Peter S Kim - Steenbock Lecture I - Fall 2026

Okay everyone, it's close to three. Uh we'll get started. Welcome to this year's uh Harry Stebach lecturesship. So this is one of the most pre prestigious honor bestowed by our department. Harry Steinbach was a pioneering leader in the field of vitamin metabolism as a member of our faculty in the earliest 20th century. During his time he discovered that vitamin D deficiency is a cause of ricketetts namely the weakening of bones in children. And then afterwards because of his research and then he were able to establish the therapeutics for the disease and he also played the instrumental role in establishing the Wisconsin Alumni Research Foundation namely Warf today which is now among the oldest and most successful technology transfer offices in the United States. And this year we're lucky to have uh professor Peter Kim from Stanford give us the steamog lecturesship. So Peter is the Virginia DK Lubic professor of biochemistry at Stanford. He started his career at the as the whitehead fellow at MIT and then went on to be the associate head of the department of biology and the Howard Hughes medical uh institute investigator. And then uh after around 10 15 years he was at MIT he be he decided to go to became the president of the merk research laboratories from 2003 and 2013 where he oversaw the development over 20 new medicines vaccines including gardicil the first vaccine for cervical cancer zostax the first vaccine for prevention of shingles in adults as well as kuda the first FDA approved and PD1 inhibitor for the treatment of cancer among many others. Yeah. And then so uh he's known for discovering the spring loaded mechanism for how proteins cause viral fusion with host cells for which he was awarded NAS award in molecular biology. And this important discovery laid the foundation for a critical aspect of protein engineering namely stabilizing the meta metastable state of virusion proteins now found in many different vaccines including the vaccines for SARS kov2 which caused the co9 pandemic and among many awards accolades uh I want to say uh he's a rare group of people who's an elected member of the national academy of sciences national academy of medicine as well as national academy of engineering ing and then uh the final words I would say he's also a tremendous mentor he has mentored um dozens of students postocs and among them six have have become members of national academy and one became the president of the HHMI and many others in leading uh institutes in biotechnology as well as pharmaceutical industry yeah without further ado let's welcome Peter >> [laughter] >> Thank you very much DA for that kind introduction. Uh it's a pleasure to be here. I've had a wonderful day uh meeting and talking with people and I'm looking forward to uh future meetings. Um this lecturesship comes with two lectures. [laughter] So I actually had to prepare two lectures uh for this trip. And so today what I'm going to talk about is uh vaccines. And tomorrow I'll talk about uh using protein language models to uh evol understand and evolve uh proteins. But today's focus is on vaccines. And I thought I would just start by saying that you know vaccines really are one of the most profound public health uh achievements of biomedical science. And I think after uh probably after clean water, one of the most coste effective ways to prevent serious disease and death is vaccines. You know, you think about it, you can give a uh a baby one or two shots of a vaccine and they can be protected for life against a disease that could potentially uh kill them. Um, importantly they affect not just the health of individuals but also the health of uh entire communities and that in turn leads to uh actually big economic uh issues with countries and there indeed where countries don't have access to vaccines they're profound economic consequences. The third bullet is really remark a remarkable statement and that is that a disease a disease called smallpox which had a 30% mortality rate uh was declared officially eradicated by the world health organization in 2011. Just think about it. We've completely eliminated a disease from earth uh using uh vaccines. And indeed there are several other vaccines that have been uh over 90% eradicated in the United States and three polio, measles and uh rebella that are targeted for elimination by the WH. Um this is a picture that I like to show of uh people inside what's called an iron lung machine. And basically when someone would get infected with polio which is causes polomiitis it causes infection of muscles. If those muscles happen to be your diaphragm muscles that you need to breathe they would literally take such people and put them in these chambers where they could control the pressure so they could breathe. Right. And this shows the effect in in the United States alone there were times where there were over 20,000 cases of polio per year. And what you see is that when the sock vaccine came along, that plummeted. Then the live oral vaccine from saving came along. And in the United States, the last indigenous case was in the 1970s. Unfortunately, even though we have excellent vaccines against polio, uh it's not yet been eradicated. Indeed in Afghanistan, Nigeria and Pakistan uh there these three countries have had polio continue and have have never uh been interrupted in terms of the um the disease even and you know it's really you can see the big change between 1988 and 2014 when the WHO and others really put in a huge effort to try and eradicate polio. And the reason why I'd like to show this slide is to point out that it's not science that's holding us back here. We have the scientific solution to eradicate polio. The problems have to do with infrastructure and with issues such as guerilla warfare where people that are trying to vaccinate other people are literally uh killed as a result. So the point and and now in this country I can make the point more generally. The the problem of the vaccines are amazing things. They are among the most cost-effective ways to reduce serious disease and death. And yet there are scientific issues and scientific problems we need to solve. But there are also non-scientific issues. And those non-scientific issues also