 Can everyone hear me? Can everyone hear me? Hello, hello, hello? Can we get started? So is this on? OK, thank you. Let's get started. So I'd like to introduce this international Steenbach lecture ship by first acknowledging the importance of Harry Steenbach, who is not only important to this lecture ship, but the entire university. Steenbach was born in rural Wisconsin and then went through all of his training at the University of Wisconsin in the Department of Biochemistry, which at the time was called agricultural chemistry. And so he was really here the entire time that he was in training. He then became a professor here and continued here. So he is really very much part of who we are here. Very early on in his career as an assistant professor, he discovered the process whereby if you irradiate food that stimulates vitamin D. And then he had the foresight and vision to want to patent that, which wasn't really done at that time. And nobody wanted to patent it for him. So he took $300 out of his own pocket and went and patented it. Soon thereafter, Quaker Oates tried to buy the rights to his technology for $10 million, which you think about this is 100 years ago. That was a lot of money. But he said, no, he didn't want to do that. Instead, he wanted to start tech transfer here. And he and eight other faculty decided to start at Worf. And so the starting budget for Worf was $900, $100 from the pockets of each of these individual faculty. And from there, we have Worf, which has supported the university and continues to support the university today. So we have a lot to thank Harry Steenbach about. I think that story is different for Hans Cleaver's, but there are some similarities. So Hans Cleaver's, who I have the real pleasure of introducing today, has really spent most of his life and training and professional time in the Netherlands. He was born in a small town in the Netherlands. Then he went to Utrecht University. He went through the undergrad. He got his MD, his PhD, all fast and furious. He then became an assistant professor after a short postdoc in Boston. He became an assistant professor. And then the rise to fame was really impressive. In two years of being an assistant professor, he was chair of the department. It's kind of amazing. Then he went on from being chair of the department to director of the Institute. Then he went on from director of the Institute to president of the Netherlands Academy of Sciences. And then on to the director of a completely new hospital system. He's an MD and a PhD. And one of the things he's done throughout his career is bring together his knowledge of very basic science and his talent for it with his deep interest in coming to therapies as well. So in all of these, he's had, of course, many leadership roles and a list of honors, which of course I'm not going to go through. In 2013, he won the very first breakthrough prize in life sciences. So it was very clear that he was at the forefront of biomedical sciences at that time. And then he was elected to the National Academy of Sciences in the US. And that's no mean feat for people from outside the US. So he comes to us with many honors, a tremendous track record, and really cool seminar. I've seen him give to seminar before. And I'm sure it's going to be different today. But I'm really looking forward to hearing about what he's going to say. And so thank you, Hans, for coming to Madison, even though it was difficult to get here. For sure. Well, thanks very much, Judith, for being a wonderful host already, even though it took me, indeed, several airplanes to get here, from Amsterdam to Atlanta to Detroit. And then finally, in Madison. What I'll try to do, so there are two seminars, one today, one tomorrow. What I'll try to do in the first one is give you a little bit of the history of the discoveries that we've made in my lab over the past 30 years. I'm still, I mean, you learn when you listen to how labs have gone through the years. And our story has been a little bit over the place. But I think in the end, it comes together as a coherent story, I hope. What you saw here was the gut, probably well known to all of you, of a mammal, a mouse, or a human being. And this is a so-called crypt. And those of you who studied histology probably remember the crypts of Lieberkuhn. Now, I actually found Lieberkuhn's thesis. He is German, but he graduated in Leiden in the Netherlands, a number of years ago, as you can see here. So in the University Museum, it turned out they had a copy of his thesis. And you can see it here. My PhD students were very happy to learn that it was only 35 pages. But then I actually told them, did you see it's in Latin? So we actually tried to get through the Latin. And I took Latin at high school, but that's a long time ago. So that didn't help me much. But Google Translate is pretty good in Latin. So we could actually understand what he'd done. And essentially, he infused wax through the veins and the arteries and the lymph vessels of organs and then let it cool down and then digest away the tissue. And through a magnifying glass, he would then sketch what he would see. And this is the inside of the gut. And these little protrusions you'll see are called the villi. He also saw them in the movie. And when he flipped the piece around, he saw vessels on the outside. I have a laser pointer here. But he also saw these little indentations. And he realized these are small spaces that protrude from the lumen of the gut towards the outside. And he called them crypts. And since then, he called them the crypts of Libakun. My lab started as an immunology lab. I returned after four years in Cox-Turrah's lab in Boston by helping clone CD3 proteins, very T-cell specific proteins that help T-cell receptors signal. And I took one of the enhancers back home to start my own lab in the hope that I could clone a transcription factor. It was then you didn't know how many genes there were. In the 90s, we thought there were 100,000 or 200,000 maybe. But we did know from model systems that genes that control development often are transcription factors. It's the Hox genes, the homeobox genes in flies. So I set out. I wanted to understand how T-cells develop from stem cells. And we used these T-cell enhancers to clone transcription factors. You can use them as antibodies to fish in libraries. And we clone this gene that we called TCF1. And it turned out to be three more members of this family. And then for about five or six years, we had no clue what they did. They were very good in DNA binding. We had all sorts of reporter plasmids where the protein would bind. But it would never control DNA. We then did cat assays. Nowadays, you would do a luciferase assays. But independent on the presence of the absence of TCF, there was no change in the transcriptional activity. We then did a two hybrid screen that some of you might know is a yeast trick to find partner proteins. You use your protein as bait in your clone genes that encode proteins that can interact with your protein of interest. And we clone two genes. Groucho was known again from flies as a transcriptional co-repressor. It was not so difficult to see that when TCF would bind DNA with its AMG box. And then Groucho would bind that this was then turned into a transcriptional co-repressor. And that turns out to be the ground state of what TCF do. This one in the beginning, we actually didn't realize, was the more interesting one. Beta-catenin in man was then known as a structural component of the ecadherin adhesion junctions. So they bind to the cytoplasm of ecadherin, crucial for adhesion. However, Armadillo was known as the endpoint of the wingless signal transduction cascade. I mean, I'll call the wind signal transduction cascade. And we then could actually show that that was the key role of beta-catenin in many of the processes that I'll show you. I won't show too much biochemistry, but very briefly. This is a cell with wind receptors that you see here, fissile LRP. The destruction complex, where APC is very important, beta-catenin is the crucial, the key molecule in this pathway. It's constantly synthesized by almost every cell in your body that is immediately captured by this complex and degraded through eucreanation and then goes into the proteasome. And signaling happens when winds occupy the receptors. It somehow disables the destruction complex. And now beta-catenin, although at the RNA level it is stable, you now see that rather than being degraded within