RESEARCH TOOLS FOR COMBATING FUTURE VIRAL OUTBREAKS

Moderator, Fierce Biotech 

Kyle Doolan, PhD, Director of Research, Integral Molecular

Kyle introduces safe, fast, and quantifiable approaches for studying emerging viral threats. The innovative approaches covered include pseudoviral particles and alanine-scanning epitope mapping. At 12:02, he introduces a new product for influenza hemagglutination inhibition assays, TiterSafe, that is a safer, off-the-shelf substitute for live virus.

Chan-Nan Chen, PhD, CEO & CSO, Single Cell Technology, Inc.

Dr. Chen presents rapid isolation of highly cross-reactive influenza A and SARS-CoV-2 antibodies using picowell B-cell discovery technology. The hits from these approaches were validated in-house using Integral Molecular’s RVPs.

James E. Crowe, Jr., MD, Director, Vanderbilt Vaccine Center, Vanderbilt University Medical Center

Dr. Crowe discusses the structural mechanisms by which neutralizing antibodies recognize influenza hemagglutinin. He proposes that breadth, potency, epitope mapping, and viral escape studies should be performed. To exemplify breadth and potency experiments, he presents collaborative work with Integral Molecular using their influenza pseudovirus neutralization assay and TiterSafe particles.

The moderator, Kyle, and Chan-Nan are joined by Robert Carnahan, Associate Director of Vanderbilt Vaccine Center, to answer questions from the attendees including:

• Can binding data be used as a substitute for neutralization data in assessing antibodies or sera samples?
• How can B-cells isolated from patient samples early in a pandemic be stored and screened for antibody activity later?
• Is it possible to achieve a pan-neutralizing influenza antibody therapeutic? What characteristics would such a therapeutic be likely to have (i.e., neutralizing vs non-neutralizing antibodies and putative epitopes)?

• Case study: Learn how RVPs facilitated clinical advancement of a SARS-CoV-2 antibody therapeutic
• Antigenicity, stability, and reproducibility of Zika reporter virus particles for long-term applications. Whitbeck et al. 2020, PLoS Neglected Tropical Diseases 14, e0008730
• Dengue Reporter Virus Particles for measuring neutralizing antibodies against each of the four dengue serotypes. Mattia et al. 2011, PLoS One 6, e27252

• An antibody targeting the N-terminal domain of SARS-CoV-2 disrupts the spike trimer. Suryadevara et al. 2022, The Journal of Clinical Investigation 132, e159062
• Convergence of a common solution for broad ebolavirus neutralization by glycan cap-directed human antibodies. Murin et al. 2021, Cell Reports 35, 108984
• Influenza virus-specific human antibody repertoire studies, Crowe 2019, The Journal of Immunlogy 202, 368

Rebekah Kuschmider:

Hello everyone. Thank you for attending today's webinar, Innovative Research Tools to Combat Biological Threats, presented by Integral Molecular. I'm Rebekah Kuschmider, and I'll be moderating this webinar.

I'd like to start off by introducing today's speakers, Kyle Doolan, director of research at Integral Molecular, Chun-Nan Chen, CEO and CESO of Single Cell Technology Incorporated, and James E. Crowe, Jr., director at Vanderbilt University Medical Center. You can read their full bios to the left side of your window by selecting the speakers tab.

Just a few notes before we begin. To access additional resources for today's presentation, please click the handouts tab button on the left side of your screen. You can access closed captions from the bottom right corner of the video player. This webinar is being recorded. It will be available to watch on demand within 24 hours.

There will be two polls during this presentation. We encourage all attendees to participate in the polls. Time permitting, we will conclude with a Q&A session. So during the presentation, you can submit any questions you have using the questions and answers tab on the left side of your screen. Okay, now let's begin. Kyle, please go ahead.

Kyle Doolan:

Hello and welcome to today's webinar on Innovative Research Tools to Combat Biological Threats. My name is Kyle Doolan, and I'm the director of research at Integral Molecular. We're excited to talk with you today and are joined by Chun-Nan Chen of Single Cell Technology and James Crowe of Vanderbilt University to talk with you about approaches to studying, understanding, and applying principles of virology to respond to viral threats.

Human viral pandemics have long been a part of human history and have generally arisen from the transmission of pathogens from infected animal reservoirs to humans. In terms of annual deaths, threats from infectious disease have been generally decreasing as the result of modern medicine, public health, and sanitation.

However, modernization has changed the conditions for new zoonotic disease emergence or the transmission of viral threats from animal reservoirs to humans. In recent years, we've seen a number of smaller viral outbreaks in the initial coronaviruses, influenza, and Ebola outbreaks, while also witnessing the emergence of SARS-CoV-2, COVID-19, into becoming a worldwide pandemic.

Expanding our scientific understanding of viruses is key element to enabling the modern medicines and therapeutics and public strategies needed to both avoid and overcome viral threats.

Studying live viruses comes with some challenges, primarily related to ensuring safety and obtaining meaningful results. As virologists, we know that we have the capability to do both. But as the photo on the left demonstrates, it comes with some trade-offs in terms of locations and throughput of work that can be performed. Working in containment facilities and examining the discrete and, quite honestly, subjective formation of plaques in a plaque-reduction neutralization test or PRNT, where here you're looking for the clear plaques within these blue fields.

