Generating Enteric Nervous System Progenitors from Human Pluripotent Stem Cells

The intrinsic innervation of the gastrointestinal (GI) tract is comprised of enteric neurons and glia, which are buried within the wall of the bowel and organized into two concentric plexuses that run along the length of the gut forming the enteric nervous system (ENS). The ENS regulates vital GI functions including gut motility, blood flow, fluid secretion, and absorption and thus maintains gut homeostasis. During vertebrate development it originates predominantly from the vagal neural crest (NC), a multipotent cell population that emerges from the caudal hindbrain region, migrates to and within the gut to ultimately generate neurons and glia in response to gut‐derived signals. Loss of GI innervation due to congenital or acquired defects in ENS development causes enteric neuropathies which lack curative treatment. Human pluripotent stem cells (hPSCs) offer a promising in vitro source of enteric neurons for modeling human ENS development and pathology and potential use in cell therapy applications. Here we describe in detail a differentiation strategy for the derivation of enteric neural progenitors and neurons from hPSCs through a vagal NC intermediate. Using a combination of instructive signals and retinoic acid in a dose/time dependent manner, vagal NC cells commit into the ENS lineage and develop into enteric neurons and glia upon culture in neurotrophic media. © 2021 The Authors. Current Protocols published by Wiley Periodicals LLC.


INTRODUCTION
The enteric nervous system (ENS) is a complex and elegantly shaped network of neurons and supporting glial cells which provides the intrinsic gut innervation to control a plethora of bowel functions, such as intestinal motility, fluid secretion, movement across the epithelium, and vascular tone (Furness, 2012). The ENS is organized into ganglionated structures forming two major plexuses, the myenteric and submucosal plexus, located within the wall of gastrointestinal (GI) tract and extending longitudinally (Furness, 2006). During early embryonic development, the majority of enteric ganglia develops from the vagal neural crest (NC; Anderson, Stewart, & Young, 2006;Le Douarin & Teillet, 1973). Vagal NC is a transient multipotent cell population emerging from the dorsal tip of the neural tube at the level of somites 1 to 7 and is marked by the expression of HOX gene paralogous groups (PG) 3-5 (Burke, Nelson, Morgan, & Tabin, 1995;Chan et al., 2005;Fu, Lui, Sham, Cheung, & Tam, 2003;Kam & Lui, 2015). Apart from generating most of the ENS, vagal NC cells also contribute to cardiovascular system structures, lung, thymus, and pancreas (Hutchins et al., 2018;Le Douarin, Creuzet, Couly, & Dupin, 2004). Following delamination, vagal NC cells emigrate to the gut, a process partially orchestrated by somite-derived retinoic acid (RA; Niederreither et al., 2003;Simkin, Zhang, Rollo, & Newgreen, 2013;Uribe, Hong, & Bronner, 2018), and commit to a developmental program, which involves highly coordinated events such as progressive ENS progenitor specification and proliferation, rostro-caudal migration, and differentiation towards neurons and glia (reviewed in Kang, Fung, & Vanden Berghe, 2021;Lake & Heuckeroth, 2013). Dysregulation of these processes could result in absence or dysfunction of enteric ganglia along various lengths of the intestine, a common cause of enteric neuropathies such as Hirschsprung's disease (HSCR; Heanue & Pachnis, 2007;Obermayr, Hotta, Enomoto, & Young, 2013).
Seminal findings on the persistence of enteric neural progenitor cells in the fetal and post-natal gut offer an autologous cell therapy as a promising strategy to treat enteric neuropathies (Burns & Thapar, 2014). However, limiting factors such as the need for a sufficient pool of ENS progenitors that could be harvested have set a roadblock to this therapeutic approach. Human pluripotent stem cells (hPSCs) offer an alternative, accessible, and infinite source of such progenitor cells (reviewed in Stamp, 2017) and a number of hPSC differentiation strategies for the production of ENS progenitors and their derivatives have been reported over the past few years (Barber, Studer, & Fattahi, 2019;Fattahi et al., 2016;Lai et al., 2017;Lau et al., 2019;Li et al., 2018;Schlieve et al., 2017;Workman et al., 2017). We have recently described the generation of NC populations corresponding to various levels along the anteroposterior (A-P) axis from hPSCs, including vagal NC (Frith et al., 2018). Based on this, we developed a step-by-step differentiation strategy for the generation of ENS progenitors via an hPSC-derived vagal NC intermediate (Frith et al., 2020). Our approach relies on the induction of a HOX-negative NC population with a default cranial identity (Hackland et al., 2017), expressing SOX10, PAX3, PAX7, which is subsequently steered toward a caudal hindbrain/cervical axial identity, reflected by the upregulation of HOX PG members (3-5), following exposure to RA (Frith et al., 2018;Frith et al., 2020). Moreover, we found that RA treatment drives the acquisition of an ENS progenitor identity marked by the induction of ASCL1, PHOX2B. These vagal NC/early ENS progenitors can colonize and differentiate efficiently within the ENS of adult mice following transplantation and thus serve as an attractive cell population for use in cell therapy approaches aiming to treat enteric neuropathies such as HSCR (Frith et al., 2020). Here we provide a detailed description of the protocol for their in vitro production from hPSCs as well as their subsequent differentiation into enteric neurons and glia (Fig. 1).

