...Changing The World

Project Description

MMRF » 2018 Vaccine Development Conference #02: The Human Immunome [2018-06-27. James E. Crowe, Jr. HVP/USC]

2018 Vaccine Development Conference Session #02: Complexity of The Human Immunome [2018-06-27. James E. Crowe, Jr., Marie-Paule Kieny. HVP/USC]

2018 Vaccine Development Conference – Session #02: Complexity of The Human Immunome — James E. Crowe, Jr., Marie-Paule Kieny.


Organized in conjunction with the Human Vaccines Project, the 1st Annual Conference on the Future of Vaccine Development was a one day event which took place at the USC Michelson for Convergent Bioscience on June 27, 2018.

By bringing together some of the world’s leading scientists in the fields of immunology, genomics, bioinformatics, and bioengineering, the Future of Vaccine Development annual conference aims to explore how the convergence of new technologies across disciplines is impacting the future of vaccine development. The conference will also honor the three inaugural winners of the Michelson Prizes for Human Immunology and Vaccine Research, both via their respective presentations and the remittance of their prizes during the Awards dinner ceremony following the conference itself.


  • James E. Crowe, Jr., MD, Director, Vanderbilt Vaccine Center, Professor of Pediatrics, Professor of Pathology, Microbiology and Immunology, Ann Scott Carell Chair.
  • Marie-Paule Kieny, PhD, Director of Research @ Inserm (Institut national de la santé et de la recherche médicale), former Assistant Director-General for Health Systems and Innovation at the World Health Organization (WHO).

Marie-Paule Kieny: So good morning everybody. I’m Marie-Paule Kieny. I’m a director of research at the French Inserm. I’m also a retired assistant director general at the World Health Organization for many years, so this is where I’ve known quite a number of you. And I am a number of other things, but I do, since I left WHO, I am a member of the board of Human Vaccine Project. So it is on this capacity that I will be chairing the first session of this interesting meeting. I really look forward to hearing what are the latest developments, and what are the promises for the vaccines of tomorrow.

So I would like to call on the first speaker for this session, so this is James Crowe. James is the director of the Vanderbilt Vaccine Center, and the Ann Scott Carell Professor of Pediatrics Pathology Microbiology and Immunology. Of course he has a long bio. I will make it a little bit shorter just to say that as we know, and this is why he is here, but Jim’s laboratory studies, the human immune response to infection to a wide range of major human pathogen, including many emerging infection.

He’s an elected member of the National Academy of Medicine, and the National Academy of Inventors, and an elected follow of many other organizations. So his laboratory was recognized in 2018, this year, as the best academic research team at the 11th Annual Vaccine Industry Excellence Awards. Congratulations.

Jim has been, of course, a recipient of many awards and prizes among which I would like to cite only the most recent ones, so this was last year, The Korsmeyer Award, ASCI and the Samuel Rosenthal Foundation Prize in Academic Pediatrics. And finally, and importantly, Jim is a principle investigator, is one of the major project of the Human Vaccine Project. Jim, over to you.

James E. Crowe Jr.: Great. Well thank you very much, and happy to tell you what our team has been doing. I love the idea of convergence, and our work, none of this would have been possible. I’m going to show you stringing together genomics, synthetic genomics, structural biology, informatics. I personally don’t know how to do any of this stuff, but we have a large team that’s put together all of these disciplines to try to understand how humans kill viruses with their immune system. The principle mechanism of immunity against virus reinfection is in antibodies, and that’s what we focus on.

I’m a little fuzzy on the green one. Oh well, okay. So we have 10 or 12 ways that we make human monoclonal antibodies from a single cell on an immune individual. This is the typical scenario where we take peripheral blood mononuclear cells by a blood draw. We expand those cells in plates. We can put out 300 plates at a time, 384-well plates and search for a needle in a haystack. We have screening assays such as killing a virus, so a neutralization assay, that would be typical. We do these screens at large scale. We’ve converted them to liquid handlers. We can do it in containment.

At any one time we’re expanding over 1,000 human monocle antibodies in the lab. These are the sorts of things that we have already done in terms of making 50 to 2,500 human monoclonals to these targets. At any one time we’re doing screening campaigns. This is what we’re doing right now. And then part of my job is just going around the world finding blood samples from people who have survived exotic things. Hopefully some things that you haven’t even heard of and don’t know how to pronounce their name because we’re trying to stay ahead of emerging infections before they happen. Get them done ahead of time, not scramble afterwards. That’s the idea.

