 Be very clear how important Harry Steenbach was and is to our campus, he very briefly was the person who figured out how to produce vitamin D from UV irradiation, patented it with his own money because nobody here was interested in patenting and then actually launched Worf with some of his colleagues. I love the fact that the initial budget of Worf was $900, $100 for each of the founding faculty members who started it. That's a great number to remember. Anyway, so however what I want to do more importantly is introduce Hans Clevers. He gave a wonderful lecture yesterday and will be continuing today on his work on stem cells and organoids. He is famous, lots of awards, lots of recognition and I'm not going to leave it at that. I said enough yesterday. Is that I want to hear from Hans. Okay, Hans, Clevers. Well, thanks again very much for having me here. I really enjoyed. I've only been here for two days but I think you are very privileged to work at a beautiful campus in a fantastic city. Also, I realized yesterday that Steenbock, as we pronounce his name, is actually a Dutch word or maybe Afrikaans and it means Capricorn. Anyway, so I don't know if that means anything. Okay, so I'll very briefly recapitulate what I talked about yesterday. Just two slides and then I'll try to come up with a number of other stories, some very recent, some not published yet, but that gives some more flavor to what we think these organoids can be used for. So what you see here, you did a remark, you shouldn't call them the champion, the champion but one because the germ cells are the real stem cells. But the gut stem cells that sit at the base of these crypts are amongst the most active stem cells of the adult body of a mammal. They divide every day, every 24 hours. They do so for the lifetime of a mouse and it only takes their daughters two days of play and maybe two or three days of traveling as a differentiated cell to reach the tips of the villi and undergo apoptosis. So this organ, at least the epithelium of this organ cell renews every five days, which if our guts do that you would produce something like 100 to 200 grams of epithelial tissue per day. So it's not much lower number than the weight of the food that you digest with this so it's massive. We found a long time ago that the winds that were then only known as developmental regulators that wind signaling drives this process and we now know that winds are important for many types of stem cells. I'll say a lot more about cancer. Bill Dove was instrumental in identifying the role of APC in colon cancer in a mouse in the ABC MIN mouse. Tumor suppressor in humans. So colon cancer comes about presumably the first step is a mutation in APC or in several other components of the pathway that activates the wind pathway. And so normal crypts need wind. Colon cancer use winds by activating mutations to grow forever. And we then combine these two insights and in a microwave experiment with Pat Brown, we found elder five. Elder five, we now believe is a marker of all types of epithelial stem cells in a mammalian body where these stem cells are active. Also, elder five, I don't show this here, we and other labs found much later the seventh transmembrane receptor is a receptor for arspondin and arspondin is an amplifier of wind signals. So alone it doesn't do anything. Winds alone give very brief and weak signals. Fantastic for development. Not good for adult stem cells. They need prolonged, robust signals. And that's where arspondin together with elder five serve as an amplifying mechanism to take a weak wind signal and make it very strong. So we designed a cocktail of three growth factor based on a lot of transgenic and knockout minds that we made in the lab. No serum, basic medium but then we do this in gels. Major gel works well, collagen also sometimes work. And a single stem cell grows out. We hoped to make many stem cells but we got these structures that we now call miniguts and essentially they are a small version, of the epithelium of the gut of a mouse or a human. You can easily do this in man as well. And they have all the cell types as far as I know all functional studies that have been done on genes, transporters, enzymes, transcription factors that are known to work in vivo can be recapitulated in these organoids. And we actually, I think we sort of we misused the word organoid when we wrote this paper about ten years ago because in the literature, organoids are defined as structures that are formed when you take an embryonic organ and you grind it up into single cells and you mix them together again and these cells will actually reform an organ and those were called organoids. So we actually then, in this paper and people have complained about the nomenclature but we now I think use organoids for structures that are generated by stem cells be it adult stem cells, that's what we do or be it IPS or ES cells that grow in 3D and recapitulate structural and functional aspects of the organ that you find interesting. So that would be our definition. So it turns out that, so this is many guts in essence we have we have developed a long list of protocols you can find them, they're mostly published now you can find them in the literature and other labs have added some as well where the major ingredients are always wind, so you need to activate the wind pathway sometimes tissues make their own wind then you only need our spondyn this is the case for the small intestinal organoids we're very lucky to not to actually pick this as our first first issue because without wind for instance a human gut doesn't grow but we would have reached nothing the second one you need to activate the tyrosine kinase receptor this can be EGF Neuroregulin IGF, FGF or combinations depending on the biology of the tissue and the third almost always you need to block BMP and teach a beta signaling now if you do that typically you don't need to start from stem cells if they exist at all you can take a little bit of tissue put them in culture in 3D add these growth factors you'll see some growth and then depending on the tissue you need to add sometimes up to 8 in total 8 small molecules or growth factors and essentially we can always come up with a they're probably not all as optimal as they can be but you can grow these tissues exponentially for at least a year without selecting for any oncogenic or otherwise deleterious mutations and you see a list here so when I worked in mouse and that's what I mostly presented yesterday we could quite easily we needed to add a few components but grow small intestine, colon, stomach liver I'll give examples pancreas, prostate Rob Kopp is in Holland's gross salivary glands actually he is now transplanting them for in my mouth with the Sika syndrome a Japanese group published taste buds so they have algae 5 stem cells you can grow them and they taste you can actually inject salt or sweet and you see that calcium signals inner ear organoids will make hair cells so people have been using them to study hair cell loss in aging and hopefully learn how you can generate an inner ear etc etc so I'll give you a number of stories where we essentially try to present a variety of different ways of which these organoids can be used this is a study that Jarno