[Applause]. [Music]. [Applause]. [Music]. So, thank you. This year's laureate, Jim Allison, was born in the small South Texas town of Alice in 1948. From an early age, he was drawn to science. He played with his chemistry set and was fascinated by dissecting frogs. His father was a doctor in the town. Jim Allison sometimes accompanied him on house calls, but when doing that, he realized that following in his father's footsteps wasn't exactly what he wanted. During high school, he ran into a teacher who would not teach evolution. Allison took this very, very seriously. To him, biology without Darwin was like physics without Newton. He therefore refused to take the class, and this put him in danger of actually failing high school. Fortunately, a compromise was reached. He fulfilled the requirements by taking college biology by correspondence with the University of Texas, where he later also received a bachelor's and PhD in biochemistry. As an immunologist, Dr. James Allison's fundamental discoveries, in parallel with others, include defining the structure of the T-cell antigen.

receptor demonstrating that CD28 provides co-stimulatory signals necessary for full T cell activation, and showing that CTLA4 is an inhibitory checkpoint, which impedes activation of T-cells. All of this was basic research, and based on his basic research, Allison was the first to propose and demonstrate that immune checkpoint blockade might be a powerful strategy to treat tumors. As we will hear more about, he was also involved in the development of, he believed him up, a monoclonal antibody, which was approved by FDA for treatment of metastatic melanoma seven years ago. Dr. Allison is a member of the National Academy of Sciences and the National Academy of Medicine. He has received numerous awards, including the Lifetime Achievement Award from the American Association of Immunologists, the Law Jay Old Award, the Novartis Award in Clinical Immunology, the Medal of Honor in Basic Research from the American Cancer Society, the Harvey Prize in Human Health, the Breakthrough Prize in Biosciences, The Wolf Prize, and in 2015, the Lasky DeBakey Clinical Medical

research award now. If he hadn't been and become a scientist, Jim Allison may well have become a musician. He has played the harmonica at least casually since his youth. He told me that he became more serious in playing the harmonica during his postdoctoral years in California, and this culminated with playing on stage with Willie Nelson at the Stingaree bar by the sea in downtown San Diego. Now, that wouldn't be Allison's last run-in with Willie Nelson and his band. It so happened that Nelson's longtime harmonica player, Mickey Raphael, read about Allison's research in the newspaper, and he then invited him to come and play with the band. So, in 2016, in what he calls one of the five best moments in his life, Allison and Raphael traded harmonica solos on stage with Willie Nelson, and they remained friends to this day. Please welcome this year's laureate, Jim Allison. [Applause] [Music] [Applause] [Music] [Applause]

[Music] Thank you, Dr. Smith, and Mr. President, and members of the selection committee, and members of the Nobel Assembly. I'm really honored to have the opportunity to speak to you all, noted guests here, and give you some insights into what we've been doing and why I've been doing it for the last several years. I'd like to dedicate this talk to the many students and fellows that have trained with me and really done the work that I'm going to be talking about over the years, and also to the doctors and the patients that were involved. It had no small risk to themselves in the early clinical trials, and finally to my partner in life and science, Dr. Pat Manish Sharma, with whom much of the work I was doing to show you is done in collaboration. Now, here's the... and so the people tried to list some of these people. Of course, they aren't all up there, but the people whose names are in red are people who contributed specifically to the new work I'm going to show you later today. And what I want

to try to do is show you the early stuff that we did leading up to the idea of checkpoint blockade. Show you where that's at and then finished with some mechanistic studies that have provided some surprising new insights and how this actually occurs. So why would you want to use immunotherapy to treat cancer? Well, I think we all know now that the cancers of disease of many diverse gene mutations. Ultimately the teeth the tumor cells become highly and unstable with respective genomic integrity and so it really can be viewed as many different cancers, each independent of histah type, each which are caused by distinct genetic mechanisms. But one of the things we've learned since the war on cancer started is that trying to target these single causative mutations with targeted small molecules always essentially leads to relapse after response and that's because there's so many mutations. By the time you target one, there's already a resistant mechanism which the tumor can use to after selection to bypass it but t-cells which has been my love ever since I heard about them in graduate

