|Home | About | Journals | Submit | Contact Us | Français|
We describe a protocol to generate 3-dimensional human intestinal tissue (called organoids) in vitro from human pluripotent stem cells. To generate intestinal organoids, pluripotent stem cells are first differentiated into FOXA2+/SOX17+ endoderm by treating the cells with ActivinA for 3 days. Following endoderm induction, the pluripotent stem cells are patterned into CDX2+ mid/hindgut tissue using FGF4 and WNT3a. During this patterning step, 3-dimensional mid/hindgut spheroids bud from the monolayer epithelium attached to the tissue culture dish. The 3-dimensional spheroids are further cultured in matrigel along with pro-intestinal growth factors, and proliferate and expand over 1–3 months to give rise to intestinal tissue, complete with intestinal mesenchyme and epithelium consisting of all of the major intestinal cell types. To date, this is the only method to efficiently direct differentiation of human pluripotent stem cells into 3-dimensional human intestinal tissue in vitro.
A critical but often underappreciated step in the development of vertebrates is the early patterning that occurs during and after gastrulation. These patterning events set up the anterior-posterior (A-P) and dorsal-ventral axes of the embryo. A-P patterning of the endoderm germ layer establishes foregut (anterior) and mid/hindgut (posterior) domains. The mid and hindgut gives rise to the small and large intestines1. Studies using several model organisms demonstrate that the FGF and WNT signaling pathways promote posterior patterning and specification of hindgut endoderm2–4. We hypothesized that the ability of FGF and/or WNT signaling to promote posterior endoderm patterning could be exploited to generate posterior mid- and hindgut lineages from human pluripotent stem cells (hPSCs). In support of this, we recently published a method that demonstrates that high levels of FGF4 and WNT3A act in synergy to pattern human PSC-derived definitive endoderm into mid/hindgut endoderm and promote a gut tube-like morphogenesis, resulting in the formation of 3-dimensional, mid/hindgut-like spheroids5. We then utilized a culture system that promotes expansion of adult intestinal stem cells6 to expand mid/hindgut spheroids into 3-demensional human intestinal tissue in culture.
We have demonstrated that formation of human intestinal tissue from PSCs is via a process that very closely mimics embryonic intestine development, giving rise to an intestinal epithelium containing absorptive enterocytes, as well as the major secretory lineages including Paneth cells, goblet cells and enteroendocrine cells. The human intestinal tissue is functional as it can secrete mucins into luminal structures, and can absorb fluorescently labeled dipeptides, suggesting a functional dipeptide transport system7,8. Intestinal tissues contain villus-like structures with microscopic brush borders as well as crypt-like structures that express intestinal stem cell markers. Lastly, we have demonstrated that this is a genetically tractable system that allows for both genetic gain- and loss-of function studies.
Generation of human intestinal tissue from hPSCs takes approximately one month: 3 days of exposure to ActivinA for definitive endoderm (DE) induction, 4 days of exposure to FGF4/WNT3A to generate mid/hindgut spheroids, and 14–28 days to allow spheroids to expand into intestinal tissue. During each of these steps, genetic manipulation and addition/removal of growth factors and/or chemical agonists and antagonists is possible to achieve different experimental outcomes. At each step, the appropriate controls should be included: during DE induction, a control excluding ActivinA is included to monitor the efficiency of DE differentiation; during the mid/hindgut patterning step, a control excluding FGF4 and or WNT3a is used to monitor both CDX2 induction as well as generation of 3-dimensional spheroids.
We believe that this culture system will be useful in a wide variety of applications, including functional studies of human intestine development, maturation, and function; studies to understand factors controlling intestinal epithelial proliferation, differentiation and renewal; a platform for high throughput analysis of drug transport and uptake studies; and infectious disease studies. In addition, using patient specific induced pluripotent stem cells (iPSCs) will allow in vitro disease modeling to study the etiology and molecular mechanisms of disease processes. Lastly, we think that this system may be a starting point for tissue replacement therapies either by direct engraftment into diseased bowel or using bioengineering approaches to generate intestinal segments in vitro.
Several groups have demonstrated that mouse embryonic stem cells and induced pluripotent stem cells have the capability to generate cells that express intestinal markers9–12. However, these methods utilize spontaneous differentiation of pluripotent stem cells into embryoid bodies (EB), which contain many cell types derived from all three primary germ layers13. More recently, it has been shown that endoderm and intestinal generation can be enhanced in EBs by using a two-step differentiation protocol14.
Our differentiation method represents a significant advance over previous protocols for several reasons; 1) This is the first demonstration of differentiation of human PSCs into intestinal tissue, 2) The protocol directs differentiation of PSCs through stages that mimic embryonic intestinal development and is highly efficient, resulting in organoids containing an epithelium, of which, more than 90% expresses CDX2 3) Three dimensional growth results in intestinal tissue with both secretory and absorptive function, 4) Organoids contain all of the intestinal epithelial cell types and similar architecture with villus- and crypt-like structures, 5) Organoids contain a stratified mesenchyme and express markers of smooth muscle and intestinal sub-epithelial myofibroblast (ISEMF) cells.
