A novel method for culturing enteric neurons generates neurospheres containing functional myenteric neuronal subtypes

Background The enteric nervous system (ENS) is comprised of neurons, glia, and neural progenitor cells that regulate essential gastrointestinal functions. Advances in high-efficiency enteric neuron culture would facilitate discoveries surrounding ENS regulatory processes, pathophysiology, and therapeutics. New method Development of a simple, robust, one-step method to culture murine enteric neurospheres in a 3D matrix that supports neural growth and differentiation. Results Myenteric plexus cells isolated from the entire length of adult murine small intestine formed ≥3000 neurospheres within 7 days. Matrigel-embedded neurospheres exhibited abundant neural stem and progenitor cells expressing Sox2, Sox10 and Msi1 by day 4. By day 5, neural progenitor cell marker Nestin appeared in the periphery of neurospheres prior to differentiation. Neurospheres produced extensive neurons and neurites, confirmed by Tubulin beta III, PGP9.5, HuD/C, and NeuN immunofluorescence, including neural subtypes Calretinin, ChAT, and nNOS following 8 days of differentiation. Individual neurons within and external to neurospheres generated depolarization induced action potentials which were inhibited in the presence of sodium channel blocker, Tetrodotoxin. Differentiated neurospheres also contained a limited number of glia and endothelial cells. Comparison with existing methods This novel one-step neurosphere growth and differentiation culture system, in 3D format (in the presence of GDNF, EGF, and FGF2), allows for ∼2-fold increase in neurosphere count in the derivation of enteric neurons with measurable action potentials. Conclusion Our method describes a novel, robust 3D culture of electrophysiologically active enteric neurons from adult myenteric neural stem and progenitor cells.


Introduction
The enteric nervous system (ENS) is comprised of a vast neural network in the gastrointestinal tract responsible for a wide array of homeostatic regulatory functions including gastrointestinal secretion, regulation of motility, absorption of nutrients, sensation of stimuli, and blood flow (Lomax et al., 2005;Furness, 2012;Fleming II et al., 2020).Enteric neurons occur in groups of ganglia interconnected by bundles of nerve fibers to form two main ganglionated plexuses, namely the myenteric plexus between the longitudinal and circular muscle layers, and the submucosal plexus under the circular muscle layer (Osorio et al., 2011).The ENS interconnected network is composed of enteric neurons, glia and neural progenitor cells coordinating adaptation to microenvironmental changes (Bondurand et al., 2003;Schäfer et al., 2009;Gulbransen et al., 2012).
In addition to its regulatory effect, the importance of ENS is emphasized by the life-threatening effects of certain enteric neuropathies like congenital Hirschsprung disease, acquired Chagas disease, diabetic neuropathy, opioid-induced bowel dysfunction, postoperative ileus, inflammatory bowel disease or acute appendicitis (Xiong et al., 2000;Villanacci et al., 2008;Bagyanszki et al., 2012;Furness, 2012).Accumulating evidence suggests that debilitating neurodegenerative conditions like Parkinson's disease are associated with neuronal damage not only in the brain but also in the ENS (Liddle et al., 2018;McQuade et al., 2021).Like neurons, the non-neuronal glial cells of the ENS also play an essential role in maintaining the integrity of the gastrointestinal tract as its dysfunction leads to inflammatory neuronal degeneration and changes in neurotransmitter expression (Bush et al., 1998;Cornet et al., 2001;Aube et al., 2006).Enteric glial cells are known to produce immunoregulatory molecules that regulate tissue repair and host defence (Progatzky et al., 2022).Clearly, the ENS is essential for normal gastrointestinal function and its pathophysiology (Spencer et al., 2020).So, a thorough understanding of the ENS is critical for elucidating its normal functional influence and the pathophysiology of the gastrointestinal tract including the side effects from pharmacological agents.Studies of enteric neurons in culture are effective models to study the pathophysiology of these enteric neuropathies.
To date only a very limited number of mouse enteric neuron culture methods are available (Smith et al., 2013;Grundmann et al., 2015;Wahba et al., 2016;Brun et al., 2018;Zhang et al., 2021), however mostly with limited yields.It is known that self-renewing multipotent ENS progenitor cells (ENSPC) are able to generate neurons and glia derived from neural crest stem cells.These ENSPCs can be harvested from the neonatal and adult gut (Kruger et al., 2002;Bondurand et al., 2003).Enteric neural stem cells has been isolated through the generation of neurospheres (Bondurand et al., 2003;Schafer et al., 2003;Almond et al., 2007).Neurospheres are aggregates of neural stem and progenitor cells that differentiate into neurons and glial cells when cultured in vitro.This allows both self-renewal and expansion of precursor cells to form cellular aggregates (Reynolds et al., 1996;Schafer et al., 2003;Almond et al., 2007).
In the present study, we describe the isolation of the myenteric plexus from adult mice to form neurospheres in 3D culture in the presence of EGF, FGF2, and GDNF growth factors.In a single step culture, these neurospheres are subsequently differentiated into neuronal subtypes that are capable of generating action potentials.The differentiated neurospheres predominantly express neurons but also contain limited glia and endothelial cells.

