Focused antibody response to influenza linked to antigenic drift

Abstract

The selective pressure that drives antigenic changes in influenza viruses is thought to originate from the human immune response. Here, we have characterized the B cell repertoire from a previously vaccinated donor whose serum had reduced neutralizing activity against the recently evolved clade 6B H1N1pdm09 viruses. While the response was markedly polyclonal, 88% of clones failed to recognize clade 6B viruses; however, the ability to neutralize A/USSR/90/1977 influenza, to which the donor would have been exposed in childhood, was retained. In vitro selection of virus variants with representative monoclonal antibodies revealed that a single amino acid replacement at residue K163 in the Sa antigenic site, which is characteristic of the clade 6B viruses, was responsible for resistance to neutralization by multiple monoclonal antibodies and the donor serum. The K163 residue lies in a part of a conserved surface that is common to the hemagglutinins of the 1977 and 2009 H1N1 viruses. Vaccination with the 2009 hemagglutinin induced an antibody response tightly focused on this common surface that is capable of selecting current antigenic drift variants in H1N1pdm09 influenza viruses. Moreover, amino acid replacement at K163 was not highlighted by standard ferret antisera. Human monoclonal antibodies may be a useful adjunct to ferret antisera for detecting antigenic drift in influenza viruses.

Figure 6. Neutralizing activities by sera and monoclonal antibodies from vaccinated individuals against the variant viruses.

(A) Sera from recently vaccinated individuals born between 1943 and 1982 (n = 28) and 1983 and 1988 (n = 13) were tested for neutralization of X179A and HA variants carrying single amino acid substitutions. Sera from two individuals (donor H and TIV-20) failed to neutralize the K163E variant of X179A, and their response was focused on the Sa site. Sera from two individuals born between 1983 and 1988 were focused on K130 (Ca2 site). Mean and 95% confidence interval are shown. (B) Serum after vaccination and the V_H 3-7 antibody T2-6A from donor H, together with serum after vaccination from another donor (TIV-20), neutralized X179A strongly but failed to neutralize the K163E virus variant. Their neutralizing activity was unaffected by the K130E substitution. T1-3B and control serum had similar neutralization titers to all the viruses. In our MN assays, the OD₄₅₀ readout was dependent on NP expression following viral entry (65). The MN assays were performed in duplicate.

Figure 5. Mapping of the HA1 K163 residue in former H1N1 viruses.

Conserved HA1 residue HA1 K163 (shown in blue), which was selected for substitution by the V_H 3-7*01 and 3-15*01 antibodies, is situated in a patch (colored gold) of surface conserved among A/California/07/2009, A/Brisbane/59/2007, A/USSR/90/1977, and A/South Carolina/1/1918 HAs. Conserved residues were mapped with PYMOL version 1.7 onto PDB:3LZG (15).

Figure 4. Mapping of K163E/Q substitutions selected by head-specific T2-6A (V_H 3-7) and head cross-reactive T3-4B (V_H 3-15) antibodies.

K130E, G155E, and G170R (shown in red) resulted in loss of binding by the 2-12C control antibody. The G170R amino acid substitution, coselected by the control antibody 2-12C, is thought to be coincidentally associated with K130E. The orange triangle indicates the position of the sialic acid–binding site. K163E/Q is shown in blue. Mutations were mapped with PYMOL version 1.7 onto PDB:3LZG (15).

Figure 3. Binding activity of HA head-specific and head cross-reactive antibodies on cells infected with the variant viruses.

A selection of head-specific (V_H 3-7, 3-33, 1-18) and head cross-reactive (V_H 3-15) antibodies and controls, including the stem cross-reactive antibody T1-3B (V_H 1-69), were tested for binding to MDCK-SIAT1 cells infected with variant viruses selected with monoclonal antibody T2-6A (V_H 3-7) that selected the K163E substitution, T3-4B (V_H 3-15) that selected K163Q, and the control 2-12C (V_H 5-51) that selected (K130E/G170R). The assay was done twice with equivalent results.

Figure 2. The 6 largest V_H-related sets of HA-binding antibodies derived from day 7 Eng195 HA-specific B cells.

(A) The average number of V_H nucleotide and amino acid replacements in the 6 largest sets of HA-binding antibodies. The representative HAI and MN results of head-specific (Head S), head cross-reactive (Head CR), and stem cross-reactive (Stem CR) antibodies are summarized in the figure. –, negative; +, EC₅₀ ~366 ng/ml; ++, EC₅₀ ~70 ng/ml; +++, EC₅₀ ~7 ng/ml for neutralization of X179A virus (A/California/07/2009 H1pdm09). (B) Neutralization titers of head-specific, head cross-reactive, and stem cross-reactive antibodies for former H1, H3, H1pdm09, and H5 viruses. Two H1 head-specific, 15-2A06 and 19-4G05 (22), and 5 stem-reactive, FI6v3 (24), V3-2G6 (23), SF70-1F02 (25), and 05-2G02 and 09-3A01 (22), reference antibodies were included. Each symbol represents an independent measurement. Some measurements overlap. MN titers were assayed at least twice for each antibody. The mean EC₅₀ for neutralization for the 6 largest sets of HA-binding antibodies in the MN assay was analyzed by 1-way ANOVA, and a P value less than 0.05 was considered significant.

Figure 1. The serological neutralization titers for former seasonal H1, H3, H1pdm09, and H5-pseudotyped influenza viruses in donor H before and after trivalent seasonal influenza vaccination.

Donor H received the 2008–2009, 2009–2010, and 2011–2012 influenza vaccines, which contain antigens from the viruses listed in Table 1. The assay was performed in duplicate and means are shown. Pseudotyped influenza viruses are more sensitive to stem antibodies than wild-type viruses. TIV, trivalent seasonal influenza.