Drost did in my lab he published it maybe two years ago Toshisato who originally did the organoid cultures fully independently in Tokyo we hadn't really communicated over this so it was essentially we did exactly the same experiment and both labs realized that the mutations that occur most frequently in human colon cancer essentially reflect very well the growth factors that we empirically use in culture so APC a key negative regulator of the wind pathway we have wind are spondent in our medium and the prediction would be that when a cell loses APC or activates beta-catenin by point mutations or hits any of the other components that that cell no longer needs wind KRAS very often mutated in human colon cancer many other mutations we have EGFNR in our culture medium the prediction would be if we activate KRAS with a single base change we no longer would need EGF Noggan is our medium is a BMP inhibitor prediction would be if we hit SMAT 4 common transcription factor to the BMP and Tj beta signaling pathways we could leave Noggan out of the medium and then we've added P53 mutant and almost all solid cancers we should be able to reconstruct what in the colon cancer field has been in terms of the Vogelram based on the paper by Eric Fierre and Vogelstein about 30 years ago so this works very well what you see here top left is human normal colon organoids you can also take small intestinal organoids so in the lab they don't really behave differently you leave wind out of the medium see it here they all die they immediately die actually it's an acute growth factor so if you take it out the cells are dead in one to two days there's not the case for many other tissues so they will need wind to overcome senescence but they don't use wind as an acute growth factor we don't know what the reason for that is but it's important to know if we target APC by four different guide RNAs some work better than others but in all cases we get perfectly round organoids we no longer need wind and when we pick one of these organoids and we sequence we see two frame shifts and each of the two APC alleles so apparently these organoids come from a single cell that we hit with CRISPR-Cas9 and took APC out so now they grow without wind they still need noggin and EGF KRAS is targeted by CRISPR-Cas9 so it's not very difficult we target close to the G12 codon we try to turn G into D so double strand break and all I go of about 100 base is enough to have the cell repair the break but at the same time replace the triplet here that goes for glycine for aspartic acid and whereas normally the cells need EGF when we activate KRAS we no longer need to put EGF in the medium this is actually known to be a very strong KRAS activator and when we go for the weak alleles we cannot leave EGF out of the medium entirely but we can titrate it down so it's clearly a relationship between the strength of the allele and the amount of growth factor that are still needed to have them grow P53 mutations can be selected by this small molecule nutlin that interacts with MDM2-P53 interaction it's very efficient in killing wild type cells however if we target P53 they survive the addition of nutlin so here we do this experiment in a stepwise fashion but you can actually do it all at the same time and by just selecting with nutlin you can then in individual organoids you'll always find P53 mutations and they have different combinations of mutations and the other genes that you targeted in this single experiment so now we have three mutations and they still need noggin if we now target SMAT4 and we leave noggin out of the medium so noggin is clearly essential but when we hit the SMAT4 gene we no longer need to add noggin they still need there's no serum in this medium they still need the matrigel so they're still dependent we think on integrin signaling you can do it in any order so cancer probably what voxelstein proposed has a series of mutations but it doesn't matter how you do it you can do it all at the same time you've done six, seven, eight genes at the same time you can make a library of all sorts of combinations of tumor suppressor mutations when we started transplanting them what we noticed is that this is orthotopic you can see the colon epithelium here so they're injected directly under the epithelium so with these three mutations there's very little if any growth they sit in the sub-epithelial space and only with the four mutations do we get invasive growth and if we wait four or five weeks to get metastasis so a very simple view and I've discussed EMT quite a bit and actually we don't really believe that EMT has anything to do with metastasis at least in this model you can see these cells are invasive they're very epithelial the metastasis they produce are also epithelial and you can find little clusters of cells in the blood of these mice and they're also epithelial so there is no EMT-MET step involved, so that's a strong belief of many people so you definitely need the four combinations of these mutations and a simple view on what metastasis really is is it makes a cell independent of its niche so you can inject these cells anywhere in the mouse and these ones and they will grow out in a tumor because they don't require much other than some extracellar matrix to grow and it was actually very nice confirmation that what we've been developing totally empirically this culture medium is exactly what nature uses when it tries to create a cancer from that same stem cell and so often when we try to define a new cocktail we have the three basic ingredients but then we look in literature see what people have used to get some growth and we look in the mutations that occur in the cancers of that particular tissue and that also often inspires the pathways that should be activated that should be blocked so here we make artificial cancers extremely well defined so there's these four mutations but nothing else which can also grow natural cancers and this is now a long series there's some here but we and other labs have published large biobanks of cancers of individual patients so the idea is this is actually Dutch healthy disease this is a colon cancer we can grow the healthy cells we can go to disease cells side by side actually the healthy cells grow better than the cancer cells and I'll get back to that you can sequence of course you can do DNA RNA sequence DNA sequence you get are very clean but they are similar to what you would get directly out of the tumor but the good thing is that you actually here's the DNA the cells in the lab growing and we can now do all sorts of interesting experiments but we can also perform drug testing and if you think about that's happening here if you think about the situation when you would have a bacterial infection in your lung what has happened for the past 60 years is that these bacteria are cultured in the petri dish some 10 antibiotics we call them a spider are put on top of these bacteria and you can exactly see what antibiotic will kill the bacteria of that individual patient if you are a cancer patient you are classified in a group so you might be stage 3 colon cancer, MSI or MSS or something like that and then based on that classification you are put into a treatment regiment a treatment protocol that