Deriving enteric nervous system (ENS) lineages from human pluripotent stem cells (hPSC) provides a scalable source of cells to study ENS development and disease, and to use in regenerative medicine. Here, a detailed in vitro protocol to derive enteric neurons from hPSCs using chemically defined culture conditions is presented.
The human enteric nervous system, ENS, is a large network of glial and neuronal cell types with remarkable neurotransmitter diversity. The ENS controls bowel motility, enzyme secretion, and nutrient absorption and interacts with the immune system and the gut microbiome. Consequently, developmental and acquired defects of the ENS are responsible for many human diseases and may contribute to symptoms of Parkinson’s disease. Limitations in animal model systems and access to primary tissue pose significant experimental challenges in studies of the human ENS. Here, a detailed protocol is presented for effective in vitro derivation of the ENS lineages from human pluripotent stem cells, hPSC, using defined culture conditions. Our protocol begins with directed differentiation of hPSCs to enteric neural crest cells within 15 days and yields diverse subtypes of functional enteric neurons within 30 days. This platform provides a scalable resource for developmental studies, disease modeling, drug discovery, and regenerative applications.
The enteric nervous system (ENS) is the largest component of the peripheral nervous system. The ENS contains more than 400 million neurons that are located within the GI tract and control nearly all functions of the gut1. Molecular understanding of the ENS development and function and its defects in enteric neuropathies requires access to a reliable and authentic source of enteric neurons. Access to human primary tissue is limited, and animal models fail to recapitulate key disease phenotypes in many enteric neuropathies. Human pluripotent stem cell (hPSC) technology has proven exceedingly beneficial in providing an unlimited source of desired cell types, especially those that are difficult to isolate from primary sources2,3,4,5,6,7. Here, we provide details of a stepwise and robust in vitro method to obtain ENS cultures from hPSCs. These scalable hPSC-derived cultures open avenues for developmental studies, disease modeling, and high-throughput drug screening and can provide transplantable cells for regenerative medicine.
Enteric neuron lineages are derived from neural crest (NC) cells following the ENS developmental path during embryogenesis. In embryos, NC cells emerge along the margins of the folding neural plate. They proliferate, migrate and give rise to many different cell types including sensory neurons, Schwann cells, melanocytes, craniofacial skeleton and enteric neurons and glia8,9,10. The cell fate decision depends on the distinct region along the anterior-posterior axis that the NC cells emerge from, i.e., cranial NC, vagal NC, trunk NC and sacral NC. The ENS develops from vagal and sacral NC cells with the former dominating the population of the enteric neurons owing to the extensive migration along the length of the bowel and colonizing the gut11.
Derivation of NC cells from PSCs commonly involves a combination of dual SMAD inhibition and WNT pathway activation12,13. Until recently, all NC induction protocols involved serum and other animal product additives in the culture conditions. Chemically undefined media not only lower the reproducibility of NC induction but also challenge mechanistic developmental studies. To overcome these challenges, Barber et al14 developed an NC induction protocol using chemically defined culture conditions and was hence advantageous over alternative methods that rely on serum-replacement factors (e.g., KSR)13,14. This was obtained by basal media replacements and optimizing a protocol originally presented by Fattahi et al to derive enteric NC cells from hPSCs. This improved system is the basis of the hPSC differentiation protocol that is presented here. It begins with induction of enteric neural crest (ENC) cells in a 15-day period by precise modulation of BMP, FGF, WNT and TGFβ signaling in addition to retinoic acid (RA). We then derive the ENS lineages by treating cells with glial cell line-derived neurotrophic factor (GDNF). All media compositions are chemically defined and yield robust enteric neuronal cultures within 30-40 days.
In addition to the monolayer cell culture system described here, alternative NC induction approaches have been developed which use free-floating embryoid-bodies15,16. Migratory cells in these cultures have been shown to express NC markers with a subset representing the vagal NC. Co-cultures with primary gut tissue have been used to enrich enteric NC precursors in these cultures. Media compositions in these studies contain a combination of different factors such as nerve growth factor, NT3 and brain-derived neurotrophic factor. At this stage, it is not fully clear how these factors might affect the enteric neuron precursor commitment identities. For an efficient comparison of the enteric NC induction in the monolayer and the embryoid-body-based cultures more data is required. Given the different culture layouts in these strategies, the use of each method should be considered and optimized according to specific application in mind.