school offer three things that really other kinds of cancer treatment do not, and the first one is specificity. I think that everybody knows that T-cells recognize peptides on the surface of the cells presented by MHC molecules that tell the immune system what's going on in them. And we know now from worker Bob Schreiber and many others that in the case of T-cells, they're focused on the mutations, products of the mutations that occur in the tumor cell, not just the ones that cause it, which is what targeted therapies do, but pretty much any of them. So you can get polyvalence, see about targeting these because the immune system doesn't know what the mutations are, doesn't know whether they're cause they were not, and doesn't matter, they just tell the immune system something's different. The second thing is what you've got T-cells, a number of them become four slowly renewing memory cells, and so they're there for the rest of your life, basically, and can be reawakened if necessary to come after the tumor again.

and finally, the immune system is very adaptable. You've probably got 50 to 70 million different T cells circulating around, which in that cost that can change with time since the immune system is generating new and still immune system can adapt spontaneously the changes in the tumor. And so it's easily a match to cancer, which is the whole theme of this talk. So I'm just going to condense a lot of work here to get about it on the same page. And in 1982, we worked out the structure of the antigen receptor heterodimer but with antibodies. A number of labs showed after that that soon after that that signaling through the T-cell receptor by itself is not sufficient to activate T-cells. You need an additional co-stimulatory signals were provided only by very specialized cells such as dendritic cells. In 1988, we showed that that receptor for those signals was a molecule called CD28, and others show that there were ligands farther B7. But we there was an awesome

also a molecule called ctla-4, which had an unknown function. It was only known that it did not was not expressed until after T cell activation. So, current view is when the t-cell gets a t-cell receptor signal and a cd28 co-stimulatory signal, you turn on two pathways to two programs shown in green. There are a lot of genes that associated whose products were associated with cell proliferation because we have to expand a few dozen cells of any particular type to hundreds of thousands of cells very quickly to get the soldiers in the field to attack the cancer. But you also started this wasn't really realized until about 1994. You also start an off program by induction of the program induced by seek like for expression, which eventually accumulates in a cell and can turn t-cells off. So, without going through all the work, one of the things that we had shown in the meantime is that tumor cells lining Mouse tumor cells have plenty of antigens, and we showed that because we put b7 molecules.

into them so that they can provide co-stimulation. They'll immediately get rejected, but normally they're not rejected because they're invisible to the immune system, not because they lack antigens, because they can't give the second signal. And so, I reasoned at the time that you can't start an immune response until tumor cells die and they specialized dendritic cells come in, pick up the bits of dying tumors, redisplay the antigens on their surface in the context of B7. So, this cross priming then initiates an immune response against ISA, against tumor cells. And but then, CD4, of course, is the normal process is induced, and so it's a kind of race, you know, the tumors had a head start, and if the tumor gets to such a mass as the T cells can't generate enough to take care of it, we thought then the tumor wins because they would be turned off. I seeked like 4, so we had the idea of just disabling the brakes by putting in an antibody to.

to look for a very simple process and just keeping the brakes on living. The T cells keep running well. We did this well anyway. What that's shown on this cartoon in the right and there's there's a couple of features there that are that are very interesting. When I first thought of this, one is that we're not we're targeting the immune system not the tumor cells so the kind of cancer doesn't make any difference. It's make any difference because you know all tumors have neo antigens and that really doesn't matter what kind of cancer say we don't have to make a new drug for every cancer cell. It's a universal therapy that can be potentially used for any kind of cancer. And the second one is since it started by cross priming you can take advantage of that priming by death to use in combination with chemotherapy radiation hormone therapy whatever kills some tumor cells and give agents that do those things properties of the immune therapy as well. And so there we go sorry and so the idea of

course is wonderful idea, but does it work? And so, this just shows one of the first experiments we did. We've got, you know, transplantable tumors growing in the backs of mice. You see, untreated, they grow. We have to euthanize the mice. Block CD28, they go faster. But if we inject antibodies to seek life for the tumor, grows for a while, then gets rejected. And so when I got this result, I was, I was amazed. Not expected to slow the tumor, but it just completely eliminated. I'm covering up this one molecule out of everything that's going on was sufficient to give tumor rejection. So just to make a long story very short, after two years delay because people in part couldn't understand how you could treat cancer by ignoring the cancer cell and treating the immune system. We worked with the company called meta axon later. He had a human seats left for antibody called epi luma map. And this is a patient in a phase one trial started about 2000. This is May 2001 actually what she started.