statistically works best for the group of patients that you are in but it doesn't ensure that actually it works for you it's never been tested for you and on average the statistics are that chemotherapy radiation works for 35 to 40% of the patients yet 100% of the patients get all the side effects and if you are put on the wrong drug it takes a few months usually before it's decided that it doesn't help you deteriorate because the tumor grows but also because the drug causes side effects so it's really bad to give the wrong drug a non-acting drug to a patient so we hope that we can develop this to much like I showed yesterday for cystic fibrosis to inspire new treatments of patients and maybe personalized treatments so the idea would be you have the tumor cells you also have the normal cells that has never been done before perform DNA sequence, RNA sequence and drug screen so it's quite easy to do a 50 compound drug screen and for colon cancer in Holland we have about 8 to 9 drugs that are regularly given to colon cancer patients you could expand this to 100 or 200 Dave Tuverson, a collaborator of us has taken over the pancreas carcinoma project and he's actually been using a much broader panel of drugs also drugs that can be given clinically but have gone out of fashion or have never been used for pancreas cancer and he had several cases, he published this recently where he found a drug that would never be considered for a pancreas cancer patient that worked on the organoids and then actually had clinical benefit when it was given to the patient who was confirmed in the clinical studies so this is what they look like normal tissue, tumor tissue for one patient for another patient the drug screening is we originally did this at the Sanger center with Matthew Garnett and his people you see them here it was done in a blinded fashion so Nutlin would presumably kill the wild type cell the P-50 wild types and all the mutants and that was very nicely picked up by the robots these three were P-50 wild type cases and these were mutants 5FU is a classical chemotherapeutic so the targeted drugs are typically given to patients who have the DNA changes like you got a BRF inhibitor and lung cancer when you have a BRF mutation but classical chemotherapeutics there are no almost no biomarkers that can be used to predict whether you will or will not respond so what we see is that we get very robust non-responders and responders to these drugs but we cannot see in the DNA why this is and this has been seen by many other people so there is no simple genetics that makes you resistant or sensitive to these common chemotherapeutics we can freeze and install these organisms and it will give us the exact same exact same readout so we believe and actually there was a paper published in Science by Vluck with Giannis last year on about 60 colon cancer patients where they used this assay and they predicted with 100% correctness non-response which is very important so you would not give a drug to a patient that will not respond that's even more important than to give a drug to a patient that does respond and they were 85% correct in predicting that the drug would work so this is far better than pathology would ever give you probably if these trials become bigger and bigger it will not be surprised if it would stay as good as it looks now we are doing several of these trials but there are multiple labs around the world now trying to see if this can be used to advise oncologists also fast enough to be meaningful in the context of treatment of a cancer patient and of our interest IWP2 is a porcupine inhibitor so IWP2 is actually being developed for several kinds of cancer it blocks the creation of wind signals so the APC mutant cancers they are no longer reliant on wind because they have a downstream mutation in APC so they don't care whether their receptors are activated but RNF43 mutants if you were here yesterday these are the E3 ligases that in a negative feedback loop block the wind pathway if there are cancers that are mutant in these E3 ligases they still require a wind but they need very very little wind they should be sensitive indeed we didn't notice this was the only case in our large panel of a cancer that was RNF43 mutant and not APC mutant and was highly sensitive to this IWP2 compound and actually this is now being taken Novartis has a similar drug and they use it in pancreas cancer and a few other cancers the sequence particularly for these E3 ligases and these patients are then candidates to be tested with a porcupine inhibitor so we made two unexpected findings the first one is illustrated by this one and the next slide what you see here is a normal human colon organoid that where we have H2BGFP in these cells so you see the nuclei you see very nicely these cell divisions and when you make movies of these organoids you'll see frequent mitosis and every mitosis is successful so they will always generate two healthy daughter cells and they exponentially grow if you make the same movies from cancers almost invariably there are aneuploid so this is turns out to be a single cell with two large nuclei when it condenses it chromosomes there might be up to a hundred chromosomes so far too many and when it then segregates these chromosomes into two nuclei you see it has lagging chromosomes lots of problems here here there is originally three nuclei that then resolve into two here both daughters don't get through mitosis they both die and this is to illustrate that when we make these movies we see that the colon cancer organoids grow slower than the wild type organoids it's very unexpected it's actually a problem because if you get a tissue sample from the tumor that has normal cells they will overgrow the tumor cells in a matter of two or three days two or three passages prostate cancer where the tumors are almost always mixed in with normal cells it has been impossible for multiple labs to grow prostate cancer organoids from the primary tumor you can grow them from metastases because they are pure cancer cells but the primary tumor after three passages you are left with wild type cells that are not the ones that you want to study so my sense of this is what we are looking at is that the normal organoids are not in their normal situation they are in a situation where we maximally stimulate their damage response so they as fast as they can produce daughter cells to repair a perceived loss of tissue normally they would do this briefly until the defect is repaired and then they go back to their resting state but apparently you can keep these cells under these conditions for years and the telomeres don't shorten under these conditions third is driven by wind as published by Kamler and we confirmed that there is a lot of homologous repair genes are induced so they take care of their genome very well so we never see senescence in these cultures the cancer cells are pretty bad at dividing and they run into all sorts of checkpoints and problems and often become apoptotic the big difference is that they are like zombies so the normal cells just run as fast as they can, they sprint but then they stop these cancer cells are much slower but they will never stop so eventually so pieces fall off like in a zombie movie