A class II MHC-targeted vaccine elicits immunity against SARS-CoV-2 and its variants

We have characterized a single-domain antibody fragment that binds class II MHC antigens (VHH_MHCII) with nanomolar affinity. Immunization of mice with an influenza A virus (IAV) HA2 antigen, conjugated to VHH_MHCII, protected mice against a lethal IAV challenge (16). To apply this vaccine platform to SARS-CoV-2, we generated a recombinant protein that consists of a fusion between VHH_MHCII and the SARS-CoV-2 RBD (Fig. 1A). VHH_MHCII-Spike_RBD was expressed in Expi293 cells and purified by means of its C-terminal His₆-tag followed by size-exclusion chromatography. The identity of the product was confirmed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE) (Fig. 1B). A 200-mL culture of Expi293 cell supernatant yielded ∼20 mg of recombinant protein.

To examine the immunogenicity of VHH_MHCII-Spike_RBD, we used a two-dose immunization regimen. Preimmune serum was collected at −17 d and C57BL/6J mice (H-2^b haplotype) were primed (day −14) intraperitoneally (i.p.) with adjuvanted (poly dIdC and anti-CD40 monoclonal antibody) VHH_MHCII-Spike_RBD (20 μg), with an equimolar amount of adjuvanted Spike_RBD (13.5 μg), or with adjuvant alone. Mice were boosted with each corresponding preparation 14 d postprime (day 0) (Fig. 1C and SI Appendix, Fig. S1A). Serum was collected from all animals 14 d postboost (day 14). IgG titers were determined by ELISA against recombinant SARS-CoV-2 Spike_RBD (Wuhan Hu-1 strain) (Fig. 1D and SI Appendix, Fig. S1B). Immunization with two doses of VHH_MHCII-Spike_RBD elicited high levels of anti-Spike_RBD antibodies in all animals, reaching mean endpoint titers in excess of 1/23,600,000, an approximate 34,000-fold increase over mice immunized with two doses of Spike_RBD. Analysis of immunoglobulin subclasses showed evidence of class switching in mice immunized with VHH_MHCII-Spike_RBD, judged by the levels of IgA, IgG1, and IgG2b detected at day 14 (Fig. 1E and SI Appendix, Fig. S1C). Mean endpoint titers were higher in mice immunized with VHH_MHCII-Spike_RBD than in mice immunized with Spike_RBD, with mean endpoint titers reaching >1/7,300 (IgM; fourfold increase over Spike_RBD), 1/8,900 (IgA; 66-fold increase over Spike_RBD), 1/3,160,000 (IgG1; 6,870-fold increase over Spike_RBD), and ∼1/90,000 (IgG2b; 87-fold increase over Spike_RBD). Immunization with VHH_MHCII-Spike_RBD induced high levels of IgA, which suggests improved mucosal protection in immunized animals. Immunization of mice with two doses of VHH_MHCII-Spike_RBD thus induces a robust humoral immune response, significantly higher than that induced by the Spike_RBD alone.

Since VHH_MHCII-Spike_RBD binds class II MHC antigens encoded by the I-A locus on target APCs, we immunized a different inbred mouse strain (BALB/cJ; H-2^d haplotype) to confirm efficacy across different MHC haplotypes. Immunization with two doses of VHH_MHCII-Spike_RBD led to comparable levels of total IgG in C57BL/6J (H-2^b) and BALB/cJ (H-2^d) mice (SI Appendix, Fig. S2 A and B). Immunoglobulin class switching was evident in both mouse strains, and mean Spike_RBD-specific IgM, IgA, IgG1, and IgG2b titers were comparable (SI Appendix, Fig. S2C).

We next measured total and subclass immunoglobulin levels at days 7, 14, and 21 in mice immunized with one or two doses of adjuvanted VHH_MHCII-Spike_RBD or control (SI Appendix, Fig. S3A). Mice immunized with two doses of VHH_MHCII-Spike_RBD quickly achieved peak IgG titers on day 7, with levels persisting until at least day 21 (SI Appendix, Fig. S3A), while animals immunized with a single dose produced lower levels of total IgG and other subclass immunoglobulins (SI Appendix, Fig. S3B).