you can see these big tumors in her long. She had metastatic melanoma that had failed every therapy, including high-dose IL-2 and a dendritic cell vaccine. She got one injection of three MiG's per kig of CT for, and I happened to visit Toni Remus, her doctor, in 2011. 10 years after she'd gotten an injection, the cat scan on the right is from her 10th year checkup after in a single injection, and she's actually still doing fine. She's almost nineteen years out now from metastatic melanoma. And I should point out that when we started this work, the median life expectancy after diagnosis with metastatic melanoma was 11 months, so it was essentially a death sentence, and there was no drug that had ever prolonged life at any level whatsoever. So there are a lot of trials done resulting in a phase 3 registration trial with overall survival as the endpoint, which took almost four and a half years to complete. And what you can see here is this is the if you liver may

have line this one this is the placebo control here and what you can see is removed the survival over about four months which would have been sufficient for FDA approval because no drug had ever done that before but as you can see here very interesting thing happens between two and three years it's a survival curve flattens out and stayed that way for the duration of the trial which means basically nobody is dying of the disease after about three years and so at by 2015 there was follow-up data of several thousand patients what you can see here is it's the same at three years and kind of flattens out stays that way for ten years and about twenty a little over twenty percent of the patients so I think for all intents and purposes for this disease people that make it past four years I really think you can consider them to be cured and they don't have to worry about you know what's going to go on I wouldn't say that lightly it's based again on several thousand patients so with this disk file up

therapy. But the question is, why isn't it a hundred percent? You know, why isn't it just a fraction? There's a number of mechanistic reasons why that might be true, but of course, there's another possibility that there were other checkpoints. And Dr. Hangzhou had discovered a molecule called PD-1, and by about 2001, it was shown there was another checkpoint actually, and it differs in a couple of important ways. One of them is that one of its ligands could actually be expressed on the surface of tumor cells and induced by gamma interferon made by the CD8 cell. What that means is that rather than working at priming like CTLA-4, it's working on fully differentiated cells at the end of their development and turning off the functional cells by interaction PD-1 on the tumor cell with PD-1 on the T cell. And I'll come back to that in detail in a while. Anyway, there is very quickly this way into clinical trials. This is a summary of a phase one trial.

done by Santa Paula and very good results in melanoma non-small cell lung cancer renal cancer no responses and colorectal cancer. Although later on Bert vogelstein shows there's a subset that has genomic instability that has microsatellite instability defects and DNA damage repair. This subset has a large number of mutations and patients with that defect do respond. And then there's also no responses to castrate-resistant prostate cancer. And I'll show you later that we think figured out why that is and you know way around it. So where do we go? Well, obviously combinations. You've got two different drugs that work at different stages of T-cell development. And so put them together. So this is a trial when if several really put together by Jed wallchuck and others. This is the IPPY lua mab only arm. And you can see here the ANA and our PD well then they will vote you're mad about it here plus and I see two like four. There's a 60% response right now when you combine those two.

functionally different checkpoints, and the last time I checked, the survival is out to about four years on these patients. So we're getting close to the point of knowing whether this is going to be as durable as if you do them abalone, and there's no reason to think that it won't be. And so I think we've gone hears of this remarkably from a almost uniformly fatal disease to being able to get meaningful responses, the last perhaps for decades as well with this combination in 60% of melanomas. And so that's the best response thing, and a lot of people say, well, this is just a melanoma drug. That's not really true. There various trials have been done, the FDA's approved these edges for non-small-cell lung cancer, for renal cell cancer, Hodgkin's lymphoma, bladder cancer, head and neck Merkel cell, which is a very even more lethal form of skin cancer, also gastric cancer, in a pallet cellular carcinoma. And I'll come back to this. This is really a remarkable approval so any tumor of any histo type.

