-CoV-2 mRNA vaccines generate rapid and broad immunity, they elicit low neutralizing antibody levels in the respiratory tract 30, 31 . Trials of intranasal vaccines in humans are ongoing and have shown mixed results 88, 89 . Nonetheless, pre-clinical studies of inhaled vaccine boosters following intramuscular mRNA vaccines in murine models showed enhanced mucosal protection due to trafficking of antigen-experienced B SM to the respiratory tract through the CXCL9/10-CXCR3 axis and enhanced differentiation to IgA-producing plasma cells 48 . Our study underscores the significance of this axis in homing and shaping immunity in the upper respiratory tract of humans. In summary, our study provides evidence for the generation and maintenance of mucosal B cell memory after mRNA vaccination and further provides a framework for evaluating immunity in the upper respiratory tract and the blood following immunization. LIMITATIONS Although study participants were heterogeneous in terms of their age, tonsil condition, time since vaccination/infection, and doses of immunization, we tried to control for age as a covariate using a linear model. Moreover, the INF cohort, which was collected in 2020-2021, had similar times from their last known immunologic exposure to surgery as the VAC cohort and were exposed to early circulating strains with spike proteins similar to the strain in the first-generation mRNA vaccines, helping control for these factors. Nonetheless, we only know the time of infection for approximately half of the infected subjects. As our study is a cross-sectional study, we are unable to assess temporal changes in immune responses over time. We also note that our INF cohort only included patients with mild or asymptomatic infection; severe infection may alter or compromise the humoral immunity generated by infection 90 . Lastly, due to the increasing percentage of infected individuals, our sample size of vaccinated only individuals was limited, which could affect our ability to fully assess correlations between clinical characteristics and immune profiles. Declarations Acknowledgements: We thank the patients and their families for their generous participation; Pedro Milanez-Almeida, Edward Schrom, Sebastian Wellford, Weiming Yu, Daniel Newman, Hiroshi Ichise, Julie Reilley, and members of the Schwartzberg lab at National Institutes of Health (NIH) for advice, discussions, and/or technical assistance; Division of Otolaryngology at Children’s National Hospital for helping with participant recruitment; and National Cancer Institute CCR Genomics Core for sequencing.This research was supported by the Intramural Research Program of the NIH. The contributions of the NIH author(s) were made as part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered Works of the United States Government. However, the findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services. The antibody response study was supported by FDA’s Perinatal Health Center of Excellence (PHCE) project grant #GCBER005 and #GCBER008 to S.K. This work utilized the computational resources of the NIH Biowulf and NIH NIAID Skyline clusters. K.H. was supported by NIAID grant R00AI159302. Author contributions: K.M., P.M., P.L.S, and Q.X. conceived and designed the study. Q.X., K.M., performed flow cytometry experiments. L.B., G.G., S.P., J. T., and S.K. performed serologic testing and analysis. Q.X. analyzed the flow cytometry data and performed statistical analyses of these data. Q.X., K.M., P.L.S, CHI, C.L., K.H. and A.J.M. designed, performed analyzed or advised CITE-seq experiments. L.S., J.K., M. S., Q.X., and K.M. performed immunofluorescence staining, obtained images, and/or analyzed data. T.E.M., A.K., D.P.G., and Q.X. performed and/or analyzed ATACseq. Q.X., K.M., R.S., and L.S. obtained or analyzed Xenium data. K.M., P.M., and H.B. developed patient recruitment materials and/or recruited participants. Q.X., P.M., C.L., S.K., L.K., C.M.B., R.S., J.S.T., S.M., P.L.S., and K.M. provided critical scientific input and/or reagents. K.M., Q.X, and P.L.S. prepared the manuscript. All authors contributed to the final review and editing of the paper. Declaration of interests : The authors have no competing interests. Data availability: All data necessary to understand and evaluate the conclusions of this paper are provided in the article and Supplementary Information. The CITE-seq and ATAC-seq data will be deposited to dbGAP (the study is in the process of getting registered). Source data will be provided with the final published manuscript. Code availability: The R scripts used in this paper are available on a GitHub repository at https://github.com/kalpanamanthiram/COVID-19-Vaccination. METHODS Participant recruitment This study was approved by the Institutional Review Board (IRB) at Children’s National Hospital (IRB protocol number 00009806). Written informed consent was obtained from parent/guardians of all enrolled participants, and assent was obtained from minor participants over 7 years of age. We recruited 21 children who underwent tonsillectomy and/or adenoidectomy at Children’s National Hospital (CNH) in Washington, DC, USA and also received at least one dose of a first generation COVID-19 mRNA vaccine. The first 17 participants were recruited from December 2021 to April 2022, and the remaining were recruited from August 2022 to September 2022. Because not all tissues or blood were available from each subject, we collected a total of 21 blood samples, 19 adenoids, and 18 tonsils from these 21 participants. No statistical methods were used to predetermine sample size. All participants had negative RT-PCR testing from a nasopharyngeal swab for SARS-CoV-2 within 72 hours of surgery. Demographic information and clinical data were collected through parental questionnaires and chart review and inputted and managed in REDCap (https://project-redcap.org/), and biologic samples were acquired in the operating room by the clinical team at CNH. Participants with no evidence of prior infection (negative for anti-nucleocapsid (NC) or anti-open reading frame 8 (ORF8) serum antibodies and/or no clinical history of PCR/antigen confirmed SARS-CoV-2 infection) were classified as the VAC group (vaccinated only, 10 subjects). Samples from the remaining 11 subjects with hybrid immunity (both post-vaccinated and infection, 11 subjects) and additional post-infection subjects without vaccination who were recruited in 2022/2023 (12 subjects) were used in some assays. We also used samples obtained from our prior study of 24 children infected with SARS-CoV-2 from 2020 to 2021 (INF), prior to the availability of vaccines for children. Collection of these samples was previously described. 23 Prior infection in these individuals was confirmed by serologic testing for anti-spike antibodies and/or the presence of SARS-COV-2-specifc B cells in the blood, tonsil, or adenoid by flow cytometry. From these 24 children, 15 PBMCs, 14 adenoids, and 22 tonsils were analyzed based on availability of cells (Supplementary Table 2). Blood and tissue collection Blood samples were obtained just prior to the surgical procedure in the operating room in serum separator tubes (BD) for serum collection and sodium heparin tubes (BD) for peripheral blood mononuclear cells (PBMCs) extraction from an intravenous line placed for anesthesia. Once received in the laboratory on the day of collection, serum separator tubes were spun at 1200 g for 10 min, and serum was aliquoted and stored at -80°C. PBMCs were isolated the day after collection by density gradient centrifugation (Lymphocyte Separation Medium, MP Biomedicals) at 1500 rpm for 30 min at room temperature (RT) with no brake and washed with PBS. If red blood cell contamination was present, cells were lysed with ammonium-chloride-potassium (ACK) buffer (Gibco). Tonsil and adenoid tissues were stored in RPMI-1640 (RPMI) media with 5% heat-inactivated fetal bovine serum (FBS, VWR), gentamicin 50 mg/mL (Gibco), and 1 × antibiotic/antimycotic solution (Gibco) on ice immediately after collection. Tissues were processed the day after collection. The tissue was mechanically disrupted and filtered through a 100 μm cell strainer to create a single cell suspension, lysed with ACK buffer, and washed with PBS three times. Cells were then stored in liquid nitrogen in the presence of FBS with 10% DMSO and thawed according to our published protocol 92 for flow cytometry, CITE-seq, and functional assays below. Seroreactivity of samples to SARS-CoV-2 spike, nucleocapsid and ORF8 by ELISA 96 well Immulon plates were coated with 20 ng/100 µL of recombinant nucleocapsid, spike RBD protein, or Open Reading Frame 8 (ORF8) from WA1/2020 in PBS overnight at 4 o C. Starting at a 1:20 dilution, serum samples were serially diluted 5-fold and applied to the coated well for 1 hr at ambient temperature. Serum samples were assayed in duplicate, as described before 13, 93 . After three washes with PBS/0.05% Tween 20, bound human IgG antibodies were detected with 1:5000 dilution of HRP-conjugated anti-human IgG Fc-specific antibody (Jackson Immuno Research). After 1 hr, plates were washed with PBST followed by PBS, and o-Phenylenediamine dihydrochloride (OPD) was added for 10 min. Absorbance was measured at 492 nm. End-point titer was determined as 3-fold above the average of the absorbance values of the binding of serum samples to blank control wells. The end-point titer is reported as the serum dilution that was above this cutoff and was calculated using Prism 9 (GraphPad Software). SARS-CoV-2 serum neutralization assay Samples were evaluated in a qualified SARS-CoV-2 pseudovirion neutralization assay (PsVNA) using SARS-CoV-2 WA1/2020 strain and Omicron BA.1 subvariant. SARS-CoV-2 neutralizing activity measured by PsVNA correlates with PRNT (plaque reduction neutralization test with authentic SARS-CoV-2 virus) in previous studies 94, 95, 96 . Neutralization assays were performed as previously described 13, 93, 95, 96, 97 . Briefly, 50 µL of SARS-CoV-2 S pseudovirions (counting ~200,000 relative light units) were pre-incubated with an equal volume of medium containing serial dilutions (starting at 1:10) of all samples at RT for 1 hr. Then, 50 µL of virus-antibody mixtures were added to 293T-ACE2-TMPRSS2 cells (10 4 cells/50 μL) in a 96-well plate. The input virus with all SARS-CoV-2 strains was the same (2 × 10 5 relative light units/50 µL/well). After a 3 hour incubation, fresh medium was added to the wells. Cells were lysed 24 hour later, and luciferase activity was measured using One-Glo luciferase assay system (Promega). The assay of each sample was performed in duplicate, and the 50% neutralization titer was calculated using Prism 9 (GraphPad Software). The limit of detection for the neutralization assay is 1:20. Two independent biological replicate experiments were performed for each sample and variation in PsVNA50 titers was <10% between replicates. High dimensional flow cytometry SARS-CoV-2 specific B cell characterization with spectral flow cytometry 5 million cells per sample of PBMC, adenoid, or tonsil were resuspended in PBS with 2% FBS and 2 mM EDTA (FACS buffer). Biotinylated probes to SARS-CoV-2 were crosslinked with fluorochrome-conjugated streptavidin in a molar ratio of 4:1. Fluorochrome-conjugated streptavidin was split into 5 aliquots and conjugated to biotinylated probes by mixing for 20 min/aliquot at 4 °C. Four fluorochrome-conjugated probes were used in this assay, including spike S1 from the Wuhan strain (BioLegend) conjugated to APC, RBD from the Wuhan strain (BioLegend) conjugated to BV421, RBD from the Omicron variant (Acro) conjugated to BUV615, and RBD from the Omicron variant (Acro) conjugated to PE. Cells were first stained with the viability dye, Zombie NIR (1:800 dilution, BioLegend), for 15 min at RT, washed twice and then incubated with the 4 fluorochrome-conjugated probes plus d-biotin (Avidity) and Brilliant Stain Buffer Plus (BD Biosciences) at 4 °C for 1 hr. Then, cells were washed twice, and resuspended with True-Stain Monocyte Blocker (1:10 in 50 mL) (BioLegend) for 5 min. Anti-CXCR3 antibody and an antibody cocktail containing the rest of the surface antibodies (Supplementary Table 11) and Brilliant Stain Buffer Plus were sequentially added directly to the cells and incubated for 5 min and 30 min at RT, respectively (200 mL total staining volume). Cells were washed three times and fixed in 1% paraformaldehyde for 20 min at RT before washing again and collecting on a spectral flow cytometer (Aurora, Cytek). Antibodies are listed in Supplementary Table 11. Gating strategies are shown in Supplementary Figure 1 – 2. A separate panel including intracellular transcription factor staining was also used to stain PBMC and tonsil cells 92 . Antibodies are listed in Supplementary Table 11. Immunophenotyping with 37 color spectral flow cytometry panel Two million PBMCs per sample and 5 million cells per adenoid or tonsil were resuspended in FACS buffer after thawing. Cells were stained and acquired as described in our prior study 23 . Antibodies are listed in Supplementary Table 11. Manual gating for both panels was conducted with FlowJo Software v.10.9.0 (BD Biosciences) as in Supplementary Figure 3 - 4. Unsupervised analysis of high dimensional flow cytometry data S1 + RBD + B cells from the SARS-CoV-2-antigen-specific B cell panel with 29 parameters were analyzed with unsupervised clustering of surface antibody staining. S1 + RBD + live CD45 + CD3 - CD14 - CD19 + B cells from INF and VAC groups were analyzed. Cells from tonsil and adenoid samples were merged and processed together, while PBMC samples were processed separately due to differences in cell populations. B cell analysis was based on surface expression of IgA, IgD, IgM, IgG, CD27, CD38, CD21, CD95, CD11c, FCRL3/FCRL5, FCRL4, CD10, CD86, CD83, CD69, CD71, CXCR3, CD85J, and CD62L. Channel values (scaled output with compensated parameters) of S1 + RBD + B cells from each sample were exported from FlowJo and then processed in R (v. 4.4.2) via Rstudio (2023.12.1+402). Each cell was given an index and labeled with its origin subject’s identification and tissue type. Data were further processed using Seurat 98, 99 (v. 5.1.0). Cell clustering was performed by applying the FindNeighbors() function 98 on a distance matrix generated from the dist function by “euclidean” method, followed by Leiden clustering on the resulting SNN graph using Seurat’s FindClusters() algorithm, with a resolution parameter of 1.0 in PBMC and 0.6 in tissues. Expression of selected markers was visualized with their mean expression in each cluster by pheatmap (v. 1.0.12), and the downstream analysis and results were processed using ggplot2 (v. 3.5.1) 100 , reshape2 (v. 1.4.4) 101 , ggpubr (v. 0.6.0) 102 , ggthemes (v. 5.1.0) 103 and tidyverse (v. 2.0.0). Statistical comparison with linear model The majority of participants underwent tonsillectomy for obstructive sleep disordered breathing due to tonsillar hypertrophy (which includes those with obstructive sleep apnea diagnosed with polysomnography); a smaller portion had recurrent or chronic tonsillitis as their primary indication. We and others have noted that age and indication for tonsillectomy influence the immune cell populations in these tissues 56, 104 ; however, none of the INF subjects had recurrent tonsillitis. This limitation prevented us from adjusting for this potential confounding factor. Sex had little effect in our cohort (evaluated with PCA analysis, data not shown). Cluster proportion, cell percentages from manual gating, or neutralizing titers in each sample were modeled linearly as a function of (1) age (“Age”), and (2) history of SARS-CoV-2 infection or vaccine exposure (“Group” includes two or three categories: VAC, INF, and CON depending on the comparison. Each comparison involves two categories at a time.). The following formula was used to estimate separate coefficients for each category ofGroup (adjusted for age): lm (Frequency/Percentage ~ Group + Age). To illustrate the extent of correction achieved by the linear model, both p -values from the two-sided Wilcoxon signed ranks test (in black) and p -values from linear model (in blue) were presented in plots built with ggplot2(v. 3.5.1) 100 . Single cell RNA sequencing Sorting of S1 + andS1 - B cells for CITE-seq We performed a combined analysis by merging data from three sets of cellular indexing of transcriptomes and epitopes (CITE-seq) experiments (B-2021, B-306-3 and B-306-4). B-2021 was performed and processed as reported in our prior study 23 . Briefly in B-2021, paired PBMC, adenoid, and tonsil samples from 3 donors (total of 9 donor-tissue samples) were assessed and ‘hashtag’ antibodies (BioLegend) were used to uniquely label each of the 9 donor-tissue samples 105 . Two additional sets of CITE-seq experiments were conducted separately to include VAC samples (B-306-3 and B-306-4). For B-306-3, 2 INF and 1 CON tonsil were processed. For B-306-4, 3 VAC and 4 INF adenoid and 2 VAC, 2 INF, and 1 CON tonsil samples were processed as three sets (B-306-4 A, B, C) on 3 separate days over the course of one week to ensure manageable sorting times on each processing day (See Supplementary Table 8 for details on which samples were processed with each experiment). Frozen tonsil and adenoid cells were thawed from liquid nitrogen as described previously 23, 92 . For each donor, 100,000-500,000 thawed tonsillar cells were reserved for bulk RNA-seq. During data analysis, individual cells were demultiplexed using donor-specific single nucleotide polymorphism (SNP) information obtained from the bulk RNA-seq data (see Methods: CITE-seq processing and demultiplexing). The remaining cells of the same tissue type from each donor were pooled together (i.e. adenoids and tonsils were pooled separately). The number of cells from each donor to pool was estimated using flow cytometry data with the aim of pooling a similar number of S1 + B cells from each sample. Pooled cells were incubated with Fc blocker at 4 °C for 10 min followed by CITE-seq and sorting antibody cocktails in the following order at 4 °C: TotalSeq anti-CXCR3, anti-CCR6, and anti-CXCR5 antibodies for 10 min, and the remaining 21 CITE-seq antibodies and fluorescence-labeled sorting antibodies and viability dye (Aqua) for an additional 30 min (Antibodies are listed in Supplementary Table 11). Cells were then washed with PBS with 0.04% BSA and sorted on a BD FACS Aria™ III sorter (BD Biosciences, San Jose, CA). S1 + B cells from each tissue pool were sorted into a 200 uL low binding PCR tube (Eppendorf) with 50 mL culture medium (10% heat-inactivated FBS (VWR), 2 nM glutamine, 0.055 mM 2-mercaptoethanol, 1% penicillin/streptomycin, 1 mM sodium pyruvate, 10 mM HEPES, 1% non-essential amino acids in RPMI with glutamine), while S1 - B cells were sorted into 1.5 mL low binding tubes (Eppendorf) with 300 ml RPMI culture medium with 20% heat-inactivated FBS. All collecting low binding tubes were precoated with RPMI media with 20% FBS overnight at 4 °C prior to use. Sorting strategy is shown in Supplementary Figure 5A. Antibodies, including barcoded and fluorescence-conjugated antibodies, are listed in Supplementary Table 11. S1⁺ cells were centrifuged and resuspended for single cell partitioning without counting. An appropriate number of S1⁻ B cells were counted, aliquoted, centrifuged and resuspended for partitioning. Antibody concentrations used for CITE-seq were inferred based on the concentrations of antibodies of the same clones titrated for use in our flow cytometry staining of tonsil/adenoid cells. CITE-seq processing and demultiplexing Library construction and sequencing for B-2021 were described previously 23 . For B-306-3 and B-306-4, sorted S1 + and S1 - B cells were mixed with the reverse transcription mix and partitioned into single cell gel-bead in emulsion (GEM) using 10x 5’ Chromium Next GEM Single Cell 5’ Reagent Kit v2 for B-306-3 and 10x 5’ Chromium Next GEM Single Cell 5' HT v2 for B-306-4 (10x Genomics). The reverse transcription step was performed in an Applied Biosystems Veriti 96-well thermal cycler (Applied Biosystems). 5’ single cell gene expression (GEX), cell surface protein (CSP), and B cell receptor (BCR) libraries were prepared as instructed by 10x Genomics user guides ( https://www.10xgenomics.com/resources/user-guides/ ; CG000330 Rev F and CG000424 Rev C for the B-306-3 and B-306-4 run, respectively). Library quality and quantities were measured using a TapeStation system (Agilent) and a Qubit fluorometer (ThermoFisher). Libraries were pooled at a concentration of 10 nM and sequenced on the Illumina platform (NovaSeqX for B-306-3 and B-306-4, Illumina) using the following read lengths: Read 1: 26 base pairs, Index 1: 10 base pairs, Index 2: 10 base pairs, Read 2: 90 base pairs. Tonsil cells saved for bulk RNA-seq from each participant were resuspended in QIAzol and RNA was extracted using RNeasy micro kit (Qiagen) and standard RNA sequencing libraries were generated using Universal Plus mRNAseq kit (TECAN Genomics). These libraries were used to generate SNP calls for each donor. Sequencing results were demultiplexed and converted to FASTQ format using Illumina bcl2fastq software. The sequencing reads were adaptor and quality trimmed and then aligned to the human genome using the splice-aware STAR aligner and SNP calls were called using bcftools 106 . The pooled single cell RNA sequencing data was demultiplexed with a combination of demuxalot (v. 0.4.1) 107 and demuxlet 108 to match cells to each donor and identify doublets. For the pool with fraternal twin subjects, CNMC 124 and CNMC 125, cells were assigned by demuxalot, then reads with associated barcodes were extracted and reassigned via demuxlet to further refine the cell identities for twin subjects. CITE-seq data analyses Data from B-2021 set were processed as described 23 . For B-306-3 and B-306-4, CellRanger (v. 7.1.0) 109 was used to map cDNA libraries to the GRCh38 genome reference (GRCh39-2020-A) and to count antibody tag features. Down-sampling was performed using the cellranger aggr pipeline (10x Genomics) to normalize sequencing depth across B cell lanes. Data were further processed using Seurat (v. 5.1.0) in R 4.4.2. Cells were demultiplexed by SNP as described above (Methods: Single cell CITE sequencing and demultiplexing). Surface protein library counts were transformed by using dsb (v. 1.0.4) 110 . For quality control, cells with less than 100 detected genes, greater than 30% mitochondrial reads, or gene counts greater than 25,000 were removed. To exclude cells with extremely high surface antibody counts, the top 0.5% of cells in the surface antibody total count distribution were removed. Cell clustering was performed by applying the FindNeighbors() function from Seuraton a distance matrix generated from the dsb-transformed surface protein data, followed by Leiden clustering on the resulting SNN graph using Seurat’s FindClusters() algorithm, with a resolution parameter of 0.3. Expression of selected genes were visualized using the ComplexHeatmap package (v. 2.22.0) 111 , and the percentage of cells per population for the S1 + and S1 - cells was plotted using ggplot 2 (v. 3.5.2) 100 . To further assess the transcriptome of B SM subsets P1 to P4, we manually annotated B cell populations using protein expression data from CITE-seq for three reasons: (1) to align our analysis with flow cytometry data; (2) CXCR3 protein and mRNA expression levels were poorly correlated (Extended Data Figure 4a); and (3) to minimize batch effects across different experimental sets. We used the removeBatchEffect function from the limma (v. 3.62.1) 112 package to correct batch effects in CSP expression within the B-306-4 set. CSP expression from B-2021, B-306-3, and B-306-4 were exported as .csv files and imported into FlowJo (v. 10.9.0) for manual gating. The gated FlowJo files were then processed using CytoML (v. 2.18.0) 113 and flowWorkspace (v. 4.18.0) 114 , and the resulting gates were merged with the CITE-seq metadata by matching identical cell barcodes. To identify differentially expressed genes among B SM P1 to P4 (Extended Data Figure 4d), we used only the GEX from B-306-4 to minimize batch effects. Differential expression analysis was performed using the FindAllMarkers function in Seurat with the MAST algorithm, incorporating 'Batch' (defined as samples processed on different days within the B-306-4, set A-C), 'Subject', and 'Tissue' as latent variables (min.pct = 0.1, min.cells.feature = 3, min.cells.group = 3, logfc.threshold = 0.1). The PseudobulkExpression function was used to normalize count data for each B SM population, yielding representative expression values for differentially expressed genes within each subset. The top 20 genes with adjusted p-values 0, as identified by the FindAllMarkers results, were visualized using pheatmap (v. 1.0.12) 115 . Pseudo-bulk and differential gene expression analysis Pseudo-bulk gene differential expression analysis and gene set enrichment analysis (GSEA) were performed as described previously 116, 117 . In brief, all sorted cells in a given cell type (S1 + or S1 - ) and tissue (adenoid or tonsil) within a donor were computationally ‘pooled’ according to their B cell subset assignment by summing all reads for a given gene. Pseudo-bulk libraries composed of only a few cells (less than 5), and with fewer than 30,000 unique molecular identifier counts after pooling per library were removed from the analysis, as they are likely not modeled properly by bulk differential expression methods. Only cell type- and tissue-specific B cell subsets with more than 3 psuedobulk libraries were included for differential comparison. Genes expressed at low levels were removed for each cell type individually using the filterByExpr function from edgeR 118 . Differentially expressed genes were identified using the limma voom 119 workflow, which models the log of the counts per million (CPM) of each gene. Scaling factors for library size normalization were calculated with the calcNormFactors function with method = ‘RLE’. Genes were ranked using the moderated T statistics for the relevant coefficient from the limma voom model. Differentially expressed genes between S1 + B SM P1 and S1 + B SM P2 cells were identified with a model accommodating paired analysis (formula: ~ 0+CD21_CXCR3+Subject_Tissue). The S1 + P1 and P2 B SM subset (“CD21_CXCR3”) and tissue-specific subject (“Subject_Tissue”) were modeled as factor variable representing the B SM population (including “B SM P1” and “B SM P2”), tissue type from a specific subject (including “tonsil” and “adenoid”), respectively. The contrasts.fit function was then used to compare the estimated means between S1 + P1 B SM and S1 + P2 B SM . Enriched gene sets were identified using the pre-ranked gene-set enrichment analysis (GSEA) algorithm implemented in the fgsea (v. 1.32.0) R package. Gene set lists used for enrichment assessment (including Gene Ontology Biological Process (GO BP), GO Cellular Component (GO CC), GO Molecular Function (GO MF), Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, the Molecular Signatures Database’s Hallmark collection, Blood Transcriptomic Modules and a few published datasets 59, 120, 121 ) were collected and pooled. P values were adjusted using the Benjamini–Hochberg method for the whole gene set list. The pathways shown in Figure 5l were manually curated from gene sets relevant to immunology. Signature scores of B cell subsets Gene sets for atMBC were collected from studies on various chronic diseases in humans and mice including malaria 122 , HIV infection 123 , systemic lupus erythematosus 124 , Sjögren's syndrome 125 , rheumatoid arthritis, common variable immunodeficiency 126 , and rodent malaria Plasmodium chabaudi 59 , as well as human tonsil 59 . Gene set for GCB signature was derived from a human tonsil B cell study 127 . cMBC signature was used from a recent study on human tonsil and murine malaria 59 . Other gene signatures used to characterize B cells were obtained from MSigDB 128 (msigdbr R package, v. 7.5.1). The AddModuleScore function of Seurat was applied with default parameters to score each signature in each B cell. To represent the levels of each signature set in the populations of interest, the median score for each tissue- and subject-specific sample within the S1 - B cell subset from the B-306-4 set was used as an estimate. The median scores of per-subject median scores for each B cell population were then displayed in the heatmap of Figure 5j. Trajectory analysis To compare the lineage differentiation of S1 + B cells in INF and VAC tonsils and adenoids, we applied Slingshot (v. 2.14.0) 129 to infer lineage trajectories. S1 + cells from B-306-4 set were used in this analysis. B cell subsets were assigned by manual gating as described in Figure 5 g and h. “PC” and “PreGC” cells were removed due to low number of cells for an accurate lineage estimation. First, uniform manifold approximation and projection (UMAP) embedding was performed on the normalized single-cell surface protein expression profile from the CITE-seq dataset to obtain a low-dimensional representation. P1-P4 B SM populations were provided as cluster input and “Naïve/USM B cell” population was appointed as the start cluster to Slingshot for trajectory reconstruction. Cells were ordered through inferred pseudotime based on gene expression to indicate their differentiation progress. Trajectories for INF and VAC groups were estimated separately. BCR sequence analysis and clonal clustering BCR repertoire sequence data were analyzed using the Immcantation ( www.immcantation.org , v. 4.5.0) framework. Starting with filtered CellRanger output, V(D)J genes for each sequence were aligned to the IMGT GENE-DB reference database obtained 2/09/24 using IgBLAST (v. 1.22.0) 130 and Change-O (v. 1.3.0) 131 . Nonproductive sequences, cells without associated constant region calls, cells identified as arising from doublets or negative droplets, and cells with multiple heavy chains were all removed. Samples within each subject were pooled and sequences were grouped into clonal clusters, which contain B cells related to each other by somatic hypermutation (SHM) from a common V(D)J ancestor. Using the hierarchicalClones function of scoper (v. 1.3.0) 132 , sequences within these groups differing by a participant-specific normalized Hamming distance threshold within the CDR3 region were defined as clones using single-linkage clustering (one subject required a slightly lower threshold than others) 133 . This threshold was determined by fitting a gamma/Gaussian mixture model to the distance to nearest sequence neighbor distribution using SHazaM(v. 1.2.0) 134 . These heavy chain-defined clonal clusters were further split if their constituent cells contained light chains that differed by V and J genes. Within each clone, germline sequences were reconstructed with D segment and N/P regions masked (replaced with “N” nucleotides) using the createGermlines function within dowser (v. 2.3) 135 . SHM was calculated as the frequency of non-ambiguous mismatches from each cell to the V gene (IMGT positions 1–312) of its reconstructed germline sequence. Paired scBCR-seq data were integrated with CITE-seq data based on matched cell barcodes. To quantify B cell clonal diversity, we calculated Simpson’s diversity for each tissue-specific sample using the alphaDiversity function of alakazam (v. 1.3.0) 131 . Lower values of Simpson’s diversity indicate a greater probability of two random sequences belonging to the same clone, consistent with more large clones. To account for differences in sequence depth, samples within each comparison were down-sampled to the same number of sequences, and the mean of 1000 such re-sampling repetitions was reported. Subject/tissue/cell type samples or populations with < 70 B cells were excluded, which led to the exclusion of all S1 + cells from CNMC 10 and CNMC 99 (both CON with no history or evidence of SARS-CoV-2 infection or vaccination). Clonal overlap among tissues or B cell subsets can be used as a measure of immunological connectivity. Clonal overlap was calculated using the Jaccard index, which for each pair of B cell subsets is the number of unique clones found in both subsets (intersect) divided by the total number of unique clones among the two subsets (union). Clones were relabeled as “S1 - ” clone when the ratio of S1⁺ to S1⁻ sorted B cells within the clone was less than 0.1. After clonal clustering, only heavy chain sequences were used for subsequent analysis. Clone expansion and B cell subset transition were estimated with indices (STARTRAC-expa and STARTRAC-tran) from the STARTRAC package (v. 0.1.0) 69, 70 . In STARTRAC-tran indices analysis, S1 + or S1 - subject specific populations with less than 10 cells were removed. Clones from INF and VAC groups were analyzed. To infer lineage trees, we estimated tree topologies, branch lengths, and subject-wide substitution model parameters using maximum likelihood under the GY94 model 136, 137 . Using fixed tree topologies estimated from the GY94 model, we then estimated branch lengths and donor-wide parameter values under the HLP19 model in IgPhyML (v. 2.0.0) 136 . Trees were visualized using dowser (v. 2.3) 135 and ggtree (v. 3.14.0) 138 . ATAC-seq ATAC-seq data processing and analysis Frozen tonsil cells were thawed and stained with antibodies listed in Supplementary Table 11 for 30 min at RT. Cells were washed with PBS twice and resuspended in RPMI with 10% FBS for sorting. 