The protocol presented here is reproducible and has been successfully tested by us and others using different hPSC (induced and embryonic) lines8,14,16.
1. Media preparation
NOTE: Concentrations mentioned throughout the protocol are final concentrations of the media components. Prepare all media under sterile conditions in a laminar flow hood, and store at 4 °C in the dark. Use within 2 weeks.
2. Coating of culture plates
3. Maintenance of hPSC culture
NOTE: All cell incubation steps are at 5% CO2 and at 37 °C in a humidified incubator.
4. Neural crest induction
NOTE: This step is schematically represented in Figure 1A. Begin NC differentiation when cells are about 80% confluent. This will be achieved within 1-2 days when cells are passaged at a 5:6 ratio (see above). Start NC differentiation on day 0 with pluripotent and fully undifferentiated hPSC cells. For a high efficiency, the colony borders should have minimal differentiating cells. Passage and plate hPSC for differentiation when colonies are large, or the center of the colony starts to thicken/darken when monitored using an inverted microscope. Paying attention to the confluency and morphology tends to be more reliable than the number of plated cells as this might vary depending on the cell line and can vary from 50,000-200,000 per cm2.
5. ENC spheroid formation
NOTE: This step is schematically represented in Figure 1B.
6. EN induction
NOTE: This step is schematically represented in Figure 1B.
This protocol provides a method to derive enteric neural crest and enteric neurons from hPSCs using chemically defined culture conditions (Figure 1A-B). Generating high-quality neurons depends on an efficient enteric neural crest induction step. This can be visually assessed by checking the morphology of the free-floating spheres that should look round with smooth surfaces with a size of approximately 0.1 – 0.4 mm as seen in Figure 1C. These spheroids should also express the neural crest marker SRY-Box transcription factor 10 (SOX10)17,18, as can be seen in Figure 1D. During enteric neuron induction phase, neurites can become visible as early as day 25-30 (Figure 1E). However, this can be delayed in differentiations with lower efficiency containing undesired mesenchymal cell contaminations. The human enteric nervous system is a collection of diverse neurons expressing different neurotransmitters. This heterogeneity is recapitulated in the in vitro protocol provided and the expression of key markers can be assessed by cell staining methods such as immunofluorescence imaging and flow cytometry according to previously described protocols13,14,19,20,21 (Figure 1F,G).
Figure 1: Deriving human ENC cells and enteric neuron subtypes from hPSCs. (A) Enteric neural crest (ENC) induction from hPSC. Protocol (days 0-12) showing media compositions for deriving ENC cells from hPSCs. (B) ENC spheroid formation and enteric neuron induction from ENC. Protocol (days 12-15) showing medium composition for ENC spheroid formation. Afterwards, enteric neurons are generated and maturated by treating with a medium containing GDNF and AA. (C) Representative phase contrast images of differentiating hPSCs at different time points of the protocol. Scale bar: 1000 µM. (D) SOX10::GFP expressing H9 iPSC-derived ENC spheroids on day 14. Scale bar: 1000 µm. (E) Phase contract image of day 33 neurons. Scale bar: 400 µm. (F-G) Immunofluorescence imaging (F) and flow-cytometry (G) of neurotransmitter markers in day 40 neurons. Scale bar: 50 µm. Abbreviations: BMP4 = recombinant human BMP4 protein; RA = retinoic acid; FGF2 = recombinant human FGF basic; GDNF = recombinant human glial cell line-derived neurotrophic factor; AA = ascorbic acid. Please click here to view a larger version of this figure.