need to be solved in order for us to realize their potential. Today I'm going to focus on the scientific issues about vaccines. And there are three uh topics that I'd like to talk about. One of them is a very old story. It's older than many of you in this audience. Uh but I'm I I'd like to tell it because it has recent implications for uh vaccines that are quite important and it's been an important part of what made the COVID 19 vaccine successful and it's also now become incorporated into modern vaccine design. And then I'd like to talk about what I call the amunof focusing problem statement. There are some vaccines that have been very hard or to date impossible to make. And uh even though we think we know exactly what we need to do, we can't actually make them. And I'd like to talk about that. And then I'd like to close with um a vignette about vaccines to protect against Ebola virus. Okay. So here's the start of the story and the start of the story goes back to uh the late Don Wy who was an amazing structural biologist at Harvard and what he was able to do was to solve the crystal structure of this protein the influenza hemoglutin protein and the reason why that was such a big deal when he solved it was that it was known in the field that influenza hemoglutin the protein found on the outside of the flu virus uh surface was nec necessary and sufficient to cause membrane fusion. So I should say that you know in order for an envelope virus to get into a cell the membrane surrounding the virus needs to fuse with the membrane surrounding the cell. And it was known that the only thing you needed to have that happen was to have this uh protein on one membrane and the receptor scyic acid on another membrane. And if you brought the two membranes together and you lowered the pH you get membrane fusion. So all the information for causing membrane fusion was in this protein. There were two puzzles. One was that it was known from chemical studies that this part of the protein down here, it's a trimer and I'm just highlighting one of the three one of the monomers, one of the protein down here, which is called a fusion peptide, that this was the first part of the protein to insert into the cell membrane as part of the fusion process. And the reason why that was a puzzle was that the cell is up here. Here's where the viral membrane is. Here's where the protein is. The cell would come down here. The receptor binding domain is up here. So the puzzle was how did this fusion peptide way down here get inserted into the cell as one of the very early steps of membrane fusion. The second puzzle was that there was no need for added energy to cause membrane fusion. There was no ATP hydraysis. nothing was uh needed for energy and that uh was a puzzle that we'll come to. Well, we uh came across this problem when we were studying coiled coils. We originally started with what were called loosing zippers showed they were coiled coils and then we went on to trying to understand coiled coils and predict them from amino acid sequence. And Chave Carr when she was a graduate student set out to take a look at those proteins that were predict had sequences that were predicted to be coiled coils but then when we looked at the crystal structure in the PDB were not coiled coils and the one really prominent example that she found is shown in this uh yellow highlight here. This yellow region of hemoglutin is very strongly predicted by the uh prediction programs to be a coiled coil even though as you can see it's uh a loop region and coiled coils I neglected to say are alpha helyses that wrap around each other with a a superhelical twist but what she did was to synthesize this yellow peptide and when she studied it in solution she showed that in fact it formed a very stable three stranded coiled to coil. And so what that led to then was a model and the model shown on the left which we called the spring-loaded model in which that yellow loop region in the native hemoglutin upon dissociation of the receptor binding domains up here formed a three-stranded coiled coil a very stable three-stranded coiled coil. The net effect would was to propel the fusion peptide from the bottom of the protein up to the top of the protein where it was in the right place at the right time to insert into the target membrane. Subsequently, Don Wy structure solved the structure of the uh low pH converted form of hemoglutin and sure enough uh that yellow region had formed a stable three-stranded coiled coil. We speculated at the time that this spring-loaded model might be applicable to other uh viruses including HIV and that turned out to be correct as I'll come to in just a just a minute. But because of the implications for HIV, we actually shifted our work and started focusing on HIV. And David Chan when he was in the lab as a postoc solved the crystal structure of the core of the HIV GP41 protein which is the equivalent of um part of H he hemoglutin and indeed what he found was that just like hemoglutin there was a central three stranded coiled coil and the fusion peptides would be located up here. taken together with a lot of other people's work and ours this led to a model for viral membrane fusion and a model is depicted here. So initially uh as shown for HIV initially there's a viral uh protein that's on the surface of the viral membrane. In this case it's the HIV GP120 and GP41 proteins. It interacts with a receptor on the native cell membrane. That interaction with the receptor ultimately leads to a confirmational change in which now the spring is sprung. The three-stranded coiled coil is formed. The fusion peptide inserts into the cell membrane. And then there's a high affinity for this region which we call the N heptad repeat region. And this region down here which is we call the C heptad repeat region. There's a high affinity for each other such that they form a trimer of hairpins. And since this part of the protein is inserted into the cell and this part of the protein is anchored in the virus, that brings the two membranes together. And somehow magically that leads to membrane fusion such that after the reaction is over, you end up with just one membrane and you've allowed for membrane fusion to occur. The remarkable thing that came out uh over the next