Vaccine production

Full-length S glycoprotein gene sequence (GenBank MN908947 nucleotides 21563–25384) for NVX-CoV2373 was synthetically produced and codon optimized for expression in Spodoptera frugiperda (Sf9) cells (GenScript) as previously described in ref. 18. Briefly, the S1/S2 furin cleavage site 682-RRAR-685 was modified to 682-QQAQ-685 and two proline substitutions were introduced at positions K986P and V987P (2P) to stabilize the full-length SARS-CoV-2 S101. For NVX-CoV2443, the following substitutions were introduced in addition to the stabilizing 3Q-2P mutations: L18F, T20N, P26S, D138Y, R190S, K417T, E484K, N501Y, D614G, H655Y, T1027I, V1176F. Matrix-MTM was supplied by Novavax AB (Uppsala, Sweden). Recombinant protein nanoparticles, Matrix-MTM and formulation buffer (25 mM sodium phosphate, 300 mM sodium chloride, 0.01% Tween 80, pH 7.2) were mixed immediately before administration.

Rhesus macaque model

Animal experiments were conducted following the guidelines and regulations of the Association for Assessment and Accreditation of Laboratory Animal Care, the Swedish Animal Welfare Agency and the European guidelines for animal care. The study was approved by the regional animal ethics committee of Northern Stockholm, institutional ethical committee “Comité d’Ethique en Expérimentation Animale du Commissariat à l’Energie Atomique et aux Energies Alternatives” and the French Ministry of Higher Education and Research. Indian rhesus macaques (Macaca mulatta, male and female, 4–6 years old at study start) used in immunization experiments (n = 12) and as control animals (n = 3) were housed at Astrid Fagraeus Laboratory at Karolinska Institutet (Stockholm, Sweden). Six naïve Chinese rhesus macaques (Macaca mulatta, male, 3–4 years old at study start), serving as additional control animals, were housed at the IDMIT animal facility (CEA, Fontenay-aux-Roses, France). The animals received intramuscular injections according to the study schedule in the left quadriceps, consisting of 5 μg protein nanoparticles (NVX-CoV2373 or NVX-CoV2443) and 50 μg Matrix-MTM. At week 66, all animals were challenged with a total dose of 8 × 105 PFU SARS-CoV-2 (isolate USA-WA1/2020 (BEI: NR-53872) or hCoV-19/Japan/TY7-503/2021 (BEI: NR-55364)). The inoculum was diluted in PBS and administered 3 mL intratracheally and 1 mL intranasally (0.5 mL in each nostril) as described previously in refs. 27,53. Heparinized peripheral blood, serum, bone marrow aspirates and bronchoalveolar lavage (BAL) were collected as depicted in Fig. 2A and Fig. 6A. Body weight and temperature were monitored at each sampling timepoint. Six of the control animals, housed at CEA, were not sampled for BAL at day 4 post challenge due to restrictions in the ethical permit.

Two female Chinese rhesus macaques (Macaca mulatta, 4–5 years old) were immunized with 75 μg Matrix-MTM or PBS as a part of another study12. To maximize sample collection, each animal received an immunization in each limb muscle (deltoid or quadriceps) and PBS injections in calves. Tissue biopsies of injection sites and dLNs were collected 24 h after immunization and stored in RNALater (Invitrogen) at −20 °C until use.

Safety measurements

Complete blood counts and clinical chemistry analyses were performed at baseline, 24 h and 14 days after the first immunization by Adlego Biomedical (Solna, Sweden). Clinical chemistry was performed using an Abaxis Vetscan VS2 3.1.35 chemistry analyzer with mammalian liver profile rotors (Triolab).

Sample processing

A standard gradient density centrifugation using Ficoll-Paque (GE Healthcare) was used to isolate peripheral blood mononuclear cells (PBMCs) from heparinized blood. PBMCs were either cryopreserved in 10% dimethyl sulfoxide (DMSO)/fetal calf serum (FCS) or immediately used for downstream applications. Heparinized bone marrow samples were processed in the same way as blood, and additionally passed through a 70 μm cell strainer before use. BAL cells were separated from the supernatant by centrifugation and filtration through a 70 μm cell strainer, then were used fresh in a T cell recall assay. BAL fluid was concentrated 10-fold using Amicon Ultra centrifugal filter units with a 30 kDa cutoff (Millipore) before downstream analysis.