Because SARS-CoV-2 variants with mutations in S that enable partial immune escape have emerged, we next examined whether serum from mice immunized with VHH_MHCII-Spike_RBD recognized recombinant Spike_RBD with the K417T, E484K, and N501Y mutations. These mutations are found individually in many S variants and in combination in the P.1 (Gamma or Brazil) variant and partially in the B.1.351 (Beta or South Africa) variant which has K417N instead of K417T (17). Immunization with two doses of adjuvanted wild-type Spike_RBD induced antibodies (day 14) with a low capacity to bind the triple-mutant Spike_RBD, while immunization with two doses of adjuvanted VHH_MHCII-Spike_RBD elicited antibodies still capable of recognizing the mutant Spike_RBD to high titers (mean endpoint titer of 1/178,000) (Fig. 1F and SI Appendix, Fig. S1D). Two doses of VHH_MHCII-Spike_RBD were required to maintain high titers of IgG against the mutant Spike_RBD (SI Appendix, Fig. S3C).

We next measured the levels of neutralizing antibodies induced by immunization with two doses of adjuvanted VHH_MHCII-Spike_RBD by using replication-competent vesicular stomatitis viruses that express eGFP and variants of the SARS-CoV-2 spike (VSV-SARS-CoV-2) in place of their native glycoprotein (18, 19). Mice immunized with two doses of adjuvant failed to neutralize VSV-SARS-CoV-2 expressing the Wuhan Hu-1+D614G spike, as well as those carrying the spikes of B.1.1.7 (Alpha or United Kingdom), P.1, and B.1.351 variants (Fig. 1G) (20). Alternatively, immunization with two doses of adjuvanted VHH_MHCII-Spike_RBD induced high-titer neutralizing antibodies against VSV-SARS-CoV-2 expressing the spikes of Wuhan Hu-1+D614G (mean effective concentration, 50% [EC₅₀] titers of 1/2,426), B.1.1.7 (1/937), P.1 (1/250), and B.1.351 (1/488). Sera from mice immunized with two doses of VHH_MHCII-Spike_RBD neutralized a clinical isolate of SARS-CoV-2 strain NL/2020 (mean EC₅₀ titers of 1/499), as measured by a quantitative RT-PCR assay (SI Appendix, Fig. S4A). Neutralizing titers were comparable to that of a monoclonal neutralizing antibody, 47D11 (1/640 to 1/1280 dilution range, 0.05 μg/mL to 0.025 μg/mL) (Fig. 1H and SI Appendix, Fig. S4B) (21). While neutralizing titers were impacted by mutations in the RBD of circulating variants, mean EC₅₀ titers remain high. Regardless, high neutralizing titers against the Wuhan Hu-1+D614G after immunization with a matched RBD suggests that VHH_MHCII-Spike_RBD carrying mutations found in the RBDs of circulating variants will significantly improve neutralizing titers against those variants. Other variants remain to be tested.

