Founder Stories: Adrian McDermott, Global Head of Immunology at Sanofi Vaccines

Welcome to the influenza webinar series hosted by the Human Vaccines Project. The Human Vaccines Project is a nonprofit public-private partnership with the goal of decoding the human immune system to accelerate the development of vaccines and immunotherapies. In order to make real progress toward the development of a universal influenza vaccine, collaboration among researchers around the world is essential. We hope this monthly webinar will foster the scientific foundation needed to advance influenza research. This series of webinars will be broadcast on the third Tuesday of each month and posted online. For a copy of today's webinar, please visit the Human Vaccines Project website at www.humanvaccinesproject.org/talks. Following the presentation, we will have time for questions from the audience. Please type your questions into the questions box in the GoTo interface. I will read these questions following the presentation. Before we begin, I'd like to call your attention to an upcoming symposium on innovative vaccines for resistant infectious diseases and emergent threats being hosted by the New York Academy of Sciences on May 22nd in New York. The symposium will include talks from some of the world's leaders in vaccine development, including Wayne Coffey, CEO of the Human Vaccine Project. Please note the abstract deadline is this Friday, March 22nd, and the early bird deadline for registration is April 19th. And now it is my pleasure to introduce Dr. Adrian McDermott. Dr. McDermott is an experimental virologist and immunologist with experience in the fields of immunogenetics, vaccinology, and infectious disease immunology. He currently serves as chief of the Vaccine Immunology Program at the Vaccine Research Center within the National Institute of Allergy and Infectious Disease. Dr. McDermott leads a multidisciplinary research program that performs basic vaccine discovery B-cell research through to the assessment and clinical evaluation of lead vaccine candidates. Today, Dr. McDermott will tell us about monitoring B-cell responses to immunogens one cell at a time. Go ahead, Dr. McDermott.

