Nonhuman primate antigenic cartography of SARS-CoV-2

Abstract

Virus neutralization profiles against primary infection sera and corresponding antigenic cartography are integral part of the COVID-19 and influenza vaccine strain selection processes. Human single variant exposure sera have previously defined the antigenic relationships among SARS-CoV-2 variants but are now largely unavailable due to widespread population immunity. Therefore, antigenic characterization of future SARS-CoV-2 variants will require an animal model, analogous to using ferrets for influenza virus. We evaluated neutralization profiles against 23 SARS-CoV-2 variants in nonhuman primates (NHPs) after single variant exposure and generated an NHP-derived antigenic map. We identified a distant antigenic region occupied by BA.2.86, JN.1, and the descendants KP.2, KP.3, and KZ.1.1.1. We also found that the monovalent XBB.1.5 mRNA vaccine induced broad immunity against the mapped antigenic space. In addition, substantial concordance was observed between our NHP-derived and two human antigenic maps, demonstrating the utility of NHPs as a surrogate for antigenic cartography in humans.

Keywords: CP: Immunology; NHP; SARS-CoV-2; antibody landscapes; antigenic cartography; antigenic evolution; neutralization titer; nonhuman primates; vaccine update.

Graphical abstract

Figure 1

Panel of analyzed SARS-CoV-2 variants (A) Evolutionary relationships of virus variants were visualized by a phylogenetic tree. Virus sequences were retrieved from the GISAID (Global Initiative on Sharing All Influenza Data) Initiative (Table S1) and illustrated with clades.nextstrain.org. (B) Spike mutations of analyzed SARS-CoV-2 variants are shown relative to the ancestral Wuhan/WIV04/reference strain. Amino acid substitutions of individual variants are indicated by green tiles and deletions by blue tiles. The figure was generated using a Jupyter Notebook from Jesse Bloom and illustrated with Adobe Illustrator.

Figure 2

Neutralization of SARS-CoV-2 variants by NHPs with different exposure histories (A) Sera were collected from unvaccinated animals after challenge with Wuhan (n = 5), Alpha (n = 6), Beta (n = 12), Gamma (n = 6), Delta (n = 16), Omicron BA.1 (n = 11), BA.2.12.1 (n = 3), or BA.4/BA.5 (n = 3/n = 9) variant and (B) from animals vaccinated with a single dose of adenoviral Wuhan (Ad26.CoV.2.S, n = 4) or Beta (Ad26.CoV.2.S.351, n = 4) vaccine, or two homologous doses of monovalent Wuhan (n = 6) or XBB.1.5 (n = 6) mRNA vaccine. NAb titers against indicated SARS-CoV-2 variants (x axis) were assessed using a luciferase-based pseudovirus neutralization assay. Each panel is labeled with the respective study cohort and colored according to the exposed antigen variant to match colors in the subsequent antigenic map. Faint thin lines show individual serum titers, while bold lines represent geometric mean titers (GMTs) with 95% confidence interval (CI) of each study cohort.

Figure 3

Antigenic cartography of SARS-CoV-2 variants and antibody landscapes following monovalent Wuhan or XBB.1.5 vaccination (A) An antigenic map of SARS-CoV-2 variants was constructed as previously described^, from NHP sera following single variant exposure by infection or vaccination. Analyzed virus variants are denoted as colored circles and individual sera are displayed by open squares in colors corresponding to their exposed antigen. Major variants are shown as bigger circles and minor variants as smaller circles. Sera located outside of the represented area are indicated by triangles. A non-zoomed map is shown in Figure S5. All virus variants are positioned relative to each other (x and y axis orientation is relative) with one grid distance within this map corresponding to one 2-fold dilution of neutralization titers. The legend shows study cohorts and sample sizes used for the construction of the antigenic map. The small number in the bottom left indicates the map stress, i.e., the error between titer measurement and map position. (B) Neutralization profiles of NHPs after two doses of monovalent Wuhan (blue) or XBB.1.5 (orange) mRNA vaccines were visualized in 3D above the antigenic map by GMT antibody landscapes. Measured GMTs against each variant are indicated as colored points above the respective variant; the surface shows the fitted antibody landscape. The z axis reflects NAb titers, with every 4-fold increase from titer 10 marked.

Figure 4

Neutralizing antibody responses of matched NHP and human cohorts NAb responses in NHP sera against Wuhan, Alpha, Beta, Delta, and Omicron BA.1, BA.2, and BA.5 variants were compared with two previously published human datasets (Tables S4 and S5). Neutralization titers in human samples were assessed using live virus in Vero-TMPRSS2/ACE2 cells (human dataset 1, Rössler et al.⁴⁶) and lentivirus neutralization assays in 293T/ACE2, same as NHP data (human dataset 2, Wilks et al.³⁶). To control titer variations due to different reactivities of individual sera and estimate assay- and species-specific effects, titers were adjusted using a Bayesian framework.^, The columns show NHP (red), human(green and blue), and, in black, the derived combined GMT (geometric mean titers) ± 95% CI (confidence interval) as bold colored line and shaded area for each serum group. The black line represents the estimated GMT per serum group across organisms after adjusting for serum, assay, and organism effects. The upper row shows GMTs of raw titers with titers <LOD (limit of detection ≤20) set to LOD/2, while the lower row shows GMTs ± 95% CI after adjusting for estimated serum, assay, and organism reactivity differences (Figure S8).

Figure 5

Comparison of NHP-derived antigenic map with two previously published human antigenic maps Arrows point to the variants’ positions in the NHP map (Figure 3) from (A) the previously published human dataset 1 map and (B) the human dataset 2 map. Virus variants are shown by colored circles, and the colors correspond to the colors used in Figure 3.