10,000 viable P1 (CXCR3 - CD21 + ), P2 (CXCR3 + CD21 + ), P3 (CXCR3 + CD21 - ) and P4 (CXCR3 - CD21 - ) B SM were sorted into 500 mL of FACS buffer using an FACSAria Fusion Sorter (BD) (cell sorting strategy is shown in Supplementary Figure 5B). Cells were pelleted and resuspended in 50 uL of transposase mixture including 25 mL 2xTD buffer (Illumina), 2.5 mL TDE1 (Illumina), 0.5 mL 1% digitonin (Promega) and 22 uL water. Tagmentation was performed by incubation at 37 °C for 30 minutes at 300 rpm. Following incubation, DNA was purified using a Qiagen MinElute kit, eluting samples in 10 uL. Purified tagmented DNA was PCR amplified using previously described primers 139 , with 12 cycles of amplification. Amplified libraries were purified using a Qiagen PCR cleanup kit and sequenced for 50 cycles (paired-end reads) on a NovaSeq 6000 (Illumina). ATAC-seq was done in three biological replicates per B SM subset (samples shown in Supplementary Table 2). ATAC-seq primers are listed in Supplementary Table 11. ATAC-seq data was processed using the chrom-seek pipeline (v. 1.0.0) 140 with --assay ATAC (https://github.com/OpenOmics/chrom-seek). Reads were trimmed with Cutadapt (v. 4.4) 141 . All reads aligning to the Encode hg38 v1 blacklist regions 142 were identified by alignment with BWA (v. 0.7.17) 143 and removed with Picard SamToFastq. Remaining reads were aligned to an hg38 reference genome using BWA. Reads with a mapQ score less than 6 were removed with SAMtools (v. 1.17) 106 and PCR duplicates were removed with Picard MarkDuplicates. Data was converted into bigwigs for viewing and normalized by reads per genomic content (RPGC) using deepTools 144 (v. 3.5.1) using the following parameters: --binSize 25 --smoothLength 75 --effectiveGenomeSize 2805636331 --centerReads --normalizeUsing RPGC. Averaged bigwigs were created using the bigwigAverage function of deepTools (v. 3.5.4) 144 . Peaks were called using macsNarrow 145 (macs v. 2.2.7.1) with the following parameters: -q 0.01 --keep-dup="all" -f "BAMPE". Differential peaks were called using DiffBind (v. 2.15.2) 146 and its Deseq2 differential caller with default parameters. Peaks were considered significant with an FDR value less than 0.1. Motif analysis was completed using the MEME suite (v. 5.5.5) 147 . Known motif enrichment analysis was accomplished using AME on a combined jolma 2013, jaspar 2018 core vertebrate non-rendundant, and HOCOMOCO (v.11) 148 full human mono database. Downstream analyses and results visualization were performed with R (v.4.4.2) and visualized with ggplot2 (v. 3.5.1) 100 . Differentially accessible regions (DAR) and pathway enrichment analysis Differential peaks from comparing P1 B SM and P2 B SM (adjusted p -value < 0.10) were selected for downstream analysis. To explore the functional significance of these DARs, pathway enrichment analysis was conducted using GREAT (Genomic Regions Enrichment of Annotations Tool, v. 4.0.4) 149, 150 . The genomic coordinates of DARs were converted to BED files and uploaded to the GREAT web server. The analysis was performed mapping to GRCh38 (UCSC hg38, Dec. 2013) and using the default association rules, which map genomic regions to nearby genes based on a basal plus extension model (5 kb upstream, 1 kb downstream, and up to 1 Mb extension from the transcription start site). Enriched terms from GO BP and GO MF, MSigDB, Reactome, and other curated databases were extracted. Pathways and terms with an FDR (adjusted p -value) < 0.05 were considered significant. Top20 GO BP pathways ranked by FDR (adjusted p -value) values were visualized with bar plot. IRF4 complex motif search with FIMO To further evaluate the enrichment of IRF4 complex motifs in P1 B SM and P2 B SM , a search with stringent EICE (GGAANNGAAA), ISRE (A/GNGAAANNGAAACT) and two AICE (0bp/4bp: 0bp – GAAATGA(G/C)TCA; 4bp – TTTCNNNNTGA(G/C)TCA) 151 motifs using Find Individual Motif Occurrences (FIMO) on the DAR sequences obtained from P1 B SM to P2 B SM comparison was performed 67 . The enrichment statistics were calculated as above using a two-tailed version of Fisher's exact test. For Extended Data Figure 4i, data for memory B cells (MBC) and plasma cells (PC) were retrieved from a published dataset (https://zenodo.org/records/8373756) 99 and loaded into Signac (v1.11.0) 152 . Plots were created with the Signac CoveragePlot function. BCR signaling assessment by phosphorylation staining Tonsil cells were thawed, rinsed with RPMI supplemented with 0.1% FBS, and rested for 80 minutes at 37 °C in the same medium. After resting, the cells were incubated with a live/dead stain in PBS for 15 minutes at RT and then washed once with PBS. This was followed by staining the cells with antibodies against CXCR3 and CXCR5 for 5 minutes at RT. Subsequently, the remaining surface antibody mix was added (Supplementary Table 11), and the cells were resuspended in FACS buffer for 20 minutes at RT 92 . Afterward, the cells were washed twice with RPMI supplemented with 0.1% heat inactivated FBS and then resuspended in pre-warmed RPMI with 10% FBS and stimulated with anti-BCR antibodies, as previously described 71, 153, 154 . Stimulation was carried out at 37 °C for 2 minutes using 10 mg/mL goat F(ab’)2 anti-human IgA/G/M (Jackson ImmunoResearch Laboratories). For detecting phosphorylated signaling intermediates, the cells were fixed and permeabilized using BD Cytofix and Phosflow Perm/Wash buffers (BD Biosciences), then stained with PE-phosphorylated Syk (p-Y348) and Alexa Fluor 488- phosphorylated PLCg2 (p-Y759) antibodies (BD Biosciences). Samples were acquired on an Aurora cytometer (Cytek), and analysis was performed using FlowJo (v. 10.9.0). Gating strategies are shown in Supplementary Figure 6A. In vitro plasmablast differentiation and proliferation assay Tonsil cells were thawed and stained with viability dye and surface antibody mix (Supplementary Table 11) at RT for 30 mins. Then cells were washed with PBS twice and resuspended in RPMI with 10% FBS at concentration of 5 million cells/mL for sorting. P1, P2, P3 and P4 B SM were sorted on Aria sorter (BD) into 0.3 mL RPMI with 20% FBS in 1.5 mL tubes (Supplementary Figure 5B). Sorted cells were centrifuged at RT for 10 mins and then were labeled with 0.5 mM CFSE (CellTrace CFSE cell proliferation kit, ThermoFisher) in 1100 μL PBS at RT in the dark for 10 mins. Then, cells were washed with pre-warmed RPMI + 10% FBS as described 153 . Allogenic B cell-depleted PBMCs were prepared using a B cell depletion kit (Dynabeads CD19 Pan B, ThermoFisher) from PBMCs of an unrelated healthy donor. The sorted memory B cells and allogeneic B cell depleted PBMC were co-cultured at a 1:9 to 2:8 ratio with 2.5 mg/mL R848 and 1000 U/mL recombinant human IL-2 for 4 days at 1 × 10 6 cells per well in a 96-well flatbottom plate. RPMI with 10% FBS was used for culture. The cells were collected and stained with antibodies against CD19, CD20, CD3, CD27, CD21, IgD, CD38, and CXCR3, fixed (Lysing Solution, BD Biosciences), permeabilized (Permeabilizing Solution 2; BD Biosciences) and stained with antibodies against IgG, IgA, IgM. The cells were acquired on an Aurora cytometer (Cytek) and the analyzed using FlowJo (v. 10.9.0). Antibodies are listed in Supplementary Table 11. Gating strategies are shown in Supplementary Figure 6B. The division index, a measure of the overall proliferative response, is the average number of divisions undergone per cell in the total population, including cells that have not undergone division. Tissue processing and staining for immunofluorescence assay Formalin-fixed, paraffin-embedded (FFPE) adenoid and tonsil tissue blocks were cut into 5 μm sections and mounted onto charged slides. Two paired tonsil and adenoid samples (from one INF donor and one VAC donor) were loaded onto the same slide. For staining, slides were deparaffinized and tissue rehydrated in deionized water. Antigen retrieval was performed by incubating slides in antigen retrieval (AR) buffer (Cepham life sciences) for 45 min in a steamer (preheated, approximately 95 °C). After 45 min, slides were taken out from the steamer and allowed to cool to RT. Sections were permeabilized, blocked for 1 hour in PBS containing 0.3 % Triton X-100 (Sigma-Aldrich), 1% bovine serum albumin (Jackson Immune Research). Sections were stained with titrated amounts of non-conjugated primary antibodies, followed by overnight incubation at 4 °C. Slides were then washed with PBS (3 times, 10 min each) and stained with the appropriate secondary antibodies for 2 hrs at RT. Slides were washed and blocked again for 1 hr at RT with a 1:10 dilution of normal mouse and rabbit or goat serum. Then, slides were stained with titrated amounts of directly conjugated antibodies for 2 hrs at RT. After three final washing steps and staining with the nuclear marker TOPRO (ThermoFisher), slides were mounted with prolonged gold anti-fade mounting media (ThermoFisher) and sealed with a glass coverslip. Antibodies are listed in Supplementary Table 11. Tissue sections were imaged (using confocal system, Leica Stellaris RTB WLL FLIM) as three-dimensional (3D) tile scans and subsequently mosaic-merged to generate a continuous representation. To minimize imaging artifacts, corrections were applied for motion-induced distortions, 3D alignment inconsistencies, and thermal drift across sequential z-sections. Additionally, crosstalk and color calibration adjustments were performed using Huygens Pro (version 24.04.0p3 64-bit, Scientific Volume Imaging BV). Image deconvolution was conducted within the same software to enhance signal resolution. The reconstructed images were further processed using Imaris (v. 10.2.0, Oxford Instruments). A combination of colocalization analysis, ChannelArithmetics, Xtension, Imaris installed machine learning-based classification, and masking techniques were employed to delineate distinct anatomical regions as additional computational channels, including follicles, germinal centers, extrafollicular regions, epithelium, and crypt structures. The Surface module of Imaris was utilized to generate cell objects, integrating nuclear signals and perinuclear regions derived from the preceding image processing steps. Quantitative data were extracted from processed files via a custom parser script, which reformatted surface object statistics into “.csv” files optimized for direct import into FlowJo (v. 10.9.0) for gating and further analysis. Statistical analysis and visualization were further processed in GraphPad Prism (v. 10.2.0392). SPACE analysis of immunofluorescence imaging data Spatial Patterning Analysis of Cellular Ensembles (SPACE) is an R package designed to identify complex spatial patterns at the cell and tissue levels 79 . Cell objects from immunofluorescence images were annotated with 8 populations as in Figure 8f by manual gating with FlowJo (v. 10.9.0) and labeled by an R-based customized script. The 10 µm radius captures close cellular associations and was chosen for SPACE “census_image” function. The number of neighborhoods was chosen to achieve 5 × tissue coverage. Covariation plots were created using the SPACE “learn_pattern” function. Spatial transcriptomic profiling with Xenium In Situ platform Slides were prepared following the manufacturer’s instructions and workflow for FFPE tissue samples (CG000578 Rev A; 10x Genomics). A 5-µm section from the tissue block containing the same paired tonsil and adenoid samples (one from INF donor and one from VAC donor) used for immunofluorescence were carefully attached to the sample area on a Xenium slide (Histoserv, MD). Samples are listed in Supplementary Table 2. Xenium slides were deparaffined and rehydrated and were then assembled into the Xenium Cassette. Deparaffinized slides in the Xenium cassette were decrosslinked (CG000580 Rev A) and immediately underwent probe hybridization, ligation, and amplification (CG000582 Rev D) using the Xenium 5000 human gene panel (Prime 5K Human Pan Tissue & Pathways Panel). With autofluorescence quenching and nuclei staining, the tissue images were captured and analyzed by the Xenium Analyzer (PN-1000569, instrument software v2.0.1.0). Regions of interest were manually selected from the scanned images. Post-run data for each slide was obtained using default parameters for downstream analysis. Data were processed by Xenium analysis software (v 2.0.0.10). The raw count matrix was pre-processed using the Seurat package (v. 5.1.0) 98, 99 in R (v. 4.4.2). For quality control, cells with at least nCount > 40, 15 nFeature > 10, cell_area > 10 & < 200 were retained 155 . Raw count data were normalized using the SCTransform function with method "glmGamPoi". Dimension reduction was performed using the runPCA function and the optimal number of principal components was selected using the ElbowPlot function. Cell clusters were determined using the FindClusters function. We annotated cell populations with human tonsil reference version 2 with Seurat (v. 5.0.2) 98 and Azimuth (v. 0.5.0) 91, 99 . Cell populations with less than 150 cells per slide were removed (granulocytes, mast, preB/T and PC/doublet cells). Some cell types were merged (B naïve includes B naive & B activated; B memory includes B memory & FCRL4/5 + B memory; CD4 Non-TFH includes CD4 Non-TFH & CD4 TCM; CD4 TFH includes CD4 TFH & CD4 TFH Mem; CD8 non-naïve includes CD8 T & CD8 TCM; gdT_MAIT includes MAIT/TRDV2 + gdT & non-TRDV2 + gdT; Mono/Macro includes Mono/Macro & Cycling myeloid; PB/PC includes PB & PC). Additional statistical analyses Correlations were analyzed using Spearman’s rank correlation test using base R and corrplot (v 0.95). Paired comparisons were performed with Wilcoxon signed ranks test with ggpubr (v. 0.6.0) 156 and visualized with ggplot2 (v. 3.5.1) 100 in R (v. 4.4.2). REFERENCES Baden, L.R. et al. Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine. N Engl J Med 384 ,403-416 (2021). Polack, F.P. et al. Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine. N Engl J Med 383 ,2603-2615 (2020). Muecksch, F. et al. Increased memory B cell potency and breadth after a SARS-CoV-2 mRNA boost. Nature 607 ,128-134 (2022). Goel, R.R. et al. mRNA vaccines induce durable immune memory to SARS-CoV-2 and variants of concern. Science 374 ,abm0829 (2021). Zhang, Z. et al. Humoral and cellular immune memory to four COVID-19 vaccines. Cell 185 ,2434-2451.e2417 (2022). Chemaitelly, H. et al. Protection from previous natural infection compared with mRNA vaccination against SARS-CoV-2 infection and severe COVID-19 in Qatar: a retrospective cohort study. The Lancet Microbe 3 ,e944-e955 (2022). Hall, V. et al. Protection against SARS-CoV-2 after Covid-19 Vaccination and Previous Infection. New England Journal of Medicine 386 ,1207-1220 (2022). Stein, C. et al. Past SARS-CoV-2 infection protection against re-infection: a systematic review and meta-analysis. The Lancet (2023). Zuo, F., Marcotte, H., Hammarström, L. & Pan-Hammarström, Q. Mucosal IgA against SARS-CoV-2 Omicron Infection. N Engl J Med 387 ,e55 (2022). Huang, N. et al. SARS-CoV-2 infection of the oral cavity and saliva. Nature Medicine 27 ,892-903 (2021). Wölfel, R. et al. Virological assessment of hospitalized patients with COVID-2019. Nature 581 ,465-469 (2020). Havervall, S. et al. Anti-Spike Mucosal IgA Protection against SARS-CoV-2 Omicron Infection. New England Journal of Medicine 387 ,1333-1336 (2022). Ravichandran, S. et al. Systemic and mucosal immune profiling in asymptomatic and symptomatic SARS-CoV-2-infected individuals reveal unlinked immune signatures. Sci Adv 7 ,eabi6533 (2021). Gaebler, C. et al. Evolution of antibody immunity to SARS-CoV-2. Nature 591 ,639-644 (2021). Sokal, A. et al. Maturation and persistence of the anti-SARS-CoV-2 memory B cell response. Cell 184 ,1201-1213.e1214 (2021). Allie, S.R. & Randall, T.D. Resident Memory B Cells. Viral Immunology 33 ,282-293 (2020). Lederer, K. et al. Germinal center responses to SARS-CoV-2 mRNA vaccines in healthy and immunocompromised individuals. Cell 185 ,1008-1024.e1015 (2022). Mudd, P.A. et al. SARS-CoV-2 mRNA vaccination elicits a robust and persistent T follicular helper cell response in humans. Cell 185 ,603-613.e615 (2022). Röltgen, K. et al. Immune imprinting, breadth of variant recognition, and germinal center response in human SARS-CoV-2 infection and vaccination. Cell 185 ,1025-1040.e1014 (2022). Turner, J.S. et al. SARS-CoV-2 mRNA vaccines induce persistent human germinal centre responses. Nature 596 ,109-113 (2021). Shehata, L. et al. Systematic comparison of respiratory syncytial virus-induced memory B cell responses in two anatomical compartments. Nature Communications 10 ,1126 (2019). Tan, H.-X. et al. Cutting Edge: SARS-CoV-2 Infection Induces Robust Germinal Center Activity in the Human Tonsil. The Journal of Immunology ,ji2101199 (2022). Xu, Q. et al. Adaptive immune responses to SARS-CoV-2 persist in the pharyngeal lymphoid tissue of children. Nature immunology 24 ,186-199 (2023). Hachim, A. et al. ORF8 and ORF3b antibodies are accurate serological markers of early and late SARS-CoV-2 infection. Nature Immunology 21 ,1293-1301 (2020). Ravichandran, S. et al. SARS-CoV-2 immune repertoire in MIS-C and pediatric COVID-19. Nature Immunology 22 ,1452-1464 (2021). Liu, C., Tsang, T.K., Sullivan, S.G., Cowling, B.J. & Yang, B. Comparative duration of neutralizing responses and protections of COVID-19 vaccination and correlates of protection. Nature Communications 16 ,4748 (2025). Abu-Raddad, L.J. et al. Effect of mRNA Vaccine Boosters against SARS-CoV-2 Omicron Infection in Qatar. New England Journal of Medicine 386 ,1804-1816 (2022). Altarawneh, H.N. et al. Protection against the Omicron Variant from Previous SARS-CoV-2 Infection. New England Journal of Medicine 386 ,1288-1290 (2022). Sun, P. et al. Asymptomatic or symptomatic SARS-CoV-2 infection plus vaccination confers increased adaptive immunity to variants of concern. iScience 25 ,105202 (2022). Azzi, L. et al. Mucosal immune response in BNT162b2 COVID-19 vaccine recipients. eBioMedicine 75 ,103788 (2022). Tang, J. et al. Respiratory mucosal immunity against SARS-CoV-2 after mRNA vaccination. Sci Immunol 7 ,eadd4853 (2022). Weisel, F. & Shlomchik, M. Memory B Cells of Mice and Humans. Annual Review of Immunology, Vol 35 35 ,255-284 (2017). Sanz, I. et al. Challenges and Opportunities for Consistent Classification of Human B Cell and Plasma Cell Populations. Frontiers in Immunology 10 (2019). Kardava, L. et al. Early human B cell signatures of the primary antibody response to mRNA vaccination. Proceedings of the National Academy of Sciences 119 ,e2204607119 (2022). Andrews, S.F. et al. Activation Dynamics and Immunoglobulin Evolution of Pre-existing and Newly Generated Human Memory B cell Responses to Influenza Hemagglutinin. Immunity 51 ,398-410.e395 (2019). Nellore, A. et al. A transcriptionally distinct subset of influenza-specific effector memory B cells predicts long-lived antibody responses to vaccination in humans. Immunity 56 ,847-863.e848 (2023). Spangler, A. et al. Early influenza virus exposure shapes the B cell response to influenza vaccination in individuals 50 years later. Immunity 58 ,728-744.e729 (2025). Pape, K.A. et al. High-affinity memory B cells induced by SARS-CoV-2 infection produce more plasmablasts and atypical memory B cells than those primed by mRNA vaccines. Cell Rep 37 ,109823 (2021). Jenks, S.A. et al. Distinct Effector B Cells Induced by Unregulated Toll-like Receptor 7 Contribute to Pathogenic Responses in Systemic Lupus Erythematosus. Immunity 49 ,725-739.e726 (2018). Moir , S. et al. Evidence for HIV-associated B cell exhaustion in a dysfunctional memory B cell compartment in HIV-infected viremic individuals. Journal of Experimental Medicine 205 ,1797-1805 (2008). Isnardi, I. et al. Complement receptor 2/CD21− human naive B cells contain mostly autoreactive unresponsive clones. Blood 115 ,5026-5036 (2010). Sutton, H.J. et al. Atypical B cells are part of an alternative lineage of B cells that participates in responses to vaccination and infection in humans. Cell Rep 34 ,108684 (2021). Weiss, G.E. et al. Atypical Memory B Cells Are Greatly Expanded in Individuals Living in a Malaria-Endemic Area1. The Journal of Immunology 183 ,2176-2182 (2009). Woodruff, M.C. et al. Extrafollicular B cell responses correlate with neutralizing antibodies and morbidity in COVID-19. Nature Immunology 21 ,1506-1516 (2020). Austin, J.W. et al. Overexpression of T-bet in HIV infection is associated with accumulation of B cells outside germinal centers and poor affinity maturation. Sci Transl Med 11 (2019). Knox, J.J. et al. T-bet+ B cells are induced by human viral infections and dominate the HIV gp140 response. JCI Insight 2 (2017). Li, H., Borrego, F., Nagata, S. & Tolnay, M. Fc Receptor–like 5 Expression Distinguishes Two Distinct Subsets of Human Circulating Tissue–like Memory B Cells. The Journal of Immunology 196 ,4064-4074 (2016). Kwon, D.-i. et al. Mucosal unadjuvanted booster vaccines elicit local IgA responses by conversion of pre-existing immunity in mice. Nature Immunology 26 ,908-919 (2025). Burton, A.R. et al. Circulating and intrahepatic antiviral B cells are defective in hepatitis B. The Journal of Clinical Investigation 128 ,4588-4603 (2018). Muehlinghaus, G. et al. Regulation of CXCR3 and CXCR4 expression during terminal differentiation of memory B cells into plasma cells. Blood 105 ,3965-3971 (2005). Davis-Porada, J. et al. Maintenance and functional regulation of immune memory to COVID-19 vaccines in tissues. Immunity 57 ,2895-2913.e2898 (2024). Oh, J.E. et al. Migrant memory B cells secrete luminal antibody in the vagina. Nature 571 ,122-126 (2019). Oh, J.E. et al. Intranasal priming induces local lung-resident B cell populations that secrete protective mucosal antiviral IgA. Science Immunology 6 ,eabj5129 (2021). Arroyo-Díaz, N.M. et al. Interferon-γ production by Tfh cells is required for CXCR3(+) pre-memory B cell differentiation and subsequent lung-resident memory B cell responses. Immunity 56 ,2358-2372.e2355 (2023). Johnson, J.L. et al. The Transcription Factor T-bet Resolves Memory B Cell Subsets with Distinct Tissue Distributions and Antibody Specificities in Mice and Humans. Immunity 52 ,842-855.e846 (2020). Lee, C.M. & Oh, J.E. Resident Memory B Cells in Barrier Tissues. Front Immunol 13 ,953088 (2022). Lau, D. et al. Low CD21 expression defines a population of recent germinal center graduates primed for plasma cell differentiation. Sci Immunol 2 (2017). Zhang, H., Meng, F., Chu, C.-L., Takai, T. & Lowell, C.A. The Src Family Kinases Hck and Fgr Negatively Regulate Neutrophil and Dendritic Cell Chemokine Signaling via PIR-B. Immunity 22 ,235-246 (2005). Gao, X. et al. Zeb2 drives the formation of CD11c+ atypical B cells to sustain germinal centers that control persistent infection. Science Immunology 9 ,eadj4748 (2024). Zurbuchen, Y. et al. Human memory B cells show plasticity and adopt multiple fates upon recall response to SARS-CoV-2. Nature Immunology 24 ,955-965 (2023). Shao, W., Wang, Y., Fang, Q., Shi, W. & Qi, H. Epigenetic recording of stimulation history reveals BLIMP1–BACH2 balance in determining memory B cell fate upon recall challenge. Nature Immunology 25 ,1432-1444 (2024). Alaterre, E. et al. Integrative single-cell chromatin and transcriptome analysis of human plasma cell differentiation. Blood 144 ,496-509 (2024). Zotos, D. et al. IL-21 regulates germinal center B cell differentiation and proliferation through a B cell-intrinsic mechanism. J Exp Med 207 ,365-378 (2010). Linterman, M.A. et al. IL-21 acts directly on B cells to regulate Bcl-6 expression and germinal center responses. J Exp Med 207 ,353-363 (2010). Kim, Y. et al. IL-21 shapes the B cell response in a context-dependent manner. Cell Rep 44 (2025). Dai, D. et al. The transcription factor ZEB2 drives the formation of age-associated B cells. Science 383 ,413-421 (2024). Ochiai, K. et al. Transcriptional regulation of germinal center B and plasma cell fates by dynamical control of IRF4. Immunity 38 ,918-929 (2013). Xu, H. et al. Regulation of bifurcating B cell trajectories by mutual antagonism between transcription factors IRF4 and IRF8. Nat Immunol 16 ,1274-1281 (2015). Ma, J. et al. A blueprint for tumor-infiltrating B cells across human cancers. Science 384 ,eadj4857 (2024). Zhang, L. et al. Lineage tracking reveals dynamic relationships of T cells in colorectal cancer. Nature 564 ,268-272 (2018). Buckner, C.M. et al. Interval between prior SARS-CoV-2 infection and booster vaccination impacts magnitude and quality of antibody and B cell responses. Cell 185 ,4333-4346.e4314 (2022). Ambegaonkar, A.A. et al. Expression of inhibitory receptors by B cells in chronic human infectious diseases restricts responses to membrane-associated antigens. Science Advances 6 ,eaba6493 (2020). Portugal, S. et al. Malaria-associated atypical memory B cells exhibit markedly reduced B cell receptor signaling and effector function. eLife 4 ,e07218 (2015). Halliley, Jessica L. et al. Long-Lived Plasma Cells Are Contained within the CD19−CD38hiCD138+ Subset in Human Bone Marrow. Immunity 43 ,132-145 (2015). Pinto, D. et al. A functional BCR in human IgA and IgM plasma cells. Blood 121 ,4110-4114 (2013). Phetsouphanh, C. et al. Immunological dysfunction persists for 8 months following initial mild-to-moderate SARS-CoV-2 infection. Nature Immunology 23 ,210-216 (2022). Cheong, J.-G. et al. Epigenetic memory of coronavirus infection in innate immune cells and their progenitors. Cell 186 ,3882-3902.e3824 (2023). Penna, G., Sozzani, S. & Adorini, L. Cutting edge: selective usage of chemokine receptors by plasmacytoid dendritic cells. The Journal of Immunology 167 ,1862-1866 (2001). Schrom, E.C. et al. Spatial Patterning Analysis of Cellular Ensembles (SPACE) finds complex spatial organization at the cell and tissue levels. Proceedings of the National Academy of Sciences 122 ,e2412146122 (2025). Iijima, N. & Iwasaki, A. T cell memory. A local macrophage chemokine network sustains protective tissue-resident memory CD4 T cells. Science 346 ,93-98 (2014). Giesecke, C. et al. Tissue Distribution and Dependence of Responsiveness of Human Antigen-Specific Memory B Cells. The Journal of Immunology 192 ,3091-3100 (2014). Joo, H.M., He, Y., Sundararajan, A., Huan, L. & Sangster, M.Y. Quantitative analysis of influenza virus-specific B cell memory generated by different routes of inactivated virus vaccination. Vaccine 28 ,2186-2194 (2010). Kim, W. et al. Germinal centre-driven maturation of B cell response to mRNA vaccination. Nature 604 ,141-145 (2022). Killingley, B. et al. Safety, tolerability and viral kinetics during SARS-CoV-2 human challenge in young adults. Nature Medicine 28 ,1031-1041 (2022). Schaefer-Babajew, D. et al. Antibody feedback regulates immune memory after SARS-CoV-2 mRNA vaccination. Nature 613 ,735-742 (2023). Wachter, B.T. et al. Prior SARS-CoV-2 infection affects adaptive immune responses to Omicron BA.4/BA.5 mRNA booster. Journal of Allergy and Clinical Immunology (2025). Bettini, E. et al. Distinct components of nucleoside-modified messenger RNA vaccines cooperate to instruct efficient germinal center responses. bioRxiv (2024). Kiyono, H. & Ernst, P.B. Nasal vaccines for respiratory infections. Nature 641 ,321-330 (2025). Alu, A. et al. Intranasal COVID-19 vaccines: From bench to bed. eBioMedicine 76 (2022). Qi, H., Liu, B., Wang, X. & Zhang, L. The humoral response and antibodies against SARS-CoV-2 infection. Nature Immunology 23 ,1008-1020 (2022). Massoni-Badosa, R. et al. An atlas of cells in the human tonsil. Immunity 57 ,379-399 e318 (2024). Xu, Q. et al. Efficient Tonsillar T Follicular Helper Cell Processing and Functional Analysis through High-dimensional Flow Cytometry , vol. 221. 1940-087X, 2025. Tang, J.J. et al. Cross-reactive immunity against the SARS-CoV-2 Omicron variant is low in pediatric patients with prior COVID-19 or MIS-C (vol 13, pg 2979, 2022). Nature Communications 13 (2022). Neerukonda, S.N. et al. Establishment of a well-characterized SARS-CoV-2 lentiviral pseudovirus neutralization assay using 293T cells with stable expression of ACE2 and TMPRSS2. Plos One 16 (2021). Ravichandran, S. et al. Antibody signature induced by SARS-CoV-2 spike protein immunogens in rabbits. Sci Transl Med 12 (2020). Tang, J.J. et al. Epitope diversity of SARS-CoV-2 hyperimmune intravenous human immunoglobulins and neutralization of variants of concern. Iscience 24 (2021). Awasthi, M., Golding, H. & Khurana, S. Severe Acute Respiratory Syndrome Coronavirus 2 Hyperimmune Intravenous Human Immunoglobulins Neutralizes Omicron Subvariants BA.1, BA.2, BA.2.12.1, BA.3, and BA.4/BA.5 for Treatment of Coronavirus Disease 2019. Clin Infect Dis 76 ,E503-E506 (2023). Hao, Y.H. et al. Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nature Biotechnology 42 (2024). Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184 ,3573-3587 e3529 (2021). Wickham, H. ggplot2 Elegant Graphics for Data Analysis Introduction. Use R ,1-+ (2009). Zhang, Z. Reshaping and aggregating data: an introduction to reshape package. Ann Transl Med 4 ,78 (2016). Kassambara, A. ggpubr: 'ggplot2' Based Publication Ready Plots. 2025. Arnold, J.B. ggthemes: Extra Themes, Scales and Geoms for 'ggplot2'. 2024. Dan, J.M. et al. Recurrent group A Streptococcus tonsillitis is an immunosusceptibility disease involving antibody deficiency and aberrant T(FH) cells. Sci Transl Med 11 (2019). Stoeckius, M. et al. Cell Hashing with barcoded antibodies enables multiplexing and doublet detection for single cell genomics. Genome Biol 19 ,224 (2018). Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience 10 (2021). Rogozhnikov, A. et al. Demuxalot: scaled up genetic demultiplexing for single-cell sequencing. bioRxiv ,2021.2005. 2022.443646 (2021). Kang, H.M. et al. Multiplexed droplet single-cell RNA-sequencing using natural genetic variation (vol 36, pg 89, 2018). Nature Biotechnology 38 ,1356-1356 (2020). Zheng, G.X. et al. Massively parallel digital transcriptional profiling of single cells. Nat Commun 8 ,14049 (2017). Mule, M.P., Martins, A.J. & Tsang, J.S. Normalizing and denoising protein expression data from droplet-based single cell profiling. Nat Commun 13 ,2099 (2022). Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 32 ,2847-2849 (2016). Ritchie, M.E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 43 ,e47 (2015). Finak, G., Jiang, W. & Gottardo, R. CytoML for cross-platform cytometry data sharing. Cytometry A 93 ,1189-1196 (2018). Finak G, J.M. flowWorkspace: Infrastructure for representing and interacting with gated and ungated cytometry data sets.; 2025. R, K. pheatmap: Pretty Heatmaps. . 2025. Liu, C. et al. Time-resolved systems immunology reveals a late juncture linked to fatal COVID-19. Cell 184 ,1836-1857 e1822 (2021). Sacco, K. et al. Immunopathological signatures in multisystem inflammatory syndrome in children and pediatric COVID-19. Nat Med 28 ,1050-1062 (2022). McCarthy, D.J., Chen, Y. & Smyth, G.K. Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. Nucleic Acids Res 40 ,4288-4297 (2012). Law, C.W., Chen, Y.S., Shi, W. & Smyth, G.K. voom: precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biology 15 (2014). Tan, H.X. et al. Lung-resident memory B cells established after pulmonary influenza infection display distinct transcriptional and phenotypic profiles. Sci Immunol 7 ,eabf5314 (2022). Yang, Y. et al. Pan-cancer single-cell dissection reveals phenotypically distinct B cell subtypes. Cell 187 ,4790-4811 e4722 (2024). Holla, P. et al. Shared transcriptional profiles of atypical B cells suggest common drivers of expansion and function in malaria, HIV, and autoimmunity. Sci Adv 7 (2021). Moir, S. et al. Evidence for HIV-associated B cell exhaustion in a dysfunctional memory B cell compartment in HIV-infected viremic individuals. J Exp Med 205 ,1797-1805 (2008). Wu, C. et al. Lupus-associated atypical memory B cells are mTORC1-hyperactivated and functionally dysregulated. Ann Rheum Dis 78 ,1090-1100 (2019). Verstappen, G.M. et al. Gene expression profiling of epithelium-associated FcRL4(+) B cells in primary Sjogren's syndrome reveals a pathogenic signature. J Autoimmun 109 ,102439 (2020). Isnardi, I. et al. Complement receptor 2/CD21 human naive B cells contain mostly autoreactive unresponsive clones. Blood 115 ,5026-5036 (2010). Attaf, N. et al. FB5P-seq: FACS-Based 5-Prime End Single-Cell RNA-seq for Integrative Analysis of Transcriptome and Antigen Receptor Repertoire in B and T Cells (vol 11, 216, 2020). Frontiers in Immunology 11 (2020). Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 102 ,15545-15550 (2005). Street, K. et al. Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC Genomics 19 ,477 (2018). Ye, J., Ma, N., Madden, T.L. & Ostell, J.M. IgBLAST: an immunoglobulin variable domain sequence analysis tool. Nucleic Acids Res 41 ,W34-40 (2013). Gupta, N.T. et al. Change-O: a toolkit for analyzing large-scale B cell immunoglobulin repertoire sequencing data. Bioinformatics 31 ,3356-3358 (2015). Nouri, N. & Kleinstein, S.H. A spectral clustering-based method for identifying clones from high-throughput B cell repertoire sequencing data. Bioinformatics 34 ,i341-i349 (2018). Gupta, N.T. et al. Hierarchical Clustering Can Identify B Cell Clones with High Confidence in Ig Repertoire Sequencing Data. J Immunol 198 ,2489-2499 (2017). Yaari, G. et al. Models of Somatic Hypermutation Targeting and Substitution Based on Synonymous Mutations from High-Throughput Immunoglobulin Sequencing Data. Frontiers in Immunology 0 (2013). Hoehn, K.B., Pybus, O.G. & Kleinstein, S.H. Phylogenetic analysis of migration, differentiation, and class switching in B cells. Plos Computational Biology 18 (2022). Hoehn, K.B. et al. Repertoire-wide phylogenetic models of B cell molecular evolution reveal evolutionary signatures of aging and vaccination. Proceedings of the National Academy of Sciences 116 ,22664-22672 (2019). Nielsen, R. & Yang, Z. Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelope gene. Genetics 148 ,929-936 (1998). ggtree: an r package for visualization and annotation of phylogenetic trees with their covariates and other associated data - Yu - 2017 - Methods in Ecology and Evolution - Wiley Online Library. (2021). Corces, M.R. et al. Lineage-specific and single-cell chromatin accessibility charts human hematopoiesis and leukemia evolution. Nat Genet 48 ,1193-1203 (2016). Jang, M.K. et al. Cell-free chromatin immunoprecipitation to detect molecular pathways in heart transplantation. Life Sci Alliance 6 (2023). Martin, M. Cutadpt Removes Adapter Sequences from High-Throughput Sequencing Reads. . EMBnet 17 ,10-12 (2011). Amemiya, H.M., Kundaje, A. & Boyle, A.P. The ENCODE Blacklist: Identification of Problematic Regions of the Genome. Sci Rep 9 ,9354 (2019). Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25 ,1754-1760 (2009). Ramirez, F. et al. deepTools2: a next generation web server for deep-sequencing data analysis. Nucleic Acids Res 44 ,W160-165 (2016). Zhang, Y. et al. Model-based Analysis of ChIP-Seq (MACS). Genome Biology 9 (2008). Ross-Innes, C.S. et al. Differential oestrogen receptor binding is associated with clinical outcome in breast cancer. Nature 481 ,389-U177 (2012). Bailey, T.L., Johnson, J., Grant, C.E. & Noble, W.S. The MEME Suite. Nucleic Acids Res 43 ,W39-49 (2015). Kulakovskiy, I.V. et al. HOCOMOCO: towards a complete collection of transcription factor binding models for human and mouse via large-scale ChIP-Seq analysis. Nucleic Acids Res 46 ,D252-D259 (2018). Tanigawa, Y., Dyer, E.S. & Bejerano, G. WhichTF is functionally important in your open chromatin data? PLOS Computational Biology 18 ,e1010378 (2022). McLean, C.Y. et al. GREAT improves functional interpretation of cis-regulatory regions. Nat Biotechnol 28 ,495-501 (2010). Glasmacher, E. et al. A Genomic Regulatory Element That Directs Assembly and Function of Immune-Specific AP-1-IRF Complexes. Science 338 ,975-980 (2012). Stuart, T., Srivastava, A., Madad, S., Lareau, C.A. & Satija, R. Single-cell chromatin state analysis with Signac (vol 18, pg 1333, 2021). Nat Methods 19 ,257-257 (2022). Kardava, L. et al. IgG3 regulates tissue-like memory B cells in HIV-infected individuals. Nat Immunol 19 ,1001-1012 (2018). Ambegaonkar, A.A., Nagata, S., Pierce, S.K. & Sohn, H. The Differentiation of Human Tonsil B Cells With the Phenotypic and Functional Characteristics of T-bet plus Atypical Memory B Cells in Malaria. Frontiers in Immunology 10 (2019). Rose, J.R. Xen_Seurat_Pipeline. GitHub.https://github.com/jrose835/Xen_Seurat_Pipeline. 2025. Kassambara, A. ggpubr: 'ggplot2' Based Publication Ready Plots version 0.4.0.from GitHub. (2020). Additional Declarations There is NO Competing Interest. Supplementary Files 20250805SupplementaryTable110.xlsx Supplemental Tables 1 - 10 20250805SupplementalTable11.xlsx Supplemental Table 11. Reagents 20250813SupplementaryFigures.docx Supplementary Figures EXTENDEDDATAFIGURES.docx Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7428491","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":509247914,"identity":"1514ca71-3a61-437f-8f7b-fa4e0442c017","order_by":0,"name":"Kalpana Manthiram","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABEklEQVRIiWNgGAWjYBACCQbmBjiHmcEASLI3MBzGr4URroWxGayF5wBJWsBCCUDr8ADJ9oNtD37usGPQbT97/HFBwT153ZlvDx4uYKiVMziAXYs0T2K7Ye+ZZAazM3mJzTMMig233c5LODyD4bgxLi1yDIltErxtzAxmB3IMm3kMEhi33c4xOMzDcCxxZgMOLfwP2yT/ttUzmJ1/A9Ziv+3mGfxapCUS26R52w4zmN2A2JK47QYPSEtNYj8u78942G4s23acx+zGG8PZQC3J286AHGZwwJgfhxaJ88nHHr5tq5YzO59j8JnnT4LttuNnjD/zVNTJseHQAgRgKR40QQO8sYndtDp8WkbBKBgFo2BkAQDLMFz9nvuqgQAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-6451-8173","institution":"NIAID/National Institutes of Health","correspondingAuthor":true,"prefix":"","firstName":"Kalpana","middleName":"","lastName":"Manthiram","suffix":""},{"id":509247915,"identity":"e99b9e38-c9d0-43a1-95a7-ebca70902fed","order_by":1,"name":"Qin Xu","email":"","orcid":"https://orcid.org/0000-0003-1812-3108","institution":"NIAID/National Institutes of Health","correspondingAuthor":false,"prefix":"","firstName":"Qin","middleName":"","lastName":"Xu","suffix":""},{"id":509247916,"identity":"3666f10d-d74b-4d1b-aa4a-8838f84f50dd","order_by":2,"name":"Lihong Shi","email":"","orcid":"","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Lihong","middleName":"","lastName":"Shi","suffix":""},{"id":509247917,"identity":"de2b4a69-949a-45d1-8b3c-eadb9bfdab11","order_by":3,"name":"Liya Wang","email":"","orcid":"","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Liya","middleName":"","lastName":"Wang","suffix":""},{"id":509247918,"identity":"1e5d1aec-bf43-4893-ac66-734b7f20fee1","order_by":4,"name":"Foo Cheung","email":"","orcid":"","institution":"Trans-NIH Center for Human Immunology, Autoimmunity, and Inflammation (CHI)","correspondingAuthor":false,"prefix":"","firstName":"Foo","middleName":"","lastName":"Cheung","suffix":""},{"id":509247919,"identity":"7d16cd06-aa37-4058-83ea-ce834f31598d","order_by":5,"name":"Aparna Kotekar","email":"","orcid":"","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Aparna","middleName":"","lastName":"Kotekar","suffix":""},{"id":509247920,"identity":"39b3154c-6436-4d07-a8d1-3edfa8aaa9e2","order_by":6,"name":"Galina Koroleva","email":"","orcid":"","institution":"NIAID/National Institutes of Health","correspondingAuthor":false,"prefix":"","firstName":"Galina","middleName":"","lastName":"Koroleva","suffix":""},{"id":509247921,"identity":"c72f017a-7327-48d8-9013-5cc7d8624561","order_by":7,"name":"Elizabeth Rice","email":"","orcid":"","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Elizabeth","middleName":"","lastName":"Rice","suffix":""},{"id":509247922,"identity":"c2f7e5c1-b749-471d-90e0-19b69859babc","order_by":8,"name":"Richard Apps","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Richard","middleName":"","lastName":"Apps","suffix":""},{"id":509247923,"identity":"aa96400b-5028-48ef-9ddc-eb87ce5818fe","order_by":9,"name":"Justin Lack","email":"","orcid":"","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Justin","middleName":"","lastName":"Lack","suffix":""},{"id":509247924,"identity":"304444dc-e6d3-4c96-96e3-09388295493d","order_by":10,"name":"Craig Martens","email":"","orcid":"","institution":"NIH","correspondingAuthor":false,"prefix":"","firstName":"Craig","middleName":"","lastName":"Martens","suffix":""},{"id":509247925,"identity":"9d1a1728-0f3e-4f1f-88cf-35df94258ebb","order_by":11,"name":"Iyadh Douagi","email":"","orcid":"https://orcid.org/0000-0002-3221-8667","institution":"NIH","correspondingAuthor":false,"prefix":"","firstName":"Iyadh","middleName":"","lastName":"Douagi","suffix":""},{"id":509247926,"identity":"eaa44a33-47f6-43c9-b4cf-7692eb937ad4","order_by":12,"name":"Can Liu","email":"","orcid":"","institution":"Yale School of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Can","middleName":"","lastName":"Liu","suffix":""},{"id":509247927,"identity":"1797a28d-67fa-4a1f-9367-3eb5d922b85b","order_by":13,"name":"Juraj Kabat","email":"","orcid":"","institution":"National Institutes of Health","correspondingAuthor":false,"prefix":"","firstName":"Juraj","middleName":"","lastName":"Kabat","suffix":""},{"id":509247928,"identity":"f49c124c-0c3d-4fd1-84f1-d6a8a499a446","order_by":14,"name":"Hengameh Behzadpour","email":"","orcid":"https://orcid.org/0000-0002-0050-8336","institution":"Children's National Hospital","correspondingAuthor":false,"prefix":"","firstName":"Hengameh","middleName":"","lastName":"Behzadpour","suffix":""},{"id":509247929,"identity":"dd9ea237-a101-44a6-995f-26051b9f3730","order_by":15,"name":"Lela Kardava","email":"","orcid":"https://orcid.org/0000-0001-9135-6026","institution":"National Institute of Allergy \u0026 Infectious Disease","correspondingAuthor":false,"prefix":"","firstName":"Lela","middleName":"","lastName":"Kardava","suffix":""},{"id":509247930,"identity":"48c1c6ac-dd49-4bbf-8ff7-a1ef12080ada","order_by":16,"name":"Tovah Markowitz","email":"","orcid":"","institution":"NIAID, NIH","correspondingAuthor":false,"prefix":"","firstName":"Tovah","middleName":"","lastName":"Markowitz","suffix":""},{"id":509247931,"identity":"fcfdbae3-0a2f-4612-bccd-be68ef4d5ffc","order_by":17,"name":"Margery Smelkinson","email":"","orcid":"https://orcid.org/0000-0001-7777-5574","institution":"NIAID","correspondingAuthor":false,"prefix":"","firstName":"Margery","middleName":"","lastName":"Smelkinson","suffix":""},{"id":509247932,"identity":"7d791356-053f-4182-9713-ff84845a6f4b","order_by":18,"name":"Kenneth Hoehn","email":"","orcid":"","institution":"Department of Biomedical Data Science, Geisel School of Medicine at Darmouth, Hanover, NH, United States","correspondingAuthor":false,"prefix":"","firstName":"Kenneth","middleName":"","lastName":"Hoehn","suffix":""},{"id":509247933,"identity":"99eaf612-7a68-48db-a453-ca612e57cbf2","order_by":19,"name":"Clarisa Buckner","email":"","orcid":"","institution":"National Institutes of Health","correspondingAuthor":false,"prefix":"","firstName":"Clarisa","middleName":"","lastName":"Buckner","suffix":""},{"id":509247934,"identity":"d4b86e56-4b7f-4357-9d0e-da56d71e2ea0","order_by":20,"name":"Dominic Golec","email":"","orcid":"","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Dominic","middleName":"","lastName":"Golec","suffix":""},{"id":509247935,"identity":"d46f86c9-66cb-4cc1-b61d-76e4082f5900","order_by":21,"name":"Lorenza Bellusci","email":"","orcid":"","institution":"Food and Drug Administration","correspondingAuthor":false,"prefix":"","firstName":"Lorenza","middleName":"","lastName":"Bellusci","suffix":""},{"id":509247936,"identity":"a1716ccb-3153-4368-85df-bd47ae99e7e7","order_by":22,"name":"Gabrielle Grubbs","email":"","orcid":"","institution":"Food and Drug Administration","correspondingAuthor":false,"prefix":"","firstName":"Gabrielle","middleName":"","lastName":"Grubbs","suffix":""},{"id":509247937,"identity":"d515d683-0a40-438e-9a2e-f567b5b802ea","order_by":23,"name":"Sara Pourhashemi","email":"","orcid":"","institution":"Food and Drug Administration","correspondingAuthor":false,"prefix":"","firstName":"Sara","middleName":"","lastName":"Pourhashemi","suffix":""},{"id":509247938,"identity":"b5c027be-5b8e-4b6a-be54-2861ba9e5e4e","order_by":24,"name":"Juanjie Tang","email":"","orcid":"","institution":"CBER/Food and Drug Administration","correspondingAuthor":false,"prefix":"","firstName":"Juanjie","middleName":"","lastName":"Tang","suffix":""},{"id":509247939,"identity":"2d94f29e-cde7-4715-8e63-259615861712","order_by":25,"name":"Asya Khleborodova","email":"","orcid":"","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Asya","middleName":"","lastName":"Khleborodova","suffix":""},{"id":509247940,"identity":"11f5db80-225f-43ac-8ff6-ccd6d9d55d5f","order_by":26,"name":"Martha Kirby","email":"","orcid":"https://orcid.org/0000-0001-7157-7964","institution":"National Human Genome Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Martha","middleName":"","lastName":"Kirby","suffix":""},{"id":509247941,"identity":"dcf64f62-3825-4bbe-9f04-af57b4381d7d","order_by":27,"name":"Rachel Sparks","email":"","orcid":"https://orcid.org/0000-0001-8216-5084","institution":"National Institute of Allergy and Infectious Diseases","correspondingAuthor":false,"prefix":"","firstName":"Rachel","middleName":"","lastName":"Sparks","suffix":""},{"id":509247942,"identity":"f903cf06-3c80-4079-b0ef-0d4bf8c2d826","order_by":28,"name":"Andrew Martins","email":"","orcid":"","institution":"Yale University","correspondingAuthor":false,"prefix":"","firstName":"Andrew","middleName":"","lastName":"Martins","suffix":""},{"id":509247943,"identity":"99ad5697-03ef-46e2-a935-8d3a3aea8e69","order_by":29,"name":"John Tsang","email":"","orcid":"https://orcid.org/0000-0003-3186-3047","institution":"Yale University","correspondingAuthor":false,"prefix":"","firstName":"John","middleName":"","lastName":"Tsang","suffix":""},{"id":509247944,"identity":"fd4fea5c-697a-4fa6-8098-b697139187d3","order_by":30,"name":"Susan Moir","email":"","orcid":"https://orcid.org/0000-0002-0163-6911","institution":"Laboratory of Immunoregulation, National Institute of Allergy and Infectious Diseases, National Institutes of Health","correspondingAuthor":false,"prefix":"","firstName":"Susan","middleName":"","lastName":"Moir","suffix":""},{"id":509247945,"identity":"caf166d3-0444-41e5-b9db-4ce220d94695","order_by":31,"name":"Surender Khurana","email":"","orcid":"https://orcid.org/0000-0002-0593-7965","institution":"Food and Drug Administration","correspondingAuthor":false,"prefix":"","firstName":"Surender","middleName":"","lastName":"Khurana","suffix":""},{"id":509247946,"identity":"638b0888-142d-47a1-a124-fb717892d549","order_by":32,"name":"Pamela Mudd","email":"","orcid":"","institution":"Children's National Hospital","correspondingAuthor":false,"prefix":"","firstName":"Pamela","middleName":"","lastName":"Mudd","suffix":""},{"id":509247947,"identity":"927cc1d7-0e28-40e6-b819-03990cc63d7f","order_by":33,"name":"Pamela Schwartzberg","email":"","orcid":"","institution":"NIAID/National Institutes of Health","correspondingAuthor":false,"prefix":"","firstName":"Pamela","middleName":"","lastName":"Schwartzberg","suffix":""}],"badges":[],"createdAt":"2025-08-21 17:46:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7428491/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7428491/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":90884694,"identity":"cc062b52-23ee-4065-975f-257391cb815b","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1316853,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSARS-CoV-2-specific B cells are found in the pharyngeal tissues and blood post-vaccination\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a) Study design and participant enrollment. CON (control), INF (post-infection), VAC (vaccinated).\u0026nbsp; S = spike, NC = nucleocapsid, ORF8 = open reading frame 8\u003c/p\u003e\n\u003cp\u003e(b) Time from vaccination(s) (shown in orange) and/or infection (shown in green) to surgery (right). Samples received for each participant and sex of participant (left).\u003c/p\u003e\n\u003cp\u003e(c) Upper panel: serum neutralizing antibody titers (PsVNA50) against the early strain WA-1 and later circulating strain B1.1.159 (omicron, BA.1) (CON N = 12, INF N = 23, VAC N = 10). Lower panel: Proportion of subjects with neutralizing antibodies against the indicated strain in INF and VAC groups.\u003c/p\u003e\n\u003cp\u003e(d) Correlation between serum neutralizing antibody titers to WA-1 or omicron BA.1 and days from last exposure (infection or vaccination) to surgery (INF N = 10, VAC N = 9) with Spearman’s test. 95% confidence intervals are shaded.\u003c/p\u003e\n\u003cp\u003e(e) Flow cytometry plots showing the percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e cells (stained with fluorescently labelled S1 and RBD tetramers) among B\u003csub\u003eSM\u003c/sub\u003e in PBMCs, adenoids, and tonsils of representative CON, INF and VAC participants.\u003c/p\u003e\n\u003cp\u003e(f) Percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e cells among CD19\u003csup\u003e+\u003c/sup\u003e B cells in PBMCs, adenoids, and tonsils of CON, INF and VAC participants (PBMC: CON N = 12, INF N = 15, VAC N = 10; adenoid: CON N = 10, INF N = 14, VAC N = 9; tonsil: CON N = 12, INF N = 22, VAC N = 7).\u003c/p\u003e\n\u003cp\u003e(g) Percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e cells among CD19\u003csup\u003e+\u003c/sup\u003e B cells in matched PBMCs, adenoids, and tonsils from INF (N = 5) versus VAC (N= 6) participants, compared with two-sided Wilcoxon signed ranks test (paired).\u003c/p\u003e\n\u003cp\u003e(h) Correlations among serum neutralizing titers to WA-1 and percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in PBMCs, adenoids, and tonsils. INF (left) and VAC (right) are shown separately (PBMC INF N = 9, VAC N = 10; tonsil INF N = 10, VAC N = 7; adenoid INF = 5, VAC = 9). Correlations assessed with Spearman’s rank correlation. Correlation coefficients (r values) are indicated by colors shown in the bar. Size of the circle indicates absolute r value. Significant p values are indicated as * p\u0026lt;0.05, ** p\u0026lt;0.01, *** p\u0026lt;0.001.\u003c/p\u003e\n\u003cp\u003e(i) Composition of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells and total CD19\u003csup\u003e+\u003c/sup\u003e B cells in PBMCs, adenoids, and tonsils of INF and VAC determined by flow cytometry. Samples with fewer than 25 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells were excluded from this figure. Mean frequencies are shown (PBMC: INF N = 11, VAC N = 9; adenoid: INF N = 14, VAC N = 8; tonsil: INF N = 22, VAC N = 6). B\u003csub\u003eSM\u003c/sub\u003e = switched memory B cells, USM = unswitched memory B cells, GCB = germinal center B cells, PC/PB = plasma cells/plasmablasts.\u003c/p\u003e\n\u003cp\u003e(j) Correlation between percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e cells among B\u003csub\u003eSM\u003c/sub\u003e in adenoid and tonsil and days from last exposure to surgery in INF and VAC groups (adenoid: INF N = 5, VAC N = 9; tonsil: INF N = 10, VAC N = 7) with Spearman’s test. 95% confidence intervals are shaded.\u003c/p\u003e\n\u003cp\u003e(k) Immunoglobulin isotype composition among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e cells and total B\u003csub\u003eSM\u003c/sub\u003e in PBMCs, adenoids, and tonsils of INF and VAC. Samples with \u0026lt; 10 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e were excluded (PBMC: INF N = 13, VAC N = 9; adenoid: INF N = 14, VAC N = 8; tonsil: INF N = 22, VAC N = 6). Mean frequencies are shown (Supplemental Table 4).\u003c/p\u003e\n\u003cp\u003eP values obtained from linear model in panels c and f correcting for participant age (in blue) or from two-sided Mann-Whitney U test (in black) are shown. Medians ± quartiles are shown in box plots. \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05 were considered significant.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/66d43721948a99e323dcbfa2.png"},{"id":90886964,"identity":"9125e612-f755-45ff-bc83-d5c9734ad9a1","added_by":"auto","created_at":"2025-09-09 10:16:52","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1434524,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eUnsupervised analyses of S1\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eRBD\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e B cells in the peripheral blood\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a) Uniform manifold approximation and projection (UMAP) of unsupervised clustering based on surface markers from flow cytometric analysis of 4348 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e CD19\u003csup\u003e+\u003c/sup\u003e B cells of PBMC from both INF (1275 cells) and VAC (3073 cells) groups. Clusters enriched in VAC group are indicated in yellow and INF group in green in the legend.\u003c/p\u003e\n\u003cp\u003e(b) Heatmap of marker/antibody expression in each cluster.\u003c/p\u003e\n\u003cp\u003e(c) Mean proportion of each cluster across subjects among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in each group.\u003c/p\u003e\n\u003cp\u003e(d) Principal component analysis (PCA) of each subject based on cluster frequencies. Each point represents an individual and is colored by group.\u003c/p\u003e\n\u003cp\u003e(e) Comparison of frequencies of each cluster among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells by group. Differentially distributed clusters are shown; data from all clusters are shown in Extended Data Figure 2a. Medians ± quartiles are shown in box plots. P values obtained from linear model correcting for participant ages (in blue) and from two-sided Mann-Whitney U test (in black) are shown. \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05 was considered significant.\u003c/p\u003e\n\u003cp\u003e(f) Distribution of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e\u0026nbsp;B cells among clusters by group on UMAP.\u003c/p\u003e\n\u003cp\u003e(g) Densities of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e\u0026nbsp;B cells in each cluster by group, with density shading indicating the concentration of cells in different areas of UMAP. Color represents cell density.\u003c/p\u003e\n\u003cp\u003e(h) Heatmaps of selected individual marker/antibody expression overlayed on UMAP.\u003c/p\u003e\n\u003cp\u003eFor all panels, INF N = 15, VAC N = 10 (Supplemental Table 5).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/98477a3eccb6e75b352cf851.png"},{"id":90884692,"identity":"8e56414a-5d30-4fc5-a53a-3be2ae580616","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1688086,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eUnsupervised analyses of S1\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eRBD\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e B cells in the adenoid and tonsil\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a) UMAP of unsupervised clustering based on surface markers from flow cytometric analysis of 20,801 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e CD19\u003csup\u003e+\u003c/sup\u003e B cells from adenoid and tonsil of both INF (14,627 cells) and VAC (6174 cells) subjects. Clusters enriched in VAC group are indicated in yellow and INF group in green in legend.\u003c/p\u003e\n\u003cp\u003e(b) Heatmap of marker/antibody expression in each cluster.\u003c/p\u003e\n\u003cp\u003e(c) Mean proportion of each cluster across subjects among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in INF and VAC tissues.\u003c/p\u003e\n\u003cp\u003e(d) PCA of each subject based on cluster frequency. Each point represents an individual and is colored by group.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(e) Frequency of subjects with ≥ 10 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e GCB cells in at least one tissue in INF and VAC participants (see Supplemental Table 7 for details).\u003c/p\u003e\n\u003cp\u003e(f) Densities of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e\u0026nbsp;B cells in each cluster by group, with density shading indicating the concentration of cells in different parts of the UMAP. Color represents cell densities.\u003c/p\u003e\n\u003cp\u003e(g) Heatmaps of selected individual marker/antibody expression overlayed on UMAP.\u003c/p\u003e\n\u003cp\u003e(h) Comparisons of cluster frequencies among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in INF and VAC groups according to tissue type. Differentially distributed clusters are shown; data from all clusters are shown in Extended Data Figure 2d. Medians ± quartiles are shown in the box plots.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e(i) Percentage of CXCR3\u003csup\u003e+\u003c/sup\u003eIgA\u003csup\u003e+\u003c/sup\u003e or CXCR3\u003csup\u003e+\u003c/sup\u003eIgG\u003csup\u003e+\u003c/sup\u003e cells among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in PBMCs, adenoid, and tonsils of INF and VAC participants. Samples with \u0026lt; 10 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e were excluded (PBMC INF N = 13, VAC N = 9; adenoid INF N = 14, VAC = 8; tonsil INF N = 22, VAC N = 6). Medians ± quartiles are shown in the box plots.\u003c/p\u003e\n\u003cp\u003eFor panels h and i, p values obtained from linear model correcting for participant ages (in blue) and from two-sided Mann-Whitney U test (in black) are shown. \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05 was considered significant.\u003c/p\u003e\n\u003cp\u003eFor all panels except i, adenoid: INF N = 14, VAC N = 9; tonsil: INF N = 22, VAC N = 7, (Supplemental Table 6).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/dcb6f652d4684c81a4743c77.png"},{"id":90884700,"identity":"8fad42f6-3339-4528-ab3e-cb1e7474593b","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1060108,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eEvaluation of B\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003eSM\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e based on CXCR3 and CD21 expression\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a) Left: Representative gating of B\u003csub\u003eSM\u003c/sub\u003e by CXCR3 and CD21 expression yielding 4 populations (P1-P4). Right: Proportions of P1 (CXCR3\u003csup\u003e-\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e), P2 (CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e), P3 (CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e-\u003c/sup\u003e), and P4 (CXCR3\u003csup\u003e-\u003c/sup\u003eCD21\u003csup\u003e-\u003c/sup\u003e) among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e and total B\u003csub\u003eSM\u003c/sub\u003e in PBMCs, adenoids, and tonsils of VAC and INF subjects. Samples with \u0026lt; 10 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e were excluded in this figure (PBMC: INF N = 13, VAC N = 9; adenoid: INF N = 14, VAC N = 8; tonsil: INF N = 22, VAC N = 6). Mean frequencies are plotted (Supplemental Table 4).\u003c/p\u003e\n\u003cp\u003e(b) Comparison of percentage of P1-P4 among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in PBMCs, adenoids, and tonsils of VAC versus INF subjects. Same samples as panel a. Medians ± quartiles are shown in the box plots.\u003c/p\u003e\n\u003cp\u003e(c-d) Histogram of selected marker/antibody expression on P1-P4 subsets among B\u003csub\u003eSM\u003c/sub\u003e in PBMC (c) and tonsil (d) sample. Representative sample among PBMC (N = 3) and tonsil (N = 3) are shown (Samples listed in Supplemental Table 2).\u003c/p\u003e\n\u003cp\u003e(e) Correlation between percentages of P1-P4 among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in PBMCs, adenoids and tonsils and serum neutralizing titers to WA.1. Correlations were assessed with Spearman’s rank correlation. Correlation coefficients (r values) are indicated by colors shown in the bar and size of the circle indicates absolute r value. INF and VAC groups combined for analysis (Samples in (a) with available neutralizing titers were analyzed. PBMC N = 22, adenoid N= 21, tonsil N = 27). * p \u0026lt; 0.05, ** p\u0026lt; 0.01.\u003c/p\u003e\n\u003cp\u003e(f) Percentage of CD69\u003csup\u003e+\u003c/sup\u003e cells among S1\u003csup\u003e+\u003c/sup\u003eRBD B\u003csub\u003eSM\u003c/sub\u003e and total B\u003csub\u003eSM\u003c/sub\u003e in PBMCs, adenoids, and tonsils of VAC versus INF subjects. Samples as those from panel a. Medians ± quartiles are shown in the box plots.\u003c/p\u003e\n\u003cp\u003eIn panels b and f, p values obtained from linear model correcting for participant ages (in blue) and from two-sided Mann-Whitney U test (in black) are shown in panels b and f. \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05 were considered significant.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/5743446a72a590268af59859.png"},{"id":90886229,"identity":"da7a12c4-e658-4fe8-a832-a13649d5e737","added_by":"auto","created_at":"2025-09-09 10:08:52","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1867318,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTranscriptomic features of tissue S1\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e B\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003eSM\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e in INF and VAC individuals\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a) UMAP of unsupervised clustering based on surface protein expression from CITE-seq of sorted 9370 S1\u003csup\u003e+\u003c/sup\u003e and 172919 S1\u003csup\u003e-\u003c/sup\u003e CD19\u003csup\u003e+\u003c/sup\u003e B cells from adenoids and tonsils (adenoid: INF N = 6, VAC N = 3, CON N = 1; tonsil: INF N = 6, VAC N = 2, CON N = 2; Supplemental Table 8).\u003c/p\u003e\n\u003cp\u003e(b) Expression of CD38 and IgD surface proteins by cluster.\u003c/p\u003e\n\u003cp\u003e(c) Distribution of S1\u003csup\u003e+\u003c/sup\u003e B cells among clusters on UMAP in INF and VAC tissues.\u003c/p\u003e\n\u003cp\u003e(d) Proportion of each cluster among S1\u003csup\u003e+\u003c/sup\u003e and S1\u003csup\u003e-\u003c/sup\u003e B cells from INF and VAC subjects. Mean proportion among each group of subjects is plotted.\u003c/p\u003e\n\u003cp\u003e(e) Heatmap showing expression of GCB signature genes among S1\u003csup\u003e+\u003c/sup\u003e B cells by cluster (samples are from sample set 306-4 in Supplemental Table 8).\u003c/p\u003e\n\u003cp\u003e(f) GCB signature gene module scores in S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e and S1\u003csup\u003e+ \u003c/sup\u003eGCB from INF versus VAC tissues compared with two sided Mann-Whitney U test. Same samples as in panel e.\u003c/p\u003e\n\u003cp\u003e(g) Representative CXCR3 and CD21 expression on concatenated S1\u003csup\u003e+ \u003c/sup\u003eB\u003csub\u003eSM\u003c/sub\u003e and S1\u003csup\u003e-\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e from INF and VAC tissues of CITE-seq sample set 306-4 (see Methods).\u003c/p\u003e\n\u003cp\u003e(h) Proportion of each B cell subset (gated manually by surface marker expression) among S1\u003csup\u003e+\u003c/sup\u003e and S1\u003csup\u003e-\u003c/sup\u003e B cells of INF and VAC subjects. Same samples as in panel a.\u003c/p\u003e\n\u003cp\u003e(i) Trajectory analysis of S1\u003csup\u003e+ \u003c/sup\u003eB cells from INF and VAC subjects using Slingshot. The naïve/unswitched memory B cell population was assigned as the starting point for trajectory construction, and B cell subsets were annotated as shown in panel h.\u003c/p\u003e\n\u003cp\u003e(j) Heatmap showing gene enrichment scores for selected pathways in S1\u003csup\u003e-\u003c/sup\u003e B cell subsets defined by manual gating of surface markers. Color represents z score. MBC = memory B cell\u003c/p\u003e\n\u003cp\u003e(k) Differentially expressed genes (DEGs) identified by comparing pooled pseudobulk libraries of P1 versus P2 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM \u003c/sub\u003ewith a linear model accommodating for paired analysis. Selected genes with adjusted p value \u0026lt; 0.05 are labeled on the volcano plot (top 40 DEGs are listed in Supplemental Table 9). Same samples as in panel a.\u003c/p\u003e\n\u003cp\u003e(l) Select pathways from gene set enrichment analysis (GSEA) of ranked genes in P2 versus P1 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e. Dot size denotes normalized enrichment score (NES) and color denotes -log10(adjusted \u003cem\u003ep \u003c/em\u003evalue).\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/aefa0482cc99ee108c9c21ea.png"},{"id":90884697,"identity":"09c02584-a55d-42f4-874f-c4ee1797bac2","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1094075,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSingle cell BCR sequencing of S1\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+ \u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eB cells in INF and VAC groups\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a) Proportion of expanded (clone size \u0026gt; 1 cell) clones among S1\u003csup\u003e+ \u003c/sup\u003eCD19\u003csup\u003e+\u003c/sup\u003e B cells from\u003csup\u003e \u003c/sup\u003eINF and VAC participants.\u003c/p\u003e\n\u003cp\u003e(b) Simpson’s diversity of S1\u003csup\u003e+\u003c/sup\u003e and S1\u003csup\u003e−\u003c/sup\u003e B cells from adenoids and tonsils from 8 INF and 4 VAC donors and S1\u003csup\u003e−\u003c/sup\u003e B cells from 2 CON. Lower Simpson’s diversity indicates a greater frequency of large clones.\u003c/p\u003e\n\u003cp\u003e(c) Clonal expansion index of S1\u003csup\u003e+\u003c/sup\u003e B cells in each B cell subset in INF and VAC tissues.\u003c/p\u003e\n\u003cp\u003e(d) Somatic hypermutation (SHM) frequency of S1\u003csup\u003e+\u003c/sup\u003e or S1\u003csup\u003e-\u003c/sup\u003e B cell subsets.\u003c/p\u003e\n\u003cp\u003e(e-f) Pairwise transition index (pTrans) showing degree of BCR clone overlap between P2 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e and S1\u003csup\u003e+ \u003c/sup\u003eB cell subsets (e) and between P3 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e and S1\u003csup\u003e+ \u003c/sup\u003eB cell subsets (f) from adenoids and tonsils of 8 INF and 4 VAC donors.\u003c/p\u003e\n\u003cp\u003e(g) Heatmap showing clonal overlap between S1\u003csup\u003e+\u003c/sup\u003e or S1\u003csup\u003e-\u003c/sup\u003e B cell subsets from one representative donor who had the largest number of sorted S1\u003csup\u003e+\u003c/sup\u003e B cells (CNMC 70, 2413 S1\u003csup\u003e+\u003c/sup\u003e B cells). Off-diagonal elements are colored by the Jaccard index indicating degree of clonal overlap and are labeled by the raw number of overlapping clones. Diagonal elements are labeled by the total number of clones within a particular subset.\u003c/p\u003e\n\u003cp\u003e(h-i) Clonal lineage trees selected from the two largest GCB bearing S1\u003csup\u003e+ \u003c/sup\u003etrees from INF (h) and VAC (i) donors. Triangles indicate S1\u003csup\u003e+\u003c/sup\u003e cells and tip color indicates B cell subset. Isotype and tissue origin of each clone are listed next to the symbol. Branch lengths represent SHM frequency/codon in VDJ sequence according to the scale bar.\u003c/p\u003e\n\u003cp\u003eSamples for a-f are listed in Supplemental Table 8; adenoid: INF N = 6, VAC N = 3, CON N = 1; tonsil: INF N = 6, VAC N = 2, CON N = 2. Significance for b-d calculated with two-sided Mann-Whitney U test. Medians ± quartiles are shown in the box plots.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/a0df1867b642eec445f36cf1.png"},{"id":90884701,"identity":"1fb9c65c-369b-4ad9-9bbc-c552eb9573e8","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1031375,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCXCR3\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003eCD21\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+ \u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e(P2) B\u003c/strong\u003e\u003csub\u003e\u003cstrong\u003eSM\u003c/strong\u003e\u003c/sub\u003e\u003cstrong\u003e are predisposed to plasma cell differentiation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a)\u0026nbsp; Phosphorylation of Syk and PLCg2 in S1⁺ B\u003csub\u003eSM \u003c/sub\u003esubsets, concatenated by group, with and without 2-minute stimulation with soluble anti-IgG, -IgA, and -IgM. Representative plots are shown (N = 10 tonsils).\u003c/p\u003e\n\u003cp\u003e(b)\u0026nbsp;\u0026nbsp;\u0026nbsp; Percentage of cells expressing either or both p-Syk and/or p-PLCg2 among P1 versus P2 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e post-BCR stimulation (percentages of p-Syk and/or p-PLCg2 from untreated conditions were subtracted from stimulated conditions). Same samples as in panel a.\u003c/p\u003e\n\u003cp\u003e(c-d) Geometric mean fluorescence intensity (gMFI) of surface IgG or IgA on their respective isotype-specific B cell subsets from PBMC (c) and tonsil/adenoid (d) (PBMC N = 37, adenoid N = 33, tonsil N = 41).\u003c/p\u003e\n\u003cp\u003e(e) gMFI of IgG on P1 versus P2 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e IgG\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in INF and VAC groups. Samples with at least 10 cells in both P1 and P2 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e IgG\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e populations were analyzed (PBMC: INF N = 6, VAC N = 8; adenoid: INF N = 12, VAC N = 7; tonsil: INF N = 17, VAC N = 6).\u003c/p\u003e\n\u003cp\u003e(f) Percentages of plasmablasts (CFSE\u003csup\u003e+\u003c/sup\u003eCD38\u003csup\u003e+\u003c/sup\u003eCD20\u003csup\u003e-\u003c/sup\u003e) after 4 days of \u003cem\u003ein vitro\u003c/em\u003e culture of sorted P1 and P2 B\u003csub\u003eSM \u003c/sub\u003efrom tonsils with B cell depleted PBMCs from an unrelated donor, R848, and IL-2. Fold change comparing P1 and P2 was calculated from paired P1 versus P2 B\u003csub\u003eSM\u003c/sub\u003e from three tonsils.\u003c/p\u003e\n\u003cp\u003e(g) Division index of CFSE-labelled P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e following \u003cem\u003ein vitro \u003c/em\u003eculture described in Figure 7f.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;P values for panels b-g were calculated with two-sided Wilcoxon signed ranks test (paired).\u0026nbsp; Samples used in this figure are listed in Supplemental Table 2. In f and g, mean ± SD are shown in the box plots.