Well, hello. Yeah, I'll get started here. It's a bit strange to speak at my computer screen. This is the first time really I've done one of these webinars. You're going to have to excuse me if I kind of drift off and end up talking to myself, which is what I do generally when my, uh, during my lab meetings when I talk to my group, I think, at times. Anyway, so, um, what I'm going to talk to you today about is kind of give you an idea about what the VRC vaccine portfolio looks like for influenza moving forward into the very near future., but also why do we do B-cell analyses on these vaccine trials. But let me start by just showing you this. This is the vaccine development cycle, as we like to call it, and of course I'm speaking to you now from the third floor of this building here, which is in Bethesda on the main NIH campus, and here this is where We design immunogens, and we do a lot of the research immunobiology that goes along with the iterative process I'm about to explain. Once we have a protein, for example, we then move that protein into the vaccine production program laboratory, and this is in Gaithersburg. It's about 11 miles up the road from here., and once they've optimised their systems, they then pass it on to the Frederick pilot plant. And this is where a lot of the— for those of you that are familiar— a lot of the paperwork and the painstaking filling of vials takes place for the vaccines I'm going to describe to you today. The vials then come back to the Vaccine Research Centre, and here we have Laura vaccinating one of the individuals that this photograph is actually taken, uh, during the H7N9 vaccine trial. Then it comes back into my hands, uh, and I run a laboratory up in Gaithersburg as well as here in Building 40, and this is where the endpoint immune monitoring assays take place. We've recently just moved up all our B-cell analytical wings, so now we have a sorter, and a couple of flow cytometers up there, plus sequencing capabilities and dozens and dozens of PCR machines. So yes, if anybody would like to come and visit us, it's a really nice facility up there. So what I'm going to start off with is the end and basically tell you that these are the takeaways I would like everyone to get from today. So there is a compelling need for a universal influenza vaccine strategy. I'll try and explain the path towards a universal vaccine or a rapid emergency response, and I'll go into that a little bit later. Then I'll show you how we do the B-cell analyses of vaccine trials and how we identify antibody lineages that may be used for pre-pandemic prophylaxis. Then I can speculate on the strategies that may preferentially elicit cross-protective antibody lineages using immunogens, and then some insights into the evolution of influenza HA stem antibody responses just as a model. We are doing head as well. So this is a schematic of influenza, and what I'm going to talk to you today about is the hemagglutinin protein. Although we are really doing a lot of work here now on neuraminidase, and I view this as the underestimated target of the immune response, I, I tend to think in conjunction with HA, and I'll go on to explain more about which parts of HA we may need in a vaccine. So the outset, I also like to think of influenza HA as a tree. So as far as a vaccine vaccine is concerned, or a circulating virus is concerned, each year the leaves change, but the trunk kind of remains the same. And likewise, on the HA molecule, we have lots of glycosylation and lots of dynamic mutations that change the face of the top of HA, whereas we have less glycosylation and a more static protein arrangement which is functionally conserved here in the stem region, as we call it. Influenza pandemics do occur, and, you know, many of us remember 2009 and the panic that surrounded that because we didn't have a vaccine at the time. But this is not without, of course, some attention from the popular press in Time and Newsweek, etc. But also, we have all these occurrences and in the media about attending county fairs in the summer and getting this H3N2 variant. I was actually paying attention during the, during the meeting in Nashville and pulled up and took a photograph of one of the slides here. But in general, we have to be prepared for a pandemic threat from influenza. So in response to this, the BRC clinical trials portfolio has included over the years H5 vaccine that we trialled, and we called that BRC310, and H7N9, a group 2 vaccine, which we call BRC315. We're just analysing now an H2N2 trial, which we call BRC316. And starting, well, started at the end of last month, we recruited the first few volunteers in this rather novel HA stem-only approach, which is on ferritin particles. And that's the, that starts in this year, and then next year we have an H10 stem going into trials. But I'll talk a little bit more about the portfolio later. Clearly, we publish all of these studies, and these are the H5 and the H7 clinical publications. And then we also try and publish all our science that we do, and this is one of the publications that we did based upon the H7N9 immunisation in humans. So I'll talk more about the, I'll talk more about the H7N9 trial as we move forward. So how do we identify influenza HA-specific memory B cells following vaccination? Well, we use specific B cell probes, and when one makes a B cell probe, clearly we know a lot about the structure of that part of HA or that part of HIV or that part of EBV, depending on what we're studying at the time. However, one of the Vague reason using that technology for the full-length HA is the fact that the B cell, the HA attaches to every single B cell. And so when I first came to the BRC in 2011, this seemed to be an insurmountable problem. However, there was one bright chap called James Whittle, and all he did was mutate one oxygen molecule in position 98, and he abrogated binding of sialic acid, which meant that now we were actually picking up specific memory B cells to HA, a novel HA at this. This was H5N1. However, and I must say the caveat with this is that if we use these tools for antibody identification we do not pull out any receptor binding site antibodies of any worth because