Dr. Teso. And we're excited to welcome our expert presenters to discuss the use of human pluripotent stem cells for the treatment of type 1 diabetes. I'd like to let my co-moderator introduce herself. Good morning. Thank you so much. We are very honored to have Dr. Melton and Dr. Chen here today to talk about human pluripotent stem cells for the treatment of type 1 diabetes. Fantastic. So here's a glance at today's agenda. We're going to provide a few announcements and then we will introduce our experts in a few moments. They'll each present lectures and will be followed by a panel discussion. Jeff, we can all hear you. Doug, can you hear me? Yes, we can hear you now. Okay. So, the presenters will be taking questions from the audience at the end of the event. So, during the event, please don't hesitate to wait until the end of the session to send in your questions. If you could go ahead and type it in the Q&A box, which is shown where that's at in the slide. That's in your control panel. Please be sure to use the Q&A box, not the chat function for the questions. We will be able to use the chat box to send you important links during the announcement segment. So, we want to keep the questions in the Q&A so we can have announcements in the chat box. So, I'm the chair of the Islet Biology Development and Function Interest Group Leadership Team for the ADA, the team who has coordinated this webinar. And I want to take a moment to thank all the members of the leadership team, including Dr. Wang, for their work throughout the year to provide opportunities to the interest group members. If you are not yet a professional member of the ADA, we encourage you to join today. ADA professional members can join three interest groups, including the Islet Biology Development and Functional Group, as well as enjoying member-exclusive webinars, webinar recordings. And additionally, recognition opportunities and volunteer leadership experiences are available. And if you will go to the chat, you can find a link to learn more about the ADA membership. So, another benefit of the ADA membership is connecting with members of the interest group on the Diabetes Pro Membership Forum. And if you go to the chat, you can find a link for that. Finally, here's a preview of upcoming live webinars that are hosted by the ADA, similar to the one that you are attending now. To register for upcoming webinars, please visit the link on your screen and in the chat. So, I'd like to introduce today's first presenter. Dr. Doug Melton is a professor and Distinguished Research Fellow at Vertex Cell and Gene Therapies. And at this point, I'll turn the time over to Dr. Melton. Okay. Thank you, Dr. Wang and Tessim, for inviting me to present today. I'm especially delighted to be joined by my colleague and friend, Dr. Shubing Chen. I will share my screen and then we can confirm that we're seeing my slides. Yes. Great. So, do we now see my slides in the correct view? Let's see here. Great. Does that work? Yes. Perfect. Great. Thank you. Again, thanks for the invitation. Thanks to the ADA. As we said, my name is Doug Melton and I'm working at Vertex Pharmaceuticals while I'm on leave from my professorship at Harvard University. And I'm pleased today to talk to you about this idea, which is perhaps obvious to many on the Zoom. Namely, can we use cells as medicines to treat type 1 diabetes instead of injecting insulin? And I'll make a few general comments to begin. I'll describe some results about making the cells. I'll describe some results with the first few patients in a clinical trial. And at the end of my talk, I'm going to talk about, let's say, a prejudicial view of where I think the field can and should go. And I think you'll find that my comments complement those of Dr. Chen, who will be talking more broadly about the use of these cells in a number of venues. So let's move here. Now, why are my slides not advancing? Let's try this. There we go. I don't think this audience needs to be reminded that for more than 100 years, people have been injecting insulin using pumps and monitors. And I show this by way of sort of reminding us all of something that has troubled me, that's come to my desk, let's say, in recent years. Which is that people who are the payers for diabetes care, the insurance companies, and even some at the FDA, seem to think that injecting insulin and having glucose monitors helping you control it is adequate for the treatment of diabetes. But if you talk to any patient, they say more or less that it's a much bigger pain in the neck. It's a daily monitoring. It's an onerous disease to deal with. And that's what we're all trying to eliminate. In other words, not just to provide insulin, but to provide it in a way that the person can lead, I would say, a higher quality of life by not having to daily, if not hourly, manage and monitor their diabetes. So some will know that my lab started on what I would say is an obvious idea more than a couple decades ago. The idea is to say that if patients are missing a cell, in this case the pancreatic beta cell, it's unlikely that gene therapy is going to do anything about that because type 1 diabetes is a multigenic disease. Those genes interacting with unknown factors in the environment, that all comes crashing down on the now famous beta cell, which makes insulin. So we set out to do what I would say is a kind of obvious thing. It took much longer than I thought it would or that anyone would like. But we were able to develop a protocol that moved embryonic stem cells or induced pluripotent stem cells through stages shown here in this diagram, which is to first instruct them to become definitive endoderm, then primitive gut tube, then one stage of pancreatic progenitors, PP1, then another PP2, then they become endocrine, and finally they become glucose responsive insulin secreting beta cells. And that's shown here in the purple cell on the right. Maybe my pointer works. There we see them. And since this has been published, and we've published lots of characterization, I'm not going to show you a lot of the data, except to say here with this histological picture, that if you compare the stem cell derived clusters we make, here just the abbreviated SC-beta, staining them for insulin in the cytoplasm and the transcription factor PDX1 in the nucleus, you can see at this level of resolution that they are indistinguishable from cadaveric or organ donor