Potent neutralization of SARS-CoV-2 variants by RBD nanoparticle and prefusion-stabilized spike immunogens

Abstract

We previously described a two-component protein nanoparticle vaccine platform that displays 60 copies of the SARS-CoV-2 spike protein RBD (RBD-NP). The vaccine, when adjuvanted with AS03, was shown to elicit robust neutralizing antibody and CD4 T cell responses in Phase I/II clinical trials, met its primary co-endpoints in a Phase III trial, and has been licensed by multiple regulatory authorities under the brand name SKYCovione^TM. Here we characterize the biophysical properties, stability, antigenicity, and immunogenicity of RBD-NP immunogens incorporating mutations from the B.1.351 (β) and P.1 (γ) variants of concern (VOCs) that emerged in 2020. We also show that the RBD-NP platform can be adapted to the Omicron strains BA.5 and XBB.1.5. We compare β and γ variant and E484K point mutant nanoparticle immunogens to the nanoparticle displaying the Wu-1 RBD, as well as to soluble prefusion-stabilized (HexaPro) spike trimers harboring VOC-derived mutations. We find the properties of immunogens based on different SARS-CoV-2 variants can differ substantially, which could affect the viability of variant vaccine development. Introducing stabilizing mutations in the linoleic acid binding site of the RBD-NPs resulted in increased physical stability compared to versions lacking the stabilizing mutations without deleteriously affecting immunogenicity. The RBD-NP immunogens and HexaPro trimers, as well as combinations of VOC-based immunogens, elicited comparable levels of neutralizing antibodies against distinct VOCs. Our results demonstrate that RBD-NP-based vaccines can elicit neutralizing antibody responses against SARS-CoV-2 variants and can be rapidly designed and stabilized, demonstrating the potential of two-component RBD-NPs as a platform for the development of broadly protective coronavirus vaccines.

Fig. 1. SARS-CoV-2 variant vaccine design and characterization.

A Graphical representation of the production and characterization of SARS-CoV-2 RBD-NP vaccines, created with BioRender.com. B Molecular surface representation of the SARS-CoV-2 HexaPro trimer in the prefusion-stabilized conformation in gray (PDB: 6XKL). N-linked glycans in dark blue. A single RBD is boxed and expanded with bound hACE2 receptor (PDB: 6M0J) and CR3022 Fab (PDB: 6W41) shown for reference. Pink spheres indicate β and γ RBD mutations (K417N/T, E484K, and N501Y). Purple spheres indicate stabilizing Rpk9 mutations (Y365F, F392W, and V395I). C Representative DLS of 8 monovalent RBD-NPs with and without Rpk9 mutations. D Structural models of 8 monovalent RBD-NPs and 3 prefusion-stabilized HexaPro trimers alongside representative nsEM micrographs of each immunogen. All graphical representations of proteins made using ChimeraX.

Fig. 2. Antigenic characterization of variant antigens.

Left, Models of RBD with bound (A) hACE2 receptor, (B) LY-CoV555 (PDB: 7KMG), and C) CR3022 Fab, depicted as in Fig. 1. Right, Binding of (A) monomeric hACE2, (B) LY-CoV555, and (C) CR3022 Fab to immobilized VOC-RBD-I53-50A trimers. Data are shown in colors and global fits as black lines, with the K_D of each interaction indicated. D Left, Molecular surface representation of SARS-CoV-2 HexaPro prefusion-stabilized trimer (PDB: 6XKL) with RBDs and NTDs highlighted in light blue and green, respectively. Right, ELISA binding titers to hACE2 receptor, RBD-specific IgG (LY-CoV555 and CR3022), and NTD-specific IgG (S2L28). All graphical representations of proteins made using ChimeraX. Error bars represent standard error of mean from three replicates.

Fig. 3. Stabilizing effects of Rpk9 mutations in variant RBD components and NPs.

A Representative melting curves of RBD-I53-50A trimers with and without Rpk9 mutations. Melting temperatures (T_m) are indicated. B Hydrogen/deuterium exchange mass spectrometry of selected RBD peptides. Left, RBD structure with observed peptides numbered and regions where the VOC-related and Rpk9 mutations reside represented in blue and pink, respectively. C SEC chromatograms of RBD-NPs in three different buffers. MS: 50 mM Tris pH 7.4, 185 mM NaCl, 100 mM arginine-HCl, 4.5% v/v glycerol, 0.75% w/v CHAPS; TAG: 50 mM Tris pH 8, 150 mM NaCl, 100 mM arginine-HCl, 5% v/v glycerol; TBS: 50 mM Tris pH 8, 150 mM NaCl. Major peak at ~10.5 mL represents assembled NP and minor peak ~17 mL represents excess VOC-RBD-I53-50A. Black triangle on the x axis represents the Wu-1-RBD-NP peak in MS. D Representative aggregation profiles of purified RBD-NPs, with temperature (T_agg) indicated.

Fig. 4. Shelf-life stability of SARS-CoV-2 variant RBD-NPs.

A Summary of SDS-PAGE and nsEM stability data over 4 weeks. N/A, not assessed. Cocktail formulations include Wu-1-RBD-NP formulated with β-RBD-NP, βRpk9-RBD-NP, γ-RBD-NP, or γRpk9-RBD-NP. Detailed RBD-NP construct information can be found in Supplementary Table 2. B Binding to hACE2-Fc over 4 weeks. Immunogens were analyzed for hACE2-Fc binding by BLI after storage at the various temperatures, normalizing against a <-70°C reference sample at each time point. C Summary of UV/vis stability data over 4 weeks, using 320/280 nm absorbance ratio as a measure of particulate scattering. D Hydrodynamic diameter and polydispersity measured by DLS over 4 weeks. Raw data provided in Supplementary File 1.

Fig. 5. Variant RBD-NP vaccines elicit high neutralizing Ab titers to SARS-CoV-2 variants in mice.

A, C Immunization scheme of VOC-RBD-NP, VOC-HexaPro, and the NP scaffold (Bare-NP) vaccines in Balb/c mice. Structural models of VOC-RBD-NPs, VOC-HexaPro, and Bare-NP vaccines. B, D Two weeks post boost neutralizing Ab titers for monovalent (RBD-NP), mosaics (m), and cocktails (c) of VOC with and without Rpk9 (+/−) mutations, Bare-NP, WT-VOC-HexaPro, and cocktails (c) of WT-VOC-HexaPro against various VOC pseudoviruses. Background neutralization is denoted by a dotted line. E Immunization scheme of VOC-RBD-NP vaccines in Darwin mice. Structural models of VOC-RBD-NP vaccines. F 0.5 week post IV boost neutralizing Ab titers for VOC-RBD-NP vaccines against various VOC pseudoviruses. Cocktail formulations including Wu-1-RBD-NP were formulated with βRpk9-RBD-NP or γRpk9-RBD-NP. All graphical representations of proteins made using ChimeraX. Detailed RBD-NP immunogen information can be found in Supplementary Table 2. Kruskal–Wallis tests were performed to compare two groups to determine whether they were statistically different. Significance is indicated with stars: ^∗p < 0.05; ^∗∗p < 0.01; and non-significant differences are not shown.