Because cellular immunity, particularly that exerted by T cells, is important for protection against and memory of SARS-CoV-2 infection, we next examined the T cell response in animals immunized with a single dose of adjuvanted Spike_RBD or VHH_MHCII-Spike_RBD (Fig. 2A) (5). Splenocytes harvested 7 d postimmunization were stimulated with a library of Spike_RBD 15-mer peptides with 11-residue overlap and evaluated by an ELISpot assay (Fig. 2B and Table 1). While mouse class I MHC products prefer peptides in the eight to nine residue range, we chose to use synthetic 15-mers, as they are known to contain shorter peptides that serve as class I MHC ligands or can be processed to yield proper class I MHC ligands (22, 23). The use of longer peptides would enable identification of both class I and class II restricted epitopes. We observed robust production of IFNγ in splenocytes from mice immunized with VHH_MHCII-Spike_RBD and, to a lesser extent, from mice immunized with Spike_RBD (Fig. 2C). IFNγ production was highest when splenocytes were stimulated with peptides 42, 47, 48, 49, and 50, indicative of at least two unique stimulatory regions. Peptide 42 contains residue E484, which is frequently mutated to K or Q in SARS-CoV-2 variants, and S494, which is sometimes mutated to P in the B.1.1.7 lineage. Peptides 47 to 50, which correspond to spike residues 503 to 529, are conserved among all Centers for Disease Control–designated variants (17). Indeed, coculture of splenocytes from VHH_MHCII-Spike_RBD–immunized mice with Spike_RBD peptides 42 and 47 to 50 resulted in secretion not only of high levels of IFNγ but also high levels of the proinflammatory cytokines IL-6 and TNFα. These levels are higher than those induced by immunization with Spike_RBD and indicate proliferation of memory CD4 and CD8 T cells (Fig. 2D) (24, 25). We did not observe high levels of IL-2 secretion. Cellular immunity elicited by immunization with VHH_MHCII-Spike_RBD is strong and may persist upon infection with circulating variants.

To distinguish between CD4 and CD8 T cells as the dominant cytokine-producing cell type, we performed intracellular cytokine staining. A significant proportion of the inflammatory cytokines are derived from the CD8 T cell compartment (Fig. 2 E and F and SI Appendix, Fig. S5). Recombinant VHH_MHCII-antigen fusion proteins are therefore subject to cross-presentation and can induce an efficacious CD8 T cell response against Spike_RBD. The presence of a strong humoral immune response that includes class switching likewise implies the contribution by CD4 T cells (Fig. 2F and SI Appendix, Fig. S5).

We next compared different routes of administration, including i.p., intramuscular (i.m.), and intranasal (i.n.) routes, to elicit an immune response. Mice were primed with adjuvant only (i.p.) or received adjuvanted VHH_MHCII-Spike_RBD (i.p., i.m., or i.n.) and boosted 14 d later with a dose of the corresponding vaccine and route of administration (Fig. 3A and SI Appendix, Fig. S6A). Mean total Spike_RBD-specific IgG in the blood, as well as mean IgM, IgG1, and IgG2b levels, were comparable across the three different routes of administration, although i.p. and i.m. administration led to an increase in mean IgA levels (Fig. 3B and SI Appendix, Fig. S6A). While i.n. administration did not induce significant levels of IgA in the blood, it remains to be determined whether these animals produced significant levels of IgA in the respiratory tract.

Because the SARS-CoV-2 vaccines in use require specific formulations and cold storage conditions, we determined whether recombinant VHH_MHCII-Spike_RBD protein would retain its efficacy at inducing an antibody response upon storage at ambient temperature or lyophilization. We immunized mice with two doses of adjuvant only, adjuvanted VHH_MHCII-Spike_RBD stored at −20 °C, 4 °C, or 25 °C for 1 wk, or lyophilized and resuspended VHH_MHCII-Spike_RBD (Fig. 3A and SI Appendix, Fig. S6B). Mean total serum IgG, IgA, IgG1, IgG2b, and IgM levels between the different conditions were comparable, indicating that neither temperature of storage nor storage in liquid form is required for the vaccine to retain its efficacy (Fig. 3C).

Because the ability to produce a robust and durable immune response can decrease with age, we also determined whether immunization of aged mice with VHH_MHCII-Spike_RBD would elicit a strong immune response (26). We immunized 8- to 12-wk-old mice with two doses of adjuvant only or adjuvanted VHH_MHCII-Spike_RBD and 72-wk-old [equivalent to humans aged 56 y to 69 y (27)] with two doses of VHH_MHCII-Spike_RBD (Fig. 3A and SI Appendix, Fig. S6C). While switching to some immunoglobulin subclasses (IgA and IgG1) decreased in aged mice, consistent with diminished T cell help, mean total IgG levels in the blood were unchanged when compared to 8- to 12-wk-old mice (Fig. 3D). This may be due, in part, to high levels of IgG2b in aged mice and slightly higher levels of IgM (SI Appendix, Fig. S6C).