islets. We have so many ways of comparing them that I'm not going to show you, except to just say that by single cell sequencing, by immunohistochemistry, by fact sorting, by transmission electron microscopy, and perhaps most importantly by functional assays in numerous animals and diabetic models, these cells are as good, if not better, than organ donor islets. And I say they may be better because it's not surprising people to know that organ donor islets are, first of all, a mixture of lots of cells. They're still endothelial cells and contaminating exocrine cells, not to mention that those cells have been isolated from a cadaver by enzymatic digestion. So if one is able to make the cells fresh, let's call it, from an embryonic stem cell, it's perhaps not surprising that those cells are more robust and likely to survive longer. So I'm going to sort of skip then all of that work that says that we've made the cells and that they work and just say, how would you then move that from a research project in a laboratory, which has, you know, undergraduates, technicians, postdocs, all working, you know, in a professional way, but not in a clinically compliant way. How do we move that to patients? And that poses, I would say, two big problems. One is to improve the protocol for manufacturing and to use clinically compliant cells, and I'll talk about that in a moment. And then I'm going to finish on this problem of immune rejection. So as everyone knows, even if you made the patient's own beta cells, they would be rejected by that patient's immune system because type one patients suffer from an autoimmune disease. So how is this sort of prosecuted, you might say? Well, since these things can't be done really in a research setting, some years ago I started a company called Sema Therapeutics, and then let's think, I should remember, I think that was about 2015. And then in 2019, this biotech company was acquired by Vertex, where I am now working as a research fellow. And so what Vertex has done in a relatively short period of time is to take the protocol, sort of diagrammed in the cartoon there at the top, and make cells that could be used in a clinical trial. So these trials, or this trial has begun, it began more than a year and a half ago, and the trial is the following. The stem cell derived islets are transplanted into the portal vein, along with non-steroidal immunosuppressants. The immunosuppressant protocol is more or less identical to the one established by James Shapiro and his colleagues in Canada for the transplantation, I might say the successful transplantation of cadaveric islets. Now, it might be a little confusing that the cartoon shows that they're making pure beta cells, and then beneath it, it says clinical trial with stem cell derived islets. I should clarify that point. The cartoon implies that we make 100% beta cells, which is not true. We make islet-like structures, which are about 40-45% beta cells, about an equal number of alpha cells, which express the counter-regulatory hormone, glucagon, and that's why we call them stem cell derived islets. And I'm pleased to say that I'm able to show you the results from two of the first patients. There are more patients than two, but I've learned here at Vertex that there's a certain time at which we can release these data. I want to draw your attention to patient one on the left-hand side of this slide. This patient received a half of the calculated dose of stem cell derived islets. I'm sorry for the nomenclature, but the stem cell derived islets are called VX880. What I'd like you to see here is on the left where it says baseline, that refers to this patient's blood glucose control before transplantation. So the study days here at the bottom are minus five to minus 25, so say about a month before transplantation. You may not be able to see it because it's sort of small on the screen, but the y-axis are the blood glucose levels in milligrams per deciliter, and as people will appreciate that the range target here is between 70 and 180 milligrams per deciliter, and those are the green dots. I want to draw your attention to how many dots from this person's continuous glucose monitor were outside of range. Here in the dark blue and the light blue, and importantly, here in the bottom, which I can barely see in kind of orange and red, which are severe hypoglycemic events. So you can see this person had hypoglycemia many times in the month before transplantation. This patient received half a dose, as I said, and now you see what I consider to be a nearly perfect result. His time in range is now 99.9%. He was injecting 34 units of insulin per day and had a hemoglobin A1c of 8.6. He now injects no insulin and his hemoglobin A1c is 5.2. I have to say that this result, I don't know if I have to say it, I might say this result was better than I could have hoped for because it was half a dose, and it's the first patient that had received the cells. So let's look at that as a positive result. Now look to the right at the second patient, which is also positive, but not nearly as extensive. You can see a significant improvement in the patient's blood glucose control, but the patient is not insulin independent and it didn't work completely. So we don't know if this is a dosage issue or difference between patients. There are lots of reasons this could be different. It's done at a different site by a different surgical team. So there are two ways to look at this. I guess it's the classic, is a glass half full or half empty? I tend to look at it as being more than half full, and that the patient one shows that this can work just the way we wanted, and we have to find out why it didn't work in patient two. I'm now going to move ahead with the conclusion that everyone may not agree with, but my conclusion is going to be that stem cell derived islets work. So let's assume that they work. We can make plenty of them. We can transplant them into the liver along with immunosuppressants. What's next? There are two things I'm focusing on now, working here at Virtex. One is to gain complete mastery over cell composition, and two, to figure out ways to eliminate the need for systemic immune suppression. I'm going to describe experiments that were done with a research line. These are not the cells that Virtex uses for patients. So the first thing is, can we gain complete mastery? This TSNE plot done by single cell sequencing