Webinar: Modeling T cell dysfunction in tumor microenvironments

0:00I'm very happy to introduce today our speaker later on, but initially our company, so creative file labs. So the topic of today's talk is modeling people, this function, tumor micro environments, and I will briefly start with quick introduction of the products and services that are related to this topic and that our company is specializing in offering. So our Envivo CAR-T platform exists of different aspects. And of course Envivo CAR-T is a new approach with many different advantages over the existing platforms and it will allow engineer T-cells directly within the patients and thereby eliminate the X-Vivo manufacturing. It would allow also for a more rapid treatment timeline where we can do same day or next day administration, which is especially relevant for a very aggressive disease.

1:01There is also reduced cost associated with this approach and more broad visibility. Envivo CAR-T technology landscape exists of different approaches. There is the viral approach using Landiviral AV-based targeted delivery to the T-cells in Diva, but then there is also non-viral approaches which include the mRNA or self-amplifying RNA as well as circular RNA, targeted LNPs as well as the Vivo genome editing. And I will zoom in a little bit more on the non-viral delivery technologies that should enable more transient, stable, scalable CAR-T generation. So we can offer IVT, so D-Tructors-Cribe mRNA and Cornic-Cole CAR constructs. We also offer self-amplifying RNA, circular RNA, the DNA and LNP delivery systems. So that economic CAR yourself is targeted LNPs. Just going a little bit more into our circular RNA, some of the results that we've obtained. So we do here show basically the preparation of a control circuit B3 as well as the circular

2:09RNA CAR GFP here, which contains the CD19 CAR. You can see here both the circularized as well as the delinearized versions of those to quality control here. We've also, then, test the B3 here, it's showing here, delivery onto the T-cells to generate the CAR-T's. So we treat the cells here with different constitutions and evaluate the viability of the cells over time, as well as the expression of the target of interest via T-NL and P delivery and electric operations. As you can see here, we have a PBS control impact as well as positive control constructs and then the different constitutions of the T-NP-3, and that is electric-perioded into the cells. And we've also evaluated, then, the CD3 positivity rate and found that it was exceeding 99

3:13percent in these different conditions that we test via effects. In addition, we've also evaluated the functional effect, so the cytotoxicity, by electroporating the CIRC B3 into the cells and then evaluating the cytotoxic effect on SHP 77 cells. As you can see here, that was also successful. So in summary, we have an NPTRA transfer system that was established for RNA production. It's a very robust RNA platform, and we've demonstrated that it can express the circular RNA for up to six days in beta. We have also expressed different constructs and then evaluated their core expression for the core and GFP in the T-NP-3 cells with more than 70 percent of double positive cells. We have functional part-p cells that can achieve more than 99 percent CD3 positivity and

4:15stable car expression. And we've also generated part-p cells that were effective in cytotoxic killing off at the HP 77 tumor cells. Furthermore, I would also like to briefly go over two more topics. One is the micro-phobic solutions that we offer so we can also custom develop chips. We enable, we provide end-to-end covering of your workflow, starting from the design of the chip optophotography as well as the precision mold making and chip modeling. We also have compatibility with multiple materials against our COPDMS, silicon columners, different type of columners. We have different fabrication methods available, including 3D printing or laser relation micro-pensioning and support. And these methods can only apply for diverse applications to provide something to I-PROP drug screening support. We also have an organ on-chip and droplet systems available.

5:17So if you're interested in that, go free tool to reach out. And then coming to the topic or cluster to today's topic, I'm sure our speaker will introduce more about the tumor micro-environment. But it's of course an interesting topic. It presents several physical and biochemical barriers that can impair both passive as well active immunotherapies. And so we have, there are three dimensions of TME suppression. There's more than ball-extras physical barriers and immunosuppression. And we are attempting to offer pre-clinical framework for your experiments by offering different vivo models they can be applied to study. So for oncology, we offer many different both carcinogen as well as propanious models, but also syngenate, xenographs and humanized models as well as two more organoids. And we have these available for different types of cancers.

6:20So again, if you're interested in any of those or if the one you're interested in, it's not listed fairly to reach out and we're happy to elaborate on that. And then we also have several other animal models available, not only mouse or red, but also NHBs for these different models that are listed here. And then with that, I would like to end the introduction of the company and then I'll hand the floor back to Kathy for introduction of our speech or two. Yeah, thank you Sarah. And now let me introduce our speaker for today, Greg Delgolf. Dr. Delgolf is a professor of immunology at the University of Pittsburgh and director of the tumor microinvolvement center at the UPMC Hillman Cancer Center. And his research examines how immune cells adjust their function in response to metabolic signals and why T cells lose their capacity once they're entered tumor. And his lobby is using this in size to develop new ways to strengthen immunotherapy, including

7:22the repurposing of existing metabolic drugs. And he also started how suppressive T cells populations are sustained within the tumor microinvolvement and how does can be used in two modulated immune function. And before Dr. Delgolf begins his presentation, I want to quickly remind our audience that we will have a Q&A session at the end of the presentation. And if you have any questions during the presentation, please type your question into the QA panel. And Dr. Delgolf will get to them after his talk is finished. Now, let's welcome Dr. Delgolf and thank you for presenting your important findings to us today. You may begin when you're ready. Great. Thanks very much for the invitation. Happy to share some work from our group that I think will be relevant to the work that the modeling ideas that have been discussed. Just briefly, let me just get my way to ready to go.

