key: cord-0297643-cshgczss
authors: Seenappa, Lochana M.; Jakubowski, Aniela; Steinbuck, Martin P.; Palmer, Erica; Haqq, Christopher M.; Carter, Crystal; Fontenot, Jane; Villinger, Francois; McNeil, Lisa K.; DeMuth, Peter C.
title: Programming the lymph node immune response with Amphiphile-CpG induces potent cellular and humoral immunity following COVID-19 subunit vaccination in mice and non-human primates
date: 2022-05-19
journal: bioRxiv
DOI: 10.1101/2022.05.19.492649
sha: 13ea95bc4eeed4fe39ee14dd06d0c4524d66da94
doc_id: 297643
cord_uid: cshgczss

Despite the success of currently authorized vaccines for the reduction of severe COVID-19 disease risk, rapidly emerging viral variants continue to drive pandemic waves of infection, resulting in numerous global public health challenges. Progress will depend on future advances in prophylactic vaccine activity, including advancement of candidates capable of generating more potent induction of cross-reactive T cells and durable cross-reactive antibody responses. Here we evaluated an Amphiphile (AMP) adjuvant, AMP-CpG, admixed with SARS-CoV-2 Spike receptor binding domain (RBD) immunogen, as a lymph node-targeted protein subunit vaccine (ELI-005) in mice and non-human primates (NHPs). AMP-mediated targeting of CpG DNA to draining lymph nodes resulted in comprehensive local immune activation characterized by extensive transcriptional reprogramming, inflammatory proteomic milieu, and activation of innate immune cells as key orchestrators of antigen-directed adaptive immunity. Prime-boost immunization with AMP-CpG in mice induced potent and durable T cell responses in multiple anatomical sites critical for prophylactic efficacy and prevention of severe disease. Long-lived memory responses were rapidly expanded upon re-exposure to antigen. In parallel, RBD-specific antibodies were long-lived, and exhibited cross-reactive recognition of variant RBD. AMP-CpG-adjuvanted prime-boost immunization in NHPs was safe and well tolerated, while promoting multi-cytokine-producing circulating T cell responses cross-reactive across variants of concern (VOC). Expansion of RBD-specific germinal center (GC) B cells in lymph nodes correlated to rapid seroconversion with variant-specific neutralizing antibody responses exceeding those measured in convalescent human plasma. These results demonstrate the promise of lymph-node adjuvant-targeting to coordinate innate immunity and generate robust adaptive responses critical for vaccine efficacy.

2 convalescent human plasma. These results demonstrate the promise of lymph-node adjuvant-targeting 28 to coordinate innate immunity and generate robust adaptive responses critical for vaccine efficacy.

The evolving pandemic of coronavirus disease 2019 has resulted in continued global health 31 challenges with worldwide waves of infection driving significant morbidity and mortality despite 32 widespread use of effective vaccines 1 . The serial emergence of viral variants of concern (VOC) with 33 increased viral infectivity 2-4 , and mutations providing partial escape from neutralizing antibodies 5-8 has 34 proven a substantial challenge to successful control of viral spread. While currently authorized vaccines 35 continue to provide valuable protection against hospitalization and severe disease, high community viral 36 transmission has been observed despite high vaccination coverage 9,10 Notably, prevention of severe 37 disease appears to persist despite evidence of potentially rapid decay in vaccine-induced neutralizing 38 antibody titers, and observations of reduced activity against recent VOC 9,11-14 . Therefore, while 39 neutralizing antibody titers are considered the major correlate of protection for authorized vaccines, 15, 16 40 effective long-term protection against the worst outcomes of disease may be dependent on robust 41 memory B and cross-reactive T cell responses capable of rapid re-activation to contain nascent viral 42 infection in the absence of sterilizing immunity. Vaccines promoting responses with these 43 immunological features may provide improved patient outcomes while also preventing the emergence 44 of future coronavirus pathogens resulting from increasing exposure to natural zoonotic reservoirs.