Innate immunoprofiling

On days 0, 1, and 14 after the first immunization, the cellular composition of the PBMCs was analyzed using flow cytometry. Freshly isolated PBMCs were stained with Live/Dead Fixable Blue Dye (Life Technologies, cat# L-23105, 1:40 dilution) and FcR blocking reagent (Miltenyi Biotec, cat# 130-059-901, 1:20 dilution) followed by a panel of antibodies: CD40 FITC (5C3, Biolegend, cat# 334306, 1:20 dilution), NKG2A PE (Z199, Beckman Coulter, cat# IM3291U, 1:40 dilution), CD80 BV421 (L307.4, BD, cat# 564160, 1:40 dilution), CCR7 PE-Dazzle594 (G043H7, Biolegend, cat# 353236, 1:50 dilution), CD123 Per-CP-Cy5.5 (7G3, BD, cat# 558714, 1:80 dilution), CD3 APC-Cy7 (SP34-2, BD, cat# 557757, 1:80 dilution), CD66 APC (TET2, Miltenyi Biotec, cat# 130-118-539, 1:80 dilution), CD70 BV786 (Ki-24, BD, cat# 565338, 1:80 dilution), HLA-DR BV650 (L243, Biolegend, cat# 307650, 1:80 dilution), CD11c PE-Cy7 (3.9, Biolegend, cat# 301608, 1:160 dilution), CD16 AF700 (38 G, BD, cat# 560713, 1:160 dilution), CD20 BV605 (2H7, Biolegend, cat# 302334, 1:160 dilution) and CD14 BV510 (M5E2, Biolegend, cat# 301842, 1:160 dilution). After washing with PBS, samples were fixed using 1% paraformaldehyde (PFA) and acquired on a BD LSRFortessa cell analyzer. The data were analyzed using FlowJo software v.10.7.1 (FlowJo).

Plasma cytokine and chemokine quantification

After the first immunization, plasma cytokines and chemokines were analyzed using the ProcartaPlex NHP Cytokine & Chemokine Panel 30plex (Thermo Fisher Scientific) according to the manufacturer’s instructions at the Affinity Proteomics core facility, SciLifeLab, Stockholm, Sweden. Samples collected at baseline, 24 h, and 14 days after the immunization were analyzed using a MagPix (Luminex) instrument, and the data were analyzed with Belysa Immunoassay Curve Fitting software (Millipore). Standard curves were generated using 5-parameter logistic curve fit.

RNA sequencing and bioinformatic analysis

At 0 and 24 h after the first immunization, whole blood was collected into PAXgene Blood RNA tubes (PreAnalytiX) and stored at −20 °C. Tissue biopsies, collected 24 h after 75 μg Matrix-MTM or PBS immunization, were placed in RNALater (Invitrogen) and stored at −20 °C. RNA isolation was performed using PAXgene Blood RNA Kit (PreAnalytiX) for blood samples and RNeasy mini kit (Qiagen) for lymph nodes according to manufacturer´s instructions. For muscle biopsies, RNA was isolated using RNeasy Fibrous Tissue mini kit (Qiagen) according to the manufacturer’s instructions. RNA integrity was checked using TapeStation RNA ScreenTape assay (Agilent Technologies) according to the manufacturer’s instructions. In preparation for Illumina sequencing, isolation of mRNA, cDNA synthesis, anchor ligation, amplification and library indexing were performed using the Illumina Stranded mRNA Prep Ligation kit according to the manufacturer’s instructions. The libraries were sequenced using an Illumina NovaSeq S6000 on one lane of the S4-300 (v1.5) flowcell with the standard paired-end read set up (2 × 150 bp), with an average sequencing depth of 38 M reads per sample.