And in terms of the kind of drug development principles, or philosophy, the idea is to find antibodies that have high potency. So if you have a monoclonal antibody that has a 50 percent maximal inhibitory concentration of one microgram per milliliter, you’ll achieve a certain level of protection, and it’ll last a certain period of time. But if you make an antibody that’s 1,000 fold better, effectively you have longer half-life. It may not be 1,000 fold, but the functional half-life will be much better. So we’re always trying to find very rare ultra-potent antibodies.

And the other idea when you’re looking at viruses, viruses have a lot of diversity. And the idea is if we can find individual monoclonals that would cover not just a single virus, but the entire class of viruses, we could cover multiple viruses with one drug, which simplifies the drug process. Also if a new virus from that family emerges, such as Zika which is a flavivirus, if we have flavivirus antibodies then we’re ready for Zika ahead of time, so this idea of breadth.

And I’m just showing you antigen antibody complexes by crystallography, single particle EM like this, or cryoEM on particles. And these are antibodies we’ve done in the last probably two or three years. Some of them are even unpublished. So all classes of viruses, flu, including those of us who are middle aged, or H2 immune, or from _____, or survivors of H5, H7, the pneumonia viruses, bad viruses like Ebola, small pox, et cetera.

So this has become pretty routine, and I think the idea here is, in terms of understanding the human immune system, each of these clones single molecules from a single cell, is a snapshot of the entire immune system. So we’re just seeing one little snippet. They could be used as a drug, but it has tools for understanding immunity. It’s just a snapshot. We know that people don’t make one clone. They make thousands, maybe millions, and that’s sort of where the project leads.

So I’m just going to spin you through. These are screens that we do this with graduate students in the lab. They come in first year out of college and they do this. So this is a screen one of the students did with eastern equine encephalitis virus, which is an alphavirus. So there’s alphaviruses all over the world that go to your brain and cause infection. And in a single survivor individual these are neutralization curves. We can completely neutralize the virus with IC50 values that are less than one nanogram per milliliter, so extraordinarily potent from a single individual.

There are alphaviruses that get your joints, so chikungunya, mayaro in South America, Ross River in Australia. There are regional alphaviruses. It would be very complicated to make a regional drug for each arthritis emerging virus. So this is an example, we took an individual from Australia. Again, isolated neutralizing antibodies, extraordinary potency, so IC50s of single nonogram per milliliter. This particular antibody not only binds to Ross River, these are actual structures, cryoEM, Ross River, mayaro for South America, and chikungunya which is all over the Caribbean, and all over the world really. So we have a single antibody that can solve all three.

This is a project we did about a month ago. Well, we did the crystallography a month ago. We took an individual who had inadvertently been inoculated with a veterinary vaccine for Hendra virus. So these are bat viruses, Hendra and Nipah. There’s a Nipah outbreak in India right now. People are dying of Nipah. Hendra goes from bats to horses and kills the horses.

So there’s a horse vaccine, and this veterinarian got inadvertently inoculated and became immune to Hendra. And we flew him into Nashville, made these antibodies. We get antibodies that cross react with both Hendra and Nipah, and again, super potent antibodies. We already have the crystal structure of one of them. Here’s the receptor binding to the Hendra protein, so this would be virus binding to its receptor. An antibody sits and basically mimics a receptor.

So again, a single individual. We have Hendra, Nipah antibody drugs, and then Rift Valley is a bunyavirus, so bunyaviruses are all over the world. This particular one is in East Africa. It’s on people’s lists as a threat to emerge. It’s happening right now in East Africa. And again, very potent neutralizing antibodies. This one is .1 nanogram per milliliter. Now we’re getting a picogram per milliliter drug.

So these snapshots tell you that the human immune system is amazing, and we only need one person to find this stuff. We don’t have to do 10,000 people to find these antibodies. And what about breadth? I mentioned some cross family. This is norovirus, so there are thousands of noroviruses that fall into geno-group one and geno-group two in blue and green. And again, from a single individual we find antibodies that go across the entire geno-group at very high potency. So norovirus is a diarrhea virus. This is a cruise ship virus. It affects military. But we also unexpectedly find antibodies that would get geno-group one and two.

