Cross-Reactive and Potent Neutralizing Antibody Responses in Human Survivors of Natural Ebolavirus Infection

Abstract

Recent studies have suggested that antibody-mediated protection against the Ebolaviruses may be achievable, but little is known about whether or not antibodies can confer cross-reactive protection against viruses belonging to diverse Ebolavirus species, such as Ebola virus (EBOV), Sudan virus (SUDV), and Bundibugyo virus (BDBV). We isolated a large panel of human monoclonal antibodies (mAbs) against BDBV glycoprotein (GP) using peripheral blood B cells from survivors of the 2007 BDBV outbreak in Uganda. We determined that a large proportion of mAbs with potent neutralizing activity against BDBV bind to the glycan cap and recognize diverse epitopes within this major antigenic site. We identified several glycan cap-specific mAbs that neutralized multiple ebolaviruses, including SUDV, and a cross-reactive mAb that completely protected guinea pigs from the lethal challenge with heterologous EBOV. Our results provide a roadmap to develop a single antibody-based treatment effective against multiple Ebolavirus infections.

Figure 1. Cross-reactive B cell responses in filovirus immune donors

Supernatants from EBV-transformed PBMC samples isolated from survivors were screened in ELISA binding assays using BDBV, EBOV or MARV GPs (A–C). Results for four BDBV survivors (A), one EBOV survivor (B) or one MARV survivor (C) are shown. Height of the bars indicates OD_{405 nm} values in ELISA binding to full-length extracellular domain of GP of the indicated virus species. Reactive supernates are color-coded based on the cross-reactivity pattern: species-specific cell lines are highlighted in black; cross-reactive lines to 2 or 3 species are shown in yellow or blue, respectively. Previous work has shown that the amino acid sequence of GP differs between BDBV and EBOV by over 34%, and between BDBV and MARV by over 72%. (D) Percentages of lines secreting antibodies specific to BDBV, EBOV or MARV GPs, or cross-reactive antibodies to BDBV and EBOV (designated BDBV/EBOV) or BDBV, EBOV and MARV (designated BDBV/EBOV/MARV) are shown. Increasing intensity of the pink cell fill color corresponds to increasing reactivity for indicated virus. See also Figure S1.

Figure 2. Cross-neutralizing antibodies from survivors of natural BDBV infection

(A) Heat map showing the binding of BDBV mAbs to a panel of filovirus GPs. The EC₅₀ value for each GP-mAb combination is shown, with dark red, orange, yellow, or white shading indicating high, intermediate, low, or no detectable binding, respectively. EC₅₀ values greater than 10,000 ng/mL are indicated by the > symbol. NAb names are highlighted in red. (B) Heat map showing the neutralization potency of BDBV GP-specific mAbs against BDBV. The IC₅₀ value for each virus-mAb combination is shown. IC₅₀ values greater than 10,000 ng/mL are indicated by the > symbol. Neutralization assays were performed in triplicate. (C) Binding of representative mAbs from six distinct binding groups to the filovirus GP. (D) Neutralization activity of representative neutralizing mAbs from three binding groups against BDBV, EBOV or SUDV. Error bars represent the SE of the experiment, performed in triplicate. See also Table S1, Data S1 and Figures S2 to S3.

Figure 3. BDBV-neutralizing antibodies target at least two distinct antigenic regions of the GP surface

Data from competition-binding assays using non-neutralizing mAbs from binding Group 1A (white background) and neutralizing mAbs from binding Groups 1A, 1B, 3A or 3B (pink background). Numbers indicate the percent binding of second mAb in the presence of the first mAb, compared to binding of un-competed second mAb. MAbs were judged to compete for the same site if maximum binding of second mAb was reduced to <30% of its un-competed binding (black boxes with white numbers). MAbs were considered non-competing if maximum binding of second mAb was >70% of its un-competed binding (white boxes with red numbers). Grey boxes with black numbers indicate an intermediate phenotype (competition resulted in between 30 and 70% of un-competed binding). Blue, purple, and green dashed lines indicate what appear to be major competition groups; the blue and purple groups overlap substantially but not completely.