Samples were preprocessed using an nf-core rnaseq pipeline (v3.7)102. STAR alignment (v2.7.10a) to the M. mulatta genome (Mmul_10) was used for genome alignment, and quantification was performed with Salmon (v1.8.0). A customized bioinformatic analysis workflow in R (v4.1.2) was used: differential gene expression analysis was performed using DESeq2 (v1.34.0), ClusterProfiler (v4.2.2) was used for Gene Set Enrichment analysis in combination with the Blood Transcriptome Module database103. To compare differentially expressed genes, a Wald test was performed with multiple hypothesis testing controlling the false discovery rate using the Benjamini-Hochberg procedure (q < 0.05).

Spike detection in plasma by MSD

An S-PLEX SARS-CoV-2 Spike Kit (K150ADJS, Meso Scale Diagnostics) was used to quantify the Spike concentration in plasma before and after immunization according to the manufacturer’s instructions at the Affinity Proteomics core facility, SciLifeLab, Uppsala, Sweden. 25 μL of plasma was used for the analysis, and the plates were read using a MESO QuickPlex SQ 120 instrument. Quantification was based on an 8-point calibration standard curve using the recombinant SARS-CoV-2 S protein included in the kit. The raw signals were converted into data expressed in femtograms per milliliter.

Binding antibody titers by ELISA

96-well half-area ELISA plates (Greiner Bio) were coated with recombinant proteins (prefusion-stabilized S (S-2P) or RBD, acquired through the Global Health Discovery Collaboratory funded by the Bill & Melinda Gates Foundation) at 1 μg/mL in PBS and incubated overnight at 4 °C. Plates were washed three times using PBS-T (PBS containing 0.05% Tween 20) and blocked with blocking buffer (PBS with 5% (w/v) skimmed milk powder) for 1 h at room temperature (RT). Duplicates of serially diluted samples in blocking buffer were added to the plate and incubated for 2 h at RT. The plates were washed three times and goat anti-monkey IgG-horseradish peroxidase (Nordic MUBio, cat# 246-GAMon/IgG(H + L), 1:20 000 dilution) in blocking buffer was added for 1 h at RT. For development, 1-Step Ultra TMB-ELISA substrate (Thermo Fisher Scientific) was added for 5 min and the reaction was stopped with 1 M H2SO4. The absorbance was measured at 450 nm with background correction at 570 nm. Data was analyzed with Prism v9.4.1 using 4-parameter logistic curve fit.

Competition ELISAs

RBD and P.1 Spike competition ELISAs were performed as above with some modifications. For the P.1 competition ELISA, the plates were coated using BV2373 protein (WA-1 Spike, Novavax). Serially diluted plasma samples were pre-incubated with 20 μg/mL of competitor protein (RBD or BV2443 (P.1 Spike)) or blocking buffer for 30 min, and then transferred to the ELISA plates for further 1.5 h incubation. The proportion of competition was calculated based on the decrease in ED50 value between the condition with and without the competitor.

For the competition assay using characterized monoclonal antibodies (mAbs), the antibody sequences were retrieved from the literature104,105,106,107,108 and recombinant antibodies were produced by GenScript (Leiden, Netherlands) or Institute for Protein Design (Seattle, WA, USA). The antibodies were biotinylated using EZ-Link Micro Sulfo-NHS-LC Biotinylation Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions.

The ELISA was performed as above, with some modifications. The experimental conditions were optimized for each competition mAb separately to maximize the dynamic range. ELISA plates were coated with 1 μg/mL S-2P for B38, C144, S2M11 and S309 and 2 μg/mL S-2P for CR3022. Serial dilutions of plasma in duplicates were added to the ELISA plates for 30 min to allow for binding of plasma antibodies, followed by an equal volume of biotinylated competitor antibodies at a predetermined concentration in blocking buffer for further 1.5 h incubation at RT. The detection was performed using Pierce high sensitivity NeutrAvidin-horseradish peroxidase (Thermo Fisher) at a 1:5000 dilution in PBS, and plates were developed using 1-Step Ultra TMB-ELISA substrate (Thermo Fisher). Competition between unbiotinylated and biotinylated competitor antibodies served as standard curve. Levels of competitor-like antibodies in the plasma were calculated by multiplying the ED50 of unbiotinylated mAbs with the ED50 of each sample.