we've mutated that region. So how do I identify antigen-specific cells? Well, we take our PBMC sample and we incubate our PBMC sample with the probe itself, and here we see a pre-vaccination— this is H3 here and H7. This is a group 2 probe set, and obviously we have an established immune response pre-vaccination against H3. We have some cross-reactive populations here, which when we initially started this with the tech Stephen Marr, going way back to 2011, I actually thought this was a compensation error. But this turns out to be the stem or some of the cross-reactive head epitopes that these B cells are recognising. So this is very useful, and this spawned a lot of publications and a lot of good work. Clearly, these individuals that we had in this vaccine trial with an H7N9 vaccine do not have any H7 head epitopes that have been recognised by their B cells. Following vaccination, they'll say 2 weeks following vaccination, you can see that the vaccine worked because now we've, we've listed memory B cells recently. We've now, ourselves and the Scripps group, published stabilized group 1 HA stem. And so now we can add this into our, into our analyses so we can pull out stem, stem memory B cells specifically. And then recently we've also got a group 2 stem. This is an example of how we can utilize these stem constructs in order to interrogate and analyze memory B cells from our vaccine trials. And here obviously this is H7 stem only, these are the cross-reactive cells that we can have an H H1 stem. So what's the workflow for sequencing these B cell receptor immunoglobulins? Well, one of the things that we've been working on quite hard recently is this BCL-6, BCL-X B cell immortalization. And when we're in the— when we're in, we're considering— we're considering looking at repetitive iterations of, of the genes that are being used, where we look at transcriptomics, for example, these are actually pretty good models, and also they serve as very good controls as well. So one of the advantages of being at the Vaccine Research Centre is the fact we have Mario Roder, and Mario has a very good relationship with Becton Dickinson. And recently we, we have this prototype S6 sorter which has 5 lasers and can sort 50 parameters, and it has other bells and whistles which enable much better index sorting. So it has 6-way index sorting, for example. So it really is a powerful tool. One of the things that having so many colors allows us to determine the exact phenotypes of the B cells. I won't talk about this today, but clearly this is something that we're very interested in, particularly from an immune monitoring perspective and an immuno— a B cell immunobiology perspective. We can take all a collection of memory B cells and we can bulk sort them. Now, this does not allow us to discern the individual pairing of the heavy and light chains, of the antibody, but it can yield gross distribution of VH or VL. So for example, we've been using bulk sorting to have a look at bone marrow, for example, where obtaining individual cells can be quite challenging because plasma cells don't carry any BCR on the cell surface, and so we can't actually pull them out using probes. But our workhorse is this single B cell sorting using the flow cytometer, and it allows us to discern individual antibody B cell receptors looking at the paired heavy and light chains. So what we do, we, we sort them into— we sort single cells, we generate cDNA, and then we do 2 rounds of PCR for the heavy, the kappa, and the lambda, and then we sequence. So actually, it turns out if you were to do a whole plate, it turns out that, that you're doing 6 rounds of PCR 3 times. So we've been, uh, with Danny Dwyer here, we've been playing around with new technologies, and I know that James Crowe's team uses this, um, uses this technique a lot. The alternative is to employ new technologies such as the 10x for sorting. 10x is a very elegant machine. We are actually working on a technique called SiteSeek but that's not really for prime time right at this moment. You can put 10x in combination with the 50-parameter sorter, and now you start to get to a, um, a really complex level of sequencing, and we are playing around with this for very rare populations, for example, in malaria or in Ebola. So the HA probes and the knowledge of the structure of HA has allowed us to isolate thousands of antibodies from our H7 trial, for example, our H5 trial. And so following the H7 or N9 vaccination, and I showed you this plot before, we can pull out many, many sequences which fall into a couple of groups. For example, Those using the heavy chain 118 and 16 and 323 we've published on before, but you can derive these from the H5 group 1 immunogen or the H7 group 2 immunogen. So what we have discovered is that clearly on, on the bottom here is the timeline of some of the more recognized pandemics going back to 1889. And we are all exposed to influenza by the age of 4 or 5. So which parts of the HA are we exposed to most? Well, there is an expansion preferentially to head epitopes, and Ravitch and others have conjectured why this may be. But it's a very immunodominant part of this molecule. And so this is actually reflected in your serum antibody concentration, where you have a lot more antibody, and a lot more antibody following infection and during infection that is directed towards the head than you would do towards the stem. However, if you are exposed to two different strains subsequently what you see, for example, H1 to pandemic H1, what you see is that because the stem is more conserved, you tend to— all those pre-existing memory B cells are preferentially amplified. And so the situation is reversed where you have a high frequency of plasma cells, memory B cells, and thus antibodies in your circulation following an exposure to a novel HA strain. And clearly this was noted in 2009 by Rammertz and colleagues and also by Raphael Medzgerud, and they published a couple of papers just looking at stem-specific responses, but also We'd really like to know what the persistence of those antibodies are to stem throughout the course of infection. And I know that Scott Hensley's done some work on this, and some other groups