Summary of ICMS Seminar Club presentation on Friday, February 25, 2022.
Title: ORGANOIDS TO MODEL HUMAN DISEASE
Speaker: Prof. Hans (J.C.) Clevers, MD, PhD, Professor of Molecular Genetics at Utrecht University, Principal Investigator at the Hubrecht Institute and the Princess Máxima Center for Pediatric Oncology and Oncode Investigator, in Utrecht, the Netherlands. From March 2022: Head of Roche Pharmaceutical Research and Early Development (pRED).
On Friday, February 25, Prof. Hans Clevers gave a presentation at Fujita Health University. He talked about how he and his group developed organoids from adult stem cells and how they are using them to research human diseases.
Recording: For members of Fujita University, a recording of the meeting (without the discussion part) is available for a three-week period at our Manabi system. Unfortunately, we cannot open the recording for a wider audience.
There were 82 participants, which for us is a record. I found it a great presentation and so did most other participants whom I asked. Beautiful descriptions came from Prof. Jim Kaufman in an e-mail to me, and he allowed me to use a few sentences from that mail for posting in this blog: — “I learned a lot in a short time, not having ever followed this field, and I was as impressed as you seem to be”; — “I had always wondered how such apparent leaps of logic allowed a person like that to make such amazing discoveries, but he very humbly made it clear that it was just one small step after another”; —”Such a keen sense of what to do–wonderful wonderful wonderful. And the stories of what they are doing with this work, particularly figuring out how SARS-2 infects epithelial cells rather than what was found for VERO cells, the three PhD students putting together the oncogenic E. coli story, and mapping the mutations against growth factors for adenocarcinomas. Spectacular! And he provides the basis for asking questions like how these stem cells make thousands of divisions without evidence for enormous accumulation of errors.” I couldn’t agree more with what Prof. Kaufman says, and, naturally (as Prof. Kaufman is one of the best writers I know), I couldn’t write it as beautifully.
What attendees also noted, and which they were also interested in from a technical point of view, were the beautiful computer animations of organoid/tissue development that Prof. Clevers showed. Professor Clevers explained that he works together with a company, Nymus 3D or Digizime, and that the process involves him making a series of drawings for an old school type of—few frames per time unit (remember the first Donald Duck movies) —movie, which then in a back-and-forth discussion with the company is shaped into a smooth computerized movie. He expressed that it was not so expensive, but such perception may differ between researchers.
I think that for all participants it was clear that Professor Clevers, despite his huge successes and many awards, stays a very nice, normal, and kind person. His humble attitude towards science is reflected in his expression that “we should never try to outsmart nature.” I am very glad that someone like him will take the helm at Roche pRED, where he will be guiding >2000 scientists in translational research.
Those among you who like reading may like to know that Prof. Clevers—an avid reader—told us (in the chat before his presentation) that his favorite (non-science) book is “The Worst Journey in the World.” This book describes the second Arctic expedition by Scott in 1910-1913 and was written by one of his team members, Apsley Cherry-Garrard.
The contents of the presentation
In the below, I summarize the contents of Prof. Clevers’ presentation. For readers who want more information, I point out that there are several beautiful videos with his presentations on the website of his group at the Hubrecht Institute.
The discovery of intestinal stem cells
Briefly summarized, the journey of Prof. Clevers to discover intestinal stem cells included a number of major discoveries including, but not only (besides what Prof. Clevers told us, I here add some information from literature):
• Finding a T cell-specific enhancer element in the CD3e gene (Clevers et al. 1989).
• Discovery of a transcription factor that bound to this enhancer element; they named this transcription factor TCF-1 (for T-cell specific transcription factor (van de Wetering et al. 1991). In the same year—submitted very shortly after the publication by the Clevers group, suggesting independent acquisition of the data—other groups found another member of this transcription factor family, called lymphoid-enhancer factor 1 (LEF1; Travis et al. 1991 and Waterman et al. 1991). Eventually, the Clevers group found that humans have four genes of this TCF/LEF family (Castrop et al. 1992).