Antibody avidity assay

An avidity ELISA was performed as above, with some modifications. After sample incubation and washing with PBS-T, the samples were treated with 2 M NaSCN (Sigma-Aldrich) or PBS for 10 min. Detection and development were performed as described. The avidity index was calculated as a ratio between ED50 values of NaSCN- and PBS-treated conditions. If ED50(NaSCN) was below the level of detection, an avidity index of 0 was assigned. On the contrary, if ED50(NaSCN) > ED50(PBS), an avidity index of 1 was assigned.

Neutralization and pseudovirus neutralization assays

Live virus neutralization assay using an authentic wild type SARS-CoV-2 virus (strain 2019-nCoV strain 2019-nCov/Italy-INMI1. European Virus Archive Global (EVAg), Marseille, France) was performed at VisMederi srl (Siena, Italy) as previously described in ref. 109, and pseudovirus particle neutralization assay using vesicular stomatitis virus (VSV)ΔG S pseudotyped virus with a luciferase reporter was performed at Nexelis (Laval, Canada) as described previously in ref. 110. All samples were assayed in duplicates.

Neutralization of variants of concern was assessed as previously described in ref. 111. Briefly, WA-1, P.1, BA.2, and BA.5 Spike-pseudotyped lentivirus particles delivering a luciferase reporter (standardized to ±100 000 RLU) were pre-incubated with three-fold serial serum dilutions for 1 h at 37°C in a black-walled 96-well plate in duplicates. 10 000 HEK293T-ACE2 cells were added to each well and the plates were incubated at 37°C and 5% CO2 for 48 h. Luminescence was measured using Bright-Glo substrate (Promega) using GloMax Navigator Luminometer (Promega). Effective 50% neutralization titer was calculated based on the luminescence of infected control wells in the absence of serum.

Systems serology

Antigen-specific antibody subclass isotypes and FcγR binding were analyzed by Luminex multiplexing in plasma samples collected at weeks 0, 2, 6, 37 and 61 from the 3-dose group. Antigens (BV2373, BV2443 (Novavax), WA-1 Spike, RBD, NTD and S2, B.1.351 Spike and RBD, B.1.617.2 Spike and RBD, B.1.1.529 Spike and RBD, HKU Spike, 229E Spike, MERS Spike (all Sino Biological)) were coupled to magnetic Luminex beads by carbodiimide-NHS ester coupling with an individual region per antigen according to manufacturer’s instructions. Antigen-coupled beads were incubated with different plasma dilutions to form immune complexes. Mouse anti-rhesus detection antibodies were added for each Ab isotype (total IgG (Southern Biotech), total IgM (Life Diagnostic), IgG1, IgG2, IgG3, IgG4, total IgA (all NIH Nonhuman Primate Reagent Resource)), followed by PE-conjugated anti-mouse Fc IgG antibody (Thermo Fisher). FcR-binding was quantified in a similar way by coupling PE-streptavidin (Agilent Technologies) to recombinant and biotinylated NHP FcRs (FcγR2A-1, FcγR2A-2, FcγR2A-3, FcγR2A-4, FcγR3A, courtesy of Duke Protein Production Facility) which were used as secondary probes. Relative antibody concentration per antigen was determined on an iQue Screener (IntelliCyt), and analysis was performed on IntelliCyt ForeCyt (v 8.1).