The method described herein does not rely on the inherent ability of ES cells to spontaneously differentiate into all tissue lineages. Furthermore, this is the first protocol to describe a defined set of growth factors that are required at each stage of differentiation to mimic embryonic development in vitro.
Despite offering such great potential, this system is not without its limitations. For example, the intestinal organoids lack several components of the intestine in vivo, such as the enteric nervous system as well as the vasculature, lymphatic and immune systems. Additionally, while all of the major epithelial cell types are generated in proportions similar to those found in vivo, and there is evidence of crypt-like domains housing stem cells, the 3-dimensional architecture is not as regular as is seen in vivo and the villus-like structures are variable from one organoid to the next.
Regardless of these drawbacks we believe that this system has extraordinary experimental utility for understanding and modeling human intestinal development, homeostasis, and disease. Moreover this system provides a viable starting point for future efforts at bioengineering human intestine.
Grow pluripotent stem cells in feeder free conditions on hESC qualified matrigel as previously described15 (National Stem Cell Bank Protocols, www.wicell.org). Briefly, hPSCs are cultured on hESC-qualified, matrigel-coated nunclon delta surface plates (6-wells) in a 5% CO2 incubator at 37°C. Cells are passaged onto new plates every 4–5 days using dispase (1mg/mL), as described by National Stem Cell Bank Protocols, www.wicell.org.
Resuspend dispase in Advanced DMEM:F12 to a final concentration of 1mg/mL. Filter sterilize using a Millipore filter sterilization tube, make 10mL aliquots and store aliquots at −20°C for up to 6 months.
Thaw hPSC qualified matrigel on ice, or overnight at 4°C. Chill sterile microcentrifuge tubes in a microcentrifuge rack at 4°C for 1 hour prior to aliquoting matrigel. Aliquot matrigel into cold microcentrifuge tubes and store at −80°C for up to 6 months. CRITICAL: Matrigel will solidify at room temperature (20–25°C), so it is important to work quickly and keep matrigel cold throughout the aliquoting process.
Coat 6 or 24 well Nunclon delta surface plates with hESC-qualified matrigel as described by manufacturer. Briefly, thaw matrigel on ice (final dilution is determined by the manufacturer and is lot-dependent) and resuspend in cold Advanced DMEM:F12 media. Add enough cold media+matrigel to the tissue culture plate so that the entire surface is covered (1mL/well for a 6 well and 0.5mL/well for a 24 well plate). Incubate plates at room temperature for at least 1 hour prior to plating hPSCs, but plates can be stored at 4°C for up to one week. CRITICAL: Matrigel will solidify at room temperature, so it is important to add matrigel to cold media prior to coating plates.
Thaw bottle of matrigel on ice or at 4°C. Once matrigel has thawed, add B27 supplement (final concentration 1×), Rspondin1 (final concentration 500ng/mL), Noggin (final concentration 100ng/mL) and EGF (final concentration 100ng/mL). Mix well by pipetting. Keeping matrigel cold at all times, make 50uL aliquots into 1.5 mL microcentrifuge tubes and store at −80°C for up to 6 months. CRITICAL: We have found that adding growth factors to matrigel is not critical, but does increase the efficiency of spheroid outgrowth.
Reconstitute growth factors in 1× PBS to the following final concentrations: Activin A 100ng/uL, EGF 50ng/uL, FGF4 500ng/uL, WNT3a 500ng/uL, Rspondin1 500ng/uL, Noggin 50ng/uL. After reconstitution, growth factors can be stored at 4°C for up to one week. For long-term storage (up to 6 months), make aliquots of growth factors and store at −80°C. After thawing an aliquot of growth factor from −80°C, store at 4°C for up to one week. We do not re-freeze growth factors once thawed.
Combine RPMI 1640, L-glutamine (final concentration 2 mM), Pen/Strep (final concentration pen 100 Units/mL; strep 100 ug/mL), Activin A (final concentration 100ng/mL). Endoderm differentiation media is best if made fresh each day, but can be stored at 4°C for 1–3 days. CRITICAL: No serum used in media on day 1.
Combine RPMI 1640, 0.2% dFBS (vol/vol) L-glutamine (final concentration 2 mM), Pen/Strep (final concentration pen 100 Units/mL; strep 100 ug/mL), Activin A (final concentration 100ng/mL). Endoderm differentiation media is best if made fresh each day, but can be stored at 4°C for 1–3 days.
Combine RPMI 1640, 2%dFBS (vol/vol) L-glutamine (final concentration 2 mM), Pen/Strep (final concentration pen 100 Units/mL; strep 100 ug/mL), Activin A (final concentration 100ng/mL). Endoderm differentiation media is best if made fresh each day, but can be stored at 4°C for 1–3 days.
Combine RPMI 1640, 2%dFBS (vol/vol) L-glutamine (final concentration 2 mM), Pen/Strep (final concentration pen 100 Units/mL; strep 100 ug/mL), FGF4 (final concentration 500ng/mL) and WNT3A (final concentration 500ng/mL). Make Mid/Hindgut differentiation media fresh on the first day of induction and store at 4°C for up to 4 days.