Pan-ebolavirus antibodies that get all of the Ebola virus species. Marburg virus, so this is – we studied a woman who traveled to Uganda, went on safari, went into a cave to look at the really cool fruit bats. There’s 40,000 fruit bats in Queen Elizabeth Park in Python Cave. Looked at the bats, nothing happened. Went home and she almost died, and ultimately was found to be Marburg immune, so we isolated antibodies from her.

This is primate studies where even as late as day five we can get a single antibody and save all the primates otherwise they die. We have the structures of these. BARDA, the US government has invested millions of dollars to escalate this. And in this past year there was an accident in a BSL4 environment. A person stuck himself with a lethal dose of Marburg so we flew in and under compassion we just treated him with this antibody.

Zika happened. We were already working on Zika because of flaviviruses. This is a potent Zika antibody cryoEM. This is a primate study where if we give the antibody before there’s no virus whereas the controls shed over a million PFU per milliliter. And if we wait to use it as a treatment on day two all the virus goes away, and the controls keep shedding. So these things not only work in small animals, but macaques.

So I’m going to transition and say these are snapshots, and the question is how complex are these repertoires. We’re making 50, 100, a couple hundred antibodies. This is starting to give us a window into what infections and vaccines do, and the question is could we use tools to go backwards and say, “Can we engineer vaccines that do this? This is the endpoint. This is what we want. Can we go backwards?”

And so I mentioned genomics. Well, the Human Genome Project resulted in fantastic instrument development, so the sequences are amazing. And I got tricked into running the genomics core in my institution for a while, and I’m running these sequencers that make billions of sequences. I said, “We should use these to do an immunology things.”

So we started sequencing the blood response of people after a vaccination. So this is an example where we gave an individual – sorry, I keep hitting the wrong button. We gave an individual a flu vaccine on this day, and we saw over 10,000 clones fire off by sequencing. And some of these we have many, many copies, so these are the top 10 clones. And if we bleed the person every day for a month we can sort of see them firing off, so we’re getting a glimmer into what the immune system looks like.

And let’s say we take one of these over time and blow it up and say, “What does that look like,” and we get these lineages. So you’re familiar with phylogenies of organisms of great apes to man sort of idea, well this is in the genome, the recombined sequence. This is a monoclonal antibody we got from a person who had been immunized, and this antibody is quite exciting. It binds and inhibits all fluA viruses, and it’s a head antibody.

And so in this person we found sequences that are point mutants of that antibody, so this is essentially the same antibody with minor affinity optimization mutations. And so this is what I’m talking about with the power of genomics. You start being able to see not just a snapshot, but you can see the entire clonal linage firing off.

And I mentioned synthetic genomics. It used to be we could just look at sequences and say, “Wow, that’s pretty amazing. Look how articulated and complex this immune response is.” But in fact, now we can synthesize DNAs in the thousands, so we’ve actually made all these antibodies and we expressed them as proteins robotically, and we can look at the function of them. And so now we can start to say, “What does the swarm of antibodies look like in a person’s body?” Rather than studying a single clone we can look at the whole population.

And so we were sort of dreaming in the beginning of the Human Vaccines Project. These are amazing technologies, and yet Wayne and the other members of the leadership group were saying, as he said in the introduction, “Let’s not just do what we’re already doing here. Let’s think of things that are out of the box and big, and we’ll lay a foundation to change how all vaccines are done.” And we had – I think the initial meeting was in La Jolla for three days, and everybody struggled and said, “Do more of what I do. Just give me the money and I’ll be good.” I think that was the first two days. And then we started really digging deep and saying, “What would change?”

And I personally had lived through the Genome Project, so I started my post doc in 1990. I was at NIH and it was amazing. Collins, and Venture and these people were just pushing, pushing, pushing, and they said, “What if we could sequence the entire genome,” and it was so visionary, exciting. And I lived through that and I remember seeing Craig Venture tell a story about it. It took 10 years to clone my first gene, and then finally I got 10 genes in there, and then finally, “Why don’t we do the whole genome” sort of thing, and that really resonated with me.