• Discovery that TCF\LEF proteins are transcriptional mediators of the Wnt pathway as they form a bipartite transcription factor together with b-catenin (or, in Drosophila, its homologue Armadillo) (Molenaar et al. 1996; van de Wetering et al. 1997; Clevers and van de Wetering 1997 ). This was simultaneously (as I understand it) discovered by the Birchmeier group (Behrens et al. 1996).
About the Wnt family (Nusse et al. 1991), the name of which is a sound-combination of the two words wing and int, I here quote Bejsovec 2006: “The Wnt gene family itself derives its name from wingless (wg), a fly gene, and int-1, a mouse gene identified as an integration site for Mouse Mammary Tumor Virus (MMTV). Loss of function mutations in wg were recovered based on their disruption of body pattern in the fly (Sharma and Chopra, 1976; Babu, 1977; Nüsslein-Volhard and Wieschaus, 1980), whereas gain of int-1 function, through retrovirally promoted expression, produced tumors in mice (Nusse and Varmus, 1982; Nusse et al., 1984).” Thus, in 1991, it was already known that the soluble messengers of this system, the Wnt factors (of which there are >10 in humans), are important for cell growth in both embryology and cancer.
• The Clevers group was the first to link Wnt signaling with adult stem cell biology, by showing that TCF4 gene disruption leads to the abolition of crypts of the small intestine (Korinek et al. 1998), and earlier by showing that TCF1 gene knockout severely disables the stem cell compartment of the thymus (Verbeek et al. 1995).
• Identification, by microarray, of genes induced by Wnt signaling in colorectal cancer cells (van de Wetering et al. 2002) and finding that 17 among them were the most interesting as potential stem cell markers as they were associated (in healthy mice) with the crypt but were not Paneth cell markers (van der Flier 2007; note their early realization that growth of tumor cells and stem cells show many similarities). Among these 17, they identified Lgr5 as the crypt-specific gene with the most promising expression pattern for being a potential marker for intestinal stem cells, as it labeled dividing cells at the crypt base (Barker et al. 2007) and earlier its expression had been found to be extinguished in the presence of a dominant-negative TCF4 that inhibited Wnt signaling (van de Wetering et al. 2002). LGR5 (leucine-rich repeat-containing G-protein coupled receptor 5) is a receptor that the Clevers group later found to enhance Wnt signaling when binding to R-spondins (de Lau et al. 2011), factors that in the intestine can be secreted from pericryptal stromal cells (Greicius et al. 2018). Other researchers later found that this was caused by the LGR5/R-spondin complex inactivating the Wnt-antagonistic molecules RNF43/ZNRF3 (Xie et al. 2013) (Fig. 1). Professor Clevers explained that LGR5 is necessary for the strong and continuous amplification of the Wnt pathway as seen in intestinal stem cells and that this is different from the nuanced regulation of Wnt pathways during embryo development.
• Providing evidence that LGR5 is indeed a marker for adult intestinal stem cells. They did this by placing a GFP marker behind the Lgr5 promoter in recombinant mice to fluorescently mark LGR5+ stem cells (Fig.2; Barker et al. 2007). Simultaneously, they also placed CreER (an estrogen receptor variant of Cre recombinase which requires tamoxifen for its activation) behind the Lgr5 promoter and used a Rosa26-lacZ reporter mouse strain in which Cre-Lox recombination induced by tamoxifen-activated-CreER (they injected the mouse with tamoxifen so that cells with an active Lgr5 promoter permanently induced LacZ expression) led to permanent blue cell progeny (Fig. 3A). All intestinal cell types were found to color blue in this system, concluding that the LGR5+ cells are the stem cell source for all the intestinal epithelial cells (Fig. 3A; Barker et al. 2007; Gehart and Clevers 2018). They showed that the LGR5+ crypt cells had true stem cell characteristics as the genetically marked LGR5+ crypt cells kept generating the other intestinal epithelial cell types for the remaining of a mouse’s life. Later, LGR5 was also found to be a marker for adult stem cells in several other tissues (Koo and Clevers 2014).