Isolation of longitudinal muscle myenteric plexus
C57BL/6 J mice aged 7-22 weeks were used in accordance with institutional Animal Care and Use Committee protocols.Animal were sacrificed using CO 2 inhalation (4 liters per minute, 2 min) followed by cervical dislocation and the abdomen of the mice were thoroughly sprayed with 70% ethanol.A longitudinal incision was made to open the abdomen.The small intestine from the ileum to duodenum was collected while removing the mesentery and placed in a 10 cm petri dish containing ~10 ml of DPBS with Pen-Strep (1x) on ice.Using gavage needle (size 20), 20 ml of DPBS/Pen-Strep (1x) solution (Supplemental Table ) was used to wash the small bowel lumen from proximal to distal end while pulling the intestine onto the needle, x2.The cleaned small bowel was then placed in a 9 cm black silicone coated dissection petri dish with ~40 ml of cold DPBS/Pen-Strep (wash) solution.
Next, using forceps and scissors open the bowel along the mesenteric line from duodenum to ileum.Then flatten the small intestine entirely using a bent forceps by gently patting the outer wall of the small intestine keeping the lumen side down using the pins to hold the specimen taut under the wash solution.The LMMP was isolated as described previously (Levin et al., 2020).Using a sterile cotton tipped applicator dipped in cold DPBS/Pen-Strep a LMMP window was created at the ileum end by gently rubbing longitudinally under a dissection stereo microscope.After creating a floating LMMP layer, rearrange the pins as needed for counter pressure and use a wet cotton tipped applicator to gently pull the LMMP segment gradually from the entire length of the small intestine.The average time to isolate the whole LMMP from a clean flattened small intestine was about 20-25 min.The LMMP was then washed once in ~20 ml of HBSS containing calcium, magnesium and Pen/Strep, x1 and then cut into 3-5 mm pieces, trim any remaining mesenteric fats using a sterile blade and forceps, then store in HBSS/Pen-Strep on ice until used for enzymatic digestion (Fig. 1A, B).

LMMP immunofluorescence
The LMMP samples were spread with pins on silicone coated black petri dishes in DPBS with Pen/Strep and immunostained at room temperature as described earlier (Levin et al., 2020).Briefly, samples were fixed in 5 ml of 4% PFA in PBS for 30 min, then washed x3 in 10 ml of DPBS and then blocked and permeabilized for 60 min in 5 ml of 5% normal goat serum and 0.1% Triton X-100 in DPBS.The LMMP was then incubated with 5 ml of blocking permeabilization (BPB) buffer containing TUBB3 or mouse IgG2a isotype control antibody at 5 µg/ml for 3 h and then washed x3 in 10 ml DPBS.The samples were then incubated with goat anti-mouse IgG F(ab') 2 -AF488 at 2 µg/ml in BPB for 1 h and washed x3 in 10 ml of DPBS.The samples were further washed in 10 ml of distilled water x2, mounted in Slow Fade diamond antifade DAPI on concave slides, edges sealed using clear acrylic nail polish and imaged using widefield Olympus BX51 fluorescence microscope.
The neurosphere neuron differentiation media is same as the neurosphere growth media but without EGF, FGF2 and Heparin.