\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/7d186cf2b3d52d62232761de.png"},{"id":90884699,"identity":"4bc8ea59-f8f3-4d14-a6be-b416aa4bd9f4","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":3846213,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCXCR3\u003c/strong\u003e\u003csup\u003e\u003cstrong\u003e+\u003c/strong\u003e\u003c/sup\u003e\u003cstrong\u003e cells and CXCL9-expressing cells co-localize in the tissue\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(a) Percentages of CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e-\u003c/sup\u003eCD57\u003csup\u003e+\u003c/sup\u003ePD1\u003csup\u003ehi\u003c/sup\u003e and CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e- \u003c/sup\u003epre-Tfh among CD4\u003csup\u003e+\u003c/sup\u003e T cells in CON (N = 10), INF (N = 14) and VAC (N = 9) adenoids. Samples listed in Supplemental Table 2.\u003c/p\u003e\n\u003cp\u003e(b) gMFI of HLA-DR expression on CD14\u003csup\u003e+ \u003c/sup\u003emonocytes, conventional dendritic cells (cDC) and plasmacytoid dendritic cells (pDC) in CON, INF and VAC adenoids. Same samples as in panel a. For panels a and b, p values obtained from linear model correcting for participant ages (in blue) and from two-sided Mann-Whitney U (in black) are shown. \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05 were considered significant. Medians ± quartiles are shown in the box plots. Comparisons of VAC versus CON were not statistically significant in both panels (not shown).\u003c/p\u003e\n\u003cp\u003e(c) Correlations between neutralizing titers to omicron BA.1 and WA-1 strains, percentages of P1-P4 among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e, CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e-\u003c/sup\u003eCD57\u003csup\u003e+\u003c/sup\u003ePD1\u003csup\u003ehi\u003c/sup\u003e among CD4\u003csup\u003e+\u003c/sup\u003e T cells, and CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e-\u003c/sup\u003e preTfh among CD4\u003csup\u003e+\u003c/sup\u003e T cells and HLA-DR gMFI of pDC, cDC, and CD14\u003csup\u003e+\u003c/sup\u003e monocytes. Data from tonsils and adenoids and INF and VAC were pooled (N = 49). Correlations were assessed with Spearman’s rank correlation. Correlation coefficients (r values) are indicated by color shown in the bar. Size of the circle indicates absolute r values. Significant p values are indicated as * p\u0026lt;0.05, ** p\u0026lt;0.01, *** p\u0026lt;0.001.\u003c/p\u003e\n\u003cp\u003e(d) Immunofluorescence images of tonsil from one INF donor, representative of paired tonsil and adenoid from one INF and one VAC donor, with major histologic regions delineated (top left). Insets show greater magnifications. Bottom panels show CD20 (blue), CD4 (purple), BCL6 (yellow), and CD138 (plasma cells and epithelial cells, white) with and without CXCR3 channel (green) shown. Arrows indicate CXCR3\u003csup\u003e+\u003c/sup\u003eCD20\u003csup\u003e+\u003c/sup\u003e B cells (arrow 1) and CXCR3\u003csup\u003e+\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e T cells (arrow 2).\u003c/p\u003e\n\u003cp\u003e(e) Images showing the localization of CXCR3⁺BCL6\u003csup\u003e-\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T and CD20\u003csup\u003e+\u003c/sup\u003e B cells (left panel) and CXCR3⁻BCL6\u003csup\u003e-\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T and CD20\u003csup\u003e+\u003c/sup\u003e B cells (right panel) in the representative tonsil from panel d.\u003c/p\u003e\n\u003cp\u003e(f) Spatial distribution of 8 cell subsets defined by CD4, CD20, BCL6, and CXCR3 expression, shown in the representative tonsil from (d) using SPACE at 10 mm scale. Peaks of local representation of each population along the latent path indicate their spatial proximity to each other. BCL6\u003csup\u003e+\u003c/sup\u003e and BCL6\u003csup\u003e-\u003c/sup\u003e cells were gated with a gap in between, which is reflected as a gap in the latent path.\u003c/p\u003e\n\u003cp\u003e(g)\u0026nbsp;\u0026nbsp;\u0026nbsp; Annotated cell types based on transcriptome are shown in nearby section of the representative tonsil from panel d run on Xenium In Situ spatial transcriptomic platform using human 5000 gene panel. 521,787,433 transcripts and 1,834,484 cells were detected after quality filtering.\u003c/p\u003e\n\u003cp\u003e(h-j) Spatial map showing the location of monocytes/macrophages (Mono/Macro) (h), activated dendritic cells (aDC) (i), pDC (j) and cycling dark zone GCB (DZ GCB) to indicate germinal centers from the representative tonsil from panel d.\u003c/p\u003e\n\u003cp\u003e(k) Spatial map showing the location and expression levels of CXCL9, both overall and specifically within mono/macro and aDC cells, in the representative tonsil from (d).\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/a2dd65878c9476ab5a8e4910.png"},{"id":92123044,"identity":"fe0d2c6f-1bf9-4bce-a81d-fe6824bf2810","added_by":"auto","created_at":"2025-09-24 23:08:32","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":16503250,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/87f08bcf-21f3-4d5a-ad1c-2aaa44e68e27.pdf"},{"id":90884696,"identity":"faeee904-8592-481d-8af9-dff2807dcb0b","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":99364,"visible":true,"origin":"","legend":"Supplemental Tables 1 - 10","description":"","filename":"20250805SupplementaryTable110.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/df5c0b5cea80c5d4f28149f5.xlsx"},{"id":90886227,"identity":"7143ea5e-a5e4-425d-9a51-2bdec6e017ee","added_by":"auto","created_at":"2025-09-09 10:08:52","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":32179,"visible":true,"origin":"","legend":"Supplemental Table 11. Reagents","description":"","filename":"20250805SupplementalTable11.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/eee3156177f773e65def0406.xlsx"},{"id":90884702,"identity":"50d1d53f-9861-4ccc-b702-0901a4f798c6","added_by":"auto","created_at":"2025-09-09 10:00:52","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":32841977,"visible":true,"origin":"","legend":"Supplementary Figures","description":"","filename":"20250813SupplementaryFigures.docx","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/be3b03399958b5a7a1e08478.docx"},{"id":90884703,"identity":"c8bb1728-1658-47b5-902f-2aac656cc609","added_by":"auto","created_at":"2025-09-09 10:00:53","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":42315481,"visible":true,"origin":"","legend":"","description":"","filename":"EXTENDEDDATAFIGURES.docx","url":"https://assets-eu.researchsquare.com/files/rs-7428491/v1/27b605dab24837455244fd11.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"SARS-CoV-2 infection and vaccination elicit distinct pharyngeal mucosal B cell responses in children","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eThe high mortality and far-reaching effects of the COVID-19 pandemic triggered the rapid development of several vaccine platforms including two mRNA-based vaccines, BNT162b2 (Pfizer) and mRNA-1273 (Moderna), which were shown to have high efficacy in preventing severe COVID-19 \u003csup\u003e1, 2\u003c/sup\u003e. These vaccines generate high-titers of serum antibodies with neutralizing activity against early and more later circulating SARS-CoV-2 strains, as well as virus-specific memory B and T cells in the peripheral blood\u003csup\u003e3, 4, 5\u003c/sup\u003e. However, the immunity afforded by these vaccines wanes with time and population-based studies suggest that the durability of immune protection following vaccination may be lower than that from prior infection with SARS-CoV-2\u003csup\u003e6, 7, 8, 9\u003c/sup\u003e. SARS-CoV-2 initially enters and infects upper respiratory tract tissues \u003csup\u003e10, 11\u003c/sup\u003e, and immunity in the upper respiratory tract, particularly mucosal IgA levels, has been shown to correlate with protection against COVID-19 \u003csup\u003e12, 13\u003c/sup\u003e. Therefore, characterizing upper respiratory tract immunity provided by these intramuscularly delivered vaccines is important to understand their efficacy and to provide insights for future vaccine development against a variety of pathogens.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA major component of protective immune memory to viruses, including SARS-CoV-2, is the development and maintenance of long-lasting, high-affinity memory B cells, which are poised to rapidly differentiate into antibody-secreting cells upon secondary antigen encounter \u003csup\u003e14, 15\u003c/sup\u003e. A subset of memory B cells in the tissues, known as tissue-resident memory B cells, do not recirculate and have been shown to provide effective localized immunity \u003csup\u003e16\u003c/sup\u003e. Secondary lymphoid organs enable the development of germinal centers (GC), which are essential for creating long-lived and high-affinity memory B cells. In these specialized structures, B cells undergo somatic hypermutation (SHM) and affinity maturation with help from T follicular helper (Tfh) cells. COVID-19 mRNA vaccines have been shown to generate SARS-CoV-2-specific Tfh and GC B cells (GCBs) in the draining axillary lymph nodes, which can persist at least 6 months\u003csup\u003e17, 18, 19, 20\u003c/sup\u003e; however, the ability of intramuscular vaccines to elicit immunity in upper respiratory tract lymphoid tissues remains an important question. Furthermore, whether immune memory in the tissue differs after vaccination versus infection is largely unknown.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe tonsils and adenoids are secondary lymphoid structures at the mucosal surface of the upper respiratory tract that contain lymphoid cells not found in the peripheral blood; viral-specific GCBs and tissue-resident memory T and B cells have been described in these mucosal tissues following respiratory infections \u003csup\u003e21, 22, 23\u003c/sup\u003e. Assessing SARS-CoV-2-specific-B cells in these tissues following intramuscular vaccination may, therefore, offer insights into the characteristics, magnitude, and durability of tissue immunity at the sites where the host first encounters the virus.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHere, we characterized SARS-CoV-2-specific B cells in the tonsils, adenoids, and peripheral blood of children immunized with a SARS-CoV-2 mRNA vaccine compared with those of children previously infected with SARS-CoV-2, taking advantage of the unique period during the COVID-19 pandemic to study immunity to an unencountered airborne pathogen and new vaccines. We found SARS-CoV-2-specific B cells, which were primarily switched memory B cells (B\u003csub\u003eSM\u003c/sub\u003e) but also some GCBs, in the tonsils and adenoids of children both post-infection and post-vaccination, indicating that immune memory can be found and maintained in the upper respiratory tract after intramuscular vaccination as well as after infection. Nonetheless, vaccination and infection generated B\u003csub\u003eSM\u003c/sub\u003e with different characteristics, including increased induction of a population defined by CXCR3 and CD21 expression post-infection. This CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e population showed a greater propensity for plasma cell differentiation and mucosal homing, had unique spatial distribution, and correlated with persistent adaptive and innate immune cell activation in mucosal lymphoid tissues. Our findings provide a framework for understanding responses to vaccination and infection, including novel parameters for assessing vaccine-induced mucosal immunity. \u0026nbsp;\u003c/p\u003e"},{"header":"RESULTS","content":"\u003cp\u003e\u003cstrong\u003eSARS-CoV-2-specific B cells are found in the pharyngeal tissues and blood post-vaccination\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo assess SARS-CoV-2-specific immunity in vaccinated individuals, we collected serum, peripheral blood mononuclear cells (PBMC), and tonsil and adenoid tissues from 21 children undergoing tonsillectomy and/or adenoidectomy from December 2021 to September 2022 at Children’s National Hospital in Washington, DC, USA, who had received at least one dose of the monovalent mRNA vaccines, BNT162b2 or mRNA-1273 \u0026nbsp;(Figure 1a, Supplemental Tables 1 and 2). These vaccines encode the ancestral Wuhan strain-like spike protein. Of these vaccinated subjects, 10 had no evidence of prior SARS-CoV-2 infection by history of positive PCR or antigen test, nor by positive serum titers against nucleocapsid (NC) and/or open reading frame 8 (ORF8), which have been used for serodiagnosis of SARS-CoV-2 infection \u003csup\u003e24, 25\u003c/sup\u003e (Supplemental Table 3). These 10 subjects comprised our vaccinated only cohort (VAC).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe compared these vaccinated children to subjects we previously recruited from late 2020 to early 2021 who had prior SARS-CoV-2 infection (INF), before the availability of vaccination for children (Figure 1a, Supplemental Tables 1 and 2) \u003csup\u003e23\u003c/sup\u003e. All had mild or asymptomatic infection. Based on the timing of sample collection and known infection dates, most of these subjects were likely infected with the D614G or alpha strains, which bear sequence similarity to the spike mRNA used in the BNT162b2 or mRNA-1273 vaccines administered to the VAC cohort. The interval from infection to surgery among INF participants was comparable to the interval from the last vaccination to surgery in the VAC group (Figure 1b, Extended Data Figure 1a). We also included a group of unvaccinated pediatric controls (CON) with no serologic or cellular evidence of prior COVID-19, who were recruited during our initial study (Figure 1a, Supplemental Table 2 and 3). To compare VAC and INF groups, we used both direct statistical comparisons (Mann-Whitney U) and linear models correcting for age. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAll VAC and INF participants had serum neutralizing antibodies to the WA-1 strain (Figure 1c, Supplemental Table 3) with no significant differences noted between the two groups. Similar to our INF cohort, VAC subjects showed a trend towards lower neutralizing titers to WA-1 with greater time from infection (Figure 1d), consistent with previous reports\u003csup\u003e26\u003c/sup\u003e. We also evaluated neutralizing titers to omicron, a variant that emerged in late 2021 and rapidly became the dominant strain, causing numerous breakthrough infections in those who were previously infected or vaccinated \u003csup\u003e27, 28\u003c/sup\u003e. Although neutralizing titers were lower to omicron than to WA-1 in both groups, they were on average higher in VAC than INF, with a higher proportion of VAC subjects having positive titers (Figure 1c, Supplemental Table 3).\u003c/p\u003e\n\u003cp\u003eWe then used fluorescently-labeled probes for the receptor-binding domain (RBD) and S1 portion of the spike protein from the original Wuhan strain to identify SARS-CoV-2-recognizing B cells (S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e) by flow cytometry (Figure 1e). As previously reported, S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e CD19\u003csup\u003e+\u003c/sup\u003e B cells were found in both tissues and blood of most INF participants (Supplemental Table 3). Notably, almost all VAC subjects also had S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e CD19\u003csup\u003e+\u003c/sup\u003e B cells in both their pharyngeal tissues and blood (Figure 1f, Supplemental Table 3). A higher frequency of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e cells was noted among B cells in peripheral blood of VAC compared to INF individuals but these trended towards a higher frequency in the adenoids of INF subjects (Figure 1f). These findings were confirmed comparing S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in the blood versus tissues within individual subjects by group (Figure 1g). Among INF subjects, the percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in the PBMCs, adenoids, and tonsils were all significantly correlated, as previously reported (Figure 1h) \u003csup\u003e23\u003c/sup\u003e. In contrast, among VAC subjects, the percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in the PBMCs correlated significantly with serum neutralizing titers, but not significantly with the percentages of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in the adenoid and tonsil (Figure 1h).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe also identified omicron-recognizing B cells (RBD-Omi\u003csup\u003e+\u003c/sup\u003e) using two fluorescently labelled RBD probes from the omicron strain (Extended Data Figure 1b-c). VAC participants had a higher proportion of detectable S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells that also recognized omicron in the blood, adenoid and tonsil compared to INF (Extended Data Figure 1d), suggesting vaccination provides broader coverage of variants than infection in both the blood and tissues\u003csup\u003e4, 19, 29\u003c/sup\u003e. \u0026nbsp;Thus, although vaccinated children had greater B cell responses in the peripheral blood compared to the mucosal tissues, we could detect SARS-CoV-2-specific B cells in the secondary lymphoid tissue of the upper respiratory tract, distal from the site of intramuscular immunization in nearly all vaccinated children.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInfection induces a higher proportion of IgA\u003csup\u003e+\u003c/sup\u003e SARS-CoV-2-specific B\u003csub\u003eSM\u0026nbsp;\u003c/sub\u003ethan vaccination\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo provide insight into SARS-CoV-2-specific B cells, we used a high dimensional flow cytometry panel consisting of 29 markers including fluorescently labelled SARS-CoV-2 probes as well as surface markers to describe B cell subsets, isotype, activation, tissue residence, and homing. In both INF and VAC, the majority of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in PBMCs, tonsils, and adenoids were B\u003csub\u003eSM\u003c/sub\u003e (Figure 1i). Unlike neutralizing titers which declined with greater time from vaccination/infection, the percentage of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in the tissues were stable (VAC) or increased (INF) with time (Figure 1j). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMucosal IgA is protective against SARS-CoV-2 infection, but mRNA vaccination has been shown to engender less mucosal IgA in the respiratory tract compared to infection \u003csup\u003e9, 12, 30, 31\u003c/sup\u003e. SARS-CoV-2-specific B\u003csub\u003eSM\u003c/sub\u003e were predominantly IgG\u003csup\u003e+\u003c/sup\u003e post-infection and post-vaccination (Figure 1k). However, INF individuals had a greater proportion of IgA\u003csup\u003e+\u003c/sup\u003e cells (and less IgG\u003csup\u003e+\u003c/sup\u003e cells) among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in PBMC and tissues compared to VAC subjects (Figure 1k, Extended Data Figure 1e). Thus, infection with early circulating strains was associated with a greater proportion of IgA\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e than vaccination, both in tissues and blood.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn humans, B\u003csub\u003eSM\u0026nbsp;\u003c/sub\u003ecan be subdivided based on expression of CD21 and CD27\u003csup\u003e32, 33\u003c/sup\u003e. CD27\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e conventional or “resting” memory B cells (cMBCs) are quiescent, highly affinity-matured cells that emerge from GC reactions. Following infection or vaccination, responding B cells become activated and display heterogenous phenotypes including downregulation of CD21 \u003csup\u003e15, 34, 35, 36\u003c/sup\u003e \u003csup\u003e37, 38\u003c/sup\u003e.\u0026nbsp;Those that express CD27 (CD27\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e-\u003c/sup\u003e) are referred to as activated memory cells (acMBC), while those that express low CD21 and CD27 are called atypical MBCs (atMBC); both are expanded in chronic infections, autoimmune diseases, and shortly after acute infection and vaccination including COVID-19 vaccines \u003csup\u003e35, 38, 39, 40, 41, 42, 43, 44\u003c/sup\u003e. The atMBCs (and some acMBCs) express CD11c, FCRL3, FCRL4, FCRL5, and CD85J, as well as Tbet \u003csup\u003e45, 46, 47\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMost of the SARS-CoV-2-specific B\u003csub\u003eSM\u003c/sub\u003e in both VAC and INF tissues and blood were cMBC (Extended Data Figure 1f-g, Supplemental Table 4). However, a larger proportion of acMBCs were noted in VAC PBMC and tonsils compared to INF, with a similar trend in the adenoids (Extended Data Figure 1f-g). Neutralizing titers correlated with the proportion of acMBCs among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in both blood and adenoids (Extended Data Figure 1h). \u0026nbsp;Furthermore, like neutralizing titers, the frequency of CD21\u003csup\u003elo/-\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e populations (acMBC and atMBC) in the blood declined with time from vaccination (Extended Data Figure 1i).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSARS-CoV-2-specific B\u003csub\u003eSM\u003c/sub\u003e in the blood are phenotypically different post-infection and post-vaccination\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further compare the phenotypes of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells between the two groups beyond traditional B\u003csub\u003eSM\u003c/sub\u003e markers, we concatenated the S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells from all VAC and INF participants and performed unsupervised clustering analyses based on expression of 19 B cell surface markers. Cells from the peripheral blood and tissues were assessed separately due to differences in cell populations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUnsupervised clustering of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells from PBMCs generated 12 phenotypically distinct clusters, 11 of which were B\u003csub\u003eSM\u003c/sub\u003e subsets (Figure 2a-c, Supplemental Table 5). VAC and INF samples largely segregated on principal component analysis (PCA) on PC2; a few VAC subjects were distinct on PC1 (Figure 2d). We compared the proportion of each cluster in the two groups by Mann-Whitney U or a linear model correcting for age (Figure 2c, e; Extended Data Figure 2a), and found that INF subjects had higher proportions of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in cluster 10 with a trend towards more in clusters 2 and 6 compared to VAC individuals; these three clusters represented CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u0026nbsp;\u003c/sub\u003e(Figure 2b, e-h). Cluster 10, the most significant cluster, represented IgA\u003csup\u003e+\u003c/sup\u003eCXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e. CXCR3 is a chemokine receptor induced by IFN-g\u0026nbsp;that directs B\u003csub\u003eSM\u003c/sub\u003e to areas of inflammation\u003csup\u003e48\u003c/sup\u003e. Although CXCR3 can be expressed on CD21\u003csup\u003e-\u003c/sup\u003e atMBCs along with CD11c and other markers regulated by Tbet \u003csup\u003e49, 50\u003c/sup\u003e, these three clusters did not express atMBC markers and were largely CD21\u003csup\u003e+\u003c/sup\u003e (Figure 2a-b, h) \u003csup\u003e5, 42\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eConversely, VAC subjects had a higher portion of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in several CD21\u003csup\u003elo\u003c/sup\u003e clusters, including clusters 1, 3, 4 and 7 (CD21\u003csup\u003elo\u003c/sup\u003eIgG\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e), and cluster 11 (CD21\u003csup\u003elo\u003c/sup\u003eIgA\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e cluster found in 2 subjects) (Figure 2a-c, e-h). Clusters 4, 7 and 11 expressed atMBC markers including CD11c, CD85J, FCRL3/5 and CD95 and had a mixture of CD27\u003csup\u003e+\u003c/sup\u003e and CD27\u003csup\u003e-\u003c/sup\u003e cells but were not CXCR3\u003csup\u003e+\u003c/sup\u003e (Figure 2b). Cluster 3 was an IgG\u003csup\u003e+\u0026nbsp;\u003c/sup\u003eCD21\u003csup\u003e-\u0026nbsp;\u003c/sup\u003eCD27\u003csup\u003e+\u003c/sup\u003e acMBC cluster that expressed variable CXCR3. Thus, unbiased analyses revealed phenotypic differences in SARS-CoV-2-specific B cells in the peripheral blood post-infection and post-vaccination, including enrichment of CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e post-infection.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDistinct SARS-CoV-2-specifc GCBs and B\u003csub\u003eSM\u003c/sub\u003e are in the pharyngeal tissues of INF and VAC subjects\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe next examined the characteristics of SARS-CoV-2-specific B cells in the pharyngeal tissues, generating 14 clusters representing naïve and unswitched memory B cells (USM), GC and pre-GC B cells (CD38\u003csup\u003e+\u003c/sup\u003eCD10\u003csup\u003e+\u003c/sup\u003eCD95\u003csup\u003e+\u003c/sup\u003eCD71\u003csup\u003e+\u003c/sup\u003e), and B\u003csub\u003eSM\u0026nbsp;\u003c/sub\u003e(Figure 3a-c, Extended Data Figure 2b-c, Supplemental Table 6). Like PBMCs, PCA analyses showed separation between VAC and INF subjects (Figure 3d).\u003c/p\u003e\n\u003cp\u003eWe previously found that SARS-CoV-2 infection induces local GC reactions in the pharyngeal lymphoid tissues\u003csup\u003e23\u003c/sup\u003e. As expected, S1\u003csup\u003e+\u003c/sup\u003e cells were observed in GCB clusters in most (15/24, 62.5%) INF individuals; however, two out of 10 (20%) VAC subjects also had at least 10 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells within GCB clusters in either the tonsil or adenoid, a threshold based on a recent study\u003csup\u003e51\u003c/sup\u003e (Figure 3e; Supplemental Table 7). These two VAC subjects had their most recent (second) vaccine doses 48 days and 232 days prior to surgery, suggesting this was not temporally associated with recent vaccination (Supplemental Table 2). INF tonsils had a higher proportion of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in the IgG\u003csup\u003e+\u003c/sup\u003e GCB cluster (cluster 6) and IgA\u003csup\u003e+\u003c/sup\u003e GCB cluster (cluster 11, by Mann-Whitney U) than VAC (Figure 3a-c, f-h; Extended Data Figure 2d), supporting the greater effectiveness of natural infection in generating and maintaining GC responses in pharyngeal mucosal tissues over vaccination.\u003c/p\u003e\n\u003cp\u003eSimilar to PBMCs, VAC tissues had a higher frequency of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells in certain clusters containing CD27\u003csup\u003e+\u003c/sup\u003eIgG\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e with lower or intermediate CD21 expression and lacking CXCR3 expression. These included a greater proportion in cluster 2 in VAC tonsils, and in cluster 7 (CD62L\u003csup\u003e+\u003c/sup\u003e) in VAC adenoids and tonsils compared to INF (Figure 3a-c, f-h; Extended Data Figure 2c-d).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn contrast, tonsils and adenoids of INF subjects contained a higher fraction of SARS-COV-2-specific B cells in CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e containing clusters compared to VAC individuals including IgA\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e populations (clusters 4 and 12) and IgG\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u0026nbsp;\u003c/sub\u003epopulations (clusters 5 and 10 in the adenoids and cluster 3 in tonsils) (Figure 3a-c, f-h, Extended Data Figure 2c-d). Increased percentages of CXCR3\u003csup\u003e+\u003c/sup\u003e IgA and IgG cells in INF subjects were confirmed by manual gating (Figure 3i). CXCR3\u003csup\u003e+\u003c/sup\u003e has been shown to be important for establishment of tissue-residenceamong memory B cells in murine lungs and production of mucosal IgA in mice, in addition to attracting cells to the mucosa \u003csup\u003e52, 53, 54\u003c/sup\u003e. Although we also observed CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in the blood, we saw a trend towards higher proportions of CXCR3\u003csup\u003e+\u003c/sup\u003e cells, particularly CXCR3\u003csup\u003e+\u003c/sup\u003eIgA\u003csup\u003e+\u003c/sup\u003e cells, among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u0026nbsp;\u003c/sup\u003eB\u003csub\u003eSM\u003c/sub\u003e in the adenoid and tonsil compared to blood, supporting the idea that CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e may preferentially home to mucosal tissues (Extended Data Figure 2e).\u003c/p\u003e\n\u003cp\u003eThus, infection and vaccination were associated with phenotypically distinct memory B cell populations with a higher proportion of CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e including CXCR3\u003csup\u003e+\u003c/sup\u003eIgA\u003csup\u003e+\u003c/sup\u003e cells post-infection, and B\u003csub\u003eSM\u003c/sub\u003e populations with lower CD21 expression and lacking CXCR3 post-vaccination in the tissues, analogous to the peripheral blood. Furthermore, a small portion of VAC individuals showed evidence of SARS-CoV-2-specific GCBs in the pharyngeal lymphoid tissue.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEvaluation of B\u003csub\u003eSM\u003c/sub\u003e based on CXCR3 and CD21 expression\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGiven the differential expression of CXCR3 and CD21 on S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e from INF and VAC individuals, we used these two markers to define 4 populations among B\u003csub\u003eSM\u003c/sub\u003e: P1 (CD21\u003csup\u003e+\u003c/sup\u003eCXCR3\u003csup\u003e-\u003c/sup\u003e); P2 (CD21\u003csup\u003e+\u003c/sup\u003eCXCR3\u003csup\u003e+\u003c/sup\u003e); P3 (CD21\u003csup\u003e-\u003c/sup\u003eCXCR3\u003csup\u003e+\u003c/sup\u003e); and P4 (CD21\u003csup\u003e-\u003c/sup\u003eCXCR3\u003csup\u003e-\u003c/sup\u003e) (Figure 4a). The proportion of P2 (CD21\u003csup\u003e+\u003c/sup\u003eCXCR3\u003csup\u003e+\u003c/sup\u003e) cells among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e was significantly higher in INF PBMCs and tissues compared to VAC (Figure 4b). In contrast, VAC individuals had a higher proportion of CXCR3\u003csup\u003e-\u003c/sup\u003e populations, P4 in blood and tissues and P1 in tissues, highlighting differences in CXCR3 expression between antigen-specific B cells in INF and VAC subjects.\u003c/p\u003e\n\u003cp\u003eWe then evaluated the expression of additional markers including transcription factors in each population in tonsil and blood. We found that atMBC markers including T-bet, CD95, CD11c, FCRL4, and FCRL3/5, were enriched primarily in P3 and P4, suggesting that these CD21\u003csup\u003e-\u003c/sup\u003e populations contained the bulk of atMBCs (Figure 4c-d). Back-gating revealed that P3 and P4 contained mixtures of atMBC and acMBC (Extended Data Figure 3a). Furthermore, the frequency of P3 among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e correlated with neutralizing titers to WA.1 (Figure 4e), consistent with correlations we saw between acMBC and neutralization titers (Extended Data Figure 1h). In contrast, most cells in P2 did not express atMBC markers but expressed an intermediate-low level of T-bet, suggestive of previous exposure to IFN-g\u0026nbsp;that may have driven and maintained expression of CXCR3 \u003csup\u003e54, 55\u003c/sup\u003e. Moreover, P1 and P2, the CD21\u003csup\u003e+\u003c/sup\u003e populations, expressed high levels of CXCR5, supporting a follicular origin, and were comprised primarily of cMBCs (Figure 4c-d and Extended Data Fig 3a).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCD69 is a marker of tissue resident memory B cells (B\u003csub\u003eRM\u003c/sub\u003e) that are poised to provide rapid local protection in response to infection or immune challenges \u003csup\u003e56\u003c/sup\u003e. In tonsil and adenoid B\u003csub\u003eSM\u003c/sub\u003e, we found that P2 had the highest frequency of CD69\u003csup\u003e+\u003c/sup\u003e cells, suggesting that this population is enriched for B\u003csub\u003eRM\u003c/sub\u003e (Figure 4d, Extended Data Figure 3b). Although CD69\u003csup\u003e+\u003c/sup\u003e S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e were noted post-infection and post-vaccination, the percentage of CD69\u003csup\u003e+\u003c/sup\u003e cells among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e was significantly higher in INF tonsils compared to VAC (Figure 4f). Thus, CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e SARS-CoV-2-specific B\u003csub\u003eSM\u003c/sub\u003e with B\u003csub\u003eRM\u003c/sub\u003e features were enriched in upper respiratory tract lymphoid tissue after infection compared to vaccination.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTissue CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e have distinct transcriptomic features\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further characterize these populations, we sorted S1\u003csup\u003e+\u003c/sup\u003e and S1\u003csup\u003e-\u003c/sup\u003e B cells from tonsils and/or adenoids of a subset of 8 INF, 4 VAC, and 2 CON subjects, including the two VAC subjects in whom S1\u003csup\u003e+\u003c/sup\u003e GCBs were identified by flow cytometry, and performed Cellular Indexing of Transcriptomes and Epitopes by Sequencing (CITE-seq) using a panel of 20 antibodies including anti-CXCR3 and CD21. In total, 9370 S1\u003csup\u003e+\u003c/sup\u003e and 172919 S1\u003csup\u003e-\u003c/sup\u003e B cells were captured and assessed for surface protein expression, gene expression and BCR sequencing at the single cell level. Unsupervised clustering based on surface protein expression revealed 6 clusters representing naïve/USM, B\u003csub\u003eSM\u003c/sub\u003e, GCB, and plasma cells/plasmablasts (PC/PB) (Figure 5a-b). S1\u003csup\u003e+\u003c/sup\u003e B cells were predominantly in a B\u003csub\u003eSM\u003c/sub\u003e cluster (cluster 1) in both INF and VAC tissues (Figure 5c-d). However, a portion of S1\u003csup\u003e+\u003c/sup\u003e B cells in both INF and VAC subjects were in cluster 2, which had a gene expression signature consistent with GCBs (Figure 5e-f), providing further support that spike-specific GCBs can be found in upper respiratory lymphoid tissues post-vaccination.