• Cells derived from an intestinal stem cell only populate the villi that are neighboring the crypt. They provided evidence for this by using different colors as permanent genetic markers for individual stem cells (Figs. 3B and 3C; Snippert et al. 2010). In this system, all crypts plus their surroundings have only one color which is evidence that there is competitive elimination between stem cells within one crypt (but not between crypts); furthermore, that these colors extend only to the top of the villi is evidence that there the cells die, and evidence was obtained that it takes about five days from generation in the crypt to traveling towards and dying at the villus top (Barker et al. 2007; Clevers 2013).
• When the crypt is injured, cells from the villus regions can migrate back into the crypt and become stem cells (Murata et al. 2020). This dedifferentiation from epithelial cells into stem cells also may be an important feature for enabling the development of more than one crypt in intestinal organoids (see the end of this post).
Organoids from intestine and other organs
The ones among you who worked with cell cultures, especially the non-commercial ones, know that they need a sensitive touch—you need to “understand” the cells and handle (e.g., passage) them with care. In the Clevers group, it was the Japanese postdoc Dr. Toshiro SATO (now professor at Keio University) who took the difficult task to try to grow the intestinal stem cells in a petri-dish. At the time, it was an unpopular assignment, because the general belief was that such was impossible. But here, the skills and dedication of Dr. Sato were combined with the accumulated knowledge in the Clevers lab that allowed the isolation of stem cells and the making of a reasonable guess about the necessary growth factors. The chosen growth medium included no serum but just nutrients and the factors:
- R-spondin1 (a Wnt signaling agonist; they wanted to stimulate the Wnt pathway, but Wnt factors themselves are very hard to work with because they are palmitoylated [a lipid chain is attached] and therefore poorly soluble)
- EGF (epidermal growth factor; this stimulates RAS signaling, which is a major pathway for cell proliferation and other processes)
- Noggin (this inhibits signaling by BMPs, and thereby inhibits a normal control of cell development; for example, when noggin is overexpressed, many more crypts are formed in the intestine [He et al. 2004])
- Matrigel (this is solubilized basement membrane matrix secreted by mouse sarcoma cells; if I understand correctly, this helps the cells with polarization and forming 3D structures)
And Dr. Sato made it work. I have come across an interview of Prof. Clevers (link) in which he explained that he hadn’t spoken to Dr. Sato about the topic for a while when he asked him how the cells were doing. “Oh, they are doing fine,” was the reply, and they had apparently been growing for three months (!). It may have needed another touch of brilliance to discover that these clumps of cells produced “mini-intestines” (more correctly, the epithelial part of the intestine), and they then were coined “organoids” (Sato et al. 2009). It was discovered that these organoids form luminal structures that have crypts surrounded by (flat) “villus-like” regions (Fig. 4) and include all the epithelial cell types that are also found in the intestine.
Using adult stem cells and similar protocols (although tweaked for the tissue in question) as originally used by Dr. Sato, organoids have now been established for a variety of tissues, amongst which: Stomach, small intestine, colon, liver, pancreas, prostate, lung, ovary, mammary gland, salivary/tear gland, taste buds, and inner ear. It is important to realize that what we conceive as an “organ” is only partially replicated in these organoids, and, for example, for the liver Prof. Clevers used the word “hepatocyte organoids” instead of “liver organoids” because they do not include bile duct cells. Prof. Clevers explained that there are developments worldwide to combine more cell/tissue-types into a single organoid, but that the circulation and nervature of in vivo systems may never be fully replicated.
The use of Organoids to model Human Disease
Organoids provide a system in between cell culture and in vivo systems and are superior in some respects to each of those. Prof. Clevers explained that organoids derived from adult stem cells can, at least in some cell systems, also be superior to those derived from induced pluripotent stem (iPS) cells as they more reliably replicate the features of adult tissue.