minutes, it lives for hours or maybe even days. And this now allows beta-catenin to go to the nucleus, bind to TCFs, and activate the transcription. So beta-catenin essentially, we proposed this a long time ago, are co-activators of these DNA binding proteins, the TCFs. And that's, I think, now generally how people believe that canonical wind signaling occurs. We also found, together with Bert Vogelstein, that when APC is lost, this is a very famous tumor suppressor, almost always mutated in colon cancer. This complex no longer works. And in the absence of a wind signal, you have high beta-catenin. And the outcome we proposed was the same. In both cases, you will activate TCF target genes. Also at the same time, when we knocked out TCF4, one of the TCF members, we lost all proliferative activity. This is a Ki67 staining. So in brown, you see proliferative cells. We lost all proliferative activity in the knockout. Crypts, implying that winds are needed to maintain crypts. And crypts are the stem cell compartments of the gut. Probably the first example where winds not only play a role in development, also in the adult body. In the maintenance of stem cells, we now know that probably holds for all epithelial stem cell types. And in colon cancer, clones arise that have lost APC. You can be here. We stay for beta-catenin. You see there's a lot of beta-catenin. There's also beta-catenin at the base of the crypts here. And this, I think, nicely illustrates the fact that physiological wind signaling that is controlled very well happens at the base of the crypts, where presumably stem cells are. And in cancer, this pathway is hyper-activated and can never be switched off. And we then proposed that possibly the processes that go wrong in colon cancer are the exact same processes that physiologically control crypts. Now, why are crypts so interesting? We then started reading very old literature from the 60s. This is a villus. These are called crypts. There's about 10 crypts that surround each villus. A mouse has about a million of these villi. We have about a billion. They presumably serve to extend the surface area to make digestion more efficient. And I was known that if you label cells here, for instance, with a BRDU, that cells that are born here from stem cells move up. It takes about two days while they proliferate to leave the crypt. They then very rapidly mature out in one of about 10 different cell types. So they're two days old here. Then they work very hard. This is where the food is. This is the lumen of the gut. For another two to three days, and by the age of four or five days, the daughter cells arise at the tips of the villi and they undergo apoptosis. So this entire tissue turns over every four or five days, with the exception, presumably, of the theoretical stem cell. This one here, it sits somewhere here that has to live as long as its host as the owner of the gut, as the mouse, three years, or a human being, 80 years. So when we saw this and we realized that wind is a crucial driver, is this process, we thought, well, we should no longer be an immunology lab. We have to be a gut lab. Actually, that's when I moved my lab from the hospital to this research institute, the Hubrecht. And we started working on this. And the attraction was two or three-fold. First of all, the speed of this system is by far the fastest stem cell system in our body. Skin would take months to do this, live for many years. Brain never does this, turn over. Neutrophils are the only cells in our body that have the same lifespan as these cells in the gut, so a limited number of days. So that was one thing, the speed, but also the design. Every cryptvillus unit is a universe in itself. And a simple glance can tell you a lot. You know that this cell is three days old, two days old, one day old. These are proliferative precursors. They sit here. These are the differentiated cells. They sit here. And there's a special differentiated cell, the panacell that they'll get back to, that instead of moving up with the others, it actually moves down. It settles at the base of the crypts and has been known for a long time. I'll show it to sterilize the crypts. It makes bactericidal proteins and presumably to protect the stem cells that it sits somewhere here. Sorry, and I have to first show you this. So we then try to find these stem cells. And it helps if you have a biomarker, because you can actually stain for these cells. You could make knock-in mice, knock-out mice. And Nick Barker set out to find such genes. And what we used was a microwave. What we did is we blocked the wind pathway in the colon cancer cell line acutely. We isolated RNA prior to that, so in the active state of the wind pathway. Eight hours later, after we blocked the pathway, we also isolated RNA. We then were very lucky to work with Pat Brown. And this slide actually has two technologies that many of you have probably not even heard of. This is a fax. A fax is like an email that you send by phone on paper. And the output here is from a so-called microwave. This was a revolution in 2000, but it is now entirely overtaken by RNA sequence approaches. Anyway, so we performed differential microarraying. And this is a list. As you can see, many of the genes were not known yet. Many ESTs, as you can see, express sequence tags. No clue what these genes were. Some of them were known. Actually, many of these genes have been turned to nice papers by various people in my lab. But this one here, GPR-49, that was later renamed LGF5, was the one that Nick then. And this actually was a two-sample experiment that has yielded something like 80 papers or something like that. So we kept on going back to this list and still is one of the best differential gene expression studies that we have seen all done in Pat Brown's lab at Stanford. So Nick then used this gene, now called LGF5, made a number of different knock-in and knock-out mice. In this particular allele that's become very popular, he has inserted GFP. So if this works, the cells that express LGF5, which I'll show you, are the stem cells at the base of the crypt, will be green. But then separated by an internal ribosome entry site, there is a second. This is a fusion protein, a crevricombinase that's fused to the estrogen receptor, a segment of that. This recombinase will recombine LOXP sites. But it's inactive in this state. When you inject the moxifen, it becomes activated. And it then can recombine LOXP sites that are elsewhere in the genome of the cell that you're targeting. Now this mouse works very well. You see three villi, five crypts. And in between these very large penne cells, it turns out there are these tiny LGF5 positive stem cells. They're green here. We had entirely missed them. They had been seen previously. I think I have an example of that. But now knowing that we can mark these cells, and they're actually not good candidates because these cells constantly proliferate. And there was a strong belief that stem cells do not proliferate, which looking back is a bit strange because that's their only business to produce daughter cells. But stem cells, probably many of you still have learned this, should be quiescent. So if you label them with a DNA label, they retain their DNA label. That's what the textbooks tell you. That's not what we find, actually. Anyway, so in those green cells, we can activate the crevricombinase. So as we cross this mouse with a cre reporter mouse, we let it become adult. And now we inject low-dose tomoxifen and we activate the cre in only a fraction of the potential stem cells. That stem cell then becomes blue because of the reporter that's also in this genetic background. And what Nick did just summarize a lot of experiments. What Nick noted was that in the first two days, he saw blue daughter cells proliferating that exited the crypt after two or three days, differentiated out in one of these 10 different cell types that I already mentioned and we'll get back to later. Day three and four of these blue cells are now exposed to the gut lumen, so they help digest the food here. They pump nutrients and liquids into the blood and limb vessels. And by day four or five, the first blue cells arrive at the tips of the villi and undergo apoptosis. Now if you don't analyze the mice after five days, but you wait half a year or a year, two years, these blue ribbons are still there, indicating that cell that we marked is a long-lived cell. Also, you can look in a ribbon and if you give a very low