but eventually they will get you as you know what happens in movies as well the second observation because we can now test side by side cancer drugs, the effect of cancer drugs on cancer cells and on normal cells from the same tissue from the same patient and almost no cancer drug is specific for the cancer cells so they are exceedingly good in killing normal cells and actually better usually in killing normal cells than cancer cells so we think that the reason that they can be used in a clinic is not so much that they are cancer specific but actually the normal cells as is exemplified by these two movies are just much better if a few cells remain to step in and repair the tissue as soon as possible and then go back to sleep again and there is a few targeted drugs that will kill the cancer but not the normal cells but almost all the drugs will kill the normal cells at least as well as cancer cells two slides on another system this is actually Jeroen's best movie that's why I showed it so you can recognize this is a stomach a human stomach these structures are called glands so there is a flat surface epithelium superficially it looks very much like the gut so this is a gland but the cell types are totally different most proliferation occurs in a region that is called the isthmus cells move up and down and the cells that move down are perietal cells these are the cells that produce protons and lower the pH to one or two perietal cells forming here then there are chief cells at the very base of the gland you see them here and chief cells make pepsinogen a very large amount of protein the highly differentiated cells however they were shown in the 60s to show some proliferative activity with 38 thymidine they take it up and we found that they expressed Troy and Nick Barker who has been in Singapore for a long time and discovered L5 in my lab found also L5 is tracing on these fully differentiated cells in a healthy stomach we noticed that very very slowly they do divide and their daughters move up and trans differentiate into other perietal cells they actually will populate the isthmus so these are really the progenitors that constantly build the stomach and even if you wait for about a year you'll see these blue cells coming out from these glands and then also we populate very different cell types that sit on the surface of the of the lumen of the gland so this is a chief cells a fully differentiated cell that sort of moon light as a stem cell now quantitatively this is not very important in a healthy stomach because most cells are 99.9% of the cells are produced in the isthmus however if you radiate mice you give them chemotherapy and probably this happens in our stomachs as well the isthmus is wiped out these are rapidly proliferating cells and then this process that is just showed speeds up enormously and what normally takes half a year to a year happens in a matter of two weeks so this is again another strategy so we think that every tissue has its own strategy its own way of building stem cell hierarchies rapidly dividing progenitor population people don't want to call these cells stem cells but then at the far end of the hierarchy there's a differentiated cell that is happy at a lower rate normally but upon damage to the crypt at a much faster rate to replenish all other cell types in this particular tissue and you can take these chief cells and under the right conditions this is a one chief cell they will rapidly de-differentiate they'll make human stomach organoids and Siena who published this paper is now in Germany injected helicobacter and could show that because there's actually low pH in these stomachs and there's a thick layer of this gastric mucins that helicobacter drills through the mucins and engages the surface of the cells much like you would see in patients so this is now she tries to explore this as a model to study the biology of helicobacter the liver there essentially are two cell types in the liver that are really the liver cells that are made from the primordial liver the hepatocytes making up 80 to 90% of the mass of the liver they are the chemical engines that perform a lot of metabolic functions and detoxification etc. they produce bile so the bile that they produce so they have an apical surface that merges with the apical surface of a neighboring cell and forms bile canalically so the bile that's secreted from the blood through these cells into the bile canalically will flow into bile ducts and bile ducts are lined by the other cell type in the liver the cholangiocytes or the bile duct cells then there's blood vessels cupra cells come from the bone marrow stellate cells or mesenchymal cells etc but these are the only real liver cells you can actually sort them this is a human donor liver so you can get parasites out of a donor liver but also cholangiocytes it's probably the most regenerative organ in a mammalian body so if you're so unlucky to have liver metastases of Francis colon cancer surgeons are now capable of removing two-thirds of your liver with the metastases the liver will then grow back in a matter of three to five weeks it doesn't really look as pretty as the original liver but it's fully functional and will contain both new parasites new bile duct cells and all the other cell types and it's believed that it's because it's demonstrated that the parasites will enter the cell cycle and will very rapidly proliferate for a brief period of time and if all of them would divide once your liver would be twice as large so again there's lots of debates over stem cells in the liver the liver repairs itself from its fully differentiated cells so I guess many people and I would be in that in that team don't believe there is a professional stem cell in the liver there are parasites, the cholangiocytes and they make up the liver so Hemi-Hepatectomy, the surgical removal of half the liver but leaving a healthy fragment in place encourages the hepatocytes to divide if the entire liver is sick due to viruses, hepatitis viruses or after intoxication or something like that, there is something that pathologists has called a long time ago the oval cell response and these are cells that appear in the damaged liver near the bile duct so pathologists believe they come from the bile duct they are small, de-differentiated they have some early bile duct markers they will proliferate and can then make both cell types, so they are like a stem cell that appears to come from the bile duct now what Miri Hoog in my lab, who is now in Cambridge first for mouse but later for human cholangiocytes showed that so these are again fully differentiated beautiful cells by AM, they are highly specialized cells but if she sorts a thousand of these cells she gets three, four hundred organoids, so fully differentiated cells in this growth factor cocktail, in a matter of a few days they de-differentiate or they re-program or whatever you would want to call it they become bi-potent stem cells and if you then transplant them they can make both hepatocytes and cholangiocytes and she actually has she is trying to publish a paper where she shows what happens at the epigenetic level and DNA methylation appears to be crucial, so de-methylation of the genomes of these cells allows them, of these cells