8:26So I think the genesis of the lab has really been trying to understand what Sarah brought up earlier, which was how does the tumor microinvolvements ultimately suppress the immune response? And I think maybe 10 or 15 years ago, one might have thought that this is really just being more about the cells that were present. There's a different kind of constellation of cells within tumors. But we now know that there's actually a lot more that's associated with why T cells are suppressed in cancer. And we think a lot of this has to do with the fact that tumors are metabolically active tissue. So the tumor microinvolvement, this kind of name that we have for this organ, that tumors solid tumors set up, is really characterized by a very distinct metabolic landscape driven mostly in part by the tumor cells activity themselves. And that creates both a sink for essential nutrients as well as a buildup of toxic byproducts.

9:31This is a melanoma patient metastasis. And this is a cutaneous metastasis. And what's special about this is this patient before their biopsy, they took a pill that contained piminid dissolved. And piminid dissolved is a way that we can mark hypoxia within tumors by just using an antigen to this adduct that forms when oxygen is low. And you can see that while this tumor does have the kind of characteristic hypoxic core, it also has hypoxia in other places as well. And that is really important to understand that the immune system, in this case these pink cells, are constantly dealing with these metabolic pressures. And we've previously shown that we can measure this either with hypoxia or with the metabolic rate of the tumor itself. And we can actually predict immunotherapy outcomes just based solely on the metabolic activity of the tumor. Hypoxia itself is an independent predictor of immunotherapy resistance outside of T cell infiltration. What we also know is that progression on immunotherapy results in hyper metabolic cancer cell states that cancer cell metabolism is plastic.

10:39And that in the face of incomplete immunity, you end up selecting for cells with higher metabolic activities. And we know that nutrient competition is a feature of cancer immunity that there are some nutrients that both tumor cells and T cells like to eat. And so there is a degree of competition. But ultimately, we can study tumors in people and we can study tumors in mice and look at the immune system all we want. Some of the questions are how much of T cell dysfunction is environmental and how much is intrinsically programmed. And these are things that are not as clear and are difficult to pull apart and such a complex system as a cancer forming in an animal or in a person. And so we started looking at T cells to try to understand a little bit more of this. And one of the things that we identified very early on was that T cells themselves aren't just products of their environments. They also change. And one of the things that we really found is that when we looked at T cells that infiltrated tumors, this is just in an experimental melanoma.

11:46You can see that there's really a very profound loss of mitochondrial mass. And this is this loss of functional mitochondria is concomitant with also a repressed competitive advantage for carbohydrates like glucose. And you can see that very, very clear comparing these peripheral T cells here to this to the tumor infiltrating T cells here. You can also see it up here in this electron micrograph right that the mitochondria of these T cells are just very difference. Small, malformed and and and of smaller mass. This occurs not just in mice and it happens in people as well. And we've observed this in multiple solid tumors, but also happens in liquid tumors as well. And the severity of the mitochondrial atrophy in that tumor infiltrating compartment is variable in the patient population. It tells us whether or not that patient has functional immunity. But what's really important to understand is that the most terminally exhausted T cells are the ones that exhibit the most severe mitochondrial dysfunction.

12:52And so when we showed this now over 10 years ago, what we were unclear of was how much of this was the exhaustion program. That's intrinsic to T cells changes their metabolism or do changes to metabolism actually cause T cells to become exhausted. There's something that we were we've been very, very interested in now and trying to pull apart and pull apart these threads. One of the pieces of information that really got us there was really identifying why T cells lost their mitochondrial mass. And one of the ways we showed is that they repressed the activity of this transcriptional co-activator called PGC-1-Alpha that programs the process by which you generate new mitochondria. And what we could show is that if we just retroviral over expressed PGC-1-Alpha in T cells, we could endow them with metabolic flexibility. We call this spare respiratory capacity, but this is essentially a read out of mitochondria that are geared for more potent activity.

13:56And it's quality of memory, of long lived cells that they have high degrees of this respiratory capacity value. So this transgenesis could do this. We could put PGC-1-Alpha into T cells and kind of rescue their activity. But more importantly, they functioned better. So this is just a model of cell therapy. We took OT1-T cells, we put them into OVA expressing tumors. And you can see that T cells that are overexpressing PGC-1-Alpha, they have elevated mitochondrial mass. And those T cells work better as a therapy. We can extend survival, we can clear some tumors, those T cells function better. So in other words, fixing the metabolic defects of exhausted T cells was enough to get them to have elevated function. So that's a really important aspect of things of understanding how all this stuff works together. And so this really brings up this concept of T cell exhaustion, which I think previously was kind of an umbrella term for T cells that don't work. But now, in the last 10 years, we've really defined that exhausted T cells are essentially a lineage of T cell differentiation that's just deleterious for cancer.

15:03But you can see this in things like in chronic viral infections, but of course in cancer. And what it really is is not just the expression of these co-inhibitory molecules, like PD1 or Tim3, but rather a progressive loss of function that starts first with the loss of poly functionality. And so, these are immune cells, or T cells make lots of cytokines at once. And this is something that gets lost. They start to window down their effector function into something that's more mono-functional, making a single cytokine or a chemokine. They also lose their ability to self-renew, which is a capacity of memory cells, right, to re-engage or proliferate first. They don't kill targets very well, despite being full of cytotoxic granules. And while they express some lineage supporting transcription factors, which you can see here, Emi's Adermin, Blimp and Talks, none of these really define the function of these cells. And it's rather a moulange of all of these different cellular networks that are being kind of hyperactivated. And ultimately, when they reach this terminal state, their apatosis prone and actually begin to acquire suppressor properties in the form of adenosine generation.