dose of the moxipen, you'll see a ribbon, one ribbon, every 10 crypts. So it's a clonal event. You can then score what different cell types there are in a ribbon and we always saw all the cell types in a ribbon. So the cell that we marked is long-lived and multipotent and by our definition was a stem cell of the system. Now there was a lot of resistance to actually multiple years because everything else was wrong. They're not rare, they're actually very abundant. As I'll show you now, they don't divide asymmetrically, they just divide. Most importantly, they divide constantly and that was really against any belief of adult stem cells. So to show you a little bit about, this is actually a complicated concept that I'll try to explain about how these cells divide throughout their days. Hugo and Lawrence, two PhD students in the lab in a collaboration with Ben Simons, a physicist in Cambridge, UK, constructed what we call the confetti mouse. It has four different colors. The mouse has no color, but when you activate creeds, a very complicated system as you can see here, but it will stochastically rearrange. And in the heterozygous stage, you get four different colors, green, yellow, red, or blue. And if you make them homozygous, you actually get 10 false colors. So this allows us to mark not all the stem cells in blue, but to let stem cells pick their own color. We cannot tell an individual stem cell which color, but stochastically, they will pick different colors. And the idea would be now that we have all these stem cells marked, and we can then, like we did earlier, lean back and see what happens. And what Nick always noticed, and Hugo and Lawrence was that sooner or later, a crypt becomes monochromatic. Now, the simple explanation would be there was only one stem cell, but that turns out not to be the case. Actually, all of those cells are stem cells. And this is, or Jeroen and Hugo, who makes these movies, predicted that this is what the mice should look like on the inside. And we indeed, to your remark, that it was a little over the top. And until we got the real images. So this is a muscle is here, food is here, and you can see, yeah, here's a blow-up. So this is one crypt resolved in one color, a next crypt in the next color, yellow, green here. You can see these parallel bands that emerge out of these crypts. And for the rest of the lifetime of a mouse, we'll sit there and look like this. Now with this mouse, we can now look at the dynamics of these stem cells. There's about 15 stem cells in every crypt. If you mark all of them, and you slice with your confocal halfway through the crypt, you see circles. So this is a crypt. The stem cells are a little bit deeper. The villi point towards you. And after a week, you can see that you see multiple colors. So it's like a pie chart because multiple different stem cells are active. They all have their own color and they make these parallel columns of cells. We wait eight weeks. Almost all of this has resolved into single colors. If you now reactivate, if you inject Tamoxifen again, because for instance, if you're a red cell, the green and yellow are gone, which you can still flip from red to blue. That's what happens here. For instance, this was a red one that becomes red and blue. But again, eight weeks later, this resolves into a monochromatic color. We've made many movies of this process, banded all the theoretical modeling. And this is a simple explanation of what happened. So you have about 15 stem cells and if you would all give them a different color, they would double the numbers. So there would be two of each color, but there's only space for 15. So by the end of the day, they divide every 23 hours. So every day they double and then half of them is kicked out. And this is done randomly. And sooner or later, some crypts actually become monochromatic in a week. Others take six months. But sooner or later, the blue one wins. This is not because the blue one was better. It was just the one that by chance won several times. And this is called neutral drift. It's very difficult to explain this. But so this looks like asymmetric division. It's asymmetric division at the population base. But if you make movies of individual cells, they just divide. And one stem cell makes two potential stem cells and it depends on where they then are. If they will be a real stem cell or if they will be a daughter cell. And I'll say more of that. Well, so this sort of proved that these are the stem cells. I actually had this slide turned out for a number of years that I got from a friend from Maastricht who also shown this slide for about 20 years. This is a human crypt. This is the lumen. So it points in this direction. These are these beautiful penicels that big granules that kill bacteria. But what I had not seen for a long time, between every penicel, there are these stem cells. And it's very regular. It's always penicel, stem cell, penicel, stem cell. So once you've seen them, they're easy to see. I then realized that Josef Pannet in 1887 in his paper where he described, so this is a crypt upside down. These are the daughter cells. This is the base of the crypt. These are the penicels. But he had also seen them with the sort of fairly crude microscopes in 1887. He had been a lot better than we had been with our confocals, et cetera, and our EMs or anyway. So he also, as you can see, he puts an S there. And I think the word stem cell officially was used in the 40s or the 50s when we had a stem cell was, you probably know better, but a little bit earlier. But I thought I had discovered the real, discover of the stem cells. But then, well, this is German, but it says, S is schmal at cellen, which means small cells. So he could have been even more famous than he is. So now that we had a mouse where we could actually see the stem cells in action, they're green. We can actually, on the outside, you can make movies by putting a glass window in the middle of a mouse and glue the guts to this. You can actually see how they divide. We could ask many, many questions. And one was asked by Paul Tete. This is a classical stem cell hierarchy as we still believe that they work. So a rare, unique stem cell that doesn't divide, then all the errors point away from the stem cell, implying that when you lose the stem cell, it's at the end of the crypt. And I guess as this comes from hemopoietic system, if you lose your hemopoietic stem cells, you'll get anemic, immunodeficient, and you'll die. But we ask if that's really true. So what we did, there was already, we had a paper, Doc Winton had a beautiful paper actually showing that these cells can revert to become stem cells. This what I'll show you is a more recent study from my lab, but it's probably the best illustration of this phenomenon of plasticity. So what Paul did is he asked, can I find a marker for the most abundant stem, the most abundant cell in the system called an enterocyte? So these are the cells with the brush borders that take up the nutrients. 95% or so of the cells on the villi are enterocytes. And indeed, alkaline phosphatase, often used by pathologists, stains beautifully these enterocytes, but it also starts staining a bit halfway the crypt. And we already knew that fates of these daughter cells are set quite early for the secretarial energy. It's actually directly after a stem cell division. So a stem cell directly becomes a secretary cell without any division. But the enterocytes of which you need many, they undergo three to four rounds of division. Meanwhile, they've already taken up their fate, then they mature and they leave the crypt. So what Paul did is he inserted Cree ER into the alkaline phosphatase locus. And the idea would now be if we now mark with the same trick, we mark these cells that sit here, what we should see, and that's exactly what he saw, is that we mark cells along the length of the villi in blue, occasionally we mark a cell in a crypt, they're sort of pretty high up, and they will wash out. So after four or five days, you still see blue cells, but by 28 days there's not a single blue cell left. That's the prediction, because we don't mark a stem cell, we mark a progenitor, and not just any progenitor, we mark a progenitor that can only make enterocytes. And that's exactly what we see here. Paul then crossed these, so this all works, he then crossed these mice to a mouse that Frette Sauvage, a genitech made. He knocked in the deuteriotoxin receptor into the other five locus. This mouse