Enzymatic digestion of LMMP and 3D neurosphere culture and differentiation
To generate neurospheres, LMMP was digested to produce myenteric plexus as described by Grundmann et. al., (2015) with modifications.Immediately after obtaining the LMMP pieces from one small intestine, the strips were transferred with forceps equally to 2 wells of a 12 well tissue culture plate (not in 1.5 ml tubes, Grundmann et. al., 2015) containing 830 μL of HBSS with calcium, magnesium and Pen/Strep 1x, 150 μL Liberase TH to a final concentration of 0.75 mg/ml, and 20 μL DNase I and allowed the digestion for 4 h at 37 • C in a rocking water bath at 30 rpm.The digested LMMP pieces from each well were then gently washed (without the very low amplitude shaking procedure, Grundmann et. al., 2015) in 5 ml of HBSS with calcium, magnesium and Pen/Strep in a 6 well plate using 1 ml pipet tip for 3 times to remove the digested cells without breaking the myenteric plexus pieces.All LMMP pieces were then pooled into a 50 ml tube containing 2 ml of NSND media, centrifuged at 200xg, 2 min at room temperature (RT).The pellet was mixed aggressively with 100 µl of NSNG using a sterile gel loading tip (bore 0.5 mm) for 20 times and then mixed with 2 ml of NSNG media.The mix was then passed through a 70 µm cell strainer and then again through a 37 µm cell strainer.Spin the filtrate at 300xg, 3 min at RT.The pellet (barely visible) was suspended in 1350 µl of ice cold matrigel and collagen I mix (4:1) by ~20 trituration using 200 µl and 1000 µl pipet tips.A 100 µl mix was spread in 35 mm MatTek glass bottom dishes with 14 mm micro well.A total of 12 dishes were plated from one small intestine LMMP samples.The dishes were warmed at 37 • C for 11 min and incubated with 3 ml of NSNG media at 37 • C and 5 % CO 2 .The media was changed every 48 or 72 h.To initiate differentiation of the neurospheres, the NSNG media was replaced with NSND media after 5-7 days in culture.Neurospheres count, growth, and differentiation were imaged using EVOS M7000 microscope (ThermoFisher Scientific).

Propidium iodide staining of 4 h digested LMMP
After 4 h digestion and 3 washes in HBSS, only a few LMMP samples were placed in 5 ml of HBSS with Ca ++ , Mg ++ and Pen/Strep containing 2 µg/ml of propidium iodide and incubated at 37 • C for 10 min (modified after Grundmann et al., 2015, 50 ng/ml;Malik et al., 2015, 20 µg/ml;Levin et al., 2020, 1.0 µg/ml).The treated LMMPs were then washed in 5 ml of HBSS for 3 times, then immediately fixed in 4% PFA in PBS at RT for 10 min.The fixed samples were then permeabilized in 0.5% Triton X-100 in 5 ml HBSS for 15 min at 37 • C, then washed in 5 ml HBSS for 3 times, then mounted in concave slides containing Slow Fade DAPI, and sealed with clear nail polish.For total cells, 0 h digested LMMP samples were permeabilized, washed, fixed, washed, stain with propidium iodide, washed, mounted and sealed as above.Confocal z-stack images were taken in Zeiss LSM 880 microscope.

Immunofluorescence of neurospheres, differentiated cells and quantitation of fluorescence area
Neurospheres and differentiated cells embedded in the matrigelcollagen matrix were fixed in 3 ml of 4% PFA in PBS immediately after removing the culture media for 30 min at 37 • C, then washed with 5 ml of warm DPBS for 3 times at RT.The samples were then blocked and permeabilized with 2.5 ml of 5% normal goat serum and 0.2% Triton X-100 (blocking buffer) for 60 min in DPBS at room temperature.The permeabilized cells were then incubated in primary (in mouse or rabbit) or isotype control antibody at 5 µg/ml in 2.5 ml of blocking buffer, for overnight at RT.The cells were then washed in 5 ml of DPBS at 37 • C DPBS for 3 times followed by incubation with the secondary antibody (goat anti-mouse or anti-rabbit IgG F(ab) 2 AF488/AF594) at 2 µg/ml in the blocking buffer for 60 min at RT.The cells were then washed in 5 ml of warm DPBS for 3 times, covered with few drops of Slow Fade DAPI, the glass bottom coverslip was then mounted on a concave slide, removed excess DAPI, and sealed with clear acrylic nail polish.Confocal z-stack images were taken in Zeiss LSM 880 microscope with AiryScan.To quantify the total fluorescence area, the maximum intensity projection fluorescence images were analyzed using a hybrid cell count software with BZ-H4C analytic application (Seki et al., (2021); https://www.keyence.com/products/microscope/fluorescencemicroscope/bz-x700/models/bz-h4c/.

Statistical analysis
Statistical tests were performed using GraphPad Prism software, Version 10.0.2.Data are presented as the means ± standard error of means (SEM).Differences between the 2 groups were examined using a unpaired t test (two-tailed), with p < 0.05 considered statistically significant.N = number of observations.