Because the anti-MHC class II nanobody used for these experiments recognizes murine antigens independent of haplotype, we also generated a version of the vaccine that could be applied in a clinical setting. We used HLA-DR4-IE–transgenic C57BL/6 IAb^null mice, which lack wild-type murine MHC class II products and instead express transgenic hybrid MHC class II molecules composed of the peptide-binding portion of human HLA-DR4 and the membrane-proximal domains of mouse I-E (DR4-IE). We used a previously characterized nanobody (VHH_hMHCII) that recognizes nearly all allelic variants of human class II MHC molecules (HLA-DR specific, with the exception of HLA-DR-03*01) (VHH_hMHCII-Spike_RBD) (28). Flow cytometry on splenocytes from HLA-DR4-IE–transgenic C57BL/6 IAb^null mice confirmed that VHH_hMHCII recognizes these hybrid DR4-IE molecules (Fig. 4A). We then generated a genetic fusion between VHH_hMHCII and either the Wuhan Hu-1 or B.1.1.7+E484K SARS-CoV-2 Spike_RBD, as an illustration of rapid and straightforward vaccine adjustment in anticipation of emerging mutations in the RBD. Both constructs expressed well in mammalian cells. A 200-mL culture of Expi293 cell supernatant yielded 15 and 12.5 mg of VHH_hMHCII-Spike_RBD (Wuhan Hu-1) and VHH_hMHCII-Spike_RBD (B.1.1.7+E484K), respectively (Fig. 4B).

To examine the immunogenicity of this vaccine, we immunized HLA-DR4-IE–transgenic C57BL/6 IAb^null mice by i.p. administration with two doses of adjuvant only or 20 μg of adjuvanted VHH_hMHCII-Spike_RBD (Wuhan Hu-1 RBD) (Fig. 4C and SI Appendix, Fig. S7A). Immunization with VHH_hMHCII-Spike_RBD elicited high-titer anti-Spike_RBD antibodies, reaching mean endpoint titers of 1/1,296,724 (total IgG), 1/17,853 (IgM), 1/208 (IgA), 1/349,669 (IgG1), and 1/5,707 (IgG2b) (Fig. 4D and SI Appendix, Fig. S7 B and C). Total serum IgG retained the ability to bind mutant, recombinant Spike_RBD with the K417T, E484K, and N501Y mutations found in the P.1 variant at mean endpoint titers of 1/17,853 (Fig. 4E and SI Appendix, Fig. S7D). We also identified three peptides, 47, 48, and 49, that elicited strong IFNγ responses in splenocytes after immunization with a single dose (Fig. 4F). Not surprisingly, CD8 T cells are the dominant cell type implicated in IFNγ production upon peptide stimulation (Fig. 4G), as the HLA-DR4-IE–transgenic C57BL/6 IAb^null mice share the H-2^b encoded class I MHC molecules.

Expi293F cells were maintained in humidified, shaking incubators at 37 °C, 8% CO₂ in Expi293 Expression Media (ThermoFisher Scientific). BSRT7/5, Vero-CCL81, Vero E6-TMPRSS2, and Vero-hACE2-TMPRSS2 cells were maintained in humidified incubators at 34 °C or 37 °C and 5% CO₂ in Dulbecco’s modified Eagle’s medium (DMEM) (Corning) supplemented with glucose, L-glutamine, sodium pyruvate, and 10% fetal bovine serum (FBS). Vero E6, Vero E6-TMPRSS2, and Vero-hACE2-TMPRSS2 cells were described previously (31). Vero E6 cells (CRL-1586, ATCC) were grown in DMEM (Gibco) supplemented with 5% fetal calf serum (Sigma), 100 U/mL penicillin (Gibco), 100 µg/mL streptomycin (Gibco), and 2 mM L-glutamine (Lonza). All commercially available antibodies utilized in this study are indicated in Table 2.