16:10They also make some IL-10 as well. Again, they're a lineage. They have an epigenetic signature that prevents them from ever becoming more memory-like. And that's really important to understand. Other things to understand is that they are not the target of checkpoint blockade-based immunotherapies. In fact, patients with high degrees of exhausted T cells are fated to do worse, not better on immunotherapy. And that PD-1 blockade really acts on these less differentiated, more progenitor-like cells. Finally, this is a progressive problem. It occurs commensurate with decreased metabolic sufficiency. But the question that we really wanted to understand was, could we really break apart and study these things? Maybe they are the right soldiers if we can identify them. And so to do this, we had to understand how to generate them. And this is where the modeling comes in. Can we model T cell exhaustion without putting them in a tumor to understand what the intrinsic metabolic defects of these cells really are? And so the way that we really went with this was a very common metabolic stress that T cells receive when they enter into tumors and to a lot of inflamed tissues, which is hypoxia.

17:20And that's going to directly affect mitochondrial activity due to the fact that you can't do oxfos without oxygen. Now, much of the work in hypoxia has been done in genetic systems by deleting the sensor HIF1-alpha or its upstream regulator VHL. What's really interesting is that hypoxia exposure just by itself can drive co-inhibitori molecule expression. It can actually cause T cells to become more permanently differentiated in an effector-like. And if you just activate T cells and grow them under hypoxia, it actually improves their ability to function once they go in vivo. But the key thing is that hypoxia has also been known to be immunosuppressive to T cells. And what we reasoned was that hypoxia is not the only signal that T cells get. And that one of the other signals that they're getting is that kind of continuous stimulation, that continuous tickling of their T cell receptor from tumor cells or infected cells. And so what we hypothesized was that hypoxia was not necessarily a signal in and of itself, but a signal modifier.

18:23And it changed the way that the T cells would see their antigen. So we developed this system using stimulatory kind of off the shelf magnetic beads the kind you would buy from anywhere really, that are coded with CD3 and CD28. And we did, and we've done this now in both mouse and human systems. We took purified CD8 T cells, activated them with these stimulatory beads. And then we either removed the beads as you're supposed to do and expand them in expansion culture. Or we continually added the beads to kind of mimic that constant stimulation. And we did that both at ambient normoxia or in a hypoxia chamber set to a hypoxia level that was very consistent with what we see in tumors and what is consistent to get that piminid dissolved to deposit into cells. And what I hope you can appreciate here is that stimulating T cells continuously does upregulate co-inhibitory receptors. And indeed hypoxia exposure alone can also upregulate other co-inhibitory receptors. But the experience of both stressors simultaneously drove the majority of cells to be kind of PD1 high and Tim3 positive.

19:25Now that's just the surface, of course. We can look at all these other markers like lag3, Tidget CD39, the transcription factor, TOTS. Again, these are all pieces of the puzzle that tell us that these cells might be exhausted. But the real proof is in the function of the cells. Right? Is that continuously stimulating the cells, even though they kind of look like they might be terminal, they still make cytokines. Hypoxia, as we've as was previously shown, can actually make cells look even better in terms of their effector function. But seeing both at the same time drove the cells to really lose those poly functionality. The ability to make multiple cytokines. And really only make gamma in or fear on their best. Now maybe we were just kind of like stunning the cells. Right? The cells were just kind of refractory to stimulation. So what we did is we actually took the cells. We put them through this five day protocol. And then we took them out for another five days outside of these conditions. And just gave them aisle two. And what we found is that that this function actually persists. Is that while the continuous stimulation cells actually recover and they are able to actually be even better effector T cells is measured by these two cytokines.

20:33If they experienced hypoxia at all during this process, even for the first five days or for the whole time, they were never able to recover their function. Which was telling us what we were inducing was a long term lineage decision. And that was something that we wanted to continue to explore. And so we did some RNA seek analysis on these four different populations. And what we found was that we weren't just getting a mix of the continuous stimulation kind of program and the hypoxia program. What we were getting is the combination of both drove unique sets of differentially expressed genes highlighted here by cluster one and seven. And really, if you can see on this PCA plot, they have completely different transcriptional profile. If we did go in and look at what these cells were doing and compared what those gene sets look like, they look the most like terminally exhausted T cells compared to all the other populations. But we also found in our RNA seek and I don't have time to talk about all this today because I want to talk about a few stories today.

21:33Is that our RNA seek did confirm a pattern of a transcriptional oppressor called blimp one, which kind of repressed mitochondrial biogenesis. And so again, so this brought us back to this PGC and alpha molecule and how this thing worked. So we went back to our adoptive transfer system where we overexpressed this molecule. And we kind of asked, well, what happens to those cells if we kind of only improve the mitochondria, what really changed? And while we could see some things that were like transcripts that were important for like cytokines and things like that, like graft versus host disease transcripts. The one thing that we really found that was really interesting was the number two gene set on this list, which is the regulation of reactive oxygen species. Because PGC and alpha doesn't just make new mitochondria, it also induces antioxidant proteins that buffer all of the raw that you're going to get from that oxfoss molecule. And indeed, what we found was that if we overexpressed PGC and alpha, those T cells were less, even though they had more mitochondria, they had less mitochondrial raw in their DNA.

