Scientist Stories: Mark Davis, Exploring Human Immunology

I think we will go ahead and get started. I know people are still joining, but hello and thank you to everyone around the world for joining us for this week's Immunome Lab Meeting, our second webinar in this exciting new series. My name is Kristin Aboud. I am the science editor at the Human Immunome Project, and I will be your moderator today. We kicked off the Immunome Lab Meeting webinar series last month with Johannes Textor of Radboud University, who presented on artificial immunological intelligence and provided some great background on historical efforts to develop an artificial immune system. So if you missed that session, I'd encourage you to visit our website, humanimmunomeproject.org, to view the recording. This week, we continue showcasing the research of leading scientists who are working at the frontiers of human immunology. And today, we have a very exciting talk for you. So without further ado, it is my great pleasure to introduce today's speaker, Dr. Mark M. Davis, Director of the Stanford Institute for Immunology, Transplantation, and Infection, a Professor of Microbiology and Immunology, and a Howard Hughes Medical Institute Investigator. Dr. Davis has received numerous honors and awards, including memberships in the National Academy of Science and the Institute of Medicine. He is well known for identifying many of the T-cell receptor genes that allow these cells to recognize a diverse repertoire of antigens. His laboratory also pioneered studies of the biochemistry, genetics, and cell biology of these molecules and T lymphocytes more generally, which play a key role in orchestrating immune responses. He and his colleagues also developed a novel way to label specific T lymphocytes according to the molecules that they recognize, which is widely used in both clinical and basic immunology studies today. His current research interests involve understanding the molecular interactions that underlie T cell recognition and the challenges of human immunology, specifically obtaining a systems-level understanding of immune responses to vaccination or infection. During today's presentation, please send me your questions using the Q&A function in Zoom. I will ask our speaker a broad selection of your questions after the presentation, and then we will have some time for discussion. With that, it is my pleasure to now turn it over to Mark Davis.

Great. Thanks, Kristen. Let's see. Right. Here we go.

Looks great.

Right.

Okay. Yeah, thanks.

Thanks for the introduction.

And it's a pleasure to be, to be part of this series. I think one of the greatest limitations of advancing human work in general and human immunology in particular has been the paucity of technologies that can give us precise information about the immune system function. And as Kristin mentioned, some years ago, we developed the peptide MHC tetramers which have been a way in which you can label T cells according to their peptide MHC specificity. Recently, we have improved that and have a kind of a next-generation multimer, which is using a ferritin backbone, which is a self-assembling 20-mer that contains a biotinylation signal sequence. And has enabled us to put 12 peptide MHCs onto a given entity.

And this has been really very much more sensitive and collects a lot more specific T cells, more than we really even thought that we could get.

And this just shows some of the staining with the same peptide MHC on the same cells.

These are the tetramer, the previous generation.

