Isolation and characterization of IgG3 glycan-targeting antibodies with exceptional cross-reactivity for diverse viral families

Abstract

Broadly reactive antibodies that target sequence-diverse antigens are of interest for vaccine design and monoclonal antibody therapeutic development because they can protect against multiple strains of a virus and provide a barrier to evolution of escape mutants. Using LIBRA-seq (linking B cell receptor to antigen specificity through sequencing) data for the B cell repertoire of an individual chronically infected with human immunodeficiency virus type 1 (HIV-1), we identified a lineage of IgG3 antibodies predicted to bind to HIV-1 Envelope (Env) and influenza A Hemagglutinin (HA). Two lineage members, antibodies 2526 and 546, were confirmed to bind to a large panel of diverse antigens, including several strains of HIV-1 Env, influenza HA, coronavirus (CoV) spike, hepatitis C virus (HCV) E protein, Nipah virus (NiV) F protein, and Langya virus (LayV) F protein. We found that both antibodies bind to complex glycans on the antigenic surfaces. Antibody 2526 targets the stem region of influenza HA and the N-terminal domain (NTD) region of SARS-CoV-2 spike. A crystal structure of 2526 Fab bound to mannose revealed the presence of a glycan-binding pocket on the light chain. Antibody 2526 cross-reacted with antigens from multiple pathogens and displayed no signs of autoreactivity. These features distinguish antibody 2526 from previously described glycan-reactive antibodies. Further study of this antibody class may aid in the selection and engineering of broadly reactive antibody therapeutics and can inform the development of effective vaccines with exceptional breadth of pathogen coverage.

Fig 1. Discovery and validation of broadly-reactive mAbs 2526 and 546.

(A) A diverse panel of HIV-1 Envs and influenza HAs were used to screen B cell receptor specificity of a donor chronically infected with HIV-1. B cell LIBRA-seq scores (LSS) for HIV-1 Env ZM197 and influenza HA NC99 identified potential broadly-reactive antibodies; red dots indicate antibodies 2526, 546, and 706. Sequence analysis revealed that all three antibodies came from the same lineage. VH and VL gene usage, CDR length, percent amino acid identity to germline, and CDR sequence for antibodies are shown. (B) 2526 and 546 cross-reactivity was validated against a diverse panel of antigens by ELISA and reactivity was compared to previously described cross-reactive mAbs. Representative ELISA area under the curve (AUC) from a set of three repeats are displayed. (C) 2526, 546, and mAb 688 bind to Nipah virus F (NiV-F) protein, and to a lesser extent Langya virus F (LayV-F) protein. Representative ELISA curves are shown.

Fig 2. 2526 and 546 bind to complex glycans on their target antigens.

(A) 2526 and 546 were shown to be glycan reactive by enzymatic treatment of antigens with PNGase-F to remove N-linked glycans or competition with 1M mannose in an ELISA format. ELISA AUC normalized to native antigen is shown. Representative ELISA area under the curve (AUC) from a set of three repeats are displayed. (B) Negative stain electron microscopy confirmed that 2526 takes on the canonical “Y” shape of antibodies instead of the “I” shape of FDG antibodies. (C) 2526, 546, and 688 show a decrease in binding to antigen made in GnT1- cells, suggesting that they preferentially target complex glycans. Closed circles indicate the specified antibody was tested against antigen made in 293F cells while open circles indicate the antibody was tested against antigen made in GnT1- cells. (D) 2526 prefers to bind to BG505.DS.SOSIP compared to BG505.SOSIP when tested with SPR. Representative SPR curves from a set of two repeats are shown.

Fig 3. 2526 does not display signs of autoreactivity.

(A) HEp-2 cell slides were immunofluorescently stained with the indicated mAbs and goat anti-human Ig-FITC secondary (green) and DAPI (blue, to stain cell nuclei). Each mAb was tested at 100, 10, and 1 μg/mL. Control human serum samples that were positive or negative for anti-nuclear antibodies are shown (right). Images are 40X magnification. (B) Autoantigens were tested for reactivity by 2526, Ab82 (negative control antibody), and A32 (positive control antibody) using a commercially available AtheNA Multi-Lyte ANA kit. Values shown in bold were considered positive.

Fig 4. X-ray crystallography identified a functional glycan-binding pocket on the side of the light chain of 2526.

(A) Crystal structure of 2526 IgG1 Fab complexed with D-Mannose is shown. The light chain is shown in cyan and the heavy chain is shown in green. Pink, blue, and red denote CDR loops. D-Mannose is shown in white and red, with its binding pocket in the crystal structure highlighted in surface representation (right) with key amino acid residues interacting with the mannose moiety shown as sticks (bottom). (B) Mutating residues in the binding pocket decreased binding of 2526 IgG3 to HIV-1 Env (KNH1209.18.DS.SOSIP), SARS-CoV-2 spike (Index strain), and influenza HA (Michigan/2016 H1). (C) Germline reversion of 2526 IgG3, whereby the amino acid sequences of the heavy and light chain were reverted to their inferred germline sequences whilst keeping the CDR3 regions mature, ablated binding to HIV-1 Env (KNH1209.18.DS.SOSIP) and SARS-CoV-2 spike (Index strain), but not to influenza HA (New Caledonia/1999 H1). Keeping the light chain mature while reverting the heavy chain to germline rescued binding. HC, heavy chain. LC, light chain. Representative ELISA area under the curve (AUC) from a set of three repeats are displayed.

Fig 5. 2526 Epitope Mapping for influenza HA, SARS-CoV-2 spike, and HIV-1 Env.

(A) 2526 was tested against an expanded panel of influenza A HA strains in ELISA and was found to preferentially bind to H1 HAs compared to H3 HAs. Binding of 2526 to an H1 stem-only stabilized construct (H1 Stem) suggested that 2526 targets the stem region. Removal of the stem glycan N23 dramatically decreased binding. WT, wild type HA. (B) 2526 was tested against full SARS-CoV-2 spike as well as subunit constructs in an ELISA format, binding strongly to the full spike as well as the N-terminal domain (NTD) construct. RBD, receptor binding domain. NTD subunit constructs with individual glycans knocked out revealed a decrease in binding of 2526 to NTD when glycan 149 was removed. (C) 2526 was tested for binding to BG505.SOSIP with and without the presence of a glycan at position 332 (left). In a competition ELISA for 2526 against a panel of HIV-1 antibodies, 2526 displayed measurable competition against itself, but not competition against a panel of known HIV-1 antibodies (right).

Fig 6. In-vitro and in-vivo functional characterization of 2526.

(A) 2526 was found to be non-neutralizing against a panel of eight HIV-1 pseudovirus strains, three influenza strains, and four SARS-CoV-2 strains. (B) 2526 was tested for potential as a prophylactic and therapeutic agent in a mouse model of influenza challenge. A schematic of both study arms is shown with the H1N1 A/Brisbane/2018 influenza strain being administered prior to antibody treatment in the therapeutic arm and vice versa for the prophylactic arm. Lung viral titers were quantified 3 days post-infection (indicated by asterisk), with no significant differences between groups that received 2526 compared to negative control groups that received PBS or the isotype control. Comparisons of weight loss between groups showed that while mice receiving either 2526 or 2526 LALA-PG (Fc effector functions knocked out) did display less weight loss compared to mice receiving PBS, they did not perform better than mice that received a negative isotype control antibody that displays no reactivity to influenza antigens.