Vanderbilt Institute of Chemical Biology



Discovery at the VICB







A New Target for Influenza Vaccine Development 


By: Carol A. Rouzer, VICB Communications
Published:  July 18, 2018


A human monoclonal antibody that binds to a novel epitope on the HA protein protects against potentially dangerous H3N2 variant viral strains.


As anyone who faithfully receives his/her flu shot every year knows, seasonal influenza is the only common infectious disease that requires a new vaccination on an annual basis. The reason for this is rapid evolution of the virus, leading to antigenic drift that enables it to escape immunity developed as a result of vaccination or exposure during previous seasons. This problem is particularly relevant in the case of the H3 strain, one of the two major influenza A subtypes that cause seasonal flu. For example, H3N2 viruses, which appeared first in 1968, cause a higher level of morbidity and mortality than most other subtypes, and they exhibit an especially high rate of antigenic drift. Furthermore, we are now also seeing the appearance of sporadic disease caused by H3N2v variant viruses of swine origin. H3N2vs are reassortants of human, swine, and avian viruses with a high potential to cause pandemic disease, and because they are antigenically distinct from seasonal viruses, current vaccines are ineffective against them. A major goal of infectious disease research is to develop a broadly effective vaccine that provides long-lasting protection; however, such a vaccine will not be successful unless it also protects against H3N2v viral strains. This led Vanderbilt Institute of Chemical Biology member James Crowe, his colleague Ian Wilson (Scripps Research Institute), and their laboratories to explore a novel human monoclonal antibody (mAb) that inactivates many human and swine H3N2 strains [S. Bangaru, et al. Nat. Comm., (2018) 9, 2669].


There are two major glycoproteins on the influenza virus surface, hemagglutinin (HA, Figure 1), and neuraminidase (NA). HA binds to target cell surface glycoproteins that contain sialic acid, initiating the first step of infection. NA's role is to cleave those same sialic acid residues from the cell surface at the end of the infection cycle so that the virus can escape. HA is the primary target of most neutralizing antibodies against influenza virus. It is a trimer of monomers, each of which is transcribed as a single protein chain that is cleaved into two chains, designated HA1 and HA2. The monomers can also be thought of as comprising a receptor-binding subdomain (RBS) and a vestigial esterase subdomain that together form a globular head in addition to a stem domain that promotes fusion of the viral membrane with endosomes after uptake into the cell. Most antibodies that target HA either bind to the RBS, thereby blocking cell attachment, or to the stem, preventing endosomal fusion. The stem domain is highly conserved, so much of the effort to develop broadly neutralizing and long-lasting vaccines have focused on this region of the protein.


FIGURE 1. Ribbon representation of the influenza HA protein taken from PDB #4JUG. Each of the three subunits is colored in an increasingly dark shade of green (HA2 chain) or orange/gold (HA1 chain). The head and stem domains are labeled. The receptor binding subdomain (RBS) is located at the top of the head.



In their effort to directly address the threat of H3N2v viruses, the Vanderbilt Vaccine Research Program with support from the NIH Vaccine and Treatment Evaluation Unit contract had immunized a group of human volunteers with an experimental vaccine based on the HA protein from a novel swine "variant" strain of H3N2 virus (the H3N2v virus strain A/Minnesota/11/2010). From antibodies generated in response to this vaccine, the Crowe laboratory developed a number of mAbs, of which H3v-47 was of particular interest. Further characterization of this mAb in the current research demonstrated that it strongly bound and neutralized the HA protein from H3N2 strains prevalent from 1989 through 2014. It also neutralized the HA from four different H3N2 swine strains tested, a property not previously observed for any human mAb.


To determine if the mAb could protect against influenza in vivo, the investigators developed a mouse model of lethal H3N2v (A/Minnesota/11/2010) infection. They found that two different doses of the H3v-47 mAb protected the mice from disease and death whether given prior to infection (preventive) or after infection (therapeutic). These findings suggested that vaccines capable of inducing formation of antibodies that act similarly to H3v-47 would offer significant protection against H3N2v-mediated disease.