This is a commercial product called dextramers, which are polymers of dextran that also are biotinylated and have peptide MHCs. And they're more sensitive than tetramers, but they're also messier because they're polymers basically. And you have considerable background in many cases as shown here. But the streamers that we, what we're talking about here with the ferritin are discrete nanoparticles essentially. And so they have more uniform characteristics and can give us, you know, cleaner, not only a greater sensitivity but also cleaner profile. And what surprised us was that they we're able to label quite a bit more, 2 or 3 times more cells. So, for instance, comparing this data, this is a flu specificity versus a p65, which is a cytomegalovirus specificity. And you can see with the tetramer you get 0.18 of the cells, but with the spheromer you get almost 0.4. Similarly with the Flu Tetramer, you get almost 3 times more cells with that reagent. So this is a big deal. I mean, you're always struggling for cell numbers in terms of this, but if you're in the business of sorting or analyzing specific T cells, you, you're always scraping for material, and to get 2 or 3 times more cells, that's That's a huge bonus. Plus, when we do the analysis of T-cell receptor sequences and so forth, we are pulling in a much broader swath of specificities. So really, I mean, it makes sense in terms of you're reaching down into lower avidity T-cell receptors. That's— you're going to get a broader picture of the repertoire. And in this paper where we introduced the technology last year, we found— we particularly focused it on looking at areas of conservation, peptides that were conserved between SARS-CoV-2 and circulating coronavirus. There are like 4 major circulating coronaviruses listed here,, and we zeroed in on areas that were conserved versus areas that were not with SARS-CoV-2. And what we found basically was that people that had a mild disease course in COVID favored these conserved residues versus people that had more severe disease did not. And typically this is— yellow here is our— are the more unique SARS-CoV-2. So, and this might reflect recent encounters with these coronaviruses, which are not, you know, basically just a common cold type of thing, nothing like COVID. So anyway, this I think adds to how we can understand the widely varying outcomes of COVID Some people never notice they have it, and other people get very sick and die. So that's about as extreme a range as you can get. So we think this partially explains that phenomenon. We've now moved on in a new paper that we hope will be accepted soon, where we have made We have one with Kari Naidoo and others here at Stanford. We had a sort of faculty-student vaccination project, which many of us, including myself, participated in. And so as soon as the vaccines came out, particularly the Pfizer vaccine with BioNTech, we got to work actually like the day before Christmas or something, and did a very nice and extensive time course. And to look at that, we made spheromers for all the predicted spike epitopes for HLA-A2, and also, which is a class 1 MHC that's quite common. And then we also made for class 2, we did a DR1501, which is also fairly common, not as common as A2, but common enough such that in the 67 people in the cohort, we were able to find at least a dozen that had the right MHC, and so that we could survey them. And what was really nice was that we got whopping responses for, in this case, spike-specific CD8 cells. Some of them at the peak were 8% of the— this particular epitope was 8% of the of all the CD8 cells in these individuals. And the other ones were not shabby either, 4%, 5%. Overall, something like 20% at the peak of all the CD8 cells were in A2 individuals were specific for one or the other of these multiple spike proteins. In the class 2 CD4 range though, it was somewhat different, not as widely distributed. There was basically one or two major epitopes for spike in for DR15 that were predicted, and this one was like 12% at the peak. In this case, the peak came a week after the boost, and this one is also really quite good too. So, we're really getting a very— we're showing that the RNA vaccines are giving us tremendously large both CD8 and CD4 responses that are specific for the spike epitopes. The vaccine only contains the spike protein, not some other ones. So, now we've made reagents that include quite a number of the other ones, like 40 in total. But the— if you compare the CD8 response between people who are getting this vaccine versus infected individuals with COVID there's a huge difference in the two, and particularly the vaccination produces almost 10-fold more CD8 cells than infection. So, it's consistent with recent reports that the virus inhibits Class I expression. We didn't see as big a difference. We saw a slight difference with the CD4 T cells, but this is pretty whopping. And also, it seems that people that— there were a number of people in the study had had COVID earlier, before they got the vaccine, and they had significantly reduced response. So, yeah, by tenfold in many cases. So it seems that having COVID reduces your ability to respond to the vaccine as well as a naive person.. And that's consistent with some other data we've had and published last year or so on HIV and hepatitis C, where these serious infections damage your immune system in significant ways that can be quite— it could last a lot longer even after the infection is over. Particularly hepatitis C, where you can basically take a drug and stop the whole disease, the sequelae of that go on for quite some time. So another topic that we've gotten into, partly because we developed the single-cell T-cell technology quite a number of years ago and enabled us to dive deeply into both human and mouse responses, And it comes back to a, uh, one of the major problems in early immunology, which is that there was a whole field of CD8 suppressor T cells that, um, was quite extensive. There were thousands of papers, a lot of people involved, but the assays were not very robust, and people managed to fool themselves quite a bit so that the whole All business collapsed, and I joked that it was like the Titanic. So we got into this just accidentally in a way, because we were— Arnold Hahn and the group was interested in celiac disease, and he did a whole project characterizing T cells especially, but other cells as well, in celiac disease, where you challenge Celiac patients that are on a gluten-free diet with some gluten, and that induces a mild form of the disease, and we're very grateful for individuals that volunteered for this study because it's, you know, you're volunteering to get sick, it's not