have done it, and we do some ourselves here. But it seems that the stem-specific responses on a population basis seem to be quite low compared to the head. And this is kind of an interesting B cell problem as why this would be, so I'll mention that again soon. So this makes stem an attractive target because of its conservation, but probably it'll only be a part of a potential universal influenza vaccine solution. And then we also have to have a look at what is the path of a stem-only construct into the clinic, because it clearly each influenza vaccine is licensed upon the ability to, to, for hemagglutination. And because you don't have a head on this and don't have receptor binding site, this is not how you would show potency or even equivalence. Let's concentrate a little bit more on the stem just briefly. There are antibodies that recognize just group 1, group 2, or those antibodies that recognize, cross-react with group 1 and group 2, and they neutralize all of these in here. On the first, on the first spike, what we see is this is a BH169 antibody footprint which is specific for group 1 neutralization. This is the same for F10 and CR6261. The recognition here is by a slightly mutated hydrophobic CDRH2. Distal to that, we have the area where the group 2 neutralizing antibodies bind, such as 8020 and 8043. And they use different heavy chains, but this recognition is by the CDR H3 genes, and they bind slightly differently. Now, those that in general, those that recognize Group 1 and Group 2, um, influenzas equally and neutralize the majority, have this kind of pattern. And, um, for me, this looks like, um, an outline of the British Isles. You'd have Cornwall down here, and you'd have Liverpool here and Edinburgh up here and Glasgow. But anyway, clearly there's no Europe next to it. So if we put the glycans on this model, we see why this distribution of footprints is what it is, and that's because in a group 2 HA, in the stem region, there's a glycan that blocks the group 1 antibodies binding. In a group 1, they have a glycan distal to the binding site, and that blocks the group 2's binding. But this group 1, group 2 kind of slides in between, in between both of these glycans, be it on a, on a group 2 or a group 1. So that's the— they're the antibodies that we're after if we really want to target HH stem. This is clearly the subject of quite a large trial with GSK now and Dr. Peter Palese. And ingeniously, what, what he did here, he took the full-length HA and put, and put on the head of a, um, the head of an influenza that the population had probably not seen. And as I showed you in the schematic before, What happens is, is that you promote the frequency and hopefully drive some of the somatic hypermutation towards the stem. So this trial is ongoing. I know that Patrick is doing some analysis of this, and it'll be very, very interesting to see what the outcome of this trial will be in a couple of years. So what I want to show you now is some of the analyses and characteristics of B cell receptor immunoglobulins derived from BRC vaccine trials. And in general, this is the schematic of— this is what we use, the schema of what we use. And I'll focus mainly on the H5, which is done in 2010, and the H7 vaccine trials. And the one that was done just this year, we're still analyzing the data from. But as I mentioned before, we have published these, and you're welcome to copy that down. So we can identify, using our probes again, we can identify plasma blasts. Yes, we can actually identify plasma blasts, whether it's a quirk of influenza and Ebola, and now we've seen it in HIV, but we can actually recognize plasmablasts using probes, because the dogma is, is that plasmablasts have downregulated their BCR and therefore they don't bind probes on them. And as I mentioned to you before, we take these cells and we single cell sort them into, into wells, and we obtain the heavy and light chains. And what we notice is that vaccination with the Group 1 influenza subtype is dominated by VH169. This was all— this was noticed by all the large immunology groups. Antonio Lanzavecchia and the Grasco group also published on this, and there's many alleles of VH169. There's like 11 alleles of VH169, and there's a nice story if you'd like to read the Grasco paper and also our paper by, uh, with Adam Wheatley, where, um, we look at the different alleles of VH169. But in general, Group 1 influenza subtypes— sorry, for Group 1 influenza subtypes, they elicit this VH, strongly VH169. And if you remember me saying just recently, a few slides ago, it was purely by the fact that they have a hydrophobic CDRH2. And that is what this is showing. So when we have a look in our H5 and H1 vaccine usage, we see VH169, and also in the recent H2H1 trial, we see a lot of VH169 as well. So there is a polymorphism at this position 54. A leucine or a phenylalanine, and, and that determines if you make a very large VH169 response. And that's in these papers here. However, because of what I've just told you about the glycans, the VH169 in general— there are some that cross-react like 9114— cannot bind to this region because it's got a whacking great glycan there. So when we look at the H7 or H3 vaccine trials, what we see is that VH169 is replaced by VH118 or by VH61. There we go. And we published that not that long ago in Science Immunology and in a couple of reviews as well. What we did find a couple of years ago, and Adam Wheatley and Gordon Joyce did this work, the majority of this work, is that we found that the population that we looked at was purified for VH118s and VH6-1, which recognize group 1 and group 2 stems. And can be found in most individuals. So we call these multi-donor classes, and these are the characteristics of the multi-donor class. And when you start looking at binding of these, what we see is that they bind across a whole range of, of HAs. And then when we put them into mice, what we see is that these can actually be quite strongly protective. So while, while the neutralization may not be fantastic, they're certainly protective using other mechanisms which we can conjecture to be ADCC. But we put them into different kinds of mice and they still, and they