A monoclonal antibody targeting the Nipah virus fusion glycoprotein apex imparts protection from disease

Abstract

Nipah virus (NiV) is a highly pathogenic paramyxovirus capable of causing severe respiratory and neurologic disease in humans. Currently, there are no licensed vaccines or therapeutics against NiV, underscoring the urgent need for the development of countermeasures. The NiV surface-displayed glycoproteins, NiV-G and NiV-F, mediate host cell attachment and fusion, respectively, and are heavily targeted by host antibodies. Here, we describe a vaccination-derived neutralizing monoclonal antibody, mAb92, that targets NiV-F. Structural characterization of the Fab region bound to NiV-F (NiV-F-Fab92) by cryo-electron microscopy analysis reveals an epitope in the DIII domain at the membrane distal apex of NiV-F, an established site of vulnerability on the NiV surface. Further, prophylactic treatment of hamsters with mAb92 offered complete protection from NiV disease, demonstrating beneficial activity of mAb92 in vivo. This work provides support for targeting NiV-F in the development of vaccines and therapeutics against NiV.IMPORTANCENipah virus (NiV) is a highly lethal henipavirus (HNV) that causes severe respiratory and neurologic disease in humans. Currently, there are no licensed vaccines or therapeutics against NiV, highlighting a need to develop countermeasures. The NiV surface displays the receptor binding protein (NiV-G, or RBP) and the fusion protein (NiV-F), which allow the virus to attach and enter cells. These proteins can be targeted by vaccines and antibodies to prevent disease. This work describes a neutralizing antibody (mAb92) that targets NiV-F. Structural characterization by cryo-electron microscopy analysis reveals where the antibody binds to NiV-F to neutralize the virus. This study also shows that prophylactic treatment of hamsters with mAb92 completely protected against developing NiV disease. This work shows how targeting NiV-F can be useful to preventing NiV disease, supporting future studies in the development of vaccines and therapeutics.

Keywords: Nipah virus; fusion glycoprotein; henipavirus; monoclonal antibodies; neutralizing antibodies.

Fig 1

The cryo-EM structure of the NiV-F–Fab92 complex. (A) The model of the NiV-F–Fab92 complex fit into the cryo-EM map (transparent surface). The relative position of the NiV envelope is shown as a black line. (B) Side view of the NiV-F–Fab92 complex, shown in cartoon representation. The NiV-F trimer is colored light blue, with the F protomer to which Fab92 is bound colored dark blue for clarity. The Fab92 heavy chain is colored dark gray and the light chain is colored light gray. Heavy chain CDRs are colored pink and light chain CDRs are colored green, according to the legend. (C) Top view of the NiV-F–Fab92 complex. The NiV-F trimer is shown with surface representation and each F protomer is colored a different shade of blue. The F protomer to which Fab92 is bound is dark blue. The Fab92 CDR loops are shown as tubes, and the heavy and light CDRs are colored shades of pink and green according to the legend. Glycans observed in the reconstruction are depicted as sticks.

Fig 2

Molecular characterization of the Fab92 epitope. (A) Close-up view of the heavy chain CDR molecular interactions with NiV-F. (B) Close-up view of light chain CDR molecular interactions with NiV-F. Fab92 is shown as a gray cartoon tube and NiV-F is shown as a dark blue cartoon. F protomers not bound by Fab92 are shown as light blue and slightly transparent. CDR loops are colored in shades of green (light chain) and pink (heavy chain), as indicated. Side chains of residues participating in interactions such as intermolecular hydrogen bonds or salt bridges are shown as sticks.