We previously reported the development of ELI-005, a protein subunit vaccine combining the Spike 46 receptor-binding-domain (RBD) protein with a lymph node targeting TLR-9 agonist, AMP-CpG, consisting 47 of a diacyl lipid conjugated to CpG DNA 17 . The RBD immunogen contains targets of humoral and cellular 48 immunity 18 with a molecular size (34 kDa) predictive of lymph node accumulation following 49 subcutaneous administration 17, 19 . While conventional vaccine adjuvants of low molecular weight (<20 50 kDa) are readily absorbed into circulating blood following peripheral administration, AMP-modification 51 is known to promote lymph node accumulation through non-covalent association with tissue albumin at 52 the injection site 20,21 . AMP-directed biodistribution of vaccine components to draining lymph nodes can 53 therefore significantly improve delivery to critical immune cells and engage mechanisms coordinating 54 the magnitude and quality of the immune response while restricting exposure to immunologically 55 irrelevant or tolerizing sites preferentially accessed by agents cleared through the blood. Previously, 56 application of this approach in ELI-005 promoted highly potent RBD-specific T cell and neutralizing 57 antibody responses following a three-dose immunization regimen in mice demonstrating the potential 58 of lymph node targeted vaccination to generate immunity with an attractive balance of cellular and 59 humoral responses 17 .

Here we describe further evaluation of ELI-005 in mice including the development of a simplified prime-61 boost regimen to rapidly induce VOC-directed cross-reactive antibody titers alongside highly potent T 62 cell responses. Through comparison of soluble and AMP-CpG, we explore the differential underlying 63 mechanisms of innate immune response activation in draining lymph nodes and their dependence on 64 effective lymph node targeting of adjuvant. We also evaluated ELI-005 for T cell and humoral 65 immunogenicity in Rhesus macaques as a predictive model for potential safety and activity in human 66 subjects. AMP-CpG showed higher levels of circulating RBD-specific CD8 + cytokine + T cells throughout the 91 assessment period, compared to animals immunized with soluble CpG or mock (Figure 1g -h). Together, 92 these results demonstrate the potency of ELI-005 prime-boost immunization and the critical role of 93 AMP-CpG to promote significantly enhanced and persistent cellular immunity in multiple tissues 94 important for potential prophylactic protection or moderation of disease severity.

Long-term evaluation of the humoral response showed rapid seroconversion after prime-boost 96 immunization for both soluble and AMP-CpG immunized groups. Serum IgG antibody titers against both 97 the immunizing antigen WH-01 RBD, as well as the Delta RBD, peaked shortly after boost immunization 98 and were maintained at high levels for the duration of this study (Figure 1i-j) . Overall, these data 99 indicate that ELI-005, adjuvanted with AMP-CpG, generates potent cross-reactive humoral immune 100 responses that persist long term at significant levels.

The persistence of circulating RBD-specific T cells post-boost suggests that ELI-005 promotes the 103 development of durable memory T cells. Analysis of this long-lived T cell population showed a 104 significantly larger population of circulating RBD-specific tetramer + cells with a CD44 + CD62L + central 105 memory phenotype among CD8 + T cells. AMP-CpG vaccine-induced responses were 2.5-fold greater than 106 in the soluble CpG comparator group (Figure 2a) , suggesting the potential for a more rapid expansion of 107 effector T cells upon future exposure to SARS-CoV-2. Consequently, we characterized the recall response 108 upon antigen re-exposure following subcutaneous administration of 10 µg WH-01 RBD, 17 weeks after 109 boost. Recall responses in circulating blood analyzed 7 days later confirmed rapid expansion of RBD-110 specific CD8 + T cells in ELI-005 immunized animals (14-fold from 1.9% pre-challenge to 26.8% post recall, 111 Figure 2b ). In contrast, recall responses in soluble CpG immunized mice were significantly lower, 112 indicating a substantial advantage for long-term memory response expansion in AMP-CpG immunized 113 animals.