few years was that this structure this trimer of hairpin structure was something that was found in many different viral uh fusion proteins. Indeed here you see for retroviruses including HIV uh for parramixa viruses uh including respiratory sensitial virus filo viruses like Ebola or the orthamixa viruses in each case there's a three-stranded coiled coil that forms that's adjacent to the fusion peptide region. So in all of these viruses there's a three-stranded coiled coil that's adjacent to the fusion peptide region. And in all cases uh there's another region be closer to the C- terminal end of the ectomain that forms the trimer of hairpins. In case of HIV it's with alpha helyses. In the case of influenza it's with a long loop. So the way in which it forms a trimer of hairpins is different but in all cases it forms a trime of hairpins which presumably brings the two membranes together as I showed you on the previous slide. And in all cases, the set the um three-stranded coiled coil is uh in the middle of the structure. This is for the protein after it's been sprung is what it looks like. But what about before it's been sprung? Well, the remarkable thing is that before it's been sprung, here's the structure I showed you already for influenza virus. This is after it's been sprung. This is before. And what I've highlighted are the regions in yellow which are the three stranded coiled coil and the regions in blue which are the part that forms the trimer of hairpins on the outside. And here you see in a native structure that the blue and the yellow are uh mixed intermixed but look here at respiratory sensitial virus where the yellow and the blue are completely separated in the native structure and yet they form a very similar trimer of hairpin structure afterwards. In the case of HIV, we have another fold in the native structure. In the case of SARS KV2, we have a very different fold as well. So, it looks as though evolution has actually come upon this solution of causing membrane fusion uh through convergent evolution and it's set up the spring in many different ways in the native structure, but after it's sprung, it's come with the same solution of forming this trimer of hairpins. Okay. So that's the introduction to the energetics which is what I want to talk about uh in this first part. As I said with an enveloped virus there's an intact lipid billayer that surrounds the genome. And as you know a cell has an intact lipid billayer. In order for this virus to infect a cell those two need to become one membrane. Thermodynamically this is a favorable process but the actual activation barrier to cause membrane fusion is quite high. The the diagrams of the transition state of going from here to here are quite fanciful and they involve somehow smooshing the the head group lipid head groups that are polar together and coming up with a way of mixing the inner leaflets of the uh of the um membrane. In the case of influenza it's known it was known that influenza virus binds to its receptor scyalic acid. It then gets endocytos and in the endoome the pH gradually decreases until it hits a pH of around pH6 5.5 to six. At that point the confirmational change that I've been talking about occurs that allows the membrane surrounding the virus to fuse with the with the endooal membrane and release the genome into the cell. It [clears throat] was thought going back to what I was talking about in terms of uh the lack of energy. It was talk thought that somehow protons must be important in providing the energy to cause this uh overcoming the activation barrier and indeed uh people were wondering what sort of how it is that the protein took advantage of protons to cause this uh transition. And so the missing link, it was thought because there wasn't a need for ATP hydraysis, the missing link was somehow going to be protons. But what Chavei showed was that actually you can get virus membrane fusion to occur at neutral pH without introducing protons if you destabilize the protein. And so here is a membrane fusion event. What you can see is that [clears throat] assay for membrane fusion if you lower the pH to pH6 or pH5.8 you get membrane fusion. You can look at the biochemistry of the confirmational change. But [clears throat] if you go at neutral pH and instead of lowering the pH if now you add a denature it like ura you [clears throat] can also cause membrane fusion. And she showed that if at neutral pH you increase the temperature to destabilize things, you can also cause membrane fusion. And as judged by the biochemical assays that we could do at the time, there was no difference that we could detect between the low pH, the high temperature or the high ura induced transitions. And so what this suggested or indicated was that actually you can trigger this membrane fusion event as in any one of in any way that destabilizes the protein. So if you destabilize the protein, you actually kick it over to the fusogenic state, which implied that the native state of the protein was not the thermodynamically most stable state, but rather when you uh denate when you destabilize the protein, you allowed a transition to a thermodynamically more stable state. And this of course was blasphemy in biochemistry to say that uh the native state of a protein is not the thermodynamically most stable state. I spent over a dozen years at MIT teaching that the amino acid sequence of a protein determines it threedimensional structure which is the thermodynamically most stable state of the protein. And yet we were saying exactly the opposite here. So how does this occur? Well, the answer again came from uh Don Wy who solved the well first I should say hemoglutin is made as a precursor as a single chain precursor and that's called HA0. It then gets proteolytically cleaved probably by furine to produce two subunits HA1 and HA2 that are dulfide bonded. When Don Wy solved the structure of HA0, what he saw that it was exactly the same as the structure he determined before of HA1 and HA2 except in the region where the cleavage occurs. And in HA0, there was an exposed loop. And in HA1, [clears throat] HA2, that loop gets cleaved and part of this protein tucks into the center of the protein, but the rest of the protein remains exactly the same. And so what's