At Integral Molecular, we've been committed to understanding viral glycoprotein structure and function in a way that retains the biological relevance. But that also enables all scientists to study these viruses, especially in conditions and locations that are impacted by viral outbreaks, that are engineered for safety of the lab workers beyond just BSL classification, but in a way that meaningfully minimizes actual risk, and that enhances our speed to respond to new viruses of interest and that is easily quantified to support more robust studies.

At Integral Molecular, we've developed technologies to explore the functional and structural aspects of viral glycoproteins that serve as the basis for cellular infection. We've developed non-replicative reporter virus particles that present viral glycoproteins and deliver a minimal reporter genome such as GFP or luciferase for the quantitative evaluation of neutralizing therapeutics.

We've developed over a hundred off-the-shelf solutions that we can ship to scientists across the world. We've also developed methods for experimentally mapping antibody epitopes of natively presented viral glycoproteins, including conformational epitopes. These epitopes provide insights into antibody mechanism of action and collectively provide insights into immunogenicity, next-gen vaccine design, and can help predict modes of viral escape.

We've been using these technologies for more than 20 years to gain fundamental insights into viral biology and to respond to viral threats as they've impacted human health. We were founded on viral technology used to elucidate the HIV entry receptors at the height of the AIDS pandemic, and have been responding in turn to viral outbreaks from dengue to SARS-CoV-2, both independently and in collaborations, culminating in peer-reviewed publications that advance our understandings of viral biology and therapy.

We've really prioritized a depth of knowledge across multiple viral families, both in our reporter virus particles and epitope mapping.

Taking a closer look at our pseudotype reporter virus particles or RVPs. They're designed to present viral envelope proteins in their conformationally relevant form on the outside while containing a non-replicative minimum genome that incorporates a reporter gene to provide quantitative readout upon particle entry.

Because of these properties, they are safe to use under more standard BSL-2 conditions. They're flexible to high throughput and automatable applications for high-content imagers, flow cytometers, and multi-well plate formats, and when accompanied by our protocols and quality control efforts, lead to technically simple operations that can be easily trained and transferred to laboratory personnel.

They've been used in all stages of therapeutic development, both by large pharma vaccine developers testing serum and clinical trials and by biotech companies screening for neutralizing antibodies and discovery.

RVPs are normally applied in a neutralization assay to evaluate the efficacy of the vaccine and therapeutics. When RVPs are incubated with sera or antibody, the recognition of the viral glycoprotein prevents interaction with the host receptors, leaving target cells unable to express the reporter.

Alternatively, when antibodies are not present or not neutralizing, the glycoprotein or proteins mediate internalization or fusion, releasing the reporter gene, and ultimately leading to downstream expression and quantitative detection. Importantly, delivery of the reporter means that the cells in which the particles have entered are unable to support viral replication, and no new particles are generated.

Here we show some typical results. In this case, for a SARS-CoV-2 GFP RVP when neutralized with benchmark antibody VHH-72. In the left figure, we show specific neutralization only by the correct pairing of antibody and particle. On the Y axis, we're showing the percentage of cells that are infected and along the X axis, as we increase the amount of neutralizing antibody, in orange, cells are protected from viral entry.

At the midpoint of the curve, we can determine the neutralizing titer or NT50. Notice that when the antibody and RVP are not matched, infection is not blocked and remains high. We find that the NT50s found from this assay are highly comparable to published literature values using live virus.

On the right, we're showing the same neutralization of SARS-CoV-2 and VHH-72 as a pair, but this time this assay was run six different times over the course of more than two years. Each curve is actually from a different independent production of our RVP product. Then the assay was performed by different operators on different days on multiple machines, and, considering that we moved laboratory locations in the middle of this time course, even different laboratory buildings.

We do see some small changes in the overall curve, but have generated very similar estimates of NT50. With our RVPs, when you open a new tube from Integral Molecular, it's going to perform the same way as the last one, whether the last one was the plate you set up that morning or one you did last year.

RVP neutralization estimates of NT50 or IC50 are highly correlated to the IC50 estimates from live virus, such as PRNT, as shown by the high correlation on the figure on the left. It runs nearly through a one-to-one correspondent and was published by a customer of ours for our SARS-CoV-2 RVPs.

This sort of experiment validates the use of RVPs as a substitute for live virus. On the right, we're showing some results from our recent meta-analysis evaluating pseudoviral technologies overall, of which the reporter viruses are one. They show that they overall correlate well with live virus in terms of IC50 across many different studies and virus types.

Interestingly, for some viral classes, the correlation statistics appears to depend more on the source or assay for the pseudotype particles than on differences observed between viral families. As scientists at Integral Molecular, we're devoted to creating the best possible materials and the best methods so that there is a strong correlation with live virus. We've seen positive correlations with other viral families, including Nipah and dengue.

The high correspondence of RVPs allows them to be used as viral substitute all along the antibody and vaccine development pathway, from discovery to approval. Here we're showing some representative publications by our customers. First, we're showing antibody discovery using SARS-CoV-2 RVPs, where multiple hits are characterized for functional neutralization.