Bead-based assays were used to quantify antibody (Ab) functionality. Ab-dependent cellular phagocytosis (ADCP)112, Ab-dependent neutrophil phagocytosis (ADNP)113, and Ab-dependent complement deposition (ADCD)114 were measured as described previously. BV2373, BV2443 (Novavax), SARS-CoV-2 WA-1, and B.1.1.529 Omicron S protein (Sino Biological) were coupled to yellow-green fluorescent (ACDP, ADNP) or non-fluorescent (ADCD) neutravidin beads (Invitrogen) and incubated with plasma samples to form immune complexes. For ADCP, cultured human monocytes (THP-1 cell line) were incubated with immune complexes for 18 h at 37 °C, during which phagocytosis occurred. For ADNP, neutrophils were isolated from fresh whole blood using EasySep Direct Human Neutrophil Isolation kit (StemCell) and incubated with immune complexes for 30 min at 37 °C. Neutrophils were stained with an anti-CD66b Pacific Blue antibody (Biolegend) prior to flow cytometry. For ADCD, lyophilized guinea pig complement (Cedarlane) was reconstituted according to the manufacturer’s instructions, diluted in a gelatin veronal buffer (Boston BioProducts) and added to the immune complexes for 50 min at 37 °C. C3 that bound to immune complexes was detected with a FITC-conjugated goat anti-guinea pig complement C3 (MP Biomedicals) antibody. Flow cytometry acquisition of all assays was performed using a Stratedigm 1300EXi cytometer. For ADCP and ADNP, phagocytosis events were gated on bead-positive cells, with additional pre-gating step for CD66b+ cells for ADNP. A phagocytosis score for ADCP and ADNP was calculated as (percentage of FITC+ cells) * (the geometric mean fluorescent intensity (gMFI) of the FITC+ cells)/10,000. ADCD was reported as gMFI of FITC-anti-C3.

Principal component analysis (PCA) was constructed with FactoMineR R package v2.7115 using all collected antibody features as variables. The co-correlate network analysis was performed using a systemsseRology package (https://github.com/LoosC/systemsseRology)45. Features were selected using the least absolute shrinkage and selection operator (LASSO) algorithm. All antibody features were used as input and the LASSO selection was repeated 100 times. Only features that were selected in 80% of the trials were used in the co-correlate network analysis. The co-correlate network was built using a threshold of absolute Spearman rho greater than 0.7 and BH-adjusted p-value lower than 0.05.

Vaccine-specific memory B cell quantification and isolation

Recombinant S-2P, RBD and BV2443 proteins were biotinylated using EZ-Link Micro Sulfo-NHS-LC Biotinylation Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Biotinylated proteins were coupled to streptavidin-conjugated fluorophores (SA-PE, SA-APC or SA-BV421) to generate molecular probes. PBMCs were stained with 100 ng fluorescent protein probes for 20 min at 4°C, followed by 7-aminoactinomycin D (7-AAD, Thermo Fisher, cat# A1310, 1:4000 dilution) and a panel of antibodies: IgM PerCP-Cy5.5 (G20-127, BD, cat# 561285, 1:40 dilution), CD3 BV510 (SP34-2, BD, cat# 740187, 1:40 dilution), CD123 BV510 (6H6, Biolegend, cat# 306022, 1:40 dilution), CD16 BV510 (3G8, BD, cat# 563830, 1:80 dilution), HLA-DR BV650 (L243, Biolegend, cat# 307650, 1:80 dilution), IgG BV786 (G18-145, BD, cat# 564230, 1:80 dilution), CD20 BV605 (2H7, Biolegend, cat# 302334, 1:160 dilution), CD14 BV510 (M5E2, Biolegend, cat# 301842, 1:60 dilution) and IgD FITC (polyclonal, Southern Biotech, cat# 2030-02, 1:160 dilution), for another 20 min at 4°C. Cells were washed with PBS with 2% heat-inactivated FCS and fixed with 1% PFA. Samples were acquired on a BD LSRFortessa cell analyzer and the data were analyzed using FlowJo software v.10.7.1 (FlowJo).

At peak timepoints (week 6 and 37), antigen-specific IgG+ B cells were single-cell sorted into 96-well plates for isolation of vaccine-specific BCRs. The procedure was similar as above, except that the PBMCs were not fixed after staining and instead resuspended in complete medium (RPMI 1640 medium with 10% heat-inactivated FCS, 100 U/mL penicillin, 100 mg/mL streptomycin, and 2 mM L-glutamine) with 7-AAD. BD Aria III Fusion cell sorter was used for single-cell sorting of Lin- HLA-DR+ CD20+ IgM- IgG+ B cells, double positive for S protein. The plates were immediately frozen on dry ice for subsequent BCR amplification. At week 6, the samples were probed using S-2P PE, S-2P APC and RBD BV421. At week 37, BV2443 PE and BV243 APC were added to the probe mix. Index sorting data is available for selected plates, to determine RBD specificity of sorted cells.

