A neutralizing single-domain antibody that targets the trimer interface of the human metapneumovirus fusion protein

Fig 1. Llama immunization with Pre-F hMPV F induces hMPV neutralizing antibodies. (A) hMPV A1-GFP isolate NL/01/00 (left panel) or hMPV A2-GFP isolate CAN97-83 (right panel) was pre-incubated with different dilutions of pre-immune or 130-BV immune serum before inoculation of Vero118 cells or LLC-MK2 cells, respectively. Twenty-four to 72 hours later, the GFP fluorescence was measured and fluorescence intensities were expressed as a percentage of a virus control without sdAb (% viral infectivity). (B) hMPV B2 (TN/83–1211) virus was pre-incubated with different dilutions of pre-immune or 130-BV immune serum before infection of LLC-MK2 cells overlayed with 0.3% avicel. After 5 days, the viral plaques were stained with mouse anti-hMPV serum. (C) hMPV A1-GFP (NL/1/00) was pre-incubated with different concentrations of 36 selected sdAb candidates (sdHMPV1-36) before inoculation of Vero118 cells. Twenty-four to 48 hours later, the GFP fluorescence was measured and fluorescence intensities were expressed as a percentage of a virus control without sdAb (% viral infectivity). The results are depicted on three different graphs to make them more clear.

Fig 2. sdHMPV16 cross neutralizes hMPV A and B but sdHMPV12 only neutralizes hMPV A strains. (A) Details on different neutralization assay formats with sdHMPV12, sdHMPV16, or Ctrl sdAb emphasizing the different cell lines, read outs, and inoculation times. Detailed information can be found in the M&M section. (B) Neutralization assay results of hMPV A1-GFP isolate NL/1/00 (left panel) or hMPV A2-GFP isolate CAN97-83 (right panel). hMPV A1-GFP (NL/1/00) virus was grown in the presence (T+) or absence (T-) of trypsin. Three different mAbs were used as controls: MF-1, MF-14, and MPE8. Fluorescence intensities are expressed as a percentage of a virus control without sdAb (% viral infectivity). (C) Overview of the IC50 values of the neutralization assays with hMPV A strains calculated from GFP or ELISA read outs for GFP or wild-type viruses, respectively. ND = not determined. (D) Neutralization assay results of hMPV B2 isolate TN/83–1211 (left), hMPV B1-GFP isolate NL/1/99, and hMPV B1 isolate SP/1/15 (table on the right). hMPV B2 (TN/83–1211) plaques were stained with mouse anti-hMPV serum and not quantified. IC50 values for hMPV B1-GFP (NL/1/99) and hMPV B1 (SP/1/15) were calculated based…

Fig 3. Binding of sdHMPV12 and sdHMPV16 to hMPV F variants assessed by SPR. The anti-Foldon antibody MF4chim was captured by protein A on a CM5 chip and used to capture the hMPV F variants 130-BV (uncleaved Pre-F A1), 115-BV (cleaved Pre-F A1), A1 Post-F (cleaved Post-F A1), or 519-BV (uncleaved Pre-F B1) via their Foldon. Serial dilutions of sdHMPV12 (A) or sdHMPV16 (B) were injected over the hMPV-F and control cells. Binding was measured with a Biacore X100 instrument as described in the Methods. Raw data and fits to the 1:1 kinetic model are shown with red and black lines, respectively. Affinity/kinetics constants and the highest nanobody concentration tested are indicated for each F protein and sdAb. The instrument detection limit for the dissociation kinetic constant (Kd) is remarked in boldface.

Fig 4. sdHMPV16 binds at the internal interface of two hMPV Pre-F protomers. (A) A ribbon model of sdHMPV12 (peach) and sdHMPV16 (green) bound to uncleaved monomeric hMPV Pre-F (blue) as viewed from the side. The unmodeled flexible protease cleavage site is represented by a dashed line. (B) Surface representations of sdHMPV12 (peach) and sdHMPV16 (green) are shown superimposed onto a ribbon model of cleaved trimeric hMPV Pre-F (PDB ID 5wb0). The sdHMPV16 surface representation is seen to clash extensively with the ribbon model of the neighboring F protomer (white). (C) Close-up of the sdHMPV12-binding site. The antigenic site I hydrophobic pocket residues are displayed as sticks with a transparent space-filling sphere representation of the atoms. Other residues important for the interaction are labeled and shown as sticks. CDR2 and the sdAb framework regions have been removed for clarity. (D) Close-up of the sdHMPV16-binding site. Important residues for the interaction are highlighted as sticks. CDR1 and the sdAb framework regions have been removed for clarity. The IMGT numbering scheme has been used to number the sdAb residues using NanobodyBuilder2 (35) via the SAbPred toolbox (36).

Fig 5. sdHMPV16 binds to infected cells and has an impact on F cleavage. (A) Flow cytometry contour plots showing representative staining of Ctrl sdAb (left panels) or sdHMPV16 (right panels) binding to hMPV-A2 GFP (CAN97-83) or hMPV B2 (TN/83–1211) infected cells detected with a rabbit anti-Histidine tag antibody. sdHMPV16 binds to hMPV-A2 GFP (CAN97-83) and hMPV B2 (TN/83–1211) infected cells but Ctrl sdAb does not. (B) Evaluation of F0 cleavage to F1 by western blot in uncleaved hMPV-A1-GFP (NL/1/00) virus preparations. Viruses were preincubated with sdAbs and then treated with different amounts of TPCK trypsin. After protease incubation samples were processed and F bands were detected using a rabbit anti-F serum. Already cleaved hMPV-A1-GFP (NL/1/00 virus preparations were also loaded onto gels as controls of almost total F0 cleavage). Numbers on the left correspond to molecular mass markers (kilodalton, KDa); and black arrowheads on the right indicate the position of F0 and F1 bands. A representative WB of 2 or 3 replicates for each sdAb is shown.

Fig 6. Prophylactic treatment with sdHMPV16-Fc reduces lung viral replication in cotton rats. Groups of 6 cotton rats were injected intramuscularly with vehicle (PBS), 5 mg/kg MPE8, 5 mg/kg, or 1 mg/kg sdHMPV16-Fc. The next day, the cotton rats were challenged intranasally with 105 PFU of hMPV A2 isolate TN/94–49 in a 0.1 mL volume. Four days after the challenge, the pulmonary viral load was determined by quantification of the lung viral RNA (L gene) compared to β-actin mRNA using RT-qPCR (A) or by quantification of viral titers (PFU/g) by plaque assay (B). The medians of the different groups were compared using a nonparametric Kruskal-Wallis test. ***0.0001 < P < .001; **0.001 < P < .01, and ns = not significant (P ≥ 0.05).