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eComparison of surface protein to mRNA levels revealed that most CXCR3\u003csup\u003e+\u003c/sup\u003e B cells had little or no detectable\u003cem\u003e\u0026nbsp;CXCR3\u003c/em\u003e transcripts (Extended Data Figure 4a). We, therefore, defined B\u003csub\u003eSM\u003c/sub\u003e populations using CXCR3 and CD21 surface markers (P1-P4) (Figure 5g) to compare their gene expression. As expected, we saw a higher proportion of P2 among S1\u003csup\u003e+\u003c/sup\u003e B cells post-infection and higher proportions of P1 and P4 (both CXCR3\u003csup\u003e-\u003c/sup\u003e) post-vaccination (Figure 5h, Extended Data Figure 4b). Trajectory analyses using Slingshot suggested distinct patterns of S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e development in INF and VAC conditions (Figure 5i). In INF tissues, two branches originating from naive B cells both landed at the P2 state, while in VAC tissues, two diverging trajectories from naïve B cells emerged, including one leading to the P4 population\u0026nbsp;(Figure 5i).\u003c/p\u003e\n\u003cp\u003eBy PCA, the CD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e populations (P1 and P2) segregated distinctly from the CD21\u003csup\u003e-\u0026nbsp;\u003c/sup\u003epopulations (P3 and P4) on PC1 (Extended Data Fig 4c). Both CD21\u003csup\u003e-\u003c/sup\u003e populations (P3 and P4) had high expression of atMBC signature genes including \u003cem\u003eTBX21, FCRL4, NKG7\u003c/em\u003e, \u003cem\u003eITGAX, and IL2RB\u003c/em\u003e \u003csup\u003e47, 57\u003c/sup\u003e as well as \u003cem\u003eHCK\u0026nbsp;\u003c/em\u003eand \u003cem\u003eFGR,\u0026nbsp;\u003c/em\u003etwo SRC family kinases linked to integrin signaling (Extended Data Figure 4d) \u003csup\u003e58\u003c/sup\u003e. P3 and P4 also had high expression of gene sets related to atMBCs, as well as antigen presentation and processing and type I and type II interferon responses, which have been previously reported to be characteristic of atMBC (Figure 5j) \u003csup\u003e39, 59, 60\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn contrast, both CD21\u003csup\u003e+\u003c/sup\u003e populations (P1 and P2) in tissues exhibited features of cMBCs including expression of \u003cem\u003eCR2\u0026nbsp;\u003c/em\u003eand \u003cem\u003eFCER2\u0026nbsp;\u003c/em\u003eand expressed high levels of the tissue-residence marker \u003cem\u003eCD69\u003c/em\u003e (Figure 5j, Extended Data Figure 4d)\u003csup\u003e59\u003c/sup\u003e. We then evaluated differences between the CXCR3\u003csup\u003e+\u003c/sup\u003e and CXCR3\u003csup\u003e-\u003c/sup\u003e cMBC (P2 and P1, respectively), which constituted the majority of the S1\u003csup\u003e+\u003c/sup\u003e cells. Differentially expressed genes (DEG) in P1 versus P2 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e revealed that cells in P2 expressed higher levels of \u003cem\u003eTCF7,\u003c/em\u003e a transcription factor associated with stem-like central memory fate and found in cMBCs \u003csup\u003e36, 39\u003c/sup\u003e (Figure 5k, Supplemental Table 9). Upon secondary challenge, B\u003csub\u003eSM\u003c/sub\u003e in the lymphoid tissue can differentiate to PB/PC or GCBs based on their transcriptomic and epigenetic profile \u003csup\u003e61\u003c/sup\u003e. DEG in P1 versus P2 S1⁺\u0026nbsp;B\u003csub\u003eSM\u003c/sub\u003e suggested distinct cell fates upon rechallenge. P2 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e expressed higher \u003cem\u003eFOS\u003c/em\u003e, an AP-1 transcription factor family member expressed highly in pre-plasmablasts\u0026nbsp;\u003csup\u003e62\u003c/sup\u003e, and expressed lower levels of \u003cem\u003eBACH2\u003c/em\u003e, a transcription factor that represses \u003cem\u003eBLIMP1\u0026nbsp;\u003c/em\u003eexpression and plasma cell differentiation and is found in memory B cells that preferentially differentiate to GCBs\u0026nbsp;\u003csup\u003e61\u003c/sup\u003e (Figure 5k). S1\u003csup\u003e+\u003c/sup\u003e P2 B\u003csub\u003eSM\u003c/sub\u003e also expressed lower \u003cem\u003eIL21R\u003c/em\u003e, which is critical for GCB differentiation, maintenance, and proliferation\u0026nbsp;\u003csup\u003e63, 64, 65\u003c/sup\u003e and the GCB marker \u003cem\u003eMEF2B\u003c/em\u003e (Figure 5k)\u0026nbsp;\u003csup\u003e66\u003c/sup\u003e. Moreover, pathway analyses revealed enrichment of genes involved PC differentiation, translation, interferon response, antigen presentation and oxidative metabolism pathways, with lower cell cycle associated genes in P2 compared to P1 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e (Figure 5l). Together, these differences suggest that P2 B\u003csub\u003eSM\u003c/sub\u003e are better poised to differentiate to plasma cells rather than GCBs compared to P1.\u003c/p\u003e\n\u003cp\u003eWe further compared the epigenetic profiles of sorted bulk P1, P2, P3 and P4 B\u003csub\u003eSM\u003c/sub\u003e from three tonsils using ATAC-seq. Again, CD21\u003csup\u003e+\u003c/sup\u003e (P1 and P2) and CD21\u003csup\u003e-\u0026nbsp;\u003c/sup\u003e(P3 and P4) populations clustered distinctly on PCA (Extended Data Figure 4e), with P3 and P4 enriched in motifs related to AP-1 transcription factors (FOS, BATF, JUN) that are associated with atMBCs and PC/PB differentiation\u003csup\u003e66\u003c/sup\u003e (Extended Data Figure 4f). Direct comparison of the cMBC groups, P1 and P2, revealed that P1 had greater overall open chromatin regions compared to P2 with a few notable exceptions including the \u003cem\u003eTCF7\u003c/em\u003e locus, which was more accessible in P2, consistent with the higher\u003cem\u003e\u0026nbsp;TCF7\u003c/em\u003e expression noted in the transcriptome (Extended Data Figure 4g, Supplemental Table 10). The enriched signaling pathways associated with differentially accessible regions in P1 versus P2 mirrored those identified in the transcriptomic analysis, including those downstream of IFN-g, BCR, and antigen receptors (Figure 5l, Extended Data Figure 4h).\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eIRF4\u0026nbsp;\u003c/em\u003eand \u003cem\u003ePRDM1\u003c/em\u003e encode transcription factors essential for PC/PB differentiation \u003csup\u003e67\u003c/sup\u003e; chromatin accessibility at these loci in P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e were similar and distinct from plasma cells and P3/P4 (Extended Data Figure 4i). However, differentially accessible regions in P2 compared with P1 had a greater proportion of AP-1-IRF composite (AICE) and interferon sequence response element (ISRE) binding motifs, which are IRF4-binding complex motifs reported to favor plasma cell differentiation. In contrast, the proportion of the Ets-IRF composite (EICE) motif, which is also a IRF4 binding complex motif but not involved in plasma cell differentiation, was comparable between P2 and P1 enriched peaks \u003csup\u003e67, 68\u003c/sup\u003e. This enrichment pattern suggests the P2 B\u003csub\u003eSM\u003c/sub\u003e population is primed for IRF4-mediated transcriptional changes driving plasma cell differentiation. Our findings suggest that among the cMBCs, P2 B\u003csub\u003eSM\u003c/sub\u003e are transcriptionally and epigenetically predisposed to antibody-secreting cell differentiation upon antigenic re-exposure compared to P1 B\u003csub\u003eSM\u003c/sub\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInfection induces greater clonal expansion among S1\u003csup\u003e+\u003c/sup\u003e B cells in pharyngeal tissues\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further understand differences between VAC and INF responses, we assessed isotype, somatic hypermutation, clonal expansion, and clonal overlap by BCR sequencing of the sorted S1\u003csup\u003e+\u003c/sup\u003e and S1\u003csup\u003e-\u003c/sup\u003e B cells characterized by CITE-seq. Single-cell BCR sequencing confirmed that the majority of S1\u003csup\u003e+\u003c/sup\u003e B cells were IgG1-class switched cells in both VAC and INF tissues, with a greater proportion of IgA1 expressing cells across all B\u003csub\u003eSM\u003c/sub\u003e subsets (P1-P4) post-infection (Extended Data Figure 5a).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA greater proportion of S1\u003csup\u003e+\u003c/sup\u003e B cells were part of expanded clones in INF compared to VAC tissues (Figure 6a). S1\u003csup\u003e+\u003c/sup\u003e B cells from INF tissues displayed lower clonal diversity compared to S1\u003csup\u003e-\u003c/sup\u003e B cells and also to S1\u003csup\u003e+\u003c/sup\u003e B cells in VAC tissues, suggesting greater antigen-driven expansion in tissues post-infection compared to post-vaccination (Figure 6b). Furthermore, in INF tissues, P2 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e had a trend towards a higher level of expansion than P1 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e; this expansion was not seen in P2 S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e post-vaccination (Figure 6c).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn tonsil, atMBCs have been noted to have slightly lower levels of SHM\u003csup\u003e59\u003c/sup\u003e; we also found that SHM frequency was lower among bulk S1\u003csup\u003e-\u003c/sup\u003e P3 B\u003csub\u003eSM\u003c/sub\u003e in the tissues. However, SHM frequency was comparable across all four P1-P4 subsets of S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in both INF and VAC (Figure 6d), revealing distinctions between SARS-CoV-2-specific cells compared to the bulk. Indeed, SHM frequency and heavy chain CDR3 amino acid length and distribution among S1\u003csup\u003e+\u003c/sup\u003e IgG\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e and GCBs were comparable between INF versus VAC tissues (Extended Data Figure 5b-d).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe then assessed BCR clonal overlap to evaluate B\u003csub\u003eSM\u003c/sub\u003e development. Using STARTRAC (single T cell analysis by RNA sequencing and TCR tracking) pairwise transition index (pTrans) \u003csup\u003e69, 70\u003c/sup\u003e, which assesses developmental links in lymphocyte populations using clonal sharing, we found that antigen-specific S1\u003csup\u003e+\u003c/sup\u003e P2 B\u003csub\u003eSM\u003c/sub\u003e exhibited high pTrans scores with S1\u003csup\u003e+\u003c/sup\u003e P1 B\u003csub\u003eSM\u003c/sub\u003e and S1\u003csup\u003e+\u003c/sup\u003e GCB (Figure 6e). Although S1\u003csup\u003e+\u0026nbsp;\u003c/sup\u003eP3 B\u003csub\u003eSM\u003c/sub\u003e were most strongly connected to GCBs, their connectivity was less strong than S1\u003csup\u003e+\u003c/sup\u003e P2 B\u003csub\u003eSM\u003c/sub\u003e (Figure 6f). Bulk S1\u003csup\u003e-\u003c/sup\u003e P2 B\u003csub\u003eSM\u003c/sub\u003e also had a strong association to S1\u003csup\u003e-\u003c/sup\u003e GCBs, whereas S1\u003csup\u003e-\u003c/sup\u003e P3 B\u003csub\u003eSM\u003c/sub\u003e were most strongly associated with S1\u003csup\u003e-\u003c/sup\u003e P4 B\u003csub\u003eSM\u003c/sub\u003e and PC/PB, further supporting distinct connections and development among these cell populations (Extended Data Figure 5e-f).\u003c/p\u003e\n\u003cp\u003eWe then looked further at BCR sequences from the tissues of one INF subject from whom we sorted a large number of S1\u003csup\u003e+\u003c/sup\u003e B cells (CNMC 70). By Jaccard index scores, we observed clonal overlap among S1\u003csup\u003e+\u003c/sup\u003e P1, P2, P3, and P4 B\u003csub\u003eSM\u003c/sub\u003e populations, with particularly strong overlap between P1 and P2. S1\u003csup\u003e+\u003c/sup\u003e P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e also had strong clonal sharing with GCBs, whereas S1\u003csup\u003e+\u003c/sup\u003e P3 and P4 B\u003csub\u003eSM\u003c/sub\u003e had GC overlap but to a lower extent than S1\u003csup\u003e+\u003c/sup\u003e P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e (Figure 6g). Evaluation of clonal trees showed S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e bearing various phenotypes (P1-P4, GCB) and isotypes emerging from a common ancestor in both INF (Figure 6h) and VAC (Figure 6i) tissues, implying broad differentiation potential or plasticity in B cell phenotypes during clonal expansion. Several clonotypes were also shared among adenoid and tonsil S1\u003csup\u003e+\u003c/sup\u003e B cells, particularly in P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e and GCB populations, supporting an immunologic connection between these two oropharyngeal lymphoid tissues, as we previously reported \u003csup\u003e23\u003c/sup\u003e (Figure 6h-i, Extended Data Figure 5g). Thus, although similar levels of SHM were noted post-infection and vaccination, infection induced greater antigen-specific B cell expansion in the tissues, particularly among P2 B\u003csub\u003eSM\u003c/sub\u003e, which likely have a GC origin.\u003cbr\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e are more responsive to BCR stimulation and poised to differentiate to plasmablasts\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo determine whether these differences reflect functional differences among P1-P4 B\u003csub\u003eSM\u003c/sub\u003e populations, we evaluated responsiveness to BCR stimulation, proliferation capacity, and ability to differentiate to antibody-secreting cells. Tonsil cells from a separate group of pediatric subjects who were either post-infection or had hybrid immunity to SARS-CoV-2 (both infected and vaccinated, Supplemental Table 2) were stimulated with soluble anti-human IgA, IgG, and IgM for 2 minutes and phosphorylation of the tyrosine kinase Syk and phospholipase Cg2 (PLC-g2), two downstream signaling proteins rapidly phosphorylated upon BCR stimulation, were measured \u003csup\u003e71\u003c/sup\u003e. As noted previously, the CD21\u003csup\u003elo\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e populations (P3 and P4) were poorly responsive to BCR stimulation with soluble immunoglobulin \u003csup\u003e72, 73\u003c/sup\u003e (Figure 7a). In contrast, the P2 B\u003csub\u003eSM\u003c/sub\u003e population had the strongest induction of p-Syk and/or p-PLCg2 among total B\u003csub\u003eSM\u003c/sub\u003e and S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e (Figure 7a-b, Extended Data Figure 6a-b). This phenotype was not limited to B\u003csub\u003eSM\u003c/sub\u003e, as the CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e (P2) subpopulation also had the strongest response to BCR stimulation among GCBs (Extended Data Figure 6c).\u003c/p\u003e\n\u003cp\u003eIn general, as B cells transition to PC/PB, membrane-associated IgG declines due to altered splicing, generating secreted IgG \u003csup\u003e36, 74\u003c/sup\u003e. We analyzed surface IgG and IgA expression (geometric mean fluorescence intensity or gMFI of anti-IgG and anti-IgA antibody staining) among B\u003csub\u003eSM\u003c/sub\u003e and PC/PB in our flow cytometry data. As expected, CD27\u003csup\u003ehi\u003c/sup\u003eCD38\u003csup\u003ehi\u003c/sup\u003e PC expressed marked reductions in surface IgG compared to B\u003csub\u003eSM\u003c/sub\u003e in both the peripheral blood and tissues (Figure 7c). Surface IgA was also reduced on plasma cells in peripheral blood, but to a lesser extent (Figure 7c) and was not reduced on tissue PC compared to GCBs and tissue B\u003csub\u003eSM\u003c/sub\u003e (Figure 7d). These observations are consistent with findings of functional surface IgA on gut and bone marrow PC in humans \u003csup\u003e75\u003c/sup\u003e and suggest that surface IgA may also play a functional role on PC in the oropharyngeal lymphoid tissue.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe then compared surface IgG expression on the two major S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e populations, P1 and P2, as a surrogate for PC differentiation \u003csup\u003e36\u003c/sup\u003e. \u0026nbsp;In INF tissues and blood, S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e IgG\u003csup\u003e+\u003c/sup\u003e P2 B\u003csub\u003eSM\u003c/sub\u003e had lower surface immunoglobulin expression than their P1 counterparts, suggesting the SARS-CoV-2-specific P2 B\u003csub\u003eSM\u003c/sub\u003e post-infection are more prone to differentiate to PC (Figure 7e). Significant differences were not noted for P1 and P2 in VAC tissues, although a trend was seen in the tonsils (Figure 7e). Moreover, in the tonsil, S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e IgG\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e from INF subjects had lower surface immunoglobulin expression compared to VAC subjects (Extended Data Figure 6d); a similar trend was noted in the adenoid, suggestive of a propensity of tissue B\u003csub\u003eSM\u003c/sub\u003e induced by infection to differentiate to PC/PB.\u003c/p\u003e\n\u003cp\u003eWe then directly assessed the ability of P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e from tissues to differentiate to plasmablasts and proliferate \u003cem\u003ein vitro\u003c/em\u003e. We sorted bulk P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e cells from tonsils of three subjects with hybrid immunity and labelled them with carboxyfluorescein succinimidyl ester (CFSE). We cultured the labelled P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e populations with B cell-depleted PBMCs of an unrelated donor, R848 (TLR7/8 agonist), and IL-2 and we assessed proliferation index and percentage of plasmablasts in culture using flow cytometry after 4 days. Due to low cell numbers and poor viability after sorting, we were unable to reproducibly culture and assess P3 and P4 B\u003csub\u003eSM\u003c/sub\u003e. Compared to P1, a greater proportion of P2 B\u003csub\u003eSM\u003c/sub\u003e differentiated to CD38\u003csup\u003e+\u003c/sup\u003eCD20\u003csup\u003elo\u003c/sup\u003e plasmablasts (Figure 7f). P2 B\u003csub\u003eSM\u003c/sub\u003e also exhibited a higher proliferation index in culture (Figure 7g). Thus, P2 B\u003csub\u003eSM\u003c/sub\u003e, which were enriched among INF samples, were better primed for plasma cell differentiation compared to P1 B\u003csub\u003eSM\u003c/sub\u003e, which may result in greater protective antibody production in the upper respiratory tract tissues upon re-exposure to SARS-CoV-2 in infected individuals compared to vaccinated individuals.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInfection, but not vaccination, is associated with persistent immunologic activation in pharyngeal tissues\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe previously found that SARS-CoV-2 infection early in the pandemic induced persistent expansion of CD4\u003csup\u003e+\u003c/sup\u003e and CD8\u003csup\u003e+\u003c/sup\u003e T cell populations associated with antiviral, GC and IFN-g\u0026nbsp;responses in the tonsils and adenoids compared to uninfected controls, with the most significant effects in adenoid \u003csup\u003e23\u003c/sup\u003e. However, whether vaccination also induces long-lasting immunologic changes in the upper respiratory tract lymphoid tissues is unclear. Consistent with our previous findings, a higher proportion of CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e-\u003c/sup\u003eCD57\u003csup\u003e+\u003c/sup\u003ePD-1\u003csup\u003ehi\u0026nbsp;\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e T cells and CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e-\u003c/sup\u003e pre-Tfh cells were noted in INF compared to CON adenoids (Figure 8a; Extended Data Figure 7a). In contrast, VAC adenoids did not exhibit changes in these T cell populations compared to CON (Figure 8a).\u003c/p\u003e\n\u003cp\u003eCOVID-19 induces persistent activation and epigenetic remodeling of innate immune cells including peripheral blood monocytes and dendritic cells well into convalescence\u003csup\u003e76, 77\u003c/sup\u003e. The frequency of CD14\u003csup\u003e+\u003c/sup\u003e monocytes/macrophages, conventional dendritic cells (cDC), and plasmacytoid dendritic cells (pDC) did not differ among the VAC, INF, and CON tissues (Extended Data Figure 7b). However, these innate immune populations expressed higher HLA-DR in INF adenoids compared to VAC (and trended toward higher expression compared to CON), revealing persistent myeloid cell activation in adenoids post-infection but not post-vaccination (Figure 8b). Unlike the tissues, peripheral blood CD14\u003csup\u003e+\u003c/sup\u003e monocytes/macrophages and cDC in INF subjects did not express higher HLA-DR post-infection, highlighting localized differences in the impact of infection in upper respiratory tract tissues compared to blood (Extended Data Figure 7c). However, the antiviral type I interferon-producing and CXCR3-expressing pDC population was slightly more activated in both INF blood and tissues (Figure 8b, Extended Data Figure 7c) \u003csup\u003e78\u003c/sup\u003e. \u0026nbsp;Thus, convalescence was associated with local activation of innate cells in tissues, which may reflect differences in the primary lymphoid site of antigen-presentation and GC response and/or antigen persistence in infection versus vaccination.\u003c/p\u003e\n\u003cp\u003eWe then assessed correlations between populations of interest. \u0026nbsp;The proportion of P2 cells among S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e was positively associated with the proportion of CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e-\u0026nbsp;\u003c/sup\u003epre-Tfh cells, an IFN-g\u0026nbsp;producing population that is involved in GC reactions and important for CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eRM\u003c/sub\u003e development in mice \u003csup\u003e23\u003c/sup\u003e\u003csup\u003e,\u003c/sup\u003e \u003csup\u003e54\u003c/sup\u003e (Figure 8c). The frequency of the CXCR3\u003csup\u003e+\u003c/sup\u003eCCR6\u003csup\u003e-\u0026nbsp;\u003c/sup\u003eCD57\u003csup\u003e+\u003c/sup\u003ePD-1\u003csup\u003ehi\u0026nbsp;\u003c/sup\u003eTfh cell population also significantly correlated with the percentage of P2 S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u0026nbsp;\u003c/sub\u003e(Figure 8c): these CXCR3\u003csup\u003e+\u003c/sup\u003e Tfh populations also correlated with CXCR3\u003csup\u003e+\u003c/sup\u003e total and/or S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e (Extended Data Figure 7d). Moreover, the proportion of P2 B\u003csub\u003eSM\u003c/sub\u003e and CXCR3\u003csup\u003e+\u003c/sup\u003e pre-Tfh cells were also positively associated with the activation levels of monocytes/macrophages, cDC, and pDC, pointing to the potential relevance of innate cell activation in development or accumulation of CXCR3\u003csup\u003e+\u003c/sup\u003e lymphocytes (Figure 8c). Thus, infection is associated with prolonged activation of innate or expansion of adaptive immune populations in tonsils/adenoids, which may contribute to the generation of CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCXCR3\u003csup\u003e+\u003c/sup\u003e lymphocytes and \u003cem\u003eCXCL9\u003c/em\u003e-expressing myeloid cells are in close proximity in the interfollicular region\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUsing multicolor immunofluorescence imaging, we compared the location of CXCR3\u003csup\u003e+\u003c/sup\u003e CD20\u003csup\u003e+\u003c/sup\u003e B and CD4\u003csup\u003e+\u003c/sup\u003e T cells and to their CXCR3\u003csup\u003e-\u0026nbsp;\u003c/sup\u003ecounterparts in 4 tissue samples, derived from tonsils and adenoids of one INF and one VAC subject (Figure 8d, Supplemental Table 2). \u0026nbsp;CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003e CD20\u003csup\u003e+\u003c/sup\u003e B cells and CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells were primarily located in the extrafollicular region (Figure 8d, Extended Data Figure 7e-f). Compared to their CXCR3\u003csup\u003e-\u003c/sup\u003e counterparts, a smaller portion of CXCR3\u003csup\u003e+\u003c/sup\u003e BCL6\u003csup\u003e-\u003c/sup\u003e B cells were in follicles, which includes the GC and surrounding mantle (Extended Data Figure 7e). A similar pattern of distribution was seen among CXCR3\u003csup\u003e+\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells (Extended Data Figure 7f). CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cells and CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003e CD20\u003csup\u003e+\u003c/sup\u003e B cells appeared co-localized or aggregated near one another compared to CXCR3\u003csup\u003e-\u003c/sup\u003e BCL6\u003csup\u003e-\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T and CXCR3\u003csup\u003e-\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003e CD20\u003csup\u003e+\u003c/sup\u003e B cells which were more diffusely distributed in the extrafollicular area (Figure 8d and e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe then assessed spatial proximity of CXCR3\u003csup\u003e+\u003c/sup\u003e and CXCR3\u003csup\u003e-\u003c/sup\u003e B and CD4\u003csup\u003e+\u003c/sup\u003e T cells using spatial patterning analysis of cellular ensembles (SPACE), which assesses spatial relationships among multiple cell populations \u003csup\u003e79\u003c/sup\u003e. Peak areas of abundance of BCL6\u003csup\u003e+\u003c/sup\u003e and BCL6\u003csup\u003e–\u003c/sup\u003e populations were distinct, clearly delineating GCs (Figure 8f). Within BCL6\u003csup\u003e-\u003c/sup\u003e regions, the CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003eCD20\u003csup\u003e+\u003c/sup\u003e B cell and CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003eCD4\u003csup\u003e+\u003c/sup\u003e T cell peaks were in close proximity to each other; similarly, in BCL6\u003csup\u003e+\u003c/sup\u003e\u0026nbsp; regions, which represent GCs, the CXCR3\u003csup\u003e+\u003c/sup\u003e BCL6\u003csup\u003e+\u003c/sup\u003e CD20\u003csup\u003e+\u003c/sup\u003e B cell and CXCR3\u003csup\u003e+\u003c/sup\u003e BCL6\u003csup\u003e+\u003c/sup\u003e CD4\u003csup\u003e+\u003c/sup\u003e T cell were closely aligned, suggesting that CXCR3\u003csup\u003e+\u003c/sup\u003e B and T cells were adjacent both within and outside GCs (Figure 8f). Moreover, compared to their CXCR3\u003csup\u003e-\u003c/sup\u003e counterparts, CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e-\u003c/sup\u003e B and T cells were closer to BCL6\u003csup\u003e+\u003c/sup\u003e populations, and in particular, closer to CXCR3\u003csup\u003e+\u003c/sup\u003eBCL6\u003csup\u003e+\u003c/sup\u003e B and T cells, further implying proximity and possible interactions among CXCR3\u003csup\u003e+\u003c/sup\u003e Tfh, GCB, and B\u003csub\u003eSM\u003c/sub\u003e populations in the GC and T-B border regions (Figure 8f).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe performed spatial transcriptomics analysis with the Xenium 5000 human gene panel of the same 4 samples from an adjacent section of tissue. With unsupervised clustering and annotation, we identified 25 cell types, which localized to particular regions (Figure 8g, Extended Data Figure 7g). CXCL9, -10, and -11 are ligands of CXCR3 and enable migration and retention of CXCR3-expressing lymphocytes in tissue \u003csup\u003e80\u003c/sup\u003e. In tonsils and adenoids, we found that macrophages/monocytes and dendritic cells were the primary expressors of \u003cem\u003eCXCL9\u003c/em\u003e, \u003cem\u003eCXCL10\u003c/em\u003e, and \u003cem\u003eCXCL11,\u0026nbsp;\u003c/em\u003ewith \u003cem\u003eCXCL9\u003c/em\u003e being the most abundantly expressed (Extended Data Figure 7h). Moreover, these innate immune cell populations were located in the interfollicular area, where we detected expression of \u003cem\u003eCXCL9\u003c/em\u003e and where we found CXCR3\u003csup\u003e+\u003c/sup\u003e B and T cells by immunofluorescence (Figure 8d, g-k). Thus, spatial analyses suggest that the CXCR3 and CXCL9/10/11 axis facilitates interaction of innate and adaptive immune cells in lymphoid tissue, which may shape both the position and characteristics of memory B cells during viral infection.\u003c/p\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eMaintenance of virus-specific adaptive immune memory in the respiratory mucosa is important for immune protection against respiratory viral infections including COVID-19.By assessing tonsils and adenoids of children undergoing tonsillectomy/adenoidectomy, we found SARS-CoV-2-specific B cells, including B\u003csub\u003eSM\u003c/sub\u003e, B\u003csub\u003eRM\u003c/sub\u003e and a few GCBs, in these upper respiratory tract lymphoid tissues following vaccination. Distribution of B\u003csub\u003eSM\u003c/sub\u003e across tissues, including mucosal tissues, after intramuscular vaccination has been noted \u003csup\u003e51, 81, 82\u003c/sup\u003e. However, whether and how these cells differ from memory cells found post-infection has not been clear. Using high-dimensional analyses, we show that immune memory is maintained in mucosal lymphoid tissues after intramuscular COVID-19 immunization, but reveal distinct features compared to those generated post-infection. In particular, we identified a subset of cMBC defined by CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e, that was enriched in infected individuals, with increased propensity for plasma cell differentiation and distinct localization within mucosal lymphoid tissues. This population may be responsible for some of the qualitative and clinical differences in humoral mucosal immunity observed with infection and vaccination.\u003c/p\u003e\n\u003cp\u003ePrior studies have found SARS-CoV-2-specfic GC and Tfh cells in the draining axillary lymph nodes following mRNA vaccination \u003csup\u003e17, 18, 19, 20, 83\u003c/sup\u003e, but no SARS-CoV-2-specific GC B cells were noted in a small number of contralateral lymph nodes \u003csup\u003e17\u003c/sup\u003e. However, spike-reactive GCBs have been noted in lung-associated lymph nodes of a few (2/18) vaccinated (and uninfected) organ donors \u003csup\u003e51\u003c/sup\u003e and tonsils of a few vaccinated adults, although whether these latter subjects were truly SARS-CoV-2 infection-naïve is unclear since some had B cells reactive to NC\u003csup\u003e60\u003c/sup\u003e. To address this concern, we measured both anti-NC antibodies and anti-ORF8 antibodies, the latter of which are more sensitive than antibodies to NC\u0026nbsp;\u003csup\u003e24\u003c/sup\u003e. We too noted the presence of spike-specific GCBs in some vaccinated participants in distal mucosal lymphoid sites, although fewer than among infected subjects. Whether these result directly in response to immunization, from cross-reactivity to common cold coronaviruses, or exposure to SARS-CoV-2 that did not result in a productive infection, perhaps due to the presence of high affinity antibodies generated by vaccination \u003csup\u003e84, 85\u003c/sup\u003e, remains an interesting question.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAlthough COVID-19 mRNA vaccines have high vaccine efficacy and prevent hospitalization and death from COVID-19, epidemiologic studies suggest that infection provides longer-lasting immunity than vaccination\u003csup\u003e6, 7, 8\u003c/sup\u003e. This may result from differences in the route of viral antigen exposure to the host’s immune system (intramuscular vs. inhaled), breadth of antigenic exposure (spike only vs. all viral antigens), nature of antigens presented (lipid nanoparticles vs. virions) or duration of antigen exposure. These factors, in turn, affect host immune responses, and, indeed, we found that infection and vaccination drive the generation of phenotypically different SARS-CoV-2-specific B cells, suggesting unique B\u003csub\u003eSM\u003c/sub\u003e developmental pathways depending on the type of exposure.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe clonal expansion and persistence of CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e post-infection suggest that IFN-g, induced by viral infection and likely produced by activated and expanded innate and Tfh cells in the tissue, may be a driving force for these developmental differences in B cell memory. The CXCR3\u003csup\u003e+\u003c/sup\u003e P2 B\u003csub\u003eSM\u003c/sub\u003e population expressed low levels of Tbet, which may reflect prior IFN-g\u0026nbsp;exposure during infection, and were distinct from atMBCs which expressed high levels of Tbet. Our results suggest that CXCR3 expression affords two key characteristics incMBCs that may contribute to enhanced mucosal protection post-infection. First, CXCR3 promotes homing and retention of B\u003csub\u003eSM\u003c/sub\u003e cells, particularly IgA\u003csup\u003e+\u003c/sup\u003e cells, to mucosal lymphoid tissues, where CXCR3\u003csup\u003e+\u0026nbsp;\u003c/sup\u003ecells bear B\u003csub\u003eRM\u003c/sub\u003e markers, as supported by murine studies \u003csup\u003e52, 54\u003c/sup\u003e. Second, we found that among cMBCs, CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e have a predilection for plasma cell differentiation upon secondary challenge compared to CXCR3\u003csup\u003e-\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM,\u0026nbsp;\u003c/sub\u003eenabling a more robust antibody response upon secondary challenge. Although CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e were enriched in the tissues, the difference in CXCR3 expression between INF and VAC was also clear in PBMCs suggesting CXCR3 may be a useful marker for evaluating potential for trafficking even in blood; this is supported by our recent findings in blood of vaccinated adults with hybrid immunity \u003csup\u003e86\u003c/sup\u003e, indicating the applicability of our findings across age groups. Moreover, although our study focused on mRNA COVID-19 vaccines, another recent study found the Ad26.COV2.S, elicited more CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e without atMBC features compared to mRNA vaccines \u003csup\u003e5\u003c/sup\u003e, suggesting similar cells may be generated by viral vaccine platforms. Nonetheless, the mRNA component of mRNA-lipid nanoparticle vaccines can stimulate IFN-g\u0026nbsp;production by Tfh cells\u0026nbsp;\u003csup\u003e87\u003c/sup\u003e, potentially explaining the presence of some CXCR3\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in the blood and tissues of vaccinated subjects in our study. Of note, the frequency of the P2 population among B\u003csub\u003eSM\u003c/sub\u003e did not correlate with serum neutralizing titers, suggesting this population reflects other aspects of immunity induced by infection and that neutralizing titers cannot be used as a proxy for this population.