And when we started having these sequencing technologies and we were seeing these lineages, I said, “Why don’t we just do the whole immunome? Why don’t we look at all of the B and T cell receptors on the planet?” And if we had that database we could understand the immune system in enormous detail. And so that was the idea. It was a moderately foolish idea, and I’ll show you why, but I think we’ve gotten ourselves a year and a half in that we think it might even be feasible.

So let me show you what we’ve been doing. A genome is really the set of all the genes. And in retrospect, what those two projects were doing was really low hanging fruit. They spent $3 billion 20 years ago and got less than 25,000 genes out of it because the genome is quite small actually. And immunome is the complete set of recombined immune receptor genes, and the scale of it is so much larger because of the combinations.

That’s why when genomes are done of organisms people always ignore these immune systems. They sort of set them to the side and say, “We have the whole genome, but we’re not looking at those.” And the reason is just to remind you they’re variable, diversity and joining genes, and you pull out from this collection. And when you do that you achieve 10 to the fourth combinatorial diversity just to make an antibody heavy chain, and this is the scale of the human genome essentially. But when you do the junctions you can add on or remove nucleotides and change the reading frame and so on. So just when you actually make the recombination you’re already to 10 to the eleventh possibilities.

And then antibody genes, unlike other genes in your body, are allowed to mutate, and in fact, they’re error prone polymerases and lack of correction, so you can get some number like 10 to the eighteenth. And this is where people said, “That’s a stupid thing to do. You can never sequence that much.” But the hypothesis is these are not all made, that we actually probably have a smaller immunome than that.

So we’ve launched into the project to say, “Well how many sequences does a person actually have?” You know, theoretically you could have some enormous number, but do we have them all? What happens over time? Are we stable, or is our immune system just flopping around? And most importantly I think for the Vaccines Project is what do we share?

So if you and I are completely different in our immune system, why do we expect that a single vaccine would be the solution for both of us? But if there are shared elements and we target our vaccines to provoke the parts of the immunome that are in all of us, then those vaccines would work better. They would work in everyone, and we would have more effective, more universal vaccines.

And of course it’s expected that age, race, gender, geography would all affect these things. And then long-term we want to perturb the immune system like a vaccine that I showed you, make a linage go off. Or what about disease states like Crohn’s disease, or rheumatoid arthritis, or multiple sclerosis? Those immunomes are probably disturbed, and could we using our healthy immunome understand and profile those diseases?

So this is what we’re doing, this is at Vanderbilt. We bring in subjects and we do what’s called a leukopheresis. So you have a large needle in one arm, and a large needle in the other. The blood all comes out, it gets filtered in the machine, we put the red blood cells back in and we keep the white blood cells and we collect something like 40 billion white blood cells, which include a lot of T and B cells.

And we have a whole crew of people who spends the day cryopreserving these things, and then we transition to the more molecular biology sequencing team and they go through these vials and they just sequence, sequence, sequence, sequence, sequence until they don’t see a lot more sequences coming back. And we spent about two years optimizing this.

And it’s really interesting if you look at the – this is Illumina brand sequencers. One of the dominant brands that is at least in our corridor. There’s also a Pac Bio and others. But the smaller instruments do 10 million reads. Then there were high seqs that were doing 100 million reads, and that’s where we started the project. And now the Nova seq instruments do a billion reads in a lane, in a flow cell.

And we’ve also been adding barcoding. This field is just continually moving. There’s a lot of methods development. If we look at a person and start saying, “What do their immunomes look like?” It’s actually hard to look at billions of sequences at one time on one slide. So we’re sort of grappling with how do you even explore it. We’ve been talking with tech folks in Silicon Valley about immersive environments where we might be able to explore and so on just because you can’t easily look at this. But this is one representation where you have V genes and J genes and the Ds that connect them, so these are all sort of clones or clonal types within one of our donors.

And so these are the data in which if we’re sequencing, sequencing, sequencing and getting more reads, we ended up with billions of raw reads, and then we have to process those for quality and so on. And in three individuals we sequenced and we ended up somewhere about 300 million clonal types in a person at one time. And then we used mathematical models to sort of finish these curves out. We’ve done that for heavy chain, for kappa light chain, lambda light chain, TCRV beta, TCRV alpha. We did it all on these people.