LMMP identification, digestion, and neurosphere culture and differentiation
To maximize neurosphere yield, LMMP was isolated from entire small intestine by dissection in silicone dishes under stereo microscope.The presence of myenteric plexus in the LMMP samples from duodenum, jejunum and ileum were confirmed using neuron specific anti-TUBB3 antibody.Neural processes, cell bodies and ganglia were apparent throughout the small intestine (Fig. 2).
To determine the viability of the cells in the myenteric plexus after 4 h of enzymatic digestion, live-dead staining using propidium iodide was used.Patches of live myenteric plexus cells were apparent following the enzymatic digestion along with occasional long nuclei smooth muscle cells (arrow head) compared to the undigested LMMP (Fig. 3A,  B).The LMMP after Liberase treatment revealed cleared areas of long nuclei smooth muscle cells and occasional presence of dead cells (≤0.3%, Fig. 3B, arrow).
The dissociated adult myenteric plexus cells from one small intestine were suspended in the matrigel-collagen mix, cultured and differentiated in 3D embedded format in twelve 14 mm microwell of 35 mm Petri dishes.Small neurospheres (~30 µm) formed within 2-3 days and becomes ~150 µm sizes by 6-7 days in culture (Fig. 4 A, B) and remained undifferentiated.Total number of neurospheres per small intestine after 5-6 days in culture was 3284 (273.7 ± 46.5 per dish, mean ± SEM, n=6).In differentiation media, the neurospheres produce extensive neurites and formed neurite network (Fig. 4C) without any subculture.

Expression of neural stem and progenitor cell markers in early neurospheres
Sox2 and Sox10 are well characterized markers of neural stem and progenitor cells and are also required for their self-renewal (Amador-- Arjona et al., 2015).(Kim et al., 2003;Pozniak et al., 2010).Both Sox2 and Sox10 were found to be highly expressed predominantly in the nuclei of the early neurospheres (Fig. 5A, B).
Additionally Msi1, the evolutionarily conserved pluripotent marker for neural stem and progenitor cells (Kanemura et al., 2001;Glazer et al., 2008) was also expressed in the growing neurospheres (Fig. 5C).These progenitor cells were distributed throughout the neurosphere body confirming the presence of these pluripotent cells in the early neurospheres.Active proliferation of neurosphere cells even at day 7 of growth was marked by presence of nuclear Ki67 expression (Fig. 5D).
Nestin, an intermediate filament protein, is a marker of multipotent neural progenitor cells and is required for their self-renewal while persisting in mature neurons (Park et al., 2010;Hendrickson et al., 2011).In early undifferentiated neurospheres nestin expression was found to be restricted to the peripheral cells and occasionally in the fibers extending out from the periphery (Fig. 6A, B).Following 5 days in differentiation media the neurosphere showed dispersed nuclei with extensively outgrowing nestin positive neurites (Fig. 6C).The survival of the neuronal progenitors and their differentiation, migration and axonal outgrowth are also regulated the p75 neurotrophin receptor (p75 NTR , Young et al., 2007;Meier et al., 2019).Examination of p75 NTR in late neurospheres confirmed its expression in the cell bodies, neurites and in the early neurons (Fig. 6D).

Neurosphere differentiation into neurons and neuronal subtypes
After 5-7 days in growth media, neurospheres developed in the matrigel-collagen matrix were grown in differentiation media for 5-8 days.Neurites were observed within 2-3 days.The differentiation of neurons in the neurospheres was confirmed by expression of neuronspecific protein Tubulin beta III specifically along the neurites (Fig. 7A or 29 days in vitro, data not shown) and the pan-neuronal marker PGP9.5 in the cell bodies and neuronal processes (Fig. 7B).The neuronal processes formed mesh-like structures around the differentiated neurospheres.Another pan-neural RNA binding protein HuD/C that does not localize to the neural processes (Desmet et al., 2014) showed cytoplasmic and nuclear localization in the differentiated neurospheres (Fig. 7C).NeuN, a neuronal specific RNA binding protein (Mullen, 1992) also showed nuclear and cytoplasmic expression following 5 days in differentiation media (Fig. 7D).Many of the HuD/C and NeuN positive cells in the differentiated neurospheres generated neurites detected by alpha Tubulin antibody (Fig. 7C, D) which colocalized with neuron-specific Tubulin beta III positive neurites (data not shown).
In addition to the differentiation of neurons in the neurosphere, the Ca 2+ -binding protein Calretinin, known to be predominantly expressed in specific neurons of the central and peripheral nervous system (Schwaller, 2014), was also detected in the cell bodies in differentiated neurospheres (Fig. 8A).Putative cholinergic neuronal marker choline acetyltransferase (ChAT) that produces acetylcholine (Granger et al., 2020) was expressed both in the neurites and cell bodies of neurospheres differentiated for 8 days (Fig. 8B).Neuronal nitric oxide synthase (nNOS), which generates nitric oxide (Echagarruga et al., 2020), was expressed both in the cell bodies and processes of neurons in differentiated neurospheres (Fig. 8C).The isotype control antibodies produced no signal (data not shown).

