A broadly protective therapeutic antibody against influenza B virus with two mechanisms of action

Abstract

Influenza B virus (IBV) causes annual influenza epidemics around the world. Here we use an in vivo plasmablast enrichment technique to isolate a human monoclonal antibody, 46B8 that neutralizes all IBVs tested in vitro and protects mice against lethal challenge of all IBVs tested when administered 72 h post infection. 46B8 demonstrates a superior therapeutic benefit over Tamiflu and has an additive antiviral effect in combination with Tamiflu. 46B8 binds to a conserved epitope in the vestigial esterase domain of hemagglutinin (HA) and blocks HA-mediated membrane fusion. After passage of the B/Brisbane/60/2008 virus in the presence of 46B8, we isolated three resistant clones, all harbouring the same mutation (Ser301Phe) in HA that abolishes 46B8 binding to HA at low pH. Interestingly, 46B8 is still able to protect mice against lethal challenge of the mutant viruses, possibly owing to its ability to mediate antibody-dependent cellular cytotoxicity (ADCC).

Figure 1. 46B8 blocks low pH-induced membrane fusion and conformational changes of HA.

(a) Hela cells expressing the B/Wisconsin/1/2010 HAs were treated with trypsin to activate HA0 and then incubated with either 46B8 or a control IgG before pH drop to 4.8 to induce cell–cell fusion. After overnight culture, representative images were obtained under a phase contrast microscope. Scale bar is 200 μm. (b) 293 T cells expressing the B/Victoria/504/2000 HA were treated with trypsin to activate HA0 and incubated with either 46B8 or a control IgG before pH drop to 4.8 to induce HA conformational changes. Cells were then treated with DTT (+DTT) or PBS (−DTT) and tested for 34B5 binding by flow cytometry. The mean fluorescence intensities were normalized to the PBS-treated cells in each group and the percent of binding is shown. Possible HA conformations at each stage are depicted above the binding data. HA1 is in blue and HA2 is in red.

Figure 2. 46B8 induces ADCC in vitro.

(a) A549 cells infected with B/Brisbane/60/2008 were labelled with 46B8, Rituximab (negative control) or Cetuximab (positive control) prior to incubation with NK cells. NK cell activation (left panel) and target cell lysis (right panel) were presented as the frequency of CD107a⁺ cells and the percent of LDH release, respectively. (b) A549 cells infected with B/Brisbane/60/2008 were labelled with 46B8 or 46B8 N297G prior to incubation with NK cells. Target cell lysis was presented as the percent of LDH release. The assays were done in duplicate. Results are representative of three independent experiments. (a,b: mean and s.e.m.)

Figure 3. 46B8 binds to conserved residues in the vestigial esterase domain of HA.

(a) Negative stain EM reconstruction (grey mesh) of 46B8 Fabs in complex with the B/Victoria/504/2000 HA trimer. Side (left panel) and overhead (right panel) views show the fit of a generic Fab model and the crystal structure of an IBV HA (PDB ID 4M40) to the EM density. HA is in different shades of blue, the heavy chain of 46B8 Fab is in red and the light chain is in yellow. (b) A detailed view of the 46B8-HA contact surface. Potential 46B8-interacting residues on HA are shown as purple sticks and labelled (numbering includes the signal sequence). Colour code is the same as in a. (c) 293 T cells expressing the WT or mutant B/Victoria/504/2000 HAs were incubated with 34B5 (top panel) or 46B8 (middle and bottom panels) at pH 7.0 (top and middle panels) or 4.8 (bottom panel). The mean fluorescence intensities (MFI) from flow cytometry profiles are shown. Mock, mock transfected cells. (d) Frequency table of potential 46B8-interacting residues. A multiple sequence alignment of 12,790 human IBV HA amino-acid sequences was used to assess the genetic diversity and calculate the frequencies of potential 46B8-interacting residues.

Figure 4. In vivo efficacy of 46B8 in influenza B mouse infection models.

(a) DBA/2 J mice were infected intranasally with a minimum lethal dose of B/Wisconsin/1/2010, B/Brisbane/60/2008, B/Victoria/504/2000, B/Russia/1/1969 or B/Massachusetts/3/1966. At 24, 48 or 72 h post infection, mice received a single treatment of 46B8 or a control IgG intravenously at 15 mg kg⁻¹. Survival curves are shown. Each group contains eight mice. Log-rank tests of all 46B8-treated groups versus the control in all infection models give P<0.05. (b) DBA/2 J mice were infected intranasally with a minimum lethal dose of B/Wisconsin/1/2010, B/Victoria/504/2000 or B/Massachusetts/3/1966. At 72 h post infection, mice received a single treatment of 46B8 or a control IgG intravenously at 5, 15 or 45 mg kg⁻¹. Survival curves are shown. Each group contains eight mice. Log-rank tests of all 46B8 doses versus the control in all infection models give P<0.05.

Figure 5. In vivo efficacy of 46B8 compared with and in combination with Tamiflu.