In a separate study, a lead antibody cocktail used the breadth of our catalog to profile neutralizing antibodies, or the antibody cocktail, across multiple variants. As an example of phase one usage, we're highlighting an evaluation of RVPs used to measure Zika neutralizing titers as a function of vaccine dose. Here up high compared to control down below.

For a phase two, three example, multiple dengue serotype RVPs were used to assess the breadth of protection from a vaccine that had received approval in a number of locations around the world, including areas with endemic infection.

To serve an even broader community of vaccine and antibody developers, we've been expanding our support for influenza research and development. We've recently launched full support for all seasonal vaccine strains including H1N1, H3N2, and both Victoria and Yamagata strain influenza B.

In addition, we've developed reporter virus particles for pandemic potential candidate vaccine strains as identified by the World Health Organization, including H5N1 and H7N9. These RVPs display functionally active HA and neuraminidase on the surface and readily infect target cell lines.

Shown on the right is the RVP displaying HA and NA from H1N1 Wisconsin/588/2019, currently used in the northern hemisphere 2023, 2024 vaccines. Here it's specifically neutralized by the benchmark antibody MEDI8852, a highly cross-reactive anti-HA antibody.

Because these are non-replicative pseudotypes that are produced from plasma-defined sequences, there is a reduced risk of genetic drift, and we're able to quickly develop or adapt to newly arising strains or even custom requests.

We're also excited to announce the development of a new type of particle, our TiterSafe influenza agglutination particles. These particles again have active HA and NA on the surface, but have been formulated to function as a substitute for live virus in the classical hemagglutination and hemagglutination inhibition HAI assays.

On the figure on the right, we show specific HAI by match serum to the B/Phuket/3073/2013 Yamagata lineage influenza B. Here we observe the binding of anti-HA-head antibodies to prevent the cross-linking of HA with sialic acid present on the red blood cells, ultimately appearing as a dot.

While in the absence of neutralizing antibodies, we see cross-linking appearing as a homogeneously filled circle. We do observe some slight protection at high serum values from other influenza B strains in the Austria serum of the Victoria lineage.

These are available as off-the-shelf reagents as a substitute for live virus and existing partially automatable workflows. These particles lack a genome and pose no recombination risk to laboratory workers, which particularly for this Yamagata lineage shown here, which has not been sequenced in humans for several years, and for which the World Health Organization recently recommended against further work with live virus to prevent accidental reintroduction back into the human population. The attributes of TiterSafe particles make them an ideal substitute.

Finally, we recommend the TiterSafe particles only for serum screening, where anti-HA antibodies disrupting the sialic acid recognition are a reasonable surrogate for neutralizing titer, as used with monoclonal antibodies has been shown to have a number of both false positive and false negative results when antibodies target particular epitopes.

To evaluate the epitopes of monoclonal antibodies, we've developed an epitope-mapping platform capable of detecting conformational epitopes, including those requiring tertiary and quaternary structure. The epitopes are useful where they provide insight into neutralizing mechanisms and strengthen therapeutic IP.

Here we perform systematic mutation of all residues to alanine and, in the case of alanine, to serine, express each of these mutants in human cells to measure binding and evaluate specific loss of function binding while controlling for protein expression and folding to identify energetically critical interactions. This provides single amino acid resolution of conformational epitopes in a high-throughput assay.

Using Using this approach, we've determined the key residues for hundreds of antiviral antibodies, including the composite images of multiple different antibodies on influenza B neuraminidase at left and full-length SARS-CoV-2 spike at right. When combined with neutralizing titer, these maps characterize the human antiviral immune response, provide insights into immunogenicity and neutralization needed for next-generation vaccine design, support hypotheses for neutralizing mechanisms of action, and help predict the potential for viral escape.

At Integral Molecular, we're scientists who are excited about collaborative science. We're excited about the technologies we've developed, and to support your work with our off-the-shelf RVP catalog, the ability to make custom RVPs on request. We can also perform neutralization and antibody epitope-mapping services to provide therapeutic programs advancement and/or provide scientific insight.

We're pleased to be joined today by people who've done just that. Happy to turn it over now to Chun-Nan Chen from Single Cell Technologies and James Crowe from Vanderbilt University to discuss their work and in part how Integral Molecular technologies enabled that work.

Chun-Nan Chen:

Hi, I'm Chun-Nan Chen, CEO of Single Cell Technology. I will start with a few questions for everyone to think about.

Though I cannot say pandemic disorder over, I think it's still appropriate to conduct a postmortem. It is easy to say after the fact that we could have shut down borders earlier to prevent widespread transmission or produce rapid antigen tests earlier. In the middle of the viral outbreak, it is hard to see the forest for the tree. These responses are reactive in nature and would not be enough.

At Single Cell Technology, our unique antibody discovery approach can be used to produce rapid medical countermeasures. I will list two items that we can do for your consideration. One is to be proactive, the other is to be rapidly reactive.

We, Single Cell Technology, specialized in antibody discovery, and you will soon find out we go about finding antibodies in a very different manner from others. We started 15 years ago, when single-cell analysis is not mainstream. We pioneered a single B-cell approach that no one else does, and it is quite outrageous.