Whole-cell patch clamp determination of action potentials in differentiated neurospheres and neurons
We performed whole-cell patch-clamp recordings to investigate the ability of the neurons in the neurospheres to generate action potentials (Fig. 9A), as well as those that migrate from the neurosphere (Fig. 9D) after 1-2 weeks in differentiation media.Notably, the presence of matrigel-collagen matrix affected our ability to readily approach the cell surface and form a gigaohm seal on the cell membrane, but in the successful attempts, we made a gigaohm seal (1-5 GΩ) and subsequently recorded neuronal activity.Cells present in the neurosphere (Fig. 9A) or away from the neurosphere (Fig. 9D) showed a wide range of resting membrane potential (-30 to − 55 mV) similar to the typical neurons as reported in the literature (Bean, 2007;Tripathy et al., 2014), and showed no spontaneous firing of action potentials.To assess whether these cells are capable of generating action potentials, we employed brief pulses of depolarizing currents and recorded membrane potential.In both instances, neurons inside and outside of the neurosphere showed a linear passive increase in the membrane potential (Fig. 9B, E, small gray traces) until the membrane reached the firing threshold (Fig. 9B, E, black traces showing before and after threshold) and generated action potentials with each subsequent increment in the magnitude of injected current pulse (Fig. 9B, E, large gray traces).In addition, we recorded the membrane voltage in response to multiple long-step currents of hyperand de-polarization polarity to confirm action potential generation.Long-step current injections reliably generated a few action potentials in a few neurons (Fig. 9C) and only a single in others (Fig. 9F) suggesting that neurons are able to generate action potentials.
Finally, to confirm whether the action potential generated upon a depolarization pulse or prolonged step current is mediated by voltagegated sodium channels, we applied 2 µm Tetrodotoxin (TTX) in the bath after recording action potential from a neuron (Fig. 10A, C).Within 5 min of TTX addition, the action potential generation was completely abolished (Fig. 10B, D).Notably the same neurons showed action potential before TTX addition (Fig. 10A, C), suggesting that the action potential is mediated by voltage-gated sodium channels identical to that of a typical neuron.These data suggest that neurons within and outside of the neurosphere are able to fire action potentials and these are mediated by voltage-gated sodium channels.

Non-neuronal cell types in neurospheres
Glial cells are central components of neurogenic niches in the embryonic and adult central nervous system (Falk and Götz, 2017).Glial cells also orchestrate many important aspects of nervous system formation and function (Allen and Lyons, 2018).Examination of early  undifferentiated neurospheres showed sporadic expression of GFAP, a glia-specific protein (Fig. 11A).Differentiated neurospheres with extensive neurite networks revealed only a limited occurrence of GFAP positive glial cells (Fig. 11B) following 7 days in differentiation media.The Glia derived GFAP mean fluorescence area was 6080 ± 1402 µm 2 compared to 52132 ± 4765 µm 2 for TUBB3 positive neurons in the differentiated neurospheres representing about 10.4% of the total neuron-glia fluorescence area (P < 0.001, n = 3).
Endothelial cell released factors are known to enhance neural progenitor cell proliferation and differentiation (Shen et al., 2004;Gama Sosa et al., 2007).Vascular endothelial cells also promote neurite outgrowth, enhanced synapse function and accelerated electrophysiological development of neurons (Wu et al., 2017;Grasman et al., 2017).Vascular endothelial cadherin expression confirmed the development of endothelial cells mainly in the core of neurospheres before differentiation while immature neurons expressing TUBB3 were mostly peripheral (Fig. 11C).Only few endothelial cells were detected in the neurospheres following 3 days in differentiation media(Fig.11D).The vascular endothelia derived cadherin mean fluorescence area was 593.5 ± 13.5 µm 2 compared to 46,413 ± 8785 µm 2 for the TUBB3 positive differentiated neurons representing about 1.3 % of the total fluorescence area (P < 0.035, n = 2).