really, they're not really benefiting, but we benefited, or the research benefited tremendously because not only did we see an increase in gluten-specific CD4 cells in the blood 6 days post-challenge, which was expected based on other people's data, but what was surprising was we also saw a real surge, an even more dramatic surge in CD8 T cells and gamma delta T cells in the same patients, and they're all coming up at the same time as the CD4 cells, and it all and going away at the same time as well. And so that was just so curious, and no one knew what to make of that, including us. And so it's something that sort of nagged at me. You know, I compare it with, you know, if you're wearing a sweater and you see a bit of yarn coming out of the sweater, your mother would have told you not to pull on that. That would be very bad for a sweater. So, but in science, if you see something that doesn't fit anyone's expectations about what should be happening, you want to pull on that. You want to look at that more closely if you can and try to figure out what's going on. Something's going on that you probably would want to know about. And so we— what brought about progress was a new postdoc came in the lab, Narisha Seligrama, and he had done his PhD on this EAE model of demyelination in mice, quite a popular model that's induced with a bigger myelin peptide. And so I asked him, could you just look at the blood of these mice and see if you see anything like what we saw in celiac, and so he did, and amazingly he saw the same thing, that 10 days after induction of the disease there was a surge of CD4 T cells, CD8 T cells, and gamma delta T cells, all at the same time point, and then a week later the same cells came back in the blood in a kind of rebound manner, a way that is actually reminiscent of relapsing-remitting MS, for example. And looking at the nervous system, you could see some of those same surges were going on. So that really said, well, okay, something's happening here. It's— and it's happening both in humans and mice. So that was very special and worthy of pursuit. And so we were able to pursue this in great detail. And what we found was that the CD4 cells, as expected, were pathogenic. The gamma delta cells make IL-17, which is also part of the pathogenesis. But the CD8 cells were a real mystery, and it took us years to figure out what was happening there. And luckily, we're collaborating with Chris Garcia, who had developed this peptide library method that that you could take a T cell receptor you were interested in, probe the library, and on a good day you can get— find out what the peptide was. So we did that with the CD8 cells and found some peptides and put those into the EAE induction model, and to our surprise, they suppressed the disease. So, it turned out that the CD8 cells here are suppressors, and that the mice, if you stimulated those CD8 cells with specific peptides, you got much— most of the mice didn't get any disease, but the few that did got a very mild form. And it turned out this was connected to a particular subset of CD8 cells that expressed the Ly49 family proteins, as originally described by Harvey Cantor at Dana-Farber. And he had identified these cells as suppressive in the context of a particular non-classical Class I molecule in mice called QA-1. But in our data, we showed that these cells were really the active principle in the suppression and were not confined to QA-1, but also could interact with the classical class I MHCs. So, and that they killed basically that in culture, we could show that these cells, the Ly49-positive CD8 cells induced in this method could kill the MOG-specific CD4 cells. And so, we published that several years ago. But really anxious to move on to human work. And this is— Narisha went off, he's now on the faculty of Washington University in St. Louis. A new postdoc, Jing Li, came in, and she has done a tremendous amount of work exploring the human side of this, since we're seeing basically the same thing happening in humans. In humans, they don't express LI49, but they express KIR, which is an analogous molecule and also is on a small subset of CD8 cells, just like Ly49 is on a small subset of maybe 5% of the CD8 cells. In mice, the cures express even smaller percentage, 1 or 2% of the CD8 cells. And so Jing got to work, and through collaborators like PJ Utz and Bill Robinson here at Stanford, we were able to immediately start looking at blood samples from autoimmune patients and could immediately see that in some lupus patients there was huge levels that instead of being 1 or 2% of CD8 cells, there were 25 or 20% of the CD8 cells. So really something was happening there. And just in general across multiple diseases, this is lupus versus MS versus celiac, while many of the patients had normal levels at the time of the blood draw, there were always some that had much higher levels, again suggesting something was going on there. And probably, but the most convincing data really came from collaborating with Nielsen-Fernandez-Becker where we were able to get gut samples from both healthy controls and patients and look for these CAR-positive CD8 cells, and we could see a clear correlation with disease severity. Even celiacs on gluten-free diet had a small, you know, generally were above normal levels but then celiacs that were having a severe disease course were much higher, so that indicates that there's a connection here, and that we imagine the connection is about if you have an autoimmune disease like celiac, you're trying to combat it with these cells, and we've also seen in vitro that we can kill, that these cure-positive CD8 cells can kill gluten-specific cells in culture, whereas cure-negative CD8 T cells do not. So, but then we did an extensive amount of single-cell gene expression and so forth using the current methods, and we just couldn't see any real difference in gene expression between— wasn't like the these CAR CD8 cells in patients were going rogue or something in some strange way. No, they were basically just the same. There could be more of them or less of them, but we didn't see any systematic difference in their gene expression. Not that, you know, there might be, and that would require a lot more investigation, but at least we didn't see anything. So, so this reminded us of a paper we published several years ago, 2015, where we made this surprising discovery that self-specific T cells were not rare in healthy people. There were tons, there are tons of self-specific T cells running around healthy people, which is Contrary to a dogma that was established decades ago in mouse models where there seemed to be a very efficient deletion in the thymus of cell-specific cells using