Figure 4. BDBV-neutralizing antibodies bind to the glycan cap or base region of GP

(A) Shown are negative-stain electron microscopy reference-free 2D class averages of Group 1A antibodies that bind both the glycan cap of GP and sGP, and Group 1B antibodies that bind the glycan cap of GP but not sGP. BDBV GP or GPΔmuc was used to generate complexes. (B) 3D reconstructions of glycan cap binders from Groups 1A and 1B reveal that these antibodies bind the glycan cap at overlapping but distinct epitopes. Top (left) and side (right) views of the complexes are shown. (C) Reference free 2D class averages of Group 1B antibodies (left) reveals that these antibodies bind an epitope below the base of GP that is flexible. In the middle image, GP is colored yellow and each Fab colored green. The right-hand panel illustrates a superimposition of crystal structures of SUDV GPΔmuc (PDB 3VEO) and Fabs (PDB 3CSY) to demonstrate how Fabs may bind to GP. (D) The composite model delineates the epitopes of the glycan cap mAbs in Group 1A or 1B. Side (above) and top (below) views are shown. (E) Docking a crystal structure of SUDV GPΔmuc (PDB 3VEO) (Bale et al., 2012), which contains a more complete model of the glycan cap region targeted by Group 1A/B mAbs, reveals how Group 1A/B mAbs target a broad region in the GP1 centered on the glycan cap, near the beginning of the mucin-like domains. Group 1B mAbs that target the base likely bind to a loop near the membrane proximal external region (MPER) that is flexible and has not yet been resolved at high resolution. TM = transmembrane region; CT = cytoplasmic tail. See also Figure S4.

Figure 5. Epitope mapping of Group 3A mAbs using saturation mutagenesis and negative stain electron microscopy

Epitope residues for three nAbs from Group 3A (BDBV270, BDBV289 and BDBV324) were identified as those for which mutation to alanine specifically reduced binding of these antibodies (A–B). GP residue W275 was common to all three nAbs, while L273 was specific for BDBV324, and Y241 was specific for BDBV289. The mutated residues are shown in space filling forms on a ribbon diagram of the EBOV GP structure, based on PDB 3CSY. (C) Binding values for nAbs and previously isolated mAbs KZ52, 2G4 and 4G7 to library clones with mutations at residues L273, W275 and Y241. The Ab reactivities against each mutant EBOV GP clone were calculated relative to reactivity with wild-type EBOV GP. (D) BDBV289 (brown) binds at the top of the viral GP near the glycan cap region. Complexes are of BDBV antibody Fab fragments bound to BDBV GPΔTM with side view (top panel) or top view (bottom panel). A representative Fab crystal structure is fit in the Fab density for each reconstruction (from PDBID 3CSY). A monomer of the GP trimer crystal structure (PDBID 3CSY) is also fit in the GP density, with white corresponding to GP1 and black to GP2. Two critical residues for binding by BDBV289 (W275 and Y241, determined using saturation mutagenesis) are highlighted in green. See also Figure S5.

Figure 6. Survival and clinical signs of EBOV-inoculated mice treated with BDBV mAbs

Groups of 5 mice in each group were injected with individual mAbs by the intraperitoneal route 1 day after EBOV challenge, using 100 μg of mAb per treatment. Animals treated with dengue virus-specific human mAb 2D22 served as controls. (A) Kaplan-Meier survival curves. (B) Body weight. (C) Illness score.

Figure 7. Survival and clinical signs of EBOV inoculated guinea pigs treated with BDBV mAbs

Groups of 5 guinea pigs per group were injected with individual mAbs by the intraperitoneal route 1 day or 1 and 3 days after EBOV challenge, using 5 mg of individual mAb (A) or 5 mg of the combination of two mAbs per treatment (B), as indicated. Animals treated with dengue virus-specific human mAb 2D22 served as controls. The survival curves are based on morning and evening observations. Mortality in the morning is shown in whole day numbers, in the evening in 1/2 day values. The body weight and illness scores are shown with one value per day. See also Table S2.