shows more or less in this research line the kind of cells that are made at the end of the stage, which I refer to as stem cell derived islets. So obviously, the cells we were most interested in were these purple cells, the stem cell derived beta cells. And we were happy to see all of these stem cell derived alpha cells here in red. What was a big surprise to us were the significant number of these blue cells here, enterochromaffin cells, which don't normally exist in adult pancreatic islets. There's evidence that they're found during fetal development, but they're not a cell type that we find, as I say, in cadaveric iodes. And not so surprisingly, there were some other cells here in brown and green that sort of didn't get with the program, didn't get the message, and become fully differentiated yet. On the whole, the purple and the red cells are the ones we want. So the question is, how do we improve the proportion of those? And here at Vertex, by adjusting the media, they have more or less eliminated all the green and the brown and the blue cells and have pretty much pure red and purple cells. But I'm going to describe experience for you with this research line where that's not the case. The proportions are shown here at the bottom. So this is our goal. We want to control the composition. I want to make beta, alpha, and somatostatin cells. I didn't show you those in the previous slide, but there are very few somatostatin cells. And not make any of these enterochromaffin cells or non-endocrine cells. And then my little cartoon here, which I'll do again with the Marvel shield-like circles, is supposed to represent some form of protection so that the cells would not be rejected by the immune system. There are many ways one could approach this problem. As I said, one could go back to the protocol and change the media. In terms of protection, Vertex, under the leadership of Chris Thanos, has developed a proprietary encapsulation device, which will go into a clinical trial. But I'm going to talk about genetic modification to try to control the composition and provide some immune evasion, if not complete immune tolerance. Now, if there are any immunologists on the call, I know that they're rolling their eyes when I say tolerance, because no one's ever achieved that. But I wouldn't say that there's any formal reason to know it's not possible. And I'll tell you a bit about our approach to that. So this is what we've done. We've adapted the work of others and made use of reagents from others to figure out how, in a large scale, we can knock out one gene at a time in the embryonic stem cell stage, and then look for genes which improve composition or provide immune protection. I'll show you some of the details of this, but just summarize it here by saying we inject or we infect the undifferentiated ES cells with a lentiviral library, which I'll show you in a moment. And we do that at a very low multiplicity of infections, so that only one in three cells, on average, gets a virus. So that's here by this asterisk, that cell, but not these two got the virus. Then we select for puramycin for cells only that have the virus, and this contains guide RNAs and the CRISPR enzymes to knock genes out. And so the nice thing about this is as we let the cells recover and grow them, if we've knocked out housekeeping genes, you know, like DNA polymerase or EF1-alpha or something, that's fine because we don't want to look at those genes anyway. So we're left with non-essential genes for growth, but each cell has one gene knocked out. We then grow those cells and differentiate them, and then we dissociate the clusters, fix the single cells in paraformaldehyde, and stain for markers. The color coding here is the one I showed you in the previous slide, where purple are the beta cells that we want, red are the alpha cells we want, blue are the cells we don't want, and terrachromaffin cells, and green are triple negative, that is, they don't stain for any of the other markers. We undo the fixation, isolate the guide DNA PCR sequence, and then ask about the proportions of cells that have had genes knocked out. So I'm going to show you here the library we use, which was developed at the Broad Institute, really just to make one point. We start, or a couple of points, I guess. We knock out every gene, 19,000 genes in the genome using four guide RNAs per gene. We did that three times to identify about 800 top candidates, and then we repeated that experiment with the 800 top candidates 18 different times. We did 18 differentiations to get statistically significant results where we have 10 guide RNAs per gene, 500 non-targeting controls, 500 intergenic controls, etc. So I would say as confident as one can be about statistically significant results, which of course are not plus minus, we're confident about these answers. And so now I'll show you here what the data looks like. This simply just shows that the experiment worked by knocking out genes already identified by us and others, which were known to be essential. So let's look here at the far left. It's well known that when you knock up pancreatic duodenal homeobox number one, you can't make beta cells. And sure enough, here you see each dot, sorry, each dot represents a single cell or single gene. And here's what the PDX knockout. So it makes you more likely to be negative for triple marker, positive for triple markers and negative for becoming a beta cell. So I'm not going to go through all of these. This is just a different way of presenting it here in this circle diagram. So if you want to make beta cells, you want to not knock out genes like this, you would rather knock out a gene like that. So I'll summarize the results by saying we've identified pathways that I had never thought about as being important for beta cell formation, pathways that allow us to eliminate the other cells. And I'll just mention what those are. It turns out the heparin sulfate pathway is important for making endocrine cells and beta cells. And that's something I'd never even read about. But if we knock out genes important for heparin sulfate synthesis, it reduces the number of the cells we want. Most interestingly to me, when we knock out genes in the ubiquitin pathway, in particular a gene called F-box L14, this protein is thought to target transcription factors for degradation. And when we knock that out, that increases the number of beta cells by a significant