Considering the robust T cell recall response observed 17 weeks after boost, an additional cohort of 115 animals was recalled 30 weeks after boost. Circulating CD8 + and CD4 + T cell responses were measured 116 prior to and 7 days post recall with 10 µg WH-01 RBD antigen. In the peripheral blood of animals 117 immunized with AMP-CpG, the percentage of polyfunctional cytokine-producing CD8 + and CD4 + T cells 118 increased 5-fold to 32% and nearly 3-fold to 8.5% post challenge, respectively . This 119 compared to just 2.3% of CD8 + and 1.3% of CD4 + T cells that displayed a cytokine + phenotype in the 120 soluble CpG group. Of note, the long-term CD8 + and CD4 + responses 32 weeks after immunization with 6 AMP-CpG persisted at levels that were higher than responses in the soluble CpG group even after they 122 had been recalled.

Cellular immune responses were analyzed in tissues important for point-of-entry and systemic 124 protection against SARS-CoV-2. Perfused lung tissue and spleens from immunized mice were analyzed 7 125 days after WH-01 RBD challenge at week 32. Antigen recall of animals originally immunized with AMP-

CpG induced a large proportion of lung resident T cells to secrete polyfunctional cytokines. In these 127 animals, 37% of CD8 + T cells ( Figure 2e ) and 12% of CD4 + T cells ( Figure 2f ) secreted cytokines, compared 128 to 14% and <2% in mice treated with soluble CpG, respectively. Similar trends were observed in spleen,

where immunization with AMP-CpG produced a mean frequency of nearly 2,500 IFNγ SFC per 10 6 130 splenocytes opposed to < 100 IFNγ SFC per 10 6 splenocytes in the soluble CpG group. These data 131 indicate that AMP-CpG-adjuvanted ELI-005 generates robust and durable T cell memory populations 132 that can quickly and potently expand upon re-exposure to SARS-CoV-2 antigens, providing potent, long-133 term cellular immunity at important sites for immune surveillance and disease moderation. Humoral 134 responses did not show an increase in serum titer at 7 days post recall, suggesting that the antibody 135 recall response may lag behind that of T cells (Supplemental Figure S1a 

The previous data demonstrated the potency and durability of ELI-005-induced immunity. However, 138 global vaccine distribution in support of mass immunization campaigns requires stability at near ambient 139 temperature without dependence on an ultra-cold storage supply chain. To examine the stability of ELI-140 005, an immunogenicity study was conducted, in which formulated vaccine doses were stored at 141 common refrigerator temperatures (4°C) for 2-4 weeks or at room temperature (22°C) for 3 days prior to 142 dosing. One week after the second immunization, animals were assessed for their cellular and humoral 143 immune responses. Cytokine producing CD8 + and CD4 + T cells in both peripheral blood and lung, as well 144 as serum antibody titers, demonstrated near-identical trends to those observed in Figure 1 where doses 145 were prepared freshly before injection (Supplemental Figure S2a suggesting an influx of APCs, poised for pathogen recognition (e.g., Tlr3/4/9, Myd88, cytosolic PRRs), 167 antigen processing and presentation (e.g., cathepsins, Tap1, β2m), and co-stimulation (eg., Cd40, Cd86).

Transcriptional evidence of innate immune cell recruitment and activation in the lymph nodes was 169 further associated with upregulated inflammatory (e.g., Tnf, Stat3), and anti-viral (e.g., Stat1, IRFs) 170 processes suggesting a potential role for these pathways in AMP-CpG-driven acute immune activation.

In contrast, minimal transcriptional changes were induced in the lymph nodes by soluble CpG consistent 172 with prior evidence showing poor lymph node accumulation and reduced immunogenicity in vivo. 

Together with the observed lymph node proteomics profile, these data indicate the essential role of 219 AMP-modification to enable lymph node delivery and potent innate immune activation by CpG DNA in 220 vivo.

Upon observing the induction of IFNβ, the strong transcriptional signature of interferon response and 222 apparent involvement of all major innate immune cell lineages in the lymph node immune response 223 following immunization with AMP-CpG, we hypothesized that activation might proceed in part through 

In contrast, genes involved in T cell anergy and exhaustion were downregulated including Siggir which 303 inhibits signal transduction of TLRs, Il18r1 and Il1r1 to negatively regulate inflammation 27 . These results

share notable similarities to those collected in mice and further confirm that immunization in NHPs with 12 AMP-CpG induces a pro-inflammatory environment to promote adaptive immunity including T cell 306 priming and activation in the draining lymph nodes.