A subset of Spike-specific BCR sequences at week 6 was acquired using FACS sorting combined with 10x Genomics single-cell RNA sequencing. Each sample was labeled using TotalSeq-C anti-human hashing antibodies (Biolegend) prior to cell sorting. Dual-S-positive IgM- B cells were sorted and processed with the Chromium Single Cell V(D)J Enrichment Kit, Human B Cell (10x Genomics), sequencing ‘5’ V(D)J’ enriched libraries and TotalSeq-C feature barcode libraries. Rhesus macaques are not directly supported by the 10x kits, so a mix of primers targeting the Ig constant region116 was spiked in during the two enrichment PCR steps (Supplementary Table 1). 5 μL of each 100 μM primer was added to MasterMix1 and MasterMix2 for enrichment steps 1 and 2, respectively, and diluted to a final volume of 50 μL. 5 μL of each MasterMix was added to the PCR mixture replacing 5 μl of nuclease-free water. 10X libraries were sequenced on a NovaSeq6000 using NovaSeq XP workflow in SP mode flowcell.

Isolation of Spike-specific B cell receptors and B cell repertoire analysis

BCR sequences were recovered from single-cell sorted B cell with reverse transcription using random hexamers (Invitrogen) as previously described in ref. 51. Superscript III reverse transcriptase kit (Invitogen) was used according to manufacturer’s instructions. Nested PCR protocol for amplification of heavy and light chain transcripts was performed as reported previously in ref. 117. PCR products were Sanger sequenced by Genewiz (Leipzig, Germany). The resulting chromatograms were pre-processed with scifer package (v0.99.3)118 and the high-quality sequences were aligned to the KIMDB rhesus database (v1.0)119 using IgDiscover (v0.15.1)120.

Raw sequencing reads from the 10x libraries were processed with CellRanger (v3.1.0, 10x Genomics). Demultiplexing was performed based on cell hashing with barcoded antibodies using Seurat R package (v4.0.6). We removed cells with less than 200 detected genes, mitochondrial reads higher than 20%, ribosomal reads lower than 5%, and all genes needed to be expressed in at least three cells. Doublets were detected and removed using DoubletFinder R package (v2.0.3). VDJ contigs were assembled using CellRanger’s --denovo option, and assembled contigs were subsequently re-aligned using IgDiscover (v0.13.0) and filtered based on default settings. Sequences were aligned to the KIMDB rhesus macaque database119. In total, 182 paired heavy and light chains were retrieved.

Detection of antigen-specific antibody-secreting cells

Antigen-specific antibody-secreting cells in the bone marrow and peripheral blood were assessed by ELISpot at indicated timepoints as previously described in ref. 51. Multiscreen IP filter plates ELISpot 96-well plates (Millipore) were activated using 35% ethanol, washed with PBS and coated with Affinity Pure goat anti-human IgG Fc fragment-specific capture antibody (Jackson ImmunoResearch, cat# 109-005-008) at 1 μg/mL in PBS. After an overnight incubation at 4°C, plates were washed with PBS and blocked using complete medium. Freshly isolated PBMCs or bone marrow mononuclear cells were plated in serial dilution and incubated overnight at 37°C and 5% CO2. After incubation, plates were washed with PBS-T and biotinylated protein probes were added for 1.5 h (0.25 μg/mL goat anti-human IgG Fc fragment-specific antibody (Jackson ImmunoResearch, cat# 109-065-008), 1 μg/mL S-2P/RBD/ovalbumin (OVA)). After another wash with PBS-T, streptavidin-conjugated alkaline phosphatase (Mabtech) was added at 1:1000 dilution for 30 min. The plates were developed using nitro blue tetrazolium 5-bromo-4-chloro-3’ indolyphosphate substrate (Mabtech) for 7 min. The spots were counted using an AID ELISpot reader (Autoimmun Diagnostika) and background subtraction was performed based on OVA wells.