still protected, so we're not sure if it is just ADCC. So Group 2, an H7 immunisation, elicited a more cross-reactive memory B cell responses than using a Group 1 immunogen. And we can see that here. So if you use our plots again and the probes, what we see is that the H7 and 9 frequency of HA really cross-reacted within the Group H5 H5N1, and then we have the H2N2s. But if you have a look at the H7s, what we see is that those that are— those ones that recognize H7 cross-bind, so they bind outside of the group, and also with— which is less so for the group 1 immunogens. We can represent this a different way, and we can look at the following, uh, H7 or H5 immunization. We see that the VH169 is really quite dominant here for the H5, which is a group 1, and very low here for the H7N9. However, we take like a 118, we could take a 61 as well, and we see that the population there's less 169, and there's a greater amount of these. And these will— and these are cross-reactive. These recognize group 1 and group 2. So in H5, of these multi-dominant classes that we've described, we get about 5%. However, when we use a group 2 immunogen such as H7N9, we get 30%. So this may have some some lessons we can learn for sequential vaccination using Group 1 and Group 2 emergence. And we have isolated a lot of sequences from the H7 trial, and I would like to emphasize that we're not an antibody discovery laboratory. So we have 600 sequences here, we have 2,700 sequences here, but we only make a few of these just to characterize them., and we're in discussions with various groups to try and characterize these more thoroughly. Now, some of the H7 sequences that we've expressed bind stem really well. This is what this, this blue plot is here. We also do pseudoneutralization in the lab, and we can see that, for example, this, this antibody here and the one below it, bind really well. But the profile of looking at pseudoneutralization is not fantastic. Even when we look at micronutrilization using an in-house assay that Barney Graham's group is, is developing, we don't see a lot of neutralization for one of the antibodies. However, again, when we put them into mice, we see very good protection studies. So for the one that shows absolutely no neutralization here, we show that it protects really well with Group 1 and Group 2. And the one that does show some neutralization across Group 1 and Group 2, it shows very good neutralization. So this work was done by Sarah, and Adrian Karangu has been developing this mic neutralization Quan Li did the pseudoneutralization for this, and Masaru helped us with the mouse study. So this is really, really encouraging for the use of these antibodies. If you have a look at our strategies that we currently have, we have the heterologous monomeric RBD. So therefore, this is, this is a number of different HAs on a particle, and we give this We've given this to mice, and we recently published this in Nature Immunology— sorry, in Nature Medicine. And we're actually working this up into being in full-length HA here, and we're doing this in collaboration with Neil King, and that's a very exciting project. Oh, this— these headless stems are the ones I was talking about at the beginning, where we are— they're going into clinical trials this year and next year. And then we have the conventional vaccines as well. So as I've said, the full-length H2 influenza and HA stem only are both on ferritin particles. So the H2 influenza trial is finished and it was very successful. That should be being published soon. And so they're both currently in phase 1 clinical trials here. So why— what is the utility of characterizing potently protective cross-reactive influenza-specific antibodies? So it informs focused immunogen design. That's one of the main things why, why we do it here. So immunogens can be epitope-based, so we can have a look at the cross-reactive HA or stem regions. We can look at specific immunoglobulin lineage-based, which is probably a strategy that's more known in HIV with eODGT8 or some of the other strategies. But there are people that are looking just to elicit the VH6-1 lineage. It can help improve the design or, and design the full-length HA approaches. So that's what I was saying, the work that we're doing with Neil King with these self-assembling nanoparticles. Particles. It can inform sequence of the vaccination in newborns, like for example, the data I showed you where we can get more multi-donor B cells with a group 2 antigen followed by group 1 rather than vice versa. There's a really hot area right now, and this passive antibody transfer pre— the way I see that it could work with some of the antibodies that we found are pre-pandemic potently protective antibodies, H7, H5, H9, and this would, would get you over that lag period before you can actually develop a conventional vaccine. Also, it can protect vulnerable populations, so the young and the old and those that are infirmed maybe, and then there's always a possibility of treatment. So the reason why I put so many question marks behind this is that there is some evidence to suggest that passively infused antibodies do not get to the areas of the lung that we actually need to protect. However, it also academically gives us some insights into the evolution of the influenza B cell and antibody responses, and I'd like to talk just at the end here about some of the aspects of this. So clearly there's a lot talked about imprinting and about original antigenic sin, and really it's just a very good demonstration of B cell memory responses. So your first exposure to a group 1 or a group 2 influenza, what that does after a number of years, that means that if you have an outbreak of a group 1 or group 2 influenza and you were initially exposed to group Group 1, then the likelihood is that you're going to recover on a population basis. A Group 2, you may not do as well. For example, with an H7N9, if you're exposed to a Group 2, you will handle a Group 2 variant quite well, whereas you may end up becoming somewhat sicker with a Group 1 exposure. And this really was the subject of a really— what's the word I'm looking for— a really provocative paper from, from, by Katie