GENERATION OF VAGAL NEURAL CREST/EARLY ENS PROGENITORS FROM hPSCs
The differentiation protocol ( Fig. 2A) begins with plating hPSCs into a culture regime that allows precise control of BMP signaling levels, achieved by simultaneous treatment with BMP4 recombinant protein along with a BMP type-I receptor inhibitor (DMH-1, an approach dubbed as Top-Down inhibition; Hackland et al., 2017), WNT signaling activation through GSK-3 inhibition mediated by the WNT agonist CHIR99021, and TFG-β pathway inhibition by employing the ALK receptor inhibitor SB431542. Following 4 days of treatment, the resulting population displays an anterior/cranial NC character reflected by the upregulation of SOX10, PAX3, PAX7, p75, and the lack of HOX gene expression. Medium supplementation with RA during days 4 to 6 induces a vagal axial identity marked by the upregulation of the corresponding HOX genes (HOXB4, HOXB5), a step which coincides with the acquisition of an ENS precursor character as evidenced by ASCL1, PHOX2B expression.
Subculture newly thawed lines for several passages (at least four) before differentiation.  All-trans retinoic acid (only add between Day 4-Day 6) b 1 mM 1 μΜ 10 μl 2 5μl 5 0μl a We recommend that neural crest induction medium is made fresh on the day of feeding. At Day 0, when hPSCs are plated into neural crest induction conditions Y-27632 2HCl, a Rho-associated coil kinase (ROCK) inhibitor, must be added at a final concentration of 10 μM to help cell attachment and increase cell survival. b All-trans retinoic acid is light sensitive, so work in the dark and keep medium out of direct light.
2. Passage hPSC cultures every 4-5 days until they reach 80%-85% confluency using ReLeSR TM , as described in the protocol provided from Stemcell Technologies.
Passaging ratios may vary between 1:6 and 1:12. For increased NC yield, passage cells at a high ratio twice per week.
ReLeSR TM -based passaging yields higher NC efficiency compared to EDTA-based dissociation solution.
4. Before starting the differentiation, coat plates with Geltrex: a. Thaw a Geltrex aliquot on ice to prevent unwanted gelling when handling small volumes. b. Dilute Geltrex 1:100 in pre-chilled DMEM/F12 (4°C), i.e., 200 μl Geltrex in 20 ml DMEM/F12. c. Transfer sufficient Geltrex solution to each well of a multi-well plate, i.e., 100-200 μl Geltrex solution per cm 2 is enough to coat the growth surface. d. Incubate Geltrex-coated plates at 37°C for a minimum of 1 hr. Alternatively, coated plates can be stored at 4°C for up to 2 weeks when sealed with Parafilm to prevent them from drying out.
6. Prepare neural crest induction medium by supplementing neural crest basal medium with small molecules and mix thoroughly (see Table 2).