Discussion
The objective of this study was the development of a novel murine culture system for characterization of enteric neurons, neurochemically defined sybtypes and their functional properties using immunofluorescence assays and electrophysiological studies following differentiation of adult myenteric neural stem/progenitor cells.Here, we describe the development of an enteric neurosphere culture in 3D format from the ganglionated myenteric plexus of adult mice that differentiates predominantly into functionally active neuronal subtypes along with limited glial and endothelial cells.The isolation of enteric nervous system cells can be challenging as the myenteric plexus is located between the longitudinal and circular smooth muscle cells.To initiate the neurosphere culture, ganglionated longitudinal muscle myenteric plexus was isolated from the small intestine of adult mice (Fig. 1) under DPBS to prevent drying of the LMMP samples during the isolation process.The average time to isolate the entire LMMP from one longitudinally open small intestine was about 20-25 min and resulted in no bacterial growth in the cultures.The starting LMMP samples were initially validated to have ganglionated myenteric plexus from duodenum, jejunum and ileum using neuron specific TUBB3 immunofluorescence analyses (Fig. 2).Usually, LMMP samples are digested first by collagenase followed by trypsin (Smith et al., 2013;Wahba et al., 2016;Zhang et al., 2021).In this study, to remove the longitudinal muscle cells from myenteric plexus, the LMMP pieces were digested in a one-step process using a cocktail of Collagenase I, Collagenase II, a neutral protease, Thermolysin (Liberase TH) and DNase I for 4 h to improve the abundance of intrinsic neural stem cells (Grundmann et al., 2015) as the myenteric ganglia do not utilize collagen but glia cells for tissue adhesion (Gershon et al., 1991).The prolonged Liberase TH digestion of LMMP samples produced only occasional dead cells (Fig. 3) along with cleared areas of muscle cells, as observed by Grundmann et al., (2015) using scanning electron microscopy and western blot analysis.Enteric neurons are terminally differentiated cells.Culture of primary neurons usually results in relatively lower yields of enteric neurons (Smith et al., 2013;Wahba et al., 2016;Brun et al., 2018), however culture and differentiation of enteric neural stem/progenitor cells produces large number of enteric neurons (Zhang et al., 2021).Usually enteric neural stem/progenitor cells are cultured in matrix uncoated plates for neurosphere growth and then separately cultured for neurosphere differentiation to produce neurons in matrix coated format (Almond et al., 2007;Binder et al., 2015;Grundmann et al., 2015;Hotta et al., 2016;Zhang et al., 2021).To improve the yield of neurospheres, we cultured the enteric neural stem/progenitor cells in 3D matrix for both growth and differentiation steps in single dish containing matrigel and collagen matrix which has shown to improve cell attachment (Wahba et al., 2016).The dissociated myenteric plexus cells from one small intestine was enough for twelve 14 mm microwell of 35 mm MatTek dish cultures and improved the total neurosphere count by 1.6-2.2fold compared to the standard matrix non-coated culture methods (Zhang et al., 2013;Zhang et al., 2021).The higher yield of neurospheres could be attributed due to the single step Liberase TH digestion which produces hundreds of myenteric networks (Grundmann et al., 2015) while avoiding excessive cell death (Fig. 3), elimination of possible loss of stem/progenitor cells from the 3D matrix during media change, and addition of GDNF in the growth media because of its essential role in the early survival and proliferation of enteric neural crest cells in the small bowel and colon (Gianino et al., 2003;Hao et al., 2009).
Within the Matrigel, formation of neurospheres were apparent within ~3 days of culture and remained undifferentiated until induced for differentiation by 6-7 days (Fig. 4).Analysis of the cellular nature of early undifferentiated neurospheres confirmed the expression of wellestablished neural stem and progenitor cell markers, Sox2 and Msi1 throughout the neurosphere (Fig. 5).Beyond self-renewal properties, Sox2 also plays a role in proper activation of neural differentiation (Amador-Arjona et al., 2015) while Msi1 modulates progenitor cell expansion (Kanemura et al., 2001;Wang et al., 2008).Also, the expression of Sox10 in early neurospheres underscored the potential of glia cell formation during neurosphere differentiation (Fig. 11) as this transcription factor is the key regulator of glial cell development (Britsch et al., 2001;Liu et al., 2020).On the contrary Nestin, a cytoskeletal intermediate filament protein of neural progenitor cells (Park et al., 2010;Bernal et al., 2018) was barely expressed in day 4 neurospheres but found to be highly distributed in the neurites of differentiated neurons (Fig. 6A-C) as observed in mature neurons of adult rat and human brain (Hendrickson et al., 2011).To overcome apoptosis-mediated neuronal losses in myenteric ganglia, new neurons are formed from the dividing precursor cells which express both nestin and p75 NTR (Kulkarni et al., 2017).Similar to nestin expression profile in growing neurospheres, both small and large neurospheres were found to be neurogenic (Young et al., 2007) as confirmed by expression of p75 NTR (Fig. 6D).