allows them to form these organoids bi-potent, so I think this is sort of a physiological re-programming that happens in the vivo open damage we can grow these organoids by confocal, they look a little weird so they have a very thin, yet by AM they are highly polarized large lumen, very thin layer so they, and if you characterize them by RNA-seq they have predominantly early cholangiocytes markers, we can actually induce them towards in the parasite fate, they are about as good as eye-habs, as hepatocytes are produced from IPS cells that don't really look like the real thing but they have the early transcription factors get all of them in and on our markers so this is how they grow and keep this in mind because I'll show you what happens when some of the genes are mutated so that these round bowls, very dynamic and they grow pretty well we can keep them growing for 5, 6, 7 months or so so binadetta artijani in the lab took these cholangiocytes to see if she could much like what I showed for colon cancer if she could now create cholangiocarcinomas and at the same time address what the roles of some of the not well characterized genes in this process that are mutated commonly but where the role is not very well known it would be nice if you can use organoids to understand the function by which a gene can act as a tumor suppressor or an oncogene I think this is a very nice example that was entirely done by binadetta so BAP1 is called BRCA1 associated protein probably has nothing to do with BRCA1 the fly homologue that looks like the homologue is called calypso that is very very clearly demonstrated to be a polycomb suppressor the ubiquitinase and it targets and this is very well documented this residue in histone 2a now there's a lot of literature on the role of BAP1 in human cancer you see a whole list of cancers we find this one interesting here for the current story and I have it here if you then look what these papers report so we think there's no good models first of all because the mouse models the null allele is embryonic lethal that doesn't help you and the systemic conditional deletion in an adult mouse causes myeloids myeloid leukemias which is not what humans get in the liver and if you read papers there's lots of functions assigned to BAP1 that would all be explaining why BAP1 is a tumor suppressor and as you can see almost all of them are functions in the cytoplasm and we get back to that but none of these studies actually refer to the fly study that clearly shows that if that gene calypso is the homologue of BAP1 it's a very well characterized tritorex component now what I did again with CRISPR she took wild type organoids colandesite organoids now also with colon targeted BAP1 works very well and she immediately noticed a very strong phenotype these are the wild type organoids and these BAP1 mutant organoids seem extremely condensed as you can see here these are from three different donors by HNE blow up of one organoid there's a little lumina looks like they really are starting to lose their epithelial context this is the wall of one of the colandesite organoids so this is where the major gel is the basal site this is the luminal site you see the cells are well polarized very thin layer there's very well defined junctions in the BAP1 mutant there is no polarization anymore also there are hardly any junctions they're very loosely held together so this we would probably classify as a classical example of EMT so you knock out a gene in the cells lose their epithelial context again that's simplified here so one stain tight junctions very close to the surface so you look on the inside of an organoid you get these beautiful stains here that you see and in the BAP1 mutant organoids really doesn't localize anywhere again implying no junctions are formed beta-catenin gives you the exact same picture and when you now make movies you remember the previous one this is what they grow like they look like these slimy monsters that actually fuse and pick up as you can see here really unpleasant creatures and there's nothing nothing epithelial about these cells now I would find it very difficult to explain all of this but she then did a lot of proteomics ATAC sequencing chromatin, modifier, etc and she essentially came up with a short list of genes about 300 genes that are up or down and actually they're both up and down and both groups appear to be directly controlled by BAP1 and by ATAC sequencing you can see definitely very big difference Claude in one here is the gene that we just stained for the ATAC peaks are entirely gone as soon as you hit BAP1 so it really looks like this is a nuclear regulator through the mechanism that has been proposed by the Zofla genetic system we find no evidence for any of the papers that were published on the human role and we should now re-express BAP1 in these mutant organoids in the cytoplasm by taking the nuclear localization signal off you can see by green that actually is expressed nicely this is a movie that runs over about 24 hours but nothing much changes they retain this mesenchymal but if you now re-express in this movie it only takes 24 hours re-express it in these, they are very nicely in a matter of 24 hours reform a normal epithelium and if you now actually with all the confocal markers we have they very nicely now are polarized and you have all the junctions and the integrins etc. where they should have them so we think that why BAP1 is too much suppressor at least in this cell type the fact that it controls a fairly although it's a chromatin regulator a fairly tight program of only 300 genes where the vast majority of these genes when the functions are known you could assign a function in epithelial aspects of cells and when you take it out you lose all of that and you end up with this massive problem here now when you transplant these sort of BAP1 or BAP1 P53 they don't really make tumors but the moment you so she then made organoids P53, SMET4, NF1 and P10 mutations all homozygous so this is typically a combination you'll see E.C. carcinoma if you transplant these as I'll show you nothing much happens but she then on top of that she mutated BAP1 again she goes from here so these look pretty normal despite the fact that they have four mutations four genes this is now adding a fifth one they immediately take on this BAP1 phenotype when you transplant the four mutation these are pathologists where is it here would call this an adenoma in the liver and as he says it looks exactly like a human adenoma this is a mouse obviously the moment you add these you get this aspect and as he says this is equivalent to carcinoma and there's a strong stromal reaction so these are mouse fibroblasts that surround these things so again showing that you can use these organoids well this I think shows that you can use these organoids to build cancer models in a very defined way but also you can solve the roles of individual genes that you can find hundreds of them in the TCGA database where the function is really not known mostly there are too much suppressors so with this approach you could very rapidly just go through these genes knock them out and see how they contribute to the carcinogenic process getting back to the to the two cell types so this one I