It is a tool we've developed and matured called AbTheneum. We have helped many clients find their next-generation therapeutic antibodies by working with various, hosts including major transgenic mouse lines, wild type mouse, rat, and human.

Here's a picture on the left of our device that is bounded to a regular-size microscope slide. It has little over 90,000 picowells since the volume is in the picoliter range. The volume for each picowell hits a sweet spot where it is small enough to allow a single cell to reach nanomolar concentration in less than 20 minutes, but it is large enough through the debris when we lyse the cells in the picowell later.

The pictures on the right shows single cells in some of the picowells. We deposit cells at appropriate concentration directly onto our device and let them sync to the bottom. Subsequently, we can capture antibody directly from these cells to form an addressable antibody array to be stored or interrogated by appropriate screening agent.

Here is our process diagram. We start with cell isolation. In the next two pandemic case studies, we have two different methods for cell isolations. We deposit antibody-secreting cells onto our device. We capture secreting antibodies from each cells onto a glass slide, forming addressable antibody arrays.

Antibodies are then screened by fluorescein-labeled molecules over antibody array and scanning with a slide scanner. All IgG sequences are captured and sequenced by NGS and mapped back to their origin. We then filter the antigen-specific antibody sequences and layer all other screening data for each single cell.

I have two case studies addressing pandemics that I will go over now. First, our next likely pandemic has been predicted to be influenza. Influenza is definitely a virus to be concerned about due to the vast diversity of different strains and subtypes. For this reason, we designed this program to discover broadly reactive anti-influenza antibodies.

We did this work in collaboration with Miltenyi Biotec. We started with three donors that received their seasonal flu shots. We isolated antigen-specific antibody-secreting cells from memory B cells of the donors. We used AbTheneum to screen antibodies against four different diverse hemagglutinin or HA proteins. We also screened antibodies for their ability to compete with a broadly neutralizing antibody, MEDI8852.

Here are the results from one of the trials for the influenza campaign. A little less than 2600 antibodies are all profiled for their antibody binding properties against the four HA proteins and the ability to compete with MEDI8852. We see that 15 of them react to all four HA proteins and eight of those compete with MEDI8852.

This campaign took 60 days to get the validated data, but that timeline could be compressed to less than 40 days with our recent breakthrough. We selected 12 antibodies from the output to validate their activity by ELISA. The binding activity screened by AbTheneum were confirmed by ELISA. We then used four additional HA protein to see if any antibodies were possibly pan-influenza-binding.

I want to highlight this one antibody that shows reactivity across the four HA proteins used in screening and four additional HA proteins. We were very pleased to demonstrate our screening method applied to this challenge. We believe discovering broadly reactive influenza antibodies is a proactive way to combat the next influenza pandemic.

The antibodies can be banked and tested against new strains and used for diagnostic or possibly therapeutic applications if they show neutralization activity. Next, I want to show an example of how to better quickly and safety validate antibodies that we discovered.

During the pandemic, we did a campaign to isolate anti-receptive binding domain, RBD antibodies, from immunized mice and recover close to 2,000 hits. These are full-length, natively paired VH and VL sequences from each cell.

From a distance, this is what 1974 hits look like. The antibodies were all screened for binding to RBD, but also their ability to block ACE-2 using AbTheneum screening.

This is what our collections of antibody looks like after extensive downselection. We screen ones that have strong ACE-2 blocking activity and the highest affinity ranks, and ended up with 74 promising leads.

We tested reconstructed anti-RBD mAbs in Integral Molecular's luciferase RVP assays. These are just four plots from four different strains. We ran the assay using multiple strains, many antibodies. When a new variant RVP assay was available, we ordered it and ran some of our antibodies.

The RVP assay is really easy to run. Plates are set up, incubated for three days, and read on a luminometer. Some of our antibodies discovered from our RBD-immunized mice showed neutralizations against all variants we tried.

We also validated the RVP assay compared to the live virus with the help from Vitalant, who runs a BSL lab in San Francisco. We saw very good response to our antibodies in their live virus trials.

To do the live virus test, we needed to have trained staff available at the BSL lab to run the test, as we are not BSL-3. The live virus assay they run was plaque-reduction neutralization test, PRNT, which can be difficult to run at scale without automated equipment. We were limited to whatever variant Vitalant has available at the time.

These two plots aren't exactly the same variants, but to show that they have good correlations, we ran two to three different variants with the live virus neutralization assay, and Vitalant did a great job and we are very pleased to see our antibodies performing well.

I think there are some hurdles to run live virus neutralization, and the RVPs make it very easy to triage across more variants, easy-to-run replicates, which can be validated after with the live virus neutralization assays.

To summarize, I've reviewed very briefly two different sources that we have used for a pandemic response. Immunized animals like the RBD-immunized mice and vaccinated donor for influenza, the same workflow from vaccinated donors could be applied to convalescent patients.

Like the strategic national stockpiles, we could be researching these possible pandemic viruses and banking tools that could help us in an outbreak. For example, our broadly reactive anti-influenza antibodies could be useful during the influenza outbreak if they're still reactive to the circulating strain.

Lastly, our discovery method preemptively sequences all IgGs. Given the novel virus, we could sequence first and screen the captured antibodies later when reagents become available.