I here list some of the examples given by Prof. Clevers for how his group (plus others) has used organoids to study human disease:
Cancer treatment success prediction:
Organoids can relatively easily be grown from epithelial (adenocarcinoma) cancer cells and retain important properties of the cancer, including sensitivities to some drugs (e.g., Vlachogiannis et al. 2018; Francies et al. 2019; Ooft et al. 2019). Naturally, it depends on the type of cancer, drug, and assay, but Prof. Clevers summarized several studies as that 70-95% of drug sensitivities found for cancer organoids correlated with the treatment response in the patient from which the cancer cells had been derived. He also summarized that for many treatments only 40% of patients have a positive response, so prescreening at the organoid level may reduce unnecessary costs and side-effects resulting from ineffective treatments.
Cancer growth science:
The Clevers group (Drost et al. 2015) and, independently, the group of Prof. Toshiro Sato (Matano et al. 2015), realized that the growth factors they needed to grow organoids (see Fig. 4) corresponded to the pathways that colorectal (and many other) cancer cells need to be permanently changed for becoming aggressive (Fearan and Vogelstein 1990). By mutation of human intestinal organoids using CRISPR/Cas9, they found that: (i) Knockout of the gene for APC, which is part of the Wnt-pathway inhibitor complex (Fig. 1A), makes the organoids independent of Wnt agonists such as R-spondin; (2) Adding a gene for constantly active KRAS (KRAS-G12D) removed the dependence on RAS-pathway stimulators such as EGF; and (3) Knockout of the transcription factor SMAD4, which is critical for BMP signaling, removed the dependence on a BMP signaling inhibitor such as noggin. Thus, by their understanding of stem cell dependencies on niche factors, the Clevers group could create cancerous organoids by genetically removing those dependencies. This very much agrees with the “Vogelgram” by Fearan and Vogelstein 1990 who found dependencies of metastasizing colorectal cancer on mutations in these types of factors together with mutations in p53. P53 is a transcription factor involved in many processes, including DNA stability and repair, and its mutation leads to an accumulation of DNA damage. The Clevers group found that in human colon organoids besides the above-mentioned mutations for APC, KRAS, and SMAD4, also p53 needed to be mutated for the organoids to show invasive growth and metastasis when orthotopically transferred to a colon of an immunodeficient mouse. Thus, even when the Wnt and Ras pathways are continuously activated, and the Bmp control is blocked, there are still control mechanisms that keep the cells from unbridled growth in vivo.
Tissue transplantation:
Very exciting is that Prof. Mamoru WATANABE, in collaboration with the Clevers group who sent them the organoid cells, could show that in a mouse colitis model (induced by dextran sulfate sodium) rectally (by enema) injected colon organoids could form healthy tissue at the ulcera (Yui et al. 2012) (Fig. 5). Prof. Clevers described this as a living “band-aid.” Such type of tissue repair is one of the most promising prospects of organoids, as they can be relatively easily grown from a patient (so the grafts won’t experience allograft/histocompatibility rejection) and have a much more robust chance for survival than single cells.
Virus replication:
Organoids can allow a more efficient and/or more natural propagation of some viruses that are otherwise hard to grow in culture, probably because in organoids the cells correspond better with the natural cell types (and their polarizations) than in a cell monolayer culture. Prof. Clevers showed us how hepatocyte organoids are used for the replication of HBV and, for the first time in vitro, allow long-term infection and development of a latent state (a state in which the virus is still present but inactive and waits for activation, sometimes many years later) (De Crignis et al. 2021). The latent state of a virus is particularly hard to fight, thus having such a system in vitro is a huge methodological advantage.