that has defects in DNA damage repair, and consequently has high low levels of neo-antigens, can be treated regardless of what kind of cancer they are. For the first time, it's ever been approval for a particular genetic defect and in a class of tumors, and so we're in a pretty good setting. We know a lot of tumors respond, not all, and not all of them is high a fraction as we would like. So to go forward really in a rational way, we need to know the cellular mechanism, Alekhya, your mechanisms of the effects, so that we can, you know, design the best way to combine things and develop strategies for identifying patients who are going to respond. And so at MD Anderson, Dr. Sharma and I built on some work that she'd been for some time and in genitourinary cancers to develop what we call the immunotherapy platform. And the goal of this is to do mechanistic studies on any patient with any kind of cancer that's being treated with any drug that would have impacted immune system, of course, we defined that as

anything, basically, and the idea is to get tissue and blood from them and analyze it by all the techniques that we can to get some insight into what's going on. And I'm going to go through the details, but so far we've got enrolled over 3,000 patients, have gotten matched samples of tissue and blood from those patients, and so I'll just show you one brief thing about what how we've used this to study what's going on in prostate cancer. So as I told you, there were no responses in a phase one with PD one with early child, so VIP ulama map there were respond there's a small number but remarkable responses of men whose PSA was undetectable and had metastasis disappear and and symptomatic relief. No thanks so this went to a phase three trial shown here, which and unfortunately failed. There was no real significant difference between the if people ooh may have only you know treatment and the placebo control, and so we wanted to find out what that was so using dr. Sharma's

approach this, which is to do with localized as treatment with localized cancers. What's called a pre-surgical trial where the patients are given a couple of doses of this case if you live in lab before they go to surgery. And so these were guys that were given two doses before they would prostatectomy. And so at the time of surgery, we get the whole organ, you know - what the pathologist needs for diagnosis. We get the whole organ and for studies, and we can pretty much do anything we want from flow cytometry to multiplex tissue sections, RNA expression, genomic sequences. It's such a really powerful technique to get data. And so this is what we found. On the left column here, these are mr. chemicals, I mean it has to chemical slides. What you can see here is before treatment, no CD4 is virtually no CDH, very few regulatory T-cells, a few memory cells, etc. But after two doses of Appeal umem AB, it goes from. I'm not sure the comparative size, but this looks a

lot like pancreatic cancer at baseline where after treatment, you could see robust infiltration with CD4, CD8, Foxp3, T-reg cells. These cells just get this activated killer cells here and macrophages. So this now becomes to look like melanoma. And so just to show you what happens though. So CT four drives the T-cells into a coal tumor but it also induces all these inhibitory markers as you can see here with PTO, PTL one, Vista, PD one, and LAG-3. So all of those are inhibitory molecules that were induced by if you live in lab treatment. So you can see the numbers of cells where you see the cells here with PD one, PDL one, and Vista. So these are two very different inhibitory molecules. When you look at the numbers, you can see the PDL one is induced on the CD8 T-cells themselves as well as on macrophages and the tumor cells whereas Vista is induced to a little bit on CD8 cells but mostly on macrophages in other

studies that I won't show you. We know that PDL one and Vista are always expressed by different subsets of macrophages. They aren't ever found on the same cell. And so this then led to propose something obvious about the at this time. Back, Mr. Barr squibbitt shut down their entire prostate program because, you know, PD one didn't do anything. And then acetyl f4 did it occasionally but not enough to be considered significant. So we went to lemon showed showed him those data said we need to give it VP to drive the t-cells in and then in a PD one or PD one logon to cover up those that inhibitory pathway that was induced by ap liver may have. And so what you can see here is one of the patients. These are PSA levels. His PSA levels have been rising for some time hitting almost 500 when he got the first dose of anesthesia for a month later the second dose a month later the third dose by then it