Generation of replication-competent VSV-SARS-CoV-2 (Wuhan Hu-1, GenBank MN908947.3) has been described (19). Additional VSV recombinants expressing eGFP and SARS-CoV-2 spike variants were prepared as follows. Briefly, spike genes of Wuhan Hu-1+D614G, B.1.1.7 (GenBank OU117158.1), B.1.351 (GenBank MZ212516.1), and P.1 (GISAID EPI_ISL_804823) were truncated to remove the C-terminal 63 nucleotides and were then cloned into an infectious molecular clone (complementary DNA [cDNA]) of VSV-eGFP in place of the native VSV G gene. To rescue the recombinant VSVs, BSRT7/5 cells were infected with vaccinia virus encoding the bacteriophage T7 RNA polymerase (vTF7-3) and subsequently transfected with T7-driven support plasmids encoding VSV N, P, L, and G, as well as the infectious molecular cDNAs. Cell supernatants were harvested ∼72 h postinfection, clarified by centrifugation (5 min and 1,000 × g), and filtered through a 0.22-μm filter. Rescue supernatants were plaque purified on Vero-CCL81, Vero E6-TMPRSS2, or Vero-hACE2-TMPRSS2 cells in the presence of 25 μg/mL cytosine arabinoside (AraC). Plaques in agarose plugs were grown on Vero-CCL81, Vero E6-TMPRSS2, or Vero-hACE2-TMPRSS2 cells also in the presence of 25 μg/mL AraC to generate P1 stocks. Working stocks were generated on Vero-CCL81 or Vero E6-TMPRSS2 cells at 34 °C. SARS-CoV-2 strain NL/2020 (EVAg, Ref-SKU 010V-03903) was propagated and titrated on Vero E6 cells using the tissue culture infective dose 50 (TCID50) endpoint dilution method.

Sequences encoding the Spike_RBD and VHH_MHCII-Spike_RBD were synthesized (Integrated DNA Technologies) as double-stranded DNA. Inserts were assembled in the pVRC vector (a gift from Stephen Harrison, Harvard Medical School). Constructs were transfected into Expi293F cells (Thermo Fischer Scientific) using Polyethyleneimine “Max” (Polysciences). Cell cultures were maintained in Expi293 Media (Thermo Fischer Scientific) at 37 °C for 4 d following transfection. Proteins were harvested by centrifugation at 5,000 × g for 30 min at 4 °C, followed by affinity chromatography with HisPur Ni-NTA Resin (Thermo Fischer Scientific) and size exclusion chromatography with a Hi-Load 16/600 S75 column (Cytivia). Protein samples (2 μg each) were prepared by boiling for 5 min in sample buffer containing 1% (wt/vol) SDS and 1% (vol/vol) BME. Samples were analyzed on 10% or 15% SDS/PAGE. Gels were stained with Instant Blue (Abcam) and destained with ddH₂O.

All animals were housed in the animal facility of Boston Children’s Hospital (BCH) and were maintained according to protocols approved by the BCH Committee on Animal Care. C57BL/6J (CD45.2+) and BALB/c mice were either purchased from the Jackson Laboratory or bred in house. DR4-IE transgenic mice were purchased from Taconic. Only female mice aged 8 to 12 wk were used in this study unless indicated otherwise.

Serum samples were collected on the indicated days and stored in BD Vacutainers; 96-well plates were coated with 2 μg/mL of either recombinant Spike_RBD or Spike_RBD (K417T, E484K, N501Y) proteins in phosphate-buffered saline (PBS) overnight at 4 °C and incubated in blocking buffer (0.05% Tween20 + 2% bovine serum albumin in PBS). Plates were incubated with diluted serum samples for 2 h at room temperature. Plates were then washed four times with PBS and incubated with goat anti-mouse IgG-HRP, anti-mouse IgM-HRP, anti-mouse IgG1-HRP, or IgG2b-HRP (SouthernBiotech) at 1:10,000 or with IgA-HRP at 1:2,000 in blocking buffer for 1 h. Plates were developed with 3,3',5,5'-tetramethylbenzidine liquid substrate reagent (Sigma). The reaction was stopped with 1 N HCl, and absorbance was read at 450 nm.

