⚗️ biochemistry

Antibodies targeting a conserved cryptic epitope at the influenza hemagglutinin head-stem interface via distinct binding modes

This study characterizes two distinct, non-neutralizing antibodies that target a conserved cryptic epitope at the influenza hemagglutinin head-stem interface, revealing their suboptimal protective efficacy and providing insights for future vaccine design.

Original authors: Stroh, G. L., Lv, H., Chen, C., Sun, Y., Pholcharee, T., Zhu, C., Teo, S. T., Mitchell, D. A., Mok, C. K. P., Wu, N. C.

Published 2026-09-28

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Stroh, G. L., Lv, H., Chen, C., Sun, Y., Pholcharee, T., Zhu, C., Teo, S. T., Mitchell, D. A., Mok, C. K. P., Wu, N. C.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

The influenza virus is a master of disguise, constantly reshaping the proteins on its surface to evade the human immune system. The most prominent of these proteins is hemagglutinin, a spike-like structure that acts as the virus's key, unlocking the door to human cells. For decades, vaccines have trained the body to recognize the top, or "head," of this spike. However, because this head region changes rapidly from year to year, the immune system must be retrained annually, and the protection often fails when the virus shifts slightly. Scientists have long sought a "universal" vaccine that could train the immune system to recognize parts of the virus that do not change, specifically the "stem" or base of the spike, which remains relatively stable across many different flu strains. The challenge is that the immune system naturally prefers to attack the flashy, changing head, often ignoring the more stable, hidden stem.

In a recent study, researchers investigated two specific antibodies found in people who had received an experimental vaccine for a dangerous bird flu strain known as H5N1. These antibodies were interesting because they managed to bind to a hidden, conserved spot where the head and stem of the virus meet, a region that is usually buried and difficult for the immune system to see. The scientists wanted to understand exactly how these antibodies latched onto the virus and whether they could offer strong protection against infection. By using advanced imaging to see the molecular details of the interaction, they discovered that while these antibodies could find this hidden spot, they did so in a way that left them unable to stop the virus from infecting cells or causing disease in living animals.

The researchers began by isolating two antibodies, named 01.z.01 and 56.e.01, from human volunteers who had participated in a clinical trial for an H5N1 vaccine. These antibodies were special because they could recognize a wide variety of flu viruses, including those from different groups that usually do not cross-react. However, when the team tested them in the lab, they found that neither antibody could stop the virus from clumping red blood cells or neutralizing the virus in a dish, a property known as neutralization. This suggested that while the antibodies could grab onto the virus, they were not blocking its ability to enter cells. To understand why, the team used a powerful imaging technique called cryo-electron microscopy to take three-dimensional pictures of the antibodies bound to the flu virus spike.

The images revealed that both antibodies were targeting a very specific, hidden area at the junction where the virus's head meets its stem. In the virus's natural, resting state, this area is completely covered up by the virus's own structure, making it invisible to the immune system. The researchers found that the antibodies could only access this spot after the virus had undergone a shape-shifting event, which happens when the virus enters an acidic environment inside a cell. This shape change is necessary for the virus to fuse with the cell and release its genetic material. The study showed that the antibodies could bind to this exposed junction, but they approached it from two very different angles. One antibody, 01.z.01, approached from the side, wrapping around the spike in a unique way that had never been seen before. The other, 56.e.01, approached from the front, similar to a previously known antibody that targets the same area.

Despite their ability to bind to this conserved spot, the antibodies proved to be weak defenders. When the researchers tested them in mice, they found that the antibodies offered very little protection against lethal doses of seasonal flu viruses. Even when the antibodies were engineered to be more effective at recruiting the immune system's cleanup crew, they only saved a small fraction of the mice. The situation was even more dire when the team tested them against a highly dangerous H5N1 strain that had recently jumped from birds to humans. Even at very high doses, neither antibody could prevent the mice from getting sick or dying. The study concluded that while these antibodies successfully identified a hidden, conserved target on the virus, their binding mode and the way they interacted with the virus did not translate into strong protection.

The findings highlight a complex reality in the search for a universal flu vaccine. Just because an antibody can find a stable, unchanging part of the virus does not mean it will be effective at stopping an infection. The study suggests that the specific angle at which an antibody binds to the virus, and how that binding affects the virus's ability to infect cells, are critical factors that determine success. The researchers noted that the antibodies they studied were likely the result of the immune system reacting to a hidden part of the virus that became exposed during the vaccine preparation process, rather than a natural, protective response. This implies that future vaccine designs might need to be carefully engineered to hide these non-protective, hidden spots, forcing the immune system to focus on other parts of the virus that can actually stop the infection. The work provides a detailed map of how these antibodies interact with the virus, offering valuable clues for designing better vaccines that can elicit truly protective responses against a wide range of flu strains.

Technical Summary: Antibodies Targeting a Conserved Cryptic Epitope at the Influenza Hemagglutinin Head-Stem Interface

Problem Statement
Influenza A virus remains a persistent pandemic threat, exemplified by the recent global spread of highly pathogenic H5N1 strains. Current seasonal vaccines primarily elicit antibodies against the highly variable head domain of the hemagglutinin (HA) glycoprotein, resulting in a lack of strong heterosubtypic (cross-reactive) immunity. While broadly neutralizing antibodies targeting conserved regions like the HA stem, receptor binding site (RBS), and trimer interface have been identified, the field seeks to understand and exploit other conserved epitopes to design universal vaccines. Specifically, there is a need to characterize antibodies isolated from H5N1 vaccinees that exhibit broad reactivity but may target under-characterized, cryptic epitopes.

