mRNA delivery of genetically encoded mosaic-8 pan-sarbecovirus RBD vaccines

Abstract

Global health remains threatened by spillovers of zoonotic SARS-like betacoronaviruses (sarbecoviruses) that could be mitigated by a pan-sarbecovirus vaccine¹. We described elicitation of potently neutralizing and cross-reactive anti-sarbecovirus antibodies by mosaic-8 nanoparticles (NPs) displaying eight different sarbecovirus spike receptor-binding domains (RBDs) as 60 copies of eight individual RBDs^2-6 (mosaic-8 RBD-NPs) or 30 copies of two "quartets," each presenting four tandemly-arranged RBDs⁷ (dual quartet RBD-NPs). To facilitate manufacture of a broadly protective mosaic-8 vaccine, we generated membrane-bound RBD quartets that can be genetically encoded and delivered via mRNA: dual quartet RBD-mRNA and dual quartet RBD-EABR-mRNA, which utilizes ESCRT- and ALIX-binding region (EABR) technology that promotes immunogen presentation on cell surfaces and circulating enveloped virus-like particles (eVLPs)⁸. Immunization with mRNA immunogens elicited equivalent or improved binding breadths, neutralization potencies, T cell responses, and targeting of conserved RBD epitopes across sarbecoviruses, demonstrating successful conversion of protein-based mosaic-8 RBD vaccines to mRNA formats. Systems serology⁹ showed that the mRNA vaccines elicited balanced IgG subclass responses with increased Fcγ receptor-binding IgGs, consistent with potentially superior Fc effector functions. A new technique, Systems Serology-Polyclonal Epitope Mapping (SySPEM), revealed distinct IgG-subclass-specific epitope targeting signatures across mRNA and protein-based vaccine modalities. These results demonstrate successful conversion of mosaic-8 RBD-NPs to mRNA or EABR-mRNA vaccines that provide easy manufacturing and enhanced protection from future pandemic sarbecovirus outbreaks.

Keywords: Deep mutational scanning; SySPEM (systems serology-polyclonal epitope mapping); epitope mapping; mRNA and protein nanoparticle vaccines; pan-sarbecovirus vaccine; pandemic preparedness; systems serology.

Figure 1.. Protein- and mRNA-based pan-sarbecovirus immunogens.

(a) Avidity hypothesis. Left: Membrane-bound strain-specific BCRs (pale yellow) using avidity to recognize a strain-specific epitope (pale yellow triangle) on antigens attached to a homotypic NP. Middle: Strain-specific BCRs binding with only a single Fab (i.e., not using avidity) to a strain-specific epitope (triangle) on an antigen attached to a mosaic NP. Right: Cross-reactive BCRs using avidity to recognize a common epitope (blue circle) presented on different antigens attached to a mosaic NP, but not to strain-specific epitopes (triangles). Only a fraction of the 60 attached RBDs are shown for clarity. (b) Sarbecovirus phylogenetic tree (made using a Jukes-Cantor generic distance model with Geneious Prime^® 2023.1.2) calculated from amino acid sequences of RBDs aligned using Clustal Omega. Viruses with RBDs included in RBD quartets and mosaic-8 RBD-NPs are indicated with an asterisk. Scale bar = phylogenetic distance of 0.08 nucleotide substitutions per site. (c) Schematics of protein-based and mRNA-based pan-sarbecovirus immunogens (not to scale; showing only a fraction of antigens for clarity). (d) Schematics of mRNA constructs encoding RBD dual quartet immunogens used in this study.

Figure 2.. Dual quartet RBD mRNA immunogens elicit cross-reactive Abs.

ELISA and neutralization results for terminal bleed serum samples (see also Extended Data Fig. 2). (a) Immunization regimen. Left: Mice were primed at day 0, boosted at day 28, and samples were collected from a terminal bleed at day 77 or 78. Right: Colors used to identify immunization cohorts and symbols indicating a sarbecovirus antigen or pseudovirus that is matched (filled in data points; gray shading around name) or mismatched (unfilled data points; black outline around name). Sarbecovirus strain names are colored in panels b and c according to clade. (b,c) Results for RBD-binding ELISAs (panel b) and pseudovirus neutralization assays (panel c) for serum samples from day 77 or 78. Dashed horizontal lines indicate detection limits for assays. Left: Geomeans of RBD-binding ELISA half-maximal effective dilution (ED₅₀) values (panel b) or neutralization half-maximal inhibitory dilution (ID₅₀) values (panel c) across a panel of viral antigens or pseudoviruses for sera from animals in each cohort (geomeans are plotted as symbols with geometric standard deviations indicated by error bars). Mean ED₅₀ or ID₅₀ values across antigens are connected by colored lines corresponding to the immunization cohort. Right: Box and whisker plots of geomean responses (ELISA ED₅₀s in panel b; neutralization ID₅₀s in panel c) with individual data points representing geomean responses across n=10 sera to a single antigen or pseudovirus. Geomean titers were compared pairwise between immunization cohorts by Tukey’s multiple comparison test with the Geisser-Greenhouse correction (as calculated by GraphPad Prism). Significant differences between cohorts are indicated by asterisks: p<0.05 = *, p<0.01 = **, p<0.001 = ***, p<0.0001 = ****.