And the numbers come out something like this. People have about 10 million clonal types if you really filter down and cluster them properly. And in the first three individuals we looked at they share as much as one to six percent between each other, and between the three they share .3 percent. And you could either say, “Well, that’s’ a very small number,” or you could say, “That’s three in a thousand clonal types three of us would share exactly.” And given the number, this is about 30,000 clonal types that these people share. And so this is what I’m suggesting, that we need to work with this part of the immunome to do universal vaccine design.

And one of the questions was questions was, “If you just sequenced would you randomly have overlapping?” So we did very sophisticated supercomputing modeling to make synthetic immunomes, and pull out three of them and say, “If I just randomly picked from pots would it happen,” and the sharing is predicted to be very, very low, so the amount of sharing we’re seeing experimentally is very unexpected and high, and super interesting.

So why is the repertoire shared? I mean why do we have the same things, and we think there’s two principle reasons. One is there’s some regulation things that are going on throughout life, so you don’t have to vaccinate or infect somebody to cause a sharing. Sharing occurs without that. I’ll show you why we think that. But then also once you’re infected with something like CMV, or EBV, and you have that your whole life, most of us who are infected with those at some point we’re all going to share some clonal types for those organisms.

So here’s why I think there’s independent of antigen. If we take our 30,000, or 29,000 shared from our first three donors, and then we do the same with cord blood, these are healthy term pregnancies, and then we start overlapping that collection with one cord blood, two, three, we see clonal types that are present in all the cord bloods and in all the adults. So you’re borne with clones, and you have them into your adult life.

And we even went to the other end and we said, “Well, let’s look at people who are over 100 years of age,” so we’re currently sequencing some of these, and we do our fancy sequencing and analysis. And, in fact, we have clonal types, and these are just the CDR3 sequences that are in all individuals. They’re in cord blood, they’re in all the adults, and they’re still there when you’re 100 years of age. So your immune system is regulating some of these clonal types, and it’s likely independent of antigen exposure, but there also is antigen exposure. You do get infected and vaccinated and exposed to things. And here’s the evidence for why I think that shapes your repertoire.

So if we look at the top 100 shared clonal types, and the number of variance, or point mutants, we see that a lot of these shared ones have large lineage sizes. So these are point mutants so that single clonal type, so all these are individual clones off that clonal type. So almost certainly this is being driven by repeated infections, or chronic infections. My time is up. You going to bring me back?

Male: It’s hard to regulate the time.

James E. Crowe, Jr.: Yeah. So to summarize, we’ve done – for people who are T cell fans we’ve done even more T cell sequencing. And the T cell in these three individuals, the T cell repertoire is even more shared. Eleven percent identical clonal types between individuals and their T cells. We’ve looked at CD4s and CD8s.

One of the questions is, “If you look a year later do these people look the same?” We’ve done that, large scale blood draws. And the first individual has at least 20 percent of the same clonal types a year later. This person six and fourteen. So there’s some dynamic features of the immune system, but there’s also some shared stuff.

So in summary, the human immunome, you know, is it possible to sequence it and to have it as a database and use it? And we’re finding that it’s much smaller than expected. People have about 10 million clonal types, therefore I believe, and we believe as a team, that it’s knowable, and that this project is doable with enough individuals. It’s highly shared in both the B cell and T cell repertoires. And there’s scientific evidence of immune regulation, which we need to understand what is the mechanism of that. And there’s antigenic exposures driving this stuff, and it’s moderately stable over time, meaning that if you vaccinated and you got that repertoire stimulated, that would persist and even be boostable.

So this is a picture of our group at Vanderbilt. Very diverse group, and some of the principle individuals who’ve done the work. But we’ve done it in context with the Human Vaccines Project, Glen and Ted run it. Bob Sinkovits did a lot of the simulations and supercomputing with us. We’ve worked very closely with a lot of the partners in HVP and vendors. So thank you.

Marie-Paule Kieny: Thanks a lot. This was really very good and very inspirational. And I’m happy to see that actually something that looks so unfeasible at the beginning ends up being feasible, so good to have bet on that. So I would like to thank the one who’s – the person who’s handling the slides. Always gives a sort of warning for the next speaker that he can’t exceed his time. So we have a time for a few questions. Yes, Stanly?