effect. So we've now made, I guess we could call them monoclonal cells with these various knockouts, are making monoclonal lines and differentiating them. And we're excited about pursuing a handful of the genes we've identified by this screen. The other thing that's in the title of this screen that I forgot to mention is that, and this won't surprise anyone, that of the 800 genes that affect beta cell or alpha cell differentiation, 24 percent of them, if I remember correctly, are transcription factors. And again, that's not a surprise because we all believe transcription factors are, you know, the master regulators of the genome. Okay, now I'm going to move to the challenge of alloerjection. I don't want to pretend that I've solved the problem of complete control, but I believe, I would say, that we're on the path to being, having complete mastery over the final composition of the stem cell derived product. Now, how do we deal with this annoying immune system? Well, we're doing things in addition to this screen, which I'll just click through here, that I would say are obvious, things that a smart undergraduate, say, would propose. And others have already done. We're knocking out beta-2-microglobulin, so to prevent presentation, expressing more PD-L1, expressing CD47. We're removing, as I said, HLA in some cases, but then adding back fetal HLA presentation as here, thinking about some complement inhibitory proteins. There's nothing novel about this approach. I think the novelty comes from figuring out what is the right combination. And we haven't solved that yet, but we're working on that. And I'm going to finish by showing you one result, if I can click to it, here, that makes me think we're on the right path. So, before talking about this slide, I want to say something else about my prejudice here. What we suffer from in the field is a very good model to know how we would actually have reduced or eliminated the immune attack. So, everyone, of course, and their cousins, uses various immunocompromised mice, and you can inject them with human PBMCs. But these are kind of, I don't know, you might call them circus trick experiments, where you set everything up, but there's no way in which it reproduces what happens in a human type 1 diabetic. Nevertheless, it's a start, and I'll show you that start here, where we're asking, can we make genetic modifications to the stem cell-derived islets that allow for survival in a xenorejection model? We've published on models where we use IL-2R gamma, you know, NSG mice and inject them with human cells, but here I'm just showing you straight up xenorejection in a mouse. So, the cells that we ejected are expressing IL-10, TGF-beta, and a modified IL-2, and what this experiment shows is that improves greatly their survival over a nine-week period, and maybe that's easiest to just see here in the bottom right, where if we put genetically unmodified or wild-type human stem cell-derived islets, and then look at them with luciferase, they're rejected within two weeks. So, you see this little light up here goes away, but the cells that we've modified in the way I just described, sorry, that we've modified in the way I just described, they survived pretty well, not entirely, you can see the numbers go down, but for nine weeks. So, what's the point I'm trying to make? I've finished with this slide by saying transplantation of stem cell-derived with islets with immunosuppression in the portal vein works. It has to be perfected, and we have to complete the clinical trial, but if I could put a green check mark, I would say number one works. Number two is going to be tested this year at Vertex using their encapsulation device, and I finished my presentation by talking about number three, like let's say the dream of a next generation where you would genetically modify the cells so you could put them into a patient without systemic immunosuppression. So, my dream would be that when a young child is diagnosed with type 1 diabetes, the endocrinologist says, you know, I'm sorry for you and your family, but I have good news for you. I have here in my freezer some cells which will control your blood sugars, and your body won't reject them for a long time. I'm going to inject them into your belly, and good luck. Get back to kindergarten, and all will be well. That sounds like a dream, and it is sort of a dream, but I don't see any reason we shouldn't aim for that. So, this is a slightly older pre-COVID picture of my group. I'm very fortunate to work at Harvard with so many talented and energetic young people. These are some of them listed by name, and I note here in black at the bottom the agencies which have supported our work, and I now show here in green that I'm now working as a fellow research fellow at Vertex. So, I will stop sharing and hand this over to my colleague and friend, Shubing Chen. Let's see if that works. Oh, while waiting for a slice from Dr. Chen, thank you so very much, Dr. Melton, for this wonderful presentation, and especially the years of groundbreaking work that bring hopes to type 1 diabetes patients. Many of my clinical trial patients are talking about your trials, and another great contribution I think you are making is that you trained many excellent scientists who are continuously moving the stem cell research forward. Our second speaker, Dr. Shubing Chen, is a great example. So, Dr. Chen got her college degree from Tsinghua University from China. She did her PhD in Scripple Institute and majored in chemistry. Her postdoc training was at Harvard with Dr. Melton we just mentioned, and Dr. Chen joined Cornell as an assistant professor in 2011, and she's currently an associate professor. The major research interest of Dr. Chen is to manipulate stem cell fate using chemical and biological approaches to generate functional tissues and organs that can be used in translational research. I think she really combined her expertise in chemistry and stem cell biology, and is doing an excellent job founded by the NIH and other funding agencies. She's very well published. She's, to me, a rising star in the stem cell field. Without further ado, Dr. Chen, please. Hi, Kongjing, thank you for a very generous introduction, and Doug, thank you for sharing the most updated, exciting results. And you remind me, I was in that picture before. And hello, everyone. It's really my great honor to have the opportunity to present in