22:38They also, which you can see tabulate here, the elevation of mitochondrial raw is kind of evidence in terminally exhausted T cells, as well as in our in vitro system. This mitochondrial raw is a product of exhaustion. And so we asked a question very clearly, very, very, very directly, was is raw sufficient to drive exhaustion? Could we actually is raw by itself without all these other features of our in vitro system? And so the way that this works is hypoxia actually kind of heredoxically induces raw, because it actually lowers the efficiency of electron transport and you get kind of reverse electron flux through this system that originates mostly at complex, at mitochondrial complex one. And so we could actually model this in vitro by using a complex three inhibitor called antimice and a typically used in oxymetry experiments, but this is a well known Ross generator in the field. And so indeed we could show this we could actually treat cells without any chronic stimulation without any hypoxia, just give them this complex three inhibitor and it would cause Ross to get elevated.

23:46And what was really cool about this experiment is we could add another inhibitor we could inhibit using wrote known, which is a complex one inhibitor and actually prevent the Ross from accumulating. Okay, so this was a really important control to have now what I'll show you here is that if we just treat the cells with non toxic doses of antimice and a we can actually induce exhaustion in the absence of all the other things like hypoxia, chronic stimulation. But in a way that if you knock out complex one, you actually rescue the phenotype, which you can see also here in the function antimice and a we lose effector function we gain it back by adding another complex by adding another inhibitor. So this was really important. It's told us that it wasn't the lack of mitochondrial function, the drove exhaustion. It was the presence of dysfunctional mitochondria that we're producing Ross. So in other words, a T cell would rather have no mitochondrial activity than dysfunctional mitochondrial activity. So that was really, really important for us to understand how was Ross working. Well, to answer this question, we used anti antioxidants to try to try to inhibit and neutralize that Ross.

24:58We do this with kind of off the shelf and antioxidants like an acetylcystine that was enough to rescue T cell function. We could also rescue a T cell function when we use an acetylcystine in our in vitro system, but the really cool thing that we did was we over express just only an antioxidant protein. This is called GPX1. It is a multiactive antioxidant protein that acts on both superoxide and peroxide, and that was enough to actually rescue T cell function in the tumor. So how is this doing? How is it doing this? How is Ross doing this? Well, Ross does a lot of things. So it can induce DNA damage, and we've now shown that we can see this acting most notably on telomeres. It can interfere with epigenetics, which I'll talk about in a couple slides, but it's also a very strong mediator of enzyme activity. One of the things that Ross is really good at doing is inhibiting phosphatases. If you ever want to check the staining of a phosphatiracine antibody, you just treat the cells with peroxide, and it inhibits tyrosine phosphatases and then it goes to the roof.

26:09And so what we reasoned is that maybe this elevated Ross in as a consequence of mitochondrial dysfunction was actually changing the signaling of these T cells. And remember, chronic activation is how these T cells get this way. And so maybe this is kind of mimicking the tyrosine signaling that you get from the T cell receptor. And that's exactly what we showed is a treatment with antimicene alone just in just inducing Ross could actually cause massive changes to the phosphatiracine signature that we see in a manner that could be fixed with why adding another inhibitor. And you can see, boss the proteome, all of these different signaling molecules are elevated if we just use 4G tenders, a panphosphatiracine antibody. This was enough to mimic chronic activation in the sense where you can see here, antimicene A, actually causes near 77 to go up. This is a well known reporter of TCR signaling. We do the same thing with a bona fide tyrosine phosphatiracine inhibitor, orthovanidate.

27:17And that was enough to drive N-fat into the nucleus, a downstream consequence of tyrosine signaling. So N-fat being in the nucleus chronically generally is associated with a nirgic or exhausted phenotypes. So in other words, the Ross caused by mitochondria was directly cross talking to the T cell receptor signaling cascade and kind of mimicking that chronic activation. But it's not that's not all that's occurring here. Transcript stones are also shaped by chromatin landscapes, which are an epigenetic phenotype. And histone modifications by nature are metabolic. If we want to open up chromatin, you acetylate the histones, which causes them to relax and allows transcription factors to bind more easily. Oxygen actually plays a critical role in demethylation reactions, the removal of these methyl groups that allow them to be modified by more... by differently charged metabolites. And so what we asked kind of similarly to this Ross phenomenon was does exposure to hypoxia also change the chromatin.

28:21And so to do this, we utilized cotton run, which is a low number chip-seq protocol. There's another version of it now that uses transposases called cotton tag. But these allow you to do both histone modifications and transcription factors from very small numbers of T cells. You can even do this at the single cell level. But this was 2018 we were doing this from small amounts of cells like anywhere between 10 and 20,000 cells, which we could get from tumors. And so just from a single tumor, we were able to generate RNA-seq data as well as things like four different histone modifications and transcription factors. And so you can see here's the Tox locus. This is associated with T cell exhaustion. You can see it's repressed in non-exhausted cells, but it's open and actively transcribed in the kind of terminally activated, our terminally exhausted T cells. So now this is kind of cool. I can show you all these different traces from all these different marks.

29:21But if you really want to look at how does the entire epigenome change you need is beyond human comprehension. So we were able to use a hidden Markov model called Chrome HMM. And when we fed it all of our data, what it predicted was this bivalent chromatin that occurred. Now bivalent chromatin, which you can see here, this is this preponderance of bivalent chromatin as read up here by Chrome HMM. This had Markov model that found this. What is bivalent chromatin? Well, silent chromatin is very methylated. Open chromatin is demethylated and contains acetylation groups. Bivalent chromatin is described as chromatin that has both activating and repressive marks. And both of those marks are methylation groups. Now, remember what I told you before? Oxygen is important in the chemical reaction that removes methyl groups. And so this was actually what we would have predicted based on the oxygen data, the oxygen tension data. So this was really important to see. Now, and now importantly, you may be saying, well, what is the relevance of going from 10% or 4% of the oxygen.