Differentiation of hPSCs into vagal NC cells/early ENS progenitors (6 days in total)
7. Day 0: Plating hPSCs for neural crest differentiation: a. Grow hPSCs until they reach ∼80%-85% confluency. b. Aspirate mTeSR TM 1 medium from the flask and add 1× PBS (200 μl/cm 2 ) to wash cell colonies. Remove PBS. c. Add Accutase solution (100 μl per cm 2 ) and incubate at 37°C for 5-7 min. Observe cells under the microscope to ensure cells have detached and a single-cell suspension is obtained.

of 20
Current Protocols

Figure 3
Comparison of hPSCs plated at an optimal (upper row) and non-optimal (bottom row) seeding density during NC induction. A sub-optimal seeding density yields lower efficiency and impure NC cultures that are morphologically distinct and appear as flat and dense colonies (dashed circle) neighboring NC cells that display a stellate-like morphology (arrows). Scale bar = 50 μm. hPSCs, human pluripotent stem cells; NC, neural crest.
d. Neutralize Accutase solution by adding pre-warmed DMEM/F12 medium and triturate cells using a serological pipet in order to dissociate them into a singlecell suspension. e. Transfer cell suspension to a conical centrifuge tube.
Cells are lifted up in suspension. Do not aspirate Accutase solution as this will result in losing the cell population.
f. Agitate conical tube to ensure adequate mixing and transfer 10 μl cell suspension to a hemocytometer. g. Count cells using the hemocytometer. h. Meanwhile centrifuge cell suspension at 300 × g for 4 min at room temperature. i. After centrifugation, aspirate supernatant carefully and dislodge cell pellet in neural crest induction medium supplemented with 10 μM Y-27632 2HCl (see Table 2).

We recommend resuspending cells at a ratio of 1 million cells per ml.
j. Plate cells at a density of 30,000 to 40,000 cells per cm 2 on a Geltrex pre-coated plate in neural crest induction medium supplemented with 10 μM Y-27632 2HCl.
For a well of a 6-well plate with a surface area of ∼10 cm 2 , transfer 300 μl of cell suspension.
Cell density is very critical for neural crest differentiation efficiency and it is cell line dependent. We highly recommend performing cell density titration for every cell line used to ensure optimal yield. Representative images of non-NC like cells compromising the protocol's efficiency due to inefficient seeding density are displayed in Figure 3.
k. Following cell plating, top up each well with neural crest induction medium supplemented with 10 μM Y-27632 2HCl to obtain a final volume of 200 μl per cm 2 . l. Place cells in a 37°C incubator and move plate back and forth, side to side to ensure cells are evenly distributed throughout the growth surface. m. Incubate cells at 37°C in a humidified atmosphere of 5% CO 2 in air. a. Feed cells by aspirating old medium and adding 200 μl freshly made neural crest induction medium without Y-27632 2HCl per cm 2 . Always add medium down the side wall of the well and not directly on top of the cell colonies.
By Day 2, cells form spider's-web like colonies and exhibit a spread out, spiky characteristic morphology due to treatment with Y-27632 2HCl (Fig. 2B) a. Refresh medium with 200 μl per cm 2 of freshly made neural crest induction medium supplemented with 1 μM all-trans retinoic acid.
We found that replacing medium at this stage to ensure sufficient levels of retinoic acid is important to maintain the expression of retinoic acid induced ENS progenitor genes ASCL1, PHOX2B.
Do not be alarmed by the increased cell death following treatment with retinoic acid; it is part of the process.
Cells begin to look darker and spiky after treatment with retinoic acid (Fig. 2B).
b. Incubate cells at 37°C for 24 hr.
11. Day 6: a. End of vagal neural crest differentiation: Cells cover 95%-100% of the growth surface forming a monolayer, appear dark and spiky (Fig. 2B). b. Cells are ready for replating in non-adherent conditions prior to the generation of enteric neurons and glia (see Basic Protocol 2).
Optional: Assess vagal neural crest induction by flow cytometry, immunocytochemistry, or quantitative polymerase chain reaction (qPCR; for details about experimental techniques refer to Frith et al., 2018;Frith et al., 2020).