was essentially at baseline, and a fourth doors here anyway. The patient's metastases completely disappeared, and it was a complete responder, symptomatic relief, and everything. So there were a few of these in the trial, so with some fiddling with the dose. It was a little bit toxic in the way that was done. It's being reopened with a few hundred patients and hoping will lead to registration for the treatment of castrate-resistant prostate cancer. And so for the rest of the talk, I want to really talk about the differences between these two. There's a tendency to say, you know, we've got two checkpoints, and one of the things that people figured out right away is that PT 1 has fewer adverse events than and seek late for blockade. And so let's just use PT 1. But the point I want to make here is that these are very different drugs. It's not just that they're but checkpoints, and you don't have to think about as I told you see July 4 is hardwired. It's part of every immune

response that your body makes whereas PD one is induced resistance like yeah I'm at a fair on most of time inducing PDL 1. It's also been shown to play a role in other settings and the protection of the developing embryo from maternal attack on fetal antigen. So we think that this is a protective mechanism. It's just co-opted by cancer cells to protect them from the same mechanism as I showed you or the cartoon shows the CTA for works mainly during priming. PD-1 on exhausted cells. CTLA4. I'm not going to show you a data but it's been shown in several settings to refute new T-cells and expand clonal diversity whereas PD-1 doesn't seem to do that. It could expand the clones that are already there and number but doesn't allow recruitment of new clones. So I'll show you see to look for primarily affects CD4 cells whereas PD one only he actually only affects CD8 cells as I showed you toast if you little map seats.

like forked drive T cells into Cole tumors PD one doesn't the responses again because it works it's priming to an acetyl i-4 offered very slow where there's rapid too pd-1 which you might expect since that's just your reactivating already differentiated cells and I mentioned the adverse events and one of the things of becoming a parent the disease recurrence after response with seek life for a very rare in melanoma at least they'll usually focal and then she could be dealt with my surgery you know whereas disease recurrence after responses with pd-1 are pretty significant is about fifteen percent or so in melanoma and twenty five percent in and lung cancer resistance mechanism to include loss of gamma interferon sensitivity and other things so in order to just sort of get a really unbiased way and although it's biased that were just looking at t-cells here and what plays a role in the response we use site off time of flight cytometry where I'm not gonna go through

I'm sure everybody's heard about it, but he uses antibodies coupled to heavy metal ions that can be distinguished very sharply, you know, mass spec. And so you can do up to forty-three parameters. And so what we did was to use one for live-dead, a few for looking at dividing cells, and then a few for barcoding because we mix everything and analyze it together, and then deconvoluted but many group things by a tizzy plot called by just putting things which look more like the cells in the grouping here. And then we can identify what those are by different markers, and then we measure different clusters of the frequency of the cells in different clusters, and then we keep track. For every mouse that we look at, we keep track of the size of the tumor when it was harvested, and then we can relate any particular changes that we see with the ones that the changes that are important to the size of the tumor. And so the way that experiments done is that we give mice to MERS give AM antibody different intervals and then

they are thirteen in the case of MC we harvested look at the tissues now. I'm going to show you just data on MC 38, which is a fairly high mutation load immunogenic tumor. The response to both these agents by themselves and but we also got the same results usually it's Gila called B 16 B l6, which is very low in mutational load it's not affected by mono therapies with these antibodies so we have to use a vaccine with it but the results were the same with both of them so this is what the MC 38 tumor looks like when you gate on lymphocytes and then our bone marrow drive cells and then t-cells our cells Express CD 3 epsilon and so here the it's just dot plots basically representing you know each dot representing a cell. Blue is mice that got control antibodies red is nice to God and I see too late for a green is - they got PD 1 and so what you can see here is that is that there's a real positive of well actually