transgenic models, but basically we could tell that this actually was exaggerated in the transgenic models and that where there is, there was some deletion of cell-specific cells, it wasn't nearly as efficient, so that leaves a lot still there. And in this paper, so we had to discuss, or thought to discuss, what— why would this be? Why would you make— take a risk of autoimmunity in these cell-specific cells by having them around instead of deleting them? And the inescapable argument was, well, you're keeping them around because you might need them in infection. And this, we followed up with some data where we could tell that these cells were normally not easy to activate, but in an infection-like circumstance could do that, especially innate immune activation and molecules that do that seem to be key. So, So then in terms of thinking of a normal role for these cells, we thought, well, okay, let's look at some infected people. And so we did, and it turned out this was at the height of the early days of the pandemic, and so there were tons of people in the hospital with COVID And in no time at all, Karine Nadeau, our collaborator, managed to give us 60-some blood samples from people in the hospital. And sure enough, we saw, especially in the more severe cases, we saw a dramatic increase in the number of CAR-positive CD8 cells. And this is the kind of full-on data. And it clearly was the trend, and the more severe COVID patients had more of these cells. So, So, this fit with a model where infection breaks tolerance so that some of these self-specific cells that normally would be repressed, if they have cross-reactivity to pathogen antigens, they can break free and help defend against the infection. In contrast, CD4 regulatory T cells, which have been— were discovered 25 or so years ago and have been the mainstay of work on T cell suppression ever since, they are not doing anything obvious in the course of COVID And so, our hypothesis is that Cells leave the thymus based on some degree of self-reactivity, and sometimes that's a very modest self-reactivity, but other times it can be quite dramatic. It can be— they can be stained with tetramers with cell-specific epitopes. These cells are potentially dangerous. They're generally kept as difficult to activate, but if you have an infection, they could be activated, and then you need to control them. To prevent autoimmunity. And so, to test this hypothesis, Jing made a conditional knockout of the main LI49 family member, LI49F, in mice. That's in 90% of the LI49-positive CD8 cells in mice. Sure enough, you infect these mice with a virus called LCMV, you get a spike in the level normally, but in the knockout mice you had very little, almost— there's almost nothing left of these cells. So, but most— and there's very little effect on the natural killer cells, which also express this gene, but the natural killer cells express a much broader range. And they're, I think, not so dependent on the lipoprotein as the CD8 cells are. But the proof of the pudding is actually if you infect the mice normally, you don't get autoimmunity in terms of antibody deposits in the kidneys in this particular model. But with the knockout mice, you do. And you get some of the other hallmarks of of this kind of knocking out this kind of gene, going back to work of Harvey Kanter, is where you see an increase in germinal center B cells and also an increase in follicular helper cells, which is interesting for other reasons. But anyway, I think this says that these cells are really important in controlling autoimmunity that could arise from infection. And we know in autoimmunity in general that in many cases, autoimmunity, clinical autoimmunity, follows a serious infection. So we think that these cells are critical in preventing autoimmunity normally, and that something is going wrong in this pathway when autoimmunity happens. And we don't know what's going wrong, But that's something we're very keen to understand. In terms of the generality of this finding, some new unpublished data has been very interesting in terms of pregnancy. Here we've collaborated again with Carine Doe, who had quite a massive number of samples from women in the second trimester pregnancy. What Jing is finding is that In general, CAR positive CD8 cells are elevated in pregnancy somewhat, so fairly subtle but noticeable. And especially women with a male fetus have even more elevated levels. So I think this is interesting in terms of the generality of these cells because— and a major problem in pregnancy, understanding pregnancy is how How is it that the immune system of pregnant women generally doesn't attack the fetus, which has all sorts of other antigens, and males as much as any? So anyway, so this promises to be a really important area, both in understanding autoimmunity and in understanding fetal tolerance and pregnancy. And then lastly, I want to mention work on immune organoids, and this got started partly because working with the Gates Foundation, we learned that many vaccines have failed in clinical trials in severe diseases like HIV and TB and malaria. We're really still looking for good vaccines in that area. And that's, that's been going on for decades. And this says we, you know, even though the SARS-CoV-2 vaccines have been really quite effective, a lot of vaccines have not. And they're just trickier diseases that have been around for a long time in most cases. And so we have to understand them a lot better. And probably make a whole new generation of vaccines that would be more effective. But how to do this? The basic problem is you can only do so much in living human beings. It's quite limited, actually. We've done a lot of flu vaccine studies, but you're up against the wall in terms of getting really much detail about this. So, it seemed really useful to try to develop a new model to understand vaccination. And this is work started by a tremendous postdoc, Lisa Weger, who's now on the faculty at UC Irvine. And Lisa found that we could take tonsils, which are readily available from both children and adults. Hundreds of thousands are whacked out every year. And they're basically huge lymph nodes and have billions of lymphocytes. And Lisa basically found that you could dissociate these cells and culture them such that they could make immune responses and form germinal centers and dark zones and light zones. But especially you could vaccinate them with particular flu vaccines and get antibodies after a week or two. Subsequent work showed that we could get what's called affinity maturation, which is these cells that have some affinity, low affinity for an antigen, go through the germinal center, are mutated, and selected for having higher affinities.