Fig 3

Glycosylation at the F2 site on NiV-F impedes Fab92 binding. (A, left panel) mAb92 binding of NiV-F or HeV-F in the presence or absence of the F2 glycan (WT and F2 mut, respectively). 293T cells were transiently transfected to express the indicated HNV-F glycoprotein prior to flow cytometry with mAb92. Binding was normalized to two anti-HNV-F polyclonal antibodies, pAb2489 and pAb2490, previously shown to have equivalent cross-reactivity to NiV and HeV-F WT and mutant constructs (36), as indicated in Materials and Methods. The polyclonal antibodies were further normalized to WT NiV-F binding for further comparison. Presented are the results from three independent biological replicates with statistics determined by one-way ANOVA with Sidak's correction for multiple comparisons (****P < 0.0001). (A, right panel) mAb92 neutralization of HNV pseudotyped particles (HNVpp) infection on U87 MG glioblastoma cells HNVpp bearing WT-G and WT-F (F2 glycan present) or F2mut (F2 glycan absent) glycoproteins were produced with a VSV∆G-RLuc system as described in Materials and Methods. HNVpp input was optimized to give output within the dynamic range of the assay. A constant amount of the indicated HNVpp was used to infect permissive U87 cells in the presence or absence of a serial fivefold dilution of mAb92. Data presented are from three independent biological replicates, each with technical duplicates, and points are presented as the mean ± SE. Neutralization curves were constrained by artificially setting the lowest mAb concentration (x-axis) to maximum infection (y-axis) to represent neutralization in the absence of any mAb. The data were then analyzed using a variable slope model with a four-parameter dose-response curve (GraphPad PRISM), and statistical significance was tested with a two-way ANOVA with Dunnett's correction for multiple comparisons. (B) Models for Fab92 and NiV-F are shown in cartoon representation and colored gray and blue, respectively. The NiV-F protomer to which Fab92 is bound is colored dark blue and the neighboring protomers are colored light blue. Residues from CDR loops L1 (dark green), L3 (light green), and H2 (pink) contact the first GlcNAc residue of the F2 glycan on two protomers of NiV-F. The side chains of residues observed to interact with the first GlcNAc of the F2 glycan [as calculated by the PDBePISA server (47) are shown as sticks and colored according to CDR loop as described in Fig. 1]. The side chain of the F2 Asn67 residue is shown as blue sticks. The first F2 GlcNAc is shown as salmon sticks. (C) The complex glycan (shown as light orange sticks) from the structure PDB 4BYH (48) was modeled onto the F2 glycan site on two protomers of NiV-F by alignment onto the first GlcNAc that was observed in the cryo-EM-derived reconstruction. The black arrows indicate the direction of the F2 glycan extending outwards away from Fab92, allowing Fab92 to access its proteinaceous epitope. (D) To investigate if the presence of a full-length F2 glycan may interfere with Fab92 binding to adjacent protomers, Fab92 was superposed onto the binding site on the adjacent NiV-F protomers. The Fab92 modeled onto the adjacent protomer is colored gray. The modeled glycan at the F2 position is shown to clash with a modeled Fab92 bound to the adjacent protomer.

Fig 4

Differences between NiV-F and HeV-F within the Fab92 binding footprint. (A) The residues on the NiV-F surface in the binding interface are colored according to the Fab92 chain involved in the contact. Residues contacted by the light chain are highlighted green, residues contacted by the heavy chain are highlighted pink, and those contacted by both chains are shown in gray. The NiV-F trimer is depicted with surface representation in white, with a single protomer that contains the Fab92 epitope in light blue for clarity. (Lower) Amino acid sequence alignment between NiV-F (Malaysia, AAV80428.1) and HeV-F (AEB21197.1) was generated by Multalin (49) and plotted by ESPript (50). Residues involved in the Fab92 interface are annotated with colored boxes below the alignment. Residues in the interface that differ between NiV-F and HeV-F are outlined by a red box. (B) Surface view of the NiV-F trimer shown in white, with the protomer with Fab92 bound shown in light blue. Residues contacted by Fab92 are colored dark blue. Residues in the epitope that are conserved between NiV-F and HeV-F are colored dark blue and those that differ are colored red. (C) Molecular interactions of Fab92 with NiV-F residue Gln70. Side chains of CDR H3 (hot pink) and Gln70 are shown as sticks. NiV-F is shown in dark blue. (D) A model for steric hinderance in Fab92 recognition of HeV-F imposed by Lys70. To better understand the interactions between Fab92 and HeV-F, the HeV-F structure [PDB ID 5EJB (13), cyan cartoon] was aligned to NiV-F using COOT. The lysine at position 70 is modeled as observed in the HeV-F crystal structure, as well as in possible rotamer configurations that introduce steric clashes with Fab92 residues Tyr33 (CDR H1, pink) and Tyr100B (CDR H3, hot pink). (E) mAb92 binding of NiV and HeV-F WT or mutant F proteins. 293T cells were transfected as described in Materials and Methods to express WT HNV-F, the double residue 70 and 74 mutant (70+74mut), the residue 70 only mutant (70mut), or the residue 74 only mutant (74mut). Data presented are the result of three independent biological replicates and are analyzed as described previously for Fig. 3A, left panel (ns, not significant, ****P < 0.0001). (F) mAb92 neutralization of HNVpp expressing WT or mutant F glycoproteins. HNVpp were produced as described in Materials and Methods to express WT G and the indicated NiV/HeV WT or mutant (70+74mut, 70mut, or 74mut) F glycoprotein. Neutralization curves were generated from three independent biological replicates done in technical duplicates. Data were analyzed as described above for Fig. 3A, right panel. WT data for binding and neutralization are repeated from Fig. 3A to facilitate interpretation of these findings in the context of the 70 and 74 double or individual mutants.