Neuronal differentiation of neurospheres was evident from expression of neuron specific markers Tubulin beta III, PGP9.5, HuD/C and NeuN proteins within 3-7 days of culture in differentiation media (Fig. 7).In LMMP dissociated 2D cultures, neuronal morphology and immunocytochemical features also became apparent after 7-10 days in culture (Smith et al., 2013;Wahba et al., 2016).Some of the major myenteric neural subtypes e.g., ChAT, Calretinin and nNOS first arrive in the myenteric plexus by embryonic day E13.5 or earlier (Bergner et al., 2014;Erickson et al., 2014).Interestingly, all these myenteric neuron subtypes appeared in the differentiated neurosphere following 8 days in culture (Fig. 8) as observed in dissociated neurosphere subtype differentiation in vitro (Binder et al., 2015;Hotta et al., 2016;Cheng et al., 2017).In matrix coated dissociated neuron culture format, expression of calretinin was reported after about 10 days in culture (Smith et al., 2013) when ChAT was observed around 21 days in vitro (Wahba et al., 2016).Enteric neurospheres are usually grown in matrix coated or non-coated dishes containing EGF and bFGF, and for cell differentiation on matrix coated plates in absence of growth factors (Almond et al., 2007;Metzger et al., 2009;Binder et al., 2015;Grundmann et al., 2015;Hotta et al., 2016;Zhang et al., 2021).The present study demonstrated that the myenteric neural stem/progenitor cells can be cultured to form neurospheres in matrix embedded format and can also be differentiated to form neurons and myenteric neural subtypes in the same dish (Figs. 7 and 8).This method also incorporated GDNF in both the growth and differentiation media for neurogenesis and differentiation because this factor promotes the proliferation, migration and neural differentiation of enteric neural crest cells (Almond et al., 2007;Hao et al., 2009;Uesaka et al., 2013;Cheng et al., 2016;McKeown et al., 2017).
The differentiated neurospheres and neurons showed the defining feature of firing action potentials and revealed a wide range of resting membrane potential (-30 to − 55mv), suggesting a heterogeneous population likely to be at different stages of maturity (Ramoa et al., 1994).At their resting membrane potential none of the cells showed any spontaneous action potential which is very common in in vitro systems  due to the absence of neuronal network connectivity.However, when we stimulated neurons using a current-clamp approach, we observed passive changes in membrane potential with subthreshold stimuli and consistent action potential firing upon reaching the threshold stimulation and subsequently higher stimuli.The membrane voltage responses following stimulation by longer hyper and depolarization currents also generated few or single action potential spike patterns which suggest a heterogeneous population likely to comprise different subtypes of neurons as seen in our immunocytochemistry results.Similar to the action potentials of the CNS neurons which are mediated by tetrodotoxin-sensitive voltage-gated sodium channels, we observed a complete blockade of action potential in the patched neurons within 5 min of tetrodotoxin application (Figs. 9 and 10), confirming their physiological similarities with CNS neurons.
In addition to neurons, differentiated neurospheres contained differentiated glial cells representing only ~10% of the total neuron-glia specific protein fluorescence (Fig. 11).The enteric neural stem cells derived from neural crest cells (Dupin et al., 2012) are present in the neurospheres and differentiate into neuron and glia cells in vivo (Hotta et al., 2013;Burns et al., 2014).Both FACS selected or unselected neural crest derived cells form neurospheres with abundant glia cells before and after differentiation (Binder et al., 2015;Cheng et al., 2017).Like peripheral glia cells, Sox10 is required for the development of enteric glia cells (Britsch et al., 2001;Paratore et al., 2001).Even in the presence of a non-limiting Sox10 expression in early neurospheres, the limited glia cell expression could be attributed to the presence of GDNF during both proliferation and differentiation of neurospheres as GDNF availability determines enteric neuron numbers by controlling ENS precursor proliferation (Gianino et al., 2003;McKeown et al., 2017;Cheng et al., 2016).In addition to glia cells, the presence of endothelial cells in the neurospheres was found to be very interesting as endothelial cells promote neural stem cell proliferation, differentiation and functional maturation of neuron through activation of VEGF signaling (Sun et al., 2010;Wu et al., 2017).It would be intriguing through future clonal analysis to determine whether the endothelial cells differentiate from a common multipotent stem cell or from endothelial progenitor cells (Jing et al., 2022).The use of ENS culture in mice models to study the function of ENS in health and disease has been instrumental to advance our present understanding of ENS physiology and disease development (Schonkeren et al., 2022).Now with the advent of this functional neurosphere culture, intestinal organoids or enteric mesenchymal cells can be cocultured to study cell-cell molecular interaction pathways which can be difficult to analyze in vivo.The availability to manipulate the developmental components of this functional ENS culture system presents an advantage to study the ENS in health and disease.