the ADA and its group. And as Kongjing mentioned, I was trained in Doug's lab, and they remind me 15 years ago at my first group meeting in Melton lab. Let's see how it works. So, since I was trained in Dr. Melton's lab, so I will continue working on human embryonic stem cells and human propotent stem cells. And we still work on pancreatic beta cells and pancreatic, as Doug mentioned, there are multiple different endocrine cells. And if we go together, now we can give a little bit fancier name, like pancreatic endocrine organoids. And in the meanwhile, we also branch out from pancreatic bimage. And we, by our own effort and collaborate with many expertise in the field, we actually expand the platform to around 14 different types of cells and organs. And then we can use this to study infectious disease, which we won't talk about today. But the second part is still diabetes. So, Doug show you very exciting data. And he really saw us the best example to focus the career to push the pancreatic beta cell, pancreatic endocrine cell to clinical study. So, after I got independent, I switched the gear a little bit to focus on understand how to study how genetic and environmental factor or non-genetic factor contribute to diabetes progression and how to establish a diabetes in dish model so that we can use it for disease modeling and drug screening. For genetic factor, we know that more than 100 genetic locus has been associated with type 1, type 2 diabetes. And for non-genetic factor, we know that aging, obesity are very important for type 2 diabetes and virus and environmental toxins that has been associated with some type 1 diabetes progression. So, of course, when we want to set up a model, we always need to start from some cell types, right? And both type 1 and type 2 diabetes, very complicated disease. And in type 1, beta cell got attacked by the immunosystem. And in type 2 diabetes, muscle adipocyte liver develop resistant. And then the beta cell try to compensate for that. They got exhausted and then died. So, in that scenario, both type 1 beta cells play a very important role for both type 1 and type 2 diabetes. And so, we decide to continue to pursue this cell type for disease modeling. So, this is a platform we established in the last several years in my lab. So, we use CRISPR-based gene editing. We can either knockout diabetes-associated genes or knockout diabetes-associated single nucleotide polymorphism. And we can also knockout the regulatory region. We create like, we call it isogenic lines. And then we differentiate into pancreatic beta-like cells. We can monitor their functional survival in vitro. And then we can challenge them with disease condition, monitor their functional survival in vitro. And we can even transplant them into mice. We create this, how is humanized mice model, and use this to challenge their survival function in vivo. And we can also add environmental factors, like different, like lipotoxic, to study lipotoxicity, glucotoxicity, and even like virus stimulation, infection. And then we use that to study the genetic network controlling diabetes progression and use it to drug screening. And I will show you some example. We really start on this work when I start building my lab at Cornell. And at that time, we have very two talent postdocs who ask the question is whether isogenic platform will be sensitive enough to study type 2 diabetes-associated genes. And at that time, we knockout three genes, CDK01, KCNG11, and KCNQ1, which are identified from the first wave of GWAS study and has been confirmed in several round of follow-up studies. And we also know these three genes are highly expressed in both beta cell and human ASL-derived beta cell. So, we use CRISPR-based gene editing to knockout to create these isogenic lines. And the first thing we did is to monitor their function. And one major function of pancreatic beta cell is to respond to glucose stimulation. And for here, we see in the VATAP cells, in the low glucose condition, we detect relatively low human C-peptide expression. And in the high glucose condition, we detect relatively high C-peptide expression. And for the CDK01, KCNQ1, or KCNG11 knockout lines, we actually see they lose the response to glucose stimulation. And it's worth to point out that when we do the isogenic system to perform disease modeling, we do see a line to clone-to-clone variation. So, it's always important to include multiple clones when we perform this kind of study. And in addition to monitor their function, we also monitor their survival. This is a very busy slide. We basically challenge the VATAP cells and CDK01 cells with high glucose to mimic glucotoxicity or high palmitate to mimic lipotoxicity. I will directly go to the summary slides. So, in the control condition, actually, we don't really see major difference between VATAP and CDK01 knockout cells. However, if we challenge them with glucose or high glucose or palmitate condition, we can see that the CDK01 knockout cells cannot handle the stress very well. So, they die more. So, after we established this platform, everyone in the lab very excited. So, they want to choose their favorite gene. So, that for me, a previous graduate student in the lab. So, she find this literature and they summarize genes associated with type 1, type 2 diabetes, as well as monogenic diabetes. I understand there are different versions of this chapter, but this is the one we use as a reference, which highlight that GLIS3 is one of the two genes associated with both type 1, type 2, and monogenic diabetes. And what is GLIS3? GLIS3 is a zinc finger transcription factor. It can function as both repressor and activator, and it's strongly expressed in pancreas, kidney, and thyroid. So, Sadaf's first knockout, the GLIS3, makes knockout lines and then differentiates into pancreatic beta-like cells, and she monitors the beta cell survival. And in the GLP-positive cell, in the GLIS3 knockout cells, we see clearly increase of cell apoptosis, which is confirmed by an XM5 staining. And at the bottom, she further confirm it by the PI staining, as well as the CLIBK3 staining. So, next, she transplanted these cells into the immunodeficient mice, and we created this humanized mice model. And then seven days later, we take the xenograft out, and we find that in the xenograft carrying the biotype cells, we see insulin-positive cells with very little CLIBK3 staining. And for the