is fine, but if you inject deuteriotoxin, you'll kill all its LG5 positive cells. And that's what you see here. So we then, so then Paul marked the enterocyte progenitor. The next, a while the cell was moving up, the next day he killed all the LG5 cells by injecting deuteriotoxin in this combined mouse. And the question is, will this cell, now that would normally go up, will it go down? And revert back into a multipotent stem cell. And then rather than just make a little train of enterocytes, will it now make a ribbon that persists for the lifetime of a mouse and contains all of the cell types rather than just enterocytes? And that's exactly what he found. This is a macroscopic picture of a gut. This is a pyrus patch here. Every dot in blue is an event where one of these enterocyte precursors that actually sit quite high in the crypt when you mark them. A day later, they're essentially out of the crypt. If you then kill the stem cells, they turn around and they become, as you can see here, they become stem cells that make the ribbons, implying that these cells and almost all cells in crypt are these progenitors. They have no problem whatsoever if needed to actually have these arrows turn around and become a stem cell again. I think it's now commonly found in almost all solid tissues that this concept doesn't hold. Many tissues like liver probably has no stem cell, just has differentiated cells. When needed, progenitors or differentiated cells can actually revert to a stem cell behavior, stem cell state, replace the last tissue and then revert back into the differentiated state. So this dogmatic thinking that every tissue in our body works like this, I don't think in our field we believe anymore. And I mean, I go to a lot of stem cell meetings, there's a lot of fights over who has the real stem cell and probably everybody is right because probably many, many cells in a tissue. We know all these cells can become stem cells depending on the challenges that you submit them to. Now the fact that we saw that these cells proliferate then told us that maybe we should be able to culture these cells. And it's Toshisato took this up, Japanese gastroenterologist, he had been trying to go crypts into the in Tokyo and we said, well, we should do this in 3D because that's what they probably feel happier in Matrigel. And we knew quite a bit about the growth factor conditions. We had, when we didn't understand the crypts, we had been making random knock-ins and knock-outs in any conceivable signaling pathway. I'll say a little bit more later. What Toshi didn't know, but it was an enormous amount of luck, is that, so this is frizzled, this is LRP. These are the wind receptors. Wind is a lipid modified molecule, is extremely difficult to work with. The commercial wind, I don't know if any of you work with wind, but it doesn't work. It doesn't activate TCF reporters or kill cells. So we've always made our own wind. There's now actually a fantastic creation from Chris Garcia that descends, if you ask for him, which is a synthetic, so it doesn't look like wind at all, but it brings the frizzled and LRP together. Fantastic mimetic of wind signals. This is LGF5, also a seven-transman receptor. Two more molecules are spawned in, is a soluble molecule that was known to be an amplifier of wind signals, and I'll show you how that works in a little movie. And these two molecules are tumor suppressor genes. They are recently discovered by my lab, but also by several other labs, FENKKONG in particular, and they are E3 ligases that can actually inactivate frizzles. And this is how it works. So here you have the classical wind receptors that every animal on the planet has. The winds are presumably secreted from very nearby. They don't really travel very far. They stick to everything. You can see the activator receptors. To beta-catenin and TCF, they will activate transcription targets. Amongst those targets are these E3 ligases that you see here, tumor suppressors. They will immediately move back to these activated receptors, put ubiquitins on, and remove them from the surface. Now, I won't bore you with the blots, the western blots, but this is why wind signaling is always weak and always very brief. That's fantastic in development, because you need to be able to switch rapidly, but stem cells need signals that are strong and will last days, and that's not what this can do on its own. This is where our spondyn comes in. Our spondyn binds with high-finitur LDR5. It can, with its other furan domain, absorb away the E3 ligase and in a cell that's undergoing active wind signaling, this negative feedback loop is broken, and now, when you read out the strength of a signal, you go from about five-fold to two to five-hundred-fold. So, an incredibly robust amplification of wind signals, but importantly, our spondyn and LDR5 on their own do nothing. They can only take that negative feedback loop away that turns a developmental signal weak short into a stem cell signal, strong and robust. We didn't know this when we started doing this experiment. So, we knew that winds are really horrible to work with and it was a company in California, Nouvello, was trying to develop our spondyns that were known to be amplifiers of wind signals, but nobody knew how. They were trying to develop as a drug. They went broke and then we were lucky to get a gram of our spondyn from this company. Nouvello, that was luck. And then, we knew we had to block BMP signaling from these mouse experience we did and we needed to add EGF. We did it in 3D. We can also do it in synthetic gels. And the idea was we take one stem cell, we give them this cocktail of three recombinant proteins, no serum for a totally defined medium with the exception of matric gel, and we make more stem cells. That was the idea. But we got something different. We got structures, clearly epithelial structures. This is how they grow. Very vigorous. We actually have them in the lab now for about nine years. This is a mouse and a mouse only lives for three years. Also, I guess people, also in my lab, but many people would believe that whatever grows out of normal tissue is not normal anymore, has, must have undergone oncogenic changes because it was a dogma that normal cells cannot proliferate longer than maybe a few weeks and then go into senescence, telomeres get short, et cetera, et cetera. This does not happen in these things. So the telomeres, if anything, get longer. We've sequenced a lot. We don't find any oncogenic mutations. We've done immense amounts of sequence on what we now call organoids, but they are remarkably genetically stable. And when Tashi started looking carefully, what these structures really are, they are like small representations of the gut epithelium. So from a single cell, you make these structures that you can passage every week, 10 folds, so they grow very rapidly in greasy. The stem cells, pennant cells are there, the rapidly dividing daughters. These are the villus domains. So all of the differentiated cell types that you see in vivo are present here, normal ratios. And by single sequencing recently, Alexander van Aurenhaard in our institute showed he could actually detect two unknown cell types in these mini-guts and when we went back into a real mouse, they existed. These were rare hormone secreting cells. So they really are complete versions of the gut epithelium. Importantly, they're not a tube. They are a closed space and actually you see the dead cells cannot leave so they stay in the aluminum. Every once in a while, these things will explode, poop out the dead cells, and then they close up again. And here you can see that you get quite easily, if you put them together, they quite easily fuse and enormous sex has been able to produce in collagen gels that have bundles. So they line up and then they fuse and you can make tubes that are two centimeters long or so with a central lumen and multiple crypts and again, all the cell types. So the self-organizing properties of these things is really remarkable. We know very little of how these cells do that, but these are probably good model systems to study that process. Now, as I said, also in our lab, we thought these cells must have undergone an oncogenic change because it's not possible that things grow so fast and will not be a cancer cell. We couldn't find oncogenic mutations and then Toshi collaborated with his old lab in Tokyo and he took a single stem cell from an RFP positive mouse or a red mouse, grew it up to