30:25Or 4% bivalent to 15%. Well, if we actually enumerate what these bivalent genes are, they're kind of all of the usual suspects that we see. Like interleukin 2, TCF7, S1PR1 and KLF2. These are all genes that are associated with stemness and kind of long lived function. And these are cells that are not expressed, but they have these repressive marks that are present on the chromatin. These are bulk, these are not single cell data, these are bulk cell data. So how did we know that the two methyl groups were actually in the same cell? We used PLA to confirm this. And you can see these spec, these specs here are the proximity ligation of both, both methyl antibodies together. They need to be close enough to trigger that proximity ligation. And you can see that only happens in the exhaustancy cells. So how is this related to hypoxia? We actually were able to develop a genetic model in the lab to reduce hypoxia in our workhorse cell line B16.

31:28And we do this by deletion of a structural component of complex one. This results in dramatically reduced oxidative consumption rates and thus less hypoxia in the tumor microenvironment. And so if we do cotton run and RNA seek from the exhausted T cells we get from these tumors, you can see they start to re-express all those bivalent genes. And it's associated with a loss of deposition of that repressive mark, which was telling us oxygen was playing a very key role in removing this repressive epigenetic mark and allowing these bivalent genes to be expressed. And indeed these kinds of tumors do much better on immunotherapy. So can we actually show that in a way that is inducible? We have two different models in the lab that we use. One is using the complex one inhibitor metformin to kind of mimic that deletion of complex one. And that isn't enough to improve responses to PD1.

32:29We can also use Vegeph inhibition. This is using the anti-angiogenic acet nib that can at low doses that normalizes the vasculature and improves oxygen delivery to the tumor, which you can see down here. And that also synergizes with checkpoint blockade in multiple models. And we've actually gone on to do clinical trials with both of these inhibitory strategies. So this is all telling us that the exposure of T cells to both hypoxia and that persistent stimulation in isolation is enough to drive a dysfunctional exhausted phenotype. But remember the T cells also sit in an environment, right? And so how much of the environment is contributing to their dysfunction? So can we model that? And so what I've told you before, what I've already told you and kind of summarizing here is that tumor cells are metabolically active and they can consume oxygen and produce raw. And that's enough to take CD8 T cells that are otherwise healthy and kind of deviate them into exhaust the T cells.

33:30And here's a number of groups, including ours that are underlined, a number of papers that show this can happen. We also know that glucose is also in competition. It can consume it can build up lactates and these types of manipulations are sufficient to take otherwise healthy CD4 T cells and enrich for T regs. And we've done this. We've shown this a number of different ways. But all of our studies in our field really rely on making these indirect manipulations to either the immune cells or the tumor cells and kind of measuring the changes. But could we model it even further, right? And that was the question that we really wanted to know. And so what are the immune cells really seeing when they enter into cancer tissue? And what they really are seeing is what's called tumor interstitial fluid. That is the kind of metabolic milieu of the tumor micro environment. And that's the media, if you will, that these T cells are kind of swimming in. And so to study this, we this this this started with a really nice collaboration with this investigator at the University of Chicago, Alex Mure.

34:32And he had really shown that you could generate a quantitative metabolomics from from interstitial fluid that you could isolate from tumors. And pancreatic cancer models. And what he really he showed an absolute so he was able to absolute quantify a couple hundred nutrients that are more enriched or depleted in interstitial fluid compared to the plasma of these of my sparing tumors. And what what he was able to do because he had absolute quantitation was to actually generate a physiologic media based off of tumor interstitial fluid called tiff media. He and I have been working together since 2020 at understanding both the cancer cell and immune cell effects of exposure to these just the nutrient fraction of tumors.

35:32And so we generated we isolated this tiff media and we just started experimenting with it in T cells. So now we're not talking about chronic activation. We're not talking about mitochondrial dysfunction. We're talking about T cells just being exposed only to the nutrient fraction of tumors. And we ask kind of what happens. So we took kind of off the shelf T cells. We activated them and get generated effectors in rich media. This is all our PMI based media. And then we just simply re stimulated those T cells either in the rich media or in the physiologic media, the media that was based off of tumor nutrients. And the results are kind of underwhelming. The T cells still produced a lot of cytokines. They made IL-2. They killed their target cells. They really didn't you could maybe draw a couple of these lines and maybe see some significance. But this didn't look like the immune dysfunction that we were seeing in cancer. But I think we all can appreciate the T cells don't simply arrive on the scene and immediately start killing tumor cells. They they spend a lot of time in the tumor tissue. And so we instead activated T cells in rich media and then passed them for varying amounts of time into the tiff media.

36:43And what was really interesting when we did this experiments was that the moment the T cells arrived in the tiff media, they stopped proliferating. You can see this here as as as as born out. They're expecting super super well in our PMI. They enter into the tiff media and they flat line. Right. Now what's really intriguing is these T cells don't die. It's just a cytostatic effect. They sit there. They still differentiate, but they don't proliferate. What's more is the more time they spend in the tiff media, the more dysfunctional they become. So one day is enough to induce a little bit of dysfunction. Three days and five days. All these cells are the same age. They're all seven days old. They've just been passed into the media at different times. But you can see five days in the tiff media. These are very poor at making multiple cytokines. So this because the media is chemically defined, we know exactly the nutrients. This is not guessing based on dogma in the field of glucose and hypoxia and lactate. We know exactly what the T cells are being exposed to. We can actually take all these metabolites that are depleted in tiff media and do add back experiments.