Gostic. And it has really spawned a lot of conversation, this paper, both in the epidemiology groups and also in the immunology groups. And I think It's also spawned a lot of grant writing as well. So clearly this wasn't new because Thomas Francis back in 1960 came up with the doctrine of original antigenic sin. And clearly what that basically says is that you remember the first flu preferentially that you were ever exposed to. So how does this feed into what we've been showing here, um, to do with vaccination? So HA stem binding responses elicited by pandemic vaccines, um, the response is largely a reactivation of pre-existing memory B cells, and this may be an underestimate, but it's about 60% of the group 2 antibodies are just recall, whereas— and you see that on this graphic— whereas maybe there's a very limited amount that are directed towards the head. So maybe most of our stem responses are recall, and we remember mostly the ones that we react against when we were children. We can see this here, even in the clones that we have. If you look at the diversity, um, pre-boost and 1-year post-boost are more or less the same when we look at H7 head. And then when we look at H7H3, which is the stem region, it's about the same. You, you may be getting a little bit more 1-year post-boost. Look at a mutation level, we see a similar, a similar pattern where we induce some somatic hypermutation. But more or less it's in the same ballpark. So if you have a look at some specific individuals, and because we do these vaccine trials at the research centre here, we do, we may have exhaustion of individuals like for these three, they were actually in both the H7N1 and H7N9 vaccine trials, so either we're paying them well or they've got some vested interest. So if we look in the H5, the group 1, these are the group 1 antibodies, and these are the, these are clonal lineages that were, that we found following H5 and following H7 vaccine. And there's these two, would like to point out these two VH6-1 class lineage 1 and lineage H3. And in this individual, the same thing, and again, two classes kind of dominate. And also what this goes in with, it's very much that the H7 bring out these multi-donor classes. Like here, this is a good example. So these are just plots showing that post-H5 and one of Lineage 1 and lineage 2, sorry, lineage 2 and lineage 3 in this individual decrease and then bounce back. Likewise for those lineages in this individual in subject 2. And this does beg the question that when is this imprinting of stem responses actually happen? Well, there's many groups now in response to the civics that are establishing child cohorts all over the world. We have some inroads to this ourselves, which may answer these questions about imprinting in stem more specifically. And clearly imprinting in head is also a really big issue. But to carry on with this, let's have a look at the cross-reactive multidonor HA lineages initiated by sequential exposure. So going back to those individuals again, subject 1 was born in 1962. And so we have to— so we can actually, we don't know the exposure history, and you're really guessing in these kind of studies. But if we infer the UCA, we can see that the UCA binds to this. Maybe it was an H2N2 It binds to the group 1. So this lineage was born on group 1, and then over the years it's adapted to be recognizing a group 2. Conversely, this lineage was born in a group 2 environment and then learned how to, or mutated how to, recognize group 1. And we can have a look at someone who was born in 1970. And say, well, they were clearly exposed to H3N2, and so they have a Group 2 lineage with this, um, BH.118, which then mutated to recognize both lineages. And again, somebody who was born in much later, in 1984, also has a Group 2, um, a lineage brought up on Group 2, which then goes to recognize Group 1. So these kind of studies are very interesting, and longitudinal studies such as these where you can actually follow children exposed and you know what that exposure is, I think, are going to yield lots and lots of information about vaccine strategies and about vaccine design, etc. So just to end, in summary, the takeaway from today is there is a compelling need for a universal vaccine strategy. Particularly when we look at pandemic preparedness. There is a pathway towards universal vaccine, and, and we can do those analyses, and that's what I've shown you today. And also with the antibodies, maybe there's a rapid emergency response that we can put into place. B-cell analyses of vaccine trials, well, besides being great fun, we can identify lots of antibody lineages that may be used for pre-pandemic prophylaxis. We can speculate on the strategies that may preferentially elicit cross-protective antibody lineages. So do you, do you go with a group 1 first and then a group 2? Do you go for stem followed by head, etc., etc.? And in the coming years, I'm sure we're going to play around very much with this. And then also it gives us insights into the evolution of influenza H1N1. HA stem-specific responses, which again gives us a lot of insight into how the immune response works in general. And I'm thinking now, because we do so much work with the EOD-GT8 protein, this may be something that we'll be looking at in a couple of years with that, with the immune responses to that particular protein. So with that, I want to really acknowledge the group These guys do all the work. It's absolutely fantastic. Sarah is the star of this. She has really kicked on where Adam Wheatley started. Adam's now at the University of Melbourne, and she has some very, very, very good students working with her. Michael Chambers just moved on, but Julie Rabb is still here. Clearly, the Viral Pathogenesis Lab with Barney and and Michelle and Masaru. We couldn't do any of these studies without them. Likewise, we couldn't do any studies without Julie Ledgerwood, Grace, and the clinical team, plus all the vaccine recipients and participants. Clearly some of them, well, 3 of them come back for more, so we must be doing something correct. So with that, I'd like to thank— thanks for the opportunity to talk to you And also, thanks very much for your attention.