DIFFERENTIATION OF hPSC-DERIVED VAGAL NC/EARLY ENS PROGENITORS TO ENTERIC NEURONS AND GLIA
The protocol (Fig. 4A) continues with re-plating vagal NC cells generated from 2D monolayer differentiation into non-adherent conditions that promote formation of 3D cell sphere aggregates. These are then subsequently re-plated in adherent conditions and in the presence of neurotrophic medium to promote further differentiation toward ENS components. During that stage, SOX10+ progenitor cells migrate out of the re-plated spheres to give rise to enteric neurons and glia. The emergence of neurons, marked by downregulation of SOX10 and induction of neural markers, precedes glial occurrence, which is accompanied by sustained expression of SOX10 and upregulation of glial markers.

No treatment or coating is required for the ultra-low attachment plates.
Gogolou et al.

of 20
Current Protocols We have used 1% agarose coated plates instead of the Corning ultra-low attachment plates for sphere formation as well but the effect of the different conditions in sphere architecture and expression was not thoroughly examined.

Sphere formation (3-4 days in total)
3. Day 6 (re-plating vagal neural crest into non-adherent conditions): a. Aspirate medium from Day 6 vagal neural crest cells/early ENS progenitors carefully to avoid disturbing the monolayer on the bottom of the well. b. Perform a quick wash with 1× PBS. c. Lift cells into single-cell suspension by adding Accutase solution (100 μl per cm 2 ) and incubate at 37°C for 5-7 min. d. Neutralize Accutase solution with pre-warmed DMEM/F12 and using a serological pipet, triturate cell suspension a few times to ensure complete cell dissociation.

Do not aspirate Accutase solution at this step, as it contains cells that have been detached.
No need for counting cells; cells are plated down at a 1:1 ratio, e.g., the equivalent of a well of a 6-well plate is transferred into a well of an ultra-low attachment plate.
e. Transfer cell suspension into a conical tube. f. Place conical tube in a centrifuge and spin at 300 × g for 4 min at room temperature. g. Aspirate and discard supernatant very carefully without disturbing cell pellet. h. Resuspend pellet in complete sphere culture medium (see Table 6).
For example: When a full 6-well plate of vagal neural crest cells/early ENS progenitors is harvested for sphere formation, the resulting cell pellet is resuspended in 12 ml complete sphere medium and 2 ml of cell suspension is transferred per well for subsequent plating.
i. Transfer cell suspension to each well of the 6-well ultra-low attachment plate.
That is, 2 ml of cell suspension transferred into each well of the non-adherent 6-well plate.
Be gentle while transferring cells to the plate; avoid scraping the bottom of the well with the pipet tip. j. Place plates into a plate holder and centrifuge at 100 × g for 2 min at room temperature.

This step collates single cells in the middle of the well to increase cell-cell contact and assist sphere formation.
Following centrifugation, cell suspension is visible at the center of the well.

k. Incubate cells at 37°C for 24 hr.
Do not disturb plates for the first 24 hr.

Day 7 (24 hr after plating):
a. Inspect cultures visually using an inverted microscope.
Floating cell aggregates appear in culture. They are quite fragile without well-defined spherical shape yet.
Do not be alarmed by the increased cell death or number of single cells floating; many of them will eventually form spheres in the next 24 hr.
b. Place plates in the incubator for another 24 hr.

Day 8 (48 hr after plating):
a. Spheres should be visible at this stage and have a better and/or round shape compared to Day 7 (Fig. 4A). b. Perform half medium change to supplement cells with fresh medium (see Table 6).
i. Swirl plate to collect spheres from the circumference to the center of the well. ii. Hold plate at an angle towards you and let spheres settle to form sediment at the bottom corner of the well. iii. Remove and discard 1 ml medium very carefully by placing the tip of a P1000 pipet at an angle facing the side of each well.
Do not aspirate all medium, as this increases the risk of losing cells, too.
Do not use a vacuum aspirator to perform medium change at this point; the suction force is likely to cause sphere loss.
c. Incubate plates at 37°C for 24 hr.
6. Day 9 and Day 10: a. Spheres are re-plated into neurotrophic medium for further differentiation towards ENS derivatives. b. If spheres have fused, forming large irregularly shaped aggregates, we prefer to break them up into smaller spheres and culture them for an additional day prior to culture in neurotrophic medium (see Troubleshooting, Table 10).
Transfer spheres to a 50-ml conical tube and triturate sphere suspension two to three times by pipetting towards the side of the tube. We advise using a serological pipet with wide tip opening, i.e., 10-ml stripette which is less harsh for the spheres.
Gentle handling of spheres is highly recommended; do not break the spheres too much, as this will result in excessive cell death.
Prior to sphere plating 7. Prepare poly-L-ornithine/Geltrex coated plates: 14 of 20 a. Coat plastic growth surface with 15 μg/ml poly-L-ornithine solution prepared in cell culture grade water (50 μg/ml poly-L-ornithine solution is recommended for glass surfaces); 200 μl per cm 2 is enough to cover the growth area. Incubate plates for a minimum of 1 hr or overnight at 37°C. b. Wash plates thoroughly twice using 1× PBS. c. Aspirate 1× PBS and coat each well of the plate with diluted Geltrex solution (1:100) as previously shown (Basic Protocol 1). d. Let plates stand for 1 hr at 37°C. 8. Prepare ENS basal medium (see Table 7). 9. Prepare ENS complete medium (see Table 8).