Memory T cell recall assay

Frequencies of S-specific memory T cells in blood and BAL were assessed using a re-stimulation assay as described previously using 2 μg/mL PepMix SARS-CoV-2 Spike overlapping peptides in DMSO (15mers with 11 amino acid overlap, JPT Peptide Technologies)51. The antibody panel used for surface staining was: CD103 FITC (2G5, Beckman Coulter, cat# B49222, 1:50 dilution), CCR7 BV421 (G043H7, Biolegend, cat# 353208, 1:50 dilution), CD8a BV711 (RPA-T8, Biolegend, cat# 301044, 1:80 dilution), CD4 PE-Cy5.5 (S3.5, Invitrogen, cat# MHCD0418, 1:80 dilution) and CD45RA BV650 (5H9, BD, cat# 740608, 1:500 dilution), and the intracellular proteins were stained using: IL-21 AF647 (3A3-N2.1, BD, cat# 560493, 1:20 dilution), IL-13 PE (JES10-5A2, BD, cat# 559328, 1:33 dilution), IL-2 BV605 (MQ1-17H12, BD, cat# 564165, 1:50 dilution), IL-17A BV785 (BL168, Biolegend, cat# 512338, 1:67 dilution), CD69 ECD (TP.1.55.3, Beckman Coulter, cat# 6607110, 1:67 dilution), CD3 APC-Cy7 (SP34-2, BD, cat# 557757, 1:200 dilution) and IFNγ AF700 (B27, Biolegend, cat# 506516, 1:200 dilution). Acquisition was performed using BD LSRFortessa cell analyzer, and the data were analyzed using FlowJo software v.10.7.1 (FlowJo).

Quantification of subgenomic RNA after challenge

Viral loads in the respiratory tract, evaluated as a copies of subgenomic (sg)E and sgN RNA in the sample, were measured as previously reported27,53. BAL fluid and nasal swabs were stored in RNAzol BD (Molecular Research Center) and PBS until use, respectively. Total RNA was extracted with the RNAzol BD column kit (Molecular Research Center). TaqMan Fast Virus 1-Step Master Mix (Applied Biosystems), gene-specific primers (sgLeadSARSCoV2_F: 5′-CGATCTCTTGTAGATCTGTTCTC-3′, E_Sarbeco_R: 5′-ATATTGCAGCAGTACGCACACA-3′, wtN_R: 5′-GGTGAACCAAGACGCAGTAT-3′) and probes (E_Sarbeco_P: 5′-FAM-ACACTAGCCATCCTTACTGCGCTTCG-BHQ1-3′, wtN_P: 5′-FAM-TAACCAGAATGGAGAACGCAGTGGG-BHQ1-3′) were used for the RT-qPCR reaction in 384-well plates (Bio-Rad). The total volume of the reaction was 20 μL, and the sample volume was 3.33 μL. CFX384 Touch Real-Time PCR Detection System (Bio-Rad) was used for amplifications. The lower limit of quantification was 50 copies/reaction.

Statistical analyses

Statistical analyses were performed using non-parametric tests due to small sample size and non-normal data distribution. Paired analyses were used where applicable. All statistical tests were two-tailed and are indicated in the respective figure legends. For comparison of paired samples between two timepoints, Wilcoxon test was used. For comparison of paired samples between three or more timepoints, Friedman’s test with Dunn’s post hoc correction was used. Repeated measures two-way ANOVA with the Geisser-Greenhouse correction and Tukey’s multiple comparisons test was used to compare neutralization of different SARS-CoV-2 variants of longitudinally collected serum samples. Kruskal-Wallis test was used for comparison of non-paired data, with Dunn’s post hoc correction when more than one comparison was performed. Correlation was assessed using Spearman correlation. Kaplan-Meier analysis to evaluate the time until virus clearance between different experimental groups was performed using log-rank test with Bonferroni correction. All statistical tests were performed in Prism v9.4.1.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.