let me start down here. You look here with NKT cells in Gamma Delta cells is essentially random. There's nothing going on, they're just equally mixed. But what you can see here is there's a real paucity of red dots compared to blue or green and amongst the T rags. So this is something that actually was first noticed by surgical Allah and Tyler Simpson in my lab. I suppose by Allen Corman's lab several years ago. And that in the mouse, the an acetal 8:4 antibody can deplete regulatory C cells, T cells because they expressed so much seats like for mechanism involving macrophages mediating antibody dependent cytotoxicity. So that's that and there's controversy as to whether this happens and humans are not, we don't think it is and I'll show you some data, you know, a little while. But then here are the cd4 effector cells. This is all T c4 cells except the ones that have Fox p3 and you can see here there's an

over-representation of red dots. So these cells are being expanded by C. T. L. A. four. When you get to the CD8 area here, you can see that overall there seems to be a big increase with P. D. -1, but there are patches where S. E. C. T. R. four is expanding cells. And so this is overall what's going on. But what you can do, of course, is do a FINA graph where you line up each cluster with all the marker genes that they expressed. So we had all the marker genes from the major cell surface at incidence and differentiation transcription factors in this panel. And so just to make this approachable more visible. And here we divide them by cell type again. No difference of the NKT cells T regs. There's two kinds. There's one kind that they all have Fox p3. One has a maturation engine called KLR g1. Those cells and the smaller population that archaeology was negative. You see here look like they're depleted by C too late for treatment. If you look at the CD4 effector cells, there's.

really two major populations there's these cells that I'm going to talk a lot about. These cells for the rest of the lecture. These cells are the one like would call them teach one like because they have T bet and so they make a lot of get'em interferon and TNF alpha but they also have AI costs which is makes them a little bit strange and I'll come back to what this is a little bit later. But those cells are very strongly associated with decrease in tumor size whereas this other population which do it just looks like an uninvited T so don't really do anything and then I activate a cd4 they don't really do anything with PT one. What you see is a preference for preferential increase in a subset of cd8 cells that have high PD one hi Tim three you're also high lag three so these are what we call exhausted they're terminally differentiated but they're also functionally exhausted they do not divide left themselves but they're dividing here when you give a tacitly and a PD one they start dividing.

and these are also very strongly associated with smaller tumor size. But what is obvious from these data is that despite the fact that they've been treated with data PD one antibody, they're still expressing PD one. So if the antibody goes away, these cells are going to stop dividing again. We haven't changed the phenotype at all. And this is also supported by some work from John Weary who's shown there's epigenetic changes in the PD one gene that lock it on in these cells. You can't do anything about that. And those are not increased significantly by CTLA-4 treatment. Then there are cells here that are PD-1 intermediate. They also are terminally differentiated and that they express the openings at civil other markers, but they're not exhausted. And you can see these are expanded by both CTLA-4 and PD1. Those also are strongly associated with decrease in tumor size. And then there's these cells which look like naive T-cells again and these cells which are they're expanded.

by seats like four as well then these cells which look like recently activated t-cells and and they're actually associated with bigger tumor size. So these are getting in the way of the responses so we've really got three populations: one cd4 which expanded just by seats left for one cd8 which is expanded just by PD one and then this group which expanded by both that associated that if those are the cells that are playing the main role in shrinkage of tumor cells. And so this just sort of summarizes that the mechanist is very you know are different and that may be why when you put these two things together everything works it works much better. We're working on that now and have some data in review but measuring these subtypes is a lot more informative than just measuring total cd4 and cd8. It's a telling who's going to respond and the other thing is that the therapeutic details didn't really matter in this experiment and so does this happen in humans well this is not obviously as

easy to do as it is it - we could only find because the standard of care now is both antibodies. We can only find one patient that was treated with just at a PD Watt, and a few that were treated just with and a fantasy too late for, and only a small group that were been treated with and a PD wand. And so, this just shows the same roof, and the point is that essentially the same is the group is expanded largely about PD one. And these are these exhausted CD aids, and then the CD4, the high cost positive side gets expanded again by seats life, or we can see here mono therapy, and then also by combination therapy. You can see here t rags are not depleted by that. I notice each leg for the other thing I want to point out is that the only control that we had here with normal donor PBMC, and I just want to point out that there are none of these cells, the ones that are effective with with c24. We can't find those using this panel of antibodies. We cannot find them in the normal blood.