37:51And that's exactly what drew a new PAN to students are two trainees that worked on this project. They did. They looked at all the metabolites that were depleted in the interstitial fluid and added them back to the tiff media. And when they did this, they found that a single nutrient was able to do it, arginine. So arginine is heavily depleted into a micro environments. It is essential for T cell function. But if arginine was added back to the tiff media, they rescued their ability to proliferate in that tiff media, which was particularly exciting to see. Although a little bit disappointing because we were kind of looking for new metabolites metabolites that might also be in competition and arginine is long been known to be immuno regulatory when in low in low concentrations. So while this was interesting, it was maybe a little bit, you know, a little bit, what's the word I'm looking for? You just kind of expected one might say, but that's where things got really, really interesting.

38:52If we took these T cells now, we've completely rescued their expansion. But then we asked them to make cytokines again. When we re-stimulated them, they actually didn't rescue their cytokine production. They were actually just as dysfunctional as the cells that were in the tiff media and that had arrested their proliferation. So this was really, really interesting. What was even more intriguing is if we added all of the depleted metabolites back to the tiff media, we still couldn't rescue the function of the cells. So this was really interesting. What this was telling us was that this wasn't about a metabolite that was missing in the tumor. This was something else going on. So again, I've been telling you for the last several minutes that tumor cells are very metabolically active and they eat up all these really important metabolites for immune cells. But that's not all that's happening. The tumor cells are metabolically active. They're eating all these things, but they're actually, and so we have this image of like soldiers marching in a desert, right?

39:53They don't have anything to thrive on, but that's not really what's true here. The tumor is also producing things and these things we know are toxic. They change the biology of immune cells. And so when we do this, it's actually creating more of a wasteland than a desert, right? And so this might be what's really happening. So maybe it's not something that's missing in the tumor is interstitial fluid media. It's something that's enriched, right? And you can actually see this. There are far, if we do it, this is a radar plot of all the metabolites that we measured. And you can see that there's actually a lot of things that are enriched in tumor interstitial fluid, not things that are missing. And so we did a lot of of add back and drop out experiments to kind of, and I don't want to like, you know, I don't want to belabor the points. And so I'm going to focus on the one that was important that was really, really exciting to us, although there are several on this list. It's right here. It's called phospholethanolamine. This is something that we had never really heard of before, and we had to learn a lot about, but this thing here was extremely enriched not just in the tumor interstitial fluid that we looked at, but in many tumor interstitial fluids that we had measured.

41:01What you can see here is a commonly upregulated metabolite and cancer, because the gene that metabolizes it that takes phospholethanolamine and turns it into something else, which I'll talk about in a second, is very frequently silenced in tumor cells when they get genomically unstable. This is when we went the loss of braka or the loss of p53, and you can see in multiple different models of tumors, including in breast cancer, pancreatic melanoma, you can see elevation of phospholethanolamine. What is phospholethanolamine? Well, it is the head group of a membrane phospholipid called phosphatidylethanolamine or PE, and PE is the second most abundant phospholipid in our mammalian cell membranes. What, where does phospholethanolamine kick in? It is part of what's called the Kennedy pathway, and the Kennedy pathway is how we denovate these membrane phospholipids, either PE or PC, and you can see there are parallel arms of this pathway.

42:03Now, I need nothing about the Kennedy pathway. I don't expect you guys to know anything about the Kennedy pathway, but what it is, but it is an important feature of what we're going to see going forward. So, here are the data that got us excited. If we took rich media and just added phospholethanolamine to it, we could, we had no impact on the ability of these T cells to proliferate, however, those T cells were profoundly dysfunctional, as dysfunctional as they would be if they entered into the tumor and residual fluid media, which you can see tabulated here. So, what was going on here? Is there something very special about phospholethanolamine? Well, maybe this is, you know, maybe this is just something special about this metabolite. It does something interesting. It's been shown maybe bind mitochondrial complexes and things like that. One of the things that we did is we just added the precursor, ethanolamine, or the product, CDP ethanolamine, back to T cells. These are not enriched in TIFF, but it was kind of interesting to see anyway.

43:04And you can see that either the precursor, ethanolamine, or the product, CDP ethanolamine, each one of these things were sufficient to induce T cell dysfunction. So, there's telling us that the flux through the pathway was what was really causing the problem here. So, what was the mechanism of what was going on here? And we actually came up with the answer serendipitously, because we were doing experiments measuring the kinds of dysfunction that we were seeing in phospholethanolamine treatments. And what we did for one particular experiment is we re-stimulated the cells through their antigen receptor, or we chemically activated the T cells, which bypassed all that upstream signaling using PMA and I and I'm Ison. And what we found was with the exact same cells, while stimulating through the antigen receptor resulted in defects, PMA and I and I'm Ison rescued all those defects, which was telling us this must be at the level of proximal signaling. Something was happening way upstream that PMA and I and I'm Ison were bypassing.

44:11So, to understand this, we had to do some signaling analysis. And what we found was that indeed we saw repressed activation, sorry, we saw the upstream signaling was totally fine, very, very proximal signaling, phospholase decay, ZAP70 and PLC gamma, but downstream that kind of distal signaling from the TCR AKT, the activation of C-June, these things were different. So, in other words, this upstream signaling was totally fine, that kind of proximal signaling that happens immediately downstream of the antigen receptor, but these downstream kind of propagated signals were changing. So, we kind of looked at this schematic, and then we looked at the Kennedy pathway, and we found some commonalities. So, you can see something really necessary to propagate signals to amplify them is a second messenger called diacyl glycerol. It's formed when PLC gamma gets activated, and diacyl glycerol becomes active, and it is sufficient to activate RAS and PKC.