These are just two ways I want to share with you on the approach to a pandemic response. I hope you reach out with any question you have about our technology. Thank you.

James Crowe, Jr.:

Hi, I'm James Crowe at Vanderbilt University Medical Center. I'm going to talk to you about human monoclonal antibodies to influenza HA and how to assess them, characterize them, group them, and understand their functionality.

So what you're seeing on the screen is a space-filling model of influenza surface protein called hemagglutinin or HA. It's typically seen because of its shape as having a head domain at the top and a stem at the bottom.

Antibodies have been isolated and characterized to many regions on the head and stem domain. So, for instance, at the top there's a bowl-shaped area that's the receptor-binding domain into which sialic acid on human cells would bind. So the virus binds to the cells at the receptor-binding site or receptor-binding domain.

At the tippy top, there's a region called the apex, which is coming straight in. Around the receptor-binding domain, there's a rim of very hypervariable residues, but a lot of antibodies will recognize that rim. If you come to the side of the head, you have vestigial esterase domain. So this looks like a legacy enzymatic-looking site that's no longer functional, but it's just a structural feature now. But you see a lot of antibodies to the side of the head.

Farther down, certainly there are many antibodies to the stem. This tends to be a more conserved area and, therefore, a lot of broad or almost universal ... Broadly protective or universal antibodies bind to this region. Even farther down on the stem, near the viral membrane derived from the cell membrane, is the anchor region. More recently, antibodies have been isolated from ... That recognize the anchor region.

So here's an example of receptor-binding domain antibodies. This is a 1918 influenza specific antibody to HA, isolated from a nearly a hundred-year-old person whose the first exposure was to 1918 influenza, a virus that killed about 50 million people. The antibody is shown in three copies of Fab, and yellow and red with heavy chain, light chain, different colors.

And so, these antibodies come in sort of a 45-degree angle, engage the receptor-binding domain generally. If you look at the footprint, you can identify critical residues from the co-crystal structure that we did with Ian Wilson's group at Scripps.

 

What's interesting, that then predicted that the 1918 binding curve activity shown in black squares also would be similar to the California 2009 pandemic, H1N1, that we all lived through. In fact, the antibody is just as good against the 2009 as it was 1918. That's because the 2009 is a complex reassortment that has elements of 1918 that came back probably from big populations.

So when you engage the receptor-binding domain, there's two ways you can simulate or block the receptor. One is using a carboxyl group that either the sialic acid receptor uses it or an antibody hypervariable loop, or complementarity determining region or CDR, will bring in an aspartic acid residue to do that.

The other way is to bring a hydrophobic residue, which is mainly represented by aromatics. That's also on the tip of an antibody. So this is an example of the aspartate mode, two different antibodies, one of ours on the left called 5J8. On the right, an antibody from a Duke and Harvard consortium shown in blue. The sialic acid, as if bound as the receptor, is shown in yellow.

What you can see if you zoom in on that, the red antibody on the left comes in a certain angle and the blue antibody comes in from an opposite angle over different REMS into the receptor-binding domain. But ultimately they both present in aspartate at the bottom, and that aspartate overlays in the same way that a sialic acid does. So that's why you get receptor mimicry using that aspartic acid.

Then the other mode I mentioned is an aromatic residue. To illustrate that, I'm going to show you some antibodies that neutralize 1957 H2N2 pandemic. So that was the second pandemic of the 20th century.

And so, we isolate three antibodies from middle-aged people who had experienced 1957 influenza during childhood. On the left, you see the aromatic residue that is in position at the bottom of the receptor-binding domain. Each of these antibodies presents that. It comes in from a different angle, but ultimately it puts the aromatic in the same place.

Again, if you look on the right, sialic acid bound in that place, you can understand why there's mimicry and also blocking by the antibodies on the left, if they were engaged of sialic acid.

Now I also mentioned the straight tippy top apex site. This is an antibody called H7.5, H7 being avian influenza. That is a threat to cause a new pandemic because very few people are immune to H7 virus.

This is just negative stain electron endoscopy of the Fabs at the top bound to the HA trimer. What was interesting, we did this work with Andrew Ward's Group, Hannah and Jesper there. The unbound HA looks very typical position, but once you put an Fab on, the HA scissors open and either ... It implies that the HA has dynamic features and that the antibody is possibly causing the opening, but more likely just catching the dynamically open HA.

That leads to the conceptual idea of what are called trimer interface antibodies. So this is an antibody called FluA-20 in which the Fab is shown in green and blue here, binding to a yellow head domain protomer. If you look on the right, the top-down view, the Fab is binding inside the hidden interface where the three protomers, one, two, and three, come together, the HA head. The epitope here is in red. So it should be hidden, but in fact the HAs must open up.

This antibody is very broad because that's a highly conserved area. And so, these are weight loss curves in mice for H1, H3N2, H5N1, and H7N9. This antibody works against all of them in studies with Michael Schotsaert at Mount Sinai and the flu group there. So very broad antibodies because the site is conserved in the trimer interface.

Now we're moving out to the side of the head. This is an antibody called H3v47. So H3 variant virus is really a pig virus that sometimes crosses over to humans even in the United States.