The DNA mutation inducing ability of colon-resident pks+ E.coli:
Pks is a 60 kb operon with >10 genes that is part of the genomic DNA of some E. coli strains found in the gut of about 20% of healthy individuals. These genes form a synthesis pathway for a polyketide called “colibactin.” Pks+ E.coli were known to induce DNA breaks in human cell cultures (Nougayrède et al. 2006). Pks+ E.coli are also used as probiotics to alleviate intestinal problems (Sonnenborn 2016), which is somewhat worrying considering the DNA damaging properties (Olier et al. 2012). The Clevers group used colon organoids to confirm that pks+ E.coli induce DNA breaks indeed and found an associated mutation motif (probably resulting from DNA breakage at colibactin-binding motifs and incorrect repair). They proposed a logical mechanism for the occurrence of these mutations and found them to be more abundant in colorectal cancer than in other cancer types (Pleguezuelos-Manzano, Puschhof, Rosendahl Huber et al. 2020); they suggest that there may be a causative correlation between the mutations and colon cancer. Therefore, Prof. Clevers believes that the use of colibactin-producing E.coli as a probiotic needs reconsideration.
COVID:
The Clevers group found that SARS-CoV-2, the COVID-19 virus, can readily infect intestinal organoids (Lamers et al. 2020). By mutation analysis, they found that this infection was not only dependent on ACE2, which is the main receptor for SARS-CoV-2, but also on TMPRSS2, a transmembrane protein that activates the virus spike protein by cleaving it (Beumer et al. 2021). In this system, they did not find any inhibitory effect of chloroquine, a drug that inhibits normal endocytosis and for a while has been advocated by some as an anti-COVID drug but probably does not have beneficial effects. Prof. Clevers explained that the entry route of the virus differs per cell type and that most virologists use Vero cells (green monkey kidney epithelial) in which endocytosis is a major route for SARS-CoV-2 entry so that spike activation by TMPRSS2 is not necessary. It would be interesting if organoid-based research can help find out why only in some people the virus can efficiently replicate in the lungs.
Some model playing
Serious readers should probably stop reading here. However, contemplating the studies by the Clevers group I kept wondering why enterocytes die at the top of the villi. Prof. Clevers said that the reason for that is not well known.
Prof. Clevers said that in organoids the lifespan of enterocytes is similar to that in vivo and that the dying of cells is likewise in the villus-like regions. So, their short lifespan (they live only 5 days or so) can’t be caused by the unusual stress of the intestinal environment, and there must be some endogenous reason.
Therefore, I came up with the below model, which is so simple that probably someone has had it before me, but I am not a specialist in this field and could not find a similar model in my (superficial) literature searches. If someone knows a similar model, please let me know, then I will add that information here.
In the model, for simplicity, I only distinguish three cell types: (1) Stem cells, (2) Paneth cells, and (3) Enterocytes. The stem cells and Paneth cells like each other and will maximize their mutual contacts by regular interspersing in the crypt, while they and the enterocytes don’t like each other as much and therefore spatially separate. But that is not a unique idea and not shown in the below figure.
Fig. 6 shows the interesting part of the model. It is based on the concepts of cellular competition (e.g., Vincent et al. 2013) and dedifferentiation.
As for the competition: I propose that at the time of generation from stem cells, the new enterocytes acquire a competitive factor (giving a blue color in Fig. 6) for pushing away other enterocytes. If there is no place for the older cells (which have less of the competitive factor and therefore lose the battle) to go, they are pushed out of the epithelial layer while undergoing apoptosis (Fig. 6A). This can explain why the cells cannot migrate into neighboring “crypt-associated-villar regions” (see Fig. 3C) because at the borders they meet younger cells that are generated by the neighboring crypt, and from those they lose the battle. This competitive system is maintained in intestinal organoids.
As for dedifferentiation: When the cells become even older (than the time they normally die because of competition), they start dedifferentiating back into stem cells, and then start making new crypts (Fig. 6B). This may explain how one organoid can have more than one crypt, and also how enterocytes can reconstitute damaged crypts in vivo. The above-mentioned competitive factor might be involved in the differentiation status, but it may also be an independent factor. In Fig. 6, I gave both a blue color.
Thus, essentially the model is based on an endogenous time-course of the gradual loss of competitive and differentiation factors (which might or might not be the same).