so I come back to that a little while. I think that's very significant because it suggests that co-stimulation may have something to do with the regulation of T cell differentiation. So I'm going to show you the evidence last less next lens to talk about. So just for those non-immunologists, there's basically two schools of thought. We know that individual transcription factors are associated with the differentiation state of T sub CD4 T cells. GATA-3, for example, is associated with what are called TH2 cells that make TH2 type cytokines. T-bet is a transcription factor that determines the TH1 fate which make gamma-interferon, TNF-alpha. ROR gamma t sends cells down the TH17 pathway to make IL-17 among other things. Then finally, Foxp3 dictates the function of these cells to make Treg cells. So this was thought as a linear process. So you've got a naive cell. It gets activated by antigen, and what it becomes depends on what's going on in

the environment at that time. But several years ago, John O'Shea of the late Bill Paul suggested that it's more of a nuanced process that their intermediates between cells. They could go back and forth or someone expects multiple transcription factors, and so it's been a lot of support for this, but not a lot of data so far. And so anybody what we're thinking missus this is done by Spencer way is that one thing that CTA Ford does on cells, of course, it can stop the response, but also attenuates the T cell signaling that negative effects lower it. So you know the real strong antigen, you might get a maximum signal straight just here at arbitrary units. But if they attenuated due to high levels of seed life for expression, it's much lower. So if you take seats like far away, and they would have give a signal strength that's much higher than you can ever achieve, you know, in wild-type max, you get supernatural, if you will, not supernatural, but super odd nights four.

levels of t-cell. I'm not trying to say these are magic cell anything, but but but you could just get a much stronger T cell receptor signal, and we've known for several years now due to work of Kim Bottomly and others that stronger signals tend to be associated with th one type cells. So we decided to look, just take C 24 away genetically and go through this same sort of process and see what we've got. Well, the green and blue dots here are wild-type and heterozygous, and the red areas are see too late for knockouts. And what's apparent by the way, these are all CD fours with that same pen, well not with the same panel, but with markers two different subsets of CD4s. What you can see is there's this group here and then this one here, which are clearly very different than the cells that are you see in wild-type mice. And so what we think is that this is represent cells that this represents differentiation of space that's prevented by the presence of C 40 cells.

in normal mice, never get a chance to go down those pathways where if you remove see to look for they, they can. And so if you go back and you look at the kind of cells that are there, yet absent the wild-type and lymph nodes and hetero, and you see these cells. These are our friends, the I cost-positive TD T bat-positive cd4s. These are the same cells you get when you treat mice or humans with, with that are seats like for its another population that's non-non-none sauce to t-cell intermediate T cell that has this one actually has T bet. And then you see this one over here as bcl6 got a three ror gamma t. So it has three different transcription factors. That's really a confused cell. It's made well maybe an intermediate on the weight or one of those other other pathways. And so there's another way to look at these rather than looking at clusters. It's called archetype analysis where the tumor you start with the T cell you want with the cell you want and then measure.

just going a little bit further away from it and measure what what cells are there until you get the ones that are the farthest away from a starting point. And so that enables you to tell the difference between cells. They might have differentiated, you know, from there to there directly, as opposed to, you know, you could tell if they follow this pathway. You end up with an archetype here, here, in here that are the furthest difference that they can be from, in this case, normal T-cell. So these are the extreme cells that are the farthest away. And so when Yuri analyzed the data about that mathematical time, you see to here that it really here out these spaces out. In this, you know, I'm secretly four-space the two that are most interesting of cluster five and cluster seventh. Cluster seven, again, as the I cost positive, PD-1 high. These are the OCD fours that we find when we treat mice with anti-seats like for these cells. They're really mixed up, guys here. And again, these are the, it's not shown ever.