going on here we think is that HA0 follows Anfinson's rules. It folds to the thermodynamically most stable state of the protein. Okay? So here's the unfolded state. HA0 folds to the thermodynamically most stable state. But then it's proteolytically cleaved. And after it's proteolytically cleaved, it doesn't now unfold and resample all the confirmational space. It's stuck in this confirmation. And the only thing that happens is the part that gets cleaved tucks into the center of the protein. And it [clears throat] turns out that that traps this protein in a metastable state. And now if you destabilize that state, it goes to the thermodynamically most stable state. And the energy that's released by the transition from the native state to the fusogenic state is used to overcome the activation barrier for membrane fusion. So it's a remarkable process and as I said convergent evolution has come up with this process through different ways of setting it up but it's a remarkable process of storing energy in a protein and using it to overcome an activation energy an activation barrier in this case membrane fusion. Subsequently uh several groups studied other viruses and showed using this rigorous definition of thermoability of metastability namely that you can trigger membrane fusion using destabilization by a bunch of different means uh and showed that in fact uh these membrane fusion proteins are metastable that is they're not folded into the thermodynamic most stable state and this is now commonly accepted in the field. So as I said the energy that's released here is used somehow to cause membrane fusion and indeed this is now a general mechanism for what are called class one membrane fusion proteins which include several of the pro proteins that I've been talking about. But why is this important for vaccine development? Well, when this because this this state is is uh metastable, it [snorts] will spontaneously at some rate transition to the more stable state and you can make that transition faster by adding denaturance or lowering the pH. In the presence of a target membrane, that transition will lead to membrane fusion. But in the absence of a target membrane, if you just have the uh protein sitting there on the viral membrane, it will if there's no membrane to fuse with, it'll lead to inactivation. And so in fact, these proteins that are metastable uh sitting on the surface of the virus will uh come over and become inactivated or if they're sitting on a vaccine by themselves, they will become inactivated. And indeed in this experiment that I showed you earlier when Chave first pre-treated the virus with acid it was now irreversibly uh denatured and or irreversibly in its uh postfusion state and it would not cause membrane fusion. Similarly when she pre-treated with ura it would not cause fusion. So it [clears throat] get it actually becomes inactivated. Well, back after we proposed this model, Judith White at the University of Virginia actually decided to test the model by introducing proline mutations in hemoglutin in places where she predicted based on the structure that prolines could be accommodate accommodated. And the reason she did that is proline of course is a very strong helix breaker. And she reasoned that if she could put the prolings in the native state, which she was able to do, then if the spring-loaded model was uh correct, this should not cause membrane fusion. And indeed with these two proline mutations, she was able to get proper formation of the trimer, proper cleavage, proper function, proper antibbody binding. But when she lowered the pH, she did not get membrane fusion or um uh viral infectivity. So basically providing a functional uh confirmation of the spring-loaded model. Well, this trick of introducing prolines into the native state of the fusion protein turned out to be very important uh in in the future. And the first example of that was with the HIV uh envelope protein. For many years, people tried to solve the crystal structure of the HIV envelope protein because it was such an important structure to try and get and everybody failed. The the envelope was just too unstable to actually get a structure. But John Moore taking a lesson from uh Judy was able to find a proline residue that when he inserted it and when he stabilized the the envelope protein with a dulfide bond leading to what he called envelope SOSIP uh with this isolucin the proline mutation which [clears throat] would be in the middle of the uh three stranded coiled coil he was able to stabilize this protein sufficiently that actually then it was the structure was able to be solved and indeed This SOSIP construct is the basis still today of efforts to try and make an HIV vaccine. And over the years, starting with Judy White's uh example in in influenza going to John Moore's and HIV envelope, there have now been many different proteins that have undergone uh protein engineering to introduce these such proline residues. And [clears throat] indeed uh Jason Mlelen, Andrew Ward and Barney Graham did this with corona viruses and [clears throat] they were studying MS and SARS which are two other corona viruses. And they found and identified two residues which they could mutate the prolings which would stabilize uh these proteins and prevent them from flipping over into the postfusion state. and actually were able to make vaccine candidates with these uh proline mutations. Then when the pandemic hit with SARS KV2, the sequence was released in in January of 2020 and by May of 2020, uh both Mlelen and David Weezer's group had the crym structures of the protein for SARS Kovv2, the spike protein, because they knew where to put the two proline residues. based on this. So they're immed being they were able to immediately move forward and get the structures and it turns out that these two proline substitutions were also incorporated into the successful COVID vaccines namely the biioentech fiser and the madna mRNA vaccines which incorporated these same two proline mutations. the uh vaccine candidates like the Oxford Astroenica or the uh Russian or or Chinese vaccines that did not incorporate this these two prolines ended up not being successful and we now know at least for one of them that that's because uh of the transition to