Yeah, thank you, Adrian. That was really interesting. At this point, we can take questions from any attendees. So if you have a question, please just type it into your question box and I'll read it aloud. And while we're waiting for people to sort of get moving typing their questions in, I'm curious if your group has any plans to further investigate the mechanisms of those protective but non-neutralizing antibodies.

Yes, we are. We're actually doing quite a bit of work on that, to tell you the truth, um, using different mice with different collaborators, um, really, um, trying to understand with LALA mutations, etc., etc. Um, we have some— again, going back to the HIV stuff, we've got a lot of experience looking at the FC portions and how to manipulate the FC portions. In HIV antibodies, and we're applying a lot of that into the influenza space.

So yeah, cool. Okay, so here's the, the next question. It says, nice talk, Adrian. Could you please comment on the sample requirements for your single-cell sequence analysis? What is the frequency and number of cells you need to be able to pull out meaningful sequences?

Well, that's a real numbers game because it depends when you do it. So if you If you look 2 weeks post-vaccine, you can maybe get away with using 10 million PBMCs, and you can pull out a really good variety of B cells, memory B cells that are using all sorts of different heavy chains. However, if you go a year later, or maybe 8 months later even, then the frequency of the circulating memory B cells has gone down so low. Now you're just looking at a smaller population, and therefore you may not pull out the, the diversity that you would pull out just post-vaccine.