just showed you these actually grow these are the best protocols they grow exceedingly well almost as if a single cell will always make an organoid your parasites have been extremely important in toxicology for instance but also many other aspects of drug development you can buy freshly isolated parasites from rejected donor livers from companies and pharma use these a lot the problem with the parasites is the moment you put them in culture they deteriorate and after about three years you can't use them and you have to open your next file so it's difficult to do real experiments there's no way you could ever do defined CRISPR experiments etc so Huili Hu in the lab I don't have her name here Huili is now back in China she came up with a protocol that was published late last year but you could take first in mouse a single parasite we know it's a parasite rose of tomato so it's red and you can see it actually has a large hepatocyte 14 micrometers or something like that it'll grow and it'll make these very differently looking organoids that we've never seen before so 3D reconstruction but also on a this is now human but also on a single plane confocal there's essentially no lumen no large lumen in these organoids you can see a dividing cell here they're not the fastest growers if you do single cell sequencing they look like a parasite some of them are proliferating others are post mitotic there's some earlier forms but there's nothing that looks like an oval cell or like a stem cell you can see they're very large they make a lot of alpha antitripsin one of the abundant proteins in the serum they make enormous amounts of albumin similar to same levels as you'd see in a freshly isolated parasite from a liver also they are highly structured much more structured than we've ever seen before what you see here is a stain for MRP2 MRP2 marks the it's actually a bile transporter it marks the apical site that forms that forms the canalically that starts inside the parasite actually flows between the parasites into the real bile duct and maybe I can see that here so here you see these large hepatocytes large cytoplasm, oval nuclei there's sometimes tetraploid as you'd see in a liver sometimes octaploid actually the tetraploid is devide we see that and you see these canalically that start inside one cell, run between the parasites and then end up in this central space that's an artifact that's where we believe the bile duct would normally reside to take the bile out of the liver and benedetta is now together with the lilahendrix is now trying to combine the two types of organoids and see if we get more complex structures where we have a parasite part and a clangicide part and build the entire system and again the stress this is done by Huili they grow best from fetal liver the older the livers are the slower they grow and so say a 60 year old donor we can grow for 3 months and then they peter out we have not seen this in other organoids interestingly in cancer liver cancer are particularly known for activating TERT mutations and so they apparently in the biology of liver cells it appears that you need to activate TERT by a mutation to have a cancer grow and that's what we see so they grow and this is the only organoid where the telomers shorten and they essentially stop growing P21, P16 go up when the telomers are totally gone these are primary hepatocytes on a number of mature albumin hepatocyte markers these are from various donors fetal and adults the various organoids and these are the bile duct organoids you can see they do not or they express very little of the genes that would be present in hepatocytes if you would run IHAPs here they would be somewhere here so they would have some expression of so IHAPs are the IPS derived hepatocytes, it would have some cytochromes it would have some albumin and never to the level of what a primary hepatocyte has they're transplantable so there's very nice models developed by Marcus Grumpy and this was done in Ypidiong's lab in New York these mice developed by Marcus are FAH-negative if you block this metabolic pathway with a small molecule the mice are fine but the moment you give them normal diet they will start destroying their hepatocytes and they have been developed by Marcus as a transplantation model and indeed if you and the cells are injected in the spleen it's not a very efficient route but they will make their way into the liver and then they make these large islands so this is a mouse and we stain for human albumin they make these large islands of hepatocytes we analyze these mice after three months but by KS67 you can see there's still many defining hepatocytes so they go on and on and we think if you would wait it longer that's what we're currently doing this island and another one here will slowly have filled up the entire liver and will actually have cured the mouse of this genetic disease and indeed if you look at albumin production so typically you have 10, 20, 30 milligrams I don't know exactly in a mouse of albumin and if transplantation you're happy if you see micrograms but these will actually get close to the milligram level here from the small islets that we now have in the liver and again we would believe if we had waited longer the lines still go up after at the 90-day time period we would have seen more and more human albumin in the circulation of these mice two more stories one is we've been trying to develop organoids for studying various parasites and pathogens so cryptosporidium is a malaria related parasite that will go through a very complicated life cycle but in one host so it has asexual cycle that comes in as an osist opens up into four sporezoides they will infect cells in the gut they will then go through these mirrored stages so this basically amplifies their numbers then after a while they will produce the microgammon that's like the testis and microgammon is like the ovaries so there will be a single cell here many cells here this will lead to a sexual so to the formation of a zygote and that zygote then makes four osis and actually each of these stages has a different number of nuclei so you can quite easily recognize them by dipestane and there's actually another movie here so this parasite is a major problem for AIDS patients or for immunodeficient patients in general also in the third world when there is malnutrition patients get problems with this parasite and there's no good in vitro models to complete this whole life cycle I think there's a firm in Texas that infects calves and they in their feces there are massive amounts of these osis you can buy them and the idea would be that we inject them inside human small intestine or colon or lung that's where you typically find these and that then we would see if you could complete the entire life cycle so here you see an osis the four zoroides come out they will find a host cell make a trophazoroide and then they go through these amplification stages type 1 meronts are formed I think there must be 8 or so of these things in here so they come out of this they will actually swim around in fact a secondary cell that should be happening here and then you get a type 2 meront so we're in the life cycle we're now here so there it goes type 2 meront now this organism is ready