Serum samples were heat inactivated at 56 °C for 30 min. Neutralization assays were performed similarly to what has been described (19, 20). Briefly, threefold serial dilutions of sera, starting with a 1:20 dilution, were performed in 384-well plates and were incubated with 10⁶ plaque-forming units (pfu) of VSV-SARS-CoV-2 expressing eGFP and the Spike of Wuhan Hu-1+D614G, B.1.1.7, P.1, or B.1.351 for 1 h at 37 °C. Vero E6 cells then were added to the human serum–virus complexes in 384-well plates at 3 × 10³ cells per well and incubated at 37 °C for 16 h. Cells were fixed at room temperature in 4% formaldehyde and then rinsed with PBS. Cells were stained at room temperature with NucRed Live 647 (Invitrogen) for 30 min. Images were acquired using an InCell 6500 confocal imager (Cytiva) to visualize nuclei and infected cells (4× objective, 1 field per well). Images were segmented using InCarta (Cytiva). Infected cells were identified by comparing them to the uninfected threshold in Spotfire (Tibco). Cells were gated based on nuclear parameters.

Additional VSV neutralization assays were performed similarly to what has been described (19). Briefly, serial dilutions of serum samples were incubated with ∼10² pfu of VSV-SARS-CoV-2 for 1 h at 37 °C. Antibody–virus complexes were then added to Vero CCL-81 cells in black 96-well plates for 7.5 h at 37 °C. Cells were then fixed in 2% (vol/vol) formaldehyde (Millipore Sigma) containing 10 mg/mL Hoechst 33342 nuclear stain (Invitrogen) for 45 min at room temperature. Following incubation, nuclear stain and fixative were removed and replaced with PBS. Images were acquired with the InCell 2000 Analyzer (GE Healthcare) automated microscope in both the DAPI and fluorescein isothiocyanate (FITC) channels to visualize nuclei and infected cells (i.e., eGFP-positive cells), respectively (4× objective, four nonoverlapping fields per well, covering the entire well). Images were analyzed using the InCell Investigator Developer Toolbox v1.9 (GE Healthcare). Nuclei and GFP-positive cells were identified and segmented from their respective images using the Object Segmentation function, which uses a variation of the top-hat approach for initial segmentation followed by a binarization operation to define the target sets. Target linking was then performed to determine the total number of GFP-positive nuclei per well. Data were processed using Prism software (GraphPad Prism 9.0).

Neutralization assays using SARS-CoV-2 were performed similarly as described (32). Briefly, Vero E6 cells were seeded 1 d prior to infection at 12,000 cells per well in a 96-well tissue culture plate. Serum samples were heat inactivated (56 °C for 30 min) and serially diluted (dilution range of 1:10 to 1:1,280) in 60 μL of DMEM supplemented with penicillin, streptomycin, L-glutamine, and 2% fetal calf serum (FCS). The diluted sera were mixed with equal volumes of 120 TCID50/60 µL of SARS-CoV-2 strain NL/2020 and incubated for 1 h at 37 °C. The virus–serum mixtures were then applied onto the previously seeded Vero E6 cells and incubated at 37 °C for 2.5 d. Cells infected with 120 TCID50/60 µL of SARS-CoV-2 strain/NL/2020 were used as a positive control. As a neutralization control, we added mAb 47D11 group, which is a human monoclonal antibody that can neutralize SARS-CoV-2. In brief, mAB 47D11 (3.14 mg/mL) was prediluted 1:100 (to 31.4 ug/mL) in DMEM (+2% FCS, PenStrep, glutamine) followed by the serial dilution indicated in the figures. This mAb 47D11 was kindly provided by B. J. Bosch, Utrecht University, Utrecht, The Netherlands. At 2.5 d postinfection, supernatant was harvested and heat inactivated at 75 °C for 30 min. Real-time RT-PCR of the SARS-CoV-2 E gene was performed on the supernatant samples as described previously (33). All work with live SARS-CoV-2 was performed in a biosafety level 3 facility in the University Medical Center Utrecht.

