Identification and structural characterization of a broadly neutralizing antibody targeting a novel conserved epitope on the influenza virus H5N1 hemagglutinin

Abstract

The unabated circulation of the highly pathogenic avian influenza A virus/H5N1 continues to be a serious threat to public health worldwide. Because of the high frequency of naturally occurring mutations, the emergence of H5N1 variants with high virulence has raised great concerns about the potential transmissibility of the virus in humans. Recent studies have shown that laboratory-mutated or reassortant H5N1 viruses could be efficiently transmitted among mammals, particularly ferrets, the best animal model for humans. Thus, it is critical to establish effective strategies to combat future H5N1 pandemics. In this study, we identified a broadly neutralizing monoclonal antibody (MAb), HA-7, that potently neutralized all tested strains of H5N1 covering clades 0, 1, 2.2, 2.3.4, and 2.3.2.1 and completely protected mice against lethal challenges of H5N1 viruses from clades 1 and 2.3.4. HA-7 specifically targeted the globular head of the H5N1 virus hemagglutinin (HA). Using electron microscopy technology with three-dimensional reconstruction (3D-EM), we discovered that HA-7 bound to a novel and highly conserved conformational epitope that was centered on residues 81 to 83 and 117 to 122 of HA1 (H5 numbering). We further demonstrated that HA-7 inhibited viral entry during postattachment events but not at the receptor-binding step, which is fully consistent with the 3D-EM result. Taken together, we propose that HA-7 could be humanized as an effective passive immunotherapeutic agent for antiviral stockpiling for future influenza pandemics caused by emerging unpredictable H5N1 strains. Our study also provides a sound foundation for the rational design of vaccines capable of inducing broad-spectrum immunity against H5N1.

Fig 1

Detection of inhibitory activity of MAb HA-7 on influenza virus entry by neutralization assays. (A) Pseudovirus neutralization assay. The data are presented as the mean percentages of neutralization ± standard deviations for duplicate wells of MAb HA-7 at concentrations of 0.7 and 10 μg/ml, respectively, against HAs of H5N1 pseudovirus covering four different clades, including homologous strain AH-HA (clade 2.3.4) and heterologous strains HK-HA (clade 0), XJ-HA (clade 2.2), QH-HA (clade 2.2), and HB-HA (clade 2.3.2.1). (B) Live-virus-based neutralization assay. Shown are data for neutralizing activities of HA-7 against strains VN/1194 (clade 1), SZ/406H (clade 2.3.4), and HK/213 (clade 1) of H5N1; S-OIV H1N1 (A/Beijing/501/2009 pdm strain); and seasonal IAVs. The neutralizing antibody titer was defined as the highest dilution of MAb that completely suppressed the virus-induced CPE in 50% of the wells. Panels A and B show representative data from three independent repeat experiments.

Fig 2

Detection of cross-protection of HA-7 neutralizing MAb against lethal H5N1 virus challenge. (A and B) Survival rate of HA-7-injected mice after challenge with 5 LD₅₀ of strain VN/1194 (clade 1) (A) or SZ/406H (clade 2.3.4) (B) of H5N1 virus. (C and D) Percent body weight change of HA-7-treated mice after challenge with 5 LD₅₀ of the VN/1194 (C) and SZ/406H (D) strains of H5N1 virus. A MAb targeting the RBD of SARS-CoV was used as the control (Ctrl). Panels A to D show representative data from two independent repeat experiments with six mice per group.

Fig 3

Virus titers and histopathological changes in the lungs of MAb-treated mice challenged with H5N1 viruses. HA-7-injected mice were challenged with 5 LD₅₀ of clade 1 strain VN/1194 and clade 2.3.4 strain SZ/406H of H5N1, and lung tissues were collected at day 5 postchallenge. (A) Detection of viral titers of VN/1194 and SZ/406H H5N1 viruses in collected lung tissues. Mice injected with SARS-CoV RBD-specific MAb were used as the negative control (Ctrl). The data are expressed as means ± standard deviations of virus titers (log₁₀ TCID₅₀/g) of lung tissues from six mice per group. The limit of detection is 1.5 log₁₀ TCID₅₀/g of tissues, as shown by the dotted line. The figure shows representative data from two independent repeat experiments. (B and C) Evaluation of histopathological changes in collected lung tissues. Lung tissues from the mice injected with a MAb to the RBD of SARS-CoV and those from uninfected mice were used as the negative control (Ctrl) and normal control (Normal), respectively. Lung tissues were stained with H&E and observed by light microscopy (magnification, ×100). Shown are representative images of histopathological changes from two independent experiments with six mice per group challenged with VN/1194 (B) and SZ/406H (C) H5N1 viruses, respectively. Serious histopathological damage was seen in the control mice, with bronchial epithelial cell degeneration, necrosis, and desquamation (solid arrows). Large amounts of exudates and severe edema with infiltration of lymphocytes and mononuclear cells were also seen around blood vessels (open arrows). However, the HA-7-immunized mice showed only mild interstitial pneumonia changes with focal broadening interstitial spaces and lymphocytic infiltration (arrowheads).