the postfusion complex. So the stabilizing mutations lock the protein into the prefusion confirmation which is of course the structure that you want to have antibodies made against if you wanted to create a good vaccine. You don't want antibodies made against the postfusion state. And so by stabilizing this prefusion confirmation, you make the native protein uh the uh target of your uh im of the vaccine uh response and you therefore enhance immunogenicity and stability and this was critical for SARS Kovv2. So this principle of stabilizing first of all the fact that these proteins are metastable and then that now induces the important principle of stabilizing that metastable state if you want to make a vaccine has become incorporated into modern vaccine development and design. And there are many different uh ways of doing this now going beyond um introducing proline residues that people are are using many people are using to make vaccines and indeed the recent success of the respiratory sensitial virus vaccine or uh is was due in very large part to stabilization and tomorrow in tomorrow's talk [clears throat] I'll talk about using protein language models to uh understand and guide protein evolution and indeed Indeed, we and others are using such protein language model approaches to also stabilize vaccine candidates. I'd like to now turn to the imuno focusing problem. As I said, there are some vaccines that have been very difficult or impossible to make and um that is in spite of the fact that we know that we have monoconal antibodies that prevent the disease. So there are several diseases in humans and many more diseases in animals or animal models where we know that if we put a monoconal antibbody on board into that person or that animal we can prevent the infectious disease. And we know exactly where that monoconal antibbody binds the virus or infectious agent. And so we know exactly what the epitope is of the antibbody that is protective. And so the problem then becomes very easy to state if you for example in HIV we have what are called broadly neutralizing antibodies that people have painstakingly isolated. These are very rare antibodies from HIV infected individuals but they've been isolated. Beautiful cryomm or crystal structures have been determined. We know exactly where the antibbody binds. So if you want to make a vaccine, the problem statement is very simple. Make a vaccine that elicits antibodies that bind exactly to this epitope. Okay? If you if you make antibodies that bind this epitope, we know that it protects uh people or humans. and and indeed in the case of HIV uh this is an experiment done in humans and with a single monoconal antibbody called VRC1 and it it it shows that you get protection in humans for the viruses that are effectively inhibited by this antibbody and yet we can't do that when we try to make a vaccine that elicits antibodies that bind to a specific epitope very often we end up with antibodies that bind elsewhere and it's what I call distractoes. The you know the antigen has something else that says look at me look at me bind here and so the immune system goes and it binds there and it doesn't bind where you want. There are situations where you have an epitope that's highly conserved among different strains of a virus and you have a monoconal antibbody that binds there and people have shown in animal models that it protects against all these different strains. And so it's very clear what you want to do if you want to make a vaccine. and make a vaccine that elicits antibodies against this highly conserved antibbody and yet we can't do that and so that's what and in the case of HIV we can't do that despite 40 years of extensive very hard work on the part of many people and so it's what it's what I call the imunoc focusing problem that is if we have a particular epitope shown here in red that we want to target uh we need to get figure out how do we get the immune response to respond just to that red far. And so people have tried different strategies. For example, uh using different um strains of the protein from different viruses that have the same red epitope to try and get the immune response. They've tried putting those on a display of of what's called a mosaic display to do this, but now in one candidate, they've tried cutting off the distraes, but that requires protein design and engineering, which is tricky to do. They've tried putting the domain that you want onto a different scaffold, but that introduces new epitopes with which includes a scaffold or they've or they've tried using epitope masking to hide try and hide areas that you don't want to get antibodies to. And uh much of my many people in my group are in fact trying to develop methods to do this to actually do imunocusing. It's basically a protein engineering problem. How do you actually engineer a protein so that parts that you don't want to be immunogenic are not immunogenic and other parts are? And I won't go through uh those different efforts but instead focus uh today on one vignette and that has to do with Ebola. So um [clears throat] Ebola virus is a virus that causes hemorrhagic fevers that uh have a very high fatality rate. So far the outbreaks have all occurred in in Africa and they occur on a regular basis unfortunately. [gasps] Uh and the predominant species that has caused um outbreaks over the years is called Ebola virus uh and is actually referred to as Ebola virus zire because that was the original isolate. Um and that's called and that's the uh orange red dots here. The size of the dots indicate the number of cases that have occurred. There's another virus related in the Ebola virus family called Sudan virus [clears throat] and it's shown in blue dots and it's the second biggest cause of major outbreaks and there's a third virus which has caused uh just a couple of outbreaks called bundi bujo virus but that's the virus that actually is responsible for the ongoing outbreak that's occurring right now uh in in DRC in Africa is the bundi [clears throat] bujo it's a very rare form of the virus and this outbreak is bad. The worst outbreak of Ebola virus was in 20 excuse me 2014. It lasted two years and caused uh over 11,000 deaths with a 39% fatality rate. So this is a bad bad virus to get