Plating spheres into ENS induction medium (12 days and beyond)
10. Day 9 and Day 10: a. Swirl plates and group spheres together at the center of the well. b. Collect spheres in a 50-ml conical tube using a 10-ml serological pipet.
Sphere collection requires gentle handling to avoid breaking up the spheres, resulting in single-cell suspension and concomitant reduced survival.
c. Let spheres settle to the bottom of the tube by gravity sedimentation for 5 min at room temperature before proceeding to step d. d. When the supernatant is clear and no cell aggregates are observed within the solution, carefully remove and discard medium without disturbing sphere pellet.
Aspirating medium with a serological pipet or a P1000 pipet is highly recommended as the suction force of an aspiration pump will disturb the pellet and result in sphere loss.
e. Resuspend sphere pellet using a 10-ml serological pipet in ENS complete culture medium (see Table 8) and distribute sphere suspension to each well. c. After day 30-35, perform half medium change biweekly (every 3-4 days) and increase medium volume from 200 μl/cm 2 to 400 μl/cm 2 .

Understanding results
The neurons tend to form at regions of high density across the well, converge, and create big clumps of neuronal cell bodies that are interconnected with axonal processes that span the whole well (Fig. 4). Neurons first appear and migrate out of the spheres around day 12 of the protocol (2 days after culture in neurotrophic medium; Fig. 4B). Between day 17-22, neural fibers are visible and cells express pan-neuronal markers such as neurofilament heavy chain (NF-H), HuCD, TUJ1, PERIPHERIN, and TRKC ( Fig. 4C and D), markers widely used in the literature for enteric neuron characterization (Table 9). Glial cells expressing S100β and SOX10 emerge around day 20-22 and grow in close vicinity to the neuronal cell bodies (Fig. 4D). At this stage, the cell cultures are comprised of a mixture of neural and glial cells that differentiate and mature further as the protocol progresses (Fig. 4E). For a more detailed expression analysis of the generated neurons, refer to the original study of Frith et al. (2020). We regularly observe emergence of contaminant cells that lack the characteristic neuronal morphology and are not immunoreactive to ENS lineage markers (ASCL1, PHOX2B, SOX10). These do not appear to influence negatively the induction of neurons and/or glia.

Complete sphere culture medium
See Table 6.

Enteric nervous system (ENS) complete culture medium
See Table 8.

Neural crest basal medium
See Table 1.

Neural crest induction medium
See Table 2.

Sphere culture basal medium
See Table 5.