Despite authorization and widespread uptake of effective vaccines, the COVID-19 pandemic continues to 309 result in global waves of infection driven by the emergence of additional variants 28 , and relatively short- 

C57Bl/6J mice (n=5) were immunized twice with 10 µg WH-01 RBD protein and 1 nmol soluble or AMP-

CpG. a-b 17 weeks post dose 2, blood was collected for a pre-recall time point, followed by subcutaneous 587 challenge with 10 µg WH-01 RBD protein on the next day, and a second blood collection 7 days later. a

Pre-challenge peripheral blood leukocytes were stained for CD44 and CD62L. Shown are percentages of 589 tetramer + CD44 + CD62L + cells among CD8 + T cells. b Peripheral blood CD8 + T cells collected before and 590 after challenge were stained with tetramer specific for Spike RBD and analyzed by flow cytometry. c-g 30 591 weeks post dose 2, mice were challenged subcutaneously with 10 µg WH-01 RBD protein and assayed 7 592 days later. CD8 + T cells collected from blood (c) and perfused lung (e), and CD4 + T cells collected from 593 blood (d) and perfused lung ( 

For NHP studies, ELISpot assays were performed using Monkey IFNγ ELISpotPLUS kits (MabTech, cat# 830 3241M-4HPW). Precoated 96-well ELISpot plates were blocked with RPMI + 10% FBS for 2 hours at room 831 temperature. 0.2x10 6 PBMCs were plated into each well and stimulated overnight with 0.4 ug per 832 peptide per well of RBD-derived OLPs for WH-01, Beta and Delta variants. The spots were developed 833 based on the manufacturer's instructions. In both animal models, PMA (50 ng/mL) and ionomycin (1 834 µM) were used as positive controls, and RPMI + 10% FBS with DMSO was used as the negative control.

Spots were scanned and quantified using an S6 ImmunoSpot analyzer (CTL).

Enzyme linked immunosorbent assay (ELISA) for antibody titers

For mouse studies, ELISA assays were performed as previously described (Steinbuck et al., 2021) .

(PE/Dazzle 594, clone: 2H7, BioLegend), anti-human CD4 (APC-Cy7, clone: OKT4, BioLegend) and anti-870 human CXCR5 (PcP-eF710, clone: MU5UBEE, ThermoFisher). Cells were fixed/permeabilized using 871 Transcription Factor Staining Buffer Set (ThermoFisher, cat# 00-5521-00) and further stained with anti-872 human Bcl-6 (PE, clone: 7D1, BioLegend) and anti-human Ki-67 (BV421, clone: 11F6, BioLegend). Sample 873 acquisition was performed on a BD FACS Symphony and data were analyzed with BD FlowJo V10 874 software.

SARS-CoV-2 pseudovirus neutralization assay for NHP sera

The SARS-CoV-2 pseudovirus assay was performed by Genecopoeia as previously described 47 . SARS-CoV- 

For comparing two experimental groups, two-tailed t-test analysis was utilized when normal distribution 889 and homogeneity of variance, determined by Levene's Test, were established. Where these assumptions 890 did not apply, the Mann-Whitney test was used instead. For comparison of multiple groups, ordinary 891 one-way ANOVA was used to compare experimental groups. NanoString statistical analysis was 892 performed using Rosalind software.

COVID-19 Vaccines and SARS-CoV-2 Transmission in the Era of New Variants: 439 A Review and Perspective

Tracking Changes in SARS-CoV-2 Spike: Evidence that D614G Increases Infectivity 441 of the COVID-19 Virus

Spike mutation D614G alters SARS-CoV-2 fitness

Evaluating the Effects of SARS-CoV-2 Spike Mutation D614G on Transmissibility

SARS-CoV-2 variants B.1.351 and P.1 escape from neutralizing antibodies

Antibody resistance of SARS-CoV-2 variants B.1.351 and B.1.1.7

Transmission, infectivity, and neutralization of a spike L452R SARS-CoV-2 variant

Efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine against SARS-CoV-2 variant 452 of concern 202012/01 (B.1.1.7): an exploratory analysis of a randomised controlled trial. The 453

Evidence for increased breakthrough rates of SARS-CoV-2 variants

BNT162b2-mRNA-vaccinated individuals

Community transmission and viral load kinetics of the SARS-CoV-2 delta 457 (B.1.617.2) variant in vaccinated and unvaccinated individuals in the UK: a prospective, 458 longitudinal, cohort study

Safety and Efficacy of Single-Dose Ad26.COV2.S Vaccine against Covid-19

Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-462

CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK

Efficacy of NVX-CoV2373 Covid-19 Vaccine against the B.1.351 Variant

Humoral and cellular immune memory to four COVID-19 vaccines

Immune correlates of protection by mRNA-1273 vaccine against SARS-CoV-2 in 469 nonhuman primates

Potential SARS-CoV-2 Immune Correlates of Protection in 471 Infection and Vaccine Immunization

A lymph node-targeted Amphiphile vaccine induces potent cellular and 33

5-fold change, red) relative to baseline values. Insignificant values with p ≥ 0.05 are 681 shown in gray. Genes were clustered into 9 groups (box insert) using Gene Ontology databases and 682 annotated at the top of the heat map: (1) adaptive immunity, (2) antigen processing and presentation

) inflammation, (6) innate immunity, (7) interferon signaling, (8) migration, 684 and (9) T cell anergy and exhaustion. c Fold-change bar graphs for selected gene transcripts

Supplementary Fig. 1 Serum antibody responses induced following SARS-CoV-2 Spike RBD 689 immunization in mice

C57Bl/6J mice (n = 10) were immunized twice with 10 µg WH-01 RBD protein and 1 nmol soluble or AMP-691

30 weeks post dose 2, mice were challenged subcutaneously with 10 µg WH-01 RBD protein

Blood serum titers of anti-SARS-CoV-2 RBD antibodies were assayed 1 day before and 7 days after antigen 693 challenge for WH-01 (a) and Delta (b). Values depicted are means ± standard deviation. ns, not significant

0001 by two-sided t-test applied to antibody titers

Supplementary Fig. 2 ELI-005 retains immunogenicity upon storage at refrigerated or ambient 698 temperatures

were admixed and 700 stored at 4°C or 22°C (RT, room temperature) for the indicated period. C57Bl/6J mice (n=5) were 701 immunized twice at week 0 and 2 and assayed 7 days after. CD8 + T cells from peripheral blood (a) and 702 perfused lung (b), as well as CD4 + T cells (c-d) from those tissues were stimulated overnight with 703 overlapping WH-01 RBD peptides

booster dose, blood serum was analyzed for anti-SARS-CoV-2 RBD antibody titers against WH-01 RBD. ns, 705 not significant, by two-sided t-test applied to T cell frequencies and antibody titers

Supplementary Fig. 3 AMP-CpG immunization induces lymph node transcriptional reprogramming 709 reflecting APC recruitment with increased potential for antigen processing and presentation

a Heatmap representation of whole lymph node mRNA analyzed by NanoString nCounter® Mouse 711

Shown are mock-subtracted, average Z-scores of gene transcript levels significantly

-fold change, blue) or upregulated (≥ 1.5-fold change, red) at 2 hours post 713 injection relative to mock immunization. Insignificant values with p ≥ 0.05 are shown in gray. Gene groups 714 follow the same numbering scheme as in fig 3. b-g Volcano plot representation of log

6 hours (d-e), and 72 hours (f-g) post injection 716 representing data from supplementary figure 3a, and figure 3b and 3f, respectively. Mock vaccines 717 contained PBS vehicle only. Dotted horizontal line represents significance threshold of p = 0.05; vertical 718 dotted lines represent fold-change limits of ± 1.5-fold change

Data reported in fig 4 were re-analyzed to determine statistical significance between groups immunized 725 with soluble and AMP-CpG, whereas statistical analysis in fig 4 compared to mock treatment. Values 726 depicted are mean ± standard deviation