45:13But what's really intriguing is that DAG also exists here. It is a reactant in the generation of this membrane phospholipid. So, this suggested that maybe by feeding phospholethanolamine into T cells, by exposing them to these heightened levels, what we're actually doing is depleting diacyl glycerol from the T cells. And that's exactly what we found, is that treatment of cells with phospholethanolamine actually cause those cells to deplete all of their diacyl glycerols, at least to biochemically active ones. We could actually measure this using an antibody to DAG as well. You can see these purple rings in the control. These are the diacyl glycerols at the membranes, and you can see that it's profoundly decreased when the cells are exposed to phospholethanolamine. We could also see this happen by flow cytometry, and we could also show that this happens in cancer and in tumor infiltrating T cells. So, this was particularly exciting to us, because it told us that it was flux through this Kennedy pathway.

46:15From feeding this metabolites, it actually, again, intersected with the immune activation ability by depleting a critical second messenger that transmits T cell receptor signals. So, can we actually manipulate this in vivo? Remember I told you earlier, PCYT2 is frequently silenced in tumor cells? So, what we did is we just overexpressed it. We put it back into B16 melanoma, again, our workhorse cell line. And indeed, if we overexpressed PCYT2, the tumors are controlled by the immune system, which is really cool to see. And it is associated with decreases in phospholethanolamine within that tumor and herstetial fluid. The T cells that we isolate from those tumors that are antigen specific, they have much higher levels of diosolglissorol, and they are able to elaborate more cytokines when we reactivate them. More important, most importantly, this is not due to something else going on. If we implant the overexpressors into immune deficient animals, they grow at the exact same rates, telling us this was indeed due to control by the immune system and not through some other feature.

47:25So, to conclude this second part of the talk, media based off of the interstitial fluid of tumors allows us to have a direct view of what the metabolites present in the tumor microenvironment have on T cell function and in fate. Tiff media induces persistent and heritable dysfunction in CDAT cells, and it confirms the importance of arginine, but also identified novel, alkyl metabolites, including phospholethanolamine, and that can induce a form of T cell dysfunction driven by amplified or one might even say pathologic flux that depletes critical second messengers like DAGs. We can overexpress this molecule on tumors and it disrupts this metabolism and promotes immunity. And I think the overall take home of this is that the metabolism of tumors is sufficient to induce T cell dysfunction. And I just wanted to highlight this is kind of, you know, there's a lot more to do here, but this is just doing that same exact absolute quantitation metabolomics and looking at two different kinds of cancer, PDAQ or melanoma.

48:31And you can see there's some things that are shared and there's some things that are not and we're very interested in exploring how this is going to look in the future. So with that, I'd like to thank the people that did the work. This is my microenvironments. This is really driven by a number really key people in the laboratory, especially Drew, you paying Kelly, Hannah and Nicole. And thank you for your attention. Happy to take questions with the remaining time. Thank you so much. That was an amazing talk. We share again. Yes, so I'm happy to move on to the one a section for the rest of the bedroom in our session. And it seems that we already have a couple of questions there. So let's get started. So first of all, could you maybe elaborate on the initial at the start of your talk to talked about the T cells, like the genus and the exhaustion of the T cells. And so there's a question of like trying to compare it to progenitor exhaust exhaust exhaust T cells versus terminally exhausted T cells.

49:37So it would be great if you could elaborate a little bit on that. Yeah, so the progenitor exhaust T cells that mean so they're named T packs, their name progenitor exhausted because they express PD one. But the fact is these are definitely the more memory long live like cells. So we start to see a few things happen at the progenitor exhaust stage, but it's very, very minimal. These cells are still very metabolically sufficient. It's really the terminally exhausted cells, those ones, those are ones that really experience that very profound metabolic dysfunction. So that's really kind of when I talk about exhausted T cells, I'm most always referring to those very terminally differentiated exhaust cells as opposed to the ones that are more the T packs. They express a lot of things that are very that promotes themness, but also a lot of metabolic genes that help them reengage that proliferate burst. Great, thank you so much. There is next question a little bit longer. Could restoring or enhancing mitochondrial fitness in dysfunctional or exhausted cells within the hypoxic and metabolically stressful tumor micro environment have unintended consequences.

50:49Specifically could excessive or disregulated mitochondrial activity alter T cell activation differentiation, metabolic state or loss production in ways that compromise T cell function promote excessive immune activation reduce functional control or potentially increase off target damage to healthy issues. And how can mitochondrial function be restored well preserved, preserving pieces of the appropriate activation thresholds and overall safety. I think that these are all important questions, you have to remember that that when we are improving mitochondrial function, all we're doing is putting gas in the tank, okay, that's very clear to under so that analogy works here. We're not taking breaks off, we're not hitting accelerators, we're just allowing the cell to regulate itself outside of that outside of the fuel in the tank. And so we have never observed over activation, you know, we get a little bit of changes in function, but even in systems where we use females females when they're hyper activated will induce Vidal I go in mice because they can kill normal melanocytes.

