A potent broadly neutralizing human RSV antibody targets conserved site IV of the fusion glycoprotein

Abstract

Respiratory syncytial virus (RSV) infection is the leading cause of hospitalization and infant mortality under six months of age worldwide; therefore, the prevention of RSV infection in all infants represents a significant unmet medical need. Here we report the isolation of a potent and broadly neutralizing RSV monoclonal antibody derived from a human memory B-cell. This antibody, RB1, is equipotent on RSV A and B subtypes, potently neutralizes a diverse panel of clinical isolates in vitro and demonstrates in vivo protection. It binds to a highly conserved epitope in antigenic site IV of the RSV fusion glycoprotein. RB1 is the parental antibody to MK-1654 which is currently in clinical development for the prevention of RSV infection in infants.

Fig. 1

In vitro binding and neutralization. a The neutralization activity of RB1 against laboratory strains RSV A (Long) and b RSV B (Washington) with palivizumab as an assay control antibody. Assays were run in duplicate, showing error bars at mean with standard deviation (SD). An IC₅₀ was calculated using a log versus response variable slope 4 parameter fit curve and represents the concentration of antibody required for a 50% reduction in the RSV infectivity in a microneutralization assay. c The neutralization activity of RB1 against laboratory strains HMPV A and B with MPE8 control antibody. Assays were run in triplicate, showing error bars at mean with standard deviation (SD). An IC₅₀ was calculated using a log versus response variable slope 4 parameter fit curve and represents the concentration of antibody required for a 50% reduction in the HMPV infectivity in a microneutralization assay. d The enzyme-linked immunoassay (ELISA) binding curves of RB1 binding to the pre-F and post-F protein conformations. The assay was run in duplicate showing error bars at mean with SD. An EC₅₀ was calculated and represents the concentration of antibody required for a 50% reduction in binding. Source data provided as a source data file

Fig. 2

Neutralization of a diverse set of clinical isolates. a The clinical isolates used to evaluate RB1 neutralization activity were displayed as a dendrogram to evaluate sequence diversity. A phylogenetic tree of 345 unique GenBank sequences is represented in the inner part of the circle while the fusion protein sequences of 46/47 RSV A and B clinical isolates are marked as spokes on the outside of the circle (1 clinical isolate contained an incomplete sequence). Source data (clinical isolate sequences and accession numbers) are provided as a source data file. b RB1 was assessed for its ability to neutralize RSV clinical isolates using an in vitro microneutralization assay. A panel of 47 RSV clinical isolates containing amino acid changes in the F fusion glycoprotein were tested in the in vitro neutralization assay. IC₅₀ was calculated and represents the concentration of antibody required for a 50% reduction in the RSV infectivity. The error bars represent the geometric mean with 95% confidence intervals

Fig. 3

Crystal structure of the RB1 Fab and RSV Pre-fusion F complex. a Interaction of RB1 with the RSV pre-F trimer. Structure of RSV pre- F trimer (protomer 1: yellow, protomer 2: orange) bound to three copies of RB1 (heavy chain: blue, light chain: cyan) as viewed down the 3-fold axis (left) and rotated 90^o to view from the side (right). b Close-look of the RB1-RSV pre-F interaction. Interaction of RB1 and RSV-F: RSV F protein is shown in ribbons on the right with the protomer1 subunit in blue and the protomer2 in green. RB1 is shown as surface representation with the CDRs colors (HC: CDR1: red, CDR2: orange, CDR3: yellow, LC: CDR1: blue, CDR2: dark purple, CDR3: light purple). c Comparison of RB1 and 101F binding. A side view of the RB1 RSV-F interaction with the 101 F interaction modeled in overlay. The RSV-F trimer: RB1 interaction is shown as colored previously (protomer 1: yellow, protomer2: orange, heavy chain: blue, light chain: cyan). The 101F structure was superimposed based on the peptide derived from residues, 427–436 of the F protein (shown as spheres in red). The resulting pose of the 101F Fab is shown in purple. d Close up view of the interaction interface between prefusion F and RB1 (left) and the superposed post-fusion F and RB1 (right). RB1 is depicted in cartoons and held fixed in both images with heavy chain in purple and light chain in cyan. The RSV-F is shown in CA ribbon representation with one strand spanning residues 464–470 depicted in cartoon representation to illustrate the shift between prefusion and post-fusion conformations near the epitope. The monomer that mediates the main interaction is colored in yellow and the other two monomers of the trimer are shown in gray. Two residues of the neighboring monomer, Glu 161 and Ser 182, are shown as sticks and no longer near the antibody in the post-fusion conformation. e A zoomed out view of the prefusion F: RB1 complex is shown on the left, in contrast to a superposed post-fusion F with RB1 on the right. Residues of F within 3.5 angstroms of the antibody in the prefusion form are shown as spheres

Fig. 4

Efficacy of RB1 in the upper and lower respiratory tracts of cotton rats. Cotton rats were administered RB1 by intramuscular injection with a dose titration and blood was collected for evaluation of serum antibody concentrations the following day. Immediately after blood sample collections, each animal was challenged intra-nasally with 1 × 10⁵ plaque-forming units (pfu) of RSV A2 or B 18537 strains. Four days post-challenge, animals were euthanized, blood collected for evaluation of antibody concentrations, and nose and lung tissue collected to assess mAb efficacy by measuring RSV infectious titers in both the nasal and pulmonary tissues. Depicted are the log pfu per gram of tissue for each strain for lung and nose tissues against the log of day 1 serum concentrations in µg/mL. Points represent observed data (5 animals per dose and 5 untreated per experiment) and lines represent non-linear regression fit of the inhibitory sigmoidal dose-response analysis with variable slope depicted with 95% confidence bands for RSV A lung (a), RSV A Nose (b), RSV B lung (c), and RSV B nose (d). e Shows similar analysis after dosing with RB1 or RB1-LALA, which contains mutations to decrease Fc function activity of antibodies, and an analysis of palivizumab as compared to palivizumab-LALA (f). The horizontal dotted lines in all graphs represent the average log pfu/g tissue for the naïve challenged controls in that experiment