The RBD of the spike protein was divided into 53 overlapping 15-mer peptides with an overlap of 11 amino acids (Table 1). The peptides were obtained from GenScript as lyophilized powder. The peptides were resuspended at a concentration of 4 mg/mL in water (with up to 5% dimethyl sulfoxide) and stored at −20 °C.

IFNγ ELISpot assays (BD ELISPOT Mouse IFNγ ELISPOT Set, BD Biosciences) were performed according to the manufacturer’s instructions. Briefly, 96-well ELISpot plates were coated with an IFNγ capture antibody (BD Biosciences, 51-2525K) in PBS overnight at 4 °C. Plates were then blocked with complete growth medium for 2 h at room temperature. Wells were then emptied, and 1 × 10⁶ splenocytes from immunized mice were added to the plates. Cells were arrayed in the presence or absence of 15-mer peptides with 11-residue overlaps derived from the SARS-CoV-2 Spike_RBD (10 μg/mL) in 200 μL of complete medium and incubated overnight at 37 °C. Plates were then washed and incubated with a biotinylated IFNγ detection antibody (BD Biosciences, 51-1818KZ) for 2 h at room temperature and incubated with streptavidin-horseradish peroxidase (BD Biosciences) for 1 h at room temperature. Plates were developed with 3-amino-9-ethyl-carbazole substrate (BD ELISPOT AEC Substrate Set) for 5 min to 30 min and dried overnight. Spots were enumerated using the KS ELISpot analysis system (Immunospot).

Three days poststimulation with Spike_RBD 15-mer peptides, cell supernatants were collected and used for ELISA to measure IFNγ, IL-6, IL-2, and TNFα production. IFNγ, IL-2, IL-6, and TNFα were measured using the BD OptEIA Mouse IFN-γ ELISA Set (BD Biosciences, 555138), BD OptEIA Mouse IL-6 ELISA Set (BD Biosciences, 555240), BD OptEIA Mouse IL-2 ELISA Set (BD Biosciences, 555148), and Mouse TNFα Uncoated ELISA kit (Invitrogen, 88-7324-22) per manufacturer's protocol.

Cells harvested from excised spleens were dispersed into RPMI 1640 (Gibco) through a 40-μm cell strainer using the back of a 1-mL syringe plunger. Cell mixtures were subjected to hypotonic lysis (NH₄Cl) to remove red blood cells, washed twice in flow cytometry buffer (2 mM (ethylenedinitrilo)tetraacetic acid and 1% FBS in PBS) and resuspended in flow cytometry buffer containing the corresponding fluorescent dye-conjugated antibodies. Staining steps were carried out at 1:100 dilutions in the presence of Fc block (Biolegend) for 30 min at 4 °C in the dark. Samples were washed twice with fluorescence-activated cell sorter (FACS) buffer before further analysis. All flow data were acquired on a FACS Fortessa flow cytometer (BD Biosciences) and analyzed using FlowJo software (Tree Star).

VHH_MHCII-Spike_RBD combined with adjuvant in PBS were flash frozen in a 2-mL Eppendorf tube. The tube was opened, placed in a 300-mL flask, and lyophilized overnight using a Freezone 4.5 Plus Freeze Dryer (Labconco). After overnight lyophilization, the resulting powder was left at room temperature for 1 wk before immunization.

All data represent at least two independent experiments. All statistical analyses were performed using Prism 6. Statistical methods used are indicated in the corresponding legend of each figure.