0001 by non-727 parametric two-tailed T test comparing immunization with soluble CpG to AMP-CpG

Flow cytometric analysis strategy and example scatter plots for surface marker and intra-cellular cytokine

Rhesus macaques (n=2/3) were immunized at week 0 and 4 with 140 µg WH-01 RBD protein admixed 742 with either 5 mg or 10 mg of AMP-CpG, or 140 µg full WH-01 Spike protein admixed with AMP-CpG at 5 743 mg. Vital signs were recorded

Sera was collected at multiple timepoints for assessment in a hematology complete blood 745 count panel. Shown are white blood cells (WBC) (c), % neutrophils (d), % lymphocytes (e), % monocytes 746 (f), and % hematocrit (g)

For mouse studies, vaccines consisted of 10 µg SARS-CoV-2 Spike RBD WH-01 protein (GenScript, cat# 751 Z03483), combined with 1 nmol AMP-CpG-7909

Antigen 754 rechallenge doses consisted of 10 µg SARS-CoV-2 Spike RBD WH-01 protein, administered subcutaneously 755 bilaterally into the tail base. For experiments testing storage temperature conditions, doses were admixed 756 and stored at 4°C or room temperature (22°C) for the indicated time. For NHP studies, vaccines consisted 757 of either SARS-CoV-2 Spike RBD WH-01 protein (GenScript; cat# Z03483) or SARS-CoV-2 WH-01 Spike 758 protein (Acro Biosystems, SPN-C5H9

For stimulation of ex vivo cell samples, OLP peptide pools of 15mers with 11 amino acid overlap were 762 generated spanning the SARS-CoV-2 Spike RBD (R319-S591, GenScript). Sequences that contained VOC 763 mutations were exchangeable with the corresponding mutated peptides due to a modular OLP pool

All animal studies were carried out under an institute-approved Institutional Animal Care and Use 767

Committee (IACUC) protocol following federal, state, and local guidelines for the care and use of animals

The indicated antigen-adjuvant combinations were administered into mice 770 subcutaneously at the base of the tail (bilaterally, 50 µl each) at weeks 0 and 2. Peripheral blood samples 771 were collected as indicated 7 days post second dose, and 1 day prior to and 7 days post antigen challenge 772 where applicable. Spleen and lung tissue was collected either 7 days post second dose or 7 days post 773 antigen challenge. Lungs were harvested following perfusion with 10 mL of PBS into the right ventricle of 774 the heart. Lung tissue was physically dissociated and digested with RMPI 1640 media

Indian-origin, 4-5 year old female rhesus macaques (Macaca mulatta) were 777 randomly allocated into 3 groups of 2 or 3 animals. All animals were housed at

Animals received immunizations subcutaneously into the upper thigh at week 0

Lymph node FNAs were collected from the inguinal lymph nodes at weeks 0 and 4 for flow cytometric 782 analysis of lymph node resident cells and 24 hours post week 10 boost for NanoString analysis

Cell activation and cytokine determination by ICS

For mouse vaccine immunogenicity studies

For mouse innate immune response analyses, surface activation marker staining and ICS analysis was 787 performed on fixed/permeabilized (BD, cat# 554714) single cell suspensions of inguinal lymph nodes 788 collected 6-48 hours post immunization. Live/Dead fixable stain (Aqua, Invitrogen, cat# L34966) was 789 used to exclude dead cells

Ly6C (BV650, clone: HK1.4, BioLegend), Ly6G (PcP-Cy5.5, clone: 1A8, BioLegend)

Dazzle594, clone: 29A1.4, BioLegend), MHCII (APC-Cy7, clone: M5/114.15.2, BioLegend), CD86 (BV785, 793 clone: GL-1, BioLegend), IFNγ (BV711, clone: XMG1.2, BioLegend), IL12p40 (PECy7, clone: C15

IL1β (PE , clone: NJTEN3, Invitrogen), IL6 (AF488, clone: MP5-20F3, Invitrogen), and IFNβ 795 (APC, ASSAYPRO, cat# 32183-05161T). Sample acquisition was performed on BD FACS Symphony and 796 data were analyzed with BD FlowJo V10 software