Okay, our next question says, what would be a strategy to break the B cell memory that we do not want?

Would it be antigen adjuvants? Oh, that's an interesting question. Um, would adjuvants play into this? One of the things about these lineage studies is what we've seen using malaria is, um, while an adjuvant can really affect a large frequency change early on, the adjuvant doesn't really really affect the long-term, the long-term antibody levels. That's what we found using some neoantigens with malaria. So while it recruits a lot of memory B cells, and I think if it's an immunodominant response, then yes, an adjuvant will work. If it's a subdominant response, then we have to really question the strategy of the adjuvant that we would use. So for example, I think using stem only on a ferritin particle, I think that's going to be hugely interesting to see if we can break that immunodominance by just giving a more or less self-adjuvanted stem and see if we can affect the frequency of, of those multi-donor class antibodies in the serum.

Okay, the next question says, great talk, Adrian. Can you explain the gray sections of the pie graphs where you were looking at the cross-reactive stem binding lineages that change over time in vaccinated subjects?

Oh, you mean those ones? Oh no. What the gray areas are here are all the other all the other lineages that don't repeat. So these are the only two lineages that have repeated in, or two or three lineages that have repeated. And so while we have information on them, it was better to tell the stories of the ones that have actually been repeated. And they, and the one subject three is probably a very small population here, and we've not picked it up any particular frequency, but we did pick it up in a higher frequency when we use Group 2, if that explains it.

Just the way you make it. Yeah. Okay, next we have, you talked about stem antibodies, but what about different classes and subclasses of antibodies? IgA stem versus head, for example? Mucosal antibodies, how do we analyze them?

Well, I didn't present that actually. We, our primers are designed to pick up IgM, IgG, and IgA, and the H2 vaccine is yielding some very interesting results in that area because a lot of the individuals who we saw who had clearly were born— in fact, the graph's in front of me— so were clearly born in the, in the time when H2N2 was in circulation. You see that they predominantly have an IgA response. And so we're looking into this a lot more now, and probably moving forward We're going to age stratify all of our vaccine studies moving forward because of this, but that's another story for another time. But I think, I think we, we do have a lot of information on the IgA versus IgG versus IgM responses, and, and that story will begin to tell probably next year.

Okay, next we have: B cell responses don't necessarily correlate with antibodies in the serum. Have you looked at the presence of IgG in the serum by mass spec, and do you have a sense for the levels of these antibodies in the serum?

You know, getting that technique off the ground— if I was talking about it this morning— getting that technique off the ground is really difficult. And that's why really, um, the biggest show in town in this is, um, is George's group, um, George Georgiou's. Um, the technology behind doing mass spec, uh, we don't have the expertise here. Interestingly, though, there are some individuals who are in the analytics section of the vaccine production who use mass spec all the time to characterize vaccines who may be interested in doing this, but we have not done that yet, and if we did do it, we'd probably do it in collaboration with George's group.

So Yeah, we're looking at it. Okay, next we have, how about the longevity of the stem-specific response of trial participants? Do they maintain a stem-specific response after seasonal influenza exposure?

Great question. Um, yes, I think, um, Scott has some data on this to say that there are individuals who do maintain a really good frequency of antibody responses against stem. And then there are other people who do not. And I can't remember his story, and I would prefer to let him tell it. But we, we've looked at this, and we, we've looked at all the trial participants, and the stem responses do go away pretty quickly. The antibody response does go away pretty quickly. That's why when we use— that's one of the criteria that we're using for looking at the stem-only constructs to see what the persistence of those antibodies is actually going to be.

Okay, and it looks like we have a couple questions on T cells, but they're the same, so I'm just going to read one. It says, can you talk a little bit about T cell-based vaccines and their role in universal influenza vaccine strategies?

Oh my goodness, well, that's not my area really, but clearly there's a lot of work being done looking at T-cell vaccines, a similar kind of, similar kind of pattern to what we see with the B-cell responses. There's a lot of immunodominant epitopes that are recognised. When you look at CD8 T cell responses. We actually have started to do some CD4 T cell responses on our vaccine recipients, and that's— and those stories are very, very interesting as well. But as far as CD8 T cell responses in the context of these vaccines, I can't really comment because we've not done that work.

So, okay, and we have one more question, so this will be the last one. It says, how does stem-based antibody perform in terms of protection or neutralization immunization as compared to antibodies specific to the head?

Right, well, that's again, that's the big— that's why I think stem would only be a component of a vaccine. And I said there was two reasons for that. One was this longevity reason. The persistence of those stem antibodies is not great. The potency of some of them is equal to some of the middling head ones, But again, the head outperforms— the head-specific antibodies outperform the stem-specific antibodies. However, if you're looking for something that would give you a long-term protection, then probably vaccine strategies such as the one I've been describing today to try and elicit more and more potent and long The durability, increase the durability and how long they last. I think that is what we can do in the next couple of years, and whether that proves to be significant in a population is only— vaccine trials are only going to, you know, with our partners is the only way to actually solve that problem. But yeah, I mean, the head dominates entirely over the stem right now.

Okay, great. Um, thank you again, Adrian, for such an interesting talk today, and thank you to everyone who joined. I hope you all have a great day.