Figure 3.. mRNA-encoded dual quartets elicit potent T cell responses.

ELISpot assay data for SARS-2 RBD-specific IFN-γ (left) and IL-4 (right) responses of splenocytes from BALB/c mice that were immunized with the indicated immunogens (immunization regimen in Fig. 2a). Results are shown as spots per 3×10⁵ cells for individual mice (colored circles) presented as the median (bars) and standard deviation (horizontal lines). Cohorts were compared by Tukey’s multiple comparison test calculated by GraphPad Prism. Significant differences between cohorts linked by horizontal lines are indicated by asterisks: p<0.05 = *, p<0.0001 = ****.

Figure 4.. Dual quartet RBD-mRNA immunogens elicit Ab responses against conserved class 4 and class 5 RBD epitopes.

(a) Sequence conservation of 16 sarbecovirus RBDs calculated using ConSurf shown on a surface representation of SARS-2 RBD (PDB 7BZ5). Class 1, 2, 3, 4, 1/4, and 5 anti-RBD Ab epitopes^,– are outlined in dots in different colors using information from representative structures of Abs bound to SARS-2 spike or RBD (C102: PDB 7K8M; C002: PDB 7K8T, S309: PDB 7JX3; CR3022: PDB 7LOP; C118: PDB 7RKV; WRAIR-2063: PDB 8EOO). (b,c) Left: Line plots for DMS results using a SARS-2 WA1 RBD library (b) and SARS-1 library (c) from the indicated number of mice immunized with the immunogens listed above each line plot. X-axis: RBD residue number. Y-axis: sum of the Ab escape of all mutations at a site (larger numbers = more Ab escape). Each line represents one antiserum with thick lines showing the average across the n = 5 or 6 sera in each group. Lines are colored according to RBD epitopes in panel a. Right: Average site-total Ab escape calculated for results from n = 5 or 6 serum samples for the indicated RBD yeast display libraries. Mice were immunized with the immunogens listed, and results were mapped to the highlighted residues on the surface of the SARS-2 WA1 RBD (PDB 6M0J). Gray indicates no escape; a gradient of red represents increasing degree of escape. Residue numbers show sites with the most escape with font colors representing different RBD epitopes (defined in panel a; class 1/4 residues are colored with fonts corresponding to class 1 or class 4 residues). The same data are shown in Table 1 and in Extended Data Figs. 3 and 4 (line and logo plots for SARS-2 WA1 and SARS-1 RBD libraries, respectively).

Figure 5.. mRNA-encoded dual quartet RBD immunogens elicit balanced IgG subclass and potent FcγR-binding responses.

MFI = median fluorescent intensity. Binding of IgG1, IgG2a, IgG2b, IgG3, FcγR2b-binding IgGs, FcγR3-binding IgGs, FcγR4-binding IgGs, and total IgG to the indicated spikes, RBDs, or non-sarbecovirus control proteins. Antigen names (y-axes) are colored according to spike or RBD clades. Matched antigens are indicated with gray shading around the name, and mismatched antigens are indicated with a black outline around the name. Each column represents binding data from an individual mouse, and immunization cohorts are separated by a vertical black line. A soluble influenza hemagglutinin trimer (CA-09 sHA) was used as a control to evaluate non-specific binding. See also Extended Data Fig. 5.

Figure 6.. SySPEM score comparisons show differences in epitope recognition across immunogen cohorts.

SySPEM scores from individual mice were determined for RBD epitopes (columns) recognized by different IgG classes (rows, IgG subclass plus different FcγR-binding IgGs) with statistical comparisons between immunogen cohorts (colors). A SySPEM value of 0 indicates that none of the IgGs in that sample were affected by the glycan addition and therefore the sample did not contain IgGs that recognize that epitope, and a SySPEM value of 100 indicates that all IgGs in that sample recognized that epitope (Extended Data Fig. 6). A SySPEM value between 0 and 100 indicates the proportion of IgGs in a sample that recognized the epitope that was blocked by glycan addition in the RBD KO mutant. Box and whisker plots of SySPEM scores with individual data points representing one mouse are shown. Significant differences between cohorts were calculated using Tukey’s HSD posthoc test and linked by vertical lines indicated by asterisks: p<0.05 = *, p<0.01 = **, p<0.001 = ***, p<0.0001 = ****. See also Extended Data Fig. 8 and Fig. 7.

Figure 7.. SySPEM UMAP analyses.

Uniform Manifold Approximation and Projection (UMAP) was used to project multi-dimensional SySPEM scores into two (panels a-e) or three (panel f) dimensions for (a) IgG1 (b) IgG2a, (c) IgG1, IgG2a, IgG2b, and IgG3, (d) Total IgG, (e) FcγR-binding IgGs, and (f) Total IgG, FcγR-binding IgGs, and IgG subclasses. Samples with similar antibody–epitope recognition profiles cluster together, with colored ellipses indicating the variance within each immunogen group and centroid markers (large Xs) showing the group mean. (f) Three-dimensional UMAP projection for all IgG subclasses and FcγR-binding IgGs with centroid markers (large Xs). See also Fig. 6, Extended Data Figs. 6, 8.