\u003c/p\u003e\n\u003cp\u003eConversely, we found that mRNA vaccination induced a greater proportion of virus-specific CD21\u003csup\u003elo\u003c/sup\u003e and particularly CXCR3\u003csup\u003e-\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e populations than infection in both blood and tissues \u003csup\u003e29, 38, 60\u003c/sup\u003e. Although some studies suggest that CD21\u003csup\u003elo\u003c/sup\u003e atMBCs have lower degrees of somatic hypermutation than cMBC and emerged from extrafollicular reactions particularly in the context of chronic viral or autoimmune conditions \u003csup\u003e39, 45\u003c/sup\u003e, we found that\u0026nbsp;SARS-CoV-2-specific CD21\u003csup\u003elo\u003c/sup\u003e populations in tonsils/adenoids had comparable levels of SHM as CD21\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e, highlighting the importance of assessing antigen-specific versus bulk cell populations and evaluating cells in patients without conditions characterized by strong extrafollicular reactions. The frequency of CD21\u003csup\u003e-/lo\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e in the blood declines with time after influenza or COVID-19 vaccination, suggesting these cells may be shorter lived \u003csup\u003e34, 35\u003c/sup\u003e. It is possible that the enrichment of SARS-CoV-2-specific CD21\u003csup\u003e-\u0026nbsp;\u003c/sup\u003ecellspost-vaccination accounts for the steeper decline in immunity months after mRNA vaccination compared to infection. Nonetheless, FCRL5\u003csup\u003e+\u003c/sup\u003eTbet\u003csup\u003e+\u003c/sup\u003e memory B cells with atMBC features have been found 1 year after influenza vaccination and contributed significantly to serum antibody levels upon secondary challenge; similarly FCRL4\u003csup\u003elo\u003c/sup\u003e atMBCs were present at least 6 months post-influenza vaccination, suggesting the picture is more complex \u003csup\u003e36, 37\u003c/sup\u003e. We also note that vaccination engendered broader variant coverage, as determined by cross-reactivity to omicron, highlighting an important feature induced by mRNA vaccination. Thus, the plasticity and function of different memory populations following vaccination versus infection remain important questions.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAs humans are confronted with new pathogens, the development of vaccinations for respiratory pathogens that elicit strong mucosal immunity, which may reduce mucosal shedding and transmission, is a priority. Although intramuscular SARS-CoV-2 mRNA vaccines generate rapid and broad immunity, they elicit low neutralizing antibody levels in the respiratory tract \u003csup\u003e30, 31\u003c/sup\u003e. Trials of intranasal vaccines in humans are ongoing and have shown mixed results \u003csup\u003e88, 89\u003c/sup\u003e. Nonetheless, pre-clinical studies of inhaled vaccine boosters following intramuscular mRNA vaccines in murine models showed enhanced mucosal protection due to trafficking of antigen-experienced B\u003csub\u003eSM\u003c/sub\u003e to the respiratory tract through the CXCL9/10-CXCR3 axis and enhanced differentiation to IgA-producing plasma cells \u003csup\u003e48\u003c/sup\u003e. Our study underscores the significance of this axis in homing and shaping immunity in the upper respiratory tract of humans. In summary, our study provides evidence for the generation and maintenance of mucosal B cell memory after mRNA vaccination and further provides a framework for evaluating immunity in the upper respiratory tract and the blood following immunization.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cu\u003eLIMITATIONS\u003c/u\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlthough study participants were heterogeneous in terms of their age, tonsil condition, time since vaccination/infection, and doses of immunization, we tried to control for age as a covariate using a linear model. Moreover, the INF cohort, which was collected in 2020-2021, had similar times from their last known immunologic exposure to surgery as the VAC cohort and were exposed to early circulating strains with spike proteins similar to the strain in the first-generation mRNA vaccines, helping control for these factors. Nonetheless, we only know the time of infection for approximately half of the infected subjects. As our study is a cross-sectional study, we are unable to assess temporal changes in immune responses over time. We also note that our INF cohort only included patients with mild or asymptomatic infection; severe infection may alter or compromise the humoral immunity generated by infection \u003csup\u003e90\u003c/sup\u003e. Lastly, due to the increasing percentage of infected individuals, our sample size of vaccinated only individuals was limited, which could affect our ability to fully assess correlations between clinical characteristics and immune profiles.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u0026nbsp;\u003c/strong\u003eWe thank the patients and their families for their generous participation; Pedro Milanez-Almeida, Edward Schrom, Sebastian Wellford, Weiming Yu, Daniel Newman, Hiroshi Ichise, Julie Reilley, and members of the Schwartzberg lab at National Institutes of Health (NIH) for advice, discussions, and/or technical assistance; Division of Otolaryngology at Children’s National Hospital for helping with participant recruitment; and National Cancer Institute CCR Genomics Core for sequencing.This research was supported by the Intramural Research Program of the NIH. The contributions of the NIH author(s) were made as part of their official duties as NIH federal employees, are in compliance with agency policy requirements, and are considered Works of the United States Government. However, the findings and conclusions presented in this paper are those of the authors and do not necessarily reflect the views of the NIH or the U.S. Department of Health and Human Services.\u0026nbsp;The antibody response study was supported by FDA’s Perinatal Health Center of Excellence (PHCE) project grant #GCBER005 and #GCBER008 to S.K. \u0026nbsp;This work utilized the computational resources of the NIH Biowulf and NIH NIAID Skyline clusters.\u0026nbsp;K.H. was supported by NIAID grant R00AI159302.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions:\u0026nbsp;\u003c/strong\u003eK.M., P.M., P.L.S, and Q.X. conceived and designed the study. Q.X., K.M., performed flow cytometry experiments. L.B., G.G., S.P., J. T., and S.K. performed serologic testing and analysis. Q.X. analyzed the flow cytometry data and performed statistical analyses of these data. Q.X., K.M., P.L.S, CHI, C.L., K.H. and A.J.M. designed, performed analyzed or advised CITE-seq experiments. L.S., J.K., M. S., Q.X., and K.M. performed immunofluorescence staining, obtained images, and/or analyzed data. T.E.M., A.K., D.P.G., and Q.X. performed and/or analyzed ATACseq. Q.X., K.M., R.S., and L.S. obtained or analyzed Xenium data. K.M., P.M., and H.B. developed patient recruitment materials and/or recruited participants. Q.X., P.M., C.L., S.K., L.K., C.M.B., R.S., J.S.T., S.M., P.L.S., and K.M. provided critical scientific input and/or reagents. K.M., Q.X, and P.L.S. prepared the manuscript. All authors contributed to the final review and editing of the paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of interests\u003c/strong\u003e: The authors have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability:\u003c/strong\u003e All data necessary to understand and evaluate the conclusions of this paper are provided in the article and Supplementary Information. The CITE-seq and ATAC-seq data will be deposited to dbGAP (the study is in the process of getting registered). Source data will be provided with the final published manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability:\u003c/strong\u003e The R scripts used in this paper are available on a GitHub repository at https://github.com/kalpanamanthiram/COVID-19-Vaccination.\u003c/p\u003e"},{"header":"METHODS","content":"\u003cp\u003e\u003cstrong\u003eParticipant recruitment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Institutional Review Board (IRB) at Children\u0026rsquo;s National Hospital (IRB protocol number 00009806). Written informed consent was obtained from parent/guardians of all enrolled participants,\u0026nbsp;and assent was obtained from minor participants over 7 years of age.\u003c/p\u003e\n\u003cp\u003eWe recruited 21 children who\u0026nbsp;underwent tonsillectomy and/or adenoidectomy at Children\u0026rsquo;s National Hospital (CNH) in Washington, DC, USA and also received at least one dose of a first generation COVID-19 mRNA vaccine. The first 17 participants were recruited from December 2021 to April 2022, and the remaining were recruited from August 2022 to September 2022. Because not all tissues or blood were available from each subject, we collected a total of 21 blood samples, 19 adenoids, and 18 tonsils from these 21 participants. No statistical methods were used to predetermine sample size. All participants had negative RT-PCR testing from a nasopharyngeal swab for SARS-CoV-2 within 72 hours of surgery. Demographic information and clinical data were collected through parental questionnaires and chart review and inputted and managed in REDCap (https://project-redcap.org/), and biologic samples were acquired in the operating room by the clinical team at CNH. Participants with no evidence of prior infection (negative for anti-nucleocapsid (NC) or anti-open reading frame 8 (ORF8) serum antibodies and/or no clinical history of PCR/antigen confirmed SARS-CoV-2 infection) were classified as the VAC group (vaccinated only, 10 subjects). Samples from the remaining 11 subjects with hybrid immunity (both post-vaccinated and infection, 11 subjects) and additional post-infection subjects without vaccination who were recruited in 2022/2023 (12 subjects) were used in some assays.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe also used samples obtained from our prior study of 24 children infected with SARS-CoV-2 from 2020 to 2021 (INF), prior to the availability of vaccines for children. Collection of these samples was previously described.\u003csup\u003e23\u003c/sup\u003e Prior infection in these individuals was confirmed by serologic testing for anti-spike antibodies and/or the presence of SARS-COV-2-specifc B cells in the blood, tonsil, or adenoid by flow cytometry. From these 24 children, 15 PBMCs, 14 adenoids, and 22 tonsils were analyzed based on availability of cells (Supplementary Table 2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBlood and tissue collection\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBlood samples were obtained just prior to the surgical procedure in the operating room in serum separator tubes (BD) for serum collection and sodium heparin tubes (BD) for peripheral blood mononuclear cells (PBMCs) extraction from an intravenous line placed for anesthesia. Once received in the laboratory on the day of collection, serum separator tubes were spun at 1200 g for 10 min, and serum was aliquoted and stored at -80\u0026deg;C. PBMCs were isolated the day after collection by density gradient centrifugation (Lymphocyte Separation Medium, MP Biomedicals) at 1500 rpm for 30 min at room temperature (RT) with no brake and washed with PBS. If red blood cell contamination was present, cells were lysed with ammonium-chloride-potassium (ACK) buffer (Gibco).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTonsil and adenoid tissues were stored in RPMI-1640 (RPMI) media with 5% heat-inactivated fetal bovine serum (FBS, VWR), gentamicin 50 mg/mL (Gibco), and 1 \u0026times; antibiotic/antimycotic solution (Gibco) on ice immediately after collection. Tissues were processed the day after collection. The tissue was mechanically disrupted and filtered through a 100 \u0026mu;m cell strainer to create a single cell suspension, lysed with ACK buffer, and washed with PBS three times. Cells were then stored in liquid nitrogen in the presence of FBS with 10% DMSO and thawed according to our published protocol \u003csup\u003e92\u003c/sup\u003e for flow cytometry, CITE-seq, and functional assays below.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSeroreactivity of samples to SARS-CoV-2 spike, nucleocapsid and ORF8 by ELISA\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e96 well Immulon plates were coated with 20 ng/100 \u0026micro;L of recombinant nucleocapsid, spike RBD protein, or\u0026nbsp;Open Reading Frame 8 (ORF8) from WA1/2020 in PBS overnight at 4 \u003csup\u003eo\u003c/sup\u003eC. Starting at a 1:20 dilution, serum samples were serially diluted 5-fold and applied to the coated well for 1 hr at ambient temperature. Serum samples were assayed in duplicate, as described before \u003csup\u003e13, 93\u003c/sup\u003e. After three washes with PBS/0.05% Tween 20, bound human IgG antibodies were detected with 1:5000 dilution of HRP-conjugated anti-human IgG Fc-specific antibody (Jackson Immuno Research). After 1 hr, plates were washed with PBST followed by PBS, and o-Phenylenediamine dihydrochloride (OPD) was added for 10 min. Absorbance was measured at 492 nm. End-point titer was determined as 3-fold above the average of the absorbance values of the binding of serum samples to blank control wells. The end-point titer is reported as the serum dilution that was above this cutoff and was calculated using Prism 9 (GraphPad Software).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSARS-CoV-2 serum neutralization assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSamples were evaluated in a qualified SARS-CoV-2 pseudovirion neutralization assay (PsVNA) using SARS-CoV-2 WA1/2020 strain and Omicron BA.1 subvariant. SARS-CoV-2 neutralizing activity measured by PsVNA correlates with PRNT (plaque reduction neutralization test with authentic SARS-CoV-2 virus) in previous studies\u003csup\u003e94, 95, 96\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNeutralization assays were performed as previously described\u003csup\u003e13, 93, 95, 96, 97\u003c/sup\u003e. Briefly, 50 \u0026micro;L of SARS-CoV-2 S pseudovirions (counting ~200,000 relative light units) were pre-incubated with an equal volume of medium containing serial dilutions (starting at 1:10) of all samples at RT for 1 hr. Then, 50 \u0026micro;L of virus-antibody mixtures were added to 293T-ACE2-TMPRSS2 cells (10\u003csup\u003e4\u003c/sup\u003e cells/50 \u0026mu;L) in a 96-well plate. The input virus with all SARS-CoV-2 strains was the same (2 \u0026times; 10\u003csup\u003e5\u003c/sup\u003e relative light units/50 \u0026micro;L/well). After a 3 hour incubation, fresh medium was added to the wells. Cells were lysed 24 hour later, and luciferase activity was measured using One-Glo luciferase assay system (Promega). The assay of each sample was performed in duplicate, and the 50% neutralization titer was calculated using Prism 9 (GraphPad Software). The limit of detection for the neutralization assay is 1:20. Two independent biological replicate experiments were performed for each sample and variation in PsVNA50 titers was \u0026lt;10% between replicates.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eHigh dimensional flow cytometry\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eSARS-CoV-2 specific B cell characterization with spectral flow cytometry\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e5 million cells per sample of PBMC, adenoid, or tonsil were resuspended in PBS with 2% FBS and 2 mM EDTA (FACS buffer). Biotinylated probes to SARS-CoV-2 were crosslinked with fluorochrome-conjugated streptavidin in a molar ratio of 4:1. Fluorochrome-conjugated streptavidin was split into 5 aliquots and conjugated to biotinylated probes by mixing for 20 min/aliquot at 4 \u0026deg;C. Four fluorochrome-conjugated probes were used in this assay, including spike S1 from the Wuhan strain (BioLegend) conjugated to APC, RBD from the Wuhan strain (BioLegend) conjugated to BV421, RBD from the Omicron variant (Acro) conjugated to BUV615, and RBD from the Omicron variant (Acro) conjugated to PE. Cells were first stained with the viability dye, Zombie NIR (1:800 dilution, BioLegend), for 15 min at RT, washed twice and then incubated with the 4 fluorochrome-conjugated probes plus d-biotin (Avidity) and Brilliant Stain Buffer Plus (BD Biosciences) at 4 \u0026deg;C for 1 hr. Then, cells were washed twice, and resuspended with True-Stain Monocyte Blocker (1:10 in 50\u0026nbsp;mL) (BioLegend) for 5 min. Anti-CXCR3 antibody and an antibody cocktail containing the rest of the surface antibodies (Supplementary Table 11) and Brilliant Stain Buffer Plus were sequentially added directly to the cells and incubated for 5 min and 30 min at RT, respectively (200\u0026nbsp;mL total staining volume). Cells were washed three times and fixed in 1% paraformaldehyde for 20 min at RT before washing again and collecting on a spectral flow cytometer (Aurora, Cytek). Antibodies are listed in Supplementary Table 11. Gating strategies are shown in Supplementary Figure 1 \u0026ndash; 2.\u003c/p\u003e\n\u003cp\u003eA separate panel including intracellular transcription factor staining was also used to stain PBMC and tonsil cells \u003csup\u003e92\u003c/sup\u003e. Antibodies are listed in Supplementary Table 11.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eImmunophenotyping with 37 color spectral flow cytometry panel\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo million PBMCs per sample and 5 million cells per adenoid or tonsil were resuspended in FACS buffer after thawing. Cells were stained and acquired as described in our prior study \u003csup\u003e23\u003c/sup\u003e. Antibodies are listed in Supplementary Table 11.\u0026nbsp;Manual gating for both panels was conducted with FlowJo Software v.10.9.0 (BD Biosciences) as in Supplementary Figure 3 - 4.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eUnsupervised analysis of high dimensional flow cytometry data\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eS1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells from the SARS-CoV-2-antigen-specific B cell panel with 29 parameters were analyzed with unsupervised clustering of surface antibody staining. S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e live CD45\u003csup\u003e+\u003c/sup\u003eCD3\u003csup\u003e-\u003c/sup\u003eCD14\u003csup\u003e-\u003c/sup\u003eCD19\u003csup\u003e+\u003c/sup\u003e B cells from INF and VAC groups were analyzed. Cells from tonsil and adenoid samples were merged and processed together, while PBMC samples were processed separately due to differences in cell populations. B cell analysis was based on surface expression of IgA, IgD, IgM, IgG, CD27, CD38, CD21, CD95, CD11c, FCRL3/FCRL5, FCRL4, CD10, CD86, CD83, CD69, CD71, CXCR3, CD85J, and CD62L. Channel values (scaled output with compensated parameters) of S1\u003csup\u003e+\u003c/sup\u003eRBD\u003csup\u003e+\u003c/sup\u003e B cells from each sample were exported from FlowJo and then processed in R (v. 4.4.2) \u003cem\u003evia\u003c/em\u003e Rstudio (2023.12.1+402). Each cell was given an index and labeled with its origin subject\u0026rsquo;s identification and tissue type. Data were further processed using Seurat\u003csup\u003e98, 99\u003c/sup\u003e(v. 5.1.0). Cell clustering was performed by applying the FindNeighbors() function\u003csup\u003e98\u003c/sup\u003eon a distance matrix generated from the dist function by \u0026ldquo;euclidean\u0026rdquo; method, followed by Leiden clustering on the resulting SNN graph using Seurat\u0026rsquo;s FindClusters() algorithm, with a resolution parameter of 1.0 in PBMC and 0.6 in tissues. Expression of selected markers was visualized with their mean expression in each cluster by pheatmap (v. 1.0.12), and the downstream analysis and results were processed using ggplot2 (v. 3.5.1)\u003csup\u003e100\u003c/sup\u003e, reshape2 (v. 1.4.4)\u003csup\u003e101\u003c/sup\u003e, ggpubr (v. 0.6.0)\u003csup\u003e102\u003c/sup\u003e, ggthemes (v. 5.1.0) \u003csup\u003e103\u003c/sup\u003e and tidyverse (v. 2.0.0).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatistical comparison with linear model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe majority of participants underwent tonsillectomy for obstructive sleep disordered breathing due to tonsillar hypertrophy (which includes those with obstructive sleep apnea diagnosed with polysomnography); a smaller portion had recurrent or chronic tonsillitis as their primary indication. We and others have noted that age and indication for tonsillectomy influence the immune cell populations in these tissues \u003csup\u003e56, 104\u003c/sup\u003e; however, none of the INF subjects had recurrent tonsillitis. This limitation prevented us from adjusting for this potential confounding factor. Sex had little effect in our cohort (evaluated with PCA analysis, data not shown).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCluster proportion, cell percentages from manual gating, or neutralizing titers in each sample were modeled linearly as a function of (1) age (\u0026ldquo;Age\u0026rdquo;), and (2) history of SARS-CoV-2 infection or vaccine exposure (\u0026ldquo;Group\u0026rdquo; includes two or three categories: VAC, INF, and CON depending on the comparison. Each comparison involves two categories at a time.). The following formula was used to estimate separate coefficients for each category ofGroup (adjusted for age): lm (Frequency/Percentage ~ Group + Age). To illustrate the extent of correction achieved by the linear model, both \u003cem\u003ep\u003c/em\u003e-values from the two-sided Wilcoxon signed ranks test (in black) and \u003cem\u003ep\u003c/em\u003e-values from linear model (in blue) were presented in plots built with ggplot2(v. 3.5.1) \u003csup\u003e100\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSingle cell RNA sequencing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eSorting of\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eS1\u003csup\u003e+\u0026nbsp;\u003c/sup\u003eandS1\u003csup\u003e-\u003c/sup\u003e\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;B cells for CITE-seq\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe performed a combined analysis by merging data from three sets of\u0026nbsp;cellular indexing of transcriptomes and epitopes (CITE-seq) experiments (B-2021, B-306-3 and B-306-4). B-2021 was performed and processed as reported in our prior study \u003csup\u003e23\u003c/sup\u003e. Briefly in B-2021, paired\u0026nbsp;PBMC, adenoid, and tonsil samples from 3 donors (total of 9 donor-tissue samples) were assessed and \u0026lsquo;hashtag\u0026rsquo; antibodies (BioLegend) were used to uniquely label each of the 9 donor-tissue samples \u003csup\u003e105\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTwo additional sets of CITE-seq experiments were conducted separately to include VAC samples (B-306-3 and B-306-4).\u0026nbsp;For B-306-3, 2 INF and 1 CON tonsil were processed. For B-306-4, 3 VAC and 4 INF adenoid and 2 VAC, 2 INF, and 1 CON tonsil samples were processed as three sets (B-306-4 A, B, C) on 3 separate days over the course of one week to ensure manageable sorting times on each processing day (See Supplementary Table 8 for details on which samples were processed with each experiment). Frozen tonsil and adenoid cells were\u0026nbsp;thawed from liquid nitrogen as described previously \u003csup\u003e23, 92\u003c/sup\u003e. For each donor, 100,000-500,000\u0026nbsp;thawed tonsillar cells were reserved for bulk RNA-seq. During data analysis, individual cells were demultiplexed using donor-specific single nucleotide polymorphism (SNP) information obtained from the bulk RNA-seq data (see Methods: CITE-seq processing and demultiplexing).\u003c/p\u003e\n\u003cp\u003eThe remaining cells of the same tissue type from each donor were pooled together (i.e. adenoids and tonsils were pooled separately). The number of cells from each donor to pool was estimated using flow cytometry data with the aim of pooling a similar number of S1\u003csup\u003e+\u003c/sup\u003e B cells from each sample. Pooled cells were incubated with Fc blocker at 4 \u0026deg;C for 10 min followed by CITE-seq and sorting antibody cocktails in the following order at 4 \u0026deg;C: TotalSeq anti-CXCR3, anti-CCR6, and anti-CXCR5 antibodies for 10 min, and the remaining 21 CITE-seq antibodies and fluorescence-labeled sorting antibodies and viability dye (Aqua) for an additional 30 min (Antibodies are listed in Supplementary Table 11). Cells were then washed with PBS with 0.04% BSA and sorted on a BD FACS Aria\u0026trade; III sorter (BD Biosciences, San Jose, CA). S1\u003csup\u003e+\u003c/sup\u003e B cells from each tissue pool were sorted into a 200 uL low binding PCR tube (Eppendorf) with 50\u0026nbsp;mL culture medium (10% heat-inactivated FBS (VWR), 2 nM glutamine, 0.055 mM 2-mercaptoethanol, 1% penicillin/streptomycin, 1 mM sodium pyruvate, 10 mM HEPES, 1% non-essential amino acids in RPMI with glutamine), while S1\u003csup\u003e-\u003c/sup\u003e B cells were sorted into 1.5 mL low binding tubes (Eppendorf) with 300\u0026nbsp;ml RPMI culture medium with 20% heat-inactivated FBS. All collecting low binding tubes were precoated with RPMI media with 20% FBS overnight at 4 \u0026deg;C prior to use.\u0026nbsp;Sorting strategy is shown in Supplementary Figure 5A.\u0026nbsp;Antibodies, including barcoded and fluorescence-conjugated antibodies, are listed in Supplementary Table 11.\u0026nbsp;S1⁺\u0026nbsp;cells were centrifuged and resuspended for single cell partitioning without counting. An appropriate number of S1⁻\u0026nbsp;B cells were counted, aliquoted, centrifuged and resuspended for partitioning. Antibody concentrations used for CITE-seq were inferred based on the concentrations of antibodies of the same clones titrated for use in our flow cytometry staining of tonsil/adenoid cells.\u0026nbsp;\u003c/p\u003e\n\u003ch4\u003e\u003cstrong\u003eCITE-seq processing and demultiplexing\u003c/strong\u003e\u003c/h4\u003e\n\u003cp\u003eLibrary construction and sequencing for B-2021 were described previously \u003csup\u003e23\u003c/sup\u003e. For B-306-3 and B-306-4, sorted S1\u003csup\u003e+\u003c/sup\u003e and S1\u003csup\u003e-\u003c/sup\u003e B cells were mixed with the reverse transcription mix and partitioned into single cell gel-bead in emulsion (GEM) using 10x 5\u0026rsquo; Chromium Next GEM Single Cell 5\u0026rsquo; Reagent Kit v2 for B-306-3 and 10x 5\u0026rsquo; Chromium Next GEM Single Cell 5\u0026apos; HT v2 \u0026nbsp;for B-306-4 (10x Genomics). The reverse transcription step was performed in\u0026nbsp;an Applied Biosystems Veriti 96-well thermal cycler (Applied Biosystems). 5\u0026rsquo; single cell gene expression (GEX), cell surface protein (CSP), and B cell receptor (BCR) libraries were prepared as instructed by 10x Genomics user guides (\u003ca href=\"https://www.10xgenomics.com/resources/user-guides/\" title=\"https://www.10xgenomics.com/resources/user-guides/\"\u003ehttps://www.10xgenomics.com/resources/user-guides/\u003c/a\u003e; CG000330 Rev F and CG000424 Rev C for the B-306-3 and B-306-4 run, respectively). Library quality and quantities were measured using a TapeStation system (Agilent) and a Qubit fluorometer (ThermoFisher). Libraries were pooled at a concentration of 10 nM and sequenced on the Illumina platform (NovaSeqX for B-306-3 and B-306-4, Illumina) using the following read lengths: Read 1: 26 base pairs, Index 1: 10 base pairs, Index 2: 10 base pairs, Read 2: 90 base pairs.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTonsil cells saved for bulk RNA-seq from each participant were resuspended in QIAzol and RNA was extracted using RNeasy micro kit (Qiagen) and standard RNA sequencing libraries were generated using Universal Plus mRNAseq kit (TECAN Genomics). These libraries were used to generate SNP calls for each donor. Sequencing results were demultiplexed and converted to FASTQ format using Illumina bcl2fastq software. The sequencing reads were adaptor and quality trimmed and then aligned to the human genome using the splice-aware STAR aligner and SNP calls were called using bcftools \u003csup\u003e106\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe pooled single cell RNA sequencing data was demultiplexed with a combination of demuxalot (v. 0.4.1) \u003csup\u003e107\u003c/sup\u003e and demuxlet \u003csup\u003e108\u003c/sup\u003e to match cells to each donor and identify doublets. For the pool with fraternal twin subjects, CNMC 124 and CNMC 125, cells were assigned by demuxalot, then reads with associated barcodes were extracted and reassigned via demuxlet to further refine the cell identities for twin subjects.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eCITE-seq data analyses\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData from B-2021 set were processed as described \u003csup\u003e23\u003c/sup\u003e. For B-306-3 and B-306-4, CellRanger (v. 7.1.0) \u003csup\u003e109\u003c/sup\u003e was used to map cDNA libraries to the GRCh38 genome reference (GRCh39-2020-A) and to count antibody tag features. Down-sampling was performed using the cellranger aggr pipeline (10x Genomics) to normalize sequencing depth across B cell lanes. Data were further processed using Seurat (v. 5.1.0) in R 4.4.2. Cells were demultiplexed by SNP as described above (Methods: Single cell CITE sequencing and demultiplexing). Surface protein library counts were transformed by using dsb (v. 1.0.4) \u003csup\u003e110\u003c/sup\u003e. For quality control, cells with less than 100 detected genes, greater than 30% mitochondrial reads, or gene counts greater than 25,000 were removed. To exclude cells with extremely high surface antibody counts, the top 0.5%\u0026nbsp;of cells in the surface antibody total count distribution were removed. Cell clustering was performed by applying the FindNeighbors() function from Seuraton a distance matrix generated from the dsb-transformed surface protein data, followed by Leiden clustering on the resulting SNN graph using Seurat\u0026rsquo;s FindClusters() algorithm, with a resolution parameter of 0.3. Expression of selected genes were visualized using the ComplexHeatmap package \u0026nbsp;(v. 2.22.0) \u003csup\u003e111\u003c/sup\u003e, and the percentage of cells per population for the S1\u003csup\u003e+\u003c/sup\u003e and S1\u003csup\u003e-\u003c/sup\u003e cells was plotted using ggplot\u003cem\u003e2\u0026nbsp;\u003c/em\u003e(v. 3.5.2) \u003csup\u003e100\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo further assess the transcriptome of B\u003csub\u003eSM\u003c/sub\u003e subsets P1 to P4, we manually annotated B cell populations using protein expression data from CITE-seq for three reasons: (1) to align our analysis with flow cytometry data; (2) CXCR3 protein and mRNA expression levels were poorly correlated (Extended Data Figure 4a); and (3) to minimize batch effects across different experimental sets. We used the removeBatchEffect function from the limma (v. 3.62.1) \u003csup\u003e112\u003c/sup\u003e package to correct batch effects in CSP expression within the B-306-4 set. CSP expression from B-2021, B-306-3, and B-306-4 were exported as .csv files and imported into FlowJo (v. 10.9.0) for manual gating. The gated FlowJo files were then processed using CytoML (v. 2.18.0) \u003csup\u003e113\u003c/sup\u003e and flowWorkspace (v. 4.18.0) \u003csup\u003e114\u003c/sup\u003e, and the resulting gates were merged with the CITE-seq metadata by matching identical cell barcodes.\u003c/p\u003e\n\u003cp\u003eTo identify differentially expressed genes among B\u003csub\u003eSM\u003c/sub\u003e P1 to P4 (Extended Data Figure 4d), we used only the GEX from B-306-4 to minimize batch effects. Differential expression analysis was performed using the FindAllMarkers function in Seurat with the MAST algorithm, incorporating \u0026apos;Batch\u0026apos; (defined as samples processed on different days within the B-306-4, set A-C), \u0026apos;Subject\u0026apos;, and \u0026apos;Tissue\u0026apos; as latent variables (min.pct = 0.1, min.cells.feature = 3,\u0026nbsp;min.cells.group = 3, logfc.threshold = 0.1). The PseudobulkExpression function was used to normalize count data for each B\u003csub\u003eSM\u003c/sub\u003e population, yielding representative expression values for differentially expressed genes within each subset. The top 20 genes with adjusted p-values \u0026lt; 0.05 and log2FC \u0026gt; 0, as identified by the FindAllMarkers results, were visualized using pheatmap (v. 1.0.12) \u003csup\u003e115\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003ePseudo-bulk and differential gene expression analysis\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePseudo-bulk gene differential expression analysis and gene set enrichment analysis (GSEA) were performed as described previously \u003csup\u003e116, 117\u003c/sup\u003e. In brief, all sorted cells in a given cell type (S1\u003csup\u003e+\u003c/sup\u003e or S1\u003csup\u003e-\u003c/sup\u003e) and tissue (adenoid or tonsil) within a donor were computationally \u0026lsquo;pooled\u0026rsquo; according to their B cell subset assignment by summing all reads for a given gene. Pseudo-bulk libraries composed of only a few cells (less than 5), and with fewer than 30,000 unique molecular identifier counts after pooling per library\u0026nbsp;were removed from the analysis, as they are likely not modeled properly by bulk differential expression methods. Only cell type- and tissue-specific B cell subsets with more than\u0026nbsp;3 psuedobulk\u0026nbsp;libraries were included for differential comparison. Genes expressed at low levels were removed for each cell type individually using the filterByExpr function from edgeR\u0026nbsp;\u003csup\u003e118\u003c/sup\u003e. Differentially expressed genes were identified using the limma voom\u0026nbsp;\u003csup\u003e119\u003c/sup\u003e workflow, which models the log of the counts per million (CPM) of each gene. Scaling factors for library size normalization were calculated with the calcNormFactors function with method\u0026thinsp;=\u0026thinsp;\u0026lsquo;RLE\u0026rsquo;.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eGenes were ranked using the moderated T statistics for the relevant coefficient from the limma voom model. Differentially expressed genes between S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e P1 and S1\u003csup\u003e+\u003c/sup\u003e B\u003csub\u003eSM\u003c/sub\u003e P2 cells were identified with a model accommodating paired analysis (formula: ~ 0+CD21_CXCR3+Subject_Tissue). The S1\u003csup\u003e+\u003c/sup\u003e P1 and P2 B\u003csub\u003eSM\u003c/sub\u003e subset (\u0026ldquo;CD21_CXCR3\u0026rdquo;) and tissue-specific subject (\u0026ldquo;Subject_Tissue\u0026rdquo;) were modeled as factor variable representing the B\u003csub\u003eSM\u003c/sub\u003e population (including \u0026ldquo;B\u003csub\u003eSM\u003c/sub\u003e P1\u0026rdquo; and \u0026ldquo;B\u003csub\u003eSM\u003c/sub\u003e P2\u0026rdquo;), tissue type from a specific subject (including \u0026ldquo;tonsil\u0026rdquo; and \u0026ldquo;adenoid\u0026rdquo;), respectively.