Conclusion
The method we describe involves the isolation of LMMP from flattened small intestine and its digestion and growth in conditions that enhance neurosphere production.This neurosphere culture method is a single step procedure for growing adult myenteric ganglia in a 3D format for neural growth and differentiation.The developed neurospheres produce myenteric neuronal subtypes and functionally active neurons within 8 days of differentiation and can be cultured for at least 4 weeks.The differentiated neurospheres also contain a limited number of enteric glia and endothelial cells.

Fig. 2 .
Fig. 2. Immunofluorescence image of LMMP showing the extent of neural network in small intestine.The neural network was detected using neuron-specific anti-TUBB3 antibody (green) showing the distribution of neural processes running between the ganglia (arrows) from duodenum (A), jejunum (B) and ileum (C).The nuclei were stained with DAPI (blue).Scale bars, 100 µm (A, B) or 200 µm (C).

Fig. 3 .
Fig. 3. Live dead staining of LMMP samples using Propidium iodide.A: LMMP fixed, permeabilized and stained before 4 h digestion showing abundant long nuclei smooth muscle cells and some round nuclei cells; B: 4 h digested LMMP stained with Propidium iodide (red) showing mostly live round nuclei cells with occasional dead cells (arrow), cleared areas of muscle cells with few long nuclei smooth muscle cells (arrow head).A, B: merge of propidium iodide and DAPI (blue).Scale bar: 100 µm.

Fig. 4 .
Fig. 4. Phase contrast images of cultured enteric neurospheres from adult myenteric plexus in 3D matrigel-collagen matrix.A: Early neurosphere by day 3 of culture; B: Developed and undifferentiated neurosphere culture by day 7; C: Differentiated neurosphere sent out profuse neurites into the matrix and formed neurite mesh by day 10 of culture.Scale bars: 150 µm.

Fig. 6 .
Fig. 6.Expression of neural progenitor cell markers in early and differentiated neurospheres.A: Occasional expression of Nestin (red) in some peripheral cells, in the filamental area, day 4. B: Most of the peripheral cells showed expression of Nestin by day 5. C: Differentiated neurospheres with dispersed nuclei and nestin positive neurites, day 11.D: Neural progenitor cell marker p75 NTR (red) expressed in the dispersed neurons and in the neurospheres, at day 10 of growth media.Counter stain DAPI (nucleus), blue and cytoplasmic alpha tubulin, green.A, B, C: merge of blue, green and red.Scale bars: 50 µm (A), 100 µm (B, C) or 200 µm (D).All neurospheres were positive to the labelling (Nestin, n = 4; p75 NTR , n = 5).

Fig. 8 .
Fig. 8. Neurospheres differentiate into neural subtypes following 8 days in differentiation media.A: Calcium binding calretinin expressed primarily in the cell bodies (red).B & C: Both ChAT (B, green) and nNOS (C, red) showed expression in cell bodies and neurites.Nuclei are stained with DAPI in blue.Scale bars: 100 µm.IgG control antibodies produced no signals (data not shown).All neurospheres were positive to the labelling (Calretinin, nNOS, n = 3; ChAT, n = 4).

Fig. 9 .
Fig. 9. Light micrograph of a single cell in a neurosphere (A) or a differentiated neuron (D) with a patch pipette recording the action potential firing ability of a single cell.B & E: Representative traces of membrane voltage in response to the short depolarization current pulses (B: 1-100 pA or E: 1-200 pA for 2 ms with 10 pA increment each time) from a patched cell within a neurosphere (A) or a differentiated neuron (D).Black traces represent membrane responses to two consecutive stimuli before (shorter peak) and after (larger peak) firing the first action potential.Gray traces below the short peak black trace show the passive response of the membrane to subthreshold stimuli.Gray traces with a large peak represent the fringe of action potential each time with a higher magnitude current than the threshold current.C & F: Representative traces of membrane voltage in response to longer hyper and depolarization current steps (C: Steps − 40 -70pA, F: − 50 -70 pA) with 10pA increment in subsequent steps for 1100 ms from a patched cell in A or D. Black traces represent membrane responses to two consecutive stimuli before (shorter peak) and after (larger peak) firing the first action potential.

Funding
C.M. was supported by NIH Research Training Program (NIH 5T32HL007849-22).D.E.L. was supported by the iTHRIV Scholars Program which is funded in part by the Natioal Center of Advancing Translational Sciences of the National Institute of Health under award numbers UL1TR003015 and KL2TR003016.D.E.L. was also supported by the a grant from the Departmet of Surgery, University of Virginia.S. R/ M. was supported by a grant from Bill & Melinda Gates Foundation

Fig. 10 .
Fig. 10.Representative traces of membrane voltage in response to the short pulse stimulation (A, B; 1-100 pA for 2 ms with 10 pA increment) or long step stimulation (C, D; − 40 -70 pA with 10 pA increment for 1100 ms) depolarization currents.Black traces represent membrane responses to two consecutive stimuli before (shorter peak) and after (larger peak) firing the first action potential.Gray traces below short peak black trace show passive response of the membrane to subthreshold stimuli.Gray traces with a large peak represent the fringe of action potential each time with a higher magnitude current than the threshold current.B, D: Representative traces of membrane voltage after incubation of cells in 2 µm Tetradotoxin for 5 min.The two black traces represent the two consecutive stimuli responses as seen in B or D, however without any action potential generation.None of the higher magnitude currents than the threshold current showed any action potential.