infected by. up until uh this year, the second biggest uh outbreak was of Ebola virus year in 2018, which also lasted two years and had a 66% fatality rate. But now with the Bundi Bujo outbreak that's occurred, you can see that in the first 120 days, it's actually become um the second biggest outbreak already. The largest outbreak was here with 28,000 cases. We're already up to almost 7,000 cases with the current outbreak and it's continuing uh pretty much unabated right now. And the issue um again is both scientific and non-scientific. Uh the scientific issue is that the the vaccines that we have right now all protect against Ebove the Zire form and uh that's because that's been the major problem so far. That's where the focus has been. But we don't have any vaccines against BDBV. And uh what's going on right now is that the WHO is testing the existing vaccines against Ebon to see whether they help with BDBV. We and others are working on trying to come up with a BDBV uh uh vaccine. Uh but in the meantime, the virus is continuing to spread. And again the issue there's there's a lot of warfare going on and also in these countries where this these out where this in the regions where this outbreak is occurring actually getting Ebola virus is about number three on the list of concerns. You know uh number one is just basic uh survival and getting food to eat. Number two is getting shot by the gorilla warfare and then u you have um have uh the spread of the virus. So again, they're both scientific and non-scientific issues that that we need to deal with. Well, the Ebola virus has like uh the virus I've been talking about a single glyoprotein that's prominent on its surface and that's called glyoprotein or GP. And as I've said already, there's a family a genus of Ebola viruses. Ebola virus zire uh Sudan and budnau are the ones that have caused outbreaks. But these others which have infected animals and actually Taiforest has infected humans once are all in this family. And what we'd like to do is figure out how to make a vaccine that covers all of these viruses so that you actually have a panbola virus. And so our efforts in this regard were uh really jumpstarted by somebody that some of you may know um who uh when he was a postocck in my lab uh really started the tremendous work that um on Ebola virus and he really focused on two particular uh issues. The first is what I've been talking about and that is that um because these different species uh have residues that are different, the current vaccines are species specific. And so we would like to do is solve that problem. But the other thing is that the current vaccines are all viralbased vectors. So they require storage at minus 80° C. And that of course is not very convenient when you're running around Africa trying to vaccinate uh people. And so uh also to try to avoid the low temperature storage issue. And what Duo did was to take two approaches to use the uh use proteinbased nanop particles u based on feritin which we knew were thermo to try and put these onto those ther uh presentations and then to use um a shielding method namely glyosillation to immunofocus the antibbody response away from the species specific areas shown here in red and toward the conserved regions of GP. >> [clears throat] >> So on the on the first issue of the feritin based nanoparticles um many years ago Keno and Gary Nabel at the NIH showed that feritin which spontaneously forms a 24 uh particle used to transport iron in bacteria it self assembles to form this nano particle made of 24 subunits And there's this beautiful three-fold axis of symmetry which actually uh geometrically aligns well with the trimer on in this case hemoglutin that three-fold axis and they showed that you could fuse these two together as a genetic construct and make a particle which displayed hemoglutin on the surface. And the advantage of using these particles is that dendritic cells will grab particles and transport them much more effectively than they will small proteins. And in addition, particles will cross-link B cell receptors on the surface of B cells leading to enhanced uh imunogenicity as compared to the isolated antigens. And so multi multime display of antigens promotes uptake by the dendritic cells and it increases bell receptor clustering and therefore signal activation. And what do Duo showed was that when he put the glyoprotein onto the nanop particles and then studied them in mice comparing the trimer to the feritin nanoparticle when he looked at binding antibodies there were better binding antibodies with the feritin nanoparticle but it wasn't a huge difference but when he looked at neutralizing activity of the anti-era there was a big difference with the glyop with the nanoparticle eliciting much better neutralizing antibodies against the virus than the isolated trimer. So basically he had a system to uh to work with uh as moving forward. So then for amunof focusing what do was to install glycans into the protein and in particular into those regions which were highly variable and the reason the way in which he did this was using uh genetic expression and the and that's because the motif that encodes for glycosillation in malian in malian cells is a sparaging any amino acid except for proline and then searing 43 amine. So this tri this three residue motif is what is recognized by the glycosillation machinery to add glycans. And so he [clears throat] screened a bunch of regions where there was variability in the protein to identify those places where he could insert genetically this motif and get the glycan added. And what he showed was that that strategy was compatible with the multivalent display on feritin. And he could add two, three or four glycans. And it turns out that the when he added these three glycans to the region of the protein which is highly variable called the glycen cap, he could get um a good stable nanop particle. [clears throat] And he went on to use this uh species in mice. And in green you can see that with his GP plus3 construct he's able to get neutralizing titers against both Eboth BDBV which we've been talking about for the current outbreak and Sudan. So he's been able to cover the [clears throat] three different species that actually caused outbreaks in humans. And