Combine Advanced DMEM:F12, B27 supplement (1× final dilution = 2mLs per 50mL media), L-glutamine (final concentration 2 mM), Pen/Strep (final concentration pen 100 Units/mL; strep 100 ug/mL), HEPES buffer (final concentration 15mM), Rspondin1 (final concentration 500ng/mL), Noggin (final concentration 100ng/mL), EGF (final concentration 100 ng/mL). Intestine growth media is best if made fresh, but can be stored at 4°C for up to one week.
PBS + 0.5% (v/v) Triton-X 100. PBST can be made ahead of time and stored at room temperature for up to 6 months.
Add 10% donkey serum to PBST (for example, 100uL donkey serum to 900uL PBST). Make fresh before each use and keep on ice.
Splitting hPSCs into 24 well dishes for differentiation. Steps 1–11, 2 hours.
Growth of hPSCs to 85–90% confluence. Step 12, 4 days.
Differentiating hPSCs into DE. Steps 13–17, 3 days.
Immunostaining. Box 1, 2 days.
Differentiating DE into hindgut. Steps 18–21, 3–4 days.
Growing spheroids into organoids. Steps 22–29, 14 days.
Splitting human intestinal tissue, Steps 30–36, 2–3 hours.
Expanding human intestinal tissue. Steps 37–41, up to 140 days. The tissue was still viable at 140 days but we have not extended culture past this time point.
This procedure outlines an extremely efficient method of directed differentiation that generates human intestinal tissue from human pluripotent stem cells in vitro. Each step of the differentiation process yields very specific and robust results. Definitive endoderm differentiation yields a population of cells that is approximately 85–90% pure by FoxA2+ Sox17+ double-positive staining5. During the 4-day mid/hindgut induction protocol, morphogenetic tissue movements in the tissue culture dish will give rise to 3-dimensional (3D) structures including epithelial tubes and spheroids. 3D structure is typically first observed at the end of the second day of induction (after 48 hours) and appears mostly as thickenings of the adherent epithelium. Within 2–3 days, spherical structures will begin to delaminate from the adherent epithelium and morphogenesis is complete by the end of 4-days at which time spheroids will contain a CDX2+ E-cadherin-positive epithelial layer surrounded by a CDX2+ E-Cadherin-negative mesenchymal layer that is strikingly similar to mouse mid/hindgut development at embryonic day (E) 8.5 (ref. 5). On the third and fourth day of the mid/hindgut induction protocol, greater than 20 spheroids are generated per well (ranging from 20–42). The robustness of spheroid generation is dependent on the efficiency of endoderm differentiation, and the density/confluency of the endoderm at the end of the 3-day endoderm induction step (Figures 1, ,22).
Using in vitro culture conditions that support the growth and expansion of adult mouse intestinal stem cells6, human hindgut spheroids proceed through a series of developmental events that mimics in vivo intestinal development5. During the first 14 days in culture, spheroids expand into organoids, giving rise to a pseudostratified epithelium that coexpresses the fetal intestinal transcription factors CDX2, SOX9 and KLF5. Over the next 14 days, the pseudostratified intestinal epithelium transitions into columnar epithelium and gives rise to intestinal tissue that contains an apical-basal orientation including a luminal surface. During this time, villus-like structures emerge into the lumen of the organoids. After approximately one month in culture, human intestinal organoids resemble a fetal mouse intestine, and ~90% of all epithelial cells express the intestinal marker CDX25. During prolonged culture (>2 months), villus-like structures emerge and express LGR5 and ASCL2, markers of adult mouse intestinal stem cells6,16.
After the first 28 days in culture, organoids can be passaged by manually cutting them in half with a scalpel. After the organoids are cut in half and re-embedded into matrigel, they will continue to grow and tissue will undergo an approximate doubling every 10–14 days. We have successfully passaged organoids for up to 8 passages spanning approximately 140 days, stopping only for an experimental end-point. It is likely that human organoids will continue to grow and expand much longer than this. It is possible that there is a less time-consuming method by which to split and expand human organoids, however, we have only performed manual passaging to date.
This work was supported by the Juvenile Diabetes Research Foundation JDRF-2-2003-530 (J.M.W.) and NIH, R01GM072915, R01DK080823A1 and S1 (J.M.W.); This work was also supported in part by PHS Grant P30 DK078392 (J.R.S) and K01 DK091415 (J.R.S.). J.C.H. is supported by an Endocrine Fellows Foundation Developmental Research Grant in Diabetes, Obesity and Fat Cell Biology. We also acknowledge core support from the Pluripotent Stem Cell Facility (supported by U54 RR025216).
Author Contributions StatementJ.M.W., J.R.S., J.C.H., and K.W.M. conceived the study and experimental design. J.M.W. and J.R.S. analyzed data and co-wrote the manuscript. J.C.H. and K.W.M. performed experiments.
Competing Financial Interests Statement
J.M.W. and J.R.S. are inventors on a patent involving the system described herein.