\u0026nbsp;The contrasts.fit function was then used to compare the estimated means between S1\u003csup\u003e+\u003c/sup\u003e P1\u0026nbsp;B\u003csub\u003eSM\u003c/sub\u003e and S1\u003csup\u003e+\u003c/sup\u003e P2\u0026nbsp;B\u003csub\u003eSM\u003c/sub\u003e.\u003c/p\u003e\n\u003cp\u003eEnriched gene sets were identified using the pre-ranked gene-set enrichment analysis (GSEA) algorithm implemented in the fgsea (v. 1.32.0) R package. Gene set lists used for enrichment assessment (including Gene Ontology Biological Process (GO BP), GO Cellular Component (GO CC), GO Molecular Function (GO MF), Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, the Molecular Signatures Database\u0026rsquo;s Hallmark collection, Blood Transcriptomic Modules and a few published datasets \u003csup\u003e59, 120, 121\u003c/sup\u003e) were collected and pooled. \u003cem\u003eP\u003c/em\u003e values were adjusted using the Benjamini\u0026ndash;Hochberg method for the whole gene set list.\u0026nbsp;The pathways shown in Figure 5l were manually curated from gene sets relevant to immunology.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eSignature scores of B cell subsets\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGene sets for atMBC were collected from studies on various chronic diseases in humans and mice including malaria \u003csup\u003e122\u003c/sup\u003e, HIV infection \u003csup\u003e123\u003c/sup\u003e, systemic lupus erythematosus \u003csup\u003e124\u003c/sup\u003e, Sj\u0026ouml;gren\u0026apos;s syndrome \u003csup\u003e125\u003c/sup\u003e, rheumatoid arthritis, common variable immunodeficiency \u003csup\u003e126\u003c/sup\u003e , and rodent malaria \u003cem\u003ePlasmodium chabaudi\u003c/em\u003e\u003csup\u003e59\u003c/sup\u003e, as well as human tonsil \u003csup\u003e59\u003c/sup\u003e. Gene set for GCB signature was derived from a human tonsil B cell study \u003csup\u003e127\u003c/sup\u003e. \u0026nbsp;cMBC signature was used from a recent study on human tonsil and murine malaria \u003csup\u003e59\u003c/sup\u003e. Other gene signatures used to characterize B cells were obtained from MSigDB \u003csup\u003e128\u003c/sup\u003e (msigdbr R package, v. 7.5.1). The AddModuleScore function of Seurat was applied with default parameters to score each signature in each B cell.\u0026nbsp;To represent the levels of each signature set in the populations of interest, the median score for each tissue- and subject-specific sample within the S1\u003csup\u003e-\u0026nbsp;\u003c/sup\u003eB cell subset from the B-306-4 set was used as an estimate. The median scores of per-subject median scores for each B cell population were then displayed in the heatmap of Figure 5j.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTrajectory analysis\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo compare the lineage differentiation of S1\u003csup\u003e+\u003c/sup\u003e B cells in INF and VAC tonsils and adenoids, we applied Slingshot (v. 2.14.0)\u0026nbsp;\u003csup\u003e129\u003c/sup\u003e to infer lineage trajectories. S1\u003csup\u003e+\u003c/sup\u003e cells from B-306-4 set were used in this analysis. B cell subsets were assigned by manual gating as described in Figure 5 g and h. \u0026ldquo;PC\u0026rdquo; and \u0026ldquo;PreGC\u0026rdquo; cells were removed due to low number of cells for an accurate lineage estimation. First, uniform manifold approximation and projection (UMAP) embedding was performed on the normalized single-cell surface protein expression profile from the CITE-seq dataset to obtain a low-dimensional representation. P1-P4 B\u003csub\u003eSM\u003c/sub\u003e populations were provided as cluster input and \u0026ldquo;Na\u0026iuml;ve/USM B cell\u0026rdquo; population was appointed as the start cluster to Slingshot for trajectory reconstruction. Cells were ordered through inferred pseudotime based on gene expression to indicate their differentiation progress. Trajectories for INF and VAC groups were estimated separately.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eBCR sequence analysis and clonal clustering\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBCR repertoire sequence data were analyzed using the Immcantation (\u003ca href=\"http://www.immcantation.org\"\u003ewww.immcantation.org\u003c/a\u003e, v. 4.5.0) framework. Starting with filtered CellRanger output, V(D)J genes for each sequence were aligned to the IMGT GENE-DB reference database obtained 2/09/24 using IgBLAST (v. 1.22.0) \u003csup\u003e130\u003c/sup\u003e and Change-O (v. 1.3.0) \u003csup\u003e131\u003c/sup\u003e. Nonproductive sequences, cells without associated constant region calls, cells identified as arising from doublets or negative droplets, and cells with multiple heavy chains were all removed. Samples within each subject were pooled and sequences were grouped into clonal clusters, which contain B cells related to each other by somatic hypermutation (SHM) from a common V(D)J ancestor. Using the hierarchicalClones function of scoper (v. 1.3.0) \u003csup\u003e132\u003c/sup\u003e, sequences within these groups differing by a participant-specific normalized Hamming distance threshold within the CDR3 region were defined as clones using single-linkage clustering (one subject required a slightly lower threshold than others) \u003csup\u003e133\u003c/sup\u003e. This threshold was determined by fitting a gamma/Gaussian mixture model to the distance to nearest sequence neighbor distribution using SHazaM(v. 1.2.0) \u003csup\u003e134\u003c/sup\u003e. These heavy chain-defined clonal clusters were further split if their constituent cells contained light chains that differed by V and J genes. Within each clone, germline sequences were reconstructed with D segment and N/P regions masked (replaced with \u0026ldquo;N\u0026rdquo; nucleotides) using the createGermlines function within dowser (v. 2.3) \u003csup\u003e135\u003c/sup\u003e.\u0026nbsp;SHM was calculated as the frequency of non-ambiguous mismatches from each cell to the V gene (IMGT positions 1\u0026ndash;312) of its reconstructed germline sequence. Paired scBCR-seq data were integrated with CITE-seq data based on matched cell barcodes.\u003c/p\u003e\n\u003cp\u003eTo quantify B cell clonal diversity, we calculated Simpson\u0026rsquo;s diversity for each tissue-specific sample using the alphaDiversity function of alakazam (v. 1.3.0) \u003csup\u003e131\u003c/sup\u003e. Lower values of Simpson\u0026rsquo;s diversity indicate a greater probability of two random sequences belonging to the same clone, consistent with more large clones. To account for differences in sequence depth, samples within each comparison were down-sampled to the same number of sequences, and the mean of 1000 such re-sampling repetitions was reported. Subject/tissue/cell type samples or populations with \u0026lt; 70 B cells were excluded, which led to the exclusion of all S1\u003csup\u003e+\u003c/sup\u003e cells from CNMC 10 and CNMC 99 (both CON with no history or evidence of SARS-CoV-2 infection or vaccination). Clonal overlap among tissues or B cell subsets can be used as a measure of immunological connectivity. Clonal overlap was calculated using the Jaccard index, which for each pair of B cell subsets is the number of unique clones found in both subsets (intersect) divided by the total number of unique clones among the two subsets (union). Clones were relabeled as \u0026ldquo;S1\u003csup\u003e-\u003c/sup\u003e\u0026rdquo; clone when the ratio of S1⁺\u0026nbsp;to S1⁻\u0026nbsp;sorted B cells within the clone was less than 0.1. After clonal clustering, only heavy chain sequences were used for subsequent analysis. Clone expansion and B cell subset transition were estimated with indices (STARTRAC-expa and STARTRAC-tran) from the STARTRAC package (v. 0.1.0) \u003csup\u003e69, 70\u003c/sup\u003e. In STARTRAC-tran indices analysis, S1\u003csup\u003e+\u003c/sup\u003e or S1\u003csup\u003e-\u003c/sup\u003e subject specific populations with less than 10 cells were removed. Clones from INF and VAC groups were analyzed.\u003c/p\u003e\n\u003cp\u003eTo infer lineage trees, we estimated tree topologies, branch lengths, and subject-wide substitution model parameters using maximum likelihood under the GY94 model \u003csup\u003e136, 137\u003c/sup\u003e. Using fixed tree topologies estimated from the GY94 model, we then estimated branch lengths and donor-wide parameter values under the HLP19 model in IgPhyML (v. 2.0.0) \u003csup\u003e136\u003c/sup\u003e. Trees were visualized using dowser (v. 2.3) \u003csup\u003e135\u003c/sup\u003e and ggtree (v. 3.14.0) \u003csup\u003e138\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eATAC-seq\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eATAC-seq data processing and analysis\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFrozen tonsil cells were thawed and stained with antibodies listed in Supplementary Table 11 for 30 min at RT. Cells were washed with PBS twice and resuspended in RPMI with 10% FBS for sorting. 10,000 viable P1 (CXCR3\u003csup\u003e-\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e), P2 (CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+\u003c/sup\u003e), P3 (CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e-\u003c/sup\u003e) and P4 (CXCR3\u003csup\u003e-\u003c/sup\u003eCD21\u003csup\u003e-\u003c/sup\u003e) B\u003csub\u003eSM\u003c/sub\u003e were sorted into 500 mL of\u0026nbsp;FACS buffer using an FACSAria Fusion Sorter (BD) (cell sorting strategy is shown in Supplementary Figure 5B). Cells were pelleted and resuspended in 50 uL of\u0026nbsp;transposase mixture including 25 mL 2xTD buffer (Illumina), 2.5 mL TDE1 (Illumina), 0.5 mL 1% digitonin (Promega) and 22 uL water. Tagmentation was performed by incubation at 37 \u0026deg;C for 30 minutes at 300 rpm. Following incubation, DNA was purified using a Qiagen MinElute kit, eluting samples in 10 uL. Purified tagmented DNA was PCR amplified using previously described primers \u003csup\u003e139\u003c/sup\u003e, with 12 cycles of amplification. Amplified libraries were purified using a Qiagen PCR cleanup kit and sequenced for 50 cycles (paired-end reads) on a NovaSeq 6000 (Illumina). ATAC-seq was done in three biological replicates per B\u003csub\u003eSM\u003c/sub\u003e subset (samples shown in Supplementary Table 2). ATAC-seq primers are listed in Supplementary Table 11.\u003c/p\u003e\n\u003cp\u003eATAC-seq data was processed using the chrom-seek pipeline (v. 1.0.0) \u003csup\u003e140\u003c/sup\u003e with --assay ATAC (https://github.com/OpenOmics/chrom-seek). Reads were trimmed with Cutadapt (v. 4.4) \u003csup\u003e141\u003c/sup\u003e. All reads aligning to the Encode hg38 v1 blacklist regions \u003csup\u003e142\u003c/sup\u003e were identified by alignment with BWA (v. 0.7.17) \u003csup\u003e143\u003c/sup\u003e and removed with Picard SamToFastq. Remaining reads were aligned to an hg38 reference genome using BWA. Reads with a mapQ score less than 6 were removed with SAMtools (v. 1.17) \u003csup\u003e106\u003c/sup\u003e and PCR duplicates were removed with Picard MarkDuplicates. Data was converted into bigwigs for viewing and normalized by reads per genomic content (RPGC) using deepTools\u0026nbsp;\u003csup\u003e144\u003c/sup\u003e (v. 3.5.1) using the following parameters: --binSize 25 --smoothLength 75 --effectiveGenomeSize 2805636331 --centerReads --normalizeUsing RPGC. Averaged bigwigs were created using the bigwigAverage function of deepTools (v. 3.5.4)\u0026nbsp;\u003csup\u003e144\u003c/sup\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003ePeaks were called using macsNarrow \u003csup\u003e145\u003c/sup\u003e (macs v. 2.2.7.1) with the following parameters: -q 0.01 --keep-dup=\u0026quot;all\u0026quot; -f \u0026quot;BAMPE\u0026quot;. Differential peaks were called using DiffBind (v. 2.15.2) \u003csup\u003e146\u003c/sup\u003e and its Deseq2 differential caller with default parameters. Peaks were considered significant with an FDR value less than 0.1. Motif analysis was completed using the MEME suite (v. 5.5.5) \u003csup\u003e147\u003c/sup\u003e. Known motif enrichment analysis was accomplished using AME on a combined jolma 2013, jaspar 2018 core vertebrate non-rendundant, and HOCOMOCO (v.11) \u003csup\u003e148\u003c/sup\u003e full human mono database. Downstream analyses and results visualization were performed with R (v.4.4.2) and visualized with ggplot2 (v. 3.5.1) \u003csup\u003e100\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eDifferentially accessible regions (DAR) and pathway enrichment analysis\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDifferential peaks from comparing P1 B\u003csub\u003eSM\u003c/sub\u003e and P2 B\u003csub\u003eSM\u003c/sub\u003e (adjusted \u003cem\u003ep\u003c/em\u003e-value \u0026lt; 0.10) were selected for downstream analysis. To explore the functional significance of these DARs, pathway enrichment analysis was conducted using GREAT (Genomic Regions Enrichment of Annotations Tool, v. 4.0.4) \u003csup\u003e149, 150\u003c/sup\u003e. The genomic coordinates of DARs were converted to BED files and uploaded to the GREAT web server. The analysis was performed mapping to GRCh38\u0026nbsp;\u003ca href=\"http://genome.ucsc.edu/cgi-bin/hgGateway?db=hg38\" target=\"_blank\"\u003e(UCSC hg38, Dec. 2013)\u003c/a\u003e and using the default association rules, which map genomic regions to nearby genes based on a basal plus extension model (5 kb upstream, 1 kb downstream, and up to 1 Mb extension from the transcription start site). Enriched terms from GO BP and GO MF, MSigDB, Reactome, and other curated databases were extracted. Pathways and terms with an FDR (adjusted \u003cem\u003ep\u003c/em\u003e-value) \u0026lt; 0.05 were considered significant. Top20 GO BP pathways ranked by FDR (adjusted \u003cem\u003ep\u003c/em\u003e-value) values were visualized with bar plot.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eIRF4 complex motif search with FIMO\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further evaluate the enrichment of IRF4 complex motifs in P1\u0026nbsp;B\u003csub\u003eSM\u003c/sub\u003e and P2 B\u003csub\u003eSM\u003c/sub\u003e, a search with stringent EICE (GGAANNGAAA), ISRE (A/GNGAAANNGAAACT) and two AICE (0bp/4bp:\u0026nbsp;0bp \u0026ndash; GAAATGA(G/C)TCA; 4bp \u0026ndash; TTTCNNNNTGA(G/C)TCA) \u003csup\u003e151\u003c/sup\u003e motifs using Find Individual Motif Occurrences (FIMO) on the DAR sequences obtained from P1\u0026nbsp;B\u003csub\u003eSM\u003c/sub\u003e to P2 B\u003csub\u003eSM\u003c/sub\u003e comparison was performed \u003csup\u003e67\u003c/sup\u003e.\u0026nbsp;The enrichment statistics were calculated as above using a two-tailed version of Fisher\u0026apos;s exact test.\u003c/p\u003e\n\u003cp\u003eFor Extended Data Figure 4i, data for memory B cells (MBC) and plasma cells (PC) were retrieved from a published dataset (https://zenodo.org/records/8373756) \u003csup\u003e99\u003c/sup\u003e and loaded into Signac (v1.11.0) \u003csup\u003e152\u003c/sup\u003e. Plots were created with the Signac CoveragePlot function.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBCR signaling assessment by phosphorylation staining\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTonsil cells were thawed, rinsed with RPMI supplemented with 0.1% FBS, and rested for 80 minutes at 37 \u0026deg;C in the same medium. After resting, the cells were incubated with a live/dead stain in PBS for 15 minutes at RT and then washed once with PBS. This was followed by staining the cells with antibodies against CXCR3 and CXCR5 for 5 minutes at RT. Subsequently, the remaining surface antibody mix was added (Supplementary Table 11), and the cells were resuspended in FACS buffer for 20 minutes at RT \u003csup\u003e92\u003c/sup\u003e. Afterward, the cells were washed twice with RPMI supplemented with 0.1% heat inactivated FBS and then resuspended in pre-warmed RPMI with 10% FBS and stimulated with anti-BCR antibodies, as previously described \u003csup\u003e71, 153, 154\u003c/sup\u003e.\u0026nbsp;Stimulation was carried out at 37 \u0026deg;C for 2 minutes using 10 mg/mL goat F(ab\u0026rsquo;)2 anti-human IgA/G/M (Jackson ImmunoResearch Laboratories). For detecting phosphorylated signaling intermediates, the cells were fixed and permeabilized using BD Cytofix and Phosflow Perm/Wash buffers (BD Biosciences), then stained with PE-phosphorylated Syk (p-Y348) and Alexa Fluor 488- phosphorylated PLCg2 (p-Y759) antibodies (BD Biosciences). Samples were acquired on an Aurora cytometer (Cytek), and analysis was performed using FlowJo (v. 10.9.0).\u0026nbsp;Gating strategies are shown in Supplementary Figure 6A.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eIn vitro\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;plasmablast differentiation and proliferation assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTonsil cells were thawed and stained with viability dye and surface antibody mix (Supplementary Table 11) at RT for 30 mins. Then cells were washed with PBS twice and resuspended in RPMI with 10% FBS at concentration of 5 million cells/mL for sorting. P1, P2, P3 and P4 B\u003csub\u003eSM\u003c/sub\u003e were sorted on Aria sorter (BD) into 0.3 mL RPMI with 20% FBS in 1.5 mL tubes (Supplementary Figure 5B). Sorted cells were centrifuged at RT for 10 mins and then were labeled with 0.5\u0026nbsp;mM CFSE (CellTrace CFSE cell proliferation kit, ThermoFisher) in 1100 \u0026mu;L PBS at RT in the dark for 10 mins. Then, cells were washed with pre-warmed RPMI + 10% FBS as described \u003csup\u003e153\u003c/sup\u003e. \u0026nbsp;Allogenic B cell-depleted PBMCs were prepared using a B cell depletion kit (Dynabeads CD19 Pan B, ThermoFisher) from PBMCs of an unrelated healthy donor. The sorted memory B cells and allogeneic B cell depleted PBMC were co-cultured at a 1:9 to 2:8 ratio with 2.5 mg/mL R848 and 1000 U/mL recombinant human IL-2 for 4 days at 1 \u0026times; 10\u003csup\u003e6\u003c/sup\u003e cells per well in a 96-well flatbottom plate. RPMI with 10% FBS was used for culture. The cells were collected and stained with antibodies against CD19, CD20, CD3, CD27, CD21, IgD, CD38, and CXCR3, fixed (Lysing Solution, BD Biosciences), permeabilized (Permeabilizing Solution 2; BD Biosciences) and stained with antibodies against IgG, IgA, IgM. The cells were acquired on an Aurora cytometer (Cytek) and the analyzed using FlowJo (v. 10.9.0). Antibodies are listed in Supplementary Table 11. Gating strategies are shown in Supplementary Figure 6B. The division index, a measure of the overall proliferative response, is the average number of divisions undergone per cell in the total population, including cells that have not undergone division.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTissue processing and staining for immunofluorescence assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFormalin-fixed, paraffin-embedded (FFPE) adenoid and tonsil tissue blocks were cut into 5 \u0026mu;m sections and mounted onto charged slides. Two paired tonsil and adenoid samples (from one INF donor and one VAC donor) were loaded onto the same slide. For staining, slides were deparaffinized and tissue rehydrated in deionized water. Antigen retrieval was performed by incubating slides in antigen retrieval (AR) buffer (Cepham life sciences) for 45\u0026thinsp;min in a steamer (preheated, approximately 95\u0026thinsp;\u0026deg;C). After 45 min, slides were taken out from the steamer and allowed to cool to RT. Sections were permeabilized, blocked for 1 hour in PBS containing 0.3 % Triton X-100 (Sigma-Aldrich), 1% bovine serum albumin (Jackson Immune Research). Sections were stained with titrated amounts of non-conjugated primary antibodies, followed by overnight incubation at 4 \u0026deg;C. Slides were then washed with PBS (3 times, 10 min each) and stained with the appropriate secondary antibodies for 2 hrs at RT. Slides were washed and blocked again for 1 hr at RT with a 1:10 dilution of normal mouse and rabbit or goat serum. Then, slides were stained with titrated amounts of directly conjugated antibodies for 2 hrs at RT. After three final washing steps and staining with the nuclear marker TOPRO (ThermoFisher), slides were mounted with prolonged gold anti-fade mounting media (ThermoFisher) and sealed with a glass coverslip. Antibodies are listed in Supplementary Table 11.\u003c/p\u003e\n\u003cp\u003eTissue sections were imaged (using confocal system, Leica Stellaris RTB WLL FLIM) as three-dimensional (3D) tile scans and subsequently mosaic-merged to generate a continuous representation. To minimize imaging artifacts, corrections were applied for motion-induced distortions, 3D alignment inconsistencies, and thermal drift across sequential z-sections. Additionally, crosstalk and color calibration adjustments were performed using Huygens Pro (version 24.04.0p3 64-bit, Scientific Volume Imaging BV). Image deconvolution was conducted within the same software to enhance signal resolution.\u003c/p\u003e\n\u003cp\u003eThe reconstructed images were further processed using Imaris (v. 10.2.0, Oxford Instruments). A combination of colocalization analysis, ChannelArithmetics, Xtension, Imaris installed machine learning-based classification, and masking techniques were employed to delineate distinct anatomical regions as additional computational channels, including follicles, germinal centers, extrafollicular regions, epithelium, and crypt structures. The Surface module of Imaris was utilized to generate cell objects, integrating nuclear signals and perinuclear regions derived from the preceding image processing steps.\u003c/p\u003e\n\u003cp\u003eQuantitative data were extracted from processed files \u003cem\u003evia\u003c/em\u003e a custom parser script, which reformatted surface object statistics into \u0026ldquo;.csv\u0026rdquo; files optimized for direct import into FlowJo (v. 10.9.0) for gating and further analysis. Statistical analysis and visualization were further processed in GraphPad Prism (v. 10.2.0392).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSPACE analysis of immunofluorescence imaging data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSpatial Patterning Analysis of Cellular Ensembles (SPACE) is an R package designed to identify complex spatial patterns at the cell and tissue levels \u003csup\u003e79\u003c/sup\u003e. Cell objects from immunofluorescence images were annotated with 8 populations as in Figure 8f by manual gating with FlowJo (v. 10.9.0) and labeled by an R-based customized script. The 10 \u0026micro;m radius captures close cellular associations and was chosen for SPACE \u0026ldquo;census_image\u0026rdquo; function. The number of neighborhoods was chosen to achieve 5 \u0026times; tissue coverage. Covariation plots were created using the SPACE \u0026ldquo;learn_pattern\u0026rdquo; function.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSpatial transcriptomic profiling with Xenium In Situ platform\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSlides were prepared following the manufacturer\u0026rsquo;s instructions and workflow for FFPE tissue samples (CG000578 Rev A; 10x Genomics). A 5-\u0026micro;m section from the tissue block containing the same\u0026nbsp;paired tonsil and adenoid samples (one from INF donor and one from VAC donor) used for immunofluorescence were carefully attached to the sample area on a Xenium slide (Histoserv, MD). Samples are listed in Supplementary Table 2. Xenium slides were deparaffined and rehydrated and were then assembled into the Xenium Cassette. Deparaffinized slides in the Xenium cassette were decrosslinked (CG000580 Rev A) and immediately underwent probe hybridization, ligation, and amplification (CG000582 Rev D) using the Xenium 5000 human gene panel (Prime 5K Human Pan Tissue \u0026amp; Pathways Panel). With autofluorescence quenching and nuclei staining, the tissue images were captured and analyzed by the Xenium Analyzer (PN-1000569, instrument software v2.0.1.0). Regions of interest were manually selected from the scanned images. Post-run data for each slide was obtained using default parameters for downstream analysis.\u003c/p\u003e\n\u003cp\u003eData were processed by Xenium analysis software (v 2.0.0.10). The raw count matrix was pre-processed using the Seurat package (v. 5.1.0) \u003csup\u003e98, 99\u003c/sup\u003e in R (v. 4.4.2). For quality control, cells with at least nCount \u0026gt; 40, 15 nFeature \u0026gt; 10, cell_area \u0026gt; 10 \u0026amp; \u0026lt; 200 were retained\u003csup\u003e155\u003c/sup\u003e. Raw count data were normalized using the SCTransform function with method \u0026quot;glmGamPoi\u0026quot;. Dimension reduction was performed using the runPCA function and the optimal number of principal components was selected using the ElbowPlot function. Cell clusters were determined using the FindClusters function.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe annotated cell populations with human tonsil reference version 2 with Seurat (v. 5.0.2) \u003csup\u003e98\u003c/sup\u003e and Azimuth (v. 0.5.0) \u003csup\u003e91, 99\u003c/sup\u003e. Cell populations with less than 150 cells per slide were removed (granulocytes, mast, preB/T and PC/doublet cells). Some cell types were merged (B na\u0026iuml;ve includes B naive \u0026amp; B activated; B memory includes B memory \u0026amp; FCRL4/5\u003csup\u003e+\u003c/sup\u003e B memory; CD4 Non-TFH includes CD4 Non-TFH \u0026amp; CD4 TCM; CD4 TFH includes CD4 TFH \u0026amp; CD4 TFH Mem; CD8 non-na\u0026iuml;ve includes CD8 T \u0026amp; CD8 TCM; gdT_MAIT includes MAIT/TRDV2 + gdT \u0026amp; non-TRDV2 + gdT; Mono/Macro includes Mono/Macro \u0026amp; Cycling myeloid; PB/PC includes PB \u0026amp; PC).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdditional statistical analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCorrelations were analyzed using Spearman\u0026rsquo;s rank correlation test using\u0026nbsp;base R and\u0026nbsp;corrplot\u0026nbsp;(v 0.95). Paired comparisons were performed with Wilcoxon signed ranks test with\u0026nbsp;ggpubr\u0026nbsp;(v. 0.6.0) \u003csup\u003e156\u003c/sup\u003e and visualized with ggplot2 (v. 3.5.1)\u0026nbsp;\u003csup\u003e100\u003c/sup\u003e in R (v. 4.4.2).\u003c/p\u003e"},{"header":"REFERENCES","content":"\u003col\u003e\n\u003cli\u003eBaden, L.R.\u003cem\u003e et al.\u003c/em\u003e Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine. \u003cem\u003eN Engl J Med\u003c/em\u003e \u003cstrong\u003e384\u003c/strong\u003e,403-416 (2021).\u003c/li\u003e\n\u003cli\u003ePolack, F.P.\u003cem\u003e et al.\u003c/em\u003e Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine. \u003cem\u003eN Engl J Med\u003c/em\u003e \u003cstrong\u003e383\u003c/strong\u003e,2603-2615 (2020).\u003c/li\u003e\n\u003cli\u003eMuecksch, F.\u003cem\u003e et al.\u003c/em\u003e Increased memory B cell potency and breadth after a SARS-CoV-2 mRNA boost. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e607\u003c/strong\u003e,128-134 (2022).\u003c/li\u003e\n\u003cli\u003eGoel, R.R.\u003cem\u003e et al.\u003c/em\u003e mRNA vaccines induce durable immune memory to SARS-CoV-2 and variants of concern. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e374\u003c/strong\u003e,abm0829 (2021).\u003c/li\u003e\n\u003cli\u003eZhang, Z.\u003cem\u003e et al.\u003c/em\u003e Humoral and cellular immune memory to four COVID-19 vaccines. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e,2434-2451.e2417 (2022).\u003c/li\u003e\n\u003cli\u003eChemaitelly, H.\u003cem\u003e et al.\u003c/em\u003e Protection from previous natural infection compared with mRNA vaccination against SARS-CoV-2 infection and severe COVID-19 in Qatar: a retrospective cohort study. \u003cem\u003eThe Lancet Microbe\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e,e944-e955 (2022).\u003c/li\u003e\n\u003cli\u003eHall, V.\u003cem\u003e et al.\u003c/em\u003e Protection against SARS-CoV-2 after Covid-19 Vaccination and Previous Infection. \u003cem\u003eNew England Journal of Medicine\u003c/em\u003e \u003cstrong\u003e386\u003c/strong\u003e,1207-1220 (2022).\u003c/li\u003e\n\u003cli\u003eStein, C.\u003cem\u003e et al.\u003c/em\u003e Past SARS-CoV-2 infection protection against re-infection: a systematic review and meta-analysis. \u003cem\u003eThe Lancet\u003c/em\u003e (2023).\u003c/li\u003e\n\u003cli\u003eZuo, F., Marcotte, H., Hammarstr\u0026ouml;m, L. \u0026amp; Pan-Hammarstr\u0026ouml;m, Q. Mucosal IgA against SARS-CoV-2 Omicron Infection. \u003cem\u003eN Engl J Med\u003c/em\u003e \u003cstrong\u003e387\u003c/strong\u003e,e55 (2022).\u003c/li\u003e\n\u003cli\u003eHuang, N.\u003cem\u003e et al.\u003c/em\u003e SARS-CoV-2 infection of the oral cavity and saliva. \u003cem\u003eNature Medicine\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e,892-903 (2021).\u003c/li\u003e\n\u003cli\u003eW\u0026ouml;lfel, R.\u003cem\u003e et al.\u003c/em\u003e Virological assessment of hospitalized patients with COVID-2019. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e581\u003c/strong\u003e,465-469 (2020).\u003c/li\u003e\n\u003cli\u003eHavervall, S.\u003cem\u003e et al.\u003c/em\u003e Anti-Spike Mucosal IgA Protection against SARS-CoV-2 Omicron Infection. \u003cem\u003eNew England Journal of Medicine\u003c/em\u003e \u003cstrong\u003e387\u003c/strong\u003e,1333-1336 (2022).\u003c/li\u003e\n\u003cli\u003eRavichandran, S.\u003cem\u003e et al.\u003c/em\u003e Systemic and mucosal immune profiling in asymptomatic and symptomatic SARS-CoV-2-infected individuals reveal unlinked immune signatures. \u003cem\u003eSci Adv\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e,eabi6533 (2021).\u003c/li\u003e\n\u003cli\u003eGaebler, C.\u003cem\u003e et al.\u003c/em\u003e Evolution of antibody immunity to SARS-CoV-2. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e591\u003c/strong\u003e,639-644 (2021).\u003c/li\u003e\n\u003cli\u003eSokal, A.\u003cem\u003e et al.\u003c/em\u003e Maturation and persistence of the anti-SARS-CoV-2 memory B cell response. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e184\u003c/strong\u003e,1201-1213.e1214 (2021).\u003c/li\u003e\n\u003cli\u003eAllie, S.R. \u0026amp; Randall, T.D. Resident Memory B Cells. \u003cem\u003eViral Immunology\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e,282-293 (2020).\u003c/li\u003e\n\u003cli\u003eLederer, K.\u003cem\u003e et al.\u003c/em\u003e Germinal center responses to SARS-CoV-2 mRNA vaccines in healthy and immunocompromised individuals. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e,1008-1024.e1015 (2022).\u003c/li\u003e\n\u003cli\u003eMudd, P.A.\u003cem\u003e et al.\u003c/em\u003e SARS-CoV-2 mRNA vaccination elicits a robust and persistent T follicular helper cell response in humans. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e,603-613.e615 (2022).\u003c/li\u003e\n\u003cli\u003eR\u0026ouml;ltgen, K.\u003cem\u003e et al.\u003c/em\u003e Immune imprinting, breadth of variant recognition, and germinal center response in human SARS-CoV-2 infection and vaccination. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e,1025-1040.e1014 (2022).\u003c/li\u003e\n\u003cli\u003eTurner, J.S.\u003cem\u003e et al.\u003c/em\u003e SARS-CoV-2 mRNA vaccines induce persistent human germinal centre responses. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e596\u003c/strong\u003e,109-113 (2021).\u003c/li\u003e\n\u003cli\u003eShehata, L.\u003cem\u003e et al.\u003c/em\u003e Systematic comparison of respiratory syncytial virus-induced memory B cell responses in two anatomical compartments. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e,1126 (2019).\u003c/li\u003e\n\u003cli\u003eTan, H.-X.\u003cem\u003e et al.\u003c/em\u003e Cutting Edge: SARS-CoV-2 Infection Induces Robust Germinal Center Activity in the Human Tonsil. \u003cem\u003eThe Journal of Immunology\u003c/em\u003e,ji2101199 (2022).\u003c/li\u003e\n\u003cli\u003eXu, Q.\u003cem\u003e et al.\u003c/em\u003e Adaptive immune responses to SARS-CoV-2 persist in the pharyngeal lymphoid tissue of children. \u003cem\u003eNature immunology\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e,186-199 (2023).\u003c/li\u003e\n\u003cli\u003eHachim, A.\u003cem\u003e et al.\u003c/em\u003e ORF8 and ORF3b antibodies are accurate serological markers of early and late SARS-CoV-2 infection. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e,1293-1301 (2020).\u003c/li\u003e\n\u003cli\u003eRavichandran, S.\u003cem\u003e et al.\u003c/em\u003e SARS-CoV-2 immune repertoire in MIS-C and pediatric COVID-19. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e,1452-1464 (2021).\u003c/li\u003e\n\u003cli\u003eLiu, C., Tsang, T.K., Sullivan, S.G., Cowling, B.J. \u0026amp; Yang, B. Comparative duration of neutralizing responses and protections of COVID-19 vaccination and correlates of protection. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e,4748 (2025).\u003c/li\u003e\n\u003cli\u003eAbu-Raddad, L.J.\u003cem\u003e et al.\u003c/em\u003e Effect of mRNA Vaccine Boosters against SARS-CoV-2 Omicron Infection in Qatar. \u003cem\u003eNew England Journal of Medicine\u003c/em\u003e \u003cstrong\u003e386\u003c/strong\u003e,1804-1816 (2022).\u003c/li\u003e\n\u003cli\u003eAltarawneh, H.N.\u003cem\u003e et al.\u003c/em\u003e Protection against the Omicron Variant from Previous SARS-CoV-2 Infection. \u003cem\u003eNew England Journal of Medicine\u003c/em\u003e \u003cstrong\u003e386\u003c/strong\u003e,1288-1290 (2022).\u003c/li\u003e\n\u003cli\u003eSun, P.\u003cem\u003e et al.\u003c/em\u003e Asymptomatic or symptomatic SARS-CoV-2 infection plus vaccination confers increased adaptive immunity to variants of concern. \u003cem\u003eiScience\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e,105202 (2022).\u003c/li\u003e\n\u003cli\u003eAzzi, L.\u003cem\u003e et al.\u003c/em\u003e Mucosal immune response in BNT162b2 COVID-19 vaccine recipients. \u003cem\u003eeBioMedicine\u003c/em\u003e \u003cstrong\u003e75\u003c/strong\u003e,103788 (2022).\u003c/li\u003e\n\u003cli\u003eTang, J.\u003cem\u003e et al.\u003c/em\u003e Respiratory mucosal immunity against SARS-CoV-2 after mRNA vaccination. \u003cem\u003eSci Immunol\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e,eadd4853 (2022).\u003c/li\u003e\n\u003cli\u003eWeisel, F. \u0026amp; Shlomchik, M. Memory B Cells of Mice and Humans. \u003cem\u003eAnnual Review of Immunology, Vol 35\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e,255-284 (2017).\u003c/li\u003e\n\u003cli\u003eSanz, I.