Epistasis reduces fitness costs of influenza A virus escape from stem-binding antibodies

Abstract

The hemagglutinin (HA) stem region is a major target of universal influenza vaccine efforts owing to the presence of highly conserved epitopes across multiple influenza A virus (IAV) strains and subtypes. To explore the potential impact of vaccine-induced immunity targeting the HA stem, we examined the fitness effects of viral escape from stem-binding broadly neutralizing antibodies (stem-bnAbs). Recombinant viruses containing each individual antibody escape substitution showed diminished replication compared to wild-type virus, indicating that stem-bnAb escape incurred fitness costs. A second-site mutation in the HA head domain (N129D; H1 numbering) reduced the fitness effects observed in primary cell cultures and likely enabled the selection of escape mutations. Functionally, this putative permissive mutation increased HA avidity for its receptor. These results suggest a mechanism of epistasis in IAV, wherein modulating the efficiency of attachment eases evolutionary constraints imposed by the requirement for membrane fusion. Taken together, the data indicate that viral escape from stem-bnAbs is costly but highlights the potential for epistatic interactions to enable evolution within the functionally constrained HA stem domain.

Keywords: HA stem; antigenic escape; epistasis; evolution; influenza A virus.

Fig. 1.

Dynamics of mutation frequency and resistance of passaged NL09 virus populations to stem-bnAbs. NL09 virus was combined with (A, C, and E) 70-1F02 or (B, D, and F) 05-2G02 antibody and serially passaged 10 times with increasing concentrations of antibody. A single replicate was performed with an initial concentration of 20 μg/mL (A and B), 2 μg/mL (C and D), or 1 μg/mL (E and F) of 70-1F02 or 05-2G02, for a total of three replicates per stem-bnAb. n.t. = not tested; the dotted line indicates the limit of detection (5%). The earliest passaged populations that showed PRNT₅₀ titers >40 µg/mL or P7 populations were sequenced. In panel F, the P6 variant was sequenced instead of the P7 variant because the titer of the P7 variant was not sufficient for sequencing. Amino acid positions are indicated with H1 numbering.

Fig. 2.

Identification of stem-bnAb escape mutations. (A and B) PRNT₅₀ titers of recombinant NL09 viruses harboring individual mutations found in the passaged NL09 variants. Concentrations of (A) 70-1F02 and (B) 05-2G02 antibodies leading to a 50% reduction in plaque counts are shown. (C) Locations of selected escape mutations on the HA of A/California/04/2009 (H1N1) (PDB ID 3LZG) (39).

Fig. 3.

By diminishing antibody binding, the identified escape mutations enable HA-membrane fusion in the presence of stem-bnAbs. (A and B) Binding between mutant HA and the stem-bnAbs (A) 70-1F02 or (B) 05-2G02. The mean and SD of four independent experiments are plotted. Bmax and Kd values with 95% CI were calculated using GraphPad Prism 9.5.0. (C) HA-mediated membrane fusion was evaluated by monitoring for syncytia formation at low pH. Red rectangles indicate that syncytia formation was observed.

Fig. 4.

Replication of recombinant viruses harboring individual escape mutations. (A) MDCK cells were infected at an MOI of 0.002 PFU/cell; (B) A549 cells and (C) NHBE cells were infected with MOI = 0.01 PFU/cell. The mean and SD of at least three independent experiments are plotted. Statistical significance was assessed by two-way ANOVA with Sidak’s multiple comparisons test. ****, < 0.0001; ***, < 0.001; **, < 0.01; *, < 0.05.

Fig. 5.

The effect of escape mutations on virus and membrane fusion. (A) Fusion pH of NL09 HA harboring individual escape mutations was measured by syncytium formation. Red rectangles mark the highest pH at which syncytia were observed. (B) Kinetics of low pH-induced conformational change of NL09 HA proteins containing single escape mutations, as measured by ELISA with a conformation-sensitive antibody. The mean and SD of five independent replicates are plotted. (C) Time to 50% dissociation of the conformation-sensitive antibody (half-life). The mean and SD of five independent replicates are plotted.

Fig. 6.

HA N129D does not markedly alter receptor specificity but enhances binding to human-like synthetic glycans. (A) Profiles of virus binding to v4.0 of the CFG glycan array are shown as the relative intensity of the fluorescence (RFU) for each glycan structure, indicating virus recognition and binding. The different linkage of sialylated glycans are highlighted by yellow = 2,3 SIA, blue = 2,3 and 2,6 SIA, and pink = 2,6 SIA. (B) The binding of each virus to human-like (6′SLN) and avian-like (3′SLN) was measured by a solid-phase binding assay. The mean and SD of five independent replicates are plotted. Statistical significance was analyzed by two-way ANOVA with Sidak’s multiple comparisons test. ****, < 0.0001; ***, < 0.001.

Fig. 7.

N129D mutation enhances the replication of NL09 viruses harboring individual escape mutations in NHBE cells and facilitates evolution in HA2 in the presence of a stem-bnAb. (A–D) Viral replication in NHBE cells. The mean and SD of at least three independent replicates are plotted. Statistical significance was assessed using a two-way ANOVA with Sidak’s multiple comparisons test. **, < 0.01; * < 0.05. (E–G) NL09 virus or (H–J) NL09-HA-N129D viruses were combined with 70-1F02 antibody and serially passaged four times with increasing concentration of antibody. Three replicates lineages were evaluated, with an initial concentration of 2 μg/mL.