Background Information
The in vitro generation of enteric neural progenitors from hPSCs holds a great promise for the development of cell-therapy-based approaches and the study of the specification of ENS lineages and hence has attracted a considerable amount of research interest. To date, differentiation strategies have either implemented embryoid body formation and culture of hPSC in the presence of EGF, FGF signals (Li et al., 2018;Workman et al., 2017) or employed a monolayer differentiation ap-proach that relies on transforming growth factor β (TGF-β) signaling suppression, bone morphogenetic protein (BMP) signaling regulation and WNT pathway stimulation to generate an NC-like population (Barber et al., 2019;Lau et al., 2019). Patterning of the in vitro derived NC to a vagal axial identity is routinely achieved by retinoic acid (RA) addition (Barber et al., 2019;Lau et al., 2019;Workman et al., 2017) while further commitment along the ENS trajectory has been mediated by co-culture with intestinal (Workman et al., Gogolou et al.
The protocol described here outlines an efficient method of directed differentiation that generates ENS progenitors from hPSCs via a vagal NC induction step following RA treatment. The derivation of NC relies on the combined WNT signaling stimulation and TFG-β pathway inhibition together with precise levels of BMP signaling. The latter is achieved by the introduction of saturating levels of exogenous BMP which is titrated by the simultaneous addition of a BMP receptor inhibitor (DMH-1) to counteract the effects of variable endogenous BMP production and the concomitant variations in NC induction (Hackland et al., 2017). The resulting population, which displays a cranial NC character (Frith et al., 2018), is then further differentiated toward a vagal NC/early ENS progenitor state following exposure to RA in a dose-dependent fashion (Frith et al., 2020). The obtained ENS precursor population is marked by the expression of ASCL1, PHOX2B, SOX10, p75, and CD49d. This approach is quicker than previously published protocols that yield equivalent ENS progenitors after 10 to 15 days of hPSC differentiation (Fattahi et al., 2016;Workman et al., 2017) and crucially the ENS progenitors can colonize the ENS of adult mice following transplantation (Frith et al., 2020). They can also be employed for the subsequent generation of electrophysiologically active enteric neurons in vitro without the requirement for an in vivo physical niche and instructive gutderived signals.

Critical Parameters
Directed differentiation of hPSCs can be very challenging and often results in a great variability in the protocol's yield among independent replicates or no induction at all (Fig.  3). We found that an inefficient vagal NC differentiation is likely to reflect a poor quality hPSC starting population, a non-optimal seeding density, or sometimes could be due to inadequate medium change and the accompanied depletion of medium components essential for the progression of the differentiation. Thus, it is very important that attention should be paid to these critical parameters, as stated below, for a successful differentiation outcome.

Quality of hPSC cultures
Culture of hPSCs should be performed following good laboratory practice guidelines.
Cultures should be monitored for bacterial and fungal contamination and should be routinely examined for mycoplasma. To avoid contamination, antibiotics such as gentamycin or penicillin-streptomycin could be added in the culture medium. Cell cultures should be examined regularly for morphological changes that indicate unwanted differentiation. Spontaneously differentiated cells grown in hPSC cultures impair differentiation efficiency and therefore, their presence should be eliminated by manual scraping. Screening for markers such as NANOG, OCT4, SOX2 using qPCR or surface proteins SSEA4, SSEA3, TRA-1-60, through flow cytometry, or immunostaining is highly recommended to assess the presence of undifferentiated hPSCs (Thomson et al., 1998;Wright & Andrews, 2009). Regular karyotyping of hPSCs should also be employed to determine genetic aberrations that might occur during propagation of stem cell cultures (Baker et al., 2016). We routinely culture hPSCs on Geltrex or vitronectin coated flasks using mTeSR TM 1 or Essential 8 medium and use ReLeSR TM as passaging reagent. Alternative maintenance conditions have not been tested and may affect the differentiation efficiency of the current protocol. We found that passaging cells using EDTAbased dissociation solution yields lower NC efficiency.

Seeding density
Seeding density determines NC efficiency in a cell-line dependent manner. Therefore, it is highly recommended that plating densities for NC induction are optimized for each cell line/end user.

Medium change
Induction and maintained expression of ENS progenitor genes (ASCL1, PHOX2B) depends on retinoic acid addition in NC medium. We found that daily medium replacement with retinoic acid is required to retain high ASCL1, PHOX2B transcript levels.

Troubleshooting
See Table 10 for common problems encountered when performing these protocols and suggested solutions.

Time Considerations
Basic Protocol 1: 6 days. Basic Protocol 2: 16 days or more. (To date, we have kept enteric neurons alive until day 100 but we have not extended culture past this time point). Prepare fresh ENS medium and minimize freeze-thaw cycles of small molecules Monitor/assess differentiation efficiency of previous steps at Day 6 and Day 9/10 Determine optimal sphere resuspension volume prior to plating for enteric neuron differentiation; test different volumes/sphere densities for plating