Macrophages were defined as CD3 -/CD19 -/NKp46 -/CD11c -/Ly6G -/CD11b + /Ly6C low ; monocytes were 798 defined as CD3 -/CD19 -/NKp46 -/CD11c -/Ly6G -/CD11b + /Ly6C high

Neutrophils were defined as CD3 -/CD19 -/NKp46 -/CD11c -/CD11b + /Ly6C med /Ly6G + ; NK 800 cells were defined as CD3 -/NKp46 + . Data was expressed as total cell number per lymph node over time, or 801 as Z-scores of the number of cells that were positive for the respective cytokine or surface marker

10 6 PBMCs/well were resuspended 803 in R10 media supplemented with anti-CD49d monoclonal antibody (clone: 9F10, BD), anti-CD28 804 monoclonal antibody (clone: CD28.2, BD), and Golgi inhibitors monensin (Fisher Scientific, cat# 805 NC0176671) and brefeldin A (Fisher Scientific, cat# 50-112-9757) and incubated at 37°C for 8 hours, then 806 maintained at 4°C overnight. The next day

clone: S3.5, Invitrogen), CD8 (AF647, clone: RPA-T8, BioLegend), CD45RA (FITC, clone: 5H9, BD

CCR7 (BV650, clone: G043H7, BioLegend), and aqua live/dead dye (Invitrogen, L34957), and 809 subsequently fixed with BD CytoFix/CytoPerm (BD, 554714). Cells were further stained with antibodies

The detection 840 antibody used was horseradish peroxidase (HRP)-conjugated goat anti-human IgG (H+L) (ThermoFisher, 841 cat# SA5-10283) at a 1:2000 dilution. The WHO International Standard for anti-SARS-CoV-2 842 immunoglobulin (anti-RBD IgG High 20/150, NIBSC) was used for reference values. Serum titers were 843 determined at an absorbance cutoff of 0.5 OD and converted into binding antibody units/mL

Gene transcript analysis by Nanostring

For mouse studies, inguinal lymph nodes were harvested from immunized C57BL/6J mice at the indicated 847 time points, processed into single cell suspensions, and lysed with RLT buffer (Qiagen

Transcriptional profiles of immune signaling were generated using the nCounter Mouse Immunology 849 Panel of 568 mouse immune response genes (NanoString Technologies). For NHP studies, lymph node 850 FNA were assessed using the nCounter NHP Immunology Panel of 770 macaque immune response genes 851 (NanoString Technologies). Transcriptional responses were assessed with nSolver software v4.0 852 (NanoString Technologies) and differential gene expression was carried out using ROSALIND software

To determine the cytokine/chemokine content of lymph nodes, animals were vaccinated and inguinal 856 lymph nodes were collected 6-48 hours post immunization. For IFNAR-1 blockade, an antagonistic mAb 857 (clone: MAR1-5A3, BioXcell) or an isotype control (clone: MOPC-21, BioXcell) were injected 858 intraperitoneally 24 hours prior to immunization

containing Mini protease inhibitor cocktail (Roche, cat# 53945000) and HALT phosphatase 860 inhibitors (Thermo Fisher, cat# 78442) was added to the intact lymph nodes prior to homogenization 861 using a TissueLyser II (Qiagen). Centrifugation-cleared lysates were analyzed with Luminex Cytokine and 862 Chemokine kits (EMDMillipore, cat# MCYTOMAG-70K and MECY2MAG-73K)

NHP antigen-specific GC B cell analysis in lymph node FNAs 864

Lymph node cells from FNA were incubated with 866 the RBD-fluorochrome complex and subsequently stained with aqua live/dead dye

anti-human IgM (FITC, clone: G20-127, BD), anti-human IgG (PE-Cy7, clone: G18-145, BD)

CD3 (AF700, clone: SP34-2, BD), anti-human PD-1 (BV650, clone: EH12.1, BD)

We thank D. M. Lidgate for expert medical writing assistance, and C. S. Davis at CSD Biostatistics Inc., A.

Bot at Capstan Therapeutics and D. J. Irvine at M.I.T. for helpful advice and discussion.