Fig 5

Prophylactic treatment with mAb92 protects Syrian golden hamsters from NiV but not HeV disease. (A) mAb92 titers were determined by ELISA assay against purified NiV-F using anti-rabbit secondary antibody. Each sample was tested in duplicate and each data point represents the end-point titer for an individual hamster. (B) Survival of control hamsters and mAb92 treated hamsters following NiV Malaysia and HeV challenge. Statistical significance was calculated with a Log-rank (Mantel-Cox) test. Each group consisted of seven animals monitored for survival. (C) Viral load (genome copies/mL) in oropharyngeal swabs following challenge with NiV Malaysia determined by qRT-PCR. (D) Viral load (genome copies/mL) in oropharyngeal swabs following challenge with HeV determined by qRT-PCR. (E) Infectious virus titer (TCID₅₀/mL) in oropharyngeal swabs collected following challenge with NiV Malaysia determined by end-point titration on Vero E6 cells. (F) Infectious virus titer (TCID₅₀/mL) in oropharyngeal swabs collected following challenge with HeV determined by end-point titration on Vero E6 cells. For (E) and (F), the dotted line represents the limit of detection for the titration assay and each point represents a swab from an individual animal at the given time point. In each plot, NiV control animals are represented by dark blue and circles, NiV mAb92 animals by light blue and triangles, HeV control animals by maroon and squares, and HeV mAb92 animals by salmon pink and diamonds.

Fig 6

Prophylactic treatment with mAb92 reduces viral loads in tissues from hamsters following challenge with NiV and HeV. (A) Viral load (genome copies/gram) in lung and brain tissues collected from hamsters challenged with NiV Malaysia at 5 DPI determined by qRT-PCR. (B) Infectious virus titer (TCID₅₀/gram) in lung and brain tissues collected from hamsters following NiV Malaysia challenge at 5 DPI, determined by end-point titration on Vero E6 cells. (C) Viral load (genome copies/gram) in lung and brain tissues collected from hamsters challenged with HeV at 4 DPI, determined by qRT-PCR. (D) Infectious virus titer (TCID₅₀/gram) in lung and brain tissues collected from hamsters following HeV challenge at 4 DPI, determined by end-point titration on Vero E6 cells. For (C) and (D), a dotted line represents the limit of detection for the titration assay. Statistical significance was determined using an unpaired t-test in GraphPad prism, v.9.1.0. Significance is denoted as follows: ns = P > .05, *P ≤ .05; **P ≤ .01; ***P ≤ .001; ****P ≤ .0001. In each plot, NiV control animals are represented by dark blue and circles, NiV mAb92 animals by light blue and triangles, HeV control animals by maroon and squares, and HeV mAb92 animals by salmon pink and diamonds.

Fig 7

Pathological and histological changes in lung tissue in control and mAb92 treated hamsters after NiV Malaysia or HeV challenge. (A) Gross pathology of lungs from representative control animals (left) and mAb92 treated animals (right), following challenge with NiV Malaysia, euthanized at 5 DPI. (B) Gross pathology of lungs from representative control animals (left) and mAb92 treated animals (middle and right), following challenge with HeV, euthanized at 4 DPI. (C) Hematoxylin and eosin (H&E) staining of lung tissue from representative control animals (left) and mAb92 treated animals (right), following challenge with NiV Malaysia, euthanized at 5 DPI (100 × bar = 100 µm). (D) H&E staining of lung tissue from representative control animals (left) and mAb92 treated animals (middle and right), following challenge with HeV, euthanized at 4 DPI (100 × bar = 100 µm). (E) In situ hybridization (ISH) stain of representative lung from control animals (left) and mAb92 treated animals (right), following challenge with NiV Malaysia, euthanized at 5 DPI (100 × bar = 100 µm). (F) ISH stain of representative lung from control animals (left) and mAb92 treated animals (middle and right), following challenge with HeV, euthanized at 4 DPI (100 × bar = 100 µm).

Fig 8

Mapping of anti-HNV-F Fab structures reveals the antigenic landscape of HNV-F. To visualize the breadth of characterized epitopes on NiV-F, the structures of mAbs in complex with HNV-F currently available in the PDB were modeled bound to the crystal structure of NiV-F. The following structures are shown: apical epitopes, mAb66 [PDB ID 6T3F (36)], mAb92 (PDB ID 8RVN), mAb 12B2 [PDB ID 7KI4 (38)], mAb 4H3 [PDB ID 7UOP (40)], and mAb 2D3 [PDB ID 7UP9 (40)]; lateral epitopes, mAb 5B3 [PDB ID 6TYS (37)], mAb 1F5 [PDB ID 7KI6 (38)], mAb 1A9 [PDB ID 7UPK (40)], and mAb 1H8 [PDB ID 7UPA (40)]; and basal epitopes, mAb 2B12 [PDB 7UPD (40)] and mAb 1H1 [PDB ID 7UPB (40)]. The NiV-F trimer is shown with surface representation, with each protomer colored a different shade of gray. The antibody Fab molecules are depicted as cartoon tube. The relative position of the NiV-F viral envelop is noted with a black line.