to go into the sexual stages so for this you need to make these micro and macrosisons and that's where they are so this one is full this one is full of the male version of this they can actually freely swim this one is the lucky one that finds the macrosisons first fertilizes it now we get these four osis formed and this is then secreted by the infected animal in the feces and ready to infect another animal or an unlucky human being with a defective immune system so this works very well so we inject it actually the first time it worked so here you see the osis so you see these little snakes coming out they are the sporozoites presumably so we saw also these spectacular things by EM they are more easily recognized so in invading sporozoites hours after we inject the osis the whole thing takes about four days or so you go through the meront type two the meront one is not here the macrogamond the zygote that makes the osis and here you have the developing osis again and you can actually so what we also did is we infected we did not inject osis we actually purified sporozoites so they cannot infect you cannot feed them to mice because they will be destroyed they need to be in an osis so we inject the sporozoites four days later we ground up the organoids and then test whether there are osis formed when you feed these organoids to mice they will get a very productive infection now we can do this two or three so in patients this is a chronic infection we can passage them two or three times but logistically it is very difficult to keep the parasites in if you passage them you break open the organoids and but this Lutolf now has a continuous gut growing this is a combination of bioengineering and stem cell biology so in a gel he has a major lumen and some side lumens if he puts miniguds in he gets a functional gut that you can flush and so has crypts stem cells has all the other cell types when you put osis in you see that they infect well they don't have his data here they infect the crypts and they sit at the crypt to villus junction and they constantly re-infect cells that come out of the crypt and then when they die they essentially poop out these osis that every day you can flush this gut and harvest osis from these and he's been following this now for a month and this gut so it's an entirely beautiful system so finally this is a paper that's conditionally rejected as it happens to most of our papers snake venom gland organoids so I don't know why this happens this was these three PhD students so they at one point they told me well you know Hans we're growing snake venom gland organoids and I said well you know snakes are reptilians so our growth vectors are not supposed to work but they work very well so somehow they lay their hands on a cape coral snake but not because we cannot kill we have very few snakes in Holland as you can realize and you're not allowed to kill snakes but from breeders we can get eggs that are close to hatching and when this little snake this cape coral snake would come out of the egg it's ready to hunt so it's a complete snake it has fangs and it has eyes and so it can actually catch a small mice so what they did is they dissected this and they took they got the venom gland out and since then there's this guy I always forget his name in Australia was named by Stingray who was always wrestling with crocodiles and what's his name Steve Urban so we have a Dutch version of Steve Urban who goes around the world wrestles with crocodiles has a big shark scar on his shoulder and so he sends us now because this works bits of snakes from around the world and actually they're beautiful organoids they've been growing for two years now you can see so this is the inside so there's this dark matter that turns out to be highly concentrated poison or venom there's also the ciliated cells we can expand them very rapidly when we drive them into differentiation by taking the growth factors out they make these venoms then by single cell sequencing we're always we had a lot of contamination problems in the lab at one point and they were worried that these were not snake and there was so they actually did the single cell sequencing got the data back and then realized there was no annotated genome so this was a project with some hurdles and and so then they realized actually the king cobra is in the same class of snakes and there was actually an annotated genome so they helped them and then they made this observation that they showed to me so this is clearly a snake and this was their single cell sequence pattern so they thought that it was actually so what we find is that there is a stem cell like population these organisms actually have elder 5 snake elder 5, they have some other markers that we know from the gut and from the parotid gland so the venom gland is presumably evolutionary derived from the parotid gland and parotid glands from humans also grow very well and then we see all these different that's actually not known so these venom glands make about 20, 30, 40 different toxin proteins that fall in families that try finger toxins they're here I don't know what these are, they're sort of a lectin type toxins they affect coagulation they affect, they kill erythrocytes there's a lot of them that have neurotoxic effects so there's probably a large number of molecules that you could explore for drugs Botox, well it doesn't come from these but I'm sure you could find botox like molecules here also we find there are different lineages so there are different lineages that make different sets of toxins that was not known and then to know whether they really are relatively small proteins highly modified and if you express the genes in hack 293 cells they make the protein but it's not a toxin because it doesn't have the right disulfide bridges like oscillation patterns so the question was do these organoids make the right toxins, can you maybe use them as a production platform to make toxins and that turns by mass pack you can see this is venom so these snakes get milked and here you see the masses of the three major proteins of this Cape Coral snake and this is the same measurement of the toxins that come out of our organoids essentially they are identical in molecular ways so we believe they are perfectly correctly produced actually they have been these three guys in the lab have been killing all sorts of things cells and with these so they really look like toxins and so I hope that actually we will be able to publish this paper somewhere and with that I named these people I'd like to thank you very much for your attention or maybe I should say one more thing so the third guy on that list Jorick Post is now growing tear glands human tear glands he can make them cry so this guy the Dutch Steve Irvin is now getting us crocodile tear glands and see if he could grow those as well anyway that's exactly how we do it for colon cancer we leave wind out of the medium 99 out of 100 colon cancers are wind independent and so that's an easy trick for pancreas cancer they invariably have RAS mutations and P53 mutations so you can have nut lint or leave EGF out but for prostate cancer there is not a common pathway that would be affected that you could use for this also no good surface marker so for