about 100 million cells in the form of these mini-guts. This is from Cologne, by the way. In Tokyo, memores people had treated mice with DSS. DSS causes colitis, is a very popular model in gastroenterological research. They infused the Dutch red organoids through the anuses of 40 mice, waited, closed the anuses a little bit of super glue, waited for a few hours, removed the super glue, and in those few hours, yeah, we cannot do this in Holland, I can tell you, in those few hours, these mini-guts will actually find these ulcers. The basal side of the epithemium is on the outside, densely packed with integrins, they will bind, open up, and like a living band-aid, they cure these mice. And the only way, this is a real experiment. So you see a colon, you see these large patches of red cells, and the only way you can really find where the Dutch red cells are against this black Japanese background is by looking for RFP. By any other means, this is functional tissue. There's no exhaustion, there was one prediction, so basically these patches will live as long as the mice. There's also no oncogenic changes, you never see polyps or worse. I think proving that indeed, you can cure these cells for a very, very long time, and they are still safe, you can give them back to their original owner. I also stress that we do this from one cell. We could have treated maybe 1,000 mice or 10,000 mice, there's no limit to the amount of tissue you can make. We can also do the liver, and we're currently involved in a big program where we try to do this in man, where we try to not wait till a liver is so sick that it needs to be replaced by a donor liver, but when it's still okay enough to give it stem cells in the form of these liver organoids, and then from a donor that is not a donor that no longer lives, but actually donor that sits at home on the couch and is watching television, because it was a small biopsy that was needed to make the stem cells. That works in a lab, human to mouse, but it's a far way off till we can do that with patients. There was something very strange because these mini guts, this is an EM picture, this is the villus-like domain, these are the crypt-like domains, and if you have seen these many, you see a very nicely developed brush border here, the enterocytes, dead cells in the lumen, as I already said, this is a crypt, so the brush border stops here, these are the undifferentiated cells, penne cells, there is no niche, and the textbooks on this organ are full of descriptions where there's a myoepithelial basket of cells that sit around the niche. So why is there a similar debating event? Why are these not perfectly around bowls, but why do you see these crypt-like structures? So the region here must be different somehow from the region here. Again, Toshi, this was entirely his idea, I should stress, notice the following, that if you slice with a confocal through the bottom of a large number of crypts here, in green you see elder five GFP cells, in black you see the penne cells, and there's almost a geometrical arrangement of these cells. So penne cells are always surrounded by stem cells, and stem cells are always surrounded by penne cells, if they really need each other. And actually, when you kill the stem cells, penne cells are extremely unhappy to touch each other, and they suck these cells from higher or back in between where they are. Oh, there is a, no, I think I, yeah, oh, what, so when you grow crypts, they always grow, a single crypt will always give you a mini-gut. If you go cut them in half, the bottom of the crypt will always give you a mini-gut, the top doesn't. Single stem cells are very inefficient in growing organoids. 1%, 2%, so 98% will fail. Toshi noticed that occasionally, he would get outgrowth, but often the first cell that was made by the stem cell was a penne cell. So he thought maybe the penne cells who are the bodyguards of the stem cells, they kill bacteria, as we have known for a long time. Maybe there were also the niche cells, maybe they also provided the growth signals that stem cells need. And to test this, what Toshi did is he realized that when we make our cell suspensions from crypts straight out of a mouse, we have to be sloppy because epithelial cells don't really like to be outside an epithelial organization. And we usually put our gate here, we get single cells, but when you put your gate here in the forward scatter plot, you actually get doublets. And on the singlet, you'll see green cells, stem cells, and in the red, he stains for penne cells. So stem cells, penne cells, and there are no cells that are green and red, because they're either a stem cell or a penne cell. But when you gate on the doublets, you'll see many events that are green and red. And when you now look at what they are, they are small stem cells stuck onto much larger penne cells. And this is almost a perfect sort. As you can see here, green purple, green purple, green purple, small stem cell on big penne cell. When you plate out the small stem cells alone, you get, as I already said, you get 1% organoids, very low take rate. Penne cells, zero. But the doublets, 50%. And here you see what happens. So a small stem cell in a penne cell. Within a day, there's a little cyst. The penne cell really is the organizers of everything. Where it sits, you'll see a bud. And in that bud, you'll see multiple penne cells. And every day, you see penne cells move up. And the next day, there's a new bud. And then you have a new crypt. And that's how they go from one crypt. Essentially, you see a bud here. You see that a bud goes out to get crypts all along. Now, we are very happy about this paper. We send it to one of the top three journals. And of course, there's always a review or three draws. Didn't like the paper very much. And because we didn't really show what the penne cells and the stem cells were doing, and we should really, the best experiment would have been to have separate one penne cell, one stem cell, and then put them together. Because maybe what we were sorting was a special penne cell stem cell doublets. That was more special than a random. That was, I think, the argument. So we could never do that experiment, because if we needed to enzymatically treat the cells to produce treat, and then by the time that they recover, they no longer are very happy. But Toshi Denkimev had this experiment. So he has yellow stem cells, from a yellow mouse, red penne cells from a red mouse. And if you make movies from separately, nothing much happens. But if you actually mix them, you get this spectacular phenomenon. And the penne cells are like dogs that actually chase down all the stem cells. And by day three or so, they totally reassemble the. So Toshi Denkimev looked at these movies forever, and we had no clue what they were seeing. But it actually shut up the referee, and the paper was published. Anyway, we also did a microwave to see what penne cells make. So these are stem cells compared to penne cells. These are all the stem cell markers that we know very well, including L5 here. Not just crucial stem cells, frizzled, of course. But we've noticed that penne cells make defensins. They were known. So they kill bacteria. Lysozyme defensins. But they make very high levels of wind three. You can see that here. They make TGValpha and EGF. TGValpha is related to EGF. And they have the notch ligands. So in the organoids, at least, they are a very potent source of wind and of some other components that we know that are crucial. But to illustrate this again, I told you that so wind activates the wind pathway, but weakly. Our spondyn on its own does nothing. But the combination gives an extremely potent but local wind signal. So what you see here is an organoid that has two buds. So penne cells are here. They make wind. In blue, X2 is a wind reporter. So if you now have our spondyn everywhere in the culture, it can only amplify the wind signals around the source of wind. So the source of wind sits here. The penne cells here and here. You see it's blue here and not here. And we think that everything that happens here is driven by this wind gradient. So there's a very steep wind gradient, high here, low here. This controls proliferation, controls differentiation, probably also the formation of crypts versus the phyllis domain. And you can block it nicely with a porcupine inhibitor. This is all fairly old, a paper that came out very recently. It builds on work of one or two years ago by Hugh Burmer and owner Basak