finally what he showed was that uh when he takes his glycosillated GP feritin construct GP plus3 and he stores it at 37° C for up to two weeks. There's no change in the thermal melting curve of the protein. So it's a stable protein that could be transmitted um at certainly refrigerated. Uh in Africa things get pretty warm. Uh sometimes much higher than 37 but even 37 for two weeks is stable. So that's where um things stood as I said um duo started with Ebola virus and he was able to show that he got cross protection against Sudan virus and buda virus at least in mice. As I said already there are vaccines already against Ebola virus. There are no vaccines that are specific for Sudan virus which I which as I said is the second leading cause of major outbreaks of Ebola virus. And so uh what Rebecca Costello went on to do was to ask whether or not we could make a Sudan virus vaccine since that is an unmet need unmet medical need. So could we make a Sudan virus vaccine but again take the strategy that Duo used to try and make it not just a Sudan virus vaccine but something that would protect against um other viruses as well and I won't go through the process that I just [clears throat] went through with Duo for how she did this but again she used a feritin nanoparticle presentation of the glyoprotein again she focused in on the glycine cap which is where the highly variable regions are here's the conservation. What she's trying to do is focus the immune response on the conserved regions. I'm just showing one of the three protoimemer highlighted in colors here. It's a trimer like the other proteins. And she ended up uh with this hyperglycillated version 2.2 that showed the best properties. And I won't walk through how she characterized it, but she was able to get this this protein on a nano particle. And when she uh immunized um mice with this with this with the wild type Sudan versus the hyper glycosillated 2.2 2 version and she looked at neutralizing titers against now all of the uh different uh Ebola viruses. Here what you see is that she gets good neutralizing titers in both cases with Sudan but then with Eb she gets much better neutralizing titers with her construct with Bundy BJO she's getting good titers and indeed she's able to get uh neutralizing titers against all uh six members of the Ebola virus family. So at least in mice and we're now going on to test it in um guinea pigs uh the uh her Sudan hypoglycosillated nano particle is a pan Ebola virus vaccine. And so um it just shows the basic principles of taking advantage of increasing immunogicity and uh using an imun focusing approach to hopefully get something which is a candidate that will be across all Ebola viruses. And with that, I'd just like to close by saying I have an amazing group of people. I've only talked about the work of some of them. Uh they're listed here. And um I'm fortunate in this current environment to have funding in particular from uh SEPY, the Coalition for Epidemic Preparedness, uh and the Gates Foundation to help support this work. Uh and with that, I'll be happy to answer any questions. Thank you. Hi, that was a really interesting talk. I was wondering if since you were showing the conserved region on those proteins is often on the side kind of where the protein is on the feritin particle. Uh do you ever have issues with like immune cells not being able to access an antibbody that's bound there? >> Yeah, great question. Um [clears throat] the answer is yes. And sometimes it works to our advantage and sometimes it works against us. If if the epitope is located on the side or even lower on the side, um we can get antibodies to bind there, but uh we but those regions tend to be uh not very immunogenic. They tend to be what I call imuno recessive. And we think that that's because even though an antibbody could get in there and you measure the affinity for the binding, it's lower by two or three fold, but it can still get in. But there is some steric hindrance. But a B cell where the B cell receptor is on the B cell membrane and in order to stimulate the immune response it needs to get into that epitope. It of course is much much much larger and so we anticipate that the accessibility there is uh going to be a problem and that may be related to why it is that they're much less immunogenic. Um, another project that Duo did which he's continuing here uh at Wisconsin is to actually figure out how to take such uh epitopes or such constructs where the epitopes down in the bottom and not very exposed and turn it upside down and he did that and it worked beautifully and he's he's continuing to do that work here. Uh so accessibility is a key issue in what in what we need to do here, right? Yeah, absolutely. Hi, that was a great talk. Um, I'm curious about like so the rise of like mRNA vaccines and is that feasible for something like Ebola? >> Yeah, great question. Um, I'll preface this by saying that mRNA vaccines have have uh had an amazing impact on uh the world. And um a major reason that we're all here in this room is thanks to mRNA vaccines. um despite what you may hear from other people. Uh and um the the advantage of mRNA vaccines is that they are fast to produce and um you can therefore change the sequence very quickly which is very hard to do with a proteinbased vaccine. Um the and now we know that the mRNA vaccine has been tested in literally hundreds of millions of people. Anytime you take a foreign substance and you put it into your body, there could be side effects and those side effects can be different for different people. So there is a risk of taking a foreign substance and injecting into your body. I think the first thing that needs to be acknowledged in this discussion is there is risk. That risk needs to be compared to the risk of the disease. Okay? And that's where I think this whole argument about mRNA vaccines has gotten completely out of whack, right? Is that the risk of the mRNA vaccine which is there and it's real is is much much much smaller than the risk of the disease. In the case of Ebola where you're talking about fatality rates of 40 to 60% um this shouldn't even be a discussion and indeed um Madna is working on an mRNA vaccine right now for BDBV uh we're also working on an mRNA construct which uh