knockout cells, we see a significant increase of CLIBK3-positive cells with a decrease of GLP-positive, insulin-positive cells. And I'll come back later why we want to create this humanized mice model. So, of course, knockout line data is very exciting, but then genomic people will ask that we usually identify single nucleotide polymorphism instead of completely loss of function mutations. So, whether the system will be sensitive enough to study SNPs. So, GLIS3, then we go through the SNPs associated with… GLIS3 SNPs associated with type 2 diabetes, and we find very interesting it's located in a very small 2KB region in the second intron of GLIS3. So, we first decide to… we go monitor its research and decide to focus on one SNPs, which is associated with type 2 diabetes, and we call it RS200. And when we go to study the SNPs, the first thing we need to do is to sequence what we call so-called wild-type cells, because many wild-type cells… the definition for SNPs is at least 1% of population should carry the risk allele. So, in essence, a lot of… we so-called wild-type cells are not necessary to be homozygous for non-risk allele. But in this case, we are pretty lucky. The wild-type cells are homozygous for the non-risk allele, so we do a point mutation to change it to risk allele A. And then, Zeping, another postdoc, pick up the project and then differentiate the cells carrying the non-risk allele CC to risk… and the isogenic line carry either non-risk allele CC and carry the risk allele AA into pancreatic beta cells, and he monitored the GLIS3 RNA expression, and indeed, we find a significant decrease of GLIS3 RNA expression in the cells carried the AA non-risk… risk allele. And then, he further validated the cell survival, and we find that when we differentiate into insulin GLP-positive cells, and when we monitor the anexin 5-positive cells… anexin 5-positive and DAP-negative cells, and in the cells carried the AA risk allele, he detects a significant increase of the percentage of anexin 5-positive cells, and we further confirm this result with immunostaining risk in phase 3. So now we've shown using isogenic system to show the phenotype or loss of function limitation of Glyph3 cause the increased cell death and as well as the single nucleotide polyphenism. So the next question will be whether we can perform some chemical screening to find some drug candidate that can rescue the phenotype. So we use a Glyph3 human ES knockout line and differentiate into pancreatic beta cells. If we treat it with DMSO control, we can clearly see the cell death. And now we look for the small molecule that can both maintain the number of insulin positive cells and decrease the cell apoptotic risk. So this is one of the lead compound we identified from the screening called galunisertine, which is a TGL beta 2 inhibitor, which can decrease the loss of Glyph3-caused cell death through a dose-dependent manner, the SC50 around 1 micromolar. Here is some primary screening data. In the DMSO treated condition, we see a very high cell death. And in the glutisertine treated condition, we see a significant decrease of cell death as well as cell apoptosis. So here I'll come back to why we want to spend time to create these humanized mice. So we want to have a system we can monitor the drug activity on human cells in vivo before we move to a clinical study. So we transplant this Glyph3 knockout human ESL-derived beta cells into the kidney capsule of the mice. And then we gave the mice drugs for a certain period of time. And then we take the xenograft out. And if we look at the xenograft treated with vehicles, we still see very strong cell death and see less insulin positive cells. For the xenograft treated with glutisertine, we see a significant decrease of cell apoptosis, both in the whole population as well as the insulin positive cells, which are quantified on the right. So now we can study one genes, right? So now what is the next step? So several years ago, we started a collaboration with Dr. Francis Collin, who has been really pioneered on the type 2 GWAS study. So we work as a group to prioritize around 20 different type 2 diabetes associated genes and create isogenic human ESL lines from all of them. So here is just the list of the genes we choose. It's based on mainly three criteria. One is their significant score in the GWAS study. And the second is whether it's highly expressed in the pancreatic beta cells. And the third one is some previous study whether it's associated with a beta cell biology. So here is a very preliminary study, and it's just a summary. I'll show you some of this data. As you can see, in each genotype, we have two clones, and we have two biotype, and this is the knockout lines. And we find that several genes, like when we knockout TCF7L2, like TNF4 alpha, we see significant decrease of cell differentiation. And we also monitor the cell survival. We challenge them with lipotoxicity condition, and we see a lot of lines. If we knockout the genes, they cannot handle the stress very well, and they show this increase of cell survival. And what the lines label is red color. That's the one that shows statistically significant difference. And we also make this isogenic line using insulin GLP reporter lines, so we purify the GLP-positive cell. We monitor the total insulin content in the GLP-positive cells. And in many lines, we also see a significant decrease of the insulin production. Finally, this is the most labor-intensive experiment, and Dongxiang also monitors the glucose-stimulating insulin secretion of all these knockout lines. And for many of these lines, we see a significant decrease of fold induction of insulin secretion between high glucose and low glucose conditions. And then we run the RNA-seq and ATAC-seq of all these knockout lines, and currently we are working with Collin's lab to perform some studies to identify the key or harbor genes control the beta cell biology. And in the last several minutes, because all this work is still on beta cells, of course beta cell is very important, but type 1 and type 2 diabetes are not only beta cells, and Doug also mentioned in his talk that immunoATAC is very important. So what we are interested in is whether we can set up an in vitro model that we can visualize how immuno cells attack beta cells. This is our very initial trial, and we are very excited about the results. And the cell type we decided to use is