The differentiation protocol described here provides a robust in vitro method to obtain enteric neurons from hPSCs within 30-40 days (Figure 1E) and enteric glia expressing glial fibrillary acidic protein, GFAP, and SOX10 in older cultures (> day 55)13,14,19,22. These neurons and glia are induced by stepwise differentiation of hPSCs into vagal and enteric neural crest cells followed by enteric neuron and glia progenitors13,14. The neural crest cells express neural crest markers, transcription factor AP-2 alpha and beta, TFAP2A and TFAP2B, respectively21. We previously showed that CD49 can be used as a marker to indicate the efficiency of hPSC-derived neural crest cells14. The efficiency and success of this method depends on a combination of factors. First, it is critical to perform all steps under sterile conditions. Testing hPSC cultures for mycoplasma contamination is important as it can adversely affect stem cell growth and is not detectable visually. Second, use of evenly and properly coated plates reduces the chance of patchy hPSC culture growth and detachment of NC cultures and older neurons from the surface of the wells. Third, for high efficiency in differentiation it is essential to start with pluripotent and fully undifferentiated hPSC cells. hPSC colonies should have minimal differentiating cells at the colony borders. hPSCs should be passaged or plated for differentiation when colonies are large, or the center of the colony starts to thicken/darken in color as observed through a phase contrast microscope. Undesired differentiated cells (usually mesenchymal cells), not only drop the overall differentiation efficiency, but they can result in detachment of neurons from the surface of the well as the cultures get older. It is therefore recommended to monitor the ENC cultures in the first 15 days carefully. Under efficient differentiation conditions dark disks appear on the surface of the culture ~day 6 which indicates efficient spheroid formation on days 12-15. In addition, in the spheroid phase (day 12-15) free-floating spheres should have smooth surfaces (Figure 1D). Maintaining the 3D spheroid phase on days 12-15 is generally sufficient to improve the purity of ENCs for the downstream differentiation of cultures into ENS lineages. However, for studies that require early and pure ENC progenitor cells populations, FACS using CD49D is recommended. CD49D is a specific surface marker for SOX10:hPSC-derived NC lineages and has been used successfully across cell lines13,14. For more information on isolating CD49D+ cells using FACS and recommendations on gating strategy readers are referred to the paper by Barber et al, Figure 2D, Box1 and Supplementary Figure 614. Fourth, to maximize efficiency, differentiation should start when hPSC cultures are monolayer and ~80% confluent. This means having a fully confluent culture on day 2. Lower confluency negatively affects the efficiency and very dense cultures tend to have a higher rate of cell death and can potentially lead to emergence of unwanted differentiated cells.
Re-plating cells on day 15 marks an important step in this protocol. This is the last cell transfer step for a monolayer neuronal culture and hence the plate layout should be chosen according to the assays planned. The number of cells transferred to each well can considerably affect the neuronal diversity and older neurons remaining attached to the surface of the wells. Generally, a cell density in the range of 250,000-350,000 cells per cm2 (about 75,000-100,000 cells per well of a 96-well plate or 500,000-700,000 per well of a 24-well plate) is desirable. Starting at day 30-40, one might consider adding fibronectin and laminin to the feeding medium (not more than once a week) to prospectively prevent detachment of neurons especially for cells planned for longer neuron cultures. At this point it is recommended to reduce the feeding frequency to once in every 5-7 days and leaving some medium in each well before adding extra fresh medium. This reduces the chance of applying unnecessary physical stress on the neuronal projections. Considering evaporation of medium (especially for wells in the border of the plate) one can remove 75 µL and add 100 µL of fresh medium to each well. Based on experience, cultures grown on polymer coverslip bottom imaging plates tend to live and stay attached to the surface of the well for a longer time. Therefore, considering the polymer coverslip bottom imaging plates for experiments requiring very old neuronal cultures such as co-cultures with other cell types is recommended. Although air-dried PO/FN/laminin coated plates might still yield in efficient differentiation, this has not been tested systematically for the purpose of this protocol. Hence, the use of freshly coated plates is recommended.
The protocol described here is based on a previously published protocol by Barber et al14. Compared to older ENC induction methods that use media containing serum-replacement factors (e.g., KSR)13,14, the current protocol media conditions are chemically defined. This is advantageous towards more consistent differentiations and allows more precise developmental studies by manipulating the culture conditions. This robust protocol reproducibly generates cultures of enteric neurons from independent induced and embryonic stem cell lines that have been tested, such as H9, UCSF4 and WTC11. A powerful aspect of the presented protocol is its capacity to generate a variety of ENS neuronal subtypes resembling those found in the human ENS. For example, expression of specific markers for different ENS neurochemical identities can be checked by flow cytometry and fluorescence imaging (Figure 1F,G). The typical range of hPSC-derived enteric neuronal populations positive for nitric oxide synthase 1 (NOS1) is 5%-20%; choline O-acetyltransferase (CHAT) is 40%-60%; gamma-aminobutyric acid (GABA) is 15%-30% and serotonin (5-HT) is 15%-30%. However, the extent to which the in vitro structural and functional connectivity between these subtypes accurately models the in vivo cellular networks requires further studies. In addition, single cell analyses to expand the transcriptional profiling of individual subtypes would not only help to better compare the in vivo and in vitro cellular identities but would also benefit developmental and molecular studies as well as future protocol optimizations towards enriching specific subtypes for drug discovery and regenerative medicine.