And so, this antibody from a vaccinee binds to the vestigial esterase domain or the side of the HA head. This is a co-crystal structure from Heng Zhang in Wilson's group. Then, finally, this is an example of the stem antibody called 8D4. Again, it's a 1918 antibody, but instead of anti-receptive binding domain like I showed you before, this one binds to the stem.

Like many canonical stem antibodies, the heavy chain shown in yellow is the only engaging protein here. The light chain shown in orange of the Fab is hanging off. That's a very typical stem antibody.

Stem also is highly conserved. So if people talk about broadly protective antibodies or almost universal antibodies, they often bind to the stem. Then this is a representation of an Fab binding at the very, very bottom of the stem region, in what's been now termed the anchor region. This is work from Julianna Han in Andrew Ward lab.

So what you see is there are many sites of vulnerability on influenza HA, and I could have given the same summary about NA, because you have antibodies that decorate all the surfaces of NA, which is the other surface protein on influenza. Some of these sites of vulnerability have high sequence conservation which bodes well for vaccine design, for broadly protective vaccines, and also for the idea of isolating antibodies that could be used prophylactically or therapeutically against a wide variety of viruses.

Now where are we now? What we need are better reagents that are easier to use and safer for doing really large-scale studies. So we need reagents that reflect the wide antigenic diversity of flu strains we call breadth, breadth of neutralization, breadth of recognition. We need safe reagents that we can use for highly pathogenic strains that normally you'd have to use in BSL-3 or above. For potency studies, we'd like to use those in BSL-2 because those labs are more available.

Also, you'd like to have systems that determine the antigenic sites without having to do co-crystallography or cryo-EM high resolution, which are very laborious studies for the purposes of epitope mapping.

The same sort of systems can be used to assess for escapes. So we all understand variants after the COVID pandemic. So flu also makes variants that escape antibody neutralization, so reagents for loss of function.

So we've done some studies with Integral Molecular. They are developing reagents that are in this direction. So these are studies we've done together with influenza B antibodies. And so, there's three antibodies here that we've made called FluB-115, 160, and 117. They were tested in a pseudovirus microneutralization assay against three different strains representing the two lineages of influenza B.

So influenza B has Yamagata and Victoria lineages. We can see that you get various breadth of reactivity of these antibodies in the pseudovirus neutralization. These pseudoviruses can be used in dilution with replicates. So you can get errors bars and nice titration curves, and you can calculate IC50 values very clearly.

You can also use particles that Integral is making. It's called TiterSafe. And so, the idea is it's safe. Instead of using virus to agglutinate red blood cells, which is the ... HAI assay is actually the only regulatory accepted correlate of protection for a vaccine and other immunotherapies for flu.

But using virus for a high-path virus is difficult. So here TiterSafe is a reagent consisting of HA and NA on particles that have an HIV core. You can replace the live virus, and that makes it safer because the virus is nonreplicating.

In this case, they were testing our antibodies 115 and 117, which are to the HA receptor-binding domain and should have activity. You can see the 115's quite potent against all the strain tested, whereas the 117 is more limited. On the bottom left, you can see the ... If you're familiar with the matrix versus pelleting result in an HAI assay with red blood cells, you get the same visual effect with these TiterSafe particles.

So I think there's a lot of development in the field that's going to make this even faster and better, and we'll find lots more antibodies that are helpful. So good science is team science. This is our group in Crowe Lab at Vanderbilt Vaccine Center.

Also mention our collaborators, Ian Wilson and Andrew Ward at Scripps. I also showed you some overlays of HTX proteomics studies from Sheng Lee at UCSD. Data on all the particles I showed you at the end is from Maya Cabot, Nathan Krump, Allee Sheetz, and Parul Ganjoo. We have funding from a number of sources. Thank you very much.

Rebekah Kuschmider:

All right. Well, thank you so much. We can move on to the Q&A now. Before we begin, I'd like to introduce Robert Carnahan, the associate director at Vanderbilt Vaccine Center, who will be joining us for this portion of the presentation. Welcome, Robert.

Reminder to everyone, you can still submit questions using the Q&A tab to the left of your screen. I see we have a lot of questions already. We'll try to get to as many as possible. All right. To begin, I believe this is a question for Kyle. Can you use binding data as a substitute for neutralization data?

Kyle Doolan:

The short answer is yes, but always be careful what you're testing for. If you're looking for neutralization, you should test for neutralization. Binding is a reasonable proxy in some instances and for some viruses. As we warned in the TiterSafe and the hemagglutinization, there you're looking at it as a competition assay, which is one step above binding. But there were also modes of failure.So you want to balance throughput with biological insight and really understanding that biology is a key aspect of whether or not that substitution is appropriate.

Rebekah Kuschmider:

All right. Thank you. The next question, this one's for Robert. Do you think it's possible to achieve a pan-neutralizing influenza antibody therapeutic? What characteristics would such a therapeutic be likely to have, for example, neutralizing versus non-neutralizing antibodies and putative epitopes?