We could show that this was happening in this system. And so, and a lot of other things that I won't be able to mention here, but paper was—

we published a paper last year on this, and it's become—

we're doing a tremendous amount of work in the lab now to follow that up. But just the basics are that we can put human immunity in a dish and test vaccines, but also understand all sorts of other things that might be going on.

And just a bit about recent unpublished data.

I told you about CD8 T cells and the paper that came out this year earlier.

And as putting it alongside of the much better understood CD4 Tregs.

So, how are we going to understand the relationship between these two? Here we've turned to the tonsil organoid system that Lisa developed. A student, Mustafa Ganezadeh, and a postdoc, Xin Chen, have worked out a really clever way to interrogate these cultures and look at the role of these two different types of regulatory cells in one package. Basically, what they've been able to do is to use CRISPR-Cas9 electroporating constructs that will knock out particular genes. And then we can look at the phenotype of the resulting cultures. And this is one example of these knockouts. And we get about 80% reduction in the level, in this case, of FOXP3. And the cool thing is that if we eliminate FOXP3, which is the main driving transcription factor for the CD4 Tregs, Now the tonsils will make autoantibodies, which they wouldn't before. So you could do all kinds of things, simulate with autoantigens, nothing, nothing will happen in the way of producing secreted antibodies. But if you knock out FOXP3, now this sort of stimulation will work. So we could tell that there's a significant barrier that the CD4 Tregs are actually a checkpoint against autoantigens. And it's also well known that in both humans and mice, gene— if you impair FOXP3, one of the characteristics of, you know, severe inflammation and so forth are the autoantibodies that suddenly happen.

