Design and characterization of protective pan-ebolavirus and pan-filovirus bispecific antibodies

Abstract

Monoclonal antibodies (mAbs) are an important class of antiviral therapeutics. MAbs are highly selective, well tolerated, and have long in vivo half-life as well as the capacity to induce immune-mediated virus clearance. Their activities can be further enhanced by integration of their variable fragments (Fvs) into bispecific antibodies (bsAbs), affording simultaneous targeting of multiple epitopes to improve potency and breadth and/or to mitigate against viral escape by a single mutation. Here, we explore a bsAb strategy for generation of pan-ebolavirus and pan-filovirus immunotherapeutics. Filoviruses, including Ebola virus (EBOV), Sudan virus (SUDV), and Marburg virus (MARV), cause severe hemorrhagic fever. Although there are two FDA-approved mAb therapies for EBOV infection, these do not extend to other filoviruses. Here, we combine Fvs from broad ebolavirus mAbs to generate novel pan-ebolavirus bsAbs that are potently neutralizing, confer protection in mice, and are resistant to viral escape. Moreover, we combine Fvs from pan-ebolavirus mAbs with those of protective MARV mAbs to generate pan-filovirus protective bsAbs. These results provide guidelines for broad antiviral bsAb design and generate new immunotherapeutic candidates.

Fig 1. Schematics of bsAbs and neutralization activity against rVSV-EBOV.

(A) Schematic of bsAbs constructs, along with their nomenclature. The first designation in bsAb names reflects their format (“DV” for DVD-Ig, “SC” for scFv-IgG, “hSC” for hinge scFv-IgG, and “AS” for asymmetric). For symmetric bsAbs, Fvs with blue domains are the ones remaining in Fab format and their name appears second in order. For example, “DV_A774-A878” is a DVD-Ig format with A774 Fvs as the green (outer) domains and A878 Fvs as the blue (inner). (B) Heat map of neutralizing half-maximal inhibitory concentration (IC₅₀) values against rVSV-EBOV for the antibody panel. mAbs or bsAbs with curves that did not cross the 50% threshold are designated as non-neutralizing (NN).

Fig 2. Neutralization of rVSV-EBOV_CL by Group II bsAbs.

Group II bsAb neutralization activity against rVSV-EBOV_cl. rVSV-EBOV was pretreated with thermolysin to generate rVSV-EBOVcl prior to incubation with serial dilution of bsAbs. IC₅₀ values indicated in the legend in parentheses.

Fig 3. bsAbs exhibit neutralization breadth across rVSV-EBOV, -SUDV, -MARV.

Heat map of neutralizing half-maximal inhibitory concentration (IC₅₀) values of the bsAb panel against VSV-EBOV, -SUDV, and -MARV. mAbs or bsAbs with curves that did not cross the 50% threshold are designated as non-neutralizers (NN). ND, not determined.

Fig 4. bsAbs exhibit neutralization breadth against authentic filoviruses.

Neutralization curves for (A) Group I and (B) Group II bsAbs against EBOV and SUDV (respective IC₅₀ values indicated in the legend in parentheses). Group III bsAb neutralization against (C) EBOV (D) SUDV and (E) MARV (IC₅₀ values again in parentheses in the legend).

Fig 5. Neutralization of viral escape mutants.

(A) Neutralization capacity of A774 against rVSVs bearing either WT EBOV/Makona GP or a partial neutralization escape variant GP (A774R). (B) Capacity of Group I bsAbs to neutralize VSV-EBOV^A774R bearing the A774 escape variant GP. (C) Capacity of Group I bsAbs to neutralize VSV-SUDV^A878R bearing the A878 escape variant GP. IC₅₀ values indicated in the legend in parentheses.

Fig 6. Binding profiles of bsAbs to EBOV GP and MARV GP by biolayer interferometry (BLI).

Sensors were loaded with antibodies then dipped into solutions with indicated concentrations of EBOV GP. Gray lines show curve fits to a 1:1 binding model.

Fig 7. Two-phase binding of mAbs and bsAbs to EBOV GP_CL.

Kinetic binding curves for parental mAbs (A) A878 and (B) MR191 against cleaved EBOV GP (GP_CL) were determined by BLI. Sensors loaded with mAb were then dipped into solutions with indicated concentrations of EBOV GP_CL. (C-H) Two phase binding experiments for parental mAbs and group III bsAbs. Sensors were loaded with parental mAb A878 (C,E,G) or MR191 (D,F,H), then sequentially dipped into an analyte containing EBOV GP_CL followed by a second antibody: (C) MR191; (D) A878; (E,F) DV_A878-MR191; or (G,H) hSC_MR191-A878.

Fig 8. BsAbs afford broad protection against lethal challenge of divergent filoviruses in murine models.

BALB/c mice challenged with 100 plaque forming units (pfu) of mouse-adapted EBOV and treated with a single dose of parental mAb A878 or A774 [100ug, ~5 mg per kilogram (mg/kg)]; bsAb [133ug, ~6.65 mg/kg, adjusted for molecular weight]; or vehicle (phosphate buffered saline, PBS) 2 days post-infection (dpi). Survival and weight loss for Group I bsAbs (A), Group II bsAbs (B), and Group III bsAbs (E). Survival and weight loss for type 1 interferon α/β receptor knockout (IFNAR^-/-) mice treated with a single dose of parental mAb MR191 [100ug, ~5 mg/kg]; Group III bsAb [133ug, ~6.65 mg/kg, adjusted for molecular weight]; or vehicle (PBS) 2 dpi with 1000 pfu of MARV-Ci67 (MARV) (C) or Ravn virus (RAVV) (D). Each treatment group consisted of 10 mice. Survival curves for each group were compared to the PBS-treated group using the Log-rank (Mantel-Cox) test. (*p<0.05; **p<0.01; ***p<0.001; ****p<0.0001).