Methodology
The study employed a multidisciplinary approach combining structural biology, biophysics, and in vivo animal models to characterize two broadly reactive human monoclonal antibodies (mAbs), 01.z.01 and 56.e.01, previously isolated from H5N1 vaccinees.

Epitope Mapping and Competition: The authors utilized enzyme-linked immunosorbent assays (ELISA) to assess binding breadth across diverse HA subtypes (Group 1 and 2). Biolayer interferometry (BLI) competition assays were conducted against known antibodies (5J8, FluA-20, and S8V1-157) to determine if 01.z.01 and 56.e.01 target known epitopes or novel ones.
Structural Biology (Cryo-EM): To elucidate the molecular recognition mechanism, the authors determined the cryo-electron microscopy (cryo-EM) structures of the H3N2 HA head domain in complex with 01.z.01 Fab (in a quadripartite complex with FluA-20 and ADI-85647 Fabs) and 56.e.01 Fab (in a complex with FluA-20 and ADI-85647 Fabs). For 56.e.01, an AlphaFold3-predicted model was fitted into the low-resolution map.
Biophysical Characterization: Binding affinities ( $K_d$ ) were measured using BLI against various HA constructs (head-only, uncleaved HA0, and acid-treated HA) to assess the accessibility of the epitope under different conformational states. Site-directed mutagenesis was performed on key HA residues (e.g., P293, T40) to determine their contribution to binding affinity and breadth.
In Vivo Efficacy: Prophylactic protection studies were conducted in BALB/c mice. Mice were treated with 01.z.01 or 56.e.01 (formatted as mouse IgG1 or IgG2c) and challenged with lethal doses of H1N1 (A/California/04/2009), mouse-adapted H3N2 (X-79), and highly pathogenic H5N1 (A/Texas/37/2024). Survival rates, weight loss, and lung viral titers were monitored.

Key Contributions and Results

Identification of a Cryptic Head-Stem Interface Epitope: Both 01.z.01 and 56.e.01 bind to a highly conserved, cryptic epitope located at the interface between the HA head and stem domains. This epitope is largely occluded in the prefusion trimer conformation and becomes accessible only upon acid-induced conformational changes or dissociation of the stem domain.
Distinct Binding Modes:
- 01.z.01: Adopts a unique "side-on" approach angle, rotated approximately 90° relative to the previously characterized antibody S8V1-157. Its HCDR3 forms an extended parallel $\beta$ -sheet with HA1 residues via backbone-backbone interactions, a feature that likely contributes to its broad recognition despite sequence variation.
- 56.e.01: Engages the same epitope with a "head-on" approach angle similar to S8V1-157, sharing germline gene usage (IGKV2-2801, IGHJ402) and an LCDR3 motif. However, its HCDR3 contains a disulfide bond and adopts a different orientation compared to S8V1-157.
Determinants of Antigenicity: Residue P293 on HA1 was identified as a critical determinant of binding breadth. Variants at this position (e.g., Ser in H11, Thr in H6/H13/H16) significantly reduced or abolished binding for both antibodies. Conversely, residue T40 showed high tolerance to mutation, indicating it is not a primary constraint for binding.
Non-Neutralizing Nature: Despite broad binding, neither antibody demonstrated neutralizing activity in hemagglutination inhibition (HAI) or microneutralization (MN) assays. This is consistent with the epitope being buried in the prefusion state, preventing the antibodies from blocking receptor binding or membrane fusion directly.
Limited In Vivo Protection:
- Fc-Dependent Effects: Protection was largely dependent on the Fc region. 01.z.01 formatted as mouse IgG2c (activating Fc $\gamma$ Rs) provided partial protection (20–100% survival) against H1N1 and H3N2, whereas the IgG1 format offered minimal protection. 56.e.01 showed weaker Fc-dependent protection overall.
- H5N1 Failure: Even at high doses (25 mg/kg), neither antibody provided protection against lethal H5N1 challenge, unlike the central stem antibody MEDI8852 which conferred 100% survival.
- Germline Reversion: Reverting both antibodies to their germline sequences resulted in a contraction of binding breadth, particularly losing binding to H1 and H2, highlighting the role of somatic hypermutation in achieving broad reactivity.

Significance
The paper claims that these findings expand the known repertoire of interactions between cross-reactive antibodies and the influenza HA head-stem interface. By revealing that distinct binding modes (side-on vs. head-on) can target the same cryptic epitope, the study underscores the structural plasticity of the immune response to this region.

Crucially, the authors conclude that antibodies targeting this specific cryptic epitope represent suboptimal vaccine responses. Their non-neutralizing nature and marginal protective efficacy in vivo, particularly against highly pathogenic H5N1, suggest that immunogens designed to elicit such antibodies may not be sufficient for universal protection. The study proposes that future immunogen design should consider stabilizing the mature, closed trimeric form of HA to shield this cryptic interface, thereby shifting the immune focus toward more protective epitopes like the RBS or the central stem. This work provides a functional and structural basis for refining next-generation influenza vaccine strategies to avoid the induction of non-neutralizing, cryptic-epitope-specific antibodies.

Technical Summary: Antibodies Targeting a Conserved Cryptic Epitope at the Influenza Hemagglutinin Head-Stem Interface

More like this