Despite the cultures containing different neuronal, and in older cultures glial cells, this protocol does not generate the other cell types naturally found in the in vivo niche of the ENS, such as muscle cells, epithelial cells, and immune cells. Studying the ENS interaction with these cell types in the context of health and disease requires investment in optimizing co-culture and intestinal organoids systems. Overall, a powerful method to generate different subtypes of the ENS in from hPSCs in vitro is described here. The scalable cultures are suitable for developmental studies of the human ENS, toxicology testing and disease modeling experiments aimed at developing drug therapies and providing transplantable cells for treating enteric neuropathies.
The authors have nothing to disclose.
The work was supported by grants from UCSF Program for Breakthrough Biomedical Research and Sandler Foundation, March of Dimes grant no. 1-FY18-394 and 1DP2NS116769-01, the NIH Director's New Innovator Award (DP2NS116769) to F.F. and the National Institute of Diabetes and Digestive and Kidney Diseases (R01DK121169) to F.F., H.M. is supported by Larry L. Hillblom Foundation postdoctoral fellowship, NIH T32-DK007418 fellowship and UCSF Program for Breakthrough Biomedical Research independent postdoctoral fellowship.
Ascorbic acid | Sigma-Aldrich | A5960 | |
B27 supplement (serum free, minus vitamin A) | Gibco | 12587-010 | |
Basement membrane matrix, Geltrex | Gibco | A14133-2 | |
BMP4 | R&D systems | 314-BP | |
Cell culture centrifuge | Eppendorf, model no. 5810R | 02262501 | |
Cell detachment solution, Accutase | Stemcell Technologies | 07920 | |
CHIR99021 | Tocris | 4423 | |
Conical tubes | USA scientific | 1475-0511, 1500-1211 | |
Differentiation base medium, Essential 6 | Life Technologies | A1516401 | |
DMEM/F-12 no glutamine | Life Technologies | 21331020 | |
EDTA | Corning | MT-46034CI | |
Feeder-free hPSC maintenance medium, Essential 8 Flex Medium Kit | Life Technologies | A2858501 | |
FGF2 | R&D systems | 233-FB/CF | |
Fibronectin | Corning | 356008 | |
GDNF | Peprotech | 450-10 | |
Hemocytometer | Hausser Scientific | 1475 | |
Human pluripotent stem cells, H9 ESC | WiCell | RRID: CVCL_1240 | |
Incubator with controlled humidity, temperature and CO2 | Thermo Fisher Scientific | Herralcell 150i | |
Inverted microscope | Thermo Fisher Scientific | EVOS FL | |
Laminar flow hood | Thermo Fisher Scientific | 1300 series class II, type A2 | |
Laminin | Cultrex | 3400-010 | |
L-glutamine supplement, Glutagro | Corning | 25-015-CI | |
MEM NEAAs | Corning | 25-025-CI | |
Multiwell plates, Falcon | BD | 353934, 353075 | |
N-2 Supplement | CTS | A1370701 | |
Neurobasal Medium | Life Technologies | 21103049 | |
PBS (Ca and Mg free) | Life Technologies | 10010023 | |
Pipette filler | Eppendorf | Z768715-1EA | |
Pipette tips | USA scientific | 1111-2830 | |
Pipettes | Fisherbrand | 13-678-11E, 13-678-11F | |
PO | Sigma-Aldrich | P3655 | |
polymer coverslip bottom imaging plates, ibidi | ibidi | 81156 | |
RA | Sigma-Aldrich | R2625 | |
SB431542 | R&D systems | 1614 | |
Trypan blue stain, 0.4% | Thermo Fisher Scientific | 15250-061 | |
Ultra-low attachment plates | Fisher Scientific | 07-200-601 | |
Y-27632 dihydrochloride | R&D systems | 1254 |