these are fox be three positive so we think then that this is evidence for the fact that cyclic 4 is not just expanding and cells were already there. It may be generating an entirely new cell by sending it down on differentiation pathway, which is not possible in the presence of seats like four. So, this is something that an interesting thing that we're working on further to get further data. But I want to come back to these now because just didn't. It's not that these were unknown. I'm just showing you all that data because that's unbiased. It's just what the machines tell us. But these cells were actually discovered about ten years ago by Pam Sharma in a pre-surgical study in bladder cancer. Anyway, I cost as a member of the seed cd28 seats life or family chose the tree. He's got a single ligand, no my Casa ligand. Normally, if you tell them immunologist, I've got an i cost positive cd4, I'll tell you it's a follicular t helper cell which makes type 2 cytokines.

or it's a regulatory t-cell, neither of which would do you any good in a tumor. And so, as I said, Pam Shaw, my first fountain needs to be increased in both tumor and blood and patients after if you live without treatment and she by shortage so this at all of the cd4 positive gamma interferon and TNF alpha cells that produced or produced in response to tumor antigens study with Jed wall shock show that that crease was associated with longer survival and then in mouth studies show that they were essential for optimal efficiency of Sita life for blockade. In other words, if you do the therapies that we had shown under conditions were to get about a 90% response, it would drop to about 50% if you try to advise the lack, neither I costs or its ligand. So it appears to play an important role and it seems to be by singling through pi/3 context to the Akt pathway and now show you that this this is a really can be

considered positive checkpoint you could target it to improve efficacy of sita for signaling. So it's a study that we did to explore that where mice are given tumor and then given anti-CTLA-4 and along with either irradiated tumor cells or radiated tumor cells that have been transduced to express a constant. And these are all the singles antibody controls. This is the response rate that you get when you have an acetal A 4 plus the radiated wild-type tumor cells, about 30 percent, but that goes up to close to 90 percent if you just put a casa ligand on the tumor cell. And if you do that at a cost knockout mice, you don't see you lose that effect. So this shows indeed that by giving in our seats like 4 and then I costs agonistic I cost signals, you get a much more powerful CD4 response. And I'm time to show you data, but it's because of about a 30-fold increase in cells that make both gamma interferon and TNF alpha. So we formed a

company has made an antibody to this, and it's in trials now. We don't know what's gonna happen, at least the preclinical letters, and it's pretty good. So I'm gonna just finish by saying that you can block multiple checkpoints, and I've shown you negative and positive. Also, there's many things being explored, you know, with our clinic viruses, local ablation to enhance the United immunity, blocking other espresso factors, vaccines are coming along, and finally genetically targeted therapies in combination with check. I think that what we're going to see in the future is that checkpoint blockade is going to become a part of essentially all therapies, and might be combined with radiation or chemo therapies or many other things in counts of tumors that don't respond well to checkpoints by themselves. So, to close and say this is what we look, you still looking for, go ahead and in cancer therapy for the last 30 years or so, and that's you treat a large number of patients and compare the median survival to the standard of care or whatever, and if you

move it over about a few months. You know that's a success, and that's certainly important to those patients who respond. But we know now from if you looma have that you can do that, but also have a response rate of 20% or so in melanoma patients that you know live for a decade after a single round of treatment. And of course, this is not data, this is aspirational, but what we hope to do is actually move that tail of the curve up as high as we can get it. There's many kinds of cancer as we can, and the good news is I think we know the basic rules. There's still some frustrations in glioblastoma, pancreatic cancer, for example, which haven't responded yet, and we need to get the response rates and the patients who do respond up to again. It's close to a hundred percent as we can, and I'm optimistic you will succeed if we keep going in with that. I'll finish and close and thank you very much for your [Applause].

Nobel Lecture: James Allison, Nobel Prize in Physiology or Medicine 2018

Welcome to the Nobel Prize’s official YouTube channel, which showcases videos about Nobel Prize-awarded achievements and Nobel Laureates.

Since 1901, the Nobel Prize has been awarded in the fields of physics, chemistry, physiology or medicine, literature and peace, while a memorial prize in economic sciences was added in 1968.

Please note that our YouTube comments are moderated. Abusive, promotional and otherwise inappropriate comments will not be posted.