T cells isolated from PBMC. So this is a co-culture system, and this is a live movie, and you can see the human ES cell derived insulin GLP beta cells, so the beta cell is in green color, and the T cells are labeled with a red color here, you see it's like a yellow color, and now you can see the T cell actually come to attack the beta cells, and the GLP signal decrease. And then in the control condition without any T cell co-culture, and the cell are very healthy, yeah. So we are, hopefully ideal case is we want to convert it to ES cell derived T cells so that we can study how different genetic factor modulates this T cell media beta cell damage. So maybe the last four minutes, I just want to share with you a new center that we got very lucky to build this year at Weill Cornell, and we call it Center for Genomic Health. So we basically combine three efforts, and this center was really built based on diabetes program, and which we were very lucky to have Dr. Doug Melton as one of our advisory in the scientific advisory board. And so, but now we kind of expanded to the Center for Genomic Health, and the goal is we want to combine three approach, one is biobanking, and the second one is stem cell models, and then we go to drug discovery. And this is the example to show you how we do the biobanking, we have, this is our diabetes program, we have T1D, T2D biobankings, and collaborate with our physician and the surgeons. We also have a large effort on the bioethics surgery samples, and then we do tissue collection, and for some samples we do iPS cell generation, and we do iPS cell differentiation, and for some sample we can do organoid banking, and we are very lucky to work with Illumina, so we actually will do whole genome sequencing for 500 type 1 and type 2 diabetes patients, so that we can correlate their clinical samples to their whole genome sequencing data to their isogenic, to their iPS cell and iPS cell derailed cells for disease modeling. So the second pipeline we built is, we call it isogenic platform, so that's, I show you a lot of preliminary data, previous data here, so now we kind of build it as an industry-style pipeline, and we are open for collaboration. So we basically choose the genes we are interested in, and then focus on the gene to knock out, or SNP to knock in, and then we create the isogenic system, differentiate them, monitor their function and survival in vitro, and monitor their function and survival in vivo, and can go to mechanism study, and then finally, if any gene which is interesting, we can go to drug screening. And last is our drug screening platform, because I was trained as a chemist, so we have a local drug repurposing screening effort, and we have several pipelines, and we are also very lucky to get a lot of support from NIDDK to identify the, perform the pilot screening and drug repurposing screening to identify small molecules that can promote beta cell proliferation function and survival. And this is our platform, we can do primary screening, go to heat confirmation, dose curve, in vitro confirmation, I say, we work with our drug TDI to do the medicinal chemistry, or outsource this service, and we can also do in vivo confirmation, it means a humanized mice model, a mechanism study, and if it's a good compound, we'll go to IP protection. So, I think I will stop here, and it's really my great honor to share with you some of our work on the, using this isogenic human ES cells, or human IPA cells to study different diseases, and we are looking forward to collaborate with people in the, from ADA group. And I want to thank people from my lab, as well as our collaborators, and the funding agency support our work. Thank you for your attention. Thank you so much, Dr. Chen, for your wonderful presentation. Now we start the question and answer section. Jeff, I can read the first question, if you want. I think you have to share a lot of secret today. The first question is from Albert Hua. Can you, Doc, can you please comment on how dose was determined in relationship to patient weight, insulin requirement, cell product characteristics, such as cell number, insulin content, GSIS? Sure. The patient dose, when I said half a dose, was a calculation based on what has been learned from years of cadaveric islet transplants. So, the short version of the answer is that all the things Albert mentioned are relevant, but it all comes down to really just simple counting the number of beta cells. A non-diabetic patient, as you probably know, has about 1 billion beta cells, and so based on that and what's known from cadaveric islet transplant, the dose was adjusted to be a little less than half of that amount. So, maybe I stated otherwise, that the potency assays and everything all come down to measuring how many beta cells are in the final drug product. The nice thing about this approach, as I'm sure everyone knows, is unlike a regular drug where dosage really matters, it's my opinion that you could put two or five times as many cells in, and that'll just be a good thing because the cells auto-regulate. Okay, in follow-up with this question, if half doses, you already achieved a beautiful outcome with half dose, are you going to still move on to the full dose? That's my additional question. Yeah, they will, and remember the second patient did not respond as well. So, we don't understand if it's unlikely to be the drug product that varies. It's more like differences in the patient. So, I think in the next few years, the goal should be to find a dose which would, say, achieve insulin independence in more than 90 percent of the patients. There are going to be patients where it won't work, either because their BMI or their insulin resistance or who knows what other medical complications, but as everyone appreciates, since it's an allogenic product, it's not an autologous product, we have to find a dose that works for most people. Doug, in regards to the same question, Jessica Pereira asks, can the immune components be the explanation for the differences between patient one and patient two, and do you know if there's any differences in the T-cell populations between those patients? Well, there's certainly differences in T-cell populations because that's true for every individual on the planet, right? And whether we would assign that as the reason for the difference, it could be a whole host of things. The main reason I show that is that to remind