in the lab. This is how we now look at the hierarchy of the gut epithelium. So there is an LGF-5 positive stem cell. There's a lot of plasticity. The LGF-5 positive stem cell essentially needs three signals, high wind, high notch, high EGF. And then if we switch off one of these three, we can actually titrate these in. And we can make all of these just like a coat. We can make 100% androcytes by this combination, 100% pannous cells by this combination. But you could never make this group of cells that I didn't mention yet, the enteroendocrine cells. They're very rare in a crypt. They're really unstudied, other than that people have stained for them in tissue. But there's almost no functional studies because they are so rare. Yet because the gut is so large, the gut is the largest hormone producing the largest endocrine organ of our body. So there's massive amounts of hormones come out of your gut. And if you look here at the names, I don't know if you recognize some of them, but they're all involved in either controlling the activity of the intestinal tract or the sensation of hunger, satiety, et cetera, et cetera. So they control metabolism. Glyp1 is actually one of the best earning drug targets. Glyp1, stabilize Glyp1. This is when sugar enters your gut. Glyp1 is secreted into your blood by this cell, goes to the pancreas, and leads to a release of insulin. So before ever the glucose goes up, your insulin is already released and is sensed by your gut. So these cells must have receptors. So the idea is that they all have receptors that look in what's in your gut and then prepare your body for whatever is coming or maybe stop eating, and that sort of thing. Very little, though. So we did single-cell sequencing. I won't take you through this. You've probably seen many of these plots. There were arguments, are there 20 lineages do all cells produce only one hormone? Can they make any hormone? Do they make combinations of hormones? What are their receptors? Nobody really knew any of this. So then Helmut Geyhardt came up with this fantastic experiment that I'll try to explain. Neurogenin 3 is a transcription factor that is when a stem cell becomes an enteroendocrine precursor, neurogenin 3 pulses in a matter of a few hours and it goes off again. Very unusual. Usually these transcription factors will stay on, but it goes on and off in two or three hours. And from then on, that cell is an enteroendocrine cell. It can still make many different decisions within that lineage, but that's what it is. In this locus, he cloned a very stable but slow folding red protein, tomato, and a rapidly folding, but he destabilized it, green protein. And the idea would be that if a cell becomes an enteroendocrine cell, it pulses neurogenin 3. Green will follow immediately, but red is slow but stays on. And by color, you get a color code, green, yellow, red, tells you exactly what the age of the cell is after this peak. Now I have to say this is never my idea. This is entirely done by Helmut. And this works in an organoid. Follow this cell. Let's wait one cycle. Yeah. So this, oh sorry, I should probably say this green here has other fluorescence in the lumen, all these dead cells. This is where the epithelium sits. It has no color. Only the enteroendocrine cells that express neurogenin 3 have a color. And as you can see, they start with green, yellow, red. And then what we didn't know, green, yellow, red. So it goes to two of these pulses. If we change the conditions, we actually found conditions to massively induce enteroendocrine cells. You see it here on the right. So they're all green, and they become yellow. And then they become not exactly timed in a very way. They become red. So it works. Also in vivo, you'll see that the green cells are at the base of the crypt. The yellow cells are halfway, and the red cells are at the top of the crypt. And then this was the plan of Helmut. So here you see, again, the villi. You see crypts. We zoom in. These are the pannet cells. These are the stem cells that you might recognize by these. This is the site where a stem cell will go out of cycle and decide to become an enteroendocrine cell. So if it does so, it will actually, there it goes. Neurogenin 3 will peak. It will be green, as you can see here. Slowly, red will follow. Green will be degraded. So here the cell is yellow. This process goes on. This takes by about 24 to 30 hours. It is slowly becoming red. It grows this axon-like structure. And you see green, yellow, and red. And that's exactly what we see. And when now, he sorts directly from these mice, he sorts enteroendocrine cells. They're produced every day. So each cell goes into a single well. The color is recorded. RNA sequence, single cell sequencing is performed. And then, simply, it's very simple to now fill this entire hierarchy with the single cell profiles of enteroendocrine cells. And to summarize, so this is all done by him in about a year and a half. So he basically then focused just to show what you can do with these kinds of approaches. He focused on transcription factors that were not known to play a role in the system, but were very early on expressed in the different lineages. Also, it helped him to sort out how many. There's one, two, three, four, five, six lineages are there. And the transcription factors in blue, he knocked out with CRISPR. And out of the 10 or so that he knocked out, six gave a very nice phenotype. So they were clearly involved in setting up one or sometimes two of these different lineages. We still don't know what the signals are. It's highly reproducible. Also, different regions of the gut have different ratios of these enteroendocrine cells. Organoids from those regions do exactly the same. So GLIP1 is only made by an ilial organoid, never by a duodenal organoid. And I think secretion is the other way around. Yeah, this is now an enormously rich source. We can now look for what kinds of receptors do these cells have. What do they actually see in the gut? What do they respond to? And I also advise a company that is a calliope in New York who's heavily interested in these cells. And they have great interest on Coca-Cola and Pepsi-Cola, because these companies know that their light drinks taste sweet to your mouth, but they don't taste sweet to your gut. So they never will lead to the feeling of satiety. And they hope to find out what actually are the signals that these cells pick up, because they will tell you that you stop eating or stop drinking. Yeah, so when we first did this, we found that if we could grow these organs, we can now essentially probably grow all epithelia of mice and humans. We thought that the biggest translational application would be regenerative medicine, as I showed briefly. That is actually a long, long ways, extremely expensive. And this probably doesn't fit with the business model of an academic lab, because you have to improve. It has a GLP, GMP. It costs probably $50 million or $100 million, so that will probably not be done by us, but should be done probably by companies. But we then actually focused on the use of these organoids. And tomorrow, I'll give more examples for cancer. But the best one that actually now is clinical practice in Holland is cystic fibrosis. And to give you some background, this is my last little story. The most common CFTR mutation is F508-DEL, which is a class 2 mutation. I can do a bit better than you. CFTR protein is produced, however it is not processed properly, and does not reach the cell surface. Lack of CFTR at the cell surface means there is no chloride transport out of epithelial cells. In addition, enac regulation is lost, resulting in excessive sodium absorption. This leads to dehydrated airway secretions, impaired mucosillary clearance, and chronic bacterial infection. A corrector molecule, such as VX809, could potentially facilitate movement of F508-DEL protein to the cell surface. Additional treatment with a potentiator, such as IVACaftor, could then activate these F508-DEL CFTR channels. I could do a whole talk like this, actually. Yeah, so what he tried to explain is that, so CF is actually the most common erratitary disease amongst Caucasians. And it's believed that it is so because carriers of mutations in this gene, the CFTR gene, is a chloride channel. There's large numbers of mutations. One is very common, the Delta 508, or is it here? It's about half of the patients in Holland, but also in the US, I believe, are homozygous for Delta 508, one of the US is