Robert Carnahan:

Sure, great question. The golden question for everyone in the field, honestly, is can we do it? I mean we know that these kind of antibodies exist, they're rare. I'll say all of the above, answer yes to the questions. Neutralizing or non-neutralizing, yes, we see both. You can have neutralizing antibodies that are broad and non-neutralizing antibodies. They all have their advantages. They can be to different epitopes.

There's an antibody being developed by Biere, a company that targets NA, for example. In contrast to what James Crowe talked about HA antibodies, you can have NA antibodies that are very broad. You can have HA antibodies like MEDI8852 and other antibodies that either are broad within A or both A and B influenza viruses.

So the answer is, yes, they all come with complexity. None of those have actually successfully got through a clinical trial. Some of them have failed at clinical trials for different reasons. Sometimes it's the bar is set too high. So do we want to prevent symptomatic disease or do we want to prevent disease progression? Those are things we have to think about.

So they target different epitopes. They're possible, they're rare. It may be that if you want to do prophylaxis, it might be very difficult to have a really broad one and you might have to have targeted ones that are seasonally brought out because they can prevent symptomatic disease rather than just disease progression.

Rebekah Kuschmider:

Okay. Well, thank you very much. Our next question is for Chun-Nan. How have you been able to or had to adopt your process to facilitate function-forward screening?

Chun-Nan Chen:

Yeah. I think I'll echo what Kyle said. Most of the time we look for binding antibodies and a lot of times the surrogate is, I would say, the blocking ... We get some blockade assay. The blockade assays, it's not a neutralization assay, but it's a pretty good surrogate.

So in our example of showing the anti-SARS-CoV-2 against the RBD, that one, we actually looked for blockade to human ACE-2 protein, and that blockade activity, I think, in a lot of cases does accurately predict the neutralization activity. So that's an example where the blockade actually will help with the screening.

But the same thing, it would be difficult to do with influenza because of the lack of suitable reagents and other restrictions. But we are working with others, including Integral Molecular, of designing functional or neutralization in the microscale. So that one will be very interesting to implement.

Rebekah Kuschmider:

All right. I think we are going to put our polls back up on the screen to give everyone a chance to answer those questions. I know some folks didn't get to do that. So let's put these back up if everybody wants to take a quick moment to read the questions again, read the possible answers, and submit your response. It's like a Q&A function for the audience as well as the Q&A for our speakers. I think we'll move on to our second polling question here shortly.

All right, this is our second polling question. Folks can choose multiple answers on this one, and then we will move back into the Q&A. All right. While our audience is finishing that up, let's get back into it. This next question is for Kyle. Kyle, can the RVPs be used for mutational or escape analysis?

Kyle Doolan:

Yes. So as the SARS-CoV-2 pandemic unfolded, we really watched this in real time as the virus mutated. In that instance, we've actually combined the epitope mapping and the RVP analysis that we've done. So in terms of mutational analysis, there we can make the RVPs for each of the alanine mutants and evaluate whether or not they escape neutralization by an antibody.

Then if you want to do the escape analysis, you're more likely to look at the epitope mapping there because you're going to define the epitope, and then figure out if directly your mutations are escaping that epitope. That being said, there's always conformational effects and things we don't quite check, so you always want to check it with the RVPs. But that's how it's gone. As we've made all the different mutants, people have tested their therapeutic in that way.

Rebekah Kuschmider:

Great, thank you. Moving to the next question, Robert, this one is for you. Your talk mentioned epitopes at the anchor region near the viral membrane. Is there a known mechanism of antibodies targeting this region? Are some of the effects of stem or anchor antibodies steric in that they prevent the clustering of HA proteins needed to facilitate fusion?

Robert Carnahan:

Yeah, several questions in there. But, yeah, so the epitopes of the anchor region are attractive because they're highly ... So if we're looking for antibodies that are broad, like sort of pan-influenza antibodies, these are a good place to start because they're conserved across most strains, many strains.

The function of this region, we think, is to tether the HA to the viral membrane. And so, it does have to do with the ability of the HA to move around. We don't exactly know how the antibodies that target this region work. We know that some of them are dependent upon what's called Fc effector function, so not pure neutralization by the antibody, but bringing in other aspects of the immune system.

Maybe there's also a component to it, as the question kind of alluded to, that there's some steric hindrance, that maybe these antibodies work by preventing the HA protein movement within that viral membrane. But there's also some thoughts that they prevent fusion. There's some other stem antibodies that have been described that show the ability to inhibit fusion. So still more work to be done I guess I would say.

Rebekah Kuschmider:

All right. Coming back to you, Kyle. With regard to this topic, how can SMOs and clinical research sites prepare for the future?

Kyle Doolan:

Yeah. I think public health is complex here. So obviously it's kind of big picture answers around what we started with, surveillance and testing. But I think that the mobilization, once we get into the central lab and clinical trials, establishing what you're going to do ahead of time in terms of what tests are important, what are the SOPs, what do we know about the fundamental biology of the viruses themselves, those types of things.

I think that the World Health Organization and the CDC are moving in that direction with prototype type class-specific viruses to study more in depth. That's the direction I think things will go in terms of a response.

Rebekah Kuschmider:

Okay. Chun-Nan, we have a question for you here. In today's talk, you discussed B cell discovery. However, T cell responses to vaccines and infection are also of considerable importance. Is your platform amenable to T cell evaluation and have you performed these types of assays previously?