To develop a vaccine that elicits antibodies similar to H3v-47, it is necessary to know exactly where on the HA protein the mAb binds. To obtain this information, the researchers conducted a competitive binding study, pitting H3v-47 against other mAbs that bind to the HA protein at known sites. The results showed that H3v-47 did not interfere with the binding of antibodies that targeted the HA stem, but it partially interfered with binding of antibodies directed against the RBS, suggesting a binding site in or near the globular head. This result was a surprise because prior work had shown that H3v-47 did not prevent binding of HA to cell surface glycoproteins. Thus, they had anticipated that it would be directed toward the stem domain rather than the RBS-containing head domain.


To determine the exact site of H3v-47 binding to HA, researchers in the Wilson laboratory obtained an X-ray crystal structure of the protein from H3N2v A/Minnesota/11/2010 complexed with the Fab fragment of the mAb (Figure 2). The results showed that H3v-47 binds to the head domain, but not in the portion where cell attachment occurs. Rather, its binding site spans the vestigial esterase domain and the adjacent portion of the head, beneath the site of glycoprotein attachment. These results demonstrated that H3v-47 exploited a protein epitope that was not utilized by any other known mAb.


FIGURE 2. Ribbon representation of the crystal structure of the HA from the H3N2v flu virus strain A/Minnesota/11/2010 bound to the Fab fragment of the H3v-47 antibody. (a) The entire structure is shown with one subunit of the HA protein colored in yellow (HA1) and orange (HA2), the Fab heavy chain in green, and the light chain in cyan. N-linked glycans are shown as colored balls. (b) A close-up of the binding site of the Fab fragment to HA. The HA receptor binding subdomain is shown in yellow, and the vestigial esterase subdomain in tan. The antigenic sites on the HA protein are highlighted in red and orange, and the Fab protein is colored as in (a). Figure reproduced under the Creative Commons Attribution 4.0 International License 4.0 from S. Bangaru, et al. Nat. Comm., (2018) 9, 2669.



Studies of the natural evolution of H3 influenza viruses revealed that in 1989, two mutations – E82K and G124S – occurred in the H3v-47 HA binding site. Both K82 and S124 are involved in interactions with H3v-47. In fact, K82E and/or G124S mutations of the HA from H3N2v A/Minnesota/11/2010 resulted in substantial reductions in H3v-47 binding affinity, particularly of the Fab fragment. These findings help to explain why H3v-47's neutralizing activity is greatest against HA proteins from H3N2 viruses that have circulated since 1989.


One mechanism used by the influenza virus to evade the immune system is acquisition of new glycosylation sites that interfere with antibody binding. Indeed, over the years, the number of such sites has increased from 2 to as many as 8 among the HAs of seasonal flu viruses. The HA from H3N2v A/Minnesota/11/2010 used to derive the H3v-47 mAb contains 4 glycosylation sites, and based on the locations of sites in other HA proteins, it is plausible that the addition of more sites might reduce H3v-47's affinity. However, when the researchers engineered an H3N2v HA protein with three additional glycosylation sites predicted to cause interference with binding, the H3v-47 mAb did not lose its ability to neutralize the protein. Thus, the investigators concluded that the virus could not easily use this mechanism to develop resistance to H3v-47.


Due to H3v-47's rather unconventional binding site, it was unclear exactly how it neutralizes the target HA. The researchers had already ruled out blockade of cell surface binding, so they next evaluated the possibility that H3v-47 prevents endosomal membrane fusion. However, the mAb was unable to block the low pH-induced conformational changes that are required for HA to promote membrane fusion. These findings suggested that H3v-47 would not prevent virus entry into the cell or endosomal processing. This led the investigators to consider the possibility that H3v-47 interferes with egress of the virus from the cell. They tested this possibility in MDCK cells, comparing the effects of H3v-47 to those of zanamivir, an NA inhibitor known to block viral egress. In support of their hypothesis, the researchers used transmission electron microscopy to show that in the presence of H3v-47, H3N2v (A/Minnesota/11/2010) remained clustered at the surface of infected MDCK cells rather than escaping freely into the surrounding medium. Nearly identical results were obtained in zanamivir-treated cells (Figure 3a). In contrast, the stem-binding mAb CR8020 had no effect on viral egress. Using a secondary antibody complexed to gold particles, the researchers found H3v-47 localized between clumped viral particles or between viral particles and the cell membrane. These findings suggest that the mAb might work by tethering the virus particles to each other or to the cell surface (Figure 3b).