51:50We never see Vidal I go in our experiments, for instance, so and that's because there's still all the regular checks and balances that are still at work in that T cell. So when we, you know, restore that the key, we're restoring mitochondrial activity to cells that have lost it, they just self regulate again, right, and they're also regulated by all the other cells. So I'd like to have the problem of over activation, you know, that these cells are working so well that we that we cure everyone, but but right now that's not exactly what happens. The other thing that's really important to is that when we are when we're changing mitochondrial fitness and by the mechanisms that we're using, it's more about oxygen tension changes or potentially putting in PGC and alpha. Heart and parcel of generating new mitochondria is not just hitting hitting the accelerator, we're filling up the tank. In other words, all the antioxidant proteins come along for the ride. So when you as you're making new mitochondria, they're also buffering their own Ross. And so we don't see if anything when we, for instance, improve mitochondrial fitness through, for instance, like PGC and alpha, those T cells actually have reduced Ross, not elevated Ross.

52:56Great. Thank you. Yeah, I think the next question is also is, I think you've already answered that. So we'll move on to some of the other questions focusing on other topics. What is the role of glutamine limitation in parking so this function and the tumor micro environment? Yes, so this is so glutamine is definitely in competition here and glutamine can play some other roles. There's some beautiful work from Jeff Wrathmel, Bob Leoni and Jonathan Powell that have really explored how glutamine can alter. The competition for glutamine can alter cells as tumors, almost get the lion share of a lot of glutamine. Glutamine is really critical in, in both energy metabolism as well as in the, as an epigenic regulation. Some of the features I talked about today, like those demethylation reactions that we see occurring in on histone proteins and on DNA. Oxygen is one cofactor, iron is another, an alpha-ketaglutarate, which is the direct byproduct of glutamine metabolism is also a critical cofactor. So all, so glutamine deprivation can also have epigenetic consequences on T cells.

54:10Great. Thank you for explaining that. Next question, what is on the inability of carate cells to target solid tumors and if the inability of those cells to target solid tumors is related to the topics you just discussed? Yes, so I mean, this is a great question. I mean, you know, we actually wrote a perspective on this a few years ago about exhaustion and carate cells. I mean, the fact is, is that why carate cells don't work in solid tumors is so multifaceted, it's impossible to pin down, right? There are some targets that, that solid tumors, that are present on solid tumors that are great targets for killing tumors. They also are shared. They're just not safe targets, right? It's just one problem, right? Persistence in forming a memory population is another problem, right? T cells also have to traffic to a solid tumor where they don't have to traffic to a liquid tumor, right? The liquid tumors in the blood or in the bone marrow places where T cells are doing just fine. But indeed, there are some cars that infiltrate solid tumors and get stuck on the hypoxic at the edge, border of the hypoxic region that they can infiltrate.

55:19That's definitely a barrier. It's at the barrier to efficacy in solid tumors. No, I think that there are multi multifactorial barriers, which is for why carate cells don't work in solid tumors. But this is one of them. Great. Okay. Another, first of all, thank you. Excellent work. Thank you for sharing. Do you see any overlap of tumor metabolites driving differential gross levels or argy distinct levels? Yeah, there's absolutely interaction because many of these nutrients can drive Ross production, either on their own or through changes to glue to toilet, glue to thiamine tablets and et cetera. We have not looked systematically. Let me just put it that way, but it is something that we're currently interested in exploring. Next question. What's exhaustive T cell viability significantly different among our PMI, CIFM and PTN, or was it fairly similar? Interestingly enough, the phenomenon that we saw in the TIF media was not exhaustion. We actually looked very, very, we, we transcriptomically, functionally, epigenically profile the T cells that were under in the TIF media or in phosphathenolamine.

56:34While they were dysfunctional, they were not canonically exhausted. They don't upregulate talks, PD1, TEM3, any of those things. They actually are dysfunctional because of this unique mechanism rather than being intrinsically unable to make cytokines. And that was really born out of that PMI and I'm isen kind of comparison, right? Is that triggering through the T cell receptor didn't produce cytokines triggering using PMI and I'm isen we did get cytokines, which means those cells were still competent to make cytokines. This is a distinct form of T cell dysfunction different than exhaustion. But to answer your question, the viability was the same. We did not see differences in viability. We did see differences in expansion in the absence of Argentine in that TIF media, but that, but they, they, they just stalled. They stopped their cytostatic, but they didn't die. Great. Thank you. Okay. I'm going to select one last question. We're running out of time. There's still several ones to go, but just the last one. Can T-reg exhaustion be at potential therapeutic strategy?

57:37Ah, great question. T-regs are my second favorite cell, maybe my first favorite cell, tough to say. That's T-regs are incredibly fascinating. But one thing that was that and this something that we've actually directly studied, we've, we've asked T-reg cells to suppress continuously. We take T-regs, we put them into a suppression assay for 72 hours, we remove them and then we give them new targets to suppress. And we've done this continuously for weeks, even from mice. And we've shown other people have shown this in the GVHD field as well in the kind of transplant setting. T-regs are inexhaustible. Right. So they, I think their phenotype is part of exhaustion because they see self-anogen all the time. So they kind of are seeing chronic angina all the time. So I would say my, my, my, my take on this is that we cannot exhaust T-regs. I think they are exhausted. They have an exhausted kind of signature to them. And so I don't think it's an ideal therapeutic strategy, but targeting T-regs remains one of those interesting ways of kind of overcoming some of these barriers.

58:43Great. Thank you so much. This was a really wonderful presentation. I was very excited to any session. And with that, I would like to, yeah, and look forward to Capy for the recruiting remarks. Yeah. Thanks Sarah and thank you Dr. Delco for your important important findings. And please note that creative bowel loves regularly invites estimated experts in life science and bowel technology to share their latest findings in our business. So we encourage you to visit our website and subscribe to our social media. So you won't miss any upcoming events. And that's for today. Thank you for joining us. And we will see you next time.