Male: Yeah, Jim, I want to go back to your demonstration of monoclonal antibodies that cross species, and might be useful for classes of viruses. Let’s take the Ebola group example you showed us. So is it the antibodies to Marburg, which normally in the infected individual, would not prevent let’s say _____. What is the target that you’re talking about? Is it a epitope on the glycoprotein that is not normally revealed during infection?

James E. Crowe, Jr.: The conserved epitopes, where are they?

Male: Yeah.

James E. Crowe, Jr.: I don’t know if I could even go back. I could show you but maybe I’ll just tell you with my body. So the glycoprotein has a head domain, and then there’s a stem domain too. It’s sort of like a lollypop, and the conserved elements are underneath the head. So the antibodies have to go in, down, and kind of go upwards, and it’s a kind of an occult site. So most people aren’t making a lot of those because to stimulate them it has to be a B cell receptor on a cell. It’s hard for a cell to go down and up, but if you get them then they’re really good antibodies.

The other difficulty in Ebola is a big glycan cap on it in a mucin like domain, so there’s a lot of obstruction of that. If you cleave that stuff off the epitope becomes much more accessible, and we can’t cleave it on the body, so what we found is a second antibody that binds to the glycan cap, and it causes a structural perturbation and it moves the glycans out of the way, and then we get 1,000 fold enhancement of neutralization to the occult site.

So we’re using – we just did a primate study where that works for Ebola, for Pan-Ebola, but you have to use both because the first one moves the sugars out of the way, and the second one comes up to the hidden epitope. So you’re absolutely right, we have to find rare antibodies to do these conserved epitopes because the viruses tend to sort of hide them in places.

Marie-Paule Kieny: Go ahead.

Male: Regarding the sharing, along with the plasma cells versus other memory cells, and what’s shared between the two compartments.

James E. Crowe, Jr.: Yeah, so the antibodies that we study in vaccine trials typically are looked at in serum or plasma, and those are secreted by long plamsa cells in the bone marrow. The studies that I’m showing you are memory B cells, or plasma cells in the periphery. And the question is, “Are those repertoires the same or are they different? So the stuff that goes in your bone marrow that really gives you serum antibodies? If it’s a different repertoire we’re studying the wrong stuff. Maybe not for drug development, but for understanding vaccines.

So we and others are – I didn’t show you, when we do the leukopheresis we’re also collecting plasma or in serum, and we’ve been capturing – like the guy I showed you actually I’m not – he’s a public figure. He’s an Ebola survivor, so we’ve taken his plasma, purified it on GP, Ebola GP on a column, cut the FAB and we sequenced the FABs so that we’re sequencing the proteins in the person, and the B cells. We did not get bone marrow from him. So we’re looking at memory B cells versus the serum protein antibodies, and we see partial overlap.

So some of them overlap, but we see unique sequences in the serum that we did not see in the deep sequencing, or in the monoclonal. So it’s an unresolved question that needs to be looked at carefully.

Marie-Paule Kieny: Okay, I take the last question. Go ahead.

Male: Kind of following up on that question. So in some of these deep immuno-profiling experiments, have you looked at tissues or lymph nodes in comparison for peripheral blood. You know, the question is how much of the information you actually accessed?

James E. Crowe, Jr.: Yeah. No, it’s a great question. So we have done that. We’ve taken tissues from 20 sites in the body. And when we started we just used pooled individuals. It was as much as 40 individuals. And if you just deep sequenced the tissues versus blood, blood is the outlier, all the tissues are memory cells predominantly, and then if you just agnostically cluster the repertoires, the mucosal sites hang together. So the GI tract and the lung look more like each other. Then you get spleen, bone marrow and lymph nodes. They kind of cluster together, and the blood is the outlier.

So the external scientific advisors last year advised us to go obtain tissues, so we have made an agreement with the organ donation folks in our hospital, and we’re collecting tissues from braindead subjects after organ procurement to do fresh cell immunomes rather than pooled. And so that we can match the person’s exact blood to their tissues. And this is not trivial, but that’s what we’re doing this year. Thank you.

Transcript curation: Alison Deshong

Back to Top