FIGURE 3. (a) Transmission electron micrographs of H3N2v virus (A/Minnesota/11/2010) from infected MDCK cells following treatment with MAb H3v-47, control mAb CR8020, zanamivir, or vehicle. The red squares highlight the clustering of virus at the cell surface in H3v-47- and zanamivir-treated cells, as opposed to free virus in the case of CR8020- and vehicle-treated cells. (b) Close up of H3v-47 viral particles attached to each other and the cell surface with H3v-47 mAb detected between them by a gold-labeled secondary antibody (red arrows). Figure reproduced under the Creative Commons Attribution 4.0 International License 4.0 from S. Bangaru, et al. Nat. Comm., (2018) 9, 2669.


In addition to directly interfering with viral function, antibodies may trigger host immune responses in vivo. One such response is antibody-dependent cellular cytotoxicity (ADCC), in which antibody binding to and cross-linking of Fc receptors on natural killer (NK) cells activates the cells to release toxic perforins, granzymes, and antiviral cytokines. Many mAbs directed against the HA stem domain have the ability to activate ADCC, leading the researchers to investigate whether or not H3v-47 also exhibits this capacity. They found that, in fact, H3v-47 could cross-link Fc receptors in vitro, and that it promoted activation of NK cells in culture using the HA from A/Sydney/5/1997 as the antigen.


Although this was the first time an mAb was found to bind to a viral HA in the location targeted by H3v-47, other mAbs are known to bind in the vestigial esterase domain. A comparison of crystal structures of the Fab portions of these mAbs bound to HA reveals, however, that they all bind at different sites, and in different orientations (Figures 4 and 5). This finding is consistent with the fact that some of these mAbs neutralize HA by mechanisms distinct from that of H3v-47. For example, HC45 interferes with cell surface binding, as its orientation places it within close proximity of HA's sialic acid-binding site, and H5M9/H5 blocks endosomal fusion. Interestingly, the CR8071/fluB mAb, which is specific to the influenza B virus, binds at the vestigial esterase domain and prevents viral egress. Thus, there is a precedent for the mechanism of action of H3v-47.


FIGURE 4. Space-filling representation of influenza virus HA protein, with each subunit shown in a different shade of gray. The receptor-binding subdomain is highlighted in tan. It is the site of attachment of many mAbs including C05, S139, HC63, and F045-092. Also shown are ribbon representations of three mAbs (H3-47, HC45, and F005-126) that bind to the vestigial esterase domain at different sites. Figure reproduced under the Creative Commons Attribution 4.0 International License 4.0 from A. M. Andrews, et al. Nat. Comm., (2018) 9, 2592.



FIGURE 5. Space-filling representation of influenza virus HA protein, with each subunit shown in a different shade of gray. Also shown are ribbon representations of three mAbs (H3-47, H5M9/H5, and CR8071/fluB) that bind to the vestigial esterase domain (highlighted in red) at different sites. Figure reproduced under the Creative Commons Attribution 4.0 International License 4.0 from A. M. Andrews, et al. Nat. Comm., (2018) 9, 2592.



In summary, H3v-47 is a human mAb that binds to a distinct epitope on the HA from H3N2 viral strains. It has very broad specificity, neutralizing HA proteins from H3N2 viruses spanning a 25-year period as well as proteins from swine-derived viruses. It protects against lethal H3N2v infection, at least in mice, by both preventing viral egress and by activating antibody-mediated cellular cytotoxicity. These results suggest that the epitope recognized by H3v-47 is a potentially very valuable target for development in the quest for a broadly neutralizing long-term influenza vaccine. This is especially important if the vaccine is intended to protect against dangerous H3N2v strains that carry a high potential for causing pandemic disease.



View Nature Communications article: A multifunctional human monoclonal neutralizing antibody that targets a unique conserved epitope on influenza HA









The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center