missing. But as you can see there, well, there are now probably three or 4,000 different mutations are known that often are only occurring in one family or in a small ethnic group or in a village. And clearly, if you want to design drugs, there is no good business model for that. The Delta 508, the drug that was just mentioned, Vertex, in California developed incredible medicines. So this is a folding problem. The channel is a very complicated molecule. Doesn't fold very well, gets stuck in the ER. And the patients cannot secrete chloride. Their epithelial dry out, they become very viscous, get chronic infections. They lose various functions of lungs, of pancreas, of liver, of gut. And this small molecule of Vertex will actually, like a chaperone, put this sort of back in the right form, in the right format, and bring the channel out to the surface. It's really amazing that this works. So it's a folding problem because of the mutation and a small molecule helps the protein to fold correctly. The reason, I should say, that CF is so common because it presumably protects against cholera infections. There were lots of cholera epidemics in Europe in the Middle Ages because the toxin that cholera secretes opens the channel and this actually leads to a massive secretion of liquid into your gut, massive diarrhea, and the patients get up to 15 or 20 liters of diarrhea per day. They dehydrate and they die. If you're a carrier of a CF mutation, you have half the dose of the channel and you're much more likely to survive a cholera infection. That's the theory behind this. That's why it is so common, but we actually use that cholera trick as I'll show you. So this again shows the population. In Holland, we have about 1,500. You'll have about 30,000 CF patients. Half of them are homozygous for this mutation. The Vertex drug, a combination of two drugs, actually will help these patients. There's another few groups that are slightly, that are not too small. One of these groups responds to only one of these two components very well, but then there's a large number of patients that have rare mutations that even if the drug would work, you could never test because you cannot do trials on two people. You need statistical evidence. Also, formally, you need to prove that you stop the deterioration of lung function of pulmonary function. These patients lose 3% of their lung function per year. You have to show that you stop the loss of 3% of pulmonary function. In kids with chronic infections. So it's almost impossible to measure. So it is very, very subjective. And because of that, Vertex has never dared to actually enter these patients. Also, they could probably never make the money. And then again, because these are very mutation-specific in the way they work, they help fault one particular, a misfolded protein version, but probably not if your mutation is there or there. So what our collaborators in the Children's Hospital cross from the end, and if a big one realized, is that you can test functionally this channel. And what you see here is actually a rectal biopsy. It's painless for kids. Actually, almost everybody who is suspected of CF gets a rectal biopsy. We get the remaining tissue. You could also do it on lung. You can do it on kidney tissue. You can do it on liver tissue. But this works the best. What you see here is a movie that runs over one hour and loops. We add cholera toxin. We open the channel. I told you that these mini-guts are closed spaces. And what happens is they basically, they get filled with liquid. So this is like the mini-gut version of cholera diarrhea. And because there is no opening, the organoid swells. And we have software that very easily measures this. This is a CF patient. I'll show him a little bit later. You can see there's almost no lumen to start with. So that's typically what we see. They look already collapsed. They grow as well as these, but there's almost no lumen. And when we add cholera toxin, they don't care because there is no channel that they could act on. And this is why CF patients will always survive cholera infection. If we now incubate this particular patient for a few hours with the vertex drug, so it was this patient, if we now add cholera toxin, we have restored the swelling assay. So it's a very simple assay. We actually now have about, we have half the Dutch patients in our biobank, almost 750 patients. And we're probably completing the whole 1500. In the next year, we're also doing this now, Europe-wide. This is our champion, Fabian. The Weereldrei door is the mode, you probably have not seen this, but the most popular Dutch talk show. This is actually a one-hit rock artist over about 20 years ago that you probably also have never heard of. He told his story. So he had a rare mutation in this CFTR gene, only shared with his aunt. Exact same mutation, his aunt is 30 years older. His aunt was not very sick. He was very, very sick. He would probably not have lived much longer. Of course, from the end, his treating doctor realized that his mutation in 3D was not so far away from the Delta-5 weight. And we set up this organoid test that he just showed, and he was the first, actually, he saw the results. He was the first to respond to this drug. He was then immediately put on the drug by Kars and within a matter of a day, he gave up like a liter of slime and he said, I can finally breathe again. And we actually had drug for about a month because another patient had stopped taking the drug. Then we ran out of the drug, or cross ran out of the drug. It was stopped again. That's when he appeared on television. He told his story. He said, well, this need to be reimbursed. You might wanna know that Vertex wanted to price the drug at $400,000 per year per kid. And if it works, it's lifelong. So that's, and our minister, so other than here in this country, our minister negotiates always overpricing of drugs and thought it was excessive, particularly because also half of the patients that have the right mutation do not respond clinically. That's not known why that is. But so you would spend enormous amount of money to only treat a subgroup of patients. And the other half could not be treated at all because there was no guidelines how they, because they were never in trials. So he told his story. And based on this, the minister then decided and the insurance companies followed. And the regulations are now such in Holland that every time that we in Utrecht was no longer us, it's now a small foundation, finds patients like him. And we have so far already found 60 in Holland that respond, although they have a strange mutation, they'll be put on the drug and they get reimbursed by insurance companies. And so far, the score is 100%. So every time we say it works in this simple essay, it works in vivo. And we're now doing this in a consortium with two or three other centers for the entirety of Europe. U.S. so far has not shown very much interest, I must say. And then finally, I think, this is finally, yeah, but we also, this is actually a bunky on coal. Again, one of these experiments that I never knew was going on. He realized, he read the CRISPR-Cas9 papers in 2012 and somehow lays hands on the construct. And this is the test that I just showed you of normal. You have the Sisypher Rosas test. You're very cyclical in P with frost colon of cholera toxin. This one does not swell, this one will swell. But what Bonkyong did is he, with CRISPR-Cas9, he, this is actually half a year after those papers were published, he caused a double strand break right next to the mutation. You see that happening here, CRISPR comes in. And then we only need a short piece of DNA, a 100 base oligo is good enough to bridge the gap, so let the cell repair the break, but at the same time, replace the mutation by the wild type tissue and cure the stem cells of, I believe, we actually, we also did the Fabian. And this is the outcome. So here you see the patient before we went through this CRISPR-Cas9 and this is after they repaired. And you can see you very nicely restore the swelling. As if you now know, we've done this extensively that if you design your CRISPR-Cas9 very well, there is no, we never see off-target. Depends a bit on the sequence. And as I should stress, was done without me knowing in my lab, Bonkyong really came up with this and was published, I think it's the first repair, the first demonstration of a repair of a genetic defect in human cells. With that, I guess I named most of the people in