(a) DBA/2 J mice were infected intranasally with a minimum lethal dose (left panels) or a high lethal dose (right panels) of B/Victoria/504/2000. At 72 h post infection, mice received a single treatment of 46B8 or a control IgG intravenously at 45 mg kg⁻¹, or Tamiflu orally at 100 mg kg⁻¹ twice a day for 5 days. Top panels: survival curves. Each group contains 12 mice. Log-rank tests of 46B8- versus Tamiflu-treated groups in both infection models give P<0.05. Bottom panels: percent of average BW of survived mice as compared with the average pre-infection weight. (b) DBA/2 J mice were infected intranasally with a high lethal dose of B/Victoria/504/2000. At 72 h post infection, mice received a single treatment of 46B8 or a control IgG intravenously at 45 mg kg⁻¹, Tamiflu orally at 100 mg kg⁻¹ twice a day for 3 days, or a combined treatment of 46B8 and Tamiflu. At day 3, 4 and 6 post infection, lung homogenates were prepared and viral titres in the homogenates were determined with a TCID₅₀ assay on MDCK cells. Titres are shown in TCID₅₀ ml⁻¹. Each group contains 10 mice. An unpaired two-tailed t-test of 46B8 versus Tamiflu and the combined treatment versus 46B8 or Tamiflu alone at day 4 and 6 give P<0.05. Dotted line: detection limit of the assay; arrow: treatment initiation point. (c) DBA/2 J mice were infected intranasally with a high lethal dose of B/Victoria/504/2000. At 72 h post infection, mice received a single treatment of 46B8 or a control IgG intravenously at 5 mg kg⁻¹, Tamiflu orally at 100 mg kg⁻¹ twice a day for 5 days, or a combined treatment of 46B8 and Tamiflu. Left panel: survival curves. Each group contains eight mice. A log-rank test of the combined treatment versus 46B8 alone gives P<0.05, and the combined treatment versus Tamiflu alone gives P=0.06. Right panel: percent of average BW of survived mice as compared with the average pre-infection weight. (a,c: mean and s.d.)

Figure 6. Properties of mutant B/Brisbane/60/2008 viruses and HAs.

(a) Plaque reduction assay. 100 plaque-forming units (pfu) of each virus was incubated with varying concentrations of 46B8 for 1 h prior to infection of MDCK cells. At 1 h post infection, the inoculum was removed and cells were overlaid with varying concentrations of 46B8 in agarose. At day 4 post infection, the numbers of plaques were counted for each virus and normalized to the number at the lowest antibody concentration. The assay was done in triplicate. (b) 293 T cells expressing the WT or mutant B/Brisbane/60/2008 HAs were incubated with 34B5 (left panel) or 46B8 (middle and right panels) at pH 7.0 (left and middle panels) or 4.8 (right panel). Flow cytometry profiles are shown. Mock, mock transfected cells. (a: mean and s.e.m.)

Figure 7. Fitness of mutant B/Brisbane/60/2008 viruses.

(a) In vitro fitness. MDCK cells were infected with the WT or mutant B/Brisbane/60/2008 viruses at MOI 0.01. Released viral genome in the supernatant was quantitated at day 4 post infection by qPCR of the viral M1 Matrix gene. Genome copy numbers are shown as histograms. The assay was done in triplicate. (b) In vivo fitness. DBA/2 J mice were infected with WT or mutant viruses at 3 × 10⁴, 1 × 10⁴ or 5 × 10³ plaque-forming units (pfu) per mouse. Survival curves (top panels) and percent of average BW as compared with the average pre-infection weight (bottom panels) are shown. Each group contains six mice. (a: mean and s.e.m.; b: mean and s.d.)

Figure 8. Both WT and mutant B/Brisbane/60/2008 viruses are sensitive to 46B8 in vivo and to Tamiflu.

(a) MDCK cells were infected with WT or mutant viruses at MOI 0.01. After removal of the virus inoculum, cells were grown in the presence of varying concentrations of oseltamivir acid for 4 days. Released viral genome in the supernatant was quantitated by qPCR of the viral M1 Matrix gene. The copy numbers of each virus were normalized to the value at the lowest oseltamivir acid concentration. The assay was done in triplicate. (b) DBA/2 J mice were infected intranasally with a minimum lethal dose of the WT or mutant viruses. At 48 h post infection, mice received Tamiflu orally at 100 mg kg⁻¹ twice a day for 5 days. Survival curves are shown. Each group contains eight mice. Log-rank tests of Tamiflu versus water treated groups for all viruses give P<0.05. (c) DBA/2 J mice were infected intranasally with a minimum lethal dose of the WT or mutant B/Brisbane/60/2008 viruses. At 72 h post infection, mice received a single treatment of 46B8 or a control IgG intravenously at 15 mg kg⁻¹. Survival curves are shown. Each group contains eight mice. Log-rank tests of 46B8 versus control IgG-treated groups for all viruses give P<0.05. (a: mean and s.e.m.)

Figure 9. 46B8 induces ADCC via both WT and mutant B/Brisbane/60/2008 HAs.

(a) A549 cells infected with WT or mutant B/Brisbane/60/2008 viruses were labelled with 46B8 prior to incubation with NK cells. NK cell activation (left panel) and target cell lysis (right panel) were presented as the frequency of CD107a⁺ cells and the percent of LDH release, respectively. The assays were done in duplicate. Results are representative of three independent experiments. (b,c) DBA/2 J mice were infected intranasally with a minimum lethal dose of the WT or mutant B/Brisbane/60/2008 viruses. At 72 h post infection, mice received a single treatment of 46B8, 46B8 N297G or a control IgG intravenously at 15 mg kg⁻¹. (b) Survival curves. Each group contains eight mice. Log-rank tests of 46B8- versus control IgG-treated groups for all viruses, 46B8 N297G- versus control IgG-treated groups for WT, B1 and C2 and 46B8- versus 46B8 N297G-treated groups for A4, B1 and C2 give P<0.05. (c) Percent of average BW of survived mice as compared with the average pre-infection weight. (a: mean and s.e.m.; c: mean and s.d.)