us that every patient's going to be a little different, and that we have no idea why patient two didn't respond as well. Could be the T-cells, as this person implies, it could be their insulin requirements, could be their behavior, could be all sorts of things. That's what a clinical trial is for, of course. I mean, I don't mean to be pedantic about it, but we've got to learn which patients will respond to a certain dose and then move towards making a generic product from that. I have another interesting question from Azette. For stem cell therapy, which one is better, RPSC islets with all type of endocrine cells or RPS cells with pure beta cells? Well, I don't know what Shubing thinks. I think that a mixture of endocrine cells is better. Having the counter-regulatory hormone there is good, and I kind of touched on the fact that most studies show that the delta cells, the somatostatin cells, improve blood glucose control, in mice anyway. But let me ask Shubing, we used to joke that we would want to make 100 beta cells, but we don't say that anymore. What do you think, Shubing? I also agree that a mix of endocrine cells will be better than the pure beta cells, but I'm actually also curious, how do you think, do you want to make each individual cell type and mix them together, or you think it's better to take advantage of spontaneous differentiation than we just use what we learned about? That's a good question. That's a good question, Shubing. I think it's the latter, that you would make a mix. It turns out there are not good methods for enriching cells at the scale you would need. So as you've used and would well appreciate, you can use like magnetic beads and miltenne and this kind of stuff to enrich cells. You can never do that at a scale where you would use to make a drug product. So we have to make the product as we want it. Dr. Chen, I have a question from Dimitri Spiropoulos. He asks, in addition to gene knockouts and SNP knock-ins, are you epigenetically manipulating the HPSCs to investigate the environment or the origin of the disease? Yeah, I think that's a very interesting question, and recently there are some approach people already develop. It's still mainly based on the CRISPR-based gene editing, but just adding different enzymes, so that can induce either methylation, acetylation, or certain regulatory regime. I think that will be an interesting direction to pursue if you're interested in any epigenetic changes at certain locus. Okay, our next question is from Lu Xin Ke. The question is, the last stage of differentiation, it takes about seven to 21 days. Do you think, sorry, the question just popped away, Jeff, can you see it? The question jumped out somehow. Yeah, I can see it. The question was, the last stage of differentiation takes seven to 21 days. Do you think that the SC beta cells are mature enough to respond to glucose challenge after seven days in the last stage, or can they be mature enough to respond to glucose challenge after seven days in the last stage, or can they be mature enough to respond to glucose challenge to glucose challenge after seven days in the last stage, or can they stay longer in the last stage, or can they stay longer than 21 days? Does that help maturity? Does that help them function better? Okay, if I understand the question, at the last stage, the cells respond to glucose and secrete insulin, and they'll do that in a sequential factor. So, they don't keep secreting insulin, they turn on and off the way normal native beta cells do. But I think the question was, if you have them in culture, how long will they continue to do that? I think we've kept them more than a month and less than two months in a simple media. So, they're surprisingly stable. They're much more stable than cadaveric islets because, of course, they're pure in a sense. But implied in the question might have also been is, are they now at their final stage? And the answer would be no. When they're put in an animal, and I'm guessing in a human, they kind of get better and better. They make more insulin messenger RNA, more insulin protein, but they don't change their fate. And that, I guess we as a community call engraftment, right? They get contact with blood vessels and the nutrients and the circadian rhythm of the host. So, they get better and better. So, I'm probably as guilty as anyone in talking about making mature or fully differentiated cells. But I then am reminded that we should ask ourselves, when are we fully mature? Like, I tend to think Xubing is fully mature, but maybe if I meet her 10 years from now, she'll say she's even more mature. So, I think probably the same thing happens to cells. In other words, maturity is probably a funny term to use. So, the next question, can you cryopreserve your stem cell islets or do you differentiate for each patient individually? They are cryopreserved. So, they are cryopreserved at a stem cell stage or at fully differentiated stage? At the end stage. Perfect. Thank you. And that's something I had nothing to do with. That's all developed at Sema Therapeutics and Vertex. I think we have time for one last question. And then any of the questions that were not answered will be sent to our two panelists and they can potentially answer by email. This is from Leslie Bayer asking, what is the role of the TPH positive entrochromatin cells in islet development? And do we see these cells during in vivo development? I can't answer that question with complete confidence, but I would say they have no role. They are very similar in their genetic pathway to beta cells. And as I said, there's a report from others, I'm sorry, I can't remember who published it, that they appear during human fetal development. But there is no convincing evidence that they are found in adult pancreatic islets. Most enterochromatin cells are in the gut. Wonderful. Well, thank you both so much for taking the time to present this information to us. This was a wonderful set of seminars. And we want to also thank all those who have joined us online for this webinar. Please watch for other webinars from the ADA, from the different interest groups, and from the, from our islet interest group as well. Thank you so much. Thanks, everybody. Thanks.

human pluripotent stem cells

type 1 diabetes

stem cell-derived islets

clinical trial

immunosuppressants

gene editing

diabetes progression

environmental factors

drug screening platform