the lab where I have lots of collaborators. I'm sure I'm not the... There's actually, there's a story behind this because when Jeroen made these movies, I was traveling and I saw this in Amsterdam at the airport and I wrote to him, this is really over the top, Jeroen, you can't have to do something about it. You could have Nemo swimming through this movie. And then he mailed back and said, okay, 400 euros. And then when they landed in Boston, Nemo was, well, anyway. Thank you very much for your attention. Well, it's not true. We actually give an extremely high wind signals, but you should not confuse wind signaling through receptors with mutations in the pathway. And if we make the comparison with insulin signaling, insulin is given liberally to hundreds of thousands of people, many millions of people. It's not presumed to cause cancer, yet there are many oncogenes in the insulin signaling pathway. So they're just giving very high levels of wind for years. We never see oncogenic mutations. Actually, the reverse happens. If you would have an APC mutant, you would not select it out because the wind signal is there, that's what we see. So, but yeah, that was the reason why Nevella went broke, because investors and advisors said, you cannot activate the wind pathway in a patient because you'll cause cancer. But I don't think there is a good reason for that. Unless you think that giving wind to a cell will cause an APC mutation. But that's not... So you're smart. Now, what I'm saying is that there's no reason why a wind signal would lead to mutations in APC. An APC mutation leads to an extremely strong wind signal in the cell, but it has nothing to do with the fact that a physiological wind signal can also do that. And somehow in the minds of, well, my mind as well, these two things are connected, high wind leads to cancer, but it's not true. You can actually just... Nevella has a science paper where they had mice that over-expressed are spawned into enormous amounts. They had guts of this size. They never got cancer. If you over-stimulate and cells proliferate for a long time, so the proliferation might lead to... Yeah, so if you drive proliferation for a very long time, that's fact, but EGF will be as bad as wind in that sense. So there's nothing special about the wind in this one, but many people believe that wind is a dangerous pathway. But I would not agree as you probably hear. Replaced by their own daughters, but their own daughters, that would be a bit semantics because their own daughters were stem cells two days earlier. But actually, we do know that normally this plasticity process doesn't happen because in this alkaline phosphatase mouse, when you label these cells, they always wash out and are gone. So normally you never, well, no, I don't think we've ever seen in a normal situation that they go down, but the moment you take the stem cells out, they do it. And as many confirmed in, well, it was already done before our experiment and confirmed by many. And so that's because there are 10 penicels. And penicels, so penicels, so I think the stem cell is a relative, is not a hardware thing. Any cell that touches a penicel automatically becomes an elder five stem cell because not just high, wind is high, EGF is around. And so, but that basically shifts the question to another problem, how do you fix the penicels? I've tried to convince people in my lab to investigate if we have no clue how it's done. It is extremely tightly controlled. Tiftereotoxin receptor in lysozyme, you can kill all the penicels. So normally, they actually, I didn't tell you this, but penicels live about eight weeks. So much longer than anything else. And they actually slowly move down. So the oldest penicel is at the very base of the crypt and that's where it dies and then it disappears. So they go down like this. Every week a crypt loses about one penicel. It produces about 2,000 cells every week, but it knows exactly to replace that one penicel. If you kill all the penicels, the next day there are 10 new penicels. So there is a counting mechanism that's extremely, you can kill them for a week every day. The penicels and the next day they are back, there's 10 penicels. So there's a counting mechanism, we don't, they're bigger, but they have a few more. In damaged situations after radiation, there are many, many more penicels. So then there is a mechanism that drives up their numbers. You have huge crypts that again, very rapidly split up in smaller crypts and that's how you get your crypts back after chemotherapy or radiation or infection or something. But so that is a big, and I think that maintaining the penicel numbers at this given 10 to 12 explains everything else. It sets up all the gradients and the numbers of stem cells, et cetera, et cetera. Yeah, so what I didn't show but what I tried to mention is actually that so there's two main branches, the anthracytes, the special thing about anthracytes is that 90% of the cells, 95% of the cells are extremely abundant, relative to the other. The secretary cells split in, so secretary cells then become goblet cells, pendant cells, enter anacrine cells. So that precursor state is also plastic, can also become, but there are far fewer of them. So quantitatively, the anthracytes are the most important but they can both do it. We have not seen good evidence that the fully differentiated cells can do this, but they're on the villas, it would be very difficult, there's actually a ring, so actually when they get out of the crypt it would be very difficult for them to get back into a crypt. In other tissues, in kidney and along, other people have shown that fully differentiated cells, maybe not very efficiently, but can in situations of damage, quite easily become stem cells. And the liver-finsity oval cell is believed to be a cholangiocytopyl duct cell that arises in damage conditions, makes new tissue and becomes an oval cell again. I think it's a very common principle that almost all solid tissues use. Yeah, that's also a good question. So are the 15 stem cells the same? So at the single cell level, by any means they are the same. When you make movies, you can see that the ones that sit high up in the cells, because it's like a little basket that sits at the rim of that basket, they have a higher chance to be kicked out. So sometimes you see a stem cell dividing into two and then they push out another one that's not even dividing. So cell division and fate has nothing to do with each other in the system. The ones that sit high up have a higher chance to be pushed out by somebody that's lower. But actually, I think it's too full lower chance to persist a long time, but they still will. So they're different in location, but we have not been able to see that they're different in any other gene expression profile or DNA methylation or anything. So it's just location. Actually, when you make movies, you see they also go up and down and they're constantly moving. And then I have to show you this. So what we already mentioned is briefly, but what we learned is that different regions of the gut will actually make different ratios of these things. That's fixed. So we cannot influence the ratios. It's a given in a particular tissue. We don't know what the signals are. What we did discover, though, is that there are, because people argued, some people said there were 12 or 14 of these cells. This is actually nicely documented by us, but also by others now. There's sort of a wind gradient higher to base, lower to villas, but there's an inverted BMP gradient. So BMP is higher to the villas. When the cells move out of it, so in the crypt, they express one hormone, the very same cell, when it enters, the villas domain will switch hormones. So actually, each of these cells has a crypt version and a villas version. And by playing around with BMP level, you can block BMP quite easily in the gut, then they keep their crypt version. And they're often the more interesting hormones, clinically. So you might think of a pill that would block BMP in your gut and you would make more glib one, which is this hormones very interesting for diabetes. But we don't know what the signals are why this cell becomes a delta cell and not a kappa cell, for instance. Looked very hard. None of the known signals have any influence on this. It looks like a hardwired genetic program that has no influence from the outside, as far as we can tell. I think it may be time to go and eat. What do you think? Have a glass of wine.