Fig 4

Detection of specificity and subtypes of MAb HA-7 by ELISA. (A) Reactivity of total IgG of HA-7 with fusion proteins containing full-length HA1 (residues 13 to 325 of H5 numbering [21, 25, 46], corresponding to HA1 residues +3 to 322 [30]) and truncated fragments of HA1 of A/Anhui/1/2005(H5N1) virus. ELISA plates were respectively coated with recombinant HA1 proteins fused with human Fc (HA1-Fc), Fd sequence (HA1-Fd), or Fd plus Fc (HA1-Fdc) and HA1 protein without Fd and Fc (HA1-His). Recombinant hIgG1-Fc2 protein (rFc), commercial human IgG Fc protein (IgG-Fc), Fd fused with HIV-1 gp41 (HIV-Fd), and SARS-CoV RBD protein were used as the controls. (B) Reactivity of total IgG of MAb HA-7 with inactivated IAVs containing H5N1, S-OIV H1N1, and seasonal influenza viruses (IAVs). (C) Detection of IgG subtypes of MAb HA-7 using recombinant HA1-His protein as the coating antigen. The data are presented as the mean absorbance at 450 nm (A₄₅₀) ± standard deviations of data from duplicate wells of MAbs at a dilution of 1:3,200. Panels A to C show data from three independent experiments. (D) Amino acid sequences for HA-7 Fab heavy (V_H) and light (V_L) chains. CDR1, CDR2, and CDR3 are labeled.

Fig 5

Epitope mapping and detection of reactivity of MAb HA-7 with denatured recombinant proteins by ELISA. (A) ELISA plates were coated with truncated recombinant HA1 protein fragments covering the full-length HA1 region, followed by the addition of MAb HA-7 to detect the reactivity of total IgG. The data are presented as mean A₄₅₀ values ± standard deviations of data from duplicate wells of MAbs at a dilution of 1:3,200. Shown are representative data from three independent repeat experiments. (B) Reactivity of MAb HA-7 with DTT-denatured recombinant HA1 protein fragments, using corresponding proteins without DTT treatment as a comparison. The data are presented as mean A₄₅₀ values ± standard deviations of data from duplicate wells of MAb at a dilution of 1:51,200.

Fig 6

Mechanism study of MAb HA-7 against H5N1 pseudovirus infection. (A) Virus binding assay with QH-HA pseudovirus-infected MDCK cells. The antibodies were diluted to 10 μg/ml, 100 μg/ml, and 1 mg/ml, respectively, and binding ability was measured by quantification of HIV p24 content by ELISA. A MAb to the RBD of SARS-CoV (33G4) and human IgG-Fc were used as controls. The data are presented as mean A₄₅₀ values ± standard deviations of data from duplicate wells. (B) Inhibition by HA-7 of the postattachment process in QH-HA pseudovirus-infected MDCK cells. A MAb to the RBD of SARS-CoV (33G4) and human IgG-Fc were used as the negative controls. No-Ab and no-HA wells were included as the system control. The data are presented as mean percentages ± standard deviations for duplicated wells. Panels A and B show representative data from three independent repeat experiments.

Fig 7

3D-EM-based model docking and epitope analysis of HA-7. (A) Docking of H5HA trimer (red) into the EM density map (yellow). (B) Docking of Fab (green) into its density map (cyan). (C) Threefold top view of the trimeric HA-7Fab/H5HA complex. (D) Side view of the trimeric HA-7Fab/H5HA complex. (E) Zoom-in side view of the HA/Fab interface. Residues 81 to 83 and 117 to 122 of HA1 (black) are in contact with Fab, while the receptor (sialic acid [blue spheres]) is not. (F) Zoom-in view from Fab showing that residues 81 to 83 and 117 to 122 of HA1 (black) lie at the center of Fab's “footprint” (yellow) on HA and comprise the core epitope. The receptor (sialic acid [light blue spheres]) is in the background.