Nanobody cocktails potently neutralize SARS-CoV-2 D614G N501Y variant and protect mice

Abstract

Neutralizing antibodies are important for immunity against SARS-CoV-2 and as therapeutics for the prevention and treatment of COVID-19. Here, we identified high-affinity nanobodies from alpacas immunized with coronavirus spike and receptor-binding domains (RBD) that disrupted RBD engagement with the human receptor angiotensin-converting enzyme 2 (ACE2) and potently neutralized SARS-CoV-2. Epitope mapping, X-ray crystallography, and cryo-electron microscopy revealed two distinct antigenic sites and showed two neutralizing nanobodies from different epitope classes bound simultaneously to the spike trimer. Nanobody-Fc fusions of the four most potent nanobodies blocked ACE2 engagement with RBD variants present in human populations and potently neutralized both wild-type SARS-CoV-2 and the N501Y D614G variant at concentrations as low as 0.1 nM. Prophylactic administration of either single nanobody-Fc or as mixtures reduced viral loads by up to 10⁴-fold in mice infected with the N501Y D614G SARS-CoV-2 virus. These results suggest a role for nanobody-Fc fusions as prophylactic agents against SARS-CoV-2.

Keywords: SARS-CoV-2; antiviral therapeutics; cryo-EM; crystallography; nanobodies.

Fig. 1.

Identification and functional characterization of neutralizing anti-RBD nanobodies. (A) The distribution of nanobody sequences from five phage panning campaigns against spike and RBD proteins. The number in the inner circle indicates the number of distinct clonal groups based on the CDR3 sequences. White indicates sequences isolated only once, and colored pie slices are proportional to the number of clonally related sequences. (B) Heatmap of nanobody ELISA binding to SARS-CoV-2 and SARS-CoV RBD, blocking activity by FRET-based assay and neutralization potency with the MNV assay. Nanobodies are ordered based on their clonal lineage as shown on the right. (C) Iso-affinity plot showing the dissociation rate constants (K_D) and association rate constants (K_a) of WNb nanobodies as measured by bio-layer interferometry. Symbols that fall on the same diagonal lines have the same equilibrium dissociation rate constants (K_D) indicated on the top and right sides of the plot. (D) Comparison of potencies in the MNV assay with nanobody affinities with the four WNb lead candidates highlighted, which have antibody affinities <0.5 nM and neutralization activity <60 nM. (E) Epitope competition experiments using bio-layer interferometry using immobilized nanobodies indicated on the left column, incubated with nanobodies indicated on the top row, preincubated with SARS-CoV-2 RBD using a 10:1 molar ratio. Binding of RBD premixed with nanobody was calculated relative to RBD binding alone, which was assigned to 100%. The green and white boxes represent noncompeting and competing nanobodies, respectively.

Fig. 2.

Antibody affinities and neutralization potencies of bivalent nanobody-Fc fusions. (A) Heatmap of WNbFc nanobodies binding to RBD variants with the range of relative EC₅₀ shown [nanomolar (nM)] (Top) and blocking of ACE2 interaction with the range of relative IC50 shown (nM) (Bottom) using the RBD global variant multiplex arrays. Squares with crosses represent nonbinders and noninhibitory nanobodies in top and bottom, respectively. (B) Bio-layer interferometry affinity measurements with immobilized WNbFc fusion and SARS-CoV-2 RBD in solution. Representative binding curves of five different RBD concentrations from 6 to 100 nM to immobilized nanobodies are plotted (solid line) and were fitted to a 1:1 binding model (dashed line). Corresponding mean ± SEM. K_D values are indicated, and representative binding curves are shown from three independent experiments. (C) Neutralization of WT or N501Y D614G SARS-CoV-2 using the MNV and PRNT assay. For MNV, the values are the geometric mean of n = 2 to 4 biological replicates. For PRNT, the value is representative of two biological replicates with duplicate technical replicates per concentration. (D) Inhibition of WT and N501Y D614G SARS-CoV-2 virus using PRNT. For PRNT, the value is representative of two biological replicates with duplicate technical replicates per concentration.

Fig. 3.

Cryo-EM structure of WNb 2–WNb 10–spike complex and crystal structure of WNb2-RBD complex. (A) Cryo-EM maps for spike bound to WNb 2 and WNb 10. (Middle and Right) Overall map for spike ectodomain with the best resolved densities for six bound WNbs (three of WNb 2 and three of WNb 10) showing all RBDs in the “up” conformation. (Left) The map of an individual RBD bound to WNb 2 and WNb 10 following focused 3D classification in Relion 3.1. (B) Cryo-EM structure of WNb 2 and WNb 10 bound to the RBD (Left) and with human ACE2 superimposed (Right). (C) The WNb 2 and WNb 10 binding footprint on SARS-CoV-2 RBD are highlighted as magenta and teal, respectively. The overlay of human ACE2 helices is shown in dark blue. (D and E) Interacting residues between WNb 10 and the RBD. (F) Crystal structure of WNb 2-RBD complex. (G–K) Interacting residues between WNb 2 and the RBD identified from the crystal structure. (L) Overlay of existing nanobody structures with WNb 2 and WNb 10. The codes in the brackets refer to the PDB identification codes.

Fig. 4.

Prophylactic efficacy of neutralizing nanobody-Fc fusions against SARS-CoV-2 infection in mice. (A–D) Mice were prophylactically administered nanobodies or isotype control via intraperitoneal injection. After 24 h, mice were challenged with SARS-CoV-2 N501Y D614G virus via the Glas-col nebulization inhalation system. Lungs were harvested at day 3 postinfection and homogenized for downstream analyses. (B) Mice received 5 mg/kg of the indicated WNbFc fusions. (C) Mice received 5, 1, or 0.2 mg/kg WNbFc 2. (D) Mice received 1 or 0.2 mg/kg of the indicated WNbFc mixtures, 5 mg/kg of isotype control, or 5 mg/kg of WNbFc 2. (B–D) Viral burden in the lung was measured by SARS-CoV-2 live virus quantification by TCID50 assay (Left) and by RT-qPCR (Right). Each circle represents an individual mouse. Dashed lines indicate assay limit of detection. The horizontal lines indicate the median values of animals in each treatment group. To compare TCID50 and RT-qPCR values between groups, a Kruskal–Wallis test across the groups was performed followed with Dunn’s post-test to obtain P values comparing the treated groups with isotype controls (B and D) or highest dose animals (C).