Chun-Nan Chen:

The short answer to the first question is yes, but the market demand forced us to focus on B cells, and for obvious reasons. We can look at T cells, but we haven't really spent a lot of time on it. But what I can say is based on the capabilities of our platform, we can look at T cells in a number of aspects. So we can look at the ... We have ways to look at the cell surface expression using microscopy. So we can look at their marker expression.

We can also capture the secreted molecules from the cytokine profile. So we can look at what kind of subtype, subset they are. We can also capture their mRNA to look at their alpha-beta circulate sequences. So based on those number of information, I think we can do a very good T cell profiling.

Rebekah Kuschmider:

All right. Another question to follow up with that for you, Chun-Nan. You discussed the idea of isolation and storage of B cells or supernatants from early outbreak patients for later screening. Over what time scales can we think about preparing and still have sample viability?

Chun-Nan Chen:

We have been able to preserve that up to, I think, right now six months with some degradation of the detection signals. But then since we preemptively sequenced the antibody, so we'll have the antibody sequence on hand, we just don't know which one is going to be reactive to the antibody, even neutralization. But we were able to test them, for example, using Integral Molecular's RVPs, if they were able to identify the virus and develop something like that.

So that could be quickly used for identifying the binding antibodies, as well as testing the neutralization after the fact. So I think that's something we can do.

Rebekah Kuschmider:

Okay. For the next question, we're going to toss this one over to Kyle. Are there any functional differences between neutralization sensitivities for antibodies close to membranes in Integral RVPs?

Kyle Doolan:

That's a good question. I think it requires a little bit of speculation. There aren't a ton that are close to membranes there. If they are denying entry, if that's their mechanism of action, there will be no difference because that is what they're functionally equivalent to. If there is some aspect that relates to their ability to bud the amount of particles that's released upon subsequent infection, we're not going to see that in this particular assay as we've designed, though we have some ideas on how to assess those sort of things. The RVPs play a role in assessing it. It's just not in the assays we described today.

Rebekah Kuschmider:

Okay. Robert, we have a question for you coming in. What is the potential of VHH therapeutics for influenza?

Robert Carnahan:

So these are also called heavy chain-only antibodies. They're often isolated from llamas and other camelid species. I guess they actually come from other species, too. So I'm not a VHH expert. I mean I think that these molecules are really interesting in that they have a very flexible binding surface so that you can target them to all kinds of different regions, even ... You're not limited because you can use display technologies to sort through them. So you're not limited by an animal immune system necessary to generate them. So those are all really great.

One of the drawbacks would be they're not human. So as therapeutics, you would have to work through the fact that they need to be humanized so they're not recognized as foreign molecules. Then often you would also probably have another level of engineering.

For example, I mentioned earlier some of the antibodies that are dependent on other aspects of the immune system. They interact through Fc region. These VHH don't typically have that region, so you'd have to engineer on a region that interacts with other parts of the immune system, and then also make sure that whole thing is not recognized as foreign to the system.

So I think it's interesting for targeting. It still has some hurdles to overcome as for actually deployment as a therapeutic.

Rebekah Kuschmider:

All right, thank you. Another question for Kyle. What are the receptor densities on RVPs? Are they similar to native virus particles?

Kyle Doolan:

Yeah. So we really want to prioritize that our RVPs as a substitute for live virus and have mostly focused on their conformational and functional aspects. In terms of density, we do know that we see different densities, and the direct comparison with live virus is a little bit hard. Quantitative measurement of receptors on the surface of a particle is not an easy task, but it is something we're looking into. But I'll say that we see differences between variants of SARS and between virus classes, and that while we're probably similar, we certainly expect there to be some differences.

Rebekah Kuschmider:

Okay. Well, I think we are coming to our final question as we're nearing the end of our time here. We did have a lot of great questions today. We couldn't get to all of them, but we will try our best to get back to everyone who submitted questions personally after the webinar. For our final question, Kyle, back to you. What are the relative value and advantages of antibody and serum controls in neutralization assays?

Kyle Doolan:

Yeah, scientist controls are key. And so, you want an assay control because beyond when you're putting an RVP together with live cells and you're therapeutic, you have three components. You're testing one of them and you want to make sure the other two work. Cell health is a major aspect of that. So some sort of assay control is really important.

I think especially in serum and other more complex mixtures, a specificity control is really important. Are we seeing this as the result of the component or are there other aspects of the assay that you're doing? Not what you're intending to change but that are changing. And so, that's why we're running some of these isotype controls often to show specificity.

Then certainly a benchmark control is useful if you're doing a marking study where you're trying to look at relative values or pin it back to an absolute value, make sure that the assay is reflecting the same quantitative number. So we really consider all of those when setting up assays.

Rebekah Kuschmider:

Okay. Well, thank you all. We have come to the end of our time today, so I'd like to thank everyone for attending this Fierce Biotech webinar for submitting so many great questions. I'd like to thank our speakers for participating and Integral Molecular for presenting today's webinar.

A recorded version of this webinar will be available for you to access within 24 hours using the same audience link that was sent to you earlier. Thank you again for joining, and we look forward to seeing you at future events.