most cancers but also some cancers are just very circumscript so if you have a good pathology it will give you a clean cancer sample and then you only have cancer cells yeah I would argue that stem cells do divide asymmetrically but it's not an intrinsic phenotype it's because of the space restrained because they can only produce daughters that go up in organoids well they move away from where the stem cells are so these stomach organoids have little buds that's where the stem cells sit and the daughter cells move out and move on to the major surface of the organoid so there is a directional it's not as clean cut as in the stomach but as directional growth you see it better in intestinal organoids than in stomach organoids in the liver or these cholangiocytes organoids are fully round bowls and there's just random cells that are more progenitors that divide that slowly grow and to make differentiated cells we take the growth factors out everything that I showed you is from a dull tissue not from embryonic stem cells so the second so Jim Wells not so far away from here he got from IPS cells or ES cells although he doesn't state explicitly his final stage is actually is R-medium so once he has and then he doesn't have the problem that many other labs will have because once you have reached the gut stage you put him in R-medium and they grow forever so it's not as if you end up with a structure that you then have to analyze so that is, but so we have no clue how to specify ES cells to this stage but we know very well when you're in this stage and so the attraction of this approach is like the bone marrow stem cells they are in a fate you cannot push them out of their fate so bone marrow stem cells will never make anything but hemopoietic cells and these cells will only make the tissue where they come from so they're safe we think there are actually our collaborators I showed yesterday you can transplant these quite well into the gut our collaborators in Japan also helped by the relaxed regulatory environment because Shinya Yamanaka they have now started what I hear to try to use human colon organoids as a treatment in IBD I'm not really sure of that because you probably don't take away the cause of IBD by giving epithelium back but it's at least a first try to see if it's safe to use these organoids in patients yeah but it's done so because you can not sure what it does it's probably a matched allergenic cause it's never acute you can if you build a large biobank you can just go like an RPS you can just go to your biobank get the best match and transplant we are involved in a big program to do this for liver because there are really no good indications for cut epithelial transplants but there is vast indications for liver transplants and that is so we're now at the stage that we can actually grow at scale so we can grow very large numbers of these cells we have all the conditions now under GMP but we keep on running into the regulatory authorities that we have to show that they're safe before we can give it to patients so this is a catch-22 situation so but that's another so as I said this is a parotid gland transplant because it's an easily accessible organ so it's easier you inject through the skin human cells in mouse work very well and Rob Koppers who does this is now getting ready to do this in patients after radiation for instance for lost parotid gland and it's terrible you know your teeth a lot of your tongue goes you cannot speak anymore it's very difficult to eat and it looks like that he can repair it with these parotid glands and they like the venom glands they grow fantastically well yeah so well many people say you have no stroma in these organoids so these are incomplete tumors I would argue that actually we give them what the stroma would normally provide so we have a much better defined although there are cells it's growth factors that are made by the cells there's no immune elements we actually but as many other labs now we can quite easily combine them by adding alphabeta T cells dendritic cells gamma delta T cells and you can recapitulate everything that's known for instance that T cells will home into these so we do it with TILs the TILs from the original tumor from which we grew the organoids often what we see is the T cells will actually move to the organoid but then get stuck on the outside and then if you add checkpoint inhibitors PdL1 antibodies is what we use they will actually go in and kill the cancer cells so but this is all non-confocal not standardized and I know from multiple other labs that because the problem with checkpoint therapy is that there's no good non-human models mice will never predict what you'll need and in vivo so you can only do clinical studies there is about 5000 clinical studies currently for checkpoint inhibitors combined with any conceivable cancer drug and the companies have no idea what they should prioritize so there are all sorts of efforts to humanize mice and we think that maybe these organoids could be a simple way in to see what the parameters are to prioritize these clinical trials because there's billions of dollars go in there and many of the overwhelming majority of trials is negative so it would be good if you have some filter before you do that so it looks like you can do it but it's early days yeah so Mary who also did the original liver organoids in my lab has described ways to grow pancreas organoids but this is the exocrine pancreas so we have not been able to there's some indication that you can push them under very specific conditions into a hormone producing lineage if you then transplant them together with a fetal mouse pancreas into a mouse you'll see insulin producing cells but this is a very awkward essay so I would not argue that we can make so in the gut we can make it very easily but there's almost the same cells in the pancreas we cannot make them having said that Ariel Zhang a collaborator of us in China appears to have found conditions to take adult islets and grow very large numbers of clones of what looks like genuine hormone-secreting cells alpha beta delta from these adults again in Matrigel with I don't know this is a growth factor condition that she's not yet revealed she's trying to publish this but we cannot do it okay maybe two more questions and then he has to get to the airport sadly so how about the one from Caroline the one from we don't have her mice we have looked at our various single cell sequencing data and we do see the expression that she reports it's very rare cells though I would argue that it doesn't matter whether they're in defined plastic environment or in Matrigel but actually so what we do with Matthias Lütholf he is the buyer engineer so he makes these designs mini chips we put our organoids on and then actually they look more like a gut the function doesn't change so organ on a chip is a little bit like lab on a chip it sounds fantastic but I don't know it would be interesting if you could combine sort of series that would be very difficult for us so like gut and liver and brain or something like that or at the microbiome but just as a replacement of these cultures I mean it's pretty tedious and not cheap and I don't think it really has advantages so I'd like to take this