\u003cem\u003e et al.\u003c/em\u003e Challenges and Opportunities for Consistent Classification of Human B Cell and Plasma Cell Populations. \u003cem\u003eFrontiers in Immunology\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e (2019).\u003c/li\u003e\n\u003cli\u003eKardava, L.\u003cem\u003e et al.\u003c/em\u003e Early human B cell signatures of the primary antibody response to mRNA vaccination. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e \u003cstrong\u003e119\u003c/strong\u003e,e2204607119 (2022).\u003c/li\u003e\n\u003cli\u003eAndrews, S.F.\u003cem\u003e et al.\u003c/em\u003e Activation Dynamics and Immunoglobulin Evolution of Pre-existing and Newly Generated Human Memory B cell Responses to Influenza Hemagglutinin. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e51\u003c/strong\u003e,398-410.e395 (2019).\u003c/li\u003e\n\u003cli\u003eNellore, A.\u003cem\u003e et al.\u003c/em\u003e A transcriptionally distinct subset of influenza-specific effector memory B cells predicts long-lived antibody responses to vaccination in humans. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e,847-863.e848 (2023).\u003c/li\u003e\n\u003cli\u003eSpangler, A.\u003cem\u003e et al.\u003c/em\u003e Early influenza virus exposure shapes the B cell response to influenza vaccination in individuals 50 years later. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e58\u003c/strong\u003e,728-744.e729 (2025).\u003c/li\u003e\n\u003cli\u003ePape, K.A.\u003cem\u003e et al.\u003c/em\u003e High-affinity memory B cells induced by SARS-CoV-2 infection produce more plasmablasts and atypical memory B cells than those primed by mRNA vaccines. \u003cem\u003eCell Rep\u003c/em\u003e \u003cstrong\u003e37\u003c/strong\u003e,109823 (2021).\u003c/li\u003e\n\u003cli\u003eJenks, S.A.\u003cem\u003e et al.\u003c/em\u003e Distinct Effector B Cells Induced by Unregulated Toll-like Receptor 7 Contribute to Pathogenic Responses in Systemic Lupus Erythematosus. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e,725-739.e726 (2018).\u003c/li\u003e\n\u003cli\u003eMoir , S.\u003cem\u003e et al.\u003c/em\u003e Evidence for HIV-associated B cell exhaustion in a dysfunctional memory B cell compartment in HIV-infected viremic individuals. \u003cem\u003eJournal of Experimental Medicine\u003c/em\u003e \u003cstrong\u003e205\u003c/strong\u003e,1797-1805 (2008).\u003c/li\u003e\n\u003cli\u003eIsnardi, I.\u003cem\u003e et al.\u003c/em\u003e Complement receptor 2/CD21\u0026minus; human naive B cells contain mostly autoreactive unresponsive clones. \u003cem\u003eBlood\u003c/em\u003e \u003cstrong\u003e115\u003c/strong\u003e,5026-5036 (2010).\u003c/li\u003e\n\u003cli\u003eSutton, H.J.\u003cem\u003e et al.\u003c/em\u003e Atypical B cells are part of an alternative lineage of B cells that participates in responses to vaccination and infection in humans. \u003cem\u003eCell Rep\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e,108684 (2021).\u003c/li\u003e\n\u003cli\u003eWeiss, G.E.\u003cem\u003e et al.\u003c/em\u003e Atypical Memory B Cells Are Greatly Expanded in Individuals Living in a Malaria-Endemic Area1. \u003cem\u003eThe Journal of Immunology\u003c/em\u003e \u003cstrong\u003e183\u003c/strong\u003e,2176-2182 (2009).\u003c/li\u003e\n\u003cli\u003eWoodruff, M.C.\u003cem\u003e et al.\u003c/em\u003e Extrafollicular B cell responses correlate with neutralizing antibodies and morbidity in COVID-19. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e,1506-1516 (2020).\u003c/li\u003e\n\u003cli\u003eAustin, J.W.\u003cem\u003e et al.\u003c/em\u003e Overexpression of T-bet in HIV infection is associated with accumulation of B cells outside germinal centers and poor affinity maturation. \u003cem\u003eSci Transl Med\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e (2019).\u003c/li\u003e\n\u003cli\u003eKnox, J.J.\u003cem\u003e et al.\u003c/em\u003e T-bet+ B cells are induced by human viral infections and dominate the HIV gp140 response. \u003cem\u003eJCI Insight\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e (2017).\u003c/li\u003e\n\u003cli\u003eLi, H., Borrego, F., Nagata, S. \u0026amp; Tolnay, M. Fc Receptor\u0026ndash;like 5 Expression Distinguishes Two Distinct Subsets of Human Circulating Tissue\u0026ndash;like Memory B Cells. \u003cem\u003eThe Journal of Immunology\u003c/em\u003e \u003cstrong\u003e196\u003c/strong\u003e,4064-4074 (2016).\u003c/li\u003e\n\u003cli\u003eKwon, D.-i.\u003cem\u003e et al.\u003c/em\u003e Mucosal unadjuvanted booster vaccines elicit local IgA responses by conversion of pre-existing immunity in mice. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e26\u003c/strong\u003e,908-919 (2025).\u003c/li\u003e\n\u003cli\u003eBurton, A.R.\u003cem\u003e et al.\u003c/em\u003e Circulating and intrahepatic antiviral B cells are defective in hepatitis B. \u003cem\u003eThe Journal of Clinical Investigation\u003c/em\u003e \u003cstrong\u003e128\u003c/strong\u003e,4588-4603 (2018).\u003c/li\u003e\n\u003cli\u003eMuehlinghaus, G.\u003cem\u003e et al.\u003c/em\u003e Regulation of CXCR3 and CXCR4 expression during terminal differentiation of memory B cells into plasma cells. \u003cem\u003eBlood\u003c/em\u003e \u003cstrong\u003e105\u003c/strong\u003e,3965-3971 (2005).\u003c/li\u003e\n\u003cli\u003eDavis-Porada, J.\u003cem\u003e et al.\u003c/em\u003e Maintenance and functional regulation of immune memory to COVID-19 vaccines in tissues. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e,2895-2913.e2898 (2024).\u003c/li\u003e\n\u003cli\u003eOh, J.E.\u003cem\u003e et al.\u003c/em\u003e Migrant memory B cells secrete luminal antibody in the vagina. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e571\u003c/strong\u003e,122-126 (2019).\u003c/li\u003e\n\u003cli\u003eOh, J.E.\u003cem\u003e et al.\u003c/em\u003e Intranasal priming induces local lung-resident B cell populations that secrete protective mucosal antiviral IgA. \u003cem\u003eScience Immunology\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e,eabj5129 (2021).\u003c/li\u003e\n\u003cli\u003eArroyo-D\u0026iacute;az, N.M.\u003cem\u003e et al.\u003c/em\u003e Interferon-\u0026gamma; production by Tfh cells is required for CXCR3(+) pre-memory B cell differentiation and subsequent lung-resident memory B cell responses. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e,2358-2372.e2355 (2023).\u003c/li\u003e\n\u003cli\u003eJohnson, J.L.\u003cem\u003e et al.\u003c/em\u003e The Transcription Factor T-bet Resolves Memory B Cell Subsets with Distinct Tissue Distributions and Antibody Specificities in Mice and Humans. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e,842-855.e846 (2020).\u003c/li\u003e\n\u003cli\u003eLee, C.M. \u0026amp; Oh, J.E. Resident Memory B Cells in Barrier Tissues. \u003cem\u003eFront Immunol\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e,953088 (2022).\u003c/li\u003e\n\u003cli\u003eLau, D.\u003cem\u003e et al.\u003c/em\u003e Low CD21 expression defines a population of recent germinal center graduates primed for plasma cell differentiation. \u003cem\u003eSci Immunol\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e (2017).\u003c/li\u003e\n\u003cli\u003eZhang, H., Meng, F., Chu, C.-L., Takai, T. \u0026amp; Lowell, C.A. The Src Family Kinases Hck and Fgr Negatively Regulate Neutrophil and Dendritic Cell Chemokine Signaling via PIR-B. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e,235-246 (2005).\u003c/li\u003e\n\u003cli\u003eGao, X.\u003cem\u003e et al.\u003c/em\u003e Zeb2 drives the formation of CD11c+ atypical B cells to sustain germinal centers that control persistent infection. \u003cem\u003eScience Immunology\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e,eadj4748 (2024).\u003c/li\u003e\n\u003cli\u003eZurbuchen, Y.\u003cem\u003e et al.\u003c/em\u003e Human memory B cells show plasticity and adopt multiple fates upon recall response to SARS-CoV-2. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e,955-965 (2023).\u003c/li\u003e\n\u003cli\u003eShao, W., Wang, Y., Fang, Q., Shi, W. \u0026amp; Qi, H. Epigenetic recording of stimulation history reveals BLIMP1\u0026ndash;BACH2 balance in determining memory B cell fate upon recall challenge. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e,1432-1444 (2024).\u003c/li\u003e\n\u003cli\u003eAlaterre, E.\u003cem\u003e et al.\u003c/em\u003e Integrative single-cell chromatin and transcriptome analysis of human plasma cell differentiation. \u003cem\u003eBlood\u003c/em\u003e \u003cstrong\u003e144\u003c/strong\u003e,496-509 (2024).\u003c/li\u003e\n\u003cli\u003eZotos, D.\u003cem\u003e et al.\u003c/em\u003e IL-21 regulates germinal center B cell differentiation and proliferation through a B cell-intrinsic mechanism. \u003cem\u003eJ Exp Med\u003c/em\u003e \u003cstrong\u003e207\u003c/strong\u003e,365-378 (2010).\u003c/li\u003e\n\u003cli\u003eLinterman, M.A.\u003cem\u003e et al.\u003c/em\u003e IL-21 acts directly on B cells to regulate Bcl-6 expression and germinal center responses. \u003cem\u003eJ Exp Med\u003c/em\u003e \u003cstrong\u003e207\u003c/strong\u003e,353-363 (2010).\u003c/li\u003e\n\u003cli\u003eKim, Y.\u003cem\u003e et al.\u003c/em\u003e IL-21 shapes the B cell response in a context-dependent manner. \u003cem\u003eCell Rep\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e (2025).\u003c/li\u003e\n\u003cli\u003eDai, D.\u003cem\u003e et al.\u003c/em\u003e The transcription factor ZEB2 drives the formation of age-associated B cells. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e383\u003c/strong\u003e,413-421 (2024).\u003c/li\u003e\n\u003cli\u003eOchiai, K.\u003cem\u003e et al.\u003c/em\u003e Transcriptional regulation of germinal center B and plasma cell fates by dynamical control of IRF4. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e,918-929 (2013).\u003c/li\u003e\n\u003cli\u003eXu, H.\u003cem\u003e et al.\u003c/em\u003e Regulation of bifurcating B cell trajectories by mutual antagonism between transcription factors IRF4 and IRF8. \u003cem\u003eNat Immunol\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e,1274-1281 (2015).\u003c/li\u003e\n\u003cli\u003eMa, J.\u003cem\u003e et al.\u003c/em\u003e A blueprint for tumor-infiltrating B cells across human cancers. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e384\u003c/strong\u003e,eadj4857 (2024).\u003c/li\u003e\n\u003cli\u003eZhang, L.\u003cem\u003e et al.\u003c/em\u003e Lineage tracking reveals dynamic relationships of T cells in colorectal cancer. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e564\u003c/strong\u003e,268-272 (2018).\u003c/li\u003e\n\u003cli\u003eBuckner, C.M.\u003cem\u003e et al.\u003c/em\u003e Interval between prior SARS-CoV-2 infection and booster vaccination impacts magnitude and quality of antibody and B cell responses. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e185\u003c/strong\u003e,4333-4346.e4314 (2022).\u003c/li\u003e\n\u003cli\u003eAmbegaonkar, A.A.\u003cem\u003e et al.\u003c/em\u003e Expression of inhibitory receptors by B cells in chronic human infectious diseases restricts responses to membrane-associated antigens. \u003cem\u003eScience Advances\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e,eaba6493 (2020).\u003c/li\u003e\n\u003cli\u003ePortugal, S.\u003cem\u003e et al.\u003c/em\u003e Malaria-associated atypical memory B cells exhibit markedly reduced B cell receptor signaling and effector function. \u003cem\u003eeLife\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e,e07218 (2015).\u003c/li\u003e\n\u003cli\u003eHalliley, Jessica L.\u003cem\u003e et al.\u003c/em\u003e Long-Lived Plasma Cells Are Contained within the CD19\u0026minus;CD38hiCD138+ Subset in Human Bone Marrow. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e43\u003c/strong\u003e,132-145 (2015).\u003c/li\u003e\n\u003cli\u003ePinto, D.\u003cem\u003e et al.\u003c/em\u003e A functional BCR in human IgA and IgM plasma cells. \u003cem\u003eBlood\u003c/em\u003e \u003cstrong\u003e121\u003c/strong\u003e,4110-4114 (2013).\u003c/li\u003e\n\u003cli\u003ePhetsouphanh, C.\u003cem\u003e et al.\u003c/em\u003e Immunological dysfunction persists for 8 months following initial mild-to-moderate SARS-CoV-2 infection. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e,210-216 (2022).\u003c/li\u003e\n\u003cli\u003eCheong, J.-G.\u003cem\u003e et al.\u003c/em\u003e Epigenetic memory of coronavirus infection in innate immune cells and their progenitors. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e186\u003c/strong\u003e,3882-3902.e3824 (2023).\u003c/li\u003e\n\u003cli\u003ePenna, G., Sozzani, S. \u0026amp; Adorini, L. Cutting edge: selective usage of chemokine receptors by plasmacytoid dendritic cells. \u003cem\u003eThe Journal of Immunology\u003c/em\u003e \u003cstrong\u003e167\u003c/strong\u003e,1862-1866 (2001).\u003c/li\u003e\n\u003cli\u003eSchrom, E.C.\u003cem\u003e et al.\u003c/em\u003e Spatial Patterning Analysis of Cellular Ensembles (SPACE) finds complex spatial organization at the cell and tissue levels. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e \u003cstrong\u003e122\u003c/strong\u003e,e2412146122 (2025).\u003c/li\u003e\n\u003cli\u003eIijima, N. \u0026amp; Iwasaki, A. T cell memory. A local macrophage chemokine network sustains protective tissue-resident memory CD4 T cells. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e346\u003c/strong\u003e,93-98 (2014).\u003c/li\u003e\n\u003cli\u003eGiesecke, C.\u003cem\u003e et al.\u003c/em\u003e Tissue Distribution and Dependence of Responsiveness of Human Antigen-Specific Memory B Cells. \u003cem\u003eThe Journal of Immunology\u003c/em\u003e \u003cstrong\u003e192\u003c/strong\u003e,3091-3100 (2014).\u003c/li\u003e\n\u003cli\u003eJoo, H.M., He, Y., Sundararajan, A., Huan, L. \u0026amp; Sangster, M.Y. Quantitative analysis of influenza virus-specific B cell memory generated by different routes of inactivated virus vaccination. \u003cem\u003eVaccine\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e,2186-2194 (2010).\u003c/li\u003e\n\u003cli\u003eKim, W.\u003cem\u003e et al.\u003c/em\u003e Germinal centre-driven maturation of B cell response to mRNA vaccination. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e604\u003c/strong\u003e,141-145 (2022).\u003c/li\u003e\n\u003cli\u003eKillingley, B.\u003cem\u003e et al.\u003c/em\u003e Safety, tolerability and viral kinetics during SARS-CoV-2 human challenge in young adults. \u003cem\u003eNature Medicine\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e,1031-1041 (2022).\u003c/li\u003e\n\u003cli\u003eSchaefer-Babajew, D.\u003cem\u003e et al.\u003c/em\u003e Antibody feedback regulates immune memory after SARS-CoV-2 mRNA vaccination. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e613\u003c/strong\u003e,735-742 (2023).\u003c/li\u003e\n\u003cli\u003eWachter, B.T.\u003cem\u003e et al.\u003c/em\u003e Prior SARS-CoV-2 infection affects adaptive immune responses to Omicron BA.4/BA.5 mRNA booster. \u003cem\u003eJournal of Allergy and Clinical Immunology\u003c/em\u003e (2025).\u003c/li\u003e\n\u003cli\u003eBettini, E.\u003cem\u003e et al.\u003c/em\u003e Distinct components of nucleoside-modified messenger RNA vaccines cooperate to instruct efficient germinal center responses. \u003cem\u003ebioRxiv\u003c/em\u003e (2024).\u003c/li\u003e\n\u003cli\u003eKiyono, H. \u0026amp; Ernst, P.B. Nasal vaccines for respiratory infections. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e641\u003c/strong\u003e,321-330 (2025).\u003c/li\u003e\n\u003cli\u003eAlu, A.\u003cem\u003e et al.\u003c/em\u003e Intranasal COVID-19 vaccines: From bench to bed. \u003cem\u003eeBioMedicine\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e (2022).\u003c/li\u003e\n\u003cli\u003eQi, H., Liu, B., Wang, X. \u0026amp; Zhang, L. The humoral response and antibodies against SARS-CoV-2 infection. \u003cem\u003eNature Immunology\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e,1008-1020 (2022).\u003c/li\u003e\n\u003cli\u003eMassoni-Badosa, R.\u003cem\u003e et al.\u003c/em\u003e An atlas of cells in the human tonsil. \u003cem\u003eImmunity\u003c/em\u003e \u003cstrong\u003e57\u003c/strong\u003e,379-399 e318 (2024).\u003c/li\u003e\n\u003cli\u003eXu, Q.\u003cem\u003e et al.\u003c/em\u003e \u003cem\u003eEfficient Tonsillar T Follicular Helper Cell Processing and Functional Analysis through High-dimensional Flow Cytometry\u003c/em\u003e, vol. 221. 1940-087X, 2025.\u003c/li\u003e\n\u003cli\u003eTang, J.J.\u003cem\u003e et al.\u003c/em\u003e Cross-reactive immunity against the SARS-CoV-2 Omicron variant is low in pediatric patients with prior COVID-19 or MIS-C (vol 13, pg 2979, 2022). \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e (2022).\u003c/li\u003e\n\u003cli\u003eNeerukonda, S.N.\u003cem\u003e et al.\u003c/em\u003e Establishment of a well-characterized SARS-CoV-2 lentiviral pseudovirus neutralization assay using 293T cells with stable expression of ACE2 and TMPRSS2. \u003cem\u003ePlos One\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e (2021).\u003c/li\u003e\n\u003cli\u003eRavichandran, S.\u003cem\u003e et al.\u003c/em\u003e Antibody signature induced by SARS-CoV-2 spike protein immunogens in rabbits. \u003cem\u003eSci Transl Med\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e (2020).\u003c/li\u003e\n\u003cli\u003eTang, J.J.\u003cem\u003e et al.\u003c/em\u003e Epitope diversity of SARS-CoV-2 hyperimmune intravenous human immunoglobulins and neutralization of variants of concern. \u003cem\u003eIscience\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e (2021).\u003c/li\u003e\n\u003cli\u003eAwasthi, M., Golding, H. \u0026amp; Khurana, S. Severe Acute Respiratory Syndrome Coronavirus 2 Hyperimmune Intravenous Human Immunoglobulins Neutralizes Omicron Subvariants BA.1, BA.2, BA.2.12.1, BA.3, and BA.4/BA.5 for Treatment of Coronavirus Disease 2019. \u003cem\u003eClin Infect Dis\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e,E503-E506 (2023).\u003c/li\u003e\n\u003cli\u003eHao, Y.H.\u003cem\u003e et al.\u003c/em\u003e Dictionary learning for integrative, multimodal and scalable single-cell analysis. \u003cem\u003eNature Biotechnology\u003c/em\u003e \u003cstrong\u003e42\u003c/strong\u003e (2024).\u003c/li\u003e\n\u003cli\u003eHao, Y.\u003cem\u003e et al.\u003c/em\u003e Integrated analysis of multimodal single-cell data. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e184\u003c/strong\u003e,3573-3587 e3529 (2021).\u003c/li\u003e\n\u003cli\u003eWickham, H. ggplot2 Elegant Graphics for Data Analysis Introduction. \u003cem\u003eUse R\u003c/em\u003e,1-+ (2009).\u003c/li\u003e\n\u003cli\u003eZhang, Z. Reshaping and aggregating data: an introduction to reshape package. \u003cem\u003eAnn Transl Med\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e,78 (2016).\u003c/li\u003e\n\u003cli\u003eKassambara, A. ggpubr: \u0026apos;ggplot2\u0026apos; Based Publication Ready Plots. 2025.\u003c/li\u003e\n\u003cli\u003eArnold, J.B. ggthemes: Extra Themes, Scales and Geoms for \u0026apos;ggplot2\u0026apos;. 2024.\u003c/li\u003e\n\u003cli\u003eDan, J.M.\u003cem\u003e et al.\u003c/em\u003e Recurrent group A Streptococcus tonsillitis is an immunosusceptibility disease involving antibody deficiency and aberrant T(FH) cells. \u003cem\u003eSci Transl Med\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e (2019).\u003c/li\u003e\n\u003cli\u003eStoeckius, M.\u003cem\u003e et al.\u003c/em\u003e Cell Hashing with barcoded antibodies enables multiplexing and doublet detection for single cell genomics. \u003cem\u003eGenome Biol\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e,224 (2018).\u003c/li\u003e\n\u003cli\u003eDanecek, P.\u003cem\u003e et al.\u003c/em\u003e Twelve years of SAMtools and BCFtools. \u003cem\u003eGigascience\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e (2021).\u003c/li\u003e\n\u003cli\u003eRogozhnikov, A.\u003cem\u003e et al.\u003c/em\u003e Demuxalot: scaled up genetic demultiplexing for single-cell sequencing. \u003cem\u003ebioRxiv\u003c/em\u003e,2021.2005. 2022.443646 (2021).\u003c/li\u003e\n\u003cli\u003eKang, H.M.\u003cem\u003e et al.\u003c/em\u003e Multiplexed droplet single-cell RNA-sequencing using natural genetic variation (vol 36, pg 89, 2018). \u003cem\u003eNature Biotechnology\u003c/em\u003e \u003cstrong\u003e38\u003c/strong\u003e,1356-1356 (2020).\u003c/li\u003e\n\u003cli\u003eZheng, G.X.\u003cem\u003e et al.\u003c/em\u003e Massively parallel digital transcriptional profiling of single cells. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e,14049 (2017).\u003c/li\u003e\n\u003cli\u003eMule, M.P., Martins, A.J. \u0026amp; Tsang, J.S. Normalizing and denoising protein expression data from droplet-based single cell profiling. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e,2099 (2022).\u003c/li\u003e\n\u003cli\u003eGu, Z., Eils, R. \u0026amp; Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e32\u003c/strong\u003e,2847-2849 (2016).\u003c/li\u003e\n\u003cli\u003eRitchie, M.E.\u003cem\u003e et al.\u003c/em\u003e limma powers differential expression analyses for RNA-sequencing and microarray studies. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e43\u003c/strong\u003e,e47 (2015).\u003c/li\u003e\n\u003cli\u003eFinak, G., Jiang, W. \u0026amp; Gottardo, R. CytoML for cross-platform cytometry data sharing. \u003cem\u003eCytometry A\u003c/em\u003e \u003cstrong\u003e93\u003c/strong\u003e,1189-1196 (2018).\u003c/li\u003e\n\u003cli\u003eFinak G, J.M. flowWorkspace: Infrastructure for representing and interacting with gated and ungated cytometry data sets.; 2025.\u003c/li\u003e\n\u003cli\u003eR, K. pheatmap: Pretty Heatmaps. . 2025.\u003c/li\u003e\n\u003cli\u003eLiu, C.\u003cem\u003e et al.\u003c/em\u003e Time-resolved systems immunology reveals a late juncture linked to fatal COVID-19. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e184\u003c/strong\u003e,1836-1857 e1822 (2021).\u003c/li\u003e\n\u003cli\u003eSacco, K.\u003cem\u003e et al.\u003c/em\u003e Immunopathological signatures in multisystem inflammatory syndrome in children and pediatric COVID-19. \u003cem\u003eNat Med\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e,1050-1062 (2022).\u003c/li\u003e\n\u003cli\u003eMcCarthy, D.J., Chen, Y. \u0026amp; Smyth, G.K. Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e40\u003c/strong\u003e,4288-4297 (2012).\u003c/li\u003e\n\u003cli\u003eLaw, C.W., Chen, Y.S., Shi, W. \u0026amp; Smyth, G.K. voom: precision weights unlock linear model analysis tools for RNA-seq read counts. \u003cem\u003eGenome Biology\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e (2014).\u003c/li\u003e\n\u003cli\u003eTan, H.X.\u003cem\u003e et al.\u003c/em\u003e Lung-resident memory B cells established after pulmonary influenza infection display distinct transcriptional and phenotypic profiles. \u003cem\u003eSci Immunol\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e,eabf5314 (2022).\u003c/li\u003e\n\u003cli\u003eYang, Y.\u003cem\u003e et al.\u003c/em\u003e Pan-cancer single-cell dissection reveals phenotypically distinct B cell subtypes. \u003cem\u003eCell\u003c/em\u003e \u003cstrong\u003e187\u003c/strong\u003e,4790-4811 e4722 (2024).\u003c/li\u003e\n\u003cli\u003eHolla, P.\u003cem\u003e et al.\u003c/em\u003e Shared transcriptional profiles of atypical B cells suggest common drivers of expansion and function in malaria, HIV, and autoimmunity. \u003cem\u003eSci Adv\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e (2021).\u003c/li\u003e\n\u003cli\u003eMoir, S.\u003cem\u003e et al.\u003c/em\u003e Evidence for HIV-associated B cell exhaustion in a dysfunctional memory B cell compartment in HIV-infected viremic individuals. \u003cem\u003eJ Exp Med\u003c/em\u003e \u003cstrong\u003e205\u003c/strong\u003e,1797-1805 (2008).\u003c/li\u003e\n\u003cli\u003eWu, C.\u003cem\u003e et al.\u003c/em\u003e Lupus-associated atypical memory B cells are mTORC1-hyperactivated and functionally dysregulated. \u003cem\u003eAnn Rheum Dis\u003c/em\u003e \u003cstrong\u003e78\u003c/strong\u003e,1090-1100 (2019).\u003c/li\u003e\n\u003cli\u003eVerstappen, G.M.\u003cem\u003e et al.\u003c/em\u003e Gene expression profiling of epithelium-associated FcRL4(+) B cells in primary Sjogren\u0026apos;s syndrome reveals a pathogenic signature. \u003cem\u003eJ Autoimmun\u003c/em\u003e \u003cstrong\u003e109\u003c/strong\u003e,102439 (2020).\u003c/li\u003e\n\u003cli\u003eIsnardi, I.\u003cem\u003e et al.\u003c/em\u003e Complement receptor 2/CD21 human naive B cells contain mostly autoreactive unresponsive clones. \u003cem\u003eBlood\u003c/em\u003e \u003cstrong\u003e115\u003c/strong\u003e,5026-5036 (2010).\u003c/li\u003e\n\u003cli\u003eAttaf, N.\u003cem\u003e et al.\u003c/em\u003e FB5P-seq: FACS-Based 5-Prime End Single-Cell RNA-seq for Integrative Analysis of Transcriptome and Antigen Receptor Repertoire in B and T Cells (vol 11, 216, 2020). \u003cem\u003eFrontiers in Immunology\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e (2020).\u003c/li\u003e\n\u003cli\u003eSubramanian, A.\u003cem\u003e et al.\u003c/em\u003e Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e102\u003c/strong\u003e,15545-15550 (2005).\u003c/li\u003e\n\u003cli\u003eStreet, K.\u003cem\u003e et al.\u003c/em\u003e Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. \u003cem\u003eBMC Genomics\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e,477 (2018).\u003c/li\u003e\n\u003cli\u003eYe, J., Ma, N., Madden, T.L. \u0026amp; Ostell, J.M. IgBLAST: an immunoglobulin variable domain sequence analysis tool. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e41\u003c/strong\u003e,W34-40 (2013).\u003c/li\u003e\n\u003cli\u003eGupta, N.T.\u003cem\u003e et al.\u003c/em\u003e Change-O: a toolkit for analyzing large-scale B cell immunoglobulin repertoire sequencing data. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e31\u003c/strong\u003e,3356-3358 (2015).\u003c/li\u003e\n\u003cli\u003eNouri, N. \u0026amp; Kleinstein, S.H. A spectral clustering-based method for identifying clones from high-throughput B cell repertoire sequencing data. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e34\u003c/strong\u003e,i341-i349 (2018).\u003c/li\u003e\n\u003cli\u003eGupta, N.T.\u003cem\u003e et al.\u003c/em\u003e Hierarchical Clustering Can Identify B Cell Clones with High Confidence in Ig Repertoire Sequencing Data. \u003cem\u003eJ Immunol\u003c/em\u003e \u003cstrong\u003e198\u003c/strong\u003e,2489-2499 (2017).\u003c/li\u003e\n\u003cli\u003eYaari, G.\u003cem\u003e et al.\u003c/em\u003e Models of Somatic Hypermutation Targeting and Substitution Based on Synonymous Mutations from High-Throughput Immunoglobulin Sequencing Data. \u003cem\u003eFrontiers in Immunology\u003c/em\u003e \u003cstrong\u003e0\u003c/strong\u003e (2013).\u003c/li\u003e\n\u003cli\u003eHoehn, K.B., Pybus, O.G. \u0026amp; Kleinstein, S.H. Phylogenetic analysis of migration, differentiation, and class switching in B cells. \u003cem\u003ePlos Computational Biology\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e (2022).\u003c/li\u003e\n\u003cli\u003eHoehn, K.B.\u003cem\u003e et al.\u003c/em\u003e Repertoire-wide phylogenetic models of B cell molecular evolution reveal evolutionary signatures of aging and vaccination. \u003cem\u003eProceedings of the National Academy of Sciences\u003c/em\u003e \u003cstrong\u003e116\u003c/strong\u003e,22664-22672 (2019).\u003c/li\u003e\n\u003cli\u003eNielsen, R. \u0026amp; Yang, Z. Likelihood models for detecting positively selected amino acid sites and applications to the HIV-1 envelope gene. \u003cem\u003eGenetics\u003c/em\u003e \u003cstrong\u003e148\u003c/strong\u003e,929-936 (1998).\u003c/li\u003e\n\u003cli\u003eggtree: an r package for visualization and annotation of phylogenetic trees with their covariates and other associated data - Yu - 2017 - Methods in Ecology and Evolution - Wiley Online Library. (2021).\u003c/li\u003e\n\u003cli\u003eCorces, M.R.\u003cem\u003e et al.\u003c/em\u003e Lineage-specific and single-cell chromatin accessibility charts human hematopoiesis and leukemia evolution. \u003cem\u003eNat Genet\u003c/em\u003e \u003cstrong\u003e48\u003c/strong\u003e,1193-1203 (2016).\u003c/li\u003e\n\u003cli\u003eJang, M.K.\u003cem\u003e et al.\u003c/em\u003e Cell-free chromatin immunoprecipitation to detect molecular pathways in heart transplantation. \u003cem\u003eLife Sci Alliance\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e (2023).\u003c/li\u003e\n\u003cli\u003eMartin, M. Cutadpt Removes Adapter Sequences from High-Throughput Sequencing Reads. . \u003cem\u003eEMBnet \u003c/em\u003e\u003cstrong\u003e17\u003c/strong\u003e,10-12 (2011).\u003c/li\u003e\n\u003cli\u003eAmemiya, H.M., Kundaje, A. \u0026amp; Boyle, A.P. The ENCODE Blacklist: Identification of Problematic Regions of the Genome. \u003cem\u003eSci Rep\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e,9354 (2019).\u003c/li\u003e\n\u003cli\u003eLi, H. \u0026amp; Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler transform. \u003cem\u003eBioinformatics\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e,1754-1760 (2009).\u003c/li\u003e\n\u003cli\u003eRamirez, F.\u003cem\u003e et al.\u003c/em\u003e deepTools2: a next generation web server for deep-sequencing data analysis. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e44\u003c/strong\u003e,W160-165 (2016).\u003c/li\u003e\n\u003cli\u003eZhang, Y.\u003cem\u003e et al.\u003c/em\u003e Model-based Analysis of ChIP-Seq (MACS). \u003cem\u003eGenome Biology\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e (2008).\u003c/li\u003e\n\u003cli\u003eRoss-Innes, C.S.\u003cem\u003e et al.\u003c/em\u003e Differential oestrogen receptor binding is associated with clinical outcome in breast cancer. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e481\u003c/strong\u003e,389-U177 (2012).\u003c/li\u003e\n\u003cli\u003eBailey, T.L., Johnson, J., Grant, C.E. \u0026amp; Noble, W.S. The MEME Suite. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e43\u003c/strong\u003e,W39-49 (2015).\u003c/li\u003e\n\u003cli\u003eKulakovskiy, I.V.\u003cem\u003e et al.\u003c/em\u003e HOCOMOCO: towards a complete collection of transcription factor binding models for human and mouse via large-scale ChIP-Seq analysis. \u003cem\u003eNucleic Acids Res\u003c/em\u003e \u003cstrong\u003e46\u003c/strong\u003e,D252-D259 (2018).\u003c/li\u003e\n\u003cli\u003eTanigawa, Y., Dyer, E.S. \u0026amp; Bejerano, G. WhichTF is functionally important in your open chromatin data? \u003cem\u003ePLOS Computational Biology\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e,e1010378 (2022).\u003c/li\u003e\n\u003cli\u003eMcLean, C.Y.\u003cem\u003e et al.\u003c/em\u003e GREAT improves functional interpretation of cis-regulatory regions. \u003cem\u003eNat Biotechnol\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e,495-501 (2010).\u003c/li\u003e\n\u003cli\u003eGlasmacher, E.\u003cem\u003e et al.\u003c/em\u003e A Genomic Regulatory Element That Directs Assembly and Function of Immune-Specific AP-1-IRF Complexes. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e338\u003c/strong\u003e,975-980 (2012).\u003c/li\u003e\n\u003cli\u003eStuart, T., Srivastava, A., Madad, S., Lareau, C.A. \u0026amp; Satija, R. Single-cell chromatin state analysis with Signac (vol 18, pg 1333, 2021). \u003cem\u003eNat Methods\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e,257-257 (2022).\u003c/li\u003e\n\u003cli\u003eKardava, L.\u003cem\u003e et al.\u003c/em\u003e IgG3 regulates tissue-like memory B cells in HIV-infected individuals. \u003cem\u003eNat Immunol\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e,1001-1012 (2018).\u003c/li\u003e\n\u003cli\u003eAmbegaonkar, A.A., Nagata, S., Pierce, S.K. \u0026amp; Sohn, H. The Differentiation of Human Tonsil B Cells With the Phenotypic and Functional Characteristics of T-bet plus Atypical Memory B Cells in Malaria. \u003cem\u003eFrontiers in Immunology\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e (2019).\u003c/li\u003e\n\u003cli\u003eRose, J.R. Xen_Seurat_Pipeline. GitHub.https://github.com/jrose835/Xen_Seurat_Pipeline. 2025.\u003c/li\u003e\n\u003cli\u003eKassambara, A. ggpubr: \u0026apos;ggplot2\u0026apos; Based Publication Ready Plots version 0.4.0.from GitHub. (2020).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7428491/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7428491/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMucosal immunity is an important correlate of protection against respiratory infections such as SARS-CoV-2. Comparing B cell responses in the upper respiratory tract following vaccination and infection may offer unique insights into mucosal immunity. Here, we characterized antigen-specific B cells in the tonsils, adenoids, and peripheral blood of children who had been infected with SARS-CoV-2 or vaccinated with SARS-CoV-2 mRNA vaccines. SARS-CoV-2-specific switched memory B cells (B\u003csub\u003eSM\u003c/sub\u003e) and germinal center B cells were found in the blood and pharyngeal lymphoid tissues after vaccination or infection. However, infection generated a higher proportion of IgA\u003csup\u003e+ \u003c/sup\u003eB\u003csub\u003eSM\u003c/sub\u003e and CXCR3\u003csup\u003e+\u003c/sup\u003eCD21\u003csup\u003e+ \u003c/sup\u003eB\u003csub\u003eSM\u003c/sub\u003e, which showed distinct spatial localization, greater clonal expansion and increased propensity for plasma cell differentiation compared to their CXCR3\u003csup\u003e-\u003c/sup\u003e counterparts, accompanied by persistent activation of innate and T follicular helper cells in the tissues. Our data provide evidence for tissue-specific B cell memory after either SARS-CoV-2 vaccination or infection, but with distinct characteristics that can influence the quality, durability, and localization of immunity.\u003c/p\u003e","manuscriptTitle":"SARS-CoV-2 infection and vaccination elicit distinct pharyngeal mucosal B cell responses in children","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-09 10:00:47","doi":"10.21203/rs.3.rs-7428491/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c723b538-874d-48f9-9e43-2bfca0b84487","owner":[],"postedDate":"September 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":54086513,"name":"Biological sciences/Immunology/Mucosal immunology"},{"id":54086514,"name":"Biological sciences/Immunology/Translational immunology"},{"id":54086515,"name":"Biological sciences/Immunology/Lymphoid tissues/Tonsils"},{"id":54086516,"name":"Biological sciences/Immunology/Adaptive immunity/Humoral immunity/Immunological memory"}],"tags":[],"updatedAt":"2025-11-12T14:21:56+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-09 10:00:47","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7428491","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7428491","identity":"rs-7428491","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}