we hope will have broader um broader neutralizing activity not just against BDBV but also against other Ebola viruses. The issue with mRNA vaccines uh is that they're expensive and they require cold storage. And so it's not a very practical solution. But in the case of an outbreak like this uh and if you have people that are willing to pay, which in this case uh is going to depend in large part on countries other than the United States, um then they're they're a great solution, right? Yeah. >> Yes. Um, is this I had another question about the feritin nanoparticle strategy. I was curious basically how densely decorated these nanoparticles need to be in order to get a good response because I I imagine the feritin itself could act as another sort of distra for the vaccine. >> Great question. So, we worry about that a lot. We've shown um actually in animals and um in a vaccine a COVID vaccine candidate actually that was taken into phase one clinical trials by a biotech company that the responses to feritin are uh low and you can actually boost with these vaccines and not get the antifarotin anti response to interfere. But that's a definitely a concern in terms of the first part of your question, the density, you know, [clears throat] certainly in terms of um the B cell receptor cross-linking, the spacing between the antigens is going to be important. And so it turns out that the spacing that the feritin nanoparticles give us is actually empirically very good for cross-linking bell receptors. But uh if you were to use different spacing or different nanop particles, you could get very different results. Pamela Bjorkman at Caltech has done some very nice studies using uh spacers made out of actually double strand DNA to start to investigate this problem. But this is something that definitely is uh an important consideration, right? [clears throat] The geometry in the space and the density, right? Also on the same topic um instead of using these uh feritin nanop particles and you know trying to engineer your epitope to you know target your immune response to a particular location. Do you think an alternative strategy that might work would be to um say evolve antigenic peptides uh because you know like what the epitope should look like and you have antibodies that could bind your target. Would that be another route to a vaccine? >> Yes, I think that that that is definitely a good uh path and people are people are trying to do that. Uh as you say in those cases where you have antibodies that you know are effective and you can use evolution to move things forward. Uh it's definitely something I think that's that's worth uh worth pursuing and people are doing that. Yeah. Absolutely. >> Yes. >> Uh thank you very much for that uh uh interesting talk. I just wanted to uh get some ideas about the Ebola in particular, which is really a very big problem in in Africa right now. [clears throat] And now that you have sort of begun to develop uh uh the vaccines against the four uh Ebola species and you have tested them in in mice [snorts] and you mentioned that you also going into the guinea pigs and I'm just curious because of the seriousness of this problem how long it is likely to take [clears throat] based on your experience. right? >> Uh to get out of the the guinea pigs and then into the humans just just to be able to help the global. >> Thank you for that question. Um there there are right now um there the current Ebola virus vaccine that works against the other species Eboth is being tested in Africa in the outbreak to see whether or not it offers any protection against Bundi BJO. So, we don't know the answer, but we'll we'll find out. And we'll find that out, I hope, uh, in the next few months. Uh, and and if it does, that'll be that'll be great. As I said, there there are two other vaccines that are currently moving forward very quickly. One is the Oxford uh based vaccine, which was used in COVID. It wasn't terribly effective, but it was used in CO and has a good um manufacturing track record. It's being produced now with the Bundi BJO protein and um it has entered phase one clinical testing. So they're starting to test humans with for safety. Similarly, Madna is making an mRNA vaccine similar to what they did with COVID using the bundjo uh construct and that also has started phase one clinical trials for safety in humans. So, uh, if we're lucky, we'll have the phase one results or the comm we community will have the phase one results for those two vaccines in the January to February time frame. And if that's the case and if they look like they're safe and if they have a good immunity immunogenicity, then the WHO plans to test them in the current outbreak. others including us are a few months behind but with something that in our case with something that we hope will be more effective in case these others are not effective. Um and um but as I said, we're we're a few months behind and every with this outbreak every day matters. So we're we're trying very hard to move quickly. But as I said, the other two outbreaks that have occurred lasted two years. And so we think actually that there's um still an important need to try and keep going until we have a good vaccine that can be deployed. I should say that that's the scientific part of the problem. The non-scientific part of this problem are very very substantial and uh that of course is um something that I I have no expertise in. Right. Yeah. But they're they're very real. Yeah. >> Yes. >> A wonderful talk. I had a quick question about the antibbody binding of your panola virus. I was wondering if you looked at specifically where these neutralizing antibodies are binding to uh the glyoprotein. Do they bind to the same place? Do they have the same footprint? Are there differences? >> Yeah, great question. I would love to know the answer to that. And uh there are there are more direct and less direct ways of trying to answer that and uh we're we're pursuing both. Uh one of them in collaboration with somebody here. [laughter] Yes. And and that's a very important question to answer from a scientific point of view in terms of what what are we actually doing that's leading to success so that we can recapitulate that. Right. Absolutely. Right. >> Okay. Thank you very much.