Broad Reactivity Single Domain Antibodies against Influenza Virus and Their Applications to Vaccine Potency Testing and Immunotherapy

Abstract

The antigenic variability of influenza presents many challenges to the development of vaccines and immunotherapeutics. However, it is apparent that there are epitopes on the virus that have evolved to remain largely constant due to their functional importance. These more conserved regions are often hidden and difficult to access by the human immune system but recent efforts have shown that these may be the Achilles heel of the virus through development and delivery of appropriate biological drugs. Amongst these, single domain antibodies (sdAbs) are equipped to target these vulnerabilities of the influenza virus due to their preference for concave epitopes on protein surfaces, their small size, flexible reformatting and high stability. Single domain antibodies are well placed to provide a new generation of robust analytical reagents and therapeutics to support the constant efforts to keep influenza in check.

Keywords: immunotherapy; influenza; nanobodies; single domain antibodies; vaccine potency testing.

Figure 2

Structure of HA and epitope footprints of HA stem binding sdAbs and conventional antibodies with similar ranges of cross-reactivity. (a) Structure of hemagglutinin (A/California/04/2009, PDB code 3LZG) [84]. One unit of the trimer is highlighted, HA1 is shown in green and HA2 is shown in orange. The fusion peptide, which, upon pH mediated conformational change, straightens and inserts into the host membrane is highlighted in cyan. The residues immediately surrounding the HA0 cleavage site are highlighted in yellow. The receptor binding site is shown in pink. (b) The same hemagglutinin structure is used to display a heat map of influenza residue conservation across 6 influenza A subtypes known to infect humans (H1, H2, H3, H5, H7 and H9) [85]. Key residues in the RBS, fusion peptide and cleavage site are highly conserved as well as many residues in the stem region at the HA1/HA2 interface. (c–f) Hemagglutinin trimer structures are used to display binding sites of sdAbs (blue), conventional antibodies (red) and areas of epitope overlap (purple). (c) The original stem binding antibody C179 is displayed on A/California/04/2009. (d) SD38 and F16v3 epitopes displayed on A/Hong Kong/1/68 HA (PDB code 4FNK) [24]. (e) SD36 and CR8043 epitopes displayed on A/Hong Kong/1/68 HA. SD38 and pan influenza A conventional antibodies have a high degree of overlap in their epitope footprints while SD36 and Group 2 stem binding conventional antibodies have sharply contrasting epitope footprints. This may indicate a difference in binding preference between conventional antibodies and sdAbs. (f) SD83 epitope displayed on B/Brisbane/60/08 (PDB code 4FQM) [70]. The SD83 epitope has a greater degree of contact with HA2 compared to SD36 and SD38. The residues included in the epitope footprint were determined based on protein structures in complex (PDB codes 6FYU, 6FYT, 6FYW, 3ZTJ, 4NM8) [69,74,86] using the PBDePISA online tool [87].

Figure 1

Relationship between sites of antibody binding and mechanisms of inhibition of infection of influenza virus. (a) Cartoon of influenza virus, its envelope proteins, the sites targeted by anti-influenza sdAbs and different sdAb formats. HA head (sdAb binding to hemagglutinin head domain), HA stem (sdAb binding to hemagglutinin stem domain), NA (sdAb binding to neuram-inidase), M2 (sdAb binding to M2 ion channel), NP (sdAb binding to nucleoprotein, a component of the vRNP complexes) (b) Cartoon depiction of parts of the influenza virus infection cycle and the various methods by which antibodies inhibit the spread of the virus depending on their epitope. The infected host cell, influenza virus, nucleus and effector cell are indicated (1) Virus attachment to host cell inhibited by HA head antibodies. (2) Endocytosis of virion. (3) pH mediated fusion of virus and host membranes inhibited by HA stem antibodies. (4) pH mediated proton influx inhibited by some M2 antibodies interfering with uncoating. (5) Transport of ribonucleoproteins to the nucleus inhibited by NP antibodies. (6) Inhibition of dissociation of new virions from host cells by anti-NA antibodies. (7) Immune effector functions such as antibody dependent cellular cytotoxicity (ADCC) mediated by Fc-linked HA stem, NA and M2 antibodies.

Figure 3

Single domain antibody formats used to optimize functional targeting of influenza virus. sdAb specific to different targets on influenza can be optimized through re-formatting: bivalent format to enhance potency [72,75,83,90,108]; multi-domain fusions to enhance potency and breadth of reactivity [74]; multi-domain Fc fusions to enhance potency, breadth of reactivity, extent half-life to reduce dosing and incorporate effector function [74]; Fc fusion to enhance potency, extend half-life to reduce dosing and incorporate effector function [74,90,109]; alternative fusion partners to enhance potency and extend half-life to reduce dosing [82,110].

Figure 4

Isolation and epitope mapping of single domain antibodies against influenza hemagglutinin (HA) and applications in vaccine potency testing. (a) Adapted from [73]. Generation of panels of high affinity sdAbs against influenza HA using immunisation of alpacas with HA from key group 1, group 2 IAV and IBV strains. Construction of a sdAb (nanobody) display library and selection for antigen specific sdAbs with both broad reactivity and lineage specific binding for use in assessment of influenza vaccine potency assays are followed by epitope mapping using yeast display and mutational scanning, which includes design of a library of HA variants, display of the library on the yeast cell surface, selection using flow cytometric cell sorting to enrich HA variants that lose binding to sdAbs but retain display of correctly folded HA on yeast cell surface. Functional loss of binding is experimentally determined to confirm residues that are energetically important and contribute to the sdAb epitope. The epitope is then correlated with HA structure. Further analysis through deep sequencing of selection outputs can then be used to identify mutational ‘hotspots’ and also ‘coldspots’ where mutations are predicted to have little effect on binding of sdAb reagents. The frequencies of mutations at each position in HA relative to the non-selected population can be determined using bioinformatic analysis. It is envisaged that this approach can be used to generate a database of epitopes corresponding to a diverse collection of sdAbs recognising HA, which upon the emergence of a new viral strain can be used to predict which sdAbs could be chosen as suitable binding reagents for applications such as diagnosis, research, immune surveillance and vaccine potency testing. (b) ELISA assay for vaccine potency testing: A sandwich pair of sdAbs is used to measure HA content in influenza vaccines samples. sdAb-1 is biotin labelled and after coating onto a streptavidin ELISA plate is used as a conformational specific capture reagent for HA either from a vaccine sample or an influenza antigen standard with a known content of antigenically active HA. sdAb-2 is the detecting reagent binding to the HA captured on the ELISA plate. sdAb is labelled with a myc epitope tag which is used to detect binding. (c) A schematic representation of a vaccine potency assay readout and how it is used to evaluate HA content using comparison to a standard curve containing a known quantity of HA.