So that's really good.

We're doing parallel experiments with the cure-positive CD8 cells and they don't have this same function, but they also— they do regulate self-specific T cells. So anyway, so that's something we're finishing up and hopefully will be interesting. The other thing we're trying to do is combine organoids. So we have the spleen and skin from the same donor using Organ Donor Network. And so that's looking interesting in terms of connecting different tissues in organoids in a way that normally happens in a whole organism. So, I think that's all I want to talk about right now, and thank the many people that have made this possible in my group, especially Elsa Sol, Lisa, Shin, Asafa, Jing, tremendous people, many other also great collaborators like Kari Nadeau, Nelson Fernandez-Becker, Chris Garcia, and really patient funding agencies like the Gates Foundation and Howard Hughes, NIH especially, and Donor Network West, which has been the source of spleens and tissue from organ donors. So I'll stop there and answer any questions.

Well, thank you so much. That was really a fascinating talk, and we have a lot of questions coming in for you. So we'll start from the top. The first question is, with the spheromers, is there a trade-off between sensitivity and specificity? And if you clone the T-cell receptors from the additional cells you get with the spheromers, are they all peptide specific?

Yeah, we did that in the original Science Immunology paper. And we haven't done it on every single thing, but in the examples that we picked out that were unique in terms of their T cell receptor sequences, they had the correct specificity when put into Jurkat reporter cells.

Okay. And the next question was, how do you adjust for the time difference between vaccination and infection for COVID with regards to the assessment of the CD8 T cell response?

Is that— you mean in terms of infection?

Yeah, I think it was the noticeable difference you saw in the CD8 T cells when the vaccination occurred after infection. How do you account for it? Time difference? Does the time seem to have any factor there between infection and vaccination?

Yeah, that's very difficult to control for because, you know, people got COVID at various times and then months later, I mean, months later, then they got the vaccine. But there were only a handful of people in that category. I think we have some new data that that then will be a more uniform timeframe of infection because these are people that were part of an infection study and a treatment study. And then some weeks or months after that study was— they recovered, they were given a vaccine. So I think some new data on that will be a little more uniform, but You kind of, again, with humans, you've got to kind of get what you can when you can.

That's right. That's right. And then also on the CD8 responses with COVID does an infection after vaccination also show an attenuated CD8 response?

I think you addressed that, but— No. Yeah, we do not have data on that. Of course, Very interesting. And definitely, I think how things will evolve. But we don't have anything on that right now.

Right. Okay. So now moving to the, the KIR-positive, CD8-positive cells observed in celiac. Were they detected in blood or intestinal biopsy samples? And the person asking is wondering if it could be considered as a biomarker of disease severity that you could potentially easily measure in the blood of patients?

Yeah, that's— I mean, we've been very interested in that. That, and we've looked, done some work to look at flares in lupus patients. We didn't really see anything that particularly correlated with the flares. We definitely The gut biopsies in celiac definitely are, there is a clear correlation. And so that could be considered a biomarker. In the blood, it's, I think we still have to, there's some subtleties there that I don't think we really understand. We have to look at some more samples. Yeah, we're looking at some, some long-term studies of MS patients with Jeff Dunn here at Stanford. And I think that's, you know, where people have a well-managed clinical cohort, getting serial samples from the same patients and correlating that with drug treatments and disease course will be critical. But that's, that, takes some time to do.

Right, right. Okay. So we had 2 questions related to the KIR expression on the CD8 cells, and it seems to be within 2 populations of intermediate and high expression. And we had one question was whether that was a technical or a biological difference.

I think it's a biological difference. And this is something that was noticed years ago that KIR-expressing CD8 cells that were activated had the lower level of KIR expression. So it seems to be a marker of a recent activation. And yeah, it's something we'd like to understand a lot more because I think it's— there's information there that could be very valuable.

But I wouldn't say we have a much of an understanding of it other than what I just said. Okay.

And is there a particular subset of the KIR-positive CD8 T cells that are more regulatory in humans, or are they quite homogenous?

Well, we're pretty sure that the KIR-positive CD8s in general are not part of the immune response against the pathogen.

Okay.

They have a much more restricted repertoire based on TCR sequencing that we've done. They're very, you know, might just be a handful of reagents on— of antigens that they're specific for. And definitely that specificity is important in terms of their activity. It's not like that's just decorative. It's definitely, I think, part of the recognizing things on self-specific cells that they're regulating. So that's how they're— I imagine without having too many specifics on this is that self-specific cells are making unusual— have some unusual gene expression characteristics. And I think those are targets.

That's what's being targeted.

My guess. And have you found any CAR-positive CD8 T cells in patients with type 1 diabetes?

Well, they're definitely there, and we're actually just, we've just been working with the Teddy Foundation, which has a great collection of samples from children that are at risk for diabetes and some of whom got diabetes. And so we have serial samples. We're just analyzing those now. In fact, I'm going to— later today I'm meeting with the postdoc, Jing, who's looking at that. So I'm hoping there'll be something exciting there.

Report back.

Yeah.

Yeah.

And what do you think the role of these CD8 cells would be in a condition like a cytokine storm?

Cytokine storm?

Yeah, uh, I think a cytokine storm, uh, might be brought about by something completely different, uh, or by a failure of these cells, uh, to act. Um, and so, uh, yeah, some— something has gone terribly wrong in a cytokine storm, but It is. It does come up in different situations, and I don't think we have a really great understanding of all those situations. Definitely CAR-T therapy can induce a cytokine storm. I'm not sure people really understand why that is, right, or what the molecular determinants are. But yeah, it's definitely something that To think about just in terms of immune dysregulation in general or inappropriate immune responses in general, I would imagine there is some role for these cells, but we don't have any specific information.

Do we know which cytokines they secrete?

Yeah, they're very big on perforin and granzyme B. And actually, they express more of that constitutively than cure-negative CD8 cells. So it seems like they're primed to kill, and that's the mechanism.

Okay. And would you expect that they would be present across all animal species or many animal species?

Well, I think we're seeing similar things in mice and humans, so I think they probably are ubiquitous, at least across mammals. But we haven't tried to do any evolutionary studies, for example. Right.

Okay. And do you think it would be possible to induce or expand antigen-specific, cure-positive CD8 T cells by using vaccines, for example?

Well, I think we It's definitely—

it would be important to expand these cells in autoimmunity as a potential treatment for autoimmunity. And I'm working with a company that's trying to do that. So, right, I think that would be the obvious thing. But of course, they could be useful for various things. But I think the most obvious one right now is, can you manipulate these cells to do a better job in terms of autoimmunity?

Which is interesting, right? Because you usually think of, to control autoimmunity, of dampening the immune response.

And this is actually— well, that's the job of these CAR CD8 cells is clearly is to kill, exactly under what circumstances is not completely clear in vivo, but they definitely have the potential to kill self-reactive cells. They don't hurt other cells. So we see that gluten-specific CD4 cells are targeted by these cells in vitro and not, say, flu-specific CD4 cells or CD4 cells generally.

Yeah.

So that's actually where the T-cell receptor specificity comes in, is they're seeing something on these cells. I think these cells are, you know, self-specific cells are imprinted in some way in the thymus, is my guess, to express antigens that normal cells don't. Yeah. And that these are the the targets.

It's very fascinating. Is there any information on how this population, this cure-positive T-cell population, changes with age?

No, we should know that. And we have aging studies that we haven't tapped into on this, but that's definitely on the agenda to look at.

Yeah. And have you looked at the CD8 positive cure cells in tumor microenvironments or lymph nodes of patients with cancer?

Yeah, no, I think that's a really interesting area. And I, and, you know, I think if you, if you say, what are you trying, what are you trying to do in immunotherapy of cancer is you're trying to activate cells that are specific for the tumor, which are normally inhibited because they're seen as self-specific. So, so these, Cure cells could actually be holding back effective responses in cancer by inhibiting those things. And we have seen in a mouse model, we have seen that in an immunotherapy where we're injecting the mice with nanoparticles to stimulate cells specific for the cancer in the tumor and inhibit or even cure the tumor in the mice. We do see that the Foxp3-positive regulatory CD4 cells decrease and these cells increase to some extent. But looking at human samples, the results have been mixed. I'm not sure why that is. I would be pretty sure these cells are involved somehow in, and probably inhibiting immunotherapy. They're not going to be helpful.

Right. That's really fascinating. We have a question on whether you could comment on the effect of checkpoint inhibitors on the cure-positive CD8s.

Right.

So that's been the kind of something we start to look at, and we've seen a few examples of patients that have a spike in the CAR-positive CD8 cells following checkpoint therapy, but it hasn't been consistent. And we have a collaborator who has a really good cohort. I mean, part of this is having really good cohorts, and we have a collaborator who's been following up on this and I haven't heard back yet on the results. But I would imagine since, you know, we got into celiac, Arnold Hahn's argument was this, the only type of autoimmunity you can legally induce in volunteers. We have certainly checkpoints are a whole other realm of autoimmunity that we're accidentally inducing in many patients. And so I really expect there'd be something here. But, and again, enhancing, in that case, enhancing the ability of these CAR-positive CD8 cells to act could be helpful there.

Yeah.

But we don't know yet.

Yeah.

Maybe one last question on the celiac, which is, does COVID infection actually exaggerate celiac symptoms?

Oh, not that I know of. We haven't looked specifically at that. It definitely in the— In the COVID studies we did do, and in the paper, the people with more severe disease had vasculitis, or one of the sequelae of COVID-19 in severe cases is vasculitis, which may or may not be similar to one of the classical autoimmune diseases. So it basically means the inflammation of the vasculature. Definitely, there were higher levels of CAR-positive CD8 cells in the patients with vasculitis. But again, whether that carries over into other types of autoimmune diseases, I don't know.

Somebody probably knows. Right. Well, thank you so much, Dr. Davis, for sharing your broad wisdom and insights on all of this fascinating work in immunology and for taking so many questions from our audience. I'm pretty sure we could probably keep you for another hour going through them, but I know that it's a very early hour for you and we really do appreciate your time. My pleasure. Thank you also to all of today's attendees for participating in the webinar and for submitting such excellent questions. We're so glad, as always, to have you join us. Us, and we look forward to many more engaged discussions. And I'll just let you know that the next Immuno Lab Meeting, which is now a monthly webinar series, will take place on Thursday, December— I'm sorry, this is the wrong date here, but it will take place on Thursday, January 12th at 10:00 AM. That day we will host Dr. Anant Madhubushi from Emory University, who will be presenting on AI Across multiple length scales predicting response to immunotherapy. There's no need to register for the webinar in advance. Please just visit our website at www.humanimmunomeproject.org to obtain the link, and please do share it with your friends and colleagues. And while you're there, please sign up for the newly launched Immunome Report, which will feature voices of researchers working at the frontiers of human immunology and AI. And provide all of the in-depth and interesting coverage you've received in the past from the COVID Report. An issue will be coming out soon, so please sign up now so you don't miss out. And finally, please visit our website and follow us on social media where we will post a recording of today's webinar. You can also find a recording of last month's webinar there